NICaS Study Review 2014

Transcription

Non Invasive Cardiac system (NICaS)
Whole Body Electrical Bio Impedance
Bio Impedance Technology
Study Review
January 2014
Table of Content
No.
Article
Page
Subject
1 Cardiac Output Monitoring - Reviewing the Evidence on Four
NICaS
4
Systems, Health DEVICES Vol. 42 No. 6 June 2013
Validation
2 Yu Taniguchi et al. Noninvasive and Simple Assessmentof
NICaS
15
Caridac Output and Pulmonary Vascular Resistance With
Validation
Whole-Body Impedance Cardiography Is Useful for Monitoring
Patients With Pulmonary Hypertension, Circulation Journal, June
12,2012
3 SG Hall, J Garcia, DF Larson and R Smith. Cardiac power index:
Cardiac Power 22
staging heart failure for mechanical circulatory support,
Index Validation
Perfusion 2012 27:456-459
4 JGF Cleland et al. Non invasive measurements of Cardiac Output
NICaS
29
9CO) and Cardiac Power Index (CPI) by whole body bio
Validation
impedance in patients with heart failure. A report from SICA-HF
study 9FP7/2007-2013/241558). European Journal of Heart
Failure Supplements (2012) 11
5 Yoseph Rozenman et al, Detection of left ventricular systolic
GGI Validation 30
dysfunction using a newly developed, laptop based, impedance
cardiography index, International Journal of Cardiology Volume
149, Issue 2: 248-250, 2 June 2011
6 Robert G. Turcott et al,Measurement Presision in the
ICD for CRT
33
Optimization of Cardiac Resynchronization Therapy, Circulation Optimization
Heart Failure 2010;3:395-404, February 22, 2010
Validation
7 Yusuke Tanino et al. Whole Body Bioimpedance Monitoring for
NICaS
44
Outpatient Chronic Heart Failure Follow up, circulation Journal
Validation
73:1074-1079, June 2009
8 Konstantin M. Heinroth et al, Impedance Cardiography: a useful
ICD for CRT
50
and reliable tool in optimization of cardiac resynchronization
Optimization
devices. Europace (2007) 9, 744-750, 11 May 2007
Validation
9 Dorinna D. Mendoza et al. Cardiac Power Output predicts
Cardiac Power 57
mortality across a broad spectrum of patients with acute
Index Validation
cardiac disease. American heart Journal 153:366-370, March
2007
10 Oscar Luis Paredes et al. Impedance Cardiography for Cardiac
NICaS
62
Output Estimation - Riliable of Wrist-to-Ankle Electrical
Validation
Configuration -. Circulation Journal 2007; 70: 1164-1168,
September 2006
11 G Cotter, A Schachner, L Sasson, H Dekel and Y Moshkovitz.
NICaS
67
Impedance cardiography revisited. Physiological Measurement
Validation
27 (2006) 817-827, July 2006
12 Ronald D. Smith et al. Value of Noninvasive Hemodynamics to
ICD for
79
Achieve Blood Pressure Control in Hypertension Subjects.
Hypertension
Hypertension 2006;47:771-777, Mar 6, 2006
Validation
2
Table of Content, Cont’d
Classification Page
No.
Article
13 Marina Leitman et al. Non-invasive measurement of cardiac
NICaS
86
output by Whole-body bio-impedance during dobutamin stress
Validation
echocardiography: Clinical implementations in patients with left
ventricular dysfunction and ischemia. The European Journal of
Heart Failure 8 (2006) 136-140
14 W. Frank Peacock et al. Impact of Impedance Cardiography on
ICD Validation
Diagnosis and Therapy of Emergent Dyspnea: The ED-IMPACT
for ER
Trial. Academic Emergency Medicine Vol. 13, No. 4:365-371, April
2006
15 Hector O. Ventura, Sandra J. Taler and John E. Strobeck.
ICD Validation
Hypertension as a Hemodynamic Disease: The Role of
for
Impedance Cardiography in Diagnostic, Prognostic and
Hypertension
Therapeutic Decision Making. American Journal of Hypertension
Vol. 18, No. 2, Part 2:26S-43S, February 2005
91
16 Guillermo Torre-Amiot et al. Whole-Body Electrical BioImpedance is accurate in Noninvasive Determination of Cardiac
Output: A Thermodilution controlled, Prospective, Double
Blinded Evaluation. European Journal of Heart Failure, June 2004
NICaS
Validation
116
17 Gad Cotter, Yaron Moshkovitz, Edo Kaluski, Amram J. Cohen,
Hilton Miller, Daniel Goor and Zvi Vered. Accurate, Noninvasive
Continuous Monitoring of Cardiac Output by Whole-Body
Electrical bioimpedance. CHEST 2004;125;1431-1440
NICaS
Validation
117
18
98
Gad Cotter et al. The role of cardiac power and systemic vascular Cardiac Power
resistancein the pathophysiology and diagnosis of patients with & Peripheral
acute congestive heart failure. The European journal of Heart
Resistance
Failure 5 (2003) 443-451
Validation
129
19 John E. Strobeck et al. Beyond the Four Quadrants: The Critical
ICD Validation
and Emergong Role of Impedance Cardiography in Heart Failure.
for CHF
Editorial, April 2004 Supplement 2
138
20 Amram J. Cohen. Non-invasive measurement of cardiac output
during coronary artery bypass grafting, European Journal of
Cardio-thoracic Surgery 14 (1998):64-69
NICaS
Validation
144
21 C. Gresham Bayne. Evaluation and management of the elderly
homebound patient with congestive heart failure
ICD Validation
for CHF
150
3
4 of 158.
)
t'k*ffiam
GC WITH THE FLOW
THERMODILUTION USING A PULMONARY ARTERY CATHETER IS CONSIDERED THE GOLD STANDARD
IN MEASURING CARDIAC OUTPUT. HOWEVER. DRAWBACKS ASSOCIATED WITH THE TECHNIQUE HAVE
PROMPTED CLINICIANS TO SEEK LESS INVASIVE OPTIONS. BUT THESE ALTERNATIVES HAVE THEIR OWN
LIMITATIONS. WE EXAMINE THE EVIDENCE ON FOUR DEVICES THAT MEASURE CARDIAC OUTPUT USING
AND NON INVASIVE TECH N IQU ES.
MI N IMALLY INVASIVE
\Ieasuring a patient's cardiac output-the amount
of blood pumped bv the heart per minute-can aid
in diagnosing diseases of the cardiovascular system
and in managing high-risk patients (e.g., those suffering from burns, sepsis, or shock) and patients
undergoing maior surgicai procedures. The cardiac
output value can be derived lrom heart rate and
stroke volume (the amount of blood pumped br-
regarding their accuracv compared to other cardiac
output monitodng techniques, particuiadr thermodilution. \Y'e also examine evidence regarding
improved clinical outcomes with use of these
derrices.
)
the heart per beat) using the relationship
cordioc oulpul = stroke volume
x
heort roie
The uaditional approach to monitoring cardiac output is pulmonarv arterv catheterization
using thermodilution as the measurement method.
.Florvever, the draw-backs associated with this technique-including the risk of pulmonarr arterv
ruprure aod air emboli5m-h21.g led clinicians ro
seek less invasive alternatives.
In our December 2009 issue,
rve looked at some
rninimallf invasive and noninvasive alternatives to
pulmonarl' arterv catheterization for cardiac output
monitoring and revierved the evidence regarding
dreir effectiveoe ss. This article setres as ao update to
that piece. We profiie four additional cardiac output
monitoring svsterns and rer.'ierv the evidence
.l78
xmrraDEvlcEs JuNE2ot3
5 of 158.
) w.*ri.ors
02013
\\:e look at the following s,rstems:
The literarure s'e revierved did
not show the ECOII to be ao acceptable alternative to the reference techniques; horvercr,
nvo studies found that the monitor rvas useful
in predicting fluid responsir.eness in abdominal
ConMed Ecollt.
and cardiac surger\r patients. \I'hen the monitor
is used in the ()R
electrosurgical equip-
"vith
F
ment, interruptions in monitoring could occur.
\\:e did flot 6od any sfudies addressing clinicai
outcomes.
Edwords Llfesciences FloTroc sensor. Previousl,v
reviewed literarure showed the FloTrac sensor to
be comparable to thermodilution in stable cardiac surgerv patients. Srudies on the latest generation of the FIoTrac sensor, horvever, [ound
that the device may be less reliable than thermo-
dilution in hemodynamicallv unstable patients
such as those undergoing abdominal surgen'
ECRI lnsiitute. Member hospitols moy reproduce this poge
for internol distribution only.
Reuiewing the
Four
Cardiac Oatput
Monitoring Slstems
Euideruce on
)
)
or iiver transplantation, those on vasopressor therapt', or those with sepsis. Tivo srudies
reported improved clinical outcomes (reduced
complications and hospital length of star) rvhen
the FloTrac sensor was used to guide fluid man
agement in abdominal surgerv patients.
Edwords Llfesciences VolumeView sensor.
The
YolumeYierv sensor is relativelv nerv; therefore,
the evidence on it is limited. \Iore independent srudies on raried patient populations are
required to determine its ef6cacr'. \\"e did not
6nd anr studies ad&essing clinical outcomes.
Nl Medicql NlCoS. SLx srudies examined in our
Literarure review found that the NICaS is an
acceptable aiternative to the reference techniques for cardiac output measurements in cardiac patients and heart lailure patieots. Tvuo of
those studies found it to be more accurate than
thoracic impedance cardiographr. \V'e did not
6nd anr- srudies addressing cl-inical outcomes.
()f
these s)'stems, only \I \Iedicalt \ICaS is
noninvasive; the remaining three svstems are coosidered minimallr- invasive.
r:'i
lli ii.ji{} i.iri\:iY n:,i.'r ii:r"l' ::'i'r,iiiiVlf;}1}i i,tii'ii.}:Li
Tiaditionally, cardiac output is measured using a
technique knorvn as pulmonar,v arterl.- thermodilurion, *,hich incorporates the use of a pulmonarv
arteq'catheter, or PlC. PuimonarY atterv thermodilution is considered the gold standard in cardiac
output monitoriag. There are two basic versions oF
this technique: bolus and continuous. In bolus, or
intermittent, thermodilution, a bolus of an iniectate
at a knorvn temperature is infused into the bloodstream, and the change in temperarure caused bv
the iniectate is measured at some dorvnstream location using temperature sensors incorporated into a
PlC. ()ne dras'back to bolus thermodilution is that
it does not provide
of
UMDNS ierm- Monilors, Bedside, Cordioc Output
I2O-1741
O2O'13 ECRI lnstitute. Member hospitols moy reproduce this poge for internol distribution only.
6 of 158.
a cootiouous measurement
w.cri.org
>
HEALTHDEVICES
JUNE2ots 179
/
guidance
cardrac output and therefore cannot detect
rapid and sudden changes in the hemodr-namic starus oi high-risk pauents.
Continuous rhermodilution inroh'es
measuring the change in the temperarure
oi blood at a dorvnstream location after
introducing a heating Elament upsream.
Ti-picallr', the heating filament is placed in
the right tentricle and pou'ered on intermittenth', and blood tcmperarure is measured in
the pulmonarr arten-. The displaled cardiac
output is an average of measurements taken
ovet three to slr minutes.
The use of a P,\C adds a degree oi risk
to rhe technique: P-\Cs are both inrzsive
and associated with a number ol complications, most of s,hich are related to the
insertion of the catheter; these .include
pulmonarr- arterl rupfure, air embolism,
and arrhr-thmias. --\dditionalh', evidence
regarding the efficacv o[ P-\Cs in improving clinical outcomes is r'.'eak; therefore,
this technique is being used on a small
and declining percentage of critical care
padenrs (Schrvann et al. 201 1, \\'heeler
et aL.2006, Wiener et al. 2001).
The &arvbacks associated s'ith pulmonanarterr thermodilution have spurred the
development
o[
less invasive alternatives
tbr monitoring cardiac outpttt. These
alternatir-e technologies, although less
invasi"'e than pulmonarv arterv thermodilution, have their orvn dtaw-backs, and the
evidence regarding their usefulness can
be limited. 111 currentlv matketed cardiac
output monitoring svstems fall into one or
more oF rhe follorving categories.
Aderiql woveform onolysis. This minimallv
invasive technique uses an arterial pressure lioe and is based on the principle that
changes in arterial blood volume during a
cardiac clcle are accompanied br changes
in pulse pressure. Stroke volume can be
determined bv conrcrting the change in
pulse pressure obtained from the arterial
pressure rvaveform to a corresPoflding
change io blood volume. Pulsc pressure
changes that cortespond to a change in
18O xrllrx
7 of 158.
DEvtcEg JUNE
2ot3
) w.xrl.ors
LIMITATIONS OF THE LITERATURE
There ore limitotions to the published sludies we revievred, ond reoders should be core{ul
vrhen drowing conclusions from our findings. One limitotion is intrinsic to fhe meosuremenl of
cordioc oulput: The occurocy o{ pulmonory oriery thermodilulion, the gold stondord method,
is highly dependent on user technique. As o result, thermodilulion's resulis ore not olrvoys
relioble. This in lurn offects the ossessmenl of ollernoiive lechniques such os those used by lhe
products we review in this orticle: Becouse of ihe possible inoccurocies in ihe PAC meosure-
menis thoiihe olternoii're techniques ore compored ogoinst, on olternotive is generolly considered to be occurote even i{ there exists up fo o 3096 discreponcy (percentoge error) rvhen
compored io the thermodllution re{erence meihod. This helps exploin the sometimes limited
ond inconclusive noture of evidence regording the occurocy o{ lhe producis in this orticle.
Thus, our review of the evidence olso took inio considerolion studies comporing the producis
io other cordioc output moniioring techniques {e.g., vrhether ihe resulls from the two products
correlote, or show ihe some irend).
Another limitotion o{ the siudies is thot, while they look oi the occurocy ond correlolion
o{ the cordioc output volue, ihey don't iypicolly provide ony insighi into lhe use o{ ihese
oliernoiive iechniques in clinicol proctice {e.g., how often they ore used, ihe iype of theropies
they guide, their clinicol outcomes). Thus, we recommenC thoi hospitois opply coution rvhen
considering these devices for use.
atterial [described under Electrical, belorv]) and
characteristics such as aorric compliance the Edrvards Lifesciences FloTrac and
\'olumeYiet'.
and peripheral resistance.
blood volume are dependent on
Sr-stems using arterial rvar.elorm analr-
sis often need to compensate for chaoges
in individual arreialf aorac characteristics
that might occr-rr during the measurement
period, due either to phrsrologic changes
or to changes associated rvith trcatment.
\Iost devices compensate for these
changes using an independent measure
of stroke volume (e.g., thermodilution).
Horvever, some devices don't use anv
external calibration and instead use proprietan- algorithms that ftequenth- update
speci6c calculation parameters based on
the patient's demographic and phvsiologic
in[ormation.
Devices that use arterial rvar.eform
might not be accurate in parienrs
rvith certain arrhr-thrnias or arterial tree
pathologies (conditions that afFect the
anah sis
shape
of
the arterial pressure s'aveform)
or in the presence of circulaton'assist
devrces such as intra-aortic balloon pumps.
Three of the s1'stems rve rer.ierved use
this technique: the Conlled EC()\I (rvluch
uses a combination of afterial wavefofm
anallsis and impedance catdiographl
O2Oi3
Esopho geo I Dopp le n f his rninirnalir inr-asive
approach uses Doppler ultrasound techniques to measure blood florv rn the aorta,
and can be used in the absence of anr'
peripl-reral or central line. In this technique,
an ultrasonic rvaye is directed torvard the
aorta. \Ieasurements are perlormed using
ultrasouod probes that are inserted into
dre esophagus and aimed at the aorta. The
fte<luencr-
of
the rvave reflected back b,v the
moving blood cells (in t1-re aorta) is difierent
ilom the origrnallr uansmined ultrasound
&equencl-; this shift in irequenci- can be
used to obtain the velocin'of the blood
in the aorta. That velocin can be used to
calculate the distance traveled br- the blood
in the aorta during sr stole, knos'n as the
stroke distance. The stroke distance is then
used to estimate stroke volume using the
cross-secrional area of the aorta, lvhich is
estimated usiog the panent! age, height, and
rveight. This technique is rypicall,v used on
inrubated patients.
In this noninrzsive techoique,
cardiac output is deriwed br applving an
electtical signal across certain areas of the
bodr- (e.g., chest, ankle) using electrodes
Electricol.
ECRI lnsiitute. Member hospitols moy reproduce ihis poge
for internql distribulion only.
placed on the patient. Changes in the
applied signal are then measured during
each cardiac cr-cle. ()bsen'ed changes in
impedance and phase shift can be related
to changes in blood volume. Der,'ices using
these techniques are n'picallr susceprible
to motion artilacts that might prove a hindrance to continuous monitoring.
()ne o[ the more cornmon electrical
techniques is knorvn as impedance cardiograph,r' (ICG), u.'hich is based on tl're
theon that changes in blood tolume in
the aorta during each cardiac cvcle tesult
in irnpedance changes that can be used
to estimate stroke volume. The impedance changes can be derived br measuring
voltage changes to ao applied electrical
signal. Traditionalh-, the ICG technique
uses electrodes placed on the Patientt neck
and chesr (knorvn as thoracic ICG). ()ne
o[ these locat-ions is the suscep- excl.range) or intracardiac shunt Qlood
Ilorv across the chambers ot the heart),
ubiliq' to noise: In addition to blood flow;
pulmonan'blood flou'can be equated to
the electrodes ma,r pick up other signals
svstemic blood llorv or cardiac ourPut'
such as airflos, through the lungs. Two
Devices using this technique mal be inacsvstems rc revierv use ICG: the Conlled
EC()\I (rvhich uses a combi'ration of ICG curate in patients with abnotmallr high
intrapulmonarv or iotracardiac shttnts
and arterial r'"aveform anaitsis) and the \I
drarvback
(rvhich can occur in thc presence
fledical \ICaS.
Rebreothing. The noninvasive rebreathing
technique is based on the principie that
intake or elirnination o[ a gas (as iodicated
bv the ditference in the concentration of
the gas entering and exiring the lungs) is
proportional to the amount of blood florv
to the lungs that is invol't'ed in the gas
exchange. The amount of gas consumed/
produced is used as an indication of pulmonarv blood flos-, and in the absence of
significant intrapulmonan shunt (blood
florv to thc lungs not invoh-ed in gas
respiratort or cardrac diseases).
Tronsthoro(ic echocordiogrophy.
oi
some
In trans-
thoracic echocardiography, or TTE, an
ulrasound probe is placed on the chest
to measure the Yelocifr o[ blood across
the aort-ic or pulmonan= vaive. This, combined s,ith the cross-sectional area oF the
valves, allot's calculation of ieft or right
stroke volume and cardiac output. This
noninvasive technique mar be inaccurate
in patients u'ith r,alr.e and florv-pattern
abnormalities.
PRODUCT REVIEWS
Pulmonarv arter't- thermodilution is considered the goid standard for cardiac ourput
measurement. Therefore, the tables that
accompanv our product reviervs comParc
the evidcnce on each product with the
evidence on thermodilution. But because
comparisons to other methods can also
be use6;1 (see the box on page 180), rhe
discussions in each revierv also take into
consideration srudies comparing the products to otier cardiac output techniques.
Coolled Corp. [101345], Utica,
(US-\); +1 (800) 4-18-6s06, +1 (31s)
supplier.
\Y
7
97 -831 5 ; g'rvs,. conmed.com
Procedure. f'he EC()]I (rvhich stands for
Endotracheal Cardiac ()utput \Ionitor) is
a minimallf inrasive srstem that measures
cardiac output and other related hemodi-namic parameters using a combination o[
trvo different techniques: ICG and arteriai
s.atetbrm analvsis. The monitor obtains
an impedance rvaveform rvith the ICG
O2OI3
8 of 158.
technique and uses this in conjunction
,x.ith the arterial s,zr.elorm analrsis tech-
elecuodes are in such ciose proximifi'to the
aorta, the EC()]I is less susceptible to noise
compared to traditional ICG methods.
The monitor consists of an endotracheal
rube with a printed elecrrode arrat that
applies the electrical signal required for ICG.
The elecrode arrav consists of a single
elecuode at the proximal end of the rube
and slr electrodes at the distai eod. \\'hen
the patient is innrbated, the distal eiectrodes
are located adjacent to the ascending aorta,
thus iacilitating aortic blood ilorv measurements. The electrode arrav also detects
the R-rr,ar-e and uses the R-R interval ro
obtain l-reart rate measurements. ,lterial
pressure data is obtained from an esisting
arterial pressure Line and, in addition to cardiac output, facilitates estimation oi other
hemodvnamic pararneters such as s1'stemic
vascular resistance and stroke volume rzriation (neit1-rer of u'hich can be obtained from
ICG).;\t the beginning oF the procedure,
the clinician must enter the patientt height,
rveight, and age, rvhich are used along with
the measured ICG signal to estimate suoke
volume. -\ccording to the vendor, since the
clinicol uses, -\ccording to the vendor, the
trC()\I is used primarilv in the OR lor
short-term conLinuous morutoring and
is indicated for use in patients rvho are
expected to be inrubated for 24 hours ot
less. .\lthough ICG technologr in general is
susceptible to motion artilacts, this mav not
be an issue with this derice since inrubated
patients in the ()R are also npicallv sedated.
nique to determine catdiac output.
ECRI lnstitute. Member hospitols moy reproduce lhis poge
for inlernol distribution
only.
Disposobles,
The monitor requites a propri-
etarr- disposable endotracheal rube.
t It.i
:3
ConMed ECOM. (lmoge courtesy of
ConMed.)
w.*ri.org > HEALTHDEVICES
JUNEZoTS
l8l
)
s"i'ton"
LITERATURE REVIEW: CONMED ECOM COMPARED TO THERMODILUTION
Study
Polienlpopulolion
Referencemethod
Outcomes
Boll et ol. 2010
40 cordioc surgery
Bolus pulmonory ortery
With o percentoge error of 5O% ond o correlotion coe{ficient o{
r : O.49, the h*o methods were not considered comporoble.
Gennort et ol.
I
Thermodilution.
With percenioge errors of 419'5 ond 5970 (bosed on orteriol Iine
lccotion), the iwo methods were not considered comporoble.
201
2 (conference
obskoct)
potienls
4 cordioc surgery
potients
lhermodilution
Mous et ol. 201 I
29 coronory orlery
byposs groft poiients
Bolus pulmonory ortery ther-
modilution ond tronsesophogeol echocordiogrophy
The percentoge error wos 5070 compored to thermodilution ond
48% compored to echocordicgrophy. However, lhe oulhors report
occeptoble trending of cordioc outpul.
Moller-Sorensen
et ol. 20'12
25 coronory ortery
thermodilution
hom 37o/o to 45% bosed on pctient
posilion, the ivro methods v/ere not considered comporoble. The
study olso reported poollrending of cordioc oulpul by the ECOM
Von der Kleii et ol
20 1 2 {conference
20 coronory ortery
byposs grofi potienis
Pulmonory orlery
The obskoct reports poor correlotion (r
byposs groft potients
VVith percentoge errors ronging
monilor.
obstroct)
thermodilution"
= 0.23) beiween the ECOM
monilor ond lhe reference method in postoperolive potients.
- The obstroct does not speci{y whot lype of thermociilulion wos used.
'^
The obslrqcl does not specifo whelher lhe melhod of ihermodilulion used vros coniinuous or rnlermiilenl
Evidence. \\e revierved eight clinical srudies that compared the EC()\I monitor to
other methods of cardiac output measurement. The srudies involved cardiac surgerr
and abdominal surgerv patients, and the
number of patients in the srudies ranged
irom 12 to'10. In five of the eight snrdies,
the ECOII monitor was compared to pulmonal\. arter\' thermodilution; these studres
are outlined in the table on this page. In
the remaining three srudies, the reierence
methods rvere transpulmonarr- thermodilu-
tion (rvhich uses a central venous catheter
rather than a P-\C) (F'ellahi er aL.201.2),
esophageal Doppler fl orgensen er al. 2Ol2),
and the FloTrac sensor @airamian et al.
201 1; see our rerierv of this slstem belorv-).
In addition, three ol the eight srudies
(Bairamian et al. 2011, Gennart et aL.2012,
\hn der Kieij et al. 201,2) rvere reported as
conference abstracts onlr; therefore, ther
ma\- not rePresent the 6nal results and conclusions oi the srudies.
In all eight srudies, the authors concluded that the device \\,-as not comparable
to the reference method. The evidence
regarding the monitort abilirl'to trend
cardiac output is inconsistent, rvith one
srudv (\Iaus et al. 2011) reportiog good
trending (based on the correlatjon benveen
182
x;arrg
9 of 158.
DEvtcEs JUNE2or3
) w.xri.org
the EC()f t monitor and the reference
techniques) and rs'o studics (Jorgensen
er a|.2012, \Ioller-Soreosen et al. 2012)
reporting poor trending. livo srudies
(Bairamian et al. 20i 1, Fellahi et al. 2012)
conclude that the techoique mav be useful
for predicting fluid responsiveness (using
stroke volume rariation). Trvo other srudies @all et al. 2010, ]Ioller Sorensen et al.
20i2) tbund that the EC()\I \i,as susceptible to interference from electrocauterrequipment used during surgeries, leading
to intermittent loss of signal during electrosurger\' and somedmes several seconds
after electrosurger\- \'v?s discontinued.
\\b d-rd not 6nd anr studies ad&essing
c[oical outcomes.
Supplier. Edrrards Li[esciences Corp.
[374501], Irvine, C-\ (US.\); +1 (800) 124(e-+e) 250-2500; rwwv.edrvards.
,
::f
.,
'.\lthough
thc authrlrs usc rhe rcrm "clcctrocrutcn,"
Procedure. The FloTrac sensor uses
arterial rvaveflorm analrsis to estimate
continuous cardiac output using the FloTrac algorithm. -\ccordiog to the vendor,
the algorithm constantlv monitors rhe arterial pressure vaveform and compensates
for changes tn arterial f aottic characteris-
tics. The FloTrac sensor can be used
with
either thc E\-1000 or thc \-igilco rnonitor
offered bv Eds'ards Lit'esciences.
The FloTrac sensor can be used onlv
on patients rv-ho har.e an existing arterial
line that can be connected to the sensor.
The sensor has nvo output lines: ()ne is
connected to the arterial pressure cable
of a bedside monitor, and the other is
connected to the E\-10o0/\'igileo monitor. The sensor, once connected, sends
arterial pressure rv'aveform information
to both monitors. The cardiac output
monitor (either the E\'1000 or the \-igileo)
uses this informat.ion in coniunction with
patient demographic information entered
br the user (e.g., age, gender, height,
rveighr) to esrimate cardiac output.
\\'e covered the FloTrac sensor in our
December 2009 Guidance -{.rticle. Since
then, the vendor has updated the device's
algorithm. The hterarure review-ed in the
*cloelierc tlrc intcnt wrs "clcctrcsurgcn."
O2013
ECRI lnstilute. Member hospiiols moy reproduce this poge
for inlernol distribution only.
LITERATURE REVIEW: EDWARDS LIFESCIENCES FLOTRAC COMPARED TO THERMODILUTION
Sludy
Bioncofiore et ol. 2Ol
De Bocker et ol. 20,l
Morque ei ol. 201
I 2l
liver tronsplontolion
potients
I
3
28 liver tronsplontotion
o1.2012
With o percentoge error o{ 37o/" ond o correlotion coelficient
of r = 0.65, lhe ogreement betueen the irvo methods vros noi
considered clinicoliy occeptoble.
pulmothermodiluiion
Continuous pulmonory ortery
ihermodilution
18 seotic shock potienls
Su et ol. 2012
Vosdev et
Oulcomes
Bclus pulmonory
nory ortery
'l
ol. 20 I 2
Reference
Bclus ond continuous
58 septic potients
Slogt el ol. 20 I 3
Tsoi et
method
ortery
thermodilution
Potient populotion
Bclus pulmonory ortery
thermodilution
9 septic shock potienis
Conlinuous pulmonory
ther'modllution
potients
ortery
\Vith o percentoge error ol 30.4o/o, the FloTroc sensorwos
considered io be ot leosi os occurote os the reference methods
With on overoll percenloge error o{ 640/o, ihe iwo methods
v/ere not considered comooroble.
With on overoll percentoge error ol 53o/o, the two methods
v/ere not considered comporoble. The oulhors conclude thot
the trending obility of the FloTroc sensor is foir.
Wilh o percenloge error ol 75o/o,the irvo melhods were nol
considered comporoble. The outnors olso report poor trending
of cordioc outpuf by ihe FloTroc sensor.
20 liver tronsplontotion
potients
thermodilution
Wilh o percentoge error of 54.93o/o, ihe outhors report poor
correiotion be,'ween the furo methods.
38 coronory ortery byposs
groft poiients
Eclus pulmonory oriery
thermodiluiion
With percentoge errors of 2Ao/o ond 2296 depending on the
monitoring site (rodiol vs. femorol), lhe turo melhods were
considered comporoble.
Evidence section belorv deals onlv rvith the
oi the updated algorithm.
perlormance
The Floftac sensor can be
used for [ong-term continuous monitoring
ot- patients rvho are alreadv connected to
an invasive blood pressure monitor rvith
Glinicol uses.
preerisdng arterial iine access.
Disposobles. The srstem requires the proprietarr', disposable FloTrac sensor. It does
not require anv proprietary catheters.
Evidenre. \\'e revierved 16 srudies published
since 2010 that compared the performance
ot- the nerver generation oi the FloTrac
sensor with other cardiac output monitoring techoiques. Seven of the 16 srudies
(see the table abor.e) compared the FloTrac
to pulmonarr- arter\- thermodilution, four
studies compared
:j
Fdwords Lifesciences FloTroc sensor.
(lmoge courtesy of Edvrords Lifesciences.)
020]3
10 of 158.
it to uanspulmonarr'
thermodilution, and the remaining 6r.e
compared it ro Doppler echocardiographr.
The srudies invoived cardiac sutgerv, lir.er
transplantation, sepsis, and abdorninal surgery patients, and the number o[ patients
in the srudies ranged from i8 to 60.
In our previous revierq the earlier generation of the FloTrac sensor',vas found
ECRI lnslitute. Member hospitols moy reproduce this poge
for inlernol dislribulion only.
to be comparable to rhermodilution in
stable cardiac surgerr patients, and the
one recent stud\- we rerierr-ed that looked
at cardiac surgerl patients (\hsdev et al.
2012) reports that this is the case with the
nerv version oF the FloTrac.
\Iost srudies on the nerver generauon
the sensor examlned its performance
in hemodlnamicallr unstable patient
populations (e.g., septic patients, liver transplantadon patients). In 10 o[ the srudies
examined in our current tevies', the authors
conclude that the latest generation of the
FloTrac sensor is not compatable to the
relerence techmques; patient populations
in these srudies include patients undergoing abdominai surgerl-, patients undergoing
liver transplantauon, and patients undergoing vasopressor therapr- (rvhich could
affect arterial compliance). -\mong srudies
comparing FloTrac to thermodilution for
patients with sepsis, one srudr (De Backer
et al. 2011) reported good accurao',
rvhereas t$-o others (\Iarque et al. 2013,
Slagt et al. 2013) reported a high petcentage
error. Tr'"'o srudies (Benes et al. 2010, \Iarer
et al. 2010) looked at patient outcomes
oi
w.eri.ors > HEALTHDEVICES
JuNE2Ol3 183
gaidance
rvhen the FloTrac sensor \\?s used for
Buidmanagementduringmajorabdominal
surger'compared to,standard fluid
man-
agemenr ePPfoacnes (usmg mean arrefral
and ..rrtr"l',,.rro,i. pressure ro
guide therapr). Both srudies shotved that
pi"..ur.
LIERATURE REVTEW: EDryARD5 LlFEsclENcEs voLUMEVtEW
coMpARED To THERMoDILUTIoN
Study
ol,
!ô1lo 9t
Polient populotion Reference method Oulcomes
5 potienls undergoing Bolus ond continu- lVith p-mentoge errors
incidence oI compLi- 39]^'J::1"-l liver tronsplonlolion
ence obstroct)
^.,d
catrons decreased rvhen the FloTrac senior
rvas used. $ote that the f lar-er studr rvas
funded br- Edtvards Littsciences.)
Ihe lcogth
oi'rt^r
Supplier Edn'ards Lifesciences Corp.
-l
(800) -+24[374501], kvrne, C-\ pS-{);
8, + I (949) 250-2500; t'rvrv.edr','ards.com
321
Procedure. The \blumeVierv sensor is
part of the \blume\ierv set and is used
in conlunction with the E\-1000 monitor
for continuous cardiac ourput morutor-
ing. The sensor uses the arterial r'"'ar-eform
analrsis technique to estimate cont-inuous
cardiac output, and the arterill s,-aveform is
obtained using the \blumeYierv femoral
catheter. Horvever, uniike the FloTrac
sensor, s'hich does not use any external calibration, the YolumeYierv sensor
requires calibration using an external
cardiac outpur measurement. Tvpicallv,
:J-',!'lT:::J "'t"ry
thermodrlutron
calibration is recommended tbr arterial rvaveform analrsis techniques For
unstable or criticallv ill patients at defined
intervals of time (can range from e'r'err
30 minutes to everl eight hours based oo
patient condition) or before undertaking anr therapeutic inten'endons. The
rcndor offers the abilin-to perform transpulmonarv thermodilution' using the
\blumeYierv set, and the resultiog cardiac
output measurements can be used to periodicallr calibrate the YolumeView sensor
measurements.
The transpulmonan' thermodiiution
technique can also be used to obtain r-o1ume measurements such as giobal
end-diastolic volume, global ejection lraction, and extra\ascular lung s,ater.
cllnicol uses. The YolumeVies, set can be
used for long-term contjnuous monitoring
of padents rvho have an existiag central
renous linc.
ot
t*,
Edwords Lifesciences VolumeView sensor.
(lmoge courtesy of Edwords Lilesciences.)
184 xglrraDEvlcts
11 of 158.
JUNE2or3
> ww.*ri.ors
- .\ salinc solutioo of kn()\\-o tcmpcriturc is injcctcd
into ao cxistinla ccntral rcnous cathctcr, end rhc chengc
in rcmpcmturc of thc injcctatc is mcasured in thc [cmo,
nl artcrr u:ing thc \irlumc\iicrv crtheter,
dilution, the VolumeView wos
nol considered comporoble lo
the re{erence methods.
measurements from the YolumeYierv sensof to those obtained from pulmonarr
arten- thermodilution rvith P^\Cs in five
patients undergoing liver transplanration (see the table above). lhe srudr[ound that the \blumeYierv results rvere
not comparable to the ralues obtained
from thermodilut-ion. This srudr. s,zs onlr
repotted as a conference abstract; thus, the
results reported mav not represent the 6nal
results and conclusions of the srudr. The
other studl', Kiefer ct al. (2012), compared
the \blumeYierv sensor rvith aflother arterial s,'ar-e[orm analrsis technique that a]so
uses transpulmonarr. thermodi.lution for
calibration. This srudr rrzs performed on 72
critjcallr ill patients and found that the VolumeYierv results *-ere comparable to those
ftom the reference method; note that the
studv $?s funded br Edrvards Lifesciences.
\\'e did not 6nd anv studies addressing
cl-inical outcomes rvith this device.
Disposobles. The YolumeYierv set, rvhich
inciudes the \blumeYiew sensor and the
\blume\-ierv [emorai catheter.
Evidence. The evidence on the \blumeYierv sensor is limited, since the derice
rvas introduced in 201 1 and is therefore
relativelr nes'. We onh- found nvo srudies
that compared the YolumeYiew sensor
rvith other cardiac outpur monitoring techniques in human patienrs. Costa er al.
(2012) compared rhe cardiac outpur
o[.49o/o
:ii
i",*lT
-':ilil"-dI
fhermo.olus!fond
conl'nuous
supplier.
\I
\Iedical US-\ [457874], -\kron,
{)H (US-\); +1 (800)
nimedical.co.il
979-2904; rvrv*r
The \ICaS (rvhich stands for
Invasive Cardiac Svstem) consists
of a laptop aod one pair oF s'rist-ankle
ICG electrodes. The st'stem is based on
the ICG technique, rvhich is based on the
theorl that changes in the voltage oF an
electrical signal applied across an arcz of
the bodr are due to changes in impedance
(rvhich in rurn are associated *'ith changes
in blood volume during each cardiac clcle).
Procedure.
\on
@2013 ECRI lnstiiute. Member hospitols moy reproduce ihis poge for inlernol disiribution only.
impedance through nvo ICG electrodesone on the patieot's $,rist and the other on
the contralateral ankle.
(e.g., special ke1'board covers) ma,v be
needed, particularlv rvhen the device is
used in the ()R.
-{t the beginning of the procedure,
the clinician must enter the patient's
gender, age, t'eight, and height. This data
is used in conjunct-ion with the impedance
data obtained from the rvrist and ankle
electrodes to estimate stroke r.olume. The
electrodes also provide heart rate and a
single-channel electrocardiogram (ECG).
The ECG s.aveform, the impedance rvaveform, and the hemodi'namic parameters
are all displared on the laptop. Standard
ECG electrodes are provided as an option
with the monitor and can be used in
addition to the ICG eiectrodes to obtain
three-channel ECG.
Disposobles. One pair
elecuodes.
Cllnicql uses. The ICG technologv is inherentll suscept-ible to motion artifacts, and
Nl Medicol NlCoS. (lmqgg courtesy oI
Nl Medicol.)
-\Iost cardiac output monitors that use this
model measure changes in thoracic impedance wlth electrodes placed on the chest;
horr,'e'r'eE the \ICaS monitor measures
LITERATURE REVIEW:
n picalh- patieots need to be still s'h.ile
measuremeflts are being takeo.'fhus, this
monitoq like others that use the ICG
model, is more suitable for spot-check
measurement of cardiac output and
other hemodvnamic parameters than for
continuous monitoring. -\nother issue
to consider is that since this monitor is
laptop-based, infection conftol risks must
be taken into account, as the presence of
a kerboard complicates cleaning of the
device. -\s a result, additional precautioas
Potient populotion
Coiter, Moshkovilz,
et ol. 2004
122 cordioc potients
Cotter, Torre-Amiot,
93 polienls wilh heorl
ol. 2004
foilure
\\:e did not Efld anv studres addressing
clinical outcomes with this device.
With o correlotion coefficient o{ r
considered comporoble
= 0.8.l, the two methods
o correlotion coeflicieni oi r
considered comporoble.
= 0.9. ihe two meihods were
Bolus pulmonory
ortery
Poredes et ol. 2006
35 cqrdioc potients
ihermodiluiion
12 of 158.
thoracic ICG, in addition to the \ICaS
monitor, to thermodilution for the same
patients, and Found the \ICaS svstem to
better correlate with thermodilution.
thermodilution
Thermodiluiion*
3
-\ll slr srudies found good correlation
betrveen the NICaS svstem and the reference method. Tivo of the srudies (Cotter
et al. 2006, Goor 2006) also compared
Oulcomes
46 cordioc polienls
O2O l
122.
lVilh o correlotion coefficient of r
ortery
thermodilution
Goor 2006 (con{erence
The obslroct does nol specify whot
to
Reference melhod
43 cordioc potienis
'
35
Bolus pulmonory
Cotter et o]t.2006
obsiroct)
Evidence. \1'e reviewrd slx srudies comParing cardiac ourput measurements from
the \ICaS monitor with those ftom other
reference methods. Five of these studies used thermodilution as the refetence
method (see the table on this page), and
the remaining srudr' (Leitman et al. 2006)
used Doppler echocardiographr'. ()ne of
the srudies (Goor 2006) rvas reported as
a conlerence abstract onlr; therefore, the
reported findings mav represent onlr- preliminarr results and not the 6na1 results
and conclusions o[ the srudr. The srudies
involved cardiac patients, and the number
of patients in the srudies ranged from
Nl MEDICAL NICAS COMPARED TO THERMODILUTION
Study
e1
of rvrist-ankle
\pe of thermodilulion
ihermodilulion
were considered comporoble.
Vy'ith
:
0.886, the two methods
were
Wilh o percenioge error ol 2OYo, lhe two methods were considered comporoble.
Wiih o percenioge error ol 19.95o/o ond o correloiion coefficieni
of r : 0.9], the iwo methods were considered comporoble.
wos used.
ECRI lnstitute. Member hospilols moy reproduce lhis poge
for internol disiribution only.
w.*ri,org
>
HEALTHDEVICES JUNEzoTA
I85
&!,,!:!{,c{"u,o,
RE
FERENC ES
ISairemianJIr, RrxrkJI:, :\llard \IU, et al. (irmpanson of stnrkc trtlumc rariation obtained from
an arrcrial prcssurc to cndorraclteal sensor svstcm. 2(111 .\nnual \[ccting of thc Internatir>nal
-\ncsthcsia Rcsearch Socicq'; 201 1 \Ia1' 21 -2-l;
\hncourcr
(Oanada).
tlall'I'lt, (iulp l)(i,
Patel
\,
cr al. (irmparison
of
tlrc cndotruchcal crrditc tlurput m<rnilor ltr
thcrmodilution in canlirc surgcn plicnts
(ttli,rlrird: I,b;; zlnt;th 2(llo ()cl.21(5):1(t2-6.
.l
I et al. Inrnropcradr c
fltrid optimization using strokc r-olumc variation in high risk surgical paticnts: results oI
prospcctirc randomizcd srudt (n7 (-az
Bcncs J, Oh1'tn I, .Utmann
2{)1{};l-1(3):Rl
lll.
lliancofiorc G, Critchlcl'L.\, Lcc .\, ct al. Iilaluation of a ncs' sr>fru-arc tersitln of rhc lrlo l rac/
Yigileo (r-crsion 3.1)2) and a comparison rvith
prcr-ious data in cirdrotic paticnts undcrgoing fir-er trrnsplanr rurgc^-- /1n.ith -'lrulg2o11
Sep;1 13(3):515-22.
()
ltcnnerJ, Clruencrvald ]I, ct al- -\ compadthird-gcncrarion scmi-inrasirc ilrrerial
rvavcftrrm analvsis rrith rhermodilution in
patients undergoing coronatr surgcn'.
21l2;2tt 12:15 lt)81. ll lpub
.\' Licttti/itlUbiltlurnul
Broch
son
of
2{)l 2 Jul 31 .l
Costa (1, Cccconct'l', Baron Q et al. Clrdiac output moniroring in cirilrotic paticnts: IiYl()(Xl
r-crsus pulmonry arrerv catltctcr--prcliminarv
data fc<rn[crcncc abstractl. lt (rili;e! (qrt
(onfirc tt. 32nd Intcrnational -svmPosium
on Inte nsivc Care and I'imcrgcncr \lcdrcine
Ilrussels llclgium. 2012; (irnlcrcncc Srarr(r-ar.
pagings):20 1 2{)320-21}1 2(}S79. .\lso atailable:
htg://ccfrrrum.com/contcnr/pdi/cc
I
0ti26.pdi
(irrtcr G, Ikrshkoritz \ Kaluski Ii, ct al. -\ccunte,
nonim-asir-e continuous monitoring of cadiac
outpur bv s'holc-bodv clectrical bioimpcdancc.
Ur.,t 2001 .\pc I 25(-1): 1.13 I -'lt).
Cottcr (1, Schachncr -\, Sasson L, ct al- Impcd
ancc carditrgraphv retisitcd. I'l1itl -\l*ti 2(l(l(t
Scp;27(9):U1 7-27.
(lottcr (1,'lirrrc--\mot (], \tred Z, ct al. Wht>lc-
bodv clcctrical bio-impedrncc is rccururc
in non invasirc de tcrmination oI cardiac
output: a tlrermodilution contnrllcd, PmsPectivc, d<rublc blind cr aluation. I lr.l I lurt I iil
.\' upll 21ll.l;3(Supp);1 9. .\lso alailablc: httP;//
cur j htiupp.oxtbrd journais.org/conrcnt/3/
supplemcnt/ I9. l.extract.
\las
-lan
,\, ct al. .\rtcrial
G,
pressurc-bascd canJiac output monitoring: a
Dc Backer I),
186
xrarrxDEvlcEs JUNE2ol3 > ww.Bri.org
13 of 158.
mulriccntcr r alidation oI thc third-gcncration
strfnvare in scptic Paricnrs. InlLn;irt (.are -\[td
2t)1 1 Itb;37(2):233--11).
li(iltl
Insriturc:
I:ndotilcheal crrJirc ourpur mrtnirrr st stem
(Oon\led (lorp.) tor dctermining carditc output
and strokc r-olume lcusrom product brict-1.
(lustom I Iotlinc Se n-icc. 3)l2 l)ec -|.
lilo'I'mc scnsor Qidrvards Litlscicnccs, l.l.(-) ftrr
noninr-asilc cardiac output mrlnitoring lcustom
product briefl. Cust()m I I()tlinc Scn ice. 201 2
Dcc 6.
\oninr-asilc cardiac slstcm (\I \Icdical. Lrd)
[<rr monitoring cardiac ()utPut lcustom pnrduct
bncf]. Custom I Iotlinc Scn-ice. 2012 l)ec 5.
Stra(l:r from thc hcart: undcrstanding cardilc
ourput monitr>ring tcchniques lguidancc erticlel.
I
kt lr h l) t ir i
2009 Dec;3ll(l 2):.3tt6--lt)
l.
YolumcYics set ([idrvards Litcscicnccs, I-1.(-)
[r>r minimellt inr rsilc crrdilc output moninrring
[custom producr brict-1. Cusrom I lorlinc -\cil'ice .
201-l I'eb 15.
l;cllahi-ll-, |ischer
\l()
Rebct
()
ct al. -\ comprrison
oI cndotracheal bioimpcdance
cardi<>gr'rphr
and transpulmonarr thcrmodilution in ctrdi:c
surgcrv padcnrs.. | (.*di o lfuiit ; I L, : . lrc l l, 2111 2
.\pr'26(?):217-22.
(]cnnart I;, lJeckcrs
\trborgh
O, et al. Impacr
accuritcr'
of cardiac ourput proi-ided bl an cnd()trachcal
bioimpcriancc dclice [c<>nfcrcncc abstr,rcrl. Cz7
S,
of;rrtcrial carhcrcr krcation tln thc
(rr,
2{)12; t(r(Suppl
l).
(irxrr l) l'cripheral rersus thoracic impcdance c,rrdi(Uriphr lconfcrcncc abstncrl. | ( ,trdiu, l iil 20{)6
.\ug; 12(6 Suppl):S5J-5.
IK. llisg.rarJ.l, ()ilsm l. (-,'mp.rrison oi
bioimpcdance and ocsophagcal I)t>pplcr cerdiac
output monitoring during abdomin:rl atlrtic surger.t-. Cit (,an 20t2;16(SuPPl l).
J<trgcnscn I
Khs-aonimit B, Bhuralanontachai l{. Prcdicrion oi
Buid rc-iponsircncss in scptic shock patienrs:
comprring -strokc rrrlumc r trriarirtn br likilrac/
Vigilur ancl rur(,mJtLd ptrlsc prcssurc r.triltion.
I Lrr.l,,bttt.;l I tt ;io I ?{}1 2 treb;29(2) :6{_9.
\, I krtcr CK, \Iirn (i, et al. (llinicLrl v'rlidarion of a nc$ the rm()dilution svstcm [<rr thc
asscssment of crrdilc ourput rnd 11)lumctdc
paramc tcrs. ( i t ( -a n 2tl 12 \ Iav 3{); I 6(3) :ll'9tt.
Kre fcr
Kor,rkc \', \imada'I', \agata I I, ct al. 'liansicnt
hcmodvnamic change and xccurJc\'oI arrcrial
blrxrd prcssurc-bascd cardi,rc ()utPut.. lrc'l,
,lnd32ttl t .\ug;t
13(2):272-a.
Kusaka \', \irshit;rni K, Iric '1, ct irl. Clinical comparis()n oi rn cchocardiograph-dcni-ed r-crsus pulsc
counte r-dcrir'.d c.trdi.rc outpur mcxsurLmrnt in
abdominrl ?()rric xntutrsm surgcr.t. | (trdiotlttrt'
l. ; t .. I n, t h 2tt I 2 .\pr,26(2) :223' (,.
"t
I-cirman \I, Suclre r t:, Kaluski l'1, ct al. \on-invasir c
mcasuremcnt of crrdirc ourput bv rvholc-bodr
bi<>impedlncc during doburaminc strcss cclttt-
crrdiogr.rpltr: clJnrc,rl implic.trioni ln Prticnt j
s ith Ic[t rrntricul;rr dtsfunction and ischaemia.
I:Lr I I hnrt litil2t){)6 \lar;tt(2):13(r'l{}.
DO YOU HAVE A QUESTION ABOUT CARDIAC OUTPUT
MONITORING?
For exomple, ore you wondering . .
)
)
)
.
Whol fociors do I need to consider when purchosing o minimolly invosive or noninvos ve
cordioc output monilor?
How widespreod is the use of cordioc outpui monitoring in clinicol proclice?
Whot is lhe most estoblished olternotive technique in monitoring cordioc ouipui?
You con get onswers to these ond couniless other quesiions through HD Advisor. Just
click the link in lhe top righi corner o{ your Heolth Devices home poge to occess your HD
Advisor doshboord ond storl your inquiry.
@ADVISOR
Ask the [xperts
O2O'13 ECRI lnsiitute. Member hospitols moy reproduce this poge for internol distribution only'
S, (iros .\, (.himot L, et al- (:ardiac outpur
monitoring in scptic slrock: criluad()n oI thc
third gener:rrion l;krtrac-\'igilcoO. .l Qin,\lonil
\Ianluc
(,q
nOil r
2( l l 3 .l
0n;27 (3):27
3
-().
\Iaus'I'^\I, llcbe r 13, llanks I).\, ct al- Oardiac outpur
dcrcrmination lnrm cndorrirchcirllr mcasured
impedancc cardiographv: cljnical evaluarion of
cndotrechcrl cardiac output monitor. / (,tnlillarn: V'r ii .' Ift t h 2ll I I ()ct;25(5):771)-5.
\Iarer,J, lloldt.J, \Icngisru .\\I, ct al. (loal-directcd
inrrxrpcntire thcrapr bascd on eutocalibnrcd
artcrirl prcssurc s alclorm analvsis rcduccs ho-.pital stav in high-nsk surgical paticnts: r rand<>mized, controlled tnal. (. i r ( ut 2o ll ); 1 -t(1 ) :l{ 1 tl.
\Icl.ean .\S, t luang S-f, Kot \I, ct al. Oomperison of
cardiac output measu(cmrnts in criticallr ill pa-
ticnts: ljlo'I r,rc/\'igileo rs transrlroracic [)oppler
cclrocardi<rgrapll'. . lart,/h Iukr,il t Ltre 2O11
-[ul;39(.+):.i9{).tl.
al. 'l'hc impacr
r>i phcnrlephrinc, cphcdrinc, and incrcascd
prcload on third-generation Yigileo-likil rac and
csophagcal l)opplcr cardiac output mcasurcmcnts. -. h c't h .. 1na13 2{) I I ( )cs1 1.3(-l):75 1 -7.
\Icng I., 'l ran \P, .\lcxandcr IlS, cr
Coburn \I, Iirics \1, er al. Performancc
of cardiac output measurcmcnr de rir-cd from
artcrial prcssurc s'ar-efrrrm anah sis in prticnts
\Ictzcldcr
S,
requirin.q high-dosc \f,s()prcss(rr
n *,th 2{l I I Jun-10(t(():77 G{1-1.
thcnpl. IJr./
;1
\ftrllcr-Sorcnsen I [, I hnsen KL, Ostcrgaard \1, ct
al. Lack oi agrermcnt and trcnding abilin oi the
endorrachcal cardiac outpur monitor comparcd
n'irh rhcrmodilurion. ;li/a ;1rut.;tLt.'iol .\',und 2012
.
\I.\, lScute J, ct al. (i:rrdiac otrtpur
mcasured br unca[bratcd arteriai prcssurc s.avc-
Slrrgt (i, de Lceurv
\pr;5(r(-1) ;-133-{t).
\krnnct X,
.\ngue I
gLncration I'kil'rac/\'igrlco docs not reliablr
rrack changcs in cardiac outpur induccd bl norepincplrrine in cridcalll ill p^ticnts. Br.[ ,Arp]rilh
2r) I 2 .\or;1 0tt(.l) :61 5-2?.
\lurdri Sll,
I Iess JR, I lcss .\, ct el. Irocuscd rapid
echocardiognphic cr-aluation r-crsus vascular
cf,thcr-brsLd [*l ;rsscssmcnt oI c:rrdirc ourput
xnd tuocri()n in criticallr' ill tnuma prticnts../
'
l)u r mt . l,u t c (.o rc .\' m3 2t) 12 \ Iav;72(5) : I I 58-6-1.
\lutoh 'I, Ishikarra'l', Kobalashi S, ct al. l)ertirr-
mancc o I tlrird-gencrati<>n Irkiliac/\'igilc<r
srstcm during hrpcnlrnamic thcrrrpy lor deleled
ccrebral isciremia aticr subarachnoid hemorrhage . .1ir4:i
-\irrr, / Iilt 2lll?;3'99.
\ational I Icarr, I-ung, and l]lood Instinrte .\cutc
lle spiraron Distrcss Slntlromc (.\Rl)S) Clinical
'lrials \cnrrrrk; \\'hcclcr .\l', llcrnard ()R, cr al.
Pulmonan'artcfl' rrrsus ccntral \ cnous cathcter
to lluidu rreatmrnt of lcutc lung injun: \ I:41 |
.\lid 2{t{t6 \la1' 25;35"1(21):221.3-2-l
Parcdcs ()1., Shire-[, Shinkc'I, ct al. Impedancc
cardiograplrv [rrr cardiac output cstimarion: rcli-
abilin oI rvrist-to-anklc elccrrodc conligurarion.
(.ir:.1 ?0o6 Sup;7tt(9):l l6+fl.
Sclrl-ann \\I, I Iitlel Z, I krutl .\, ct al. I-ack of
eftectitcncss ol thc pulmonan' artcn' cathcrcr in cardiac :urgerr. .'luitb t7nQg2\l1l
\or';l
lorm
\,.lozn iak \1, ct al. 'l hird-
analvsis br rcccntlr rclcased st>fnrare r-cr-
sion 3.()2 r'crsus thcrmodilution in scptic shock.
I Uin .\[uir (tn\ar 2013 .\pc27(2):l7l-7.
Su
l](-,'t-.ai Yl', Chcn C\', ct al, (lardiac ()utput deriled fiom artcrial pressurc q-ar-cfrrrm analvsis in
padcnrs undcrgoing lir-cr transplantation: i'alidiu
oi r third-gcncrarion dcticc. 'l-ratpltnl Prut2ll12
\Iar;$t(2):'12-l lt.
'lsai Yl; Su lXl, Lin (l(1, ct al. O;rrdirc ourput dcrii-td
trom arrcrial prcssure t avciorm enallsis: r alidltrrn oi tlrc rhir,.l-gcncr:tion sol-ro'rre in peticnrs
undcrgoing r>rthotopic lir-er rransplantation.
'
l raulllail Pro: 2ll12 \Ier;-l-l(2):-1.33-7.
Van dcr Klcii S(-, Koolen B, (icrritse B, cr rr1. (-linical cr:rluatir>n ot- a nes' cndotrachcal impcd-
ince cardiogralhl mcthod. tlu,,/fu.,it 2l)12
.lul{t7 (7):729-33.
\ esdei- S, (-hauhan S, (ihoudhun \1, ct al. \rtcrixl
pressure s ar-cfrrrm dcrir-cd cardiac output I;kr'lrrc/\'igilco srstcm (third gcncration sofnvarc):
comparison of nvo monitoring sitcs r-ith thc
tlrcrmr>dilution cardiac ourput../ (l)r -\Iqnit (.rmlrl 2012.\pr;26(2): 1 1 5-2t).
\\'icncr
R-S,
\\llch I I(]. 'lrcnds in rhc
nan irtcn'catheter in the United
?ottl.
1.. 1,\
l,'1 2r)l )7 -f ul 25;29i1({)
usc
of pulmo-
Srates, 1993-
:t23t. lD
1.j(5):99{-11){)2.
OUR LATEST SURVEYS
Toke our physiologit moniloring syslem survey. We wont your thoughls on your physiologic
moniioring syslem's eose of use, functionolity, ond reliobility. Toke o {ew minules to fill out our
user experience survey, ovoiloble through your member home poge or ot hfips://survey.ecri.oq/
Survey.ospx?i=2178o362f1.244O2c9d0159tc11629Occ. When done, you'll be outomoticolly
enfered to win on iPod Mini.
See the resuhs of our CT suiyey. The results of our CT system user experience survey ore
now ovoiloble. Find out how your peers ronked lhe iop vendors ond models on lhe morket in
terms o{ lheir functions, {eolures, ond seruice. View the resulb ot https://members2.ecri.org/
ir
O2013
14 of 158.
Componenls/UserExperience/Poges/ct_sconners/userexp_moin.ospx.
ECRI lnst:tule. Member hospitols moy reproduce ihis poge
for internol distribulion only
w.xri.ors > HEATTHDEVICE5
JUNE2ol3 187
Advance Publication by-J-STAGE
Circulation Journal
Official Journal of the Japanese Circulation Society
http://www. j-circ.or.jp
Noninvasive and Simple Assessment of Cardiac Output
and Pulmonary Vascular Resistance With Whole-Body
Impedance Cardiography Is Useful for Monitoring
Patients With Pulmonary Hypertension
Yu Taniguchi, MD; Noriaki Emoto, MD, PhD; Kazuya Miyagawa, MD, PhD;
Kazuhiko Nakayama, MD, PhD; Hiroto Kinutani, MD; Hidekazu Tanaka, MD, PhD;
Toshiro Shinke, MD, PhD; Ken-ichi Hirata, MD, PhD
Background: Right heart catheterization (RHC) is the gold standard for the diagnosis of pulmonary hypertension
(PH) and a useful tool for monitoring PH. However, there are some disadvantages in the regular use of RHC because
it is invasive. Noninvasive methods for monitoring hemodynamics are needed to manage patients with PH. In this
study, we aimed to evaluate the reliability of noninvasive hemodynamic assessment with whole-body impedance
cardiography (Non-Invasive Cardiac System [NICaS]) for PH.
Methods and Results: We investigated 65 consecutive patients undergoing RHC. Two-thirds of them had pulmonary arterial hypertension and one-third had chronic thromboembolic PH; 25% of the patients were receiving medical therapy. Cardiac output (CO) was estimated by NICaS (NI-CO), thermodilution (TD-CO), and the Fick method
(Fick-CO). There was a strong correlation between NI-CO and TD-CO (r=0.715, P<0.0001) and Fick-CO (r=0.653,
P<0.0001). Noninvasive pulmonary vascular resistance (PVR) was estimated using a conventional invasive equation
with NI-CO, mean pulmonary arterial pressure was calculated by echocardiographic measurement, and pulmonary
capillary wedge pressure was estimated at 10 mmHg in all cases. NICaS-derived PVR was very strongly correlated
with invasive PVR (TD-PVR: r=0.704, P<0.0001; Fick-PVR: r=0.702, P<0.0001).
Conclusions: Noninvasive measurement of CO and PVR using NICaS and echocardiography is a useful tool for
the assessment of PH.
Key Words: Cardiac output; Noninvasive assessment; Pulmonary hypertension; Pulmonary vascular resistance;
Whole-body impedance cardiography
P
ulmonary arterial hypertension (PAH) is a progressive
disease characterized by elevated pulmonary vascular
resistance (PVR) because of pulmonary vascular remodeling. This leads to a decrease in cardiac output (CO) and
ultimately death. Recently, targeted medical therapy for PAH
patients with endothelin-receptor antagonists, phosphodiesterase-5 inhibitors, and prostacyclin analogs has been established,1 and the prognosis of PAH has improved.2 However,
there is no universally accepted consensus on the treatment
goals or follow-up strategy for PAH patients. Right heart catheterization (RHC) is not only the gold standard for the diagnosis of PAH, but is also a useful tool for monitoring PAH, and
is recommended 3–6 months after new treatments and in the
case of clinical worsening.1 Hemodynamic monitoring with
RHC is predictive of survival and effective in a goal-oriented
treatment strategy,3,4 and has been recommended by a recent
guideline;1 however, there are some disadvantages in the regular use of RHC as a follow-up procedure, especially with
regard to invasiveness. Noninvasive and less complicated
methods for monitoring hemodynamics are needed to manage
patients with pulmonary hypertension (PH). Less invasive
hemodynamic monitoring has recently been suggested as feasible in some situations.5 The Non-Invasive Cardiac System
(NICaS; NI Medical, Hod-Hasharon, Israel) is a device for
calculating CO noninvasively with whole-body impedance
cardiography (ICGWB). The NICaS-derived CO (NI-CO) has
been shown to be as reliable as the RHC-derived CO and is
applicable for the noninvasive assessment of cardiac function
Received February 3, 2013; revised manuscript received May 1, 2013; accepted May 2, 2013; released online June 12, 2012 Time for
primary review: 42 days
Division of Cardiovascular Medicine, Department of Internal Medicine, Kobe University Graduate School of Medicine, Kobe (Y.T., N.E.,
K.M., K.N., H.K., H.T., T.S., K.H.); Clinical Pharmacy, Kobe Pharmaceutical University, Kobe (N.E., K.N.), Japan
Mailing address: Noriaki Emoto, MD, PhD, Division of Cardiovascular Medicine, Department of Internal Medicine, Kobe University
Graduate School of Medicine, 7-5-1 Kusunoki, Chuo-ku, Kobe 650-0017, Japan. E-mail: [email protected]
ISSN-1346-9843 doi: 10.1253/circj.CJ-13-0172
All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected]
15 of 158.
TANIGUCHI Y et al.
Table 1. Clinical Characteristics of All Patients at Initial
Hospitalization
Table 2. Hemodynamic Measurements
Age (years)
62±14
Female (%)
39 (65)　TD-CO (L/min)
4.92±1.56
Fick-CO (L/min)
3.87±1.24
Echo-CO (L/min)
4.34±1.11
NI-CO (L/min)
4.40±1.32
Diagnosis (%)
PAH
38 (63)　IPAH
12 (20)　CTD-PAH
24 (40)　Po-PAH
2 (3)　PH associated with respiratory disorders
3 (5)　20 (33)　CTPH
WHO-fc (%)
1
1 (1.7)
2
22 (37.3)
3
32 (54.2)
4
4 (6.8)
Treatment (%)
25 (24)　Bosentan
13 (22)　Sildenafil
14 (23)　Beraprost
10 (17)　Hemodynamic variables
sPAP (mmHg)
53.9±21.3
mPAP (mmHg)
31.7±12.0
RAP (mmHg)
3.7±4.2
PCWP (mmHg)
7.0±4.3
CO (TD) (L/min)
4.90±1.62
CO (Fick) (L/min)
3.92±2.08
PVR (TD) (dyn · s–1 · cm–5)
433±244
PVR (Fick) (dyn · s–1 · cm–5)
581±344
HR (beats/min)
73±11
CO, cardiac output; CTD-PAH, collagen tissue disease associated PAH; CTEPH, chronic thromboembolic pulmonary hypertension; HR, heart rate; IPAH, idiopathic PAH; mPAP, mean PAP;
PAH, pulmonary arterial hypertension; Po-PAH, portal PAH;
PCWP, pulmonary capillary wedge pressure; PVR, pulmonary
vascular resistance; RAP, right atrial pressure; sPAP, systolic
pulmonary arterial pressure; TD, thermodilution; WHO-fc, World
Health Organization functional class.
in patients with left-sided chronic heart failure,6,7 but its feasibility in patients with PH has not been evaluated. The purpose
of this study was to evaluate the reliability of noninvasive
measurement of CO and PVR with ICGWB in patients with PH.
Editorial p ????
Methods
This study was approved by Kobe University Hospital Institutional Review Board and the patients provided written informed
consent to participate.
Patients
We enrolled 65 consecutive patients with known or suspected
pulmonary hypertension hospitalized in Kobe University Hospital from April 2010 to August 2011. All patients who were
scheduled for RHC without fulfilling one of the exclusion
criteria were eligible for the study. The exclusion criteria included restlessness and/or unstable patient condition, severe
aortic valve regurgitation and/or aortic stenosis, aortic aneurysm, heart rate >130 beats/min, intra- and extracardiac shunts,
severe peripheral vascular disease, severe pitting edema, sep-
16 of 158.
Parameter
Value
TD-PVR (dyn · s–1 · cm–5)
446±249
Fick-PVR (dyn · s–1 · cm–5)
583±362
Echo-PVR (dyn · s–1 · cm–5)
660±363
NI-PVR (dyn · s–1 · cm–5)
544±316
Echo-CO, echocardiography-derived cardiac output; Echo-PVR,
echocardiography-derived PVR; Fick-CO, cardiac output derived
by the modified Fick method; Fick-PVR, PVR derived by modified
Fick method; NI-CO, NICaS-derived cardiac output; NI-PVR,
NICaS with echocardiography-derived PVR; PVR, pulmonary
vascular resistance; TD-CO, thermodilution-derived cardiac output;
TD-PVR, thermodilution-derived PVR.
sis, and dialysis, all of which interfere with the accurate measurement of impedance-derived CO with NICaS, as previously described.6 Patients with elevated pulmonary capillary
wedge pressure (PCWP >15 mmHg) on RHC were also excluded; 5 patients were excluded because of the presence of an
intra-cardiac shunt; 29 patients were reevaluated 3–6 months
after new treatments or at clinical worsening.
Hemodynamics
Hemodynamic data were derived from standard RHC in all
patients using a 6Fr Swan-Ganz catheter (Baxter Healthcare,
Irvine, CA, USA). The catheter was introduced into the pulmonary artery under fluoroscopic guidance. Mean pulmonary
arterial pressure (mPAP), systolic and end-diastolic pulmonary
arterial pressure (sPAP and dPAP), mean right atrial pressure,
and PCWP were measured. CO was measured using the following techniques.
Thermodilution-Derived CO (TD-CO) A 5-ml bolus of iced
5% glucose solution was injected 5 times at the same rate. The
results of 3 injections within 15% of their extreme disparity
were averaged to derive the TD-CO value.
Modified Fick method (Fick-CO) Blood samples were obtained from systemic and pulmonary arteries. All samples were
measured for oxygen saturation with the same device (Radiometer ABL 715, Copenhagen, Denmark).
NI-CO These measurements were performed simultaneously with the measurement of TD-CO and Fick-CO during
RHC. The measurement of NI-CO followed the method as
previously reported:6 an alternating electrical current of 1.4 mA
with a 30-kHz frequency is passed through the patient via 2
pairs of tetrapolar electrodes – 1 pair placed on the wrist above
the radial pulse, and the other pair placed on the contralateral
ankle above the posterior tibialis arterial pulse. If the arterial
pulses in the legs are either absent or of poor quality, the second pair of electrodes is placed on the contralateral wrist.
The NICaS apparatus calculates the stroke volume (SV) by
Frinerman’s formula:8
SV = dR/R × ρ × L2/Ri × (α + β)/β × KW × HF
where dR is the impedance change; R is the basal resistance;
ρ is the blood electrical resistivity; L is the patient’s height; Ri
is the corrected basal resistance according to sex and age; KW
is a correction factor for weight according to ideal values; HF
is the hydration factor, which takes into account the body
water composition. α+β is equal to the ECG R-R wave interval
Noninvasive Assessment of PH
Figure 1. Linear correlation analysis between TD-CO and NI-CO (A), Fick-CO and NI-CO (B), TD-CO and Fick-CO (C), TD-CO
and Echo-CO (D), and Fick-CO and Echo-CO (E). TD-CO, thermodilution-derived cardiac output; NI-CO, NICaS-derived cardiac
output; Fick-CO, cardiac output derived by the modified Fick method; Echo-CO, echocardiography-derived cardiac output.
and β is the diastolic time interval. To calculate the CO, SV is
multiplied by the heart rate. Because the NI-CO values are
calculated every 20 s, the average of 3 measurements obtained
consecutively during 60 s of monitoring is considered to be the
NI-CO value for each individual case.
Echocardiography
Echocardiography was performed using a Vivid 5 system and
a 3.5-MHz transducer (GE Vingmed Ultrasound AS, Horten,
Norway). Two-dimensional Doppler examinations were performed in the usual manner. CO was measured by tracing the
left ventricular ejection flow (Echo-CO). Echo-sPAP was estimated from the peak velocity of the tricuspid regurgitation
jet plus estimated right atrial pressure (Echo-RAP).9
17 of 158.
Measurement of PVR
PVR (dyn · s–1 · cm–5) was calculated using RHC from the equation:
PVR = 80 × (mPAP–PCWP)/COTD, Fick (TD-PVR, Fick-PVR).
PVR was also estimated noninvasively using a combination
of NICaS and echocardiography, and by echocardiography
alone.10 Echo-mPAP was calculated as Echo-sPAP×0.61+
2 mmHg, as previously described,11 and noninvasive PVR was
calculated as 80×(Echo-mPAP–PCWP)/COEcho, NI (Echo-PVR,
NI-PVR). PCWP for the calculation of noninvasive PVR was
estimated at 10 mmHg in all cases.10
Statistical Analysis
Quantitative data are presented as mean ± SD. The correlations
TANIGUCHI Y et al.
Figure 2. Bland-Altman plots with mean difference (solid line) ±2SD (dotted line) comparing TD-CO and NI-CO (A), Fick-CO and
NI-CO (B), TD-CO and Echo-CO (C), and Fick-CO and Echo-CO (D). TD-CO, thermodilution-derived cardiac output; NI-CO,
NICaS-derived cardiac output; Fick-CO, cardiac output derived by the modified Fick method; Echo-CO, echocardiography-derived
cardiac output; SD, standard deviation.
among TD-CO, Fick-CO, Echo-CO, and NI-CO and between
Echo-mPAP and mPAP measured by RHC (RHC-mPAP)
were determined by calculating the Spearman’s rank correlation coefficient. P<0.05 was considered to be significant.
Agreement between methods was analyzed by the Bland-Altman method.12 The limits of the agreement were expressed as
the mean ± SD. The 95% confidence intervals (CIs) of the bias
were also calculated. Receiver-operating characteristic (ROC)
curves were generated for the detection of elevated PVR defined as > 240 dyn · s–1 · cm–5 (3 Wood units [WU]). The area
under the curve (AUC), cut-off value, sensitivity, and specificity were estimated by the ROC curves. All statistical analyses
were performed using GraphPad Prism version 5 (GraphPad
Software, La Jolla, CA, USA).
Results
The baseline characteristics of all patients at initial hospitalization are summarized in Table 1. Approximately two-thirds
of the patients had PAH (World Health Organization [WHO]
classification of PH group 1) and the other one-third of the
patients had chronic thromboembolic PH (CTEPH: WHO
group 4); 5% of the patients were classified as WHO group 3.
At enrollment, 24% of the patients were receiving medical
therapy.
Relationships Among Parameters
The mean CO values from all measurements in these sub-
18 of 158.
jects for TD-CO, Fick-CO, Echo-CO, and NI-CO were 4.92±
1.56 L/min, 3.87±1.24 L/min, 4.34±1.11 L/min, and 4.40±
1.32 L/min, respectively (Table 2). A significant and very
strong correlation was observed between TD-CO and NI-CO
(r=0.715, P<0.0001) and between TD-CO and Fick-CO (r=
0.795, P<0.0001) by 2-tailed Spearman’s rank correlation test
(Figure 1). There was a strong correlation between Fick-CO
and NI-CO (r=0.653, P<0.0001). However, the correlation
between Echo-CO and TD-CO or Fick-CO was significant but
not strong (r=0.512 or 0.461, P<0.0001, respectively). The
differences between 2 measurements were plotted according
to the Bland-Altman method (Figure 2). The mean bias and
limits of agreement between TD-CO and NI-CO, Fick-CO and
NI-CO, and TD-CO and Fick-CO were 0.50±1.08 (–1.61 to
2.61) L/min, and –0.54±1.04 (–2.57 to 1.49) L/min, and 1.02±
0.86 (–0.68 to 2.71) L/min, respectively. The limits of agreement between TD-CO and Echo-CO, and Fick-CO and EchoCO were 0.64±1.33 (–1.97 to 3.26) L/min and –0.42±1.18
(–2.73 to 1.88) L/min, respectively. There was no clear difference in the measurements of CO among the patients with idiopathic PAH, collagen tissue disease associated PAH or
CTEPH (Figure S1).
Comparison of Invasive and Noninvasive Measurement of
mPAP and PVR
The mean values of all measurements of invasive mPAP and
Echo-mPAP were 32.9±1.28 mmHg and 43.0±1.59 mmHg, respectively. There was a very strong correlation between inva-
sive mPAP and Echo-mPAP (r=0.703, P<0.0001; Figure 3A).
The limits of agreement between invasive mPAP and EchomPAP were –9.63±10.2 (–29.6 to 10.4) mmHg (Figure 3B).
The mean values of all measurements of TD-PVR, FickPVR, Echo-PVR, and NI-PVR were 446±249 dyn · s–1 · cm–5,
583±362 dyn · s–1 · cm–5, 660±363 dyn · s–1 · cm–5, and 644±316 dyn · s–1 · cm–5, respectively (Table 2). There were significant
and very strong correlations between TD-PVR and NI-PVR
(r=0.704, P<0.0001), between Fick-PVR and NI-PVR (r=0.702,
P<0.0001), and between TD-PVR and Fick-PVR (r=0.942,
P<0.0001) (Figures 4A,D). However, the correlation between
PVR measured by invasive methods and Echo-PVR was
not as strong (r=0.602 or 0.603, P<0.0001, respectively;
Figures 4G,J) as that between invasive methods and NI-PVR.
Figure 4B and dyn · s–1 · cm–5 shows the Bland-Altman plots of
the differences between TD-PVR, Fick-PVR, and NI-PVR.
The limits of agreement between TD-PVR and NI-PVR,
Fick-PVR and NI-PVR, and TD-PVR and Fick-PVR were
–195±265 (–715 to 326) dyn · s–1 · cm–5, –35±325 (–673 to 603)
dyn · s–1 · cm–5, and –135±164 (–457 to 187) dyn · s–1 · cm–5,
respectively. The limits of agreement between TD-PVR and
Echo-PVR, and Fick-PVR and Echo-PVR were –191±266
(–713 to 330) dyn · s–1 · cm–5, and –33±341 (–703 to 635)
dyn · s–1 · cm–5, respectively (Figures 4H,K). The AUC for NIPVR to detect increased PVR >240 dyn · s–1 · cm–5 (3 WU)
against TD and Fick-PVR were 0.84 (95% CI, 0.72–0.96) and
0.92 (95% CI, 0.84–0.99), respectively (Figures 3C,F), and
optimal cut-off values were 411 dyn · s–1 · cm–5 (sensitivity:
81.3%, specificity: 75%) and 400 dyn · s–1 · cm–5 (sensitivity:
80.3%, specificity: 100%), respectively. The AUC for EchoPVR against TD and Fick-PVR were lower: 0.75 (95% CI,
0.57–0.92) and 0.83 (95% CI, 0.66–0.99) (Figures 4I,L) compared with that for NI-PVR against TD and Fick-PVR.
Discussion
Figure 3. Linear correlation analysis between RHC-mPAP
and Echo-mPAP (A). Bland-Altman plots with mean difference
(solid line) ±2SD (dotted line) comparing RHC-mPAP and
Echo-mPAP (B). RHC, right heart catheterization; mPAP,
mean pulmonary arterial pressure; Echo-mPAP, mPAP calculated by echocardiography; SD, standard deviation.
We report on the reliability of a noninvasive and simple method of assessing CO and PVR using ICGWB in patients with PH.
Previous reports have indicated the feasibility of hemodynamic assessment using various methods in comparison with
RHC in a range of clinical settings;13 however, a reliable method for the assessment of PH has not yet been established. In
particular, there are few studies that have addressed the noninvasive assessment of hemodynamics in PH. Hemodynamic
assessment using cardiac magnetic resonance or echocardiography have been shown to be reliable,10,14–17 but these methods
require expensive equipment and trained operators. Thoracic
impedance cardiography has been used for the measurement of
CO noninvasively in PAH,18 and its reliability was shown to be
compromised in cardiac patients in a meta-analysis.19
We demonstrated strong correlations among the NICaS, TD,
and the Fick methods for the measurement of CO. Although
the limits of agreement between NI-CO and TD-CO or FickCO estimated by the Bland-Altman approach were not small,
they were acceptable when compared with previous reports.13
Therefore, we believe that NICaS can be a reliable tool for the
noninvasive assessment of CO in PH. However, compared
with NI-CO, the correlation between Echo-CO and TD-CO or
Fick-CO was weaker and the limits of agreements were larger.
The relative inaccuracy of CO measured by echocardiography
was consistent with a previous report,20 and may be a consequence of using the Doppler method, severe tricuspid regurgitation, and operator-dependency.
We also demonstrated the feasibility of noninvasive and
simple measurement of PVR using a combination of NICaS
and echocardiography. Kouzu et al showed that tricuspid regurgitant pressure gradient (TRPG)/right ventricular timevelocity integral (TVI) is reliable for the estimation of PVR.21
Although TRPG/TVI has been confirmed as a reliable method
for estimating PVR,16 accurate measurement of TVI needs a
skilled operator. Lindqvist et al reported the accuracy of a
simple Doppler-derived measurement of PVR with the conventional invasive equation in patients with PH;10 however,
their study excluded patients with severe tricuspid regurgitation, which causes inaccuracy in CO measurement using echocardiography. In the present study, we estimated PVR using
the conventional invasive equation with the combination of
NI-CO and Echo-mPAP. There was very strong correlation
not only between invasive mPAP and Echo-mPAP, but also
between invasive PVR and NI-PVR. Furthermore, a stronger
correlation of PVR was found for the use of NI-CO compared
with Echo-CO. The limits of agreements estimated by the
Bland-Altman analysis were large, but comparable with previous reports that showed the feasibility of PVR derived by
echocardiography against invasive PVR.21,22 Furthermore, the
high AUC, sensitivity, and specificity for NI-PVR to detect
increased PVR >240 dyn · s–1 · cm–5 (3 WU) also indicates the
reliability of noninvasive PVR assessment using NICaS.
In our study, the value for TD-CO was significantly higher
than the CO values with other methods, including Fick-CO,
and therefore, the value of TD-PVR was underestimated. This
19 of 158.
TANIGUCHI Y et al.
Figure 4. Diagnostic reliability of NI-PVR compared with invasive PVR. Linear correlation analysis between TD-PVR and NI-PVR
(A), Fick-PVR and NI-PVR (D), TD-PVR and Echo-PVR (G), and Fick-PVR and Echo-PVR (J). Bland-Altman plot with mean difference (solid line) ±2SD (dotted line) comparing TD-PVR and NI-PVR (B), Fick-PVR and NI-PVR (E), TD-PVR and NI-PVR (H), and
Fick-PVR and NI-PVR (K). Area under the receiver-operating characteristics curve with 95% confidence interval for NI-PVR and
Echo-PVR to detect increased PVR (>240 dyn · s–1 · cm–5 [3 WU]) against TD-PVR (C,I) and Fick-PVR (F,L). PVR, pulmonary vascular resistance; NI-PVR, NICaS with echocardiography-derived PVR; TD-PVR, thermodiluyion-derived PVR; Fick-PVR, PVR derived
by the modified Fick method; Echo-PVR, echocardiography-derived PVR.
could be caused by overestimation of the value of TD-CO in
the presence of low CO, consistent with previous reports.23
Study Limitations
The main limitation of this study was the need to measure the
Doppler parameter for estimating NI-PVR. Proper alignment
of the ultrasound beam is crucial for the Doppler parameter to
20 of 158.
be determined appropriately. This may have resulted in bias in
the measurement of NI-PVR. In our study, the Doppler parameter needed in order to estimate NI-PVR was only TRPG, and
there was no patient in whom we were unable to obtain that
value. Second, we used the conventional invasive equation for
estimating NI-PVR. We had to estimate PCWP at 10 mmHg
in all cases as previously reported,10 which may also have re-
sulted in the measurement of NI-PVR; however, in general, a
wide variation in PCWP is not usually observed among patients with PH. Third, because CO measurement using NICaS
in patients with cardiac shunts is known to be unreliable,24 we
excluded cases of PAH associated with cardiac shunts. Fourth,
noninvasive estimation of CO and PVR with NICaS was feasible; however, there were some patients who had large divergence between NI-CO or NI-PVR and invasive CO or PVR.
Further studies are needed to clarify the factors that lead to
inaccurate measurements of CO and PVR. Fifth, in our study,
the number of patients with WHO functional class 4 was
small. Most patients were WHO functional class 2 or 3. The
reliability of NICaS in patients with severe PH is to be examined in future studies. Finally, the study sample size was relatively small and originated from a single center. We believe
that a larger, multicenter study is needed to appropriately
confirm the reliability of the method.
Although recent advances in treatment options and management have improved the outcomes for patients with PH, treatment goals and follow-up strategy are still not well defined.
Hemodynamic monitoring with RHC is recommended in a
goal-oriented treatment strategy for PH1 and is the gold standard for the assessment of PAH; however, the invasiveness of
RHC is a critical factoring its regular use as a follow-up procedure. A noninvasive, accurate, and simple method is required for the management of patients with PH. We have
demonstrated noninvasive measurement of CO and PVR using
only simple parameters. Echo-sPAP is needed to estimate
PVR, but Echo-sPAP has been established as a simple, reliable screening parameter for PH.25 This noninvasive, reliable,
and simple assessment can be a useful tool for monitoring and
managing patients with PH.
Conclusion
Noninvasive measurement of CO and PVR using NICaS is as
reliable as invasive RHC. This simple assessment could help
physicians to manage their patients with PH.
Disclosures
None.
References
1. Galie N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, Barbera
JA, et al. Guidelines for the diagnosis and treatment of pulmonary
hypertension: The Task Force for the Diagnosis and Treatment of
Pulmonary Hypertension of the European Society of Cardiology (ESC)
and the European Respiratory Society (ERS), endorsed by the International Society of Heart and Lung Transplantation (ISHLT). Eur
Heart J 2009; 30: 2493 – 2537.
2. Benza RL, Miller DP, Barst RJ, Badesch DB, Frost AE, McGoon
MD. An evaluation of long-term survival from time of diagnosis in
pulmonary arterial hypertension from the REVEAL Registry. Chest
2012; 142: 448 – 456.
3. Fukumoto Y, Shimokawa H. Recent progress in the management of
pulmonary hypertension. Circ J 2011; 75: 1801 – 1810.
4. Hoeper MM, Markevych I, Spiekerkoetter E, Welte T, Niedermeyer
J. Goal-oriented treatment and combination therapy for pulmonary
arterial hypertension. Eur Respir J 2005; 26: 858 – 863.
5. Alhashemi JA, Cecconi M, Hofer CK. Cardiac output monitoring: An
integrative perspective. Crit Care 2011; 15: 214.
6. Paredes OL, Shite J, Shinke T, Watanabe S, Otake H, Matsumoto D,
et al. Impedance cardiography for cardiac output estimation: Reliability of wrist-to-ankle electrode configuration. Circ J 2006; 70: 1164 – 1168.
7. Tanino Y, Shite J, Paredes OL, Shinke T, Ogasawara D, Sawada T, et
al. Whole body bioimpedance monitoring for outpatient chronic heart
failure follow up. Circ J 2009; 73: 1074 – 1079.
21 of 158.
8. Cohen AJ, Arnaudov D, Zabeeda D, Schultheis L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output during
coronary artery bypass grafting. Eur J Cardiothorac Surg 1998; 14:
64 – 69.
9. Yeo TC, Dujardin KS, Tei C, Mahoney DW, McGoon MD, Seward
JB. Value of a Doppler-derived index combining systolic and diastolic time intervals in predicting outcome in primary pulmonary
hypertension. Am J Cardiol 1998; 81: 1157 – 1161.
10. Lindqvist P, Soderberg S, Gonzalez MC, Tossavainen E, Henein MY.
Echocardiography based estimation of pulmonary vascular resistance
in patients with pulmonary hypertension: A simultaneous Doppler
echocardiography and cardiac catheterization study. Eur J Echocardiogr 2011; 12: 961 – 966.
11. Chemla D, Castelain V, Humbert M, Hebert JL, Simonneau G,
Lecarpentier Y, et al. New formula for predicting mean pulmonary
artery pressure using systolic pulmonary artery pressure. Chest 2004;
126: 1313 – 1317.
12. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; 1: 307 – 310.
13. Mantha S, Roizen MF, Fleisher LA, Thisted R, Foss J. Comparing
methods of clinical measurement: Reporting standards for bland and
altman analysis. Anesth Analg 2000; 90: 593 – 602.
14. Inaba T, Yao A, Nakao T, Hatano M, Maki H, Imamura T, et al.
Volumetric and functional assessment of ventricles in pulmonary
hypertension on 3-dimensional echocardiography. Circ J 2012; 77:
198 – 206.
15. Garcia-Alvarez A, Fernandez-Friera L, Mirelis JG, Sawit S, Nair A,
Kallman J, et al. Non-invasive estimation of pulmonary vascular resistance with cardiac magnetic resonance. Eur Heart J 2011; 32: 2438 – 2445.
16. Abbas AE, Fortuin FD, Schiller NB, Appleton CP, Moreno CA, Lester
SJ. A simple method for noninvasive estimation of pulmonary vascular resistance. J Am Coll Cardiol 2003; 41: 1021 – 1027.
17. Kang KW, Chang HJ, Kim YJ, Choi BW, Lee HS, Yang WI, et al.
Cardiac magnetic resonance imaging-derived pulmonary artery distensibility index correlates with pulmonary artery stiffness and predicts functional capacity in patients with pulmonary arterial hypertension. Circ J 2011; 75: 2244 – 2251.
18. Yung GL, Fedullo PF, Kinninger K, Johnson W, Channick RN. Comparison of impedance cardiography to direct Fick and thermodilution
cardiac output determination in pulmonary arterial hypertension. Congest Heart Fail 2004; 10: 7 – 10.
19. Raaijmakers E, Faes TJ, Scholten RJ, Goovaerts HG, Heethaar RM.
A meta-analysis of published studies concerning the validity of thoracic impedance cardiography. Ann NY Acad Sci 1999; 873: 121 – 127.
20. Fisher MR, Forfia PR, Chamera E, Housten-Harris T, Champion HC,
Girgis RE, et al. Accuracy of Doppler echocardiography in the hemodynamic assessment of pulmonary hypertension. Am J Respir Crit
Care Med 2009; 179: 615 – 621.
21. Kouzu H, Nakatani S, Kyotani S, Kanzaki H, Nakanishi N, Kitakaze
M. Noninvasive estimation of pulmonary vascular resistance by Doppler echocardiography in patients with pulmonary arterial hypertension. Am J Cardiol 2009; 103: 872 – 876.
22. Selimovic N, Rundqvist B, Bergh CH, Andersson B, Petersson S,
Johansson L, et al. Assessment of pulmonary vascular resistance by
Doppler echocardiography in patients with pulmonary arterial hypertension. J Heart Lung Transplant 2007; 26: 927 – 934.
23. Tournadre JP, Chassard D, Muchada R. Overestimation of low cardiac output measured by thermodilution. Br J Anaesth 1997; 79: 514 – 516.
24. Kauppinen PK, Koobi T, Hyttinen J, Malmivuo J. Segmental composition of whole-body impedance cardiogram estimated by computer simulations and clinical experiments. Clin Physiol 2000; 20:
106 – 113.
25. Taleb M, Khuder S, Tinkel J, Khouri SJ. The diagnostic accuracy of
Doppler echocardiography in assessment of pulmonary artery systolic
pressure: A meta-analysis. Echocardiography 2013; 30: 258 – 265.
Supplementary Files
Supplementary File 1
Figure S1. Linear correlation between NI-CO and TD-CO (A) or
Fick-CO (B) in IPAH ( , red line), CTD-PAH ( , blue line) and
CTEPH ( , green line).
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-13-0172
Perfusion
http://prf.sagepub.com/
Cardiac power index: staging heart failure for mechanical circulatory support
SG Hall, J Garcia, DF Larson and R Smith
Perfusion 2012 27: 456 originally published online 13 June 2012
DOI: 10.1177/0267659112450933
The online version of this article can be found at:
http://prf.sagepub.com/content/27/6/456
Published by:
http://www.sagepublications.com
Additional services and information for Perfusion can be found at:
Email Alerts: http://prf.sagepub.com/cgi/alerts
Subscriptions: http://prf.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Nov 1, 2012
OnlineFirst Version of Record - Jun 13, 2012
What is This?
22 of 158.
Downloaded from prf.sagepub.com at Tel Aviv University on September 8, 2013
7659112450933Hall SG et al.Perfusion
Cardiac power index: staging heart failure for
mechanical circulatory support
Perfusion
27(6) 456–461
© The Author(s) 2012
Reprints and permission: sagepub.
co.uk/journalsPermissions.nav
DOI: 10.1177/0267659112450933
prf.sagepub.com
SG Hall, J Garcia, DF Larson and R Smith
Abstract
Cardiac power output has been shown to quantify cardiac reserve. Cardiac reserve is defined as the difference between
basal and maximal cardiac performance. We compared cardiac power index to other commonly used hemodynamic
parameters to validate its usefulness to stage heart failure patients and determine the optimal time for implantation of
mechanical circulatory support. A retrospective study of twenty-eight heart failure patients implanted with mechanical
circulatory support was analyzed at three levels of drug therapy. Subjects were further separated into two categories:
survived versus deceased. Cardiac power index was the only statistically significant hemodynamic parameter that identified
cardiac reserve (p<0.05) in this patient population. These results showed that a cardiac power index at or below 0.34
Watts/m2 resulted in increased mortality rate, ninety days post-implantation.
Conclusion: Cardiac reserve was a determinant of post-device survival; therefore, these data suggest that device implantation should occur prior to the 0.34 Watts/m2 threshold.
Keywords
cardiac power output; cardiac reserve; mechanical circulatory support; hemodynamics; cardiac power index
Introduction
Cardiac power output is the hydraulic energy required by the
heart to provide enough blood flow to the systemic circulation. Cardiac power output measured at exercise reflects
the peak cardiac performance attainable by the heart.
Comparing maximum cardiac power output with resting
cardiac power output represents the heart’s cardiac reserve.
In heart failure patients, cardiac reserve is severely limited.1
Patients performing exercise tests or positive inotrope infusion tests tend to show little difference in cardiac performance due to poor cardiac reserve. It has been demonstrated
in several studies that heart failure patients with higher cardiac reserves have better survival rates. As heart failure progresses in these patients, cardiac reserve continues to decline.2
Once cardiac performance can no longer be improved
through inotropic drug therapy, heart failure patients may
require mechanical circulatory support. Because end-organ
damage is a predictor of poor prognosis in patients who
receive mechanical circulatory support, it is important to
implant devices before heart failure has become so severe
that irreparable systemic damage has occurred. Symptomatic
assessment, drug tolerance, and myocardial oxygen consumption are commonly used to determine the severity
of heart failure, but are unable to provide all the information necessary to reflect true cardiac reserve. Myocardial
oxygen consumption provides an indirect measurement of
cardiac output; however, given that cardiac power output
23 of 158.
incorporates cardiac output as well as systemic pressures,
this distinguishes it as a rational marker of cardiac performance when compared to myocardial oxygen consumption. Cardiac power output has been used to assess cardiac
performance and prognosis in cardiogenic shock and acute
and chronic heart failure patients.3,4 The purpose of this
study was to retrospectively analyze the application of cardiac power output to stage the severity of heart failure in
patients who have received mechanical circulatory support.
Methods
Twenty-eight patients, twenty-four male and four female,
with a history of heart failure were referred to the
Circulatory Sciences Graduate Perfusion Program, The University of
Arizona, Tucson, Arizona, USA
Corresponding author:
Doug F. Larson
4402 Arizona Health Sciences Center
1501 North Campbell Avenue
Tucson, AZ 85724, USA
Email: [email protected]
Presented at the 33rd Annual Seminar of The American Academy of
Cardiovascular Perfusion, New Orleans, Louisiana, 26-29 January 2012
457
Hall SG et al.
University of Arizona Medical Center for mechanical circulatory support. Patients were admitted due to increased
heart failure symptoms or were medically transferred
from outside institutions. The studied population consisted of New York Heart Association (NYHA) classes III
and IV and American College of Cardiology (ACCAHA) stages C and D. Patient data was taken from July
2008 to April 2011. Approval of the data review was
granted by the University of Arizona Medical Center,
Tucson and the Institutional Committee on Human
Research.
Patient data was collected at three levels of drug therapy from right heart catheterizations, Swan-Ganz catheters, and/or echocardiography exams. Hemodynamic
data was initially collected when patients were first admitted to the intensive care unit and before they received inotropic drug therapy. Patients were then assessed when
they were first placed on inotropic support. The initial
inotropic dose was documented within four hours of
administration. The most common inotropic drugs prescribed included one or more of the following: dobutamine, milrinone, and dopamine. Throughout the
patients’ admission, medical management was adjusted
based on declining heart function and/or drug tolerance.
The third level of drug therapy assessed was considered
the peak inotropic dose prescribed where current inotropic doses were increased or inotropic type was changed
before the patients became non-responsive to medication
and required mechanical circulatory support. Patients
were further subdivided into two group: survived versus
deceased. Patients in the survived group survived at least
ninety days after receiving mechanical circulatory support. The deceased group consisted of patients who died
within ninety days of receiving mechanical circulatory
support. Survival was assessed at ninety days to account
for delayed organ failure due to prolonged hypoperfusion.
Six hemodynamic parameters were compared to
assess the remaining cardiac reserve in the twenty-eight
heart failure patients studied: heart rate(HR) – beats per
minute (bpm); mean arterial pressure – mmHg, [(2 x
diastolic) + systolic]/3; stroke volume – ml, (cardiac
output(CO)/HR); ejection Fraction - %, (stroke volume
(SV)/end-diastolic volume (EDV)) x 100; cardiac index
(CI) - L/min/m2, [(HR x SV)/body surface area (BSA)];
and cardiac power index (CPI) - W/m2, [(MAP x CO) x
0.0022]/BSA.
hemodynamic standards were tested to determine which
hemodynamic parameter gave a better representation of
the remaining contractile reserve. All data are presented as
mean ± SEM.
Results
Of the twenty-eight patients studied, twenty-one survived and seven died within ninety days of receiving an
implanted device. Six of the seven patients died from
end-organ failure and one died from systemic inflammatory response syndrome within six months of receiving
mechanical circulatory support. The twenty-one patients
who survived ninety days from receiving mechanical circulatory support were classified as “destination therapy”
or “bridge to transplant” patients.
Baseline hemodynamic values, seen in Table 1, indicated abnormal values for all subjects, with the exception
of heart rate and mean arterial pressure. Baseline values
were recorded as patients were on standard heart failure
medications (ß-blocker, diuretics, angiotensin-convertingenzyme (ACE)-inhibitors, and angiotensin receptor
blockers (ARBs)). The other hemodynamic parameters
(ejection fraction, left ventricular stroke work index, cardiac output, cardiac index, and cardiac power) were all
below healthy adult ranges. Table 2 displays documented
mean inotropic therapy while patients were in the intensive care unit.
Heart rate and mean arterial pressure (Figure 1A &
1B) were not found to be statistically significant between
the “survived” and “deceased” groups. Heart rate increased
Table 1. Baseline patient characteristics
Characteristic (n=28)
Datum
Statistics
Age (year)
BSA (m2)
HF Etiology
Ischemic
Non-Ischemic
ICD
HR (BPM)
MAP (mmHg)
SVR (dyn·s/cm5)
CVP (mmHg)
PCWP (mmHg)
LVSWI (g/m2/beat)
EF (%)
CO (L/min)
53 ± 2.5
2.0 ± 0.03
13 (46%)
15 (54%)
19 (68%)
81 ± 7.5
82 ± 4.6
1724 ± 197
21 ± 2.1
29 ± 1.5
15 ± 1.92
22 ± 3.5
3.14 ± 0.3
Statistical analysis was done using Data Analysis and
Statistical Software Stata 11 (StataCorp LP, College Station,
TX, USA). An independent two-sample t-test with unequal
variance was used to identify if contractile reserve still
existed in the failing heart. Mechanical circulatory support
Abbreviations. BSA: body surface area; ICD: implantable cardioverter defibrillator; HF: heart failure; HR: heart rate; MAP: mean arterial pressure;
SVR: systemic vascular resistance; CVP: central venous pressure; PCWP:
pulmonary capillary wedge pressure; LVSWI: left ventricular stroke work
index; EF: ejection fraction; CO: cardiac output.Values are expressed as
mean ± SEM or number (% within studied population).
24 of 158.
458
Perfusion 27(6)
Table 2. Administered inotropes
Positive Inotrope
Initial Dose (mcg/kg/min)
Peak Dose (mcg/kg/min)
P-value
Dobutamine
Milrinone
Dopamine
4.7
0.47
4.5
6.0
3.23
3.78
>0.05
>0.05
>0.05
Inotropic doses are expressed as mean values for the two levels of inotropic drug therapies studied.
in both the “survived” and “deceased” groups when compared across all three levels of drug therapy. Mean arterial
pressure of the “deceased” group declined despite inotropic drug intervention. Those in the “survived” group
maintained a consistent level of mean arterial pressure
across all three levels of drug therapy.
Stroke volume was well below the healthy adult ranges
(Figure 1D), but was not found to be statistically significant between the “survived” and “deceased” groups.
Stroke volume did not increase in the “deceased” group
after the initial inotropic drug therapy; however, when
peak inotropic drug therapy was administered, stroke
volume decreased. On the other hand, the “survived”
group displayed an increase in stroke volume at each
level of drug therapy.
Left ventricular ejection fraction (Figure 1C) derived
from echocardiography was not significantly different
between the “survived” or “deceased” groups. When
placed on the initial inotropic treatment, both the “survived” and “deceased” groups had an increased ejection
fraction; however, the “deceased” was less (31.50 ± 7.5%
to 22.76 ± 13.17 %). There was a large variance in the
“deceased” group due to a select few having diastolic
heart failure with preserved ejection fraction. Those with
diastolic heart failure had poor cardiac performance, but
because their chamber size was unchanged or even
decreased despite ventricular wall thickening, the overall
percentage of blood ejected from the heart with each
contraction was still elevated due to a decrease in overall
ventricular chamber size.
Cardiac output was used to calculate cardiac index in
order to account for the distribution of blood across the
body surface area. In Figure 1E, the “survived” and
“deceased” patients presented with a cardiac index of
1.65 ± 0.83 and 1.52 ± 0.13 L/min/m2, respectively. With
the initial inotropic treatment, the “survived” group’s
cardiac index increased to 2.14 ± 0.54 L/min/m2, but
remained constant with the peak inotropic treatment at
2.17 ± 0.47 L/min/m2. The “deceased” group displayed a
moderate increase in cardiac index during the initial inotropic treatment (1.86 ± 0.62 L/min/m2), however, this
decreased on peak inotropic support (1.69 ± 0.26 L/min/
m2). Cardiac index was not a functional hemodynamic
marker to identify if contractile reserve remained or
whether it was able to detect survival outcomes between
the survived and deceased groups.
25 of 158.
Cardiac power output was used to calculate cardiac
power index to adjust for blood volume distribution
across the body surface area. The baseline administered
inotropic dose increased cardiac power index, but
revealed no statistical difference when compared to the
deceased group (Figure 1F). When peak inotropic doses
were administered, cardiac power index continued to
decline in the deceased group, but was able to distinguish the contractile reserve between the two groups.
The mean cardiac power index of patients who survived
was 0.34 ± 0.07 W/m2 at peak inotropic drug therapy. A
cardiac power index below 0.34 W/m2 identified patients
who did not survive ninety days past the implantation
of mechanical circulatory support because all patients
in the deceased group had a cardiac power index below
0.34 W/m2. Comparing cardiac power index between
the survived and deceased groups showed significantly
increased mortality rates once values fell below this
point (p<0.05).
Discussion
Current hemodynamic guidelines use left ventricular
ejection fraction and myocardial oxygen consumption as
criteria to assess cardiac and systemic degeneration for
possible mechanical circulatory support. As shown by
Marmor et al., the present study demonstrates that ejection fraction did not display any significant differences
between the “survived” and “deceased” groups during
maximal cardiac performance resulting from peak inotropic drug therapy.9 In addition, results from prior
myocardial oxygen consumption studies on patients in
end-stage heart failure have shown to be inconclusive
when peak cardiac function was measured.3,10,11 Patients
with NYHA classes III and IV symptoms are severely
compromised when tested physically for myocardial oxygen consumption due to progressive muscle deconditioning, poor vascular circulation, compromised lung
capacity, and anemia. Although ejection fraction and
myocardial oxygen consumption tests have their respective roles in measuring specific parameters of heart function, they may not be completely reliable assessments of
cardiac function in these severely compromised patients.
Right heart catheterization provides very useful information in gathering pressure and volume changes in specific areas of the heart. This method requires that patients
459
Hall SG et al.
Figure 1. Hemodynamic parameters of the survived and deceased.
*Normal hemodynamic values are from reference numbers 5-8. Values are expressed as mean ± standard error of the mean. *p=0.009.
be placed in a supine position without simulating peak
cardiac contractility. This is a significant limitation, as
assessing heart function in this manner does not allow for
measuring the difference between basal and maximal cardiac performance in order to determine true cardiac
26 of 158.
reserve. Resting values alone do not provide a sufficient
amount of information on cardiac performance in healthy
versus diseased hearts; however, positive inotropic stimulation that achieves peak contractility is when diseased
hearts can become distinguishable from healthy hearts.
460
Perfusion 27(6)
With the limited physical and physiological abilities of
this population, an effective method to assess true cardiac
reserve is through the use of stress tests using inotropic
drug infusion. This method has proven to be equivalent
to results produced by maximal exercise tests.9,12 In
severely compromised patients, this may be the only
practical method to use in order to determine existing
contractile energy. The current study used this concept by
gathering data while patients were placed on initial inotropic support in order to stimulate the remaining contractile reserve, if any existed. As heart failure progressed
in these patients and inotropic support was increased or
altered, therapeutic saturation was met, eventually, due to
the heart already functioning at its maximum contractile
capacity. This indicated a lack of existing reserve in these
patients, due to their basal cardiac performance equaling
their maximum.
Cardiac power index was the only hemodynamic
parameter that distinguished between the cardiac reserve
of patients who survived ninety days post-implant versus
those who died. Although invasive inotropic stimulation
may not be the first choice to assess cardiac contractile
reserve, measuring peak cardiac power index invasively
may provide information on whether pharmacological
support is still efficacious and if any contractile reserve
still remains. Non-invasive methods to evaluate cardiac
power output have been used for over five years and have
been shown to effectively determine the severity of heart
failure.3,8,10,13-16 With the availability of non-invasive
equipment utilizing an inert gas re-breathing technique,
patients in earlier stages of heart failure can be readily
assessed for existing cardiac reserve.17 Cardiac power
index can be a useful prognostic tool in order to determine if cardiac reserve exists in these patients and when
to intervene with pharmacological support or mechanical
circulatory support before the systemic effects of heart
failure become irreparable.
Limitations
Because this was a single-center study, our data were
non-randomized and our cohort of patients was relatively small. Due to the novelty of the population, having a large number of subjects is a difficult limitation to
overcome when performing studies on patients receiving mechanical circulatory support. The variation in
medical record charting between patients also limited
the number of hemodynamic values that could be evaluated. Subject data were collected either by right heart
catheterizations, Swan-Ganz catheters, and/or echocardiography exams, but these methods were inconsistent
across the sample population. A future prospective
study would be beneficial to reduce the lack of consistency between available subject data and their collection
methods.
27 of 158.
Conclusions
In conclusion, our data suggest that attaining cardiac
index and mean arterial pressure, either invasively or
non-invasively, and applying these values to the cardiac
power index formula, can determine the severity of heart
failure based on the heart’s maximal contractile reserve.
Peak cardiac power index values may also be a useful tool
to determine the appropriate therapeutic approach necessary for these patients. Cardiac power index has proven
to recognize remaining contractile reserve in heart failure patients in comparison to other commonly used
hemodynamic parameters. In heart failure patients categorized as NYHA classes III and IV or ACC-AHA stages
C and D, a peak cardiac power index below 0.34 W/m2,
with or without the use of drug therapy should immediately be considered for implantation of mechanical circulatory support. Patients at or below a cardiac power
index of 0.34 W/m2 have an increased incidence of mortality.
Acknowledgement
This study was supported the Steinbronn Heart Failure Research
Award to DFL.
Conflict of Interest Statement
None declared.
References:
1. Williams SG, Barker D, Goldspink DF, Tan LB. A reappraisal of concepts in heart failure: central role of cardiac
power reserve. Arch Med Sci 2005; 1: 65–74.
2. Cotter G, Williams SG, Z Vered Z, Tan LB. Role of cardiac power in heart failure. Curr Opin Cardiol 2003; 18:
215–222.
3. Lang CC, Karlin P, Haythe J, Lim TK, Mancini DM. Peak
cardiac power output, measured noninvasively is a powerful predictor of outcome in chronic heart failure. Circ
Heart Fail 2009; 2: 33–38.
4. Cotter G, Moshkovitz Y, Kaluski E, et al. The role of cardiac power and systemic vascular resistance in the pathophysiology and diagnosis of patients with acute congestive
heart failure. Eur J Heart Fail 2003; 5: 443–451.
5. Pfisterer ME, Battler A, Zaret BL. Range of normal values
for left and right ventricular ejection fraction at rest and
during exercise assessed by radionuclide angiocardiography. Eur Heart J 1985; 6: 647–655.
6. Mayer SA, Lin J, Homma S, et al. Myocardial injury and
left ventricular performance after subarachnoid hemorrhage. Stroke 1999; 30: 780–786.
7. Parry, Andrew. Inotropic drugs and their uses in critical
care. Nurs Crit Care 2012; 17: 19–27.
8. Bromley PD, Hodges LD, Brodie DA. Physiological range
of peak cardiac power output in healthy adults. Clin
Physiol Funct Imaging 2006; 26: 240–246.
9. Marmor A, Jain D, Cohen LS, Nevo E, Wackers FJ,
Zaret BL. Left ventricular peak power during exercise:
461
Hall SG et al.
a noninvasive approach for assessment of contractile reserve.
J Nucl Med 1993; 34: 1877–1885.
10. Leitman M, Sucher E, Kaluski E, et al. Non-invasive measurement of cardiac output by whole-body bio-impedance
during dobutamine stress echocardiography: clinical
implications in patients with left ventricular dysfunction
and ischaemia. Eur J Heart Fail 2006; 8: 136–140.
11.Wilson JR, Rayos G, Yeoh TK, Gothard P, Bak K.
Dissociation between exertional symptoms and circulatory function in patients with heart failure. Circulation
1995; 92: 47–53.
12. Marmor A, Schneeweiss A. Prognostic value of noninvasively obtained left ventricular contractile reserve in
patients with severe heart failure. J Am Coll Cardiol 1997;
29: 422–428.
13. Jakovljevic DG, George RS, Donovan G, et al. Comparison
of cardiac power output and exercise performance in
patients with left ventricular assist devices, explanted
28 of 158.
14.
15.
16.
17.
(recovered) patients, and those with moderate to severe
heart failure. Am J Cardiol 2010; 105: 1780–1785.
Jakovljevic DG, Donovan G, Nunan D, et al. The effect of
aerobic versus resistance exercise training on peak cardiac power output and physical functional capacity in
patients with chronic heart failure. Int J Cardiol 2010; 145:
526–528.
Shelton RJ, Ingle L, Rigby AS, Witte KK, Cleland JG, Clark
AL. Cardiac output does not limit submaximal exercise
capacity in patients with chronic heart failure. Eur J Heart
Fail 2010; 12: 983–989.
Nicholls D, O'Dochartaigh C, Riley M. Circulatory power-a new perspective on an old friend. Eur Heart J 2002; 23:
1242–1245.
Agostoni P, Cattadori G, Apostolo A, et al. Noninvasive
measurement of cardiac output during exercise by inert
gas rebreathing technique: a new tool for heart failure
evaluation. J Am Coll Cardiol 2005; 46: 1779–1781.
Non invasive measurements of Cardiac Output (CO) and Cardiac Power
Index (CPI) by whole body bio impedance in patients with heart failure. A
report from SICA- HF study (FP7/2007-2013/241558)
1
1
1
1
1
1
1
1
P Pellicori , E Wright , P Costanzo , S Smith , S Rimmer , J Hobkirk , A Torabi , T Mabote , J
1
1 1
Warden , JGF Cleland , University of Hull, Department of Academic Cardiology - Hull - United
Kingdom,





29 of 158.
Objectives: Haemodynamic dysfunction is often used as part of the definition for heart
failure (HF), predicts an adverse outcome and could be an important target for therapy
but is rarely measured in routine practice, perhaps because simple, effective, inexpensive
technology is lacking. We assessed the ability of whole body electrical bioimpedance to
measure haemodynamics non-invasively.
Methods: Patients and controls enrolled in the Studies Investigating Co-morbidities
Aggravating Heart Failure (SICA) were included in this analysis if in sinus rhythm. Stroke
volume (SV), cardiac output (CO), cardiac power index (CPI) and total body water (TBW)
were measured non-invasively using whole body bio-impedance (NICaS) after two
minutes rest in the supine position. Results were compared amongst HF patients
according to tercile of amino-terminal pro-brain natriuretic peptide (NT-proBNP) and left
ventricular ejection fraction (LVEF).
Results: The median age of the 51 patients with HF was 71 years (IQR:63--77), 15 were
women (29%), median LVEF was 36% (IQR:29-44) and median body mass index (BMI)
2
was 30kg/m (IQR:26-33). The median age of 15 control subjects was 66 years (IQR:562
75), 4 were women (27%) and median BMI was 29kg/m (IQR:26-37). As expected,
compared with controls, patients with HF had higher plasma NTproBNP, worse renal
2
2
function, lower CPI (0.64±0.25w/m vs. 0.47±0.18w/m ;p< 0.01) but TBW was similar
(47±9% vs. 46±8%;p=0.74).
Patients were divided into terciles of NTproBNP (lower and upper limits of the mid-tercile
were 373 and 1063ng/L). Patients in the highest tercile of NTproBNP had lower BMI
2
2
2
(32±5kg/m vs. 29±5kg/m vs.28±4kg/m respectively ; p=0.02), LVEF (42±5% vs. 37±9%
2
2
2
vs. 29±8%; p<0.01), CPI (0.52±0.2w/m vs. 0.50±0.2w/m vs. 0.38±0.1w/m ; p=0.04) and
poorer renal function (Creatinine: 91±28μmol/l vs. 113±35μmol/l vs. 130±47μmol/l;
p=0.02) and increased TBW (41±6% vs. 47±7% vs. 48±9%;p=0.02). No differences in
CO or SV were found. In the lowest tercile of LVEF, SV (83±24ml vs. 67±23ml
vs.58±9ml;p<0.01), CO (5.7±1.9l/min vs. 4.3±1.6l/min vs. 3.9±0.8l/min; p <0.01) and CPI
2
2
2
(0.58±0.2W/m vs. 0.45± 0.2W/m vs. 0.38±0.1W/m ; p<0.01) were all significantly lower,
but TBW was similar across terciles (43±6% vs. 46±7% vs. 48±9%; p= 0.28).
Conclusion: In patients with HF and sinus rhythm, whole body bio-impedance might be
a useful method of monitoring the haemodynamic severity of heart failure that is quick,
simple and inexpensive. Whether it is as or more useful than NTproBNP as a marker of
outcome awaits the results of large long-term studies. (European Journal of Heart Failure
Supplements ( 2012 ) 11 ( S1 ), S91)
IJCA-13355; No of Pages 3
International Journal of Cardiology xxx (2011) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Cardiology
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j c a r d
Letter to the Editor
Detection of left ventricular systolic dysfunction using a newly developed, laptop
based, impedance cardiographic index
Yoseph Rozenman a,⁎,1, Renee Rotzak a, Robert P. Patterson b,1
a
b
Heart Institute, E. Wolfson Medical Center, Holon, Israel, Affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, Israel
Medical School, University of Minnesota, USA
a r t i c l e
i n f o
Article history:
Received 8 February 2011
Accepted 10 February 2011
Available online xxxx
Keywords:
Impedance Cardiography
Congestive heart failure
Systolic dysfunction
Screening
Laptop
Congestive heart failure (CHF) is a major cause of morbidity and
mortality worldwide [1]. The development of left ventricular systolic
dysfunction (LVSD) is a marker of poor prognosis. Mild reduction in EF
progresses with time and when EF gets to b40%, most patients
develop CHF and their prognosis is dismal [2,3]. Simple, cheap
techniques that will reliably detect mild reduction of the EF, typical for
ALVSD, will thus be of enormous value to reduce the occurrence of
CHF and cardiac mortality [3–7].
Impedance Cardiography (ICG) is a noninvasive method of
determining hemodynamic status and reliable estimates of myocardial contractility (and EF) can be obtained using indices based on
systolic time intervals [8–10]. We have recently developed an index—
Granov Goor Index (GGI)—that combines time interval and impedance parameters in order to identify subjects with LVSD. The GGI is
obtained from a regional ICG signal—measured using wrist and ankle
electrodes [10–13].
In this manuscript we determined the accuracy of GGI in
identifying subjects with LVSD (in a “training” cohort of 100
individuals) and validated the findings in an additional cohort of
201 subjects.
⁎ Corresponding author at: Cardiology Department, E. Wolfson Medical Center, POB
5, Holon, 58100, Israel. Tel.: +972 3 5028404; fax: +972 3 5028130.
E-mail address: [email protected] (Y. Rozenman).
1
Consultants of NI Medical Ltd.
The regional ICG signal is obtained from 2 pairs of tetra-polar
electrodes: one on the wrist, above the radial artery, and the other on
the contra-lateral ankle above the posterior tibial artery. The detailed
description of the system and measured parameters can be reviewed
in previous publications [10–13]. The GGI is designed to assess the
systolic contractile function of the LV and is obtained from the
following formula:
GGI ¼ ΔR=R α HRðcorrectedÞ
where R is the basal impedance and ΔR is the change in impedance, α
is the time to peak ΔR and HR is the heart rate. The GGI takes into
account both ΔR as an estimate of stroke volume and α as a time
interval parameter. HR and R are used as normalization factors so that
the result is a dimensionless index that reflects the systolic function of
the LV.
The entire system (electrode, signal analysis and calculations) can
replace the CD ROM in a regular computer, transforming any laptop
into a simple diagnostic device—ideal for screening.
Echocardiography was performed using GE Vivid 7 (GE Vingmed
Ultrasound A/S, Norway). EF was calculated using the biplane
Simpson's technique [14]. Echocardiography and ICG were performed
on the same day.
Patients who were referred for screening echocardiography at the
outpatient clinic of the E. Wolfson Medical Center were considered for
this trial.
The study protocol conforms to the ethical guidelines of the 1975
Declaration of Helsinki and was approved by the institution's human
research committee.
Optimal cutoff value for the GGI for detection of LVSD was
determined from the Receiver Operating Characteristic (ROC) curve of
the training set. Sensitivity specificity and predictive accuracy of GGI
to detect LVSD were determined using standard technique. Statistical
analysis was performed using MedCalc® software (Broekstraat,
Belgium version 9.5.2.0).
Tables 1 and 2 describe the demographic echocardiographic and
ICG results. The superiority of GGI in detecting LVSD is confirmed
using ROC curve analysis (Fig. 1). The area under curve (AUC) of ROC
curves of ΔR/R and α as predictors of LVSD were 0.84 and 0.75
respectively, significantly lower (p b 0.001) than the value for GGI
(0.97 CI 0.92 to 0.99). For a GGI cutoff value of 10 the sensitivity,
specificity positive and negative predictive values are 86%, 100%, 100%
and 96% respectively.
0167-5273/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ijcard.2011.02.029
Please cite this article as: Rozenman Y, et al, Detection of left ventricular systolic dysfunction using a newly developed, laptop based,
30 impedance
of 158. cardiographic index, Int J Cardiol (2011), doi:10.1016/j.ijcard.2011.02.029
2
Y. Rozenman et al. / International Journal of Cardiology xxx (2011) xxx–xxx
Patients in the validation set were slightly older with a tendency
for lower prevalence of smoking. Echocardiographic and ICG values
were similar in these two sets of patients (Tables 1 and 2). A GGI cutoff
of 10 was 89% sensitive and 96% specific for detection of LVSD,
confirming the excellent results of the training set. Positive and
negative predictive values were 78% and 98% respectively.
False positives and false negatives of GGI as a predictor of EF are
extremely rare (Fig. 2) reflecting the accuracy of GGI to detect LVSD.
Careful observation of the data points suggests that rather than a
simple correlation, the relation between EF and GGI is more like a step
function with a GGI b 10 corresponding to a wide range of EF b 55%
while a GGI ≥ 10 corresponds to varying levels of EF ≥ 55%.
In this study we were able to demonstrate that a simple regional
impedance based technology can reliably detect LVSD. Regional (as
compared to thoracic) ICG enables more accurate determination of
stroke volume and cardiac output [10–12]. The addition of α, a time
interval parameter, with the resulting GGI adds significantly to
accuracy of this method to identify subjects with LVSD (increasing
the AUC from 0.84 to 0.97). Most importantly, also those at the
asymptomatic stage of mild LVSD are accurately detected.
Some of the energy of the LV contraction produces forward blood
flow during systole, while a significant amount is briefly stored as
potential energy in the distended arteries—maintaining the forward
flow during diastole [15]. The information provided by α relates to
this stored energy while the ΔR is related to the stroke volume. The
incorporation of both these parameters into the GGI makes the GGI a
more reliable index of the energy generated by the contraction of the
LV. Time interval parameters were used in the past from the ICG curve
(or its derivative) to try to estimate the EF. Despite significant
correlation [9] the coefficient values (typically approximately 0.5) are
not high enough for clinical purpose. Thus accurate estimation of the
value of EF is limited but, as we have showed, the GGI very accurately
determines whether the LVEF is normal or abnormal (step function
rather than simple correlation).
The natural history of patients with ALVSD is dismal [4]. Compared
with those with normal LV, patients with ALVSD had 4–5-fold
increase risk of CHF and death. Early detection of ALVSD requires a
search for CAD as a possible etiology. When CAD is causing reversible
dysfunction prompt revascularization is essential to improve outcome; preventing CHF and early arrhythmic death [16–19]. In those
with irreversible ALVSD; further deterioration can be attenuated by
the use of appropriate drug therapy [3].
Since early intervention in subjects with ALVSD is beneficial,
screening programs need to be implemented to identify these
individuals. Echocardiography—the “gold standard” in the assessment
of LVSD—is impractical for general screening since it is very expensive.
Thus an effective screening program should use an inexpensive
diagnostic test (such as GGI) to identify those who are likely to have
ALVSD.
Table 2
Echocardiographic findings of subjects in the training and validation cohorts.
Echocardiography
LVEF (%)
LVEDD (cm)
LVESD (cm)
IVS (cm)
LVPW (cm)
LA diameter (cm)
MR (%)
LVH (%)
LA area (cm2)
Diastolic dysfunction (%)
ICG
Cardiac index (l/m2)
Stroke index (cc/m2)
TPR(dynes-sec-cm−5)
GGI
GGI b10 (%)
All
Training cohort Validation cohort p
56 ± 9
4.8 ± 0.6
3.0 ± 0.8
1.0 ± 0.1
1.0 ± 0.1
3.8 ± 0.6
27
22
18.9 ± 5
19
55 ± 9.6
4.9 ± 0.6
2.9 ± 0.7
1.0 ± 0.16
1.0 ± 0.1
3.8 ± 0.55
27
26
18.5 ± 4.9
14
57 ± 9.2
4.8 ± 0.6
3.0 ± 0.8
1.0 ± 0.16
1.0 ± 0.8
3.8 ± 0.6
26
19
19.0 ± 5.0
20
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
3.5 ± 0.8
50 ± 9.3
1299±341
11.6 ± 2.3
17
3.5 ± 0.9
50 ± 9.5
1280 ± 350
11.6 ± 2.6
18
3.6 ± 0.75
49 ± 9.2
1308 ± 337
11.6 ± 2.1
16
NS
NS
NS
NS
NS
LVEDD and LVESD are left ventricular end diastolic and end systolic diameters. IVS and
LVPW are LV septal and posterior wall thicknesses. MR—mitral regurgitation (minimal
or mild), LVH—left ventricular hypertrophy, LA—left atrium and TPR—total peripheral
resistance.
Even though we excluded patients with symptoms of suggestive of
heart disease the decision to refer patients for echocardiography
biases the population so that data was not collected from the true
target population—asymptomatic individuals with no prior history of
heart disease. However the likelihood that this will change the
observed relation between GGI and EF is small.
The GGI can accurately detect LVSD in subjects at risk (referred for
echocardiography). The GGI can be easily obtained from regional ICG
signal using any laptop by replacing the CD ROM with the measuring
hardware and software. This laptop based device is cheap and easy to
operate making it an attractive tool to screen for LVSD. Additional
trials are required to prove the efficacy of this device in the target
population of asymptomatic patients at risk.
Table 1
Clinical characteristics of subjects in the training and validation cohorts.
Number
Male (%)
Age (mean ± SD)
BMI (kg/m2), (mean ± SD)
Hypertension (%)
Diabetes (%)
Dyslipidemia (%)
Smoke (%)
Family history of
CAD (%)
Asymptomatic CAD (%)
BMI—body mass index.
All
Training
cohort
Validation
cohort
p
301
56
63 ± 9
28.2 ± 3.5
44
20
51
20
52
100
65
60 ± 10
28 ± 4.5
41
27
57
28
58
201
52
64 ± 11
28 ± 4.4
44
17
48
16
49
NS
P = 0.0027
NS
NS
NS
NS
P b 0.055
NS
29
28
29
NS
Fig. 1. Receiver operating characteristic (ROC) curve of ΔR/R, α × HR and GGI as
diagnostic tests for LVSD (EF b 55%).
impedance
31
of 158.cardiographic index, Int J Cardiol (2011), doi:10.1016/j.ijcard.2011.02.029
Y. Rozenman et al. / International Journal of Cardiology xxx (2011) xxx–xxx
3
Fig. 2. Scatter plot of EF vs. GGI demonstrating the diagnostic power GGI (with cutoff value of 10) to detect LVSD (EF b 55%) (data from the entire cohort, n = 301).
Acknowledgement
This study was supported by NI Medical Ltd. Israel.
The authors of this manuscript have certified that they comply
with the Principles of Ethical Publishing in the International Journal of
Cardiology [20].
References
[1] Rosamond W, Flegal K, Friday G, et al. Heart disease and stroke statistics: 2007
update: a report from the American Heart Association Statistics Committee and
Stroke Statistics Subcommittee. Circulation 2007;115:e69–e171.
[2] Goldberg LR, Jessup M. Stage B heart failure: management of asymptomatic left
ventricular systolic dysfunction. Circulation 2006;113:2851–60.
[3] Schocken DD, Benjamin EJ, Fonarow GC, et al. Prevention of heart failure: a
scientific statement from the American Heart Association Councils on Epidemiology and Prevention, Clinical Cardiology, Cardiovascular Nursing, and High Blood
Pressure Research; Quality of Care and Outcomes Research Interdisciplinary
Working Group; and Functional Genomics and Translational Biology Interdisciplinary Working Group. Circulation 2008;117:2544–65.
[4] Wang TJ, Evans JC, Benjamin EJ, Levy D, LeRoy EC, Vasan RS. Natural history of
asymptomatic left ventricular systolic dysfunction in the community. Circulation
2003;108:977–82.
[5] Redfield MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett Jr JC.
Plasma brain natriuretic peptide to detect preclinical ventricular systolic or
diastolic dysfunction: a community-based study. Circulation 2004;109:3176–81.
[6] McMurray JV, McDonagh TA, Davie AP, et al. Should we screen for asymptomatic
left ventricular dysfunction to prevent heart failure? Eur Heart J 1998;19:842–6.
[7] Galasko GI, Barnes SC, Collinson P, Lahiri A, Senior R. What is the most costeffective strategy to screen for left ventricular systolic dysfunction: natriuretic
peptides, the electrocardiogram, hand-held echocardiography, traditional echocardiography, or their combination? Euro Heart J 2006;27:193–200.
[8] Cotter G, Moshkovitz Y, Kaluski E, et al. Accurate, noninvasive continuous
monitoring of cardiac output by whole-body electrical bioimpedance. Chest
2004;125:1431–40.
[9] Feng S, Okuda N, Fujinami T, Takada K, Nakano S, Ohte N. Detection of impaired left
ventricular function in coronary artery disease with acceleration index in the first
derivative of the transthoracic impedance change. Clin Cardiol 1988;11:843–7.
[10] Cotter G, Schachner A, Sasson L, Dekel H, Moshkovitz Y. Impedance cardiography
revisited. Physiol Measures 2006;27:817–27.
[11] Paredes OL, Shite J, Shinke T, et al. Impedance cardiography for cardiac output
estimation. Reliability of wrist-to-ankle electrode configuration. Circ J 2006;70:
1164–8.
[12] Goor DA, Frinerman E, Granov I, Patterson R. Peripheral versus thoracic impedance
cardiography. J Cardiac Failure, Heart Failure Society of America (HFSA), 10th
Annual Scientific Meeting, September, 10–13; 2006. p. S54–5.
[13] Rozenman Y, Goor DA, Rotzak R. Progress report on non-invasive diagnosis of
asymptomatic left ventricular systolic dysfunction. Abstract J Card Fail 2009;15:
68.
[14] Gottdiener JS, Bednarz J, Devereux RD, et al. American Society of Echocardiography: recommendations for use of echocardiography in clinical trials, a report from
the American Society of Echocardiography's Nomenclature and Standards
Committee and the task force on echocardiography in clinical trials. J Am Soc
Echocardiogr 2004;17:1086–119.
[15] Hardy CJ. Assessment of arterial elasticity by cardiovascular MRI. In: Kwong RY,
Libby P, editors. Cardiovascular magnetic resonance imaging (contemporary
cardiology). Totowa, New Jersey: Humana Press; 2008. p. 695–710.
[16] Gheorghiade M, Sopko G, De Luca, et al. Navigating the crossroads of coronary
artery disease and heart failure. Circulation 2006;114:1202–13.
[17] Pagano D, Lewis ME, Townend JN, Davies P, Camici PG, Bonser RS. Coronary
revascularization for postischaemic heart failure: how myocardial viability affects
survival. Heart 1999;82:684–8.
[18] Canty Jr JM, Suzuki G, Banas MD, Verheyen F, Borgers M, Fallavollita JA.
Hibernating myocardium. Chronically adapted to ischemia but vulnerable to
sudden death. Circ Res 2004;94:1142–9.
[19] Allman KC, Shaw LJ, Hachamovitch R, Udelson JE. Myocardial viability testing and
impact of revascularization on prognosis in patients with coronary artery disease
and left ventricular dysfunction: a meta-analysis. J Am Coll Cardiol 2002;39:
1151–8.
[20] Shewan LG, Coats AJ. Ethics in the authorship and publishing of scientific articles.
Int J Cardiol 2010;144:1–2.
32 impedance
of 158. cardiographic index, Int J Cardiol (2011), doi:10.1016/j.ijcard.2011.02.029
Measurement Precision in the Optimization of Cardiac Resynchronization Therapy
Robert G. Turcott, Ronald M. Witteles, Paul J. Wang, Randall H. Vagelos, Michael B. Fowler
and Euan A. Ashley
Circ Heart Fail. 2010;3:395-404; originally published online February 22, 2010;
doi: 10.1161/CIRCHEARTFAILURE.109.900076
Circulation: Heart Failure is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX
75231
Copyright © 2010 American Heart Association, Inc. All rights reserved.
Print ISSN: 1941-3289. Online ISSN: 1941-3297
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circheartfailure.ahajournals.org/content/3/3/395
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation: Heart Failure can be obtained via RightsLink, a service of the Copyright Clearance Center, not
the Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation: Heart Failure is online at:
http://circheartfailure.ahajournals.org//subscriptions/
33 of 158.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on December 9, 2013
Measurement Precision in the Optimization of Cardiac
Resynchronization Therapy
Robert G. Turcott, MD, PhD; Ronald M. Witteles, MD, FACC;
Paul J. Wang, MD; Randall H. Vagelos, MD, FACC;
Michael B. Fowler, MB, FRCP, FACC; Euan A. Ashley, MRCP, DPhil
Background—Cardiac resynchronization therapy improves morbidity and mortality in appropriately selected patients.
Whether atrioventricular (AV) and interventricular (VV) pacing interval optimization confers further clinical
improvement remains unclear. A variety of techniques are used to estimate optimum AV/VV intervals; however, the
precision of their estimates and the ramifications of an imprecise estimate have not been characterized previously.
Methods and Results—An objective methodology for quantifying the precision of estimated optimum AV/VV intervals was
developed, allowing physiologic effects to be distinguished from measurement variability. Optimization using multiple
conventional techniques was conducted in individual sessions with 20 patients. Measures of stroke volume and dyssynchrony
were obtained using impedance cardiography and echocardiographic methods, specifically, aortic velocity-time integral,
mitral velocity-time integral, A-wave truncation, and septal-posterior wall motion delay. Echocardiographic methods yielded
statistically insignificant data in the majority of patients (62%– 82%). In contrast, impedance cardiography yielded statistically
significant results in 84% and 75% of patients for AV and VV interval optimization, respectively. Individual cases
demonstrated that accepting a plausible but statistically insignificant estimated optimum AV or VV interval can result in worse
cardiac function than default values.
Conclusions—Consideration of statistical significance is critical for validating clinical optimization data in individual
patients and for comparing competing optimization techniques. Accepting an estimated optimum without knowledge of
its precision can result in worse cardiac function than default settings and a misinterpretation of observed changes over
time. In this study, only impedance cardiography yielded statistically significant AV and VV interval optimization data
in the majority of patients. (Circ Heart Fail. 2010;3:395-404.)
Key Words: cardiac resynchronization therapy 䡲 pacing optimization 䡲 AV delay 䡲 interventricular interval
䡲 echocardiography 䡲 impedance cardiography
C
ardiac resynchronization therapy (CRT) decreases morbidity and mortality in populations with reduced cardiac
function and conduction abnormalities1–3; however, approximately one third of these patients do not experience benefit.4
Atrioventricular (AV) and interventricular (VV) pacing interval
optimization improve hemodynamics acutely5–7 and may enhance the response to CRT relative to default interval settings
among both the traditional responder and nonresponder groups.
To date, however, results of prospective randomized trials have
been mixed.8 –12
Clinical Perspective on p 404
The clarification of fundamental issues in pacing interval
optimization is still at an early stage. For example, the extent
to which optimum AV/VV intervals change with body
position and exertion remains unclear,13–16 and data assessing
the stability of optima over time have been inconsistent.17,18
Even the definitions of pacing intervals vary among manufacturers, with identical programmed intervals corresponding
to very different ventricular pace timing.19 Perhaps the most
important fundamental weakness is the lack of statistical tools
to quantitatively evaluate the significance of the measured
data and to characterize the precision of the estimated
optimum AV/VV interval.11,20 –22
In this study, we examined multiple commonly used
AV/VV interval optimization techniques. Central to our
approach is the recognition that the underlying dependence of
cardiac function on AV/VV interval is obscured to some
degree by measurement noise and that the optimum interval
identified by a given technique is in fact an estimation of the
true physiologic optimum. The degree to which measured
optimization data demonstrate a significant dependence on
pacing interval and the precision of estimated optima were
rigorously evaluated using new statistical tools based on
Received August 6, 2009; accepted January 20, 2010.
From the Division of Cardiovascular Medicine and Center for Biomedical Informatics Research, Stanford University School of Medicine, Stanford,
Calif.
Correspondence to Euan A. Ashley, MRCP, DPhil, Cardiomyopathy Center, Stanford University School of Medicine, Division of Cardiovascular
Medicine, Falk CVRC, 300 Pasteur Dr, MC5406, Stanford, CA 94305-5406. E-mail [email protected]
© 2010 American Heart Association, Inc.
Circ Heart Fail is available at http://circheartfailure.ahajournals.org
34 of 158.
DOI: 10.1161/CIRCHEARTFAILURE.109.900076
395
Downloaded from http://circheartfailure.ahajournals.org/
by guest on December 9, 2013
396
May 2010
Circ Heart Fail
Table 1.
Patient No.
Patient Demographics
Age, y
Sex
EF
Etiology
1
80
M
31
Ischemic
2
73
M
30
Ischemic
3
70
M
28
Ischemic
4
66
M
29
Idiopathic
5
62
M
42
Toxic
6
57
F
65
HCM*
7
55
F
70
HCM*
8
80
M
26
Ischemic, AS, MR
9
45
F
20
Toxic
10
71
M
25
MR
11
54
M
24
AS
12
57
F
23
MR†
13
54
M
27
Idiopathic
14
60
M
27
Ischemic
15
59
F
59
Idiopathic
16
75
M
33
Idiopathic‡
17
77
F
18
Idiopathic§
18
60
M
20
Ischemic
19
49
M
48
Ischemic, AV block
20
66
F
52
HCM
AS indicates aortic stenosis; EF, ejection fraction; HCM, hypertrophic
cardiomyopathy; MR, mitral regurgitation.
*Dual-chamber device.
†No LV capture.
‡Frequent bigeminy.
§No atrial lead.
bootstrapping, a computational approach that makes knowledge of the underlying statistical properties of the data
unnecessary.23
Methods
Patient Selection
Data were obtained from consecutive patients referred for clinical
pacing interval optimization with institutional review board approval. Patients with dual-chamber or biventricular pacemakers or
implantable cardioverter defibrillators were included without regard
to underlying etiology (Table 1).
Optimization Techniques
AV/VV interval optimization was conducted using multiple techniques, as follows:
1. Noninvasive, beat-to-beat estimates of stroke volume (SV)
were acquired continuously using impedance cardiography
(ICG; BioZ, CardioDynamics, San Diego, Calif) (Figure
1A).24 –26 With ICG, changes in thoracic impedance are measured using surface electrodes and then processed by a proprietary algorithm to estimate SV and other hemodynamic parameters. Each test interval was delivered for 60 seconds, with
all beats recorded during the last 30 seconds included in the
analysis.
2. The remaining techniques were derived from echocardiography using a Philips iE33 System (Philips International B.V.,
Amsterdam, The Netherlands). The aortic velocity-time integral (A-VTI), which is directly proportional to SV, was
obtained by numerically integrating the ejection velocity envelop obtained by continuous-wave Doppler directed in line
35 of 158.
with aortic flow in the apical 5-chamber view (Figure 1B).20,21
For this and the other echocardiographic techniques, a 10- to
20-second equilibration period followed each programming
change. Data then were recorded over 1 to 2 respiratory cycles,
with premature and postpremature beats excluded.
3. Mitral inflow velocity-time integral (M-VTI), which is directly
proportional to inflow volume, was obtained using pulsedwave Doppler with the sample volume placed just apical to the
mitral leaflets in the apical 4-chamber view (Figure 1C).20,21
4. In contrast to the techniques described above, which attempt to
characterize cardiac function by estimates of forward flow,
septal-posterior wall-motion delay (SPWMD) provides an
assessment of left ventricular (LV) mechanical synchrony.20
As shown in Figure 1D, the transducer signal was directed
across the septum and the posterior LV wall in the parasternal
long-axis view. Color M-mode Doppler was used to highlight
the relative motion of the 2 walls, with minimum lag taken to
represent optimal ventricular synchrony.
5. A-wave truncation identifies the optimum AV delay as the
shortest pacing interval that avoids truncation of the A-wave21
and is based on the same Doppler waveform as M-VTI. Figure
1E shows an A-wave truncated at 150 milliseconds and
untruncated at 180 milliseconds.
Optimization techniques were compared in terms of their ability to
detect underlying physiologic changes with pacing interval. Specifically, the statistical significance of the data was quantitatively
estimated, as described below. Because A-wave truncation yields a
binary assessment at each test interval (A wave is or is not
truncated), it is not amenable to the analytic paradigm used for the
other techniques. Therefore, for A-wave truncation, we report the
number of times each of the 3 readers, blinded to other results, was
able to estimate an optimum pacing interval.
Summary statistics were obtained based on all patients referred for
clinical optimization and, in addition, with hypertrophic cardiomyopathy patients excluded. Because ICG allows a greater number of
data points to be obtained compared with echocardiographic methods, the AV interval optimization analysis was repeated for ICG data
using the same number of measurements that were obtained in the
corresponding A-VTI data set at each test interval.
Statistical Methods
A third-degree polynomial was fit to the data, and the location of
the maximum was taken as the estimated optimum interval. For
SPWMD, the location of the minimum absolute value of the
polynomial was taken as the optimum. Use of a continuous function,
such as a polynomial, allows interpolation between test intervals and
averaging both at and across test intervals provided that the number
of free parameters of the function is smaller than the number of
unique test intervals.
The test for statistical significance was based on the formulation of
the alternative hypothesis that the measured data do not depend on
pacing interval. The probability of obtaining the observed data under
this null hypothesis was estimated using bootstrapping.22,23 A test
statistic s was defined to be equal to the area bounded above by the
best-fitting polynomial and below by the minimum value of the
polynomial, as illustrated in Figure 2. The test statistic is not unique;
other measures also would be acceptable if they possess the property
that their value varies depending on how “physiological” the
optimization data are. Specifically, if an underlying physiologic
optimum exists, and if the range of test intervals was appropriately
selected to span the optimum, then the test statistic should be larger
than what would be obtained under the null hypothesis in which there
is no dependence on pacing interval. The greater the difference in
cardiac function at the optimum pacing interval and the extreme of
the test range, the greater the value of s. For each optimization data
set, 1000 surrogate bootstrapped data sets were generated by randomly selecting data points from the original data set with replacement and without regard to test interval. This process yields
surrogate data sets with the same number of data points at each test
interval as the original but replaces the original mapping between
Turcott et al
Optimization of Cardiac Resynchronization Therapy
397
Figure 1. Raw optimization data. A, Impedance cardiography is measured noninvasively using surface electrodes and provides
estimates of multiple hemodynamic parameters, including stroke volume (SV). B, Aortic velocity-time integral is measured with
continuous-wave Doppler at the left ventricular outflow tract and is directly proportional to ejected SV. C, Mitral velocity-time integral (M-VTI) is measured with pulsed-wave Doppler echocardiography and is directly proportional to inflow volume. D, Septalposterior wall-motion delay is obtained using tissue Doppler imaging. The VV interval that minimizes the delay is taken as representing optimum synchrony. E, A-wave truncation uses the same images as M-VTI to identify the pacing interval at which the
A-wave becomes truncated by ventricular systole. The image shows the A-wave truncated when a 150-ms AV delay is used and
untruncated with a 180-ms AV delay.
36 of 158.
398
Circ Heart Fail
May 2010
Figure 2. Estimation of statistical significance using bootstrapping. Measures of cardiac function using impedance cardiography (ICG)
and aortic velocity-time integral (A-VTI) recorded from the same patient at the same optimization session. A test statistic s is defined as
the area that is bounded above by the best-fitting polynomial and below by the minimum of the polynomial. The greater the difference
in cardiac function at best and worst test intervals, the larger the value of s. Bootstrapping is used to obtain surrogate data sets (bottom) from the original data that preserve the amplitude statistics of the original data but lack the underlying dependence on cardiac
function. The fraction p of bootstrapped tests statistics s⬘ that exceed s reflects the likelihood that the observed data occur in the
absence of a dependence of cardiac function on test interval. Here, the ICG data are highly significant (P⬍0.001) and demonstrate a
physiologically plausible dependence on pacing interval. The data are well described by the continuous function despite the delivery of
test intervals in a random order and the relatively close spacing between test intervals. Furthermore, the 120-ms test interval delivered
4 times over the course of data acquisition shows good repeatability compared with the range of SVs over all test intervals. In contrast,
the A-VTI data are consistent with the null hypothesis: The large relative noise masks the underlying physiological dependence on pacing interval, and the location of the optimum identified by A-VTI is an artifact of the statistical variability. In this patient, accepting the
optimum identified by A-VTI would result in worse cardiac function than the default interval (120 milliseconds). Solid curve indicates
best-fitting third-degree polynomial; X, location of estimated optimum pacing interval; horizontal line, 95% CI of estimated optimum. SV
indicates stroke volume.
measured data and test interval with a random pairing. Each
surrogate bootstrapped data set thus represents a single realization of
data that would be observed under the null hypothesis while
preserving the amplitude statistics of the original data. The test
statistic s⬘ was calculated for each surrogate data set, and the fraction
that was greater than or equal to that of the original data estimates the
probability that a test statistic at least as large as that associated with
the original data would be observed under the null hypothesis. For
Pⱕ0.05, the null hypothesis was rejected, and the data were
interpreted as demonstrating a statistically significant dependence on
pacing interval.
The 95% CI of the estimated optimum pacing interval was
obtained by generating an additional 1000 bootstrapped data sets in
which the original data were randomly selected with replacement
while preserving the mapping to test interval. In this case, with the
mapping between test interval and measured data retained, the
collection of best-fitting polynomials of the bootstrapped data sets
reflects the variability in the best-fitting polynomial of the original
data that is attributable to the statistical variability in the measurements. The locations of the maxima of the surrogate data sets were
ordered from smallest to largest, and the cutoff point of the smallest
37 of 158.
and largest 2.5% of the surrogate optima were taken as the 95% CI
of optimum estimated from the original data.
Assessment of A-VTI Variability
The variability of the measured A-VTI data was quantified by
calculating the average and SD over all measurements made at an
AV delay of 120 ms, which by protocol were acquired over at least
1 respiratory cycle at 4 different times during the recording session.
To allow comparison with previously published values, a transformation was derived that converts the coefficient of variation (defined
as SD divided by mean) to the average difference of successive
measurements. Specifically, ␮y⫽2CV 冑2/␲, where ␮y is the average
difference in successive measurements, and CV is the coefficient of
variation. This result is based on the transformation yi⫽兩xi1⫺xi2兩/␮i,
where the xi represents successive A-VTI measurements at a given
pacing interval in a given patient; yi, the normalized difference in
A-VTI measures in a given patient, and the subscript i, patientspecific values. The derivation assumes that xi1 and xi2 are independently drawn from a Gaussian distribution whose mean ␮i and SD
␴i can vary among patients but whose coefficient of variation CVi
remains fixed for all patients, that is, CVi⫽␴i/␮i⫽constant. Data
acquisition and device programming are summarized in Table 2.
Turcott et al
Table 2. Summary of Data Acquisition, Analysis, and
Pacemaker Programming
Data acquisition
AV and VV interval optimization were conducted independently, with
simultaneous biventricular pacing during AV optimization and an AV
delay of 120 ms during VV optimization
Data were recorded from each patient at a single session in the
following order: ICG (AV), ICG (VV), A-VTI (AV), M-VTI/A-Wave truncation
(AV), A-VTI (VV), SPWMD (VV)
Consecutive beat-to-beat measures of cardiac function were recorded
over at least 1 respiratory cycle
Test intervals were delivered in random order with 20 –30 ms spacing
The 120 ms AV delay and 0 ms VV interval were repeated 4 times over
the duration of the recording to assess stability
Data Analysis
A 3rd degree polynomial was fit to the data, with the location of the
maximum serving as the estimated optimum (For SPWMD, the optimum
was the location with the minimum absolute delay)
A P value reflecting the statistical significance of measured data was
obtained
The 95% CI of the estimated optimum was obtained
For A-wave truncation, the shortest AV delay that avoided A-wave
truncation was selected as optimum, and the fraction of patients for
whom an optimum could be determined was noted
Pacemaker Programming
The statistical significance, 95% CI, and physiologic plausibility of the
measured data were considered
The pacemaker was left at default settings (AV delay ⫽ 120 ms, VV
interval ⫽ 0 ms) for statistically insignificant data sets, wide 95% CIs
that included the default setting, or physiologically implausible data.
Otherwise, the estimated optimum pacing interval was programmed.
Results
Patient demographics are presented in Table 1. Of the 20
sequential, unique patients referred for clinical pacing optimization, 18 had biventricular pacemakers or implantable
cardioverter defibrillators and 2 had dual-chamber devices.
The LV lead in 1 patient failed to capture, and another patient
with chronic atrial fibrillation had a biventricular device
399
without an atrial lead. Patients who underwent AV delay
optimization were in sinus rhythm with the AV delay initiated
by an atrial sensed-event in all cases.
AV delay optimization data from a single patient are
presented in Figure 3. As with the ICG data shown here (left
panel), when statistically significant data were obtained they
typically exhibited an inverted U appearance, indicating that
the estimated optimum interval was in the interior of the
range of test intervals. Statistically insignificant data typically
had a best-fitting polynomial that was flat relative to the
intrinsic variability of the data.
The average and SD of the measured A-VTI data were
calculated over all beats (10 to 31, median 20) recorded from
each patient during AV optimization at the 120-ms test
interval. The ranges of the per-patient average and SD of the
A-VTI data were 6.7 to 56 cm and 0.73 to 5.1 cm, respectively. The median sample coefficient of variation was 0.075,
which corresponds to an average normalized difference in
successive measurements of ␮y⫽0.12, consistent with previous reports of ␮y⫽0.1⫾0.1.27
The VV optimization data shown in Figure 4 are from the
same patient whose AV optimization results are presented in
Figure 3. As with this individual, for most patients, VV
optimization yielded insignificant results more frequently
than did AV optimization.
The number of statistically significant data sets is summarized in Table 3. For both AV and VV optimization, A-VTI,
M-VTI, and SPWMD yielded data that were indistinguishable from the null hypothesis (ie, failed to demonstrate a
statistically significant dependence on pacing interval) the
majority of the time. As shown in Table 3, this finding
remained true whether patients with hypertrophic cardiomyopathy were included or excluded from the analysis. ICG was
significantly different from the null hypothesis 84% of the
time for AV delay optimization and 75% of the time for VV
interval optimization. Excluding patients with hypertrophic
cardiomyopathy from the analysis, the null hypothesis could
be rejected 81% and 75% of the time, respectively.
Figure 3. AV delay optimization. Data were obtained from an individual patient at a single optimization session. Impedance cardiography (ICG) and aortic velocity-time integral (A-VTI) yielded statistically significant results, although the 95% CI of the estimated optimum
was much narrower for ICG than for A-VTI. Mitral velocity-time integral (M-VTI) data were not statistically significant. Solid curve indicates best-fitting third-degree polynomial; X, location of estimated optimum pacing interval; horizontal line, 95% CI of estimated
optimum.
38 of 158.
400
Circ Heart Fail
May 2010
Figure 4. VV interval optimization. Data were obtained from the same patient and optimization session as those presented in Figure 3.
Negative VV intervals indicate that the left ventricle is paced first, with extreme values corresponding to unichamber pacing. Excluding
the unichamber LV pacing from the septal-posterior wall-motion delay (SPWMD; VV⫽⫺140 ms) would yield statistically significant
results, although the physiological plausibility of extremely early right ventricular pacing it predicts is questionable, especially in light of
the insignificant results from impedance cardiography (ICG) and aortic velocity-time integral (A-VTI). Solid curve indicates best-fitting
third-degree polynomial; X, location of estimated optimum pacing interval; horizontal line, 95% CI of estimated optimum.
Among the statistically significant data sets, the optimum
AV delay estimated by ICG differed from the default value
(120 ms) by an average magnitude of 57 ms, and 63% of the
estimates differed from default by at least 50 ms. Among the
statistically significant, ICG-derived VV interval data sets,
estimated optima differed from the default value (0 ms) by an
average of 52 ms, and 75% of the estimates differed by at
least 30 ms. Repeating the ICG AV interval optimization
analysis using the same number of measurements at each test
interval as the A-VTI recordings continued to yield statistically significant results in the majority of patients, with 81%
of the reduced ICG data sets having P values ⬍0.05.
An optimum AV delay could be estimated by 3 independent readers using A-wave truncation 69%, 75%, and 94% of
the time. The estimates of each reader are compared with ICG
results in Table 4. The relationship between optima predicted
by ICG and A-wave truncation was variable, with the Pearson
correlation coefficient ranging from 0.02 to 0.67 for the 3
readers.
Although the analysis presented above is based on unique
patient visits, 1 patient underwent both ICG and echocardiographic optimization on 2 separate occasions separated by 3.5
months. As shown in Figure 5, in both cases ICG yielded
similar-appearing, statistically significant results. The estimates of the optimum pacing intervals were precise, with
narrow 95% CIs, and had good concordance, with 91 ms at
the first optimization and 67 ms 3.5 months later. In contrast,
in neither optimization session did A-VTI yield statistically
significant results. The wide 95% CI of the initial A-VTI
optimum predicted that subsequent optimization attempts
would be unlikely to identify a similar optimum interval.
Discussion
In this study, multiple clinically accepted AV and VV interval
optimization techniques were used in single sessions in a
heterogeneous group of 20 patients. For 62% to 86% of
patients, A-VTI, M-VTI, and SPWMD yielded data that were
statistically indistinguishable from the null hypothesis; that is,
39 of 158.
the majority of the time these measures yielded data that did
not show a significant dependence on AV or VV interval.
With A-wave truncation, it was possible for 3 independent
readers to estimate an optimum AV delay 69%, 75%, and
94% of the time. ICG performed better than the echocardiography methods, yielding statistically significant data 84%
of the time for AV delay optimization and 75% of the time for
VV interval optimization. Notably, the echocardiography
techniques tested here represent the most commonly used
approaches to AV/VV interval optimization.28
In previous studies, SPWMD failed to predict a response to
CRT,29,30 and interval optimization using A-VTI neither
improved clinical outcomes12 nor yielded acute data that were
distinguishable from a negative control.31 The poor precision
of these techniques demonstrated in this study may account
for the lack of clinical utility seen in the earlier work.
AV/VV interval optimization traditionally has been conducted without consideration of the intrinsic variability of the
measured data. Test intervals often are delivered in a nonrandom order, sweeping systematically from one end of the test
range to the other, and the AV or VV interval associated with
the best average measure of cardiac function typically is
taken at face value to represent the underlying physiologic
optimum.11,20 Although efficient, the traditional approach has
important drawbacks. A nonrandom order of test intervals
allows systematic error to be introduced into the measured
data in a way that cannot be subsequently corrected by
averaging (eg, with subtle drift of the Doppler angle over the
course of data collection). Furthermore, without considering
the intrinsic variability of the measured data, it is not possible
to characterize the precision of the estimated optimum
AV/VV interval.
In the absence of knowledge about its precision, an
estimated optimum AV/VV interval may in fact be spurious
and associated with worse cardiac function than the
population-derived, default setting (eg, a sensed AV delay of
120 milliseconds) (Figures 2, 3, and 5). In addition, although
multiple studies have suggested that optimum pacing inter-
Turcott et al
Table 3.
P Value Associated With Measured Optimization Data
AV Optimization
Patients
ICG
A-VTI
M-VTI
ICG
A-VTI
SPWMD
2
⬍0.001
0.18
0.03
0.26
0.03
0.40
⬍0.001
⬍0.001
0.018
0.51
Estimated Optima: ICG and A-Wave Truncation
ICG Estimated
Optima, ms
VV Optimization
Individual patients
1
Table 4.
Patient
No.
401
Optimum
A-Wave
Truncation, ms
95% CI
Reader 1
Reader 2
Reader 3
0.002
1
168
156
177
180
150
210
0.25
2
240
240
240
150
150
180
206
193
213
150
120
*
210
3
⬍0.001
0.013
0.21
0.12
0.31
0.10
3
4
⬍0.001
0.001
0.007
0.72
0.12
0.003
4
197
187
206
240
180
0.42
5
93
60
117
180
120
210
91
83
97
120
150
120
5
0.09
0.37
0.50
⬍0.001
0.14
6
⬍0.001
0.68
6
7
⬍0.001
0.06
7
128
121
138
8
⬍0.001
0.24
0.001
8
200
177
270
180
*
240
0.24
9
178
170
197
120
150
120
0.57
10
60
60
60
240
*
*
0.93
11
60
60
60
150
120
120
9
10
0.008
0.004
0.98
0.59
0.05
0.91
0.61
⬍0.001
0.092
⬍0.001
⬍0.001
0.19
0.41
0.007
11
⬍0.001
0.17
0.44
12
0.016
0.04
0.17
12
60
60
60
180
*
*
173
125
190
150
180
120
0.73
13
0.023
0.50
0.44
0.054
0.41
0.054
13
14
⬍0.001
0.77
0.057
0.002
0.58
0.11
14
164
157
172
180
150
80
123
118
131
240
*
*
218
30
248
*
*
180
18
77
66
86
19
30
30
30
120
60
90
20
240
180
240
15
⬍0.001
16
0.051
0.003
0.14
0.006
0.073
0.24
15
0.05
0.16
0.002
0.061
0.081
16
0.075
17
⬍0.001
17
18
⬍0.001
19
⬍0.001
20
⬍0.001
0.32
⬍0.001
⬍0.001
All patients
Significant, n
16
6
3
12
2
3
Attempted, n
19
16
14
16
14
14
Significant, %
84
38
21
75
14
21
Significant, n
13
6
3
12
2
3
Attempted, n
16
14
14
16
14
14
Significant, %
81
42
21
75
14
21
HCM excluded
HCM indicates hypertrophic cardiopmyopathy; AV, atrioventricular; VV,
interventricular; ICG, impedance cardiography; A-VTI, aortic velocity-time
integral; M-VTI, mitral velocity-time integral; SPWMD, septal-posterior wallmotion delay.
vals change over time,12,17,18 the lack of characterization of
the precision of the estimated optima makes it impossible to
determine whether the observed change reflects changes in
underlying physiology or is simply due to statistical variability in measured data31 (Figure 5).
The fact that any optimization technique carries some
degree of imprecision requires caution when designing randomized prospective trials. Two electrogram-based optimization techniques sponsored by competing manufacturers are
currently undergoing clinical trials. In 1 design,32 patients are
randomized to the experimental optimization technique or a
control arm that may or may not include AV and VV interval
optimization. If optimization is performed, it is conducted at
the physician’s discretion without a requirement to examine
the precision of the measured data, raising the possibility of a
programmed AV or VV interval that yields worse cardiac
function than population-derived default settings. By design,
40 of 158.
Percent interpretable was as follows: Reader 1, 94%; Reader 2, 69%;
Reader 3, 75%. Correlation coefficients were as follows: Reader 1, 0.02;
Reader 2, 0.67; Reader 3, 0.48.
*Reader indicated that A-wave truncation could not be reliably identified. ICG
indicates impedance cardiography; CI, confidence interval.
the trial will not address the question of whether the experimental technique is superior to default settings. In contrast,
another clinical trial33,34 uses a 3-arm design in which patients
are randomized to the experimental technique, a control arm
in which population-derived default settings are used, and a
conventional optimization arm in which a specific optimization technique is uniformly used. This study design allows
direct comparisons between the experimental technique and
both default settings and the specific conventional optimization method, and for comparison of the conventional technique to default settings.
Statistical significance testing may provide a useful way to
evaluate the quality of competing optimization techniques. It
avoids prespecifying a gold standard, and although demonstration of improved clinical outcomes in well-designed trials
ultimately is required, the approach presented here offers a
way to narrow the very wide field of plausible optimization
techniques without the resource requirements of a prospective
clinical trial.
Although a rigorous, quantitative analysis is desirable,
statistical significance can be evaluated informally by plotting the measured data against pacing interval. With test
intervals delivered in a random order, if the data exhibit the
expected inverted U shape and are tightly clustered about the
overall curve, then one can be confident that the estimation of
optimum AV/VV interval is precise; repeating the process
402
Circ Heart Fail
May 2010
Figure 5. Stability of estimated optima
over time. Impedance cardiography and
aortic velocity-time integral were
obtained from the same patient at baseline and 3.5 months later. In this patient,
the ICG data were highly significant, the
appearance of the data on repeat optimization recapitulated the baseline data,
the estimated optima were concordant
between the 2 sessions, and the 95%
CIs of the estimates were narrow. In
contrast, the A-VTI data did not show a
statistically significant dependence on
pacing interval, and the estimated
optima were highly discordant between
the 2 optimization sessions. Accepting
the A-VTI-estimated optima without a
consideration of their statistical significance or precision would lead to the
erroneous conclusion that the optimum
AV delay had changed in the intervening
3.5 months.
likely would yield a similar result. On the other hand, if the
plot is relatively flat compared with the intrinsic variability of
the measured data, then the estimate of the underlying
optimum is imprecise and heavily influenced by measurement variability. In this case, repeating the optimization
process likely would yield a very different estimated optimum AV/VV interval. The statistical tools used here along
with plotting capability have been made available on the
Internet.35
In the majority of patients examined in the present study,
ICG generated precise estimates and A-wave truncation
yielded data from which an optimum could be inferred,
although often with significant interreader variability and
marginal correlation with the ICG-predicted optima. Notably,
neither ICG nor A-wave truncation has been clinically validated in prospective interval optimization studies. In addition,
unanswered fundamental questions include whether the physiologic optimum interval evolves over time and whether it
changes between rest and exertion or between supine and
upright posture. A study that compares estimated optimum
intervals obtained at both supine rest and upright exertion in
the same patient, perhaps using motion-tolerant ICG, would
add important insight into these basic questions. In the
absence of such data and given the theoretical potential
benefit of pacing optimization, our approach is to accept an
estimated optimum interval if quantitative and qualitative
analyses suggest that the estimate has good precision. Particular attention should be paid to the effect of outliers and the
overall shape of the curve compared with the measurement
variability.
41 of 158.
That ICG continued to yield statistically significant data in
the majority of patients, even when using an identical number
of data points as the A-VTI analysis, suggests that it has a
superior intrinsic signal-to-noise ratio and that the acquisition
time can be substantially shortened from the 60 s per test
interval that was used in this study. The superior noise
properties of ICG may be partly due to the automatic and
objective nature of data acquisition and analysis in contrast to
A-VTI, which requires the sonographer to physically hold the
probe in a fixed position and the reader to manually demarcate the envelope of the velocity waveform.
A wide variety of optimization techniques have been
advocated, including multiple approaches to the assessment
of systolic function, diastolic function, and electrical and
mechanical synchrony.9,19 –21,24 –26,30,36 – 40 Although each has a
rationale that is mechanistically plausible, consideration of
the neurohormonal derangements of heart failure and the
therapeutic interventions that have been successful lead us to
view SV and its surrogates as parameters that when optimized
are most likely to translate into clinical benefit. Specifically,
it is now well established that ameliorating the effects of
sympathetic tone in these patients leads to improved clinical
outcomes.41 For a given cardiac output, maximizing SV
would minimize sympathetic tone. Indeed, the effects on the
neurohormonal system of increased mechanical efficiency
may contribute to the salutary effects of CRT, which remains
the only contractility-enhancing intervention demonstrated to
prolong life.42 Theoretical arguments may not account for
important effects, however. For example, an increase in SV at
the expense of greater oxygen consumption may not benefit
Turcott et al
the patient with ischemic heart failure. Ultimately, any
proposed optimization technique must be validated by randomized prospective trials with hard clinical end points, a
goal which to date has not been frequently achieved.8 –12
Limitations
The study was based on a small and heterogeneous population comprising successive patients referred for clinical pacing interval optimization. Multiple conventional optimization
techniques were examined for their ability to yield statistically significant results. This property is necessary but not
sufficient for improved clinical outcomes, which were not
examined in this study.
Conclusions
Optimization of AV and interventricular intervals in CRT
requires assessment of the variability of the measured data.
Accepting an estimated optimum without considering its
precision may result in worse cardiac function than default
settings in the individual patient and confound results in
clinical trials. In the small, heterogeneous pacemaker population examined here, echocardiographic techniques yielded
statistically insignificant data in the majority of patients. In
contrast, ICG yielded precise estimates of the optimum AV
and VV interval in most patients. Further research is necessary to confirm these results, to validate the accuracy of the
impedance-predicted optima, and to demonstrate clinical
improvement with pacing interval optimization compared to
population-derived default settings.
Sources of Funding
Dr Turcott was supported by a Heart Failure Society of America
Research Fellowship and by the National Library of Medicine
(Biomedical Informatics Training Grant LM 07033). Dr Ashley was
supported by the National Institutes of Health (grant K08
HL083914).
Disclosures
Dr Witteles has received honoraria from Medtronic. Dr Wang has
received honoraria and research support from and has served as a
consultant or advisor to Boston Scientific, Medtronic, and St Jude
Medical. Dr Fowler has received honoraria from Medtronic and
Boston Scientific.
References
1. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De Marco T,
Carson P, DiCarlo L, DeMets D, White BG, DeVries DW, Feldman AM.
Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med. 2004;350:
2140 –2150.
2. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappenberger L, Tavazzi L. The effect of cardiac resynchronization on morbidity
and mortality in heart failure. N Engl J Med. 2005;352:1539 –1549.
3. McAlister FA, Ezekowitz J, Hooton N, Vandermeer B, Spooner C,
Dryden DM, Page RL, Hlatky MA, Rowe BH. Cardiac resynchronization
therapy for patients with left ventricular systolic dysfunction: a systematic
review. J Am Med Assoc. 2007;297:2502–2514.
4. Yu CM, Wing-Hong Fung J, Zhang Q, Sanderson JE. Understanding
nonresponders of cardiac resynchronization therapy— current and future
perspectives. J Cardiovasc Electrophysiol. 2005;16:1117–1124.
5. Breithardt OA, Stellbrink C, Franke A, Balta O, Diem BH, Bakker P,
Sack S, Auricchio A, Pochet T, Salo R. Acute effects of cardiac resynchronization therapy on left ventricular Doppler indices in patients with
congestive heart failure. Am Heart J. 2002;143:34 – 44.
42 of 158.
403
6. Auricchio A, Stellbrink C, Block M, Sack S, Vogt J, Bakker P, Klein H,
Kramer A, Ding J, Salo R, Tockman B, Pochet T, Spinelli J. Effect of
pacing chamber and atrioventricular delay on acute systolic function of
paced patients with congestive heart failure. Circulation. 1999;99:
2993–3001.
7. Kass DA, Chen CH, Curry C, Talbot M, Berger R, Fetics B, Nevo E.
Improved left ventricular mechanics from acute VDD pacing in patients with
dilated cardiomyopathy and ventricular conduction delay. Circulation. 1999;
99:1567–1573.
8. Boriani G, Muller CP, Seidl KH, Grove R, Vogt J, Danschel W,
Schuchert A, Djiane P, Biffi M, Becker T, Bailleul C, Trappe HJ.
Randomized comparison of simultaneous biventricular stimulation
versus optimized interventricular delay in cardiac resynchronization
therapy: The Resynchronization for the HemodYnamic Treatment for
Heart Failure Management II Implantable Cardioverter Defibrillator
(RHYTHM II ICD) study. Am Heart J. 2006;151:1050 –1058.
9. Morales MA, Startari U, Panchetti L, Rossi A, Piacenti M. Atrioventricular delay optimization by Doppler-derived left ventricular dP/dt
improves 6-month outcome of resynchronized patients. Pacing Clin
Electrophysiol. 2006;29:564 –568.
10. Rao RK, Kumar UN, Schafer J, Viloria E, De Lurgio D, Foster E.
Reduced ventricular volumes and improved systolic function with cardiac
resynchronization therapy: a randomized trial comparing simultaneous
biventricular pacing, sequential biventricular pacing, and left ventricular
pacing. Circulation. 2007;115:2136 –2144.
11. Sawhney NS, Waggoner AD, Garhwal S, Chawla MK, Osborn J, Faddis
MN. Randomized prospective trial of atrioventricular delay programming
for cardiac resynchronization therapy. Heart Rhythm. 2004;1:562–567.
12. Boriani G, Biffi M, Muller CP, Seidl KH, Grove R, Vogt J, Danschel W,
Schuchert A, Deharo JC, Becker T, Boulogne E, Trappe HJ. A prospective randomized evaluation of VV delay optimization in CRT-D
recipients: echocardiographic observations from the RHYTHM II ICD
study. Pacing Clin Electrophysiol. 2009;32(suppl 1):S120 –S125.
13. Ritter P, Daubert C, Mabo P, Descaves C, Gouffault J. Haemodynamic
benefit of a rate-adapted A-V delay in dual chamber pacing. Eur Heart J.
1989;10:637– 646.
14. Melzer C, Korber T, Theres H, Nienaber CA, Baumann G, Ismer B. How
can the rate-adaptive atrioventricular delay be programmed in atrioventricular block pacing? Europace. 2007;9:319 –324.
15. Scharf C, Li P, Muntwyler J, Chugh A, Oral H, Pelosi F, Morady F,
Armstrong WF. Rate-dependent AV delay optimization in cardiac resynchronization therapy. Pacing Clin Electrophysiol. 2005;28:279 –284.
16. Mehta D, Gilmour S, Ward DE, Camm AJ. Optimal atrioventricular delay
at rest and during exercise in patients with dual chamber pacemakers: a
non-invasive assessment by continuous wave Doppler. Br Heart J. 1989;
61:161–166.
17. Zhang Q, Fung JW, Chan YS, Chan HC, Lin H, Chan S, Yu CM. The role
of repeating optimization of atrioventricular interval during interim and
long-term follow-up after cardiac resynchronization therapy. Int
J Cardiol. 2008;124:211–217.
18. Valzania C, Biffi M, Martignani C, Diemberger I, Bertini M, Ziacchi M,
Bacchi L, Rocchi G, Rapezzi C, Branzi A, Boriani G. Cardiac resynchronization therapy: variations in echo-guided optimized atrioventricular and
interventricular delays during follow-up. Echocardiography. 2007;24:
933–939.
19. Burri H, Sunthorn H, Shah D, Lerch R. Optimization of device programming for cardiac resynchronization therapy. Pacing Clin Electrophysiol. 2006;29:1416 –1425.
20. Agler DA, Adams DB, Waggoner AD. Cardiac resynchronization therapy
and the emerging role of echocardiography (part 2): the comprehensive
examination. J Am Soc Echocardiogr. 2007;20:76 –90.
21. Jansen AH, Bracke FA, van Dantzig JM, Meijer A, van der Voort PH,
Aarnoudse W, van Gelder BM, Peels KH. Correlation of echo-Doppler
optimization of atrioventricular delay in cardiac resynchronization
therapy with invasive hemodynamics in patients with heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol.
2006;97:552–557.
22. Turcott RG, Ashley EA. Statistical methods for AV/VV optimization
[abstract]. J Cardiac Failure. 2008;14:S65.
23. Efron B, Tibshirani R. Statistical data analysis in the computer age.
Science. 1991;253:390 –395.
24. Heinroth KM, Elster M, Nuding S, Schlegel F, Christoph A, Carter J,
Buerke M, Werdan K. Impedance cardiography: a useful and reliable tool
in optimization of cardiac resynchronization devices. Europace. 2007;9:
744 –750.
404
Circ Heart Fail
May 2010
25. Braun MU, Schnabel A, Rauwolf T, Schulze M, Strasser RH. Impedance
cardiography as a noninvasive technique for atrioventricular interval optimization in cardiac resynchronization therapy. J Interv Card Electrophysiol.
2005;13:223–229.
26. Tse HF, Yu C, Park E, Lau CP. Impedance cardiography for atrioventricular interval optimization during permanent left ventricular pacing.
Pacing Clin Electrophysiol. 2003;26:189 –191.
27. Zuber M, Toggweiler S, Quinn-Tate L, Brown L, Amkieh A, Erne P. A
comparison of acoustic cardiography and echocardiography for optimizing pacemaker settings in cardiac resynchronization therapy. Pacing
Clin Electrophysiol. 2008;31:802– 811.
28. Gras D, Gupta MS, Boulogne E, Guzzo L, Abraham WT. Optimization of
AV and VV delays in the real-world CRT patient population: an international survey on current clinical practice. Pacing Clin Electrophysiol.
2009;S236 –S239.
29. Beshai JF, Grimm RA, Nagueh SF, Baker JH II, Beau SL, Greenberg SM,
Pires LA, Tchou PJ. Cardiac-resynchronization therapy in heart failure
with narrow QRS complexes. N Engl J Med. 2007;357:2461–2471.
30. Chung ES, Leon AR, Tavazzi L, Sun JP, Nihoyannopoulos P, Merlino J,
Abraham WT, Ghio S, Leclercq C, Bax JJ, Yu CM, Gorcsan J III, St John
Sutton M, De Sutter J, Murillo J. Results of the Predictors of Response to
CRT (PROSPECT) trial. Circulation. 2008;117:2608 –2616.
31. Turcott RG, Das AK, Ashley EA. Re: A prospective randomized
evaluation of VV delay optimization in CRT-D recipients: echocardiographic observations from the RHYTHM II ICD study. Pacing Clin
Electrophysiol. 2009;32:1360 –1361; author reply 1361–1362.
32. FREEDOM: A Frequent Optimization Study Using the QuickOpt
Method. http://www.clinicaltrials.gov. Identifier: NCT00418314.
Accessed January 2, 2010.
33. Stein KM, Ellenbogen KA, Gold MR, Lemke B, Lozano IF, Mittal S,
Spinale FG, van Eyk JE, Waggoner AD, Meyer TE. SmartDelay
34.
35.
36.
37.
38.
39.
40.
41.
42.
Determined AV Optimization: a comparison of AV delay methods used
in cardiac resynchronization therapy (SMART-AV): rationale and design.
Pacing Clin Electrophysiol. 2010;33:54 – 63.
Comparison of AV Optimization Methods Used in Cardiac Resynchronization Therapy (CRT) (SMART-AV). http://www.clinicaltrials.gov.
Identifier: NCT00677014. Accessed January 2, 2010.
Stanford University. Optimize CRT (Development Version) Available at
http://optimizecrt.stanford.edu. Accessed January 2, 2010.
Bertini M, Ziacchi M, Biffi M, Martignani C, Saporito D, Valzania C,
Diemberger I, Cervi E, Frisoni J, Sangiorgi D, Branzi A, Boriani G.
Interventricular delay interval optimization in cardiac resynchronization
therapy guided by echocardiography versus guided by electrocardiographic QRS interval width. Am J Cardiol. 2008;102:1373–1377.
van Gelder BM, Meijer A, Bracke FA. The optimized V-V interval
determined by interventricular conduction times versus invasive measurement by LV dP/dt(max). J Cardiovasc Electrophysiol. 2008;19:
939 –944.
Burri H, Sunthorn H, Somsen A, Zaza S, Fleury E, Shah D, Righetti A.
Optimizing sequential biventricular pacing using radionuclide ventriculography. Heart Rhythm. 2005;2:960 –965.
Vidal B, Tamborero D, Mont L, Sitges M, Delgado V, Berruezo A,
Diaz-Infante E, Tolosana JM, Pare C, Brugada J. Electrocardiographic
optimization of interventricular delay in cardiac resynchronization
therapy: a simple method to optimize the device. J Cardiovasc Electrophysiol. 2007;18:1252–1257.
Turcott RG, Pavek TJ. Hemodynamic sensing using subcutaneous photoplethysmography. Am J Physiol Heart Circ Physiol. 2008;295:
H2560 –H2572.
Brophy JM, Joseph L, Rouleau JL. Beta-blockers in congestive heart
failure. A Bayesian meta-analysis. Ann Intern Med. 2001;134:550 –560.
Kass DA. Rescuing a failing heart: putting on the squeeze. Nat Med.
2009;15:24 –25.
CLINICAL PERSPECTIVE
Cardiac resynchronization therapy improves morbidity and mortality in appropriately selected patients. Whether further
clinical benefit is possible with atrioventricular and interventricular pacing interval optimization remains unclear. Tools to
assess the statistical significance of the measured optimization data have not been available previously. In the study
reported here, an objective methodology for quantifying the statistical precision of estimated optimum pacing intervals was
developed and applied to a number of commonly used optimization techniques. Many of the techniques did not yield
statistically significant data in a majority of patients referred for atrioventricular and interventricular interval optimization,
a finding that raises questions about the ability of pacing interval optimization to enhance clinical outcomes. The data
demonstrate that accepting an estimated optimum interval without consideration of its statistical significance can result in
worse cardiac function than default settings and can lead to the erroneous conclusion that the physiological optimum has
changed over time. These results highlight the importance of evaluating the precision of measured data when conducting
pacing interval optimization for the individual patient and when interpreting the results of clinical trials.
43 of 158.
ORIGINAL ARTICLE Heart Failure
Circ J 2009; 73: 1074 – 1079
Whole Body Bioimpedance Monitoring for Outpatient
Chronic Heart Failure Follow up
Yusuke Tanino, MD; Junya Shite, MD; Oscar L Paredes, MD; Toshiro Shinke, MD;
Daisuke Ogasawara, MD; Takahiro Sawada, MD; Hiroyuki Kawamori, MD;
Naoki Miyoshi, MD; Hiroki Kato, MD; Naoki Yoshino, MD; Ken-ichi Hirata, MD
Background: Although cardiac output index (CI), stroke volume index (SVI), and total systemic vascular resistance (TSVR) are important hemodynamic parameters for the prognosis of chronic heart failure (CHF), they are
difficult to measure in an outpatient setting. Whole body bioimpedance monitoring using a Non-Invasive Cardiac
System (NICaS) allows for easy, non-invasive estimation of these parameters. Here, whether NICaS-derived
hemodynamic parameters are clinically significant was investigated by relating them to other conventional cardiovascular functional indices, and by evaluating their predictive accuracy for CHF readmission.
Methods and Results: Study subjects of 68 patients with CHF were enrolled in the study immediately upon
discharge from the hospital. NICaS-derived CI, -SVI, and -TSVR values obtained at an outpatient clinic were
significantly related with left ventricular ejection fraction (LVEF) measured by echocardiography, serum B-type
natriuretic peptide (BNP), and exercise tolerance. During the 100±98 days follow-up, 15 patients were readmitted
to our hospital for CHF recurrence. Multivariate analysis indicated that LVEF, NICaS-derived CI, NICaS-derived
SVI, and plasma BNP were significant indicators (receiver operating characteristic curve cut-off point, LVEF: 37%,
NICaS-derived CI: 2.49 L · min–1 · m–2, NICaS-derived SVI: 27.2 ml/m2, plasma BNP: 344 pg/ml) for readmission.
Conclusions: Hemodynamic parameters derived by NICaS are applicable for the non-invasive assessment of
cardiac function in outpatient CHF follow up. (Circ J 2009; 73: 1074 – 1079)
Key Words: Bioimpedance; Cardiac output; Congestive heart failure; Non-invasive cardiac monitoring system;
Prediction of readmission
T
he growing geriatric population and increased number
of survivors of acute heart failure have dramatically
increased the number of outpatients with chronic
heart failure (CHF).1–2 Hemodynamic parameters, such as
stroke volume (SV), cardiac output (CO), total systemic
vascular resistance (TSVR), and plasma level of B-type
natriuretic peptide (BNP) are thought to be important for
predicting the long-term prognosis and guiding the optimal
treatment of CHF.3–5 Until recently, hemodynamic parameters could only be obtained using the invasive thermodilution method with a Swan–Ganz catheter placed in the pulmonary artery. A non-invasive and low-cost method for
measuring hemodynamic parameters would be useful for
the clinical assessment of CHF in an outpatient setting.
Several non-invasive technologies for measuring hemodynamics are now available. Non-Invasive Cardiac System
(NICaS) (NI Medical; Hod-Hasharon, Israel) is a new
device for calculating SV, CO, and TSVR utilizing whole
body bioimpedance cardiography with electrodes placed on
one wrist and on the contralateral ankle. We previously
(Received September 8, 2008; revised manuscript received January 4,
2009; accepted January 20, 2009; released online April 16, 2009)
Division of Cardiovascular Medicine, Department of Internal Medicine, Kobe University Graduate School of Medicine, Kobe, Japan
Conflict of Interest: none declared
Mailing address: Junya Shite, MD, Division of Cardiovascular and
Respiratory Medicine, Department of Internal Medicine, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku,
Kobe 650-0017, Japan. E-mail: [email protected]
All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected]
established that NICaS-derived CO (NI-CO) is a reliable
parameter when compared with thermodilution CO or modified Fick CO6
The purpose of the present study was to evaluate the
feasibility of using NICaS in outpatients with CHF and to
assess whether NICaS-derived hemodynamic parameters in
CHF are clinically significant by relating them to other conventional cardiovascular functional indices, and to assess
the predictive accuracy of NICaS-derived parameters for
CHF-related readmission.
Methods
Patient Selection and Exclusion Criteria
In the present study, patients who met the following criteria were enrolled: (1) hospitalized because of acute CHF;
(2) had a stable condition after optimal medical therapy
including diuretics, vasodilators, ACE-inhibitors; (3) had no
renal dysfunction (defined as a creatinine value >2.0 mg/dl);
and (4) had no significant severe valve diseases. The cardinal manifestations of heart failure were dyspnea and fatigue
caused by cardiac dysfunction.
The study population comprised 68 CHF patients (56
men, 12 women, mean age 64.9±10.8 years); 20 patients
were classified as New York Heart Association (NYHA)
class I and 48 were NYHA class II patients (Table 1). The
Institutional Ethics Committee of the Kobe University
Hospital approved the study protocol and all patients provided informed consent to participate in the study.
In most patients, CHF coexisted with multiple underlying heart diseases, as shown in Table 1. Three patients
Circulation Journal Vol.73, June 2009
44 of 158.
1075
Bioimpedance Monitoring for Outpatient Heart Failure
had atrial fibrillation when CO was measured. In addition,
20 cases (29.4%) had a moderate degree of mitral valve
regurgitation, 7 cases (10.3%) had a moderate degree of
aortic valve regurgitation, and 1 case (1.5%) had moderate
aortic valve stenosis. The cardiovascular medicines taken
by the patients comprised: diuretics (n=44, 68.7%), digitalis
(n=7, 10.9%), angiotensin-converting enzyme inhibitors (n=
28, 43.7%), angiotensin receptor blockers (n=42, 65.6%),
beta blockers (n=53, 82.8%), nitrates (n=8, 12.5%), calcium
channel blockers (n=13, 20.3%), and oral inotropic agents
(n=10, 15.6%).
The exclusion criteria for the NICaS measurements
included restlessness and/or unstable patient condition,
severe aortic valve regurgitation and/or aortic stenosis,
aortic aneurysm, heart rate above 130 beats/min, intra- and
extra-cardiac shunts, severe peripheral vascular disease,
severe pitting edema, sepsis, and dialysis, all of which interfere with the accurate measurement of impedance-derived
SV,7 as previously described.
All patients were followed up at the outpatient clinic of
the Cardiovascular Division of Kobe University hospital and
had no walking disability. We measured the cardiac parameters once the patients’ attented the outpatient department,
which was the first routine visit to our hospital 1 month after
discharge from the hospital for the first heart failure admission. At the same visit, serum BNP levels were measured
and a questionnaire about exercise tolerance was administered.
Cardiac Indicators Derived by NICaS
To measure NI-CO, an alternating electrical current of
1.4 mA with a 30 kHz frequency was passed through the
patient via 2 pairs of tetrapolar electrodes, 1 pair placed on
the wrist above the radial artery, and the other pair placed
on the contralateral side above the posterior tibialis artery.
If the arterial pulses were either absent or of poor quality, a
second pair of electrodes was placed on the contralateral
site. The NICaS apparatus calculated the SV using Tsoglin
and Frinerman’s formula:8
SV = dR/R ×ρ× L2/Ri × (α+β)/β× KW × HF
where dR is the change in impedance, R is the basal resistance, ρ is the blood electrical resistivity, L is the patient
height, Ri is the corrected basal resistance according to
gender and age,8–12 KW is a correction of weight according
to ideal values,9–13 HF is a hydration factor that takes into
account body water composition,11α+β is equal to the ECG
R–R wave interval, andβis the diastolic time interval.
Because the NI-CO results are calculated every 20 s, the
average of 3 measurements obtained consecutively during
60 s of monitoring was considered to be the NI-CO value
for each individual case. By using the NI-CO value, the following parameters were calculated: NI-CO index (NI-CI =
NI-CO/body surface area), SV index (NI-SVI = NI-CI/heart
rate), and TSVR (NI-TSVR = Mean arterial pressure/CO*80
[dyne · s–1 · cm–5]).
Plasma BNP Measurements
Blood samples were obtained from the patient in a resting
condition after NICaS measurement. The samples were
drawn into plastic syringes, transferred to chilled siliconized
disposable tubes containing aprotinin (1,000 kallikrein inactivator in units/ml; Ohkura Pharmaceutical, Kyoto, Japan)
and ethylenediaminetetraacetic acid (1 mg/ml), immediately
placed on ice, and then centrifuged at –48°C. An aliquot of
45 of 158.
Table 1. Patient Characteristics
No. of patients
68 Men
56 (82.4%)
12 (17.6%)
Women
Age (years)
64.9±10.8
NYHA I
20 (29.4%)
NYHA II
48 (70.6%)
Hypertension 8 (11.8%)
Diabetes mellitus
17 (25.0%)
Ischemic heart disease
18 (26.5%)
Dilated cardiomyopathy
28 (41.2%)
NYHA, New York Heart Association classification
the plasma was immediately frozen at –80°C and thawed
only once at the time of the assay, which was performed
within 1 week. The plasma BNP concentration was measured
using a specific immunoradiometric assay kit (Shionogi
Co, Osaka, Japan), as previously reported.14
Echocardiography Measurement of the Ejection Fraction
(EF) and the Ratio of Early-to-Atrial Filling
of Transmitral Flow (E/A)
B-mode recordings were obtained with a commercially
available instrument (SONOS 5500™; Agilent Technologies
Inc, Palo Alto, CA, USA) operating at 2.5 MHz. Two-dimensional imaging examinations were performed in the standard
manner from apical 4 and 2 chamber views, and the left ventricular (LV) EFs were measured from B-mode images
according to the modified Simpson’s method. Also, pulsedDoppler findings of transmitral flow were obtained to calculate the E/A.
Exercise Tolerance Estimation
Each patient underwent a questionnaire-based interview
for the estimation of the exercise tolerance threshold (ET)
according to a specific activity scale15
MedCalc version 9.6 (MedCalc Software, Mariakerke,
Belgium) was used for data analysis. The quantitative data
are expressed as mean ± SD. A Student’s t-test was used for
descriptive statistics. The Mann–Whitney U-test was used
for non-parametric data sets. A two-tailed Pearson’s correlation was used to compare the relationship between LVEF,
log BNP, and NI-CI, NI-SVI, and NI-TSVR. Spearman’s
coefficient of rank correlation (rho) was used to estimate the
probability of correlations with ET. We used the multiple
regression analysis in order to detect the predictors for CHF
readmission, and also used receiver operating characteristic
curve (ROC) analysis for BNP, NI-CI, NI-SVI, NI-TSVR,
LVEF and ET to determine the thresholds between the readmission group and the non-readmission group. A two-tailed
P-value of less than 0.05 was considered to be significant.
Results
Relationship Between NICaS-Derived Hemodynamic
Parameters and LVEF, Plasma logBNP and ET
There were significant but weak correlations between
LVEF and NI-CI (LVEF =22.70+5.21*NI-CI, r=0.3148, P=
0.0089), NI-SVI (LVEF =22.80+0.38*NI-SVI, r=0.3257,
P=0.0067), NI-TSVR (LVEF =46.02–0.003*NI-TSVR, r=
0.2636, P=0.0298; Figure 1). There were significant moderate correlations between log BNP and NI-CI (log BNP =
1076
TANINO Y et al.
70
70
A
EF (%)
EF (%)
50
50
40
40
30
30
20
20
10
B
60
60
1.5
2.0
2.5
3.0
3.5
4.0
4.5
10
5.0
10
20
30
NI-CI (L·min–1 ·m–2)
y = 22.70 + 5.21 x
70
40
50
60
NI-SVI (ml/m2)
r=0.3148 P=0.0089
y = 22.80 + 0.38 x
r=0.3257 P=0.0067
C
EF (%)
60
50
40
30
20
10
1000
2000
3000
4000
5000
6000
NI-TSVR (dynes·cm–5 ·min–1)
y = 46.02 – 0.003 x
r=–0.2636 P=0.0298
Figure 1. Scatter plots showing the correlation between left ventricular ejection fraction (EF) and Non-Invasive Cardiac
System (NICaS) derived cardiac output index (NI-CI) [A]; EF and NICaS derived stroke volume index (NI-SVI) [B], EF and
NICaS derived total systemic vascular resistance (NI-TSVR) [C]. The 2-tailed Pearson’s correlation test (r) was used for the
analyses.
A
1000
100
10
1
B
10000
BNP (pg/ml)
BNP (pg/ml)
10000
1000
100
10
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1
10
20
NI-CI (L·min ·m )
–1
Log(y) = 3.20 – 0.38 x
BNP (pg/ml)
10000
r=–0.4585
30
40
50
60
NI-SVI (ml/m2)
–2
P<0.001
Log(y) = 3.20 – 0.028 x
r=–0.4737
P<0.0001
C
1000
100
10
1
1000
2000
3000
4000
5000
6000
Log(y) = 1.33 + 0.0003 x
r=0.4803
P<0.0001
Figure 2. Scatter plots showing the correlation between the logarithmic serum B-type natriuretic peptide (BNP) levels
with NI-CI [A]; BNP and NI-SVI [B], BNP and NI-TSVR [C]. The 2-tailed Pearson’s correlation test (r) was used for the
analyses. NI, non-invasive; CI, cardiac output index; SVI, stroke volume index; TSVR, total systemic vascular resistance.
46 of 158.
1077
10
A
10
9
8
ET (METs)
ET (METs)
8
7
6
7
6
5
5
4
4
3
3
2
B
9
1.5
2.0
2.5
3.0
3.5
4.0
4.5
2
5.0
10
20
30
NI-CI (L·min–1 ·m–2)
Spearman’s coefficient of rank correlation (rho) = 0.412 P=0.0025
10
9
40
50
60
NI-SVI (ml/m2)
Spearman’s coefficient of rank correlation (rho) = 0.575 P<0.0001
C
ET (METs)
8
7
6
5
4
3
2
1000
2000
3000
4000
5000
6000
Spearman’s coefficient of rank correlation (rho) = –0.357 P=0.0087
Figure 3. Scatter plots showing the correlation between exercise tolerance threshold (ET) and NI-CI [A]; ET and NISVI [B], and ET and NI-TSVR [C]. Spearman’s coefficient of rank correlation test (rho) was used for analysis. NI, noninvasive; CI, cardiac output index; SVI, stroke volume index; TSVR, total systemic vascular resistance.
Predictive Factors for CHF Readmission
During the 100±98 days (95%CI for the mean: 46–154)
follow-up, 15 patients were readmitted to our hospital
because of CHF recurrence (readmission group) whereas
53 patients were not readmitted because of CHF recurrence
(no-readmission group). There was a significant difference in
NI-CI between the readmission and non-readmission groups
(1.86±0.37 L · min–1 · m–2 vs 2.92±0.63 L · min–1 · m–2, P<
0.0001). Similarly, there were significant differences between
groups in NI-SVI (24.2±7.9 ml · min–1 · m–2 vs 40.4 ml · min–1 · m–2, P<0.0001), NI-TSVR (3,786±1,162 dynes · cm–5 · m–2 vs
2,420±583 dynes · cm–5 · m–2, P=0.0007), LVEF (27.4±8.4%
vs 39.3±11.7%, P=0.0005), ET (2.89–3.11 metabolic equivalents vs 5.00–7.00 metabolic equivalents in 95%CI for
median, P<0.0001), and BNP (696.5 pg/ml [95%CI; 487.5–
995.1] vs 98.1 pg/ml [95%CI; 69.8–138.1], P<0.0001).
Receiver operator characteristic curves for sensitivity, specificity, positive predictive value, and negative predictive
value for readmission for each parameter are shown in
Table 2. Multivariate analysis for readmission showed that
LVEF measured by echocardiography, NI-CI, NI-SVI, and
47 of 158.
4.0
3.5
3.0
2.5
E/A
3.20–0.38*NI-CI, r=–0.4585, P=0.0001), NI-SVI (log BNP =
3.20–0.028*NI-SVI, r=–0.4737, P<0.0001), and NI-TSVR
(log BNP =1.33+0.0003*NI-TSVR, r=0.4803, P<0.0001;
Figure 2). There was a significant correlation between ET
and NI-CI (rho =0.412, P=0.0025), NI-SVI (rho =0.575, P<
0.0001), and NI-TSVR (rho =–0.357, P=0.0087; Figure 3).
There was no significant relationship between ET and E/A
(Figure 4).
2.0
1.5
1.0
0.5
0.0
2
4
6
ET (METs)
8
P=NS
10
Figure 4. Scatter plots showing the correlation between exercise
tolerance threshold (ET) with E/A of transmitral flow derived by
Doppler echocardiogram. Spearman’s coefficient of rank correlation
test (rho) was used for analysis.
plasma BNP were significant predictors of readmission
(Table 3).
Discussion
We have previously reported the reliability of NICaSderived hemodynamic parameters when compared with
those obtained by the Swan–Ganz catheter. Overall, 2-tailed
1078
TANINO Y et al.
Table 2. Receiver Operating Curve Analysis for Readmission
Sensitivity
(%)
Specificit (%)
BNP 93.3
NI-CI
100 NI-SVI 80 NI-TSVR
100 LVEF 86.6
ET
100 88.7
79.3
98.1
68.6
64.2
71.4
PPV
(%)
NPV
(%)
AUC
Cut-off
70.0
97.9
57.7
100 92.3 94.4
46.7
100 40.6 94.4
52.0
100 0.95
0.93
0.91
0.89
0.80
0.88
344 pg/ml
2.49 L · min–1 · m–2
27.2 ml/m2
2,597 dyne · cm–5 · min–1
37%
4 METs
PPV, positive predictive value; NPV, negative predictive value; AUC, area under curve; BNP, B-type natriuretic peptide; NI-CI,
NICaS derived cardiac index; NI-SVI, NICaS derived stroke volume index; NI-TSVR, NICaS derived total systemic vascular resistance; LVEF, left ventricular ejection fraction; ET, exercise tolerance threshold.
Table 3. Multivariate Analysis for Factors of Readmission
Age
Gender
LVEF
NI-CI
NI-SVI
BNP
Heart rate
SBP
OR (95% confidence interval P value
1.0429 (0.9530–1.1414)
0.3235 (0.0593–1.7647)
0.8918 (0.8205–0.9693)
0.0318 (0.0023–0.4421)
0.8121 (0.6914–0.9540)
1.0068 (1.0019–1.0117)
1.0456 (0.9905–1.1038)
1.0118 (0.9711–1.0542)
0.3611 0.1923 0.007069
0.01023 0.01129 0.006053
0.1064 0.5750 OR, odds ratio; SBP, systolic blood pressure. Other abbreviations as in
Table 2.
Pearson’s correlation and Bland–Altman limits of agreement between NI-CO and thermodilution CO were r=0.91
and –1.06 and 0.68 L/min in 2SD and between NI-CO and
Fick CO, r=0.80 and –1.52 and 0.88 L/min in 2SD, respectively.6 In the present study, NICaS-derived hemodynamic
parameters were significantly correlated with LVEF, plasma
log BNP, and ET. Furthermore, NICaS-derived CI and SVI
were significant parameters for predicting the future recurrence of heart failure. An NICaS-derived CI of less than
2.49 L · min–1 · m–2 had 100.0% sensitivity and 79.3% specificity, and an NICaS-derived SVI of less than 27.2 ml · min–1 · m–2 had 80.0% sensitivity and 98.1% specificity. In
the readmission group, no cases of a preserved EF greater
than 55% were observed. ET was significantly worse in the
readmission group than in the non-readmission group. The
overall average EF was 36.68%, and the average EF was
27.40% in the readmission group and 39.30% in the nonreadmission group. In this study, the NICaS-derived indices
seem to be good predictors of readmission with CHF second
to BNP.
Serum BNP levels can be a good surrogate indicator for
the long-term prognosis of CHF patients.16 Logeart et al.
reported that a serum BNP of more than 350 ng/L at predischarge predicted a worse prognosis in outpatients with
CHF.17 According to our data, BNP had 93.3% sensitivity
and 88.7% specificity for readmission when its cut-off value
was 344 pg/ml, indicating that BNP had good predictive
power. Thus, in the present study, NICaS did not have a
decisive advantage over serum plasma BNP levels for predicting readmission. In clinical use, however, NICaS has
sufficient predictive power. BNP measurement requires a
peripheral venous blood sample, but some patients might
consider repeated blood sample collection an invasive procedure. NICaS has the advantage of not being invasive
because it only requires some electrodes to be placed on the
patient’s skin, similar to an electrocardiogram. It is also
important that NICaS is compatible with the conventional
hemodynamic indicators derived by the Swan–Ganz catheter. Other hemodynamic indices including CI, SV are used
when patients are in the intensive care unit. Once the patients
are moved to the general ward, however, these indices are
not readily available, although there are some limited
modalities similar to echocardiography. Even when the
patients are in the general ward or an outpatient department,
the hemodynamic indices derived by NICaS can be easily
taken and linked directly with the hemodynamic data derived
by conventional methods. Therefore, another advantage of
NICaS is that patient hemodynamic data obtained inside
and outside the intensive care unit can be reported using the
same scale, allowing for easy comparison.
According to the Ministry of Health, Labour, and Welfare
in Japan, the number of outpatients with CHF has been
increasing over the past several years.1–2 Recent studies have
been initiated to better determine the actual prevalence of
CHF in Japanese communities.18,19 Therefore, a less invasive method that is easier to use, sufficiently accurate, and
inexpensive is needed to categorize CHF patients according to their disease severity or probability of readmission.
The exercise tolerance test is a major index for estimating a
patient’s practical cardiovascular performance, but there
are a number of patients who cannot perform an exercise
test because of their age, disability, or poor muscle conditioning. In such cases, an alternative evaluation standard
of cardiovascular performance, such as NICaS, should be
utilized.
Study Limitations
It is widely believed that exercise capacity is related to
LV diastolic function20 but not to the LVEF, which in the
present study was closely correlated to CI and SVI. We did
not, however, find a significant relationship between E/A
and ET. Also, in our study subjects, 3 patients were thought
to have a purely diastolic dysfunction because they had a
preserved EF of more than 55% and an E/A with the transmitral Doppler echocardiogram of less than 1 during their
first visit to the outpatient department, and were not in the
readmission group. Thus, in our study subjects, diastolic
dysfunction did not have a significant effect, probably
because of the small number of subjects used.
Although NICaS might be useful for estimating cardiac
performance in patients who are not exercise tolerant or who
have difficulty walking, it might not be reliable when applied
to patients who match the exclusion criteria described above,
including those with arteriosclerosis obliterans and severe
pitting edema of the limbs.
48 of 158.
Conclusions
NICaS-derived hemodynamic parameters obtained by
monitoring whole body bioimpedance are applicable for the
non-invasive assessment of cardiac function for the follow
up of outpatients with CHF.
References
1. Okamoto H, Tsutsui H. The epidemiology of heart failure in Japan.
Nihon Rinsho Extra Issue 2007; 913: 49 – 54.
2. Okamoto H, Kitabatake A. Epidemiology of heart failure in Japan.
Nihon Rinsho 2003; 61: 709 – 714.
3. Solomon SD, Anavekar N, Skali H, McMurray JJ, Swedberg K,
Yusuf S, et al. Influence of ejection fraction on cardiovascular outcomes in a broad spectrum of heart failure patients. Circulation 2005;
112: 3738 – 3744.
4. Koglin J, Pehlivanli S, Schwaiblmair M, Vogeser M, Cremer P,
vonScheidt W. Role of brain natriuretic peptide in risk stratification
of patients with congestive heart failure. J Am Coll Cardiol 2001; 38:
1934 – 1941.
5. Berger R, Huelsman M, Strecker K, Bojic A, Moser P, Stanek B, et
al. B-type natriuretic peptide predicts sudden death in patients with
chronic heart failure. Circulation 2002; 105: 2392 – 2397.
6. Paredes OL, Shite J, Shinke T, Watanabe S, Otake H, Matsumoto D,
et al. Impedance cardiography for cardiac output estimation: Reliability of wrist-to-ankle electrode configuration Circ J 2006; 70: 1164 – 1168.
7. Cotter G, Moshkovitz Y, Kaluski E, Kohen A, Miller H, Goor D, et
al. Accurate, noninvasive continuous monitoring of cardiac output by
whole-body electrical bioimpedance. Chest 2004; 125: 1431 – 1440.
8. Cohen AJ, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J,
coronary artery bypass grafting. Eur J Surg 1998; 14: 64 – 69.
9. Organ LW, Bradham GB, Gore DT, Lozier SL. Segmental bioelectrical
impedance analysis: Theory and application of a new technique. J Appl
Physiol 1994; 77: 98 – 112.
10. Lukaski H, Bolonchuk WW, Hall CB, Siders WA. Validation of tetrapolar bioelectrical impedance method to assess human body composition. J Appl Physiol 1986; 60: 1327 – 1332.
49 of 158.
1079
11. Hoffer EC, Meador CK, Simpson DC. A relationship between whole
body impedance and total body water volume. Ann NY Acad Sci 1970;
170: 452 – 461.
12. Ward LC, Heitmann BL, Craid P, Stroud D, Azinge EC, Jebb S, et al.
Association between ethnicity, body mass index, and bioelectrical
impedance: Implications for the population specificity of prediction
equations. Ann NY Acad Sci 2000; 904: 199 – 204.
13. Hamwi GT, Danowski TS. Changing dietary concepts in diabetes
mellitus: Diagnosis and treatment. New York: American Diabetes
Association; 1964; 73 – 78.
14. Yasue H, Yoshimura M, Sumida H, Kikuta K, Kugiyama K, Jougasaki
M, et al. Localization and mechanism of secretion of B-type natriuretic peptide in comparison with those of A-type natriuretic peptide
in normal subjects and patients with heart failure. Circulation 1994;
90: 195 – 203.
15. Sasayama S, Asanoi H, Ishizaka S, Miyagi K. Evaluation of functional
capacity of patients with congestive heart failure. In: Yasuda H,
Kawaguchi H, editors. New aspects in the treatment of failing heart.
Tokyo: Springer-Verlag; 1992; 113 – 117.
16. Tsutamoto T, Wada A, Maeda K, Hisanaga T, Mabuchi N, Hayashi
M, et al. Plasma brain natriuretic peptide level as a biochemical
marker of morbidity and mortality in patients with asymptomatic or
minimally symptomatic left ventricular dysfunction: Comparison with
plasma angiotensin II and endothelin-1. Eur Heart J 1999; 20: 1799 – 1807.
17. Logeart D, Thabut G, Jourdain P, Chavelas C, Beyne P, Beauvais F.
B-Type natriuretic peptide assay for identifying patients at high risk
of re-admission after decompensated heart failure. J Am Coll Cardiol
2004; 43: 635 – 641.
18. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita
A; The JCARE-CARD Investigators. Clinical characteristics and
outcome of hospitalized patients with heart failure in Japan. Circ J
2006; 70: 1617 – 1623.
19. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita
A; JCARE-GENERAL Investigators. Characteristics and outcomes
of patients with heart failure in general practices and hospitals. Circ J
2007; 71: 449 – 454.
20. Parthenakis FI, Kanoupakis EM, Kochiadakis GE, Skalidis EI, Mezilis
NE, Simantirakis EN. Left ventricular diastolic filling pattern predicts cardiopulmonary determinants of functional capacity in patients
with congestive heart failure. Am Heart J 2000; 140: 338 – 344.
Europace (2007) 9, 744–750
doi:10.1093/europace/eum086
Impedance cardiography: a useful and reliable tool in
optimization of cardiac resynchronization devices
Konstantin M. Heinroth*, Marcel Elster, Sebastian Nuding, Frithjof Schlegel, Arnd Christoph,
Justin Carter, Michael Buerke, and Karl Werdan
Department of Medicine III/Cardiology, Martin-Luther-University Halle-Wittenberg, Ernst-Grube-Straße 40,
06097 Halle (Saale), Germany
Received 18 October 2006; accepted after revision 7 April 2007; online publish-ahead-of-print 11 May 2007
KEYWORDS
Aims Optimizing cardiac resynchronization therapy (CRT) devices has become more complex since
modification of both atrioventricular (AV) and interventricular (VV) stimulation intervals has become possible. The current paper presents data from the routine use of impedance cardiography (IC)-based cardiac
output (CO) measurements to guide the optimization of AV- and VV-interval timing of CRT devices.
Methods and results Forty-six patients with heart failure (left ventricular ejection fraction ,35%, New York
Heart Asociation (NYHA) III–IV) and left bundle branch block (.130 ms) in sinus rhythm were evaluated 3–5
days after implantation of a CRT device by means of IC. CO was measured without pacing and with biventricular pacing using a standard protocol of VV- and AV-interval modification from 260 to þ60 ms and 80 to
140 ms, respectively, in 20 ms steps. Mean CO without pacing was 3.66 + 0.85 L/min and significantly
increased to 4.40 + 1.1 L/min (P , 0.05) with simultaneous biventricular pacing and an AV interval of
120 ms. ‘Optimizing’ both VV and AV intervals further increased CO to 4.86 + 1.1 L/min (P , 0.05).
Maximum CO was measured in most patients with left ventricular pre-excitation. The proportion of ‘nonresponders’ to CRT was reduced by 56% following AV- and VV-interval modification using IC guidance.
Conclusion Modification of both AV and VV intervals in patients with a CRT device significantly improves CO
compared with standard simultaneous biventricular pacing and no pacing. IC is a useful non-invasive technique for guiding this modification. Marked variability of optimal AV and VV intervals between patients
requires optimization of these intervals for each patient individually.
Introduction
Cardiac resynchronization therapy (CRT) has developed from
an experimental method1 to an established adjunctive
treatment for patients with advanced heart failure. CRT
aims to improve cardiac output (CO) by lessening the interand intra-ventricular conduction delay caused in part by
left bundle branch block (LBBB).2,3 CRTreduces clinical symptoms of heart failure and hospitalization,4–6 improves haemodynamic parameters (including ventricular performance),7
and has been shown to reduce mortality.8–10 Acceptance of
this therapy is reflected by its incorporation into current
guidelines for the management of heart failure.11,12
Successful CRT in any given patient depends upon many
variables, such as the appropriate positioning of the LV
lead,13–15 and also on achieving optimal biventricular stimulation timing.16,17 This latter variable has attracted particular interest recently in an attempt to reduce the substantial
proportion (20–30%) of patients who derive no apparent
benefit from CRT, despite meeting appropriate implantation
criteria and having had technically straightforward device
implantation (‘non-responders’).18
* Corresponding author. Tel: þ49 345 557 2601; fax: þ49 345 557 2072.
E-mail address: [email protected]
CRT devices have become significantly more complex
recently. Many now have the facility to programme different
atrioventricular (AV) and interventricular (VV) stimulation
timing intervals.19,20 Data exist to show that tailoring
these intervals to suit the patient in hand further improves
the haemodynamic benefits brought by CRT. However, questions remain as to which method of haemodynamic assessment is best for guiding the adjustment of the device
settings to suit any given post-implant patient.
In the measurement of cardiac function, invasive methods
(e.g. dp/dt estimation of contractility or the thermodilution
method for CO) are the gold standards but are not suitable
for routine use during routine follow-up of patients with
implanted CRT devices. Non-invasive methods used in optimizing device settings include echocardiographic techniques,21–24 radionucleotide ventriculography,25 finger
photoplethysmography,26 and more recently, impedance
cardiography (IC).6,27
IC is an established technique for haemodynamic assessment and is capable of calculating CO on a beat-to-beat
basis.28 It relies upon changes in impedance (resistance) to
current flow through the chest between strategically
placed electrodes. Given that most current takes the path
& The European Society of Cardiology 2007. All rights reserved. For Permissions, please e-mail: [email protected]
50 of 158.
Downloaded from http://europace.oxfordjournals.org/ by guest on May 25, 2012
Cardiac resynchronization
therapy;
Non-invasive optimization;
Impedance cardiography
745
of least resistance (principally the blood-filled aorta) and
that the impedance changes with blood flow, these variables
can be used to calculate CO. IC has been used in patients
receiving dual chamber pacemakers29–31 and in the assessment of patients with heart failure.6,32–34 IC as a technique
may supplement or even replace invasive measurements and
Doppler echocardiography6,27,35 in the optimization of
implanted CRT devices.
This paper presents data from the routine use of IC in
optimizing CRT by adjusting AV and VV intervals to obtain
maximum CO in post-implant patients. To our knowledge,
this is the first report on the combined manipulation of
AV- and VV-interval timing in biventricular pacing devices
using IC.
Methods
Patient characteristics
Optimization of biventricular pacing was performed using a commercially available system for IC (Task Force Monitor Systems,
CNSystems, Graz, Austria) as described by Braun et al. 6 Two electrodes were placed bilaterally to the inferior chest wall in combination
with one electrode at the neck. Low-amplitude high-frequency
current was delivered via these surface electrodes, and transthoracic impedance (resistance) to this current flow was measured.
Changes in transthoracic impedance (mainly influenced by changes
of systolic aortic blood flow) were measured by means of four
additional surface electrodes: one pair placed bilaterally to the
sternum and the second pair bilaterally to the abdomen. Cardiac
output was calculated on a beat-to-beat basis from the transthoracic impedance signal.37 Figure 1 shows an example of the IC
measurement acquired by the Task Force Monitor System.
Device implantation
All patients received a CRT device in combination with a
cardioverter–defibrillator (Contak Renewal, Guidant, St. Paul, MN,
USA) except one patient who received only a CRT device (Contak
Renewal). Forty-five of the 46 CRT devices were implanted in our
catheterization laboratory by a cardiologist and a cardiac surgeon.
The other patient received a device with epicardial leads implanted
during surgical mitral valve reconstruction. Device implantation was
similar in all cases. Initially the right ventricular (RV) lead was
placed in the RV apex, then the coronary sinus lead was positioned
using the over-the-wire technique, and finally the right atrial lead
was implanted. The device was programmed in DDD-Mode with a
lower rate limit at 40–50 bpm, producing atrial synchronous
Table 1 Patient characteristics
Patients
Age (years)
Gender (female/male)
% CAD/% DCM
QRS
PQ
LVEF
51 of 158.
46
63 + 9
9/37
20/80
187 + 26 ms
197 + 40 ms
27 + 8%
Figure 1 Example of the impedance cardiography measurement
acquired by the Task Force Monitor System. There are two paced
beats and one beat without pacing. From top to bottom: lead II
from the surface ECG, blood pressure, and the first derivative dZ/
dt of the impedance. DBP, diastolic blood pressure; SBP, systolic
blood pressure; B, aortic valve opening; dZ/dt max, maximum of
the first derivative of the impedance signal; X, aortic valve closure.
Forty-six consecutive patients (37 males, 9 females, mean age
63 + 9 years) with heart failure [left ventricular ejection fraction
(LVEF) ,35%; NYHA III–IV], LBBB (.130 ms), and sinus rhythm
were evaluated 3–5 days after implantation of a CRT device.
Baseline characteristics of study patients are detailed in Table 1.
All patients had been receiving optimal (guideline compliant11,12)
medical treatment for heart failure for at least 1 month before
device implantation. Evaluation of patients before CRT included
12-lead surface ECG. Echocardiography was undertaken preimplant, and post-CRT device optimization according to a standard
template was performed with a System FiVe (GE VingMed) machine
coupled to a 2.5 MHz transducer. In all patients, measurement of
left ventricular (LV) dimensions, LVEF and evaluation for valve
disease, particularly mitral regurgitation was undertaken.
Transmitral flow was assessed using pulsed wave Doppler imaging
in the apical 4-chamber view. Coronary angiography was performed
in patients who had not previously had this done or where patients
had symptoms of active coronary disease requiring further imaging.
The echocardiography and coronary angiography form part of our
standard patient ‘work-up’ for device implantation, but these
results were not part of the prospectively chosen datasets gathered
for the purpose of the study.
biventricular tracking of the intrinsic sinus rhythm (VAT-Mode).36
The AV interval was set at 120 ms as a standard value6 without LV
or RV pre-excitation (VV interval ¼ 0). This ‘standard’ pacing
set-up was kept from implantation until optimization.
746
Pacing study protocol
There were no significant differences between the heart
rates without pacing (75.1 + 10.7 bpm), with initial simultaneous biventricular pacing (74.7 + 10.9 bpm), or optimized biventricular pacing (74.8 + 10.2 bpm, P . 0.05).
Thus, results for CO obtained at each CRT set-up were unaffected by the heart rate.
Cardiac output without pacing vs. simultaneous
biventricular pacing vs. optimized biventricular
pacing
The alternation between three cycles of no pacing with simultaneous biventricular pacing during stabilization and equilibration revealed (across all patients) a mean increase in CO
of 21.6 + 22.1% (P , 0.05) with pacing. The mean range
(i.e. variation) between the three recordings (in each
patient) without pacing was 10.1% and between the three
recordings with pacing was 11.8%.
Figure 2 shows data obtained for a typical patient at
different VV intervals with a fixed AV interval of 120 ms. In
this patient, the maximum CO was measured during LV preexcitation of 20 and 40 ms relative to RV stimulation. Still,
earlier LV pre-excitation, simultaneous biventricular
pacing, and RV pre-excitation resulted in a lower CO.
Data collected from all patients at each stimulation set-up
with VV optimization at a fixed AV interval of 120 ms and
with AV optimization at a VV interval set to zero are
shown in Figure 3.
In detail, the maximum CO was achieved with the following VV intervals (when combined with an AV interval of
120 ms): 260 ms in 15 pts, 240 ms in 6 pts, 220 ms in
9 pts, +0 ms in 5 pts, þ20 ms in 2 pts, þ40 ms in 2 pts,
and þ60 ms in 3 pts, so the majority of patients (30 of
46 patients) achieved peak CO with LV pre-excitation. The
remaining four patients achieved maximum CO with simultaneous biventricular pacing, two with an AV interval of
80 ms, and two with an AV interval of 140 ms.
The combination of an AV interval of 120 ms and LV
pre-excitation of 40 ms yielded the highest mean CO of
Statistics
All data are presented as means + standard error. Statistical
analysis was performed using multiple analysis of variance,
Kruskall–Wallis test, and Fisher’s exact test to compare
more than two sets of data. For a comparison of two sets
of data, a student’s t-test was performed. A P-value of
, 0.05 was considered to be significant. Data was processed
using commercially available software (Statgraphics Plus for
Windows).
Results
Cardiac resynchronization therapy
Forty-six CRT devices were successfully implanted, 45 by the
transvenous approach and 1 by the transthoracic approach
during surgical mitral valve repair. The LV lead was placed
in a posterolateral (n ¼ 35) or an anterolateral (n ¼ 10)
side branch. There were no complications.
52 of 158.
Figure 2 Cardiac output measured by impedance cardiography at
different ventriculo-ventricular intervals at an atrioventricular
interval of 120 ms in a representative patient. Optimum cardiac
output was yielded at a left ventricular pre-excitation of 240 and
220 ms.
To optimize the CRT set-up, all patients were examined in
the supine position in a silent environment to reduce the
impact of sympathetic activation by external stimuli.
A standard protocol involving a period of stabilization and
equilibration followed by VV-interval optimization and then
AV-interval optimization was employed in all patients.
Pacemakers were programmed in a DDD-Mode with a lower
rate limit of 40 bpm to avoid effects of atrial pacing on
the AV interval.36 During data acquisition, telemetry
between the implanted device and the programmer was disconnected to prevent interference with the measurement of
impedance. AV-interval values relate to the atrio-RV stimulation interval, and VV intervals relate to interventricular
interval with negative figures implying LV pre-excitation.
The first stage of the pacing protocol was a period of
stabilization and equilibration. The baseline CO without
pacing was recorded, and this was alternated three times
with ‘standard’ simultaneous biventricular pacing with an
AV interval of 120 ms for 50 s at each setting. Once consistent values for CO in both modes were confirmed, we proceeded to the ‘optimization’ stage of the pacing protocol.
With the AV interval fixed at 120 ms, VV intervals were
adjusted through a set of seven steps ranging from LV preexcitation of 260 ms to an LV delay of þ60 ms in steps of
20 ms (as reported previously38). The CO was recorded for
50 s at each setting. Then, the VV interval was set at zero
while the AV interval was adjusted from 80 to 140 ms stepwise in 20 ms increments. Again, CO at each setting was
recorded. We found that 10 s were sufficient for stabilization
in values after any change in the stimulation set-up. Finally,
the device was set at whichever combination of AV and VV
intervals produced highest CO in that patient.
We defined an increase in CO of greater than 10% above
baseline (i.e. without pacing) as a ‘positive response’ to
biventricular pacing, in keeping with data previously published.39 Where maximum CO was identified by IC to occur
at very short AV intervals (i.e. A to RV or A to LV of
,80 ms), an echocardiogram was performed to exclude a
truncated A wave caused by atrial contraction against
closed AV valves.
K.M. Heinroth et al.
747
Figure 3 Cardiac output measured by impedance cardiography at different ventriculo-ventricular and atrioventricular intervals for all
patients. Data are presented as mean + standard deviation. Maximum cardiac output was yielded at a left ventricular pre-excitation of
240 ms and an atrioventricular interval of 120 ms.
4.57 + 1.1 L/min. Similar values were obtained by LV preexcitation of 20 and 60 ms (4.51 + 1.3 and 4.50 + 1.2 L/
min, respectively). The optimal pacing set-up varied
widely from patient to patient. As a result of this variability,
there was no significant difference in CO when the results
from different AV and VV intervals were compared. No
single combination of AV and VV intervals could be recommended for application to the whole study population,
because no single combination of intervals showed statistically significant superiority over other combinations.
For the population as a whole, when the CO obtained with
‘optimized’ biventricular pacing (i.e. the CO measured at
whichever AV and VV intervals produced the highest value
in that patient) is compared with the CO obtained with
‘standard’ simultaneous biventricular pacing and compared
with the CO obtained with no pacing, a statistically significant difference is seen (Figure 4). CO without pacing
was 3.66 + 0.85 L/min. CO increased to 4.40 + 1.1 L/min
(P , 0.05) with simultaneous biventricular pacing using a
standard AV interval of 120 ms.
The mean increase in CO changing from no pacing to simultaneous biventricular pacing was 21.6 + 22.1%, P , 0.05).
53 of 158.
‘Optimizing’ VV and AV intervals further increased the mean
CO to 4.86 + 1.1 L/min. This corresponds to an increase in
mean CO of 32.8% without pacing and an increase of 11.2%
compared with simultaneous biventricular pacing, all three
set-ups resulting in significantly different CO (P , 0.05,
Figure 4).
Taking an increase in CO 10% as a definition of a positive
haemodynamic response to CRT,39 16 of the 46 patients
(35%) were ‘non-responders’ with standard simultaneous
biventricular CRT. In five of these patients, CO with standard
biventricular pacing was lower than without any pacing. The
mean CO in this group of 16 ‘non-responders’ (with standard
simultaneous biventricular CRT) was 100.8 + 4.9% (where
CO without pacing was 100%). After optimization, 9 of
these 16 patients (who were ‘non-responders’ to standard
simultaneous biventricular CRT) experienced an increase
of 10% in CO to become ‘responders’. In this group of
nine patients, the mean CO increased significantly from
101.3 + 3.6% to 121 + 5.1% (P 0.05). In the seven patients
who failed to respond despite ‘optimization,’ the mean CO
was 106.6 + 2.6% with optimized pacing. Nevertheless,
after modification of AV and VV intervals to produce the
Figure 4 Cardiac output measured by impedance cardiography without pacing, with simultaneous biventricular pacing (and an atrioventricular interval of 120 ms), and with ‘optimized’ biventricular pacing. Asterisk represents P ,0.05, simultaneous biventricular pacing vs. no
pace; ash represents P ,0.05, optimized biventricular pacing vs. simultaneous biventricular pacing and vs. no pace.
748
maximum CO, the number of haemodynamic ‘nonresponders’ was reduced to 7 of the 46 patients, i.e. the
number of non-responders was reduced by 56% (Figure 5).
Discussion
In the present study, we described the use of IC to guide
optimization of the AV and VV timing following CRT. To our
knowledge, this is the first report of both AV and VV manipulation using IC as a guide in CRT optimization. We found IC to
be simple to apply and capable of yielding rapid results in a
reliable and reproducible fashion. We found that the
optimal CRT-timing settings varied substantially between
patients underlining the need to optimize each patient’s
device individually in order to gain most benefit from
device. Using this technique, we were able to significantly
improve CO by manipulating both AV and VV intervals. By
optimizing both settings, we were able to further improve
the CRT-related increase in CO in ‘responders’ and, furthermore, could produce a significant increase in CO in patients
who, with ‘standard’ or simultaneous biventricular pacing,
had demonstrated no increase in CO (‘non-responders’). In
part, this reflects the importance of ‘electrical repositioning’ of the LV lead by VV-interval manipulation. Thus,
using this technique, we were able to reduce the rate of
non-responders to CRT by 56% (from 35 to 15%) by IC haemodynamic criteria.
Assessment of optimal biventricular stimulation:
current methods
Although invasive measurements of CO and other parameters remain the ‘gold standard’ for the evaluation of
heamodynamics, they are not suitable for routine follow-up
and optimization of CRT device settings because such invasive procedures are unpleasant for the patient and have
potentially serious complications.8,40
54 of 158.
Most units favour non-invasive methods for the optimization of CRT devices. The use of echocardiographic techniques for this purpose has received more attention in the
literature, although there is no consensus as to which of
the many echo-based parameters is the best surrogate for
CO in the context of CRT optimization.
Mitral41 and aortic42 valve Doppler velocity time integrals
(VTI) as well as several other24 echocardiographic parameters have been assessed as an alternative to invasive
measures and as a surrogate for CO in the optimization of
pacing devices. Recently, there has been promising data
using advanced tissue Doppler imaging techniques and realtime 3-D echo (particularly in selecting the optimal lead
positioning for biventricular pacing), but despite these
advances, the echocardiographic lead optimization of CRT
remains complex and time consuming. It requires an experienced operator and has significant problems relating to
reproducibility and objectivity.
Impedance cardiography: a reliable
alternative technique?
Several studies have demonstrated that IC is capable of providing a reliable and accurate measurement of CO when
compared with invasive methods.28,32,43 The utility of IC
has been shown in patients with decompensated cardiac
failure33 and in optimizing dual chamber pacemakers.44
A direct comparison of IC and echo techniques in optimizing
AV intervals for pacemakers has been made in several
studies30,31 with a recurring theme being that optimal AV
intervals calculated by IC tend to be shorter than those
obtained by echocardiography.45 Recently, Braun et al. 6 provided data using IC to guide the manipulation (primarily) of
AV intervals and showed that IC-based optimization is comparable with transaortic VTI, but IC was felt to be more sensitive to small changes in CO and easier to apply.
Limitations of impedance cardiography
The limitations of IC was usefully reviewed by Kinderman.46
In a study of 14 patients with dual chamber pacemakers,
data from IC overestimated CO if very short AV intervals
were programmed. This phenomenon seems to be attributable to a decrease in the thoracic impedance caused by a
retrograde flow into the great thoracic veins induced by
atrial contraction against closed AV valves. In these cases,
optimization of AV interval solely according to the impedance signal would result in a truncated A wave of the transmitral flow, causing potentially deleterious effects. In our
study, if IC indicated that optimum CO was seen at more
than one AV or VV interval, the longer AV interval and the
shorter VV interval were programmed. If IC indicated
optimum CO at short AV intervals (below 80 ms), transmitral
flow was checked by Doppler echocardiography to exclude a
truncated A wave. Of the 46 patients in the present study,
this echocardiographic ‘check’ was required in two,
neither of whom required a lengthening of AV interval.
Importance of optimizing both
VV and AV intervals
The early generations of CRT devices enabled programming
of the AV interval only.
Figure 5 Comparison of the percentage of ‘non-responders’ to
cardiac resynchronization therapy (definition see text) with simultaneous biventricular pacing (with an atrioventricular interval of
120 ms) vs. atrioventricular- and ventriculo-ventricular-interval
optimized biventricular pacing according to the cardiac output
measured by impedance cardiography for the whole study
population.
55 of 158.
still higher CO values, and hence hopefully reduce the
number of non-responders. Further studies are required.
Conclusions
The present paper is the first to report the utility of IC in the
optimization of both AV and VV intervals in biventricular
pacemaker set-up in routine clinical practice. Using IC, we
demonstrated that manipulating both AV and VV intervals
is feasible and may result in a significant improvement in
the haemodynamic response to CRT compared with ‘standard’ simultaneous biventricular pacing and no pacing. The
results show the importance of optimizing CRT set-up in
each patient individually, since the variability between
patients’ means that no single set of intervals can be identified as being suitable for all or even most patients.
In our opinion, IC is a promizing method of CRT optimization and may compare well with other techniques in
routine use. It carries less risk of complications than invasive
techniques and is more comfortable for the patient. The
validity and reliability of IC has now been reported by
several authors.28,31–33,43 Notwithstanding its limitations,
which include a tendency to relative CO overestimation
and the potential for selecting inappropriately short AV
intervals, IC has an emerging role in the optimization of
CRT parameters in routine clinical practice. Further work
is required to make best use of IC derived data in combining
AV and VV intervals to ideally suit any given patient.
Conflict of interest: none declared.
References
1. Leclercq C, Cazeau S, Le BH, Ritter P, Mabo P, Gras D et al. Acute hemodynamic effects of biventricular DDD pacing in patients with end-stage
heart failure. J Am Coll Cardiol 1998;32:1825–31.
2. Cohen TJ, Klein J. Cardiac resynchronization therapy for treatment of
chronic heart failure. J Invasive Cardiol 2002;14:48–53.
3. Cazeau S, Alonso C, Jauvert G, Lazarus A, Ritter P. Cardiac resynchronization therapy. Europace 2004;5(Suppl. 1):S42–8.
4. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh E et al.
Cardiac resynchronization in chronic heart failure. N Engl J Med
2002;346:1845–53.
5. Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C et al.
Effects of multisite biventricular pacing in patients with heart failure
and intraventricular conduction delay. N Engl J Med 2001;344:873–80.
6. Braun MU, Schnabel A, Rauwolf T, Schulze M, Strasser RH. Impedance
cardiography as a noninvasive technique for atrioventricular interval
optimization in cardiac resynchronization therapy. J Interv Card
Electrophysiol 2005;13: 223–229.
7. Bakker PF, Meijburg HW, de Vries JW, Mower MM, Thomas AC, Hull ML
et al. Biventricular pacing in end-stage heart failure improves functional
capacity and left ventricular function. J Interv Card Electrophysiol 2000;
4:395–404.
8. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De MT et al.
Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 2004;
350:2140–50.
9. Salukhe TV, Dimopoulos K, Francis D. Cardiac resynchronisation may
reduce all-cause mortality: meta-analysis of preliminary COMPANION
data with CONTAK-CD, InSync ICD, MIRACLE and MUSTIC. Int J Cardiol
2004;93:101–3.
10. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappenberger L
et al. The effect of cardiac resynchronization on morbidity and mortality in
heart failure. N Engl J Med 2005;352:1539–49.
11. Hoppe UC, Bohm M, Dietz R, Hanrath P, Kroemer HK, Osterspey A et al.
Guidelines for therapy of chronic heart failure. Z Kardiol 2005;
94:488–509.
Prospective randomized trial data have shown that
AV-interval optimization not only improves the haemodynamic response to CRT, but also improves NYHA functional
class and quality of life scores.47
The current generation of devices allows the manipulation
of both AV and VV intervals.
Since activation of the interventricular septum is mainly
influenced by the RV lead, modification of the VV interval
affects not only inter-ventricular dyssynchrony, but also
intra-ventricular dyssynchrony of the LV. The capacity to
manipulate VV intervals may thus further improve the
haemodynamic response to CRT.
There is increasing evidence that sequential biventricular
pacing is superior to simultaneous biventricular pacing in
many patients with a CRT device.
Perego et al. examined the impact of sequential biventricular pacing while invasively monitoring dp/dt in both ventricles simultaneously. Simultaneous ventricular pacing
produced an increase in dp/dt of 29% over baseline, but
sequential pacing produced a significantly greater increase
of 35% (P , 0.01). The mean optimal VV interval was LV preexcitation of 25 ms, and there was no detriment to RV function.38 Similar findings were reported by van Gelder et al. 20
This pattern of haemodynamic improvement with both AV
and VV optimization has been verified using Doppler echocardiography,48 3-D echocardiography49 and radionuclide
ventriculography.25 The largest benefit is obtained in most
cases by LV pre-excitation. Our study is the first to validate
the previous data, showing the additional benefit of optimizing both AV and VV intervals using IC as the method of
assessment.
Where no facility for optimizing CRT devices in individual
patients exists, we would recommend an AV interval of
120 ms and LV pre-excitation of 20–40 ms as empirical
intervals in CRT programming. However, due to significant
heterogeneity between patients, we were unable to identify
any single set of AV and VV intervals as being superior and
thus suitable for application to the whole CRT population
on an empirical basis. This variation between patients is to
be expected, since there is a substantial heterogeneity
between patients of conduction pattern, ventricular size,
scar tissue formation, ventricular function (especially with
respect to regional wall motion abnormalities) CRT lead placement positions, and other variables. The facility to manipulate VV intervals in effect allows the operator to
‘electrically reposition’ the LV lead to help overcome some
of these variations to maximize the haemodynamic response
to the device.
This work has been an early, descriptive study of the feasibility of using IC to manipulate both AV and VV intervals in
optimizing CRT devices. Further work is required before
this technique can enter mainstream use. For instance, in
any given patient, since AV and VV intervals are likely to
be interdependent, the ideal IC protocol would test every
AV-interval setting against every VV-interval setting and
vice versa. With current equipment, this would be impractical. Another simpler method of combining these interdependent values of AV and VV intervals would be to identify the
optimal VV interval (i.e. A to LV interval) and keeping this
fixed (rather than re-setting this to zero as our protocol
did), and then adjusting the AV (i.e. A to V interval) to identify the optimal CO. We suspect that this method of combined optimization of AV and VV intervals by IC may yield
749
750
56 of 158.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
dual-chamber pacemakers using impedance cardiography and Doppler
echocardiography. J Clin Basic Cardiol 1999;2:237–40.
Belardinelli R, Ciampani N, Costantini C, Blandini A, Purcaro A. Comparison of impedance cardiography with thermodilution and direct
Fick methods for noninvasive measurement of stroke volume and
cardiac output during incremental exercise in patients with ischemic cardiomyopathy. Am J Cardiol 1996;77:1293–301.
Albert NM, Hail MD, Li J, Young JB. Equivalence of the bioimpedance and
thermodilution methods in measuring cardiac output in hospitalized
patients with advanced, decompensated chronic heart failure. Am J
Crit Care 2004;13:469–79.
Adachi H, Hiratsuji T, Sakurai S, Tada H, Toyama T, Naito S et al.
Impedance cardiography and quantitative tissue Doppler echocardiography for evaluating the effect of cardiac resynchronization therapy: a
case report. J Cardiol 2003;42:37–42.
Santos JF, Caetano F, Parreira L, Madeira J, Cardoso P, Fonseca N et al.
Tissue Doppler echocardiography for evaluation of patients with ventricular resynchronization therapy. Rev Port Cardiol 2003;22:1363–71.
Bernheim A, Ammann P, Sticherling C, Burger P, Schaer B, Brunner-La
Rocca HP et al. Right atrial pacing impairs cardiac function during resynchronization therapy: acute effects of DDD pacing compared to VDD
pacing. J Am Coll Cardiol 2005;45:1482–7.
Ovsyshcher I, Furman S. Impedance cardiography for cardiac output estimation in pacemaker patients: review of the literature. Pacing Clin
Electrophysiol 1993;16:1412–22.
Perego GB, Chianca R, Facchini M, Frattola A, Balla E, Zucchi S et al.
Simultaneous vs. sequential biventricular pacing in dilated cardiomyopathy: an acute hemodynamic study. Eur J Heart Fail 2003;5:305–13.
Bleeker GB, Bax JJ, Fung JW, van der Wall EE, Zhang Q, Schalij MJ et al.
Clinical versus echocardiographic parameters to assess response to
cardiac resynchronization therapy. Am J Cardiol 2006;97:260–3.
Auricchio A, Stellbrink C, Sack S, Block M, Vogt J, Bakker P et al. The
Pacing Therapies for Congestive Heart Failure (PATH-CHF) study: rationale, design, and endpoints of a prospective randomized multicenter
study. Am J Cardiol 1999;83:130D–5D.
Jansen AH, Bracke FA, van Dantzig JM, Meijer A, van der Voort PH,
Aarnoudse W et al. Correlation of echo-Doppler optimization of atrioventricular delay in cardiac resynchronization therapy with invasive hemodynamics in patients with heart failure secondary to ischemic or idiopathic
dilated cardiomyopathy. Am J Cardiol 2006;97:552–7.
Kerlan JE, Sawhney NS, Waggoner AD, Chawla MK, Garhwal S, Osborn JL
et al. Prospective comparison of echocardiographic atrioventricular
delay optimization methods for cardiac resynchronization therapy.
Heart Rhythm 2006;3:148–54.
Drazner MH, Thompson B, Rosenberg PB, Kaiser PA, Boehrer JD, Baldwin
BJ et al. Comparison of impedance cardiography with invasive hemodynamic measurements in patients with heart failure secondary to ischemic
or nonischemic cardiomyopathy. Am J Cardiol 2002;89:993–5.
Ovsyshcher I, Zimlichman R, Katz A, Bondy C, Furman S. Measurements of
cardiac output by impedance cardiography in pacemaker patients at rest:
effects of various atrioventricular delays. J Am Coll Cardiol
1993;21:761–7.
Santos JF, Parreira L, Madeira J, Fonseca N, Soares LN, Ines L. Non invasive hemodynamic monitorization for AV interval optimization in patients
with ventricular resynchronization therapy. Rev Port Cardiol
2003;22:1091–8.
Kindermann
M.
Impedanzkardiograpie.
Herzschrittmacherther
Elektrophysiol 2004;15:1102–9.
Sawhney NS, Waggoner AD, Garhwal S, Chawla MK, Osborn J, Faddis MN.
Randomized prospective trial of atrioventricular delay programming for
cardiac resynchronization therapy. Heart Rhythm 2004;1:562–7.
Riedlbauchova L, Kautzner J, Fridl P. Influence of different atrioventricular and interventricular delays on cardiac output during cardiac resynchronization therapy. Pacing Clin Electrophysiol 2005;28(Suppl. 1):
S19–23.
Sogaard P, Egeblad H, Pedersen AK, Kim WY, Kristensen BO, Hansen PS
et al. Sequential versus simultaneous biventricular resynchronization
for severe heart failure: evaluation by tissue Doppler imaging.
Circulation 2002;106:2078–84.
12. Hunt SA. ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American
College of Cardiology/American Heart Association Task Force on
Practice Guidelines (Writing Committee to Update the 2001 Guidelines
for the Evaluation and Management of Heart Failure). J Am Coll
Cardiol 2005;46:e1–82.
13. Butter C, Auricchio A, Stellbrink C, Fleck E, Ding J, Yu Y et al. Effect of
resynchronization therapy stimulation site on the systolic function of
heart failure patients. Circulation 2001;104:3026–9.
14. Gold MR, Auricchio A, Hummel JD, Giudici MC, Ding J, Tockman B et al.
Comparison of stimulation sites within left ventricular veins on the
acute hemodynamic effects of cardiac resynchronization therapy. Heart
Rhythm 2005;2:376–81.
15. Albertsen AE, Nielsen JC, Pedersen AK, Hansen PS, Jensen HK, Mortensen
PT. Left ventricular lead performance in cardiac resynchronization
therapy: impact of lead localization and complications. Pacing Clin
Electrophysiol 2005;28:483–8.
16. Auricchio A, Ding J, Spinelli JC, Kramer AP, Salo RW, Hoersch W et al.
Cardiac resynchronization therapy restores optimal atrioventricular
mechanical timing in heart failure patients with ventricular conduction
delay. J Am Coll Cardiol 2002;39:1163–9.
17. Doshi RN. Optimizing resynchronization therapy: can we increase the
number of true responders? J Cardiovasc Electrophysiol 2005;16(Suppl.
1):S48–51.
18. Birnie DH, Tang AS. The problem of non-response to cardiac resynchronization therapy. Curr Opin Cardiol 2006;21:20–6.
19. O’Donnell D, Nadurata V, Hamer A, Kertes P, Mohammed W. Long-term
variations in optimal programming of cardiac resynchronization therapy
devices. Pacing Clin Electrophysiol 2005;28(Suppl. 1):S24–6.
20. van Gelder BM, Bracke FA, Meijer A, Lakerveld LJ, Pijls NH. Effect of optimizing the VV interval on left ventricular contractility in cardiac resynchronization therapy. Am J Cardiol 2004;93:1500–3.
21. Breithardt OA, Sinha AM, Franke A, Hanrath P, Stellbrink C.
Echocardiography in cardiac resynchronization therapy: identification
of suitable patients, follow-up and therapy optimization. Herz 2003;
28:615–27.
22. Bax JJ, Ansalone G, Breithardt OA, Derumeaux G, Leclercq C, Schalij MJ
et al. Echocardiographic evaluation of cardiac resynchronization
therapy: ready for routine clinical use? A critical appraisal. J Am Coll
Cardiol 2004;44:1–9.
23. Lane RE, Chow AW, Chin D, Mayet J. Selection and optimisation of biventricular pacing: the role of echocardiography. Heart 2004;90(Suppl.
6):vi10–6.
24. Melzer C, Knebel F, Ismer B, Bondke H, Nienhaber CA, Baumann G et al.
Influence of the atrio-ventricular delay optimization on the intra left ventricular delay in Cardiac Resynchronization Therapy. Cardiovasc
Ultrasound 2006;4:5.
25. Burri H, Sunthorn H, Somsen A, Zaza S, Fleury E, Shah D et al. Optimizing
sequential biventricular pacing using radionuclide ventriculography.
Heart Rhythm 2005;2:960–5.
26. Butter C, Stellbrink C, Belalcazar A, Villalta D, Schlegl M, Sinha A et al.
Cardiac resynchronization therapy optimization by finger plethysmography. Heart Rhythm 2004;1:568–75.
27. Tse HF, Yu C, Park E, Lau CP. Impedance cardiography for atrioventricular
interval optimization during permanent left ventricular pacing. Pacing
Clin Electrophysiol 2003;26:189–91.
28. Fortin J, Habenbacher W, Heller A, Hacker A, Grullenberger R, Innerhofer J
et al. Non-invasive beat-to-beat cardiac output monitoring by an improved
method of transthoracic bioimpedance measurement. Comput Biol Med
2006;36:1185–203.
29. Ovsyshcher I, Gross JN, Blumberg S, Furman S. Precision of impedance
cardiography measurements of cardiac output in pacemaker patients.
Pacing Clin Electrophysiol 1992;15:1923–6.
30. Kindermann M, Frohlig G, Doerr T, Schieffer H. Optimizing the AV delay in
DDD pacemaker patients with high degree AV block: mitral valve Doppler
versus
impedance
cardiography.
Pacing
Clin
Electrophysiol
1997;20:2453–62.
31. Kolb HJ, Böhm U, Mende M, Neugebauer A, Pfeiffer D, Rother T. Assessment of the optimal atrioventricular delay in patients with
Cardiac power output predicts mortality across a
broad spectrum of patients with acute cardiac disease
Dorinna D. Mendoza, MD, Howard A. Cooper, MD, and Julio A. Panza, MD Washington, DC
Background Cardiac power output (CPO) is a novel hemodynamic measurement that represents cardiac
pumping ability. The prognostic value of CPO in a broad spectrum of patients with acute cardiac disease undergoing
pulmonary artery catheterization (PAC) has not been examined.
Methods
Consecutive patients with a primary cardiac diagnosis who were undergoing PAC in a single coronary
care unit were included. The relationship between initial CPO [(mean arterial pressure cardiac output [CO])/451] and
inhospital mortality was evaluated. CPO was analyzed both as a dichotomous variable (using a cutoff value previously
established among patients with cardiogenic shock) and as a continuous variable.
Results Data were available for 349 patients. The mean CPO was 0.88 F 0.37 W. The inhospital mortality rate
was significantly higher among patients with a CPO V0.53 W (n = 53) compared with those with a CPO N0.53 W (n = 296)
(49% vs 20%, P b .001). In separate multivariate analyses, both CPO and CO were associated with inhospital mortality.
However, when both terms were included simultaneously, CPO remained strongly associated with mortality (odds ratio
0.63, 95% CI 0.43-0.91, P = .01), whereas CO did not (odds ratio 1.05, 95% CI 0.75-1.48, P = .78).
Conclusions
Cardiac power output is a strong, independent predictor of inhospital mortality in a broad spectrum
of patients with primary cardiac disease undergoing PAC. (Am Heart J 2007;153:366270.)
Pulmonary artery catheterization (PAC) remains in
common usage in the coronary care unit (CCU) setting
despite questions regarding the use of this procedure in
critically ill patients.1-3 The decision to use PAC is based
on the concept that invasive hemodynamic measurements provide a more precise understanding of the
status of the cardiovascular system, which in turn allows
close monitoring of the response to therapeutic interventions. However, traditional hemodynamic measurements obtained from PAC, such as cardiac output (CO),
do not fully describe the fundamental state of the
cardiovascular system in the setting of severe illness.4 In
particular, because of its critical dependence on preload
and afterload, measured CO provides a limited estimation of ventricular function.
Cardiac power output (CPO) is a novel hemodynamic
measure that is the product of simultaneously measured
CO and mean arterial pressure (MAP).4,5 By incorporating
both the pressure and flow domains of the cardiovascular
From the Division of Cardiology, Washington Hospital Center, Washington, DC.
This work was supported by the Cardiovascular Research Institute, Washington Hospital
Center, Washington, DC.
Submitted July 13, 2006; accepted November 21, 2006.
Reprint requests: Howard A. Cooper, MD, Division of Cardiology, Washington Hospital
Center, 110 Irving Street NW, Suite NA1103, Washington, DC 20010.
E-mail: [email protected]
0002-8703/$ - see front matter
n 2007, Mosby, Inc. All rights reserved.
doi:10.1016/j.ahj.2006.11.014
57 of 158.
system, CPO is an integrative measure of cardiac
hydraulic pumping ability. Cardiac power output has
been shown to be a powerful predictor of mortality in
patients with chronic heart failure (HF) and in those with
cardiogenic shock.6-8 Recently, a report from the SHOCK
Trial Registry found that CPO was the strongest independent hemodynamic correlate of inhospital mortality
in patients with cardiogenic shock.4 We hypothesized
that CPO would provide important prognostic information across a broad spectrum of patients with acute
cardiac disease. Accordingly, we analyzed the relationship between CPO and short-term mortality in a consecutive series of patients undergoing PAC in a single CCU.
Methods
Study design and data collection
The institutional review board for human research at the
Washington Hospital Center approved this research protocol.
Data were collected prospectively for consecutive patients
admitted to the CCU between October 2002 and February
2006. Based on all available information, the attending
physician determined the primary clinical diagnosis requiring
admission to the CCU. Patients with a noncardiac primary
diagnosis were excluded from this analysis. Persistent cardiogenic shock was determined to be present by using conventional clinical criteria of hypotension and signs of peripheral
hypoperfusion in the presence of pulmonary congestion that
did not rapidly resolve. Left ventricular ejection fraction (EF)
was visually estimated from 2-dimensional echocardiography
American Heart Journal
Volume 153, Number 3
obtained during the index admission, or from left ventricular
angiography (when available) if echocardiography was
not performed.
For each patient, the first set of hemodynamic measurements obtained in the CCU was used for analysis. Measurements included those made while receiving support with
inotropes and/or an intra-aortic balloon pump. Cardiac output
was measured by the thermodilution method, taking the
average of 3 to 5 readings. Mean arterial pressure was
determined by invasive arterial monitoring if available; otherwise, automated sphygmomanometry was used. Cardiac
power output was calculated as MAP CO H 451.4
Treatment decisions were made without specific regard to
CPO measurements.
Statistical considerations
Previously, the SHOCK investigators determined that the
cutoff value of CPO that most accurately predicted inhospital
mortality in their patient cohort was 0.53 W.4 Accordingly, we
analyzed CPO both as a dichotomous variable (N0.53 W vs
V0.53 W) and as a continuous variable (per 0.2-W increments).
Baseline variables were compared using standard statistical
tests. The relationship between baseline variables and inhospital mortality was assessed with univariate logistic regression
analysis. Multivariate logistic regression analysis was performed
to examine the independent relationship between CPO (as a
continuous variable) and inhospital mortality. A comprehensive logistic regression model was constructed including the
following covariates: age, sex, race (white or nonwhite),
treated diabetes mellitus, current or recent smoking, chronic
renal failure requiring dialysis, prior coronary artery bypass
grafting, intra-aortic balloon counterpulsation, mechanical
ventilation, inotrope use, persistent cardiogenic shock, EF,
heart rate, central venous pressure, pulmonary capillary wedge
pressure, systemic vascular resistance, and CPO. A reduced
model was then created that included only those covariates
significantly associated with inhospital mortality. Cardiac
output was subsequently forced into this reduced model. A
2-sided P value less than .05 was considered to represent
statistical significance.
Results
Baseline characteristics
Complete data were available for 349 patients. The
primary admission diagnoses are listed in Table I. The
most common diagnoses were acute myocardial infarction, cardiomyopathy, coronary atherosclerosis, and
unstable angina. Baseline patient characteristics and
hemodynamics are presented in Table II. The mean age
was 64 years, 59% were men, and 39% were of nonwhite
race. Patients had severe illness, as evidenced by the fact
that 50% required an intra-aortic balloon pump, 51%
required mechanical ventilation, and 55% required
intravenous inotropic support.
The CPO ranged from 0.17 to 2.48 W, with a mean of
0.88 F 0.37 W. Compared with patients with a CPO
N0.53 W (n = 296), patients with a CPO V0.53 W (n = 53)
were older (69 vs 63 years, P = .004) and more likely to
58 of 158.
Mendoza et al 367
Table I. Primary diagnosis
Primary diagnosis (N = 349)
ST-segment elevation myocardial infarction
Non–ST-segment elevation
myocardial infarction
Ischemic cardiomyopathy
Nonischemic cardiomyopathy
Coronary atherosclerosis
Unstable angina
Cardiac arrest
Acute myocarditis
Hypertrophic cardiomyopathy
Ventricular fibrillation
Tako-Tsubo cardiomyopathy
Aortic stenosis
Atrial fibrillation
Mitral regurgitation
Mitral stenosis
Pericardial tamponade
Ventricular tachycardia
Complete heart block
Supraventricular tachycardia
No. of patients (%)
122 (35)
79 (23)
38
33
17
16
15
4
4
4
3
2
2
2
2
2
2
1
1
(11)
(9)
(5)
(5)
(4)
(1)
(1)
(1)
(0.9)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
(0.3)
(0.3)
be women (64% vs 34%, P b .001). As expected, those
with CPO V0.53 W had a lower MAP, CO, and cardiac
index, and a higher systemic vascular resistance (all
P b .001). Of note, the EF was similar among patients
with CPO N0.53 W and patients with CPO V0.53 W (31%
vs 28%, P = not significant).
Outcomes
Patients with a CPO V0.53 W had a significantly higher
inhospital mortality rate than those with a CPO N0.53 W
(49% vs 20%, P b .001), corresponding to a positive
predictive value of 49% and a negative predictive value
of 80%. On univariate analysis, predictors of mortality
were age, diabetes mellitus, need for mechanical
ventilation, persistent cardiogenic shock, MAP, CO,
cardiac index, heart rate, central venous pressure, and
CPO (Table III). In the final reduced multivariate model,
CPO remained independently associated with a reduced
risk of inhospital mortality (odds ratio [OR] 0.65, 95% CI
0.54-0.78, P b .001) (Table IV). Other independent
predictors were heart rate ( P = .01), diabetes mellitus
( P = .02), and the need for mechanical ventilation
( P b .001). When CO was forced into this model, the
relationship between CPO and inhospital mortality was
not substantially altered (OR 0.63, 95% CI 0.43-0.91,
P = .01), whereas there was no significant relationship
between CO and mortality (OR 1.05, 95% CI 0.75-1.48,
P = .78).
Cardiogenic shock
Persistent cardiogenic shock was diagnosed in
73 patients (21%). Mortality in patients with persistent
March 2007
368 Mendoza et al
Table II. Baseline characteristics and initial hemodynamic measurements
Characteristic
Age (y)
Male, n (%)
Nonwhite race, n (%)
History of
Coronary artery bypass grafting, n (%)
Percutaneous coronary intervention, n (%)
Diabetes mellitus, n (%)
Smoking, n (%)
Chronic renal failure, n (%)
Intra-aortic balloon pump, n (%)
Mechanical ventilation, n (%)
Acute myocardial infarction, n (%)
Persistent cardiogenic shock, n (%)
Mean arterial pressure (mm Hg)
Cardiac output (L/min)
Cardiac index (L min 1 m 2)
Systemic vascular resistance (dyne s/cm5)
Heart rate (beats/min)
Pulmonary capillary wedge pressure (mm Hg)
Central venous pressure (mm Hg)
Any inotrope, n (%)
Ejection fraction (%)
All patients (N = 349)
CPO >0.53 (n = 296)
CPO <
_ 0.53 (n = 53)
64 F 14
206 (59)
135 (39)
63 F 14
188 (64)
113 (38)
69 F 11
18 (34)
22 (42)
61 (18)
46 (13)
79 (23)
27 (8)
14 (4)
175 (50)
177 (51)
201 (58)
73 (21)
80 F 13
4.9 F 1.8
2.5 F 0.8
1297 F 597
88 F 19
20 F 7
13 F 7
193 (55)
30 F 13
51 (17)
37 (13)
67 (23)
25 (8)
13 (4)
155 (52)
147 (50)
175 (59)
60 (20)
82 F 13
5.2 F 1.7
2.6 F 0.8
1207 F 535
90 F 18
19 F 7
12 F 5
161 (54)
31 F 13
Table III. Univariate logistic regression results for inhospital
mortality
Variable
Age
Male sex
Nonwhite race
Prior coronary artery bypass grafting
Prior percutaneous coronary intervention
Diabetes mellitus
Smoker
Chronic renal failure
Acute myocardial infarction
Intra-aortic balloon pump
Mechanical ventilation
Persistent cardiogenic shock
Any inotrope
MAP
CO
Cardiac index
Systemic vascular resistance
Heart rate
Pulmonary capillary wedge pressure
Central venous pressure
EF
CPO
OR (95% CI)
1.02
0.75
0.74
1.15
1.13
1.94
0.37
0.89
1.04
1.12
4.30
2.26
1.18
0.96
0.74
0.47
1.00
1.01
1.01
1.05
1.00
0.66
(1.00-1.04)
(0.46-1.22)
(0.44-1.24)
(0.61-2.16)
(0.56-2.30)
(1.12-3.36)
(0.11-1.27)
(0.23-3.14)
(0.63-1.71)
(0.69-1.84)
(2.46-7.52)
(1.29-3.94)
(0.72-1.93)
(0.93-0.98)
(0.63-0.88)
(0.32-0.68)
(1.00-1.00)
(1.00-1.03)
(0.98-1.05)
(1.01-1.10)
(0.98-1.02)
(0.55-0.79)
(19)
(17)
(23)
(4)
(2)
(38)
(57)
(49)
(25)
F 11
F 0.7
F 0.4
F 668
F 19
F7
F 12
(60)
F 12
.004
b.001
.65
.77
.37
1.0
.40
.70
.05
.35
.17
.48
b.001
b.001
b.001
b.001
.23
.15
.12
.42
.14
Table IV. Multivariate logistic regression analysis for
inhospital mortality
P
.03
.24
.25
.66
.73
.02
.11
.81
.88
.64
b.001
.004
.52
b.001
b.001
b.001
.44
.03
.45
.01
.98
b.001
cardiogenic shock was higher than mortality among
patients without this condition (37% vs 21%, P b .001).
After excluding patients with persistent cardiogenic
shock, CPO remained independently associated with
inhospital mortality (OR 0.64, 95% CI 0.50 to 0.81,
P b .001).
59 of 158.
10
9
12
2
1
20
30
26
13
70
2.8
1.6
1824
86
21
15
32
28
P
Variable
CPO
Heart rate
Diabetes mellitus
OR (95% CI)
0.65
1.02
2.10
3.76
(0.54-0.78)
(1.01-1.04)
(1.13-3.89)
(2.08-6.82)
P
b.001
.01
.02
b.001
Discussion
To our knowledge, the current study is the largest
reported analysis of CPO in a broad spectrum of
patients with critical cardiac illness undergoing PAC.
The major findings were the following: (1) inhospital
mortality was substantially higher in patients with a low
initial CPO, (2) CPO was independently associated with
inhospital mortality after controlling for important
covariates, and (3) the association between CPO and
inhospital mortality was stronger than that of the more
traditional CO.
Our findings suggest that measurement of CPO is
useful for risk stratification in patients with primary
cardiac disease undergoing PAC in the CCU setting.
CPO represents the rate of energy input that the
systemic vasculature receives from the heart, incorporating both pressure and flow domains of the cardiovascular system.4 It is therefore logical that such a
measure of cardiac pumping ability would predict
outcomes for patients with cardiogenic shock and
severe HF, as has been demonstrated previously.4,6-8
Volume 153, Number 3
Our analysis extends these findings to a broader
spectrum of patients with critical cardiac illness,
including those with acute coronary syndromes, cardiomyopathies of various etiologies, valvular heart
disease, cardiac arrhythmias, and complicated post–
percutaneous coronary intervention courses.
As expected, patients in our analysis had CPO values
considerably higher (mean 0.88 W) than those in a
more select group with acute myocardial infarction and
cardiogenic shock in the SHOCK trial registry (mean
0.62 W). Despite the substantial differences in patient
populations and absolute CPO values, however, the
relationship between CPO and inhospital mortality was
remarkably similar in the 2 analyses: an OR of 0.65 per
0.2-W increase in CPO in the current study compared
with an OR of 0.60 in the SHOCK trial registry.4
Furthermore, among the subset of our patients diagnosed with persistent cardiogenic shock (n = 73), the
positive and negative predictive values using a cutoff of
0.53 W were nearly identical in the 2 studies (62% vs
58% and 68% vs 71%, respectively). This concordance
of results suggests both that the relationship between
CPO and short-term mortality is consistent across a
broad spectrum of cardiac diagnoses and a wide range
of CPO values, and is reproducible across studies.
Cardiac power output is calculated from the more
familiar hemodynamic variables, CO and MAP. Therefore, from a statistical standpoint, it cannot provide
additional prognostic information beyond that contained
within the 2 individual components. However, CPO
combines information from these 2 separate hemodynamic axes (flow and pressure) into a single, easily
understandable variable that can potentially be targeted
for therapy. For this reason, the concept of CPO appears
to provide clinically useful information above and
beyond that provided by standard hemodynamic variables alone.
Recent studies have questioned the use of PAC in
guiding treatment in critically ill patients, including
patients with severe CHF.3,9-13 Possible explanations for
these negative results include complications related to
the PAC procedure itself (arrhythmias, infection);
misinterpretation of the obtained data; incorrect action
in response to even appropriately interpreted data; or
unanticipated deleterious effects of treatments implemented in response to these data. In patients with
cardiac disease, treatments are typically targeted to the
CO (or cardiac index) and pulmonary capillary wedge
pressure. However, CO is a measure of cardiovascular
flow, not cardiac contractility, whereas pulmonary
capillary wedge pressure is a measure of intracardiac
pressure not directly representing cardiac performance.1 Theoretically, therefore, CPO—a measure of
cardiac pumping ability—might be a more appropriate
target for medical therapy. In addition, CPO can be
simply and accurately determined noninvasively in most
60 of 158.
Mendoza et al 369
patients by using whole-body electrical bioimpedance
to determine CO, thereby eliminating any direct
complications of PAC.14,15 Clinical studies in which
noninvasively determined CPO is targeted for treatment
appear to be indicated.
Our study has several potential limitations. First, this
was a single-center retrospective analysis, subject to the
limitations inherent in such a study design. However,
measurements of CPO were not used by the CCU
physicians to make therapeutic decisions. Hence, we
are confident that the observed prognostic value of
CPO was not influenced by bias introduced from
knowledge of this information. Second, although we
attempted to control for important differences in
baseline characteristics between patients, residual confounding might remain. Third, we were unable to
distinguish between hemodynamic measurements made
with and without support from inotropes and/or an
intra-aortic balloon pump. Because both types of
interventions are likely to increase CPO, this might
affect the relationship between CPO and mortality.
However, this relationship was maintained after including the use of these interventions in the multivariate model. Indeed, the relationship between CPO and
mortality might be stronger in the absence of such
support measures. Finally, we must emphasize that our
findings demonstrate an association between CPO and
outcome without providing a direct assessment of
whether CPO should be used as a clinical end point for
therapeutic interventions.
Conclusions
Cardiac power output is a novel and useful
prognostic tool in patients with acute cardiac disease
undergoing PAC in the CCU setting. Further studies
in which CPO is targeted for treatment appear to
be warranted.
References
1. Connors Jr AF, Speroff T, Dawson NV, et al. The effectiveness of
right heart catheterization in the initial care of critically ill patients.
SUPPORT Investigators. JAMA 1996;276:889 - 97.
2. Dalen JE, Bone RC. Is it time to pull the pulmonary artery catheter?
JAMA 1996;276:916 - 8.
3. Binanay C, Califf RM, Hasselblad V, et al. Evaluation study of
congestive heart failure and pulmonary artery catheterization
effectiveness: the ESCAPE trial. JAMA 132005;294:1625 - 33.
4. Fincke R, Hochman JS, Lowe AM, et al. Cardiac power is the
strongest hemodynamic correlate of mortality in cardiogenic shock:
a report from the SHOCK trial registry. J Am Coll Cardiol
2004;44:340 - 8.
5. Cotter G, Williams SG, Vered Z, et al. Role of cardiac power in
heart failure. Curr Opin Cardiol 2003;18:215 - 22.
6. Williams SG, Cooke GA, Wright DJ, et al. Peak exercise cardiac
power output; a direct indicator of cardiac function strongly
March 2007
370 Mendoza et al
7.
8.
9.
10.
11.
predictive of prognosis in chronic heart failure. Eur Heart J
2001;22:1496 - 503.
Tan LB, Littler WA. Measurement of cardiac reserve in cardiogenic
shock: implications for prognosis and management. Br Heart J
1990;64:121 - 8.
Tan LB. Cardiac pumping capability and prognosis in heart failure.
Lancet 1986;2:1360 - 3.
Sandham JD, Hull RD, Brant RF, et al. A randomized, controlled trial
of the use of pulmonary-artery catheters in high-risk surgical
patients. N Engl J Med 2003;348:5 - 14.
Richard C, Warszawski J, Anguel N, et al. Early use of the
pulmonary artery catheter and outcomes in patients with shock and
acute respiratory distress syndrome: a randomized controlled trial.
JAMA 2003;290:2713 - 20.
Harvey S, Harrison DA, Singer M, et al. Assessment of the clinical
effectiveness of pulmonary artery catheters in management of
61 of 158.
12.
13.
14.
15.
patients in intensive care (PAC-Man): a randomised controlled trial.
Lancet 2005;366:472 - 7.
Rhodes A, Cusack RJ, Newman PJ, et al. Bennett ED. A randomised,
controlled trial of the pulmonary artery catheter in critically ill
patients. Intensive Care Med 2002;28:256 - 64.
Wheeler AP, Bernard GR, Thompson BT, et al. Pulmonary-artery
versus central venous catheter to guide treatment of acute lung
injury. N Engl J Med 2006;354:2213 - 24.
Cotter G, Moshkovitz Y, Kaluski E, et al. Accurate, noninvasive
continuous monitoring of cardiac output by whole-body electrical
bioimpedance. Chest 2004;125:1431 - 40.
Leitman M, Sucher E, Kaluski E, et al. Non-invasive measurement of
cardiac output by whole-body bio-impedance during dobutamine
stress echocardiography: clinical implications in patients with left
ventricular dysfunction and ischaemia. Eur J Heart Fail 2006;8:
136 - 40.
Circ J 2006; 70: 1164 – 1168
Impedance Cardiography for Cardiac Output Estimation
Reliability of Wrist-to-Ankle Electrode Configuratio
Oscar Luis Paredes, MD; Junya Shite, MD; Toshiro Shinke, MD;
Satoshi Watanabe, MD; Hiromasa Otake, MD; Daisuke Matsumoto, MD;
Yusuke Imuro, MD; Daisuke Ogasawara, MD;
Takahiro Sawada, MD; Mitsuhiro Yokoyama, MD
Background Non-invasive measurement of cardiac output (CO) may become an important modality for the
treatment of heart failure. Among the several methods proposed, impedance cardiography (ICG) has gained
particular attention. There are 2 basic technologies of ICG: thoracic and whole-body ICG whereby the electrodes
are applied either to the chest or to the limbs. The present study is aimed to test the effectiveness of the NonInvasive Cardiac System (NICaS), a new ICG device working with a wrist-to-ankle configuration
Methods and Results To evaluate the reliability of NICaS derived CO (NI-CO), 50 CO measurements were
taken simultaneously with thermodilution (TD-CO) and modified Fick (Fick-CO) in 35 cardiac patients, with the
TD-CO serving as the gold-standard for the evaluation. Overall, 2-tailed Pearson’s correlation and Bland-Altman
limits of agreement between NI-CO and TD-CO were r=0.91 and –1.06 and 0.68 L/min and between Fick-CO
and TD-CO, r=0.80 and –1.52 and 0.88 L/min, respectively. Good correlation was observed in patients with loading conditions altered by nitroglycerin and also in patients with moderate valvular diseases.
Conclusion Agreement between NI-CO and TD-CO is within the boundaries of the FDA guidelines of bioequivalence. NI-CO is applicable for non-invasive assessment of cardiac function. (Circ J 2006; 70: 1164 –
1168)
Key Words: Cardiac output; Impedance cardiography; Thermodilution method
S
everal studies suggest the importance of cardiac power
output calculation, which is derived from cardiac
output (CO) and mean blood pressure, to predict the
prognosis in heart failure patients not only in hospital but
also in the outpatient setting.1–3 CO measured by the thermodilution method with a Swan-Ganz catheter placed in
the pulmonary artery has become one of the most widely
accepted and used methods of monitor cardiac function,
despite its certain limitations.1,3,4 A noninvasive and low
cost method for measuring CO would be relevant for the
widespread clinical use of cardiac power output.
Some noninvasive techniques of measuring CO have been
proposed over the past years. The indirect Fick method of
re-breathing carbon dioxide5,6 and Doppler flow measurement of the left ventricular outflow tract have been shown
to be accurate;7 however, their applications require expensive equipments and trained operators. Other promising
results have been observed with devices based on electrical
bioimpedance technology,8 and 2 basic technologies of
impedance cardiography (ICG) are currently in use. The
first is called whole-body ICG9,10 (ICGWB), which was
introduced in 1948,11 in which the electrodes are placed on
the distal portion of the limbs. The second one is thoracic
(Received March 30, 2006; revised manuscript received June 1,
2006; accepted June 21, 2006)
Division of Cardiovascular and Respiratory Medicine, Department of
Internal Medicine, Kobe University Graduate School of Medicine,
Kobe, Japan
Mailing address: Junya Shite, MD, Division of Cardiovascular and
Respiratory Medicine, Department of Internal Medicine, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku,
Kobe 650-0017, Japan. E-mail: [email protected]
ICG (ICGT), which was introduced in 1964, and the electrodes are placed on the root of the neck and on the lower
chest. When the CO is measured in subjects with healthy
hearts, the results from both these technologies are usually
reliable, but the reliability of CO measurements taken by
ICGT is compromised in patients with cardiac diseases.12–16
According to the Food and Drug Administration (FDA)
standard of bio-equivalence,17 the disparity between 2 tech-
Fig 1. Schema of Non-Invasive Cardiac System, showing the impedance box, which is connected to the patient via proprietary electrodes
and the computer that interprets the collected data. Three electrodes
are applied to patient’s chest for the ECG monitor. One pair of impedance electrodes is applied to the left wrist and another pair to the right
ankle (tetra polar mode).
Circulation Journal Vol.70, September 2006
62 of 158.
1165
Wrist-to-Ankle Impedance for CO Measurement
Table 1 Statistical Analysis of CO Measurements Performed by NI-CO and Indirect Fick-CO Compared With TD-CO
All measurements in 35 patients
(n=50)
NI-CO Fick-CO
vs
vs
TD-CO TD-CO
Correlation
Bias (L/min)
SD (L/min)
Lower level of agreement (L/min)
Upper level of agreement (L/min)
Average CO ± SD (L/min)
0.911a
–0.18
0.43
–1.06
0.68
TDCO
NICO
Measurements after NTG injection
(n=15)
FickCO
0.801b
–0.32
0.60
–1.52
0.88
NI-CO Fick-CO
vs
vs
TD-CO TD-CO
0.962a
–0.15
0.50
–1.16
0.85
4.18±1.01 4.36±1.03 4.05±0.89
TDCO
NICO
FickCO
0.822b
–0.34
0.69
–1.73
1.04
3.59±0.76 3.85±0.85 3.67±0.79
Values are mean ± SD. Correlation was calculated using Pearson’s 2-tailed test.
1ap<0.0001, 1bp<0.0001, 2ap<0.0001, 2bp=0.0001.
CO, cardiac output; NI-CO, Non-Invasive Cardiac System derived CO; Fick-CO, modified Fick derived CO; TD-CO, thermodilution-derived CO; NTG, nitroglycerine.
nologies should not exceed the range of 20%. The purpose
of this study was to evaluate the reliability and feasibility of
the new Non-Invasive Cardiac System (NICaS: NI Medical,
Hod-Hasharon, Israel), which calculates the CO by measuring ICG in a tetra polar mode, derived from electrodes
placed on one wrist and the contra-lateral ankle (Fig 1).
Methods
The Trial
This study prospectively enrolled 35 patients. The thermodilution-derived CO (TD-CO) was measured using a
Swan-Ganz catheter (Baxter Healthcare, Irvine, CA, USA),
followed immediately by the modified Fick (Fick-CO) and
with the NICaS (NI-CO). In 15 subjects, a second round of
CO measurements was carried out using the 3 technologies
after 2–4 mg nitroglycerin injection into the arteries. Thus,
a total of 50 comparative CO measurements were performed. The Institutional Ethics Committee of the Kobe
University Hospital approved the study protocol and all
patients gave their consent.
Patient Selection and Exclusion Criteria
All cardiac disease patients who were scheduled for
routine right heart catheterization or emergency procedures
that required the deployment of a Swan Ganz catheter for
continuous cardiac function monitoring were eligible for
the study, unless they fulfilled one of the exclusion criteria
The exclusion criteria for the NI-CO measurements included: restlessness and/or chaotic patient condition, severe
aortic valve regurgitation and/or aortic stenosis, aortic
aneurysm, heart rate above 130 beats/min, intra- and extracardiac shunts, severe peripheral vascular disease, severe
pitting edema, sepsis and dialysis, all of which interfere
with the proper measurements of impedance derived stroke
volume.4
The study population comprised 35 patients (17 men, 18
women, mean age 65.5±13.7 years). In most patients there
was coexistence of multiple underlying heart diseases,
including hypertension (n= 15), diabetes mellitus (n= 13),
coronary artery disease (n= 21) and idiopathic dilated cardiomyopathy (n=3). Twelve patients presented with congestive heart failure (CHF) and 4 patients had atrial fibrillatio
when CO was measured. In addition, our study subjects
included 7 cases of moderate degree of valve regurgitation
(aortic regurgitation 2 cases, mitral regurgitation 4 cases,
mild tricuspid regurgitation 1 case) and 7 cases of moderate
63 of 158.
aortic valve stenosis.
Measuring CO
All operators were unaware of the CO results obtained
by the various measuring techniques.
TD-CO Right heart catheterization using a 6 or 8Fr
Swan-Ganz catheter was performed according to the standard institutional protocol. The catheter was advanced to
the pulmonary artery under fluoroscopic guidance and verified with the pressure waveforms registered on the polygraph. TD-CO was measured 5 times by injecting 5 ml
bolus of iced 9% saline solution at the same rate. Thereafter, the 3 results of the saline injections that were within
15% of their extreme disparity were averaged for the TDCO result.
Fick-CO For arterial oxygen saturation, blood samples
were obtained from the arterial access sheath, and for
venous oxygen saturation, blood samples were withdrawn
using the distal edge lumen of the Swan-Ganz catheter
placed in the pulmonary artery. All samples were immediately measured for oxygen saturation using the same device
(Radiometer ABL 715, Copenhagen, Denmark).
NI-CO To measure the CO with the NICaS apparatus,
an alternating electrical current of 1.4 mA with a 30 kHz
frequency is passed through the patient via 2 pairs of
tetrapolar electrodes, one pair placed on the wrist above the
radial pulse, and the other pair on the contralateral ankle
above the posterior tibialis arterial pulse. If the arterial
pulses in the legs are either absent or of poor quality, the
second pair of electrodes is placed on the contralateral
wrist. The NICaS apparatus calculates the stroke volume
by Frinerman’s formula:18
Stroke volume = dR/R ×ρ× L2 /Ri × (α+β)/β× KW× HF
where dR is the impedance change, R is the basal resistance,ρis the blood electrical resistivity, L is the patient’s
height, Ri is the corrected basal resistance according to
gender and age,18–23 KW is a correction of weight according
to ideal values,19–24 HF is the hydration factor, which takes
into account the body water composition,21 and α+β is
equal to the ECG R–R wave interval, andβis the diastolic
time interval.
Because the NI-CO results are calculated every 20 s, the
average of 3 measurements obtained consecutively during
60 s of monitoring was considered to be the NI-CO value
for each individual case.
1166
PAREDES O L et al.
Fig 2. Scatter plots of cardiac output (CO) showing the correlation between Non-Invasive Cardiac System (NICaS)
derived CO (NI-CO) with thermodilution-derived CO (TD-CO) [A]; modified Fick derived CO (Fick-CO) with TD-CO
[B] and NI-CO with Fick-CO [C]. The 2-tailed Pearson’s correlation test (r) was used for the analyses.
Fig 3. [A] Bland-Altman scatter plot of difference against average of cardiac output (CO) results measured by NonInvasive Cardiac System (NICaS) and thermodilution method. [B] Bland-Altman scatter plot of difference against
average of CO results measured by the modified Fick and thermodilution methods. NI-CO, NICaS derived CO; TD-CO,
thermodilution-derived CO; Fick-CO, modified Fick derived CO
The quantitative data are expressed as mean ± SD. For descriptive statistic, Student’s t-test was used. To compare the
results of NI-CO, Fick-CO and TD-CO, 2-tailed Pearson’s
correlation and the Bland-Altman25 limits of agreement
were used. The gold-standard for determining accuracy of
the results was the TD-CO. Values of p<0.05 were consid-
ered to be significant
Results
The average values of CO in the study subjects for TDCO, NI-CO and Fick-CO were 4.18±1.01 L/min, 4.36±
1.03 L/min, and 4.05±0.89 L/min, respectively. There were
64 of 158.
1167
Wrist-to-Ankle Impedance for CO Measurement
no significant differences between the 3 groups (Table 1).
The overall results of the Pearson correlation analysis were
as follows: NI-CO vs TD-CO: r=0.91, p<0.0001; Fick-CO
vs TD-CO: r= 0.80, p<0.0001 and NI-CO vs Fick-CO: r=
0.85, p<0.0001 (Fig 2). The Bland-Altman 2-standard deviation limit of agreement between the NI-CO and TD-CO
was ±0.87 (–1.06 and 0.68) L/min, and the agreement
between the Fick-CO and TD-CO was ±1.20 (–1.52 and
0.88) L/min (Fig 3). The calculated percentage of disparity
between the NI-CO and TD-CO would thus be 19.95%
(0.87 L/min ÷ 4.36 L/min), which was less than that between
Fick-CO and TD-CO (29.63% [1.20 L/min ÷ 4.05 L/min]).
Nevertheless, there are 3 cases in this series in which the
disparity between the NI-CO and TD-CO was greater than
20%, indicating that the disparity here is not equal, but
close to FDA bio-equivalence (Note that the mathematical
model of the FDA for determining bio-equivalence was not
used here. Yet the simple model of limits of agreement,
which was used, offers an acceptable appraisal of the good
interrelationships between the 2 statistical approaches).
When we analyzed the subgroup of measurements before and after nitroglycerine injection to alter vascular resistance, identical changes in CO were observed with the
NI-CO and TD-CO (TD-CO: 4.42±1.00 L/min → 3.59±
0.76 L/min, NI-CO: 4.07±1.07 L/min → 3.85±0.85 L/min
(Fig 4)). The relation between NI-CO and TD-CO after
nitroglycerine injection was r= 0.96, p<0.0001, n= 15. All
other statistical details are summarized in Table 1.
In the other subgroups of moderate degree of valvular
regurgitation (n= 7) and aortic valve stenosis (n= 7), the
correlation of NI-CO with TD-CO was r=0.92, p<0.0001,
with a lower limit of agreement of –0.97 L/min and upper
limit of agreement of 0.73 L/min.
Discussion
According to Bland and Altman26 and Raaijmakers et
al,15 when averages of repeated measurement results are
used to compare the performance of a new medical device
with a gold-standard, there is an overestimation of the correlation coefficient. In the present investigation the preferable single measurement design was used; namely, each test
consisted of a TD-CO measurement, followed immediately
by a NI-CO and a Fick-CO measurement. In 20 patients,
only 1 study was performed, whereas in the remaining 15
patients 2 studies were done: before and after nitroglycerine injection. However, each of the second tests was conducted in the manner of an independent comparative
measurement.
According to the definition of the FDA, if there is bioequivalence between the gold-standard and a new technol-
Fig 4. Changes in cardiac output in 15 patients with repeated measurements, pre and post nitroglycerine (NTG) administration. The
graph compares the results obtained by Non-Invasive Cardiac System
derived cardiac output (NI-CO) and thermodilution-derived cardiac
output (TD-CO). Values are mean ± SD.
ogy, all the comparative results should be within a range of
20% disparity. Previous studies reported limits of agreement between ICGT-CO vs TD-CO of –2.2 to 2.2 L/min at
best,27–29 ICGWB-CO (JR Medical-Tallinn, Estonia) vs TDCO of –1.37 to 1.87 L/min,10 and bipolar NI-CO vs TD-CO
of –1.25 to 1.30 L/min (Table 2).4 Based on these reports,
it can be calculated that the disparity between ICGT-CO,
ICGWB-CO, bipolar NI-CO vs TD-CO is 40%, 32.4% and
26%, respectively. Our result for tetra polar NI-CO vs TDCO (–1.06 to 0.68 L/min, disparity 20%) is better than
those results. The reason for the better result with NICaS
may be related to the most upgraded calculation formula.
An important fringe benefit of this trial is the data produced by the Fick-CO technology. This method, which
enjoys increasing popularity among practical cardiologists,
is still considered controversial.30 According to the present
results, the limits of agreement between Fick-CO and TDCO are –1.52 and 0.88 L/min (average ±1.20 L/min). In the
presence of a mean Fick-CO of 4.05 L/min, the disparity
between the 2 technologies is 30%, better than that of ICGT
but inferior to that of NI-CO.
Among the established exclusion criteria related to the
use of the NICaS are: severe aortic stenosis, in which the
NI-CO is usually underestimated, and significant aortic regurgitation, in which the NI-CO tends to be overestimated.
In the present trial, however, 7 cases of moderate aortic
stenosis and 2 with mild – moderate aortic regurgitation
were involved, indicating that NICaS is applicable in cases
of mild to moderate aortic valve disease.
From the 15 measurements that were obtained after nitroglycerine injection, we observed a decrease in the mean CO
Table 2 Summary of Previous Reports of the Accuracy of CO Measurements by ICG (ICGT-CO, ICGWB-CO, and
NI-CO vs TD-CO)
Authors
Method
Condition
Year published
Limits of agreement between
ICG and TD (L/min)
Drazner M, et al27
Van De Water JM, et al28
Leslie S J, et al29
Koobi T, et al10
Cotter G, et al4
ICGT
ICGT
ICGT
ICGWB
ICGNICaS
Bipolar
CHF
CABG
CHF
CABG
CABG, CHF
2002
2003
2004
1997
2004
–2.2 to 2.2
–2.36 to 1.99
–2.2 to 2.2
–1.37 to 1.87
1.25 to 1.30
Values are mean. ICG, impedance cardiography; ICGT, thoracic ICG; ICGWB, whole-body ICG; TD, thermodilution; CHF, congestive
heart failure; CABG, coronary artery bypass graft; ICGNICaS, Non-Invasive Cardiac System ICG. Other abbreviations as in Table 1.
65 of 158.
1168
PAREDES O L et al.
levels. However, the accuracy of the results remained unaltered in NI-CO, which suggests that NICaS is also applicable even when the arterial resistance has changed.
Clinical Implications
Recent studies have shown that the calculation of cardiac
power output (CO × mean arterial pressure) and systemic
vascular resistance are important for the management of
various cardiac diseases.1–3 Cardiac power output was found
to be the strongest independent predictor of in-hospital
mortality in patients admitted with cardiogenic shock2 and
is an important tool for assessing the clinical response to
drug therapy. In addition, there is enough evidence that
ambulatory monitoring of cardiac power would benefit patients with CHF, resulting in better titration of medication
and possibly less readmission to hospital. By introducing
NICaS apparatus, wide-spread clinical use of cardiac
power calculation would become feasible.
Conclusion and Limitations of NICaS
The present study indicates that NICaS performs at least
as accurately as the thermodilution method. However, the
reliability of the NICaS method depends on an alignment
with exclusion criteria. This allows for the use of NICaS in
approximately 80–85% of patients needing the examination.
References
1. Nohria A, Mielniczuk L, Warner L. Evaluation and monitoring of
patients with acute heart failure syndromes. Am J Cardiol 2005;
96(Suppl): 32G – 40G.
2. Fincke R, Hochman J, Lowe A, Menon V, Slater JN, Webb JG, et al
(for the SHOCK investigators). Cardiac power is the strongest hemodynamic correlate of mortality in cardiogenic shock: A report from
the SHOCK trial registry. J Am Coll Cardiol 2004; 40: 340 – 348.
3. Cotter G, Moshkovitz Y, Kaluski E, Milo O, Nobikov Y, Schneeweiss
A, et al. The role of cardiac power and systemic vascular resistance
in the pathophysiology and diagnosis of patients with acute congestive heart failure. Eur J Heart Fail 2003; 5: 443 – 451.
4. Cotter G, Moshkovitz Y, Kaluski E, Kohen A, Miller H, Goor D, et
al. Accurate, noninvasive continuous monitoring of cardiac output by
whole-body electrical bioimpedance. Chest 2004; 125: 1431 – 1440.
5. Haryadi DG, Orr JA, Kuck K, Mc James S, Westenskow DR. Partial
CO2 rebreathing indirect Fick technique for non-invasive measurement of cardiac output. J Clin Monit Comput 2000; 16: 361 – 374.
6. Jover JL, SoroM, Belda FJ, Aguilar G, Caro P, Ferrandis R. Measurement of cardiac output after cardiac surgery: Validation of a partial
carbon dioxide rebreathing (NICO) system in comparison with continuous thermodilution with a pulmonary artery catheter. Rev Esp
Anestesiol Reanim 2005; 52: 256 – 262 (in Spanish).
7. Turner MA. Doppler-based hemodynamic monitoring: A minimally
invasive alternative. AACN Clin Issues 2003; 14: 220 – 231.
8. Moshkovitz Y, Kaluski E, Milo O, Vered Z, Cotter G. Recent developments in cardiac output determination by bioimpedance: Comparison with invasive cardiac output and potential cardiovascular applications. Curr Opin Cardiol 2004; 19: 229 – 237.
9. Tischenko MI. Estimation of stroke volume by integral rheogram of
the human body. Sechenov Phisiological J 1973; 59: 1216 – 1224 (in
Russian).
10. Koobi T, Kaukinen S, Turjanmaa VM. Cardiac output can be reliably
measured noninvasively after coronary artery bypass grafting operation. Crit Care Med 1999; 27: 2206 – 2211.
11. Kedrov AA. An attempt to quantify assessment of the central and
peripheral circulation by electrometrical method. Klin Med 1948; 26:
32 – 51 (in Russian).
12. Patterson RP, Kubicek WG, Kinnen E, Witsoe DA, Noren G. Development of an electrical impedance plethysmography system to monitor cardiac output. Proceedings of the 1st Annual Rocky Mountains
Bioengineering Symposium 1964; 56 – 71.
13. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA, Mattson RH.
Development and evaluation of an impedance cardiac output system.
Aerosp Med 1966; 37: 1208 – 1212.
14. Kubicek WG, Kottke FJ, Ramos MU, Patterson RP, Witsol DA,
Labree JW, et al. The Minnesota impedance cardiograph: Theory and
applications. Biomed Eng 1974; 9: 410 – 416.
15. Raaijmakers E, Faes ThJC, Scholten RJPM, Goovaerts HG, Heethaar
R. A meta-analysis of published studies concerning the validity of
thoracic impedance cardiography. Ann NY Acad Sci 1999; 873: 121 –
134.
16. Handelsman H. Public health service reassessment: Measuring cardiac output by electrical bioimpedance. Health Technology Assessment Reports, US Dept Health and Human Services, Public Agency
for Health Care Policy and Research 1991; 6: 1 – 13.
17. Guidance for Industry. Statistical approach to establishing bioequivalence: US department of health and human services. Food and Drug
Administration, Center For Drug Evaluation and Research (CDER)
January 2001, BP.
18. Cohen AJ, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J,
coronary artery bypass grafting. Eur J Surg 1998; 14: 64 – 69.
19. Organ LW, Bradham GB, Gore DT, Lozier SL. Segmental bioelectrical impedance analysis: Theory and application of a new technique. J
Appl Physiol 1994; 77: 98 – 112.
20. Lukaski H, Bolonchuk WW, Hall CB, Siders WA. Validation of tetrapolar bioelectrical impedance method to assess human body
composition. J Appl Physiol 1986; 60: 1327 – 1332.
21. Hoffer EC, Meador CK, Simpson DC. A relationship between whole
body impedance and total body water volume. Ann NY Acad Sci
1970; 170: 452 – 461.
22. Lamberts R, Visser KR, Zijlstra WG. Impedance cardiography.
Assen, The Netherlands; Van Gorkum 1984; 21 – 94.
23. Ward LC, Heitmann BL, Craid P, Stroud D, Azinge EC, Jebb S, et al.
Association between ethnicity, body mass index, and bioelectrical
impedance: Implications for the population specificity of prediction
equations. Ann NY Acad Sci 2000; 904: 199 – 204.
24. Hamwi GT, Danowski TS. Changing dietary concepts in diabetes
mellitus: Diagnosis and treatment. New York: American Diabetes
Association 1964; 73 – 78.
25. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; I: 307 –
310.
26. Bland JM, Altman DG. Correlation, regression and repeated data.
BMJ 1994; 308: 896.
27. Drazner M, Thompson B, Rosenberg P, Kaiser PA, James MSN,
Boehrer D, et al. Comparison of impedance cardiography with invasive hemodynamic measurements in patients with heart failure secondary to ischemic or nonischemic cardiomyopathy. Am J Cardiol
2002; 89: 993 – 995.
28. Van De Water JM, Miller TW, Vogel RL, Mount BE, Dalton ML.
Impedance cardiography: The next vital sign technology? Chest
2003; 123: 2028 – 2033.
29. Leslie S J, McKee S, Newby DE, Webb DJ, Denvir MA. Non-invasive measurement of cardiac output in patients with chronic heart
failure. Blood Press Monit 2004; 9: 277 – 280.
30. Dhingra VK, Fenwick JC, Walley KR, Chittock D, Ronco JJ. Lack
of agreement between thermodilution and Fick cardiac output in critically ill patients. Chest 2002; 122: 990 – 997.
66 of 158.
INSTITUTE OF PHYSICS PUBLISHING
Physiol. Meas. 27 (2006) 817–827
PHYSIOLOGICAL MEASUREMENT
doi:10.1088/0967-3334/27/9/005
Impedance cardiography revisited
G Cotter1, A Schachner2, L Sasson2, H Dekel2 and Y Moshkovitz3
1 Divisions of Clinical Pharmacology and Cardiology, Duke University Medical Center, Durham,
NC, USA
2 Angela & Sami Sharnoon Cardiothoracic Surgery Department, Wolfson Medical Center, Israel
3 Department of Cardiac Surgery, Assuta Hospital, Petah Tikva, Israel
Received 7 February 2006, accepted for publication 9 June 2006
Published 5 July 2006
Online at stacks.iop.org/PM/27/817
Abstract
Previously reported comparisons between cardiac output (CO) results in
patients with cardiac conditions measured by thoracic impedance cardiography
(TIC) versus thermodilution (TD) reveal upper and lower limits of agreement
with two standard deviations (2SD) of approximately ±2.2 l min−1, a 44%
disparity between the two technologies. We show here that if the electrodes
are placed on one wrist and on a contralateral ankle instead of on the chest, a
configuratio designated as regional impedance cardiography (RIC), the 2SD
limit of agreement between RIC and TD is ±1.0 l min−1, approximately
20% disparity between the two methods. To compare the performances of
the TIC and RIC algorithms, the raw data of peripheral impedance changes
yielded by RIC in 43 cardiac patients were used here for software processing
and calculating the CO with the TIC algorithm. The 2SD between the TIC
and TD was ±1.7 l min−1, and after annexing the correcting factors of the
RIC formula to the TIC formula, the disparity between TIC and TD further
declined to ±1.25 l min−1. Conclusions: (1) in cardiac conditions, the RIC
technology is twice as accurate as TIC; (2) the advantage of RIC is the use
of peripheral rather than thoracic impedance signals, supported by correcting
factors.
Keywords: cardiac output measurements, thoracic bioimpedance, whole-body
bioimpedance, impedance cardiography
Introduction
Three basic technologies are currently in use for impedance cardiography (ICG): (1) the
thoracic ICG (TIC), where the electrodes are placed on the root of the neck and the lower
part of the chest, being the dominant method in the market (Patterson et al 1964, Kubicek
0967-3334/06/090817+11$30.00 © 2006 IOP Publishing Ltd Printed in the UK
67 of 158.
817
818
G Cotter et al
et al 1966, 1974); (2) the whole-body ICG (ICGWB), where four pairs of electrodes are used,
one pair on each limb (Tischenko 1973, Koobi et al 1999); (3) the regional ICG (RIC), a
technology which is used by the NICaS (noninvasive cardiac system). In this technology,
which is the subject of this report, only two pairs of electrodes are used, performing best
when placed on one wrist and on the contralateral ankle (Cohen et al 1998, Cotter et al 2004,
Torre-Amione et al 2004).
Two comprehensive reviews of the literature on clinical experience in measuring the
cardiac output (CO) by TIC determined that in patients with cardiac conditions the TIC-CO
results are unreliable (Raaijmakers et al 1999, Handelsman 1991). According to Patterson
(1985) and Wang et al (2001), a number of sources in the chest, such as the lungs, vena cava, and
systemic and pulmonary arterial vasculatures, generate systolic impedance reductions, while
the heart generates signals of increased impedance. In addition to these multifarious sources of
Z,4 variations in the electrical conductivities between the sources of impedance changes and
the TIC electrodes (Kim et al 1988, Kauppinen et al 1998), and between the various impedance
sources (Wtorek 2000) have been described. These model experimentations indicated that
the thoracic Z is not a reliable signal for calculation of the SV due to the multiple sources
of dZ/dt (Kim et al 1988, Wang and Patterson 1995, Kauppinen et al 1998, Wtorek 2000),
providing the explanations for the above-mentioned unsatisfactory clinical results obtained by
TIC (Raaijmakers et al 1999, Handelsman 1991).
In this report, an attempt is made to defin the differences between the peripheral and
thoracic impedance signals, and based on this, to explain the differences in the performance
of RIC and TIC.
As we are capable of saving raw data from the wrist–ankle (peripheral) impedance
signals, we were able to use the peripheral impedance waveforms and base impedance values
to calculate stroke volumes, using various algorithms that have been associated with TIC
calculations. This enabled us to prove that (1) the performance of RIC is twice as accurate
as reported TIC results; (2) the reasons for this are as follows: (a) the impedance changes
which are yielded by the limb electrodes are more suitable than the impedance changes of
the thoracic electrodes for calculating the stroke volume and (b) the use of properly designed
coefficient improved the accuracy of the CO results by at least an additional 25%.
Methods
The data for this project were gathered from two patient series. In both, comparisons were made
between cardiac output results measured by the NICaS versus thermodilution. One series,
which was studied in hospital A, consisted of 30 patients who were studied immediately upon
arrival at the ICU following an open heart operation. In 11 (36%), despite the intravenous
administration of adrenalin, cardiac index (CI) was lower than 2.5 l min−1 m−2. The second
series included 13 cases of acute heart failure that were studied in hospital B. CI was lower
than 2.5 l min−1 m−2 in seven (54%), and in the combined group of 43 cases of the two
hospitals, it was lower than 2.5 l min−1 m−2 in 18 (43%).
The purpose of this study was to use peripheral impedance waveforms to calculate stroke
volume by means of four different ICG algorithms and to compare each of these SV values
with the thermodilution SV result.
Of the 55 and 31 studies conducted in hospitals A and B, respectively, raw data were
successfully retrieved from only the last 30 consecutive patients of hospital A and the last 13
4
In the ICGWB and RIC, where the impedance changes are depicted in the periphery, the impedance value is
automatically converted into the real parts (R0 and R) of the measured impedance signals (Lamberts et al 1984).
68 of 158.
819
α+β
ECG
∆R
α
Figure 1. Unedited print of the ECG and R waveforms to demonstrate the definition of α
and β.
consecutive cases of hospital B. It quickly became apparent that these 43 source data were
retrieved from only one of the two NICaS devices that were being used in the two hospitals. The
reason for this was that we did not realize that the software designed for source data retrieval
was not implemented in one of the two NICaS devices that were used interchangeably during
these trials.
Measuring the CO
Thermodilution. The standard techniques for measuring the thermodilution CO (TDCO)
used in these two hospitals have been reported elsewhere (Cotter et al 2004).
RIC (NICaS). To measure the CO with the NICaS (NICO), which is a tetrapolar device, two
electrodes are placed on a wrist above the radial pulse and two on the contralateral ankle above
the posterior tibialis arterial pulse. If the arterial pulses in the legs are either absent or of poor
quality, the second pair of electrodes is placed on the contralateral wrist. The NICaS device
calculates the SV by means of the following Frinerman formula5 (Cohen et al 1998):
SV =
RρL2 (α + β)Kw × HF
,
RRi β
(2)
where SV is the cardiac stroke volume (ml), R is the resistance change during the cardiac
cycle (), R is the basal resistance (), Ri is a corrected basal R (), ρ is the blood electrical
resistivity ( cm), L is the patient’s height (cm), Kw is a correcting factor for the body weight,
HF is the hydration factor related to the body water composition and (α + β)/β is the ratio
of the systolic time plus the diastolic time divided by the diastolic time of the R waveform
(figur 1).
5
The present formula is principally the same as the formula in Cohen et al (1998), but is written differently. The
original formula was
SV =
2 R
IB × Hcorr
(α + β)
Hctcorr
× Kel × Kweight ×
×
.
Ksex
R
β
(1)
The values of the hematocrit (Hct) and Na( (el) are now represented by ρ, which is the electrical resistivity of the blood.
Ksex age, which affects the basal resistance (R), is now represented by R i, and IB (index body) is now represented by
HF (hydration factor). The correction of H 2 is no longer included in the formula, but in patients whose arms are
disproportionately long, the electrodes should be placed 5 cm proximally to their regular position.
69 of 158.
820
G Cotter et al
Definition and principles of the correcting coefficients of the Frinerman formula
The basic impedance part of Frinerman’s formula consists of the following Bonjer equation
(1950):
dRρL2
,
(3)
R2
to which Frinerman’s correcting factors were added. These coefficient were designed to
neutralize the individual effects of gender, age, body water composition and anthropomorphic
variabilities of each patient.
According to RIC studies, the basal R0 is higher in females than in males, a fact which
is in accordance with the literature (Organ et al 1994, Lukaski et al 1986, Hoffer et al 1970,
Lamberts et al 1984, Ward et al 2000), and it tends to rise with age (Organ et al 1994).
Based on reported values of basal resistances related to sex and age (Organ et al 1994,
Lukaski et al 1986, Ward et al 2000), a coefficien is calculated to determine an ideal value
of R0 for the studied patient. Ideal here means an ideal weight compared to height in healthy
condition. By dividing the measured R0 of the patient by the calculated ideal R0, a correcting
coefficien is obtained. By multiplying the measured R by the correcting coefficient the Ri
variable is determined and placed in the denominator of the formula.
Similarly, a correcting coefficien (Kw) for the body weight is calculated by dividing the
measured weight by the calculated weight according to Hamwi’s formula (1964) of ideal
weight. Kw is required for compensation of the reduced electrical conductance in the fat, and
this coefficien is placed in the numerator of the formula, affecting the value of R. In a similar
fashion, the hydration factor, HR, is calculated by the patient’s body water composition, and
this coefficien is placed in the numerator of the formula, either to reduce or to increase R in
over- or under-hydrated patients, respectively.
The variable (α + β)/β is based on the Windkessel principle (Frank 1926, Faes et al
1999), where a distinction is made between the aortic infl w α, which is incurred by the
systolic left ventricular ejection, and the aortic outfl w (α + β), which extends throughout the
cardiac cycle (figur 1).
SV =
Comparative calculation of CO by software simulation
The impedance raw data of the compiled 43 investigative patients were retrieved and their
CO were calculated by the following formulae: (1) Frinerman’s formula, as measured by
the NICaS; (2) Bonjer’s equation of 1950 (Bonjer 1950), which was expressed differently by
Patterson’s f rst version of the TIC formula (Patterson et al 1964):
2
L
Z;
(4)
V = ρ
Z
(3) Patterson’s formula of the f rst derivative (Patterson 1965):
2
L
dZ
,
×T ×ρ×
SV =
dt
Z
(5)
where SV is the cardiac stroke volume (cm), dZ/dt is the peak of the f rst derivative of the
impedance change during systole ( s−1), T is the cardiac ejection time (s), L is the length
between voltage pick-up electrodes (cm) and Z is the base impedance value (), which is still
used by the TIC technology (Kubicek et al 1974); (4) a combined formula which includes
Patterson’s equation together with Frinerman’s correcting factors (table 1, figure 2 and 3):
70 of 158.
821
11.0
ICG-CO (lit/min)
9.0
7.0
5.0
3.0
1.0
1.0
3.0
5.0
7.0
9.0
11.0
TDCO (lit/min)
Frinerman NICaS
Bonjer
Kubicek-Patterson(first derivative)
Kubicek & Frinerman Correcting Factors
Figure 2. Scatter plot comparing CO results calculated by four impedance algorithms versus
thermodilution.
Table 1. Correlations and limits of agreement between four algorithms versus TD. Comparison of
four different algorithms in 43 patients for measuring cardiac output (CO) of the same raw data:
(a) Frinerman (NICaS); (b) Bonjer; (c) Patterson–Kubicek; (d) Patterson–Kubicek–Frinerman.
Each of these was compared with thermodilution.
TD
Average CO (l min−1)
SD
Correlation with TD
p value versus NICaSa
Bias
SD of Bias
LL agreement
UL agreement
5.12
1.952
Frinerman
NICaS
5.05
2.043
0.969
–
−0.0698
0.509
−1.088
0.949
Bonjer
Patterson–
Kubicek
5.33
5.00
2.218
1.781
0.841
0.897
<0.001 <0.0005
0.2107 −0.1116
1.203
0.864
−2.196 −1.840
2.617
1.617
Patterson–Kubicek–
Frinerman
5.09
1.986
0.950
0.33
−0.0302
0.624
−1.277
1.217
SD = standard deviation; LL agreement = lower limit of agreement; UL agreement = upper limit
of agreement.
a Bonferroni adjusted p values for testing equality of correlation NICaS TD as compared with Bonjer
TD (<0.0001), Patterson–Kubicek TD (<0.0005), Patterson–Kubicek–Frinerman TD (0.33).
SV =
dR × T
ρ × L2 (α + β)
×
× Kw × HF,
dt × R
Ri β
(6)
where Frinerman’s R was replaced by dZ/dt × T and (α + β)/β was deleted.
Patient selection
Restlessness, aortic insufficien y, abdominal aneurysm, heart rate above 130 beats min−1,
arrhythmia with significan irregular heart rate, severe peripheral vascular disease, two or more
71 of 158.
822
G Cotter et al
(B) Bonjer
2.0
2.0
+2SD
0.0
-1.0
-2SD
-2.0
Difference(Bonjer-TD) CO
lit/min
3.0
1.0
lit/min
Difference(NICaS-TD) CO
(A) Frinerman
3.0
+2SD
1.0
0.0
-1.0
-2.0
- 2SD
-3.0
-3.0
0
2
4
6
8
10
0
12
2
8
10
12
3.0
2.0
+2SD
1.0
0.0
-1.0
-2SD
-2.0
0
2
4
6
8
10
Average(Patterson-Kubicek+TD)/2 CO lit/min
12
Difference(Patterson-KubicekFrinerman-TD) CO lit/min
Difference(Patterson-KubicekTD) CO lit/min
6
(D) Patterson-Kubicek-Frinerman
(C) Patterson-Kubicek
3.0
-3.0
4
Average(Bonjer+TD)/2 CO lit/min
Average(NICaS+TD)/2 CO lit/min
2.0
+2SD
1.0
0.0
-1.0
-2SD
-2.0
-3.0
0
2
4
6
8
10
12
Average(Patterson-Kubicek-Frinerman+TD)/2 CO lit/min
Figure 3. Scatter plots of Bland–Altman difference-against-average of the CO results calculated
by the four impedance algorithms versus thermodilution results (based on table 1).
pitting edema and dialysis all interfere with the proper measurements of the SV; therefore, this
patient population, which amounts to 15% of the candidates, is a priori excluded from the ICG
studies. Also excluded are RIC measurements of the CO in thoracoabdominal operations, such
as esophagectomies and pancreatectomies, where plasma loss and heavy intravenous volume
loads due to significan bleeding undermine the stability of the basal R and the reliability of
the CO results (Critchley et al 1996).
Statistical analysis
Thermodilution (TD) was used as the ‘gold standard’, and agreement between the four ICG
formulae and TD was analyzed according to the Bland–Altman approach (Bland and Altman
1986).
Results6
The results of the four algorithms revealed a rather similar average of the NICO in the range
of 5 l min−1 (table 1). The Pearson correlation coefficient were r = 0.969, 0.841, 0.897 and
0.950 with the formulae of Frinerman, Bonjer (1950), Patterson–Kubicek (Patterson 1965,
Kubicek et al 1974) and the combined Patterson–Kubicek–Frinerman formula, respectively
(table 1, figure 2 and 3). The Bland–Altman 2SD lower and upper limits of agreement with
TD were −1.088 and 0.949, −2.196 and 2.617, −1.840 and 1.617, and −1.277 and 1.217,
according to the same order of formulae as above (table 1, figur 3).
6
To calculate the SV with the Bonjer, Patterson–Kubicek and Patterson–Kubicek–Frinerman formulae, the raw data
of the 43 patients were e-mailed to Professor Robert Patterson, the inventor of the TIC technology (Patterson et al
1964, Patterson 1965), who made these calculations at the University of Minnesota.
72 of 158.
823
Discussion
To compare the performance of the currently used TIC technology with RIC, we used three
recent reports representing TIC results in patients with chronic stable congestive heart failure,
which revealed similar values (Drazner et al 2002, Van De Water et al 2003, Leslie et al 2004)
and could serve as a paradigm of the performance of TIC in the presence of cardiac conditions.
A relatively recent publication by Sageman et al (2002) could not be included here because
(a) only CI values were provided, (b) these values are characteristic of patients with normal
cardiac functions and (c) the patients’ CO results were averages of up to 20 measurements, an
approach which is not used clinically.
The average cardiac output in each of the three TIC series was in the range of 5 l min−1.
Comparison in each series of the results with thermodilution (TD) revealed a Bland–Altman
limit of agreement with a 2SD of approximately ±2.2 l min−1—a 44% deviation compared to
TD (Drazner et al 2002, Van De Water et al 2003, Leslie et al 2004). Despite the fact that the
present series did not include cases of chronic congestive heart failure, it is comparable with
the TIC studies because (a) the patients were actively managed for cardiac conditions, (b) a
CI < 2.5 l min−1 m−2 was observed in 43% of the cases and (c) the average CO was in the
range of 5 l min−1, as in the TIC paradigm.
When the Patterson–Kubicek algorithm was fed by impedance raw data yielded by the
peripheral wrist–ankle rather than by thoracic electrodes, the disparity compared to TD was
±1.7 l min−1, which is a 34% difference (figure 2 and 3). When the Patterson–Kubicek
algorithm was reinforced by the correcting factors of the NICaS formula and the peripheral
raw data were used, there was a further decline in the disparity between ICG and TD to the
range of ±1.25 l min−1, a 25% difference. This is almost as good as the 20% disparity between
the NICaS and the TD.
Judging by the 20% deviation between the NICaS and TD CO results, and by the 44%
deviation between the reported TIC and TD CO results, it is evident that the accuracy of the
NICaS is higher by a factor of more than 2, than the reported values of TIC (Raaijmakers
et al 1999, Handelsman 1991, Drazner et al 2002, Van De Water et al 2003, Leslie et al 2004).
Moreover, according to the FDA standard of bioequivalence (Guidance for Industry 2001),
the comparative results of new and gold-standard technologies should be contained within a
range of 20% disparity. This 20% range is determined by the 10% repeatability rate of each
of the two methods. Judging by figur 3(A), the agreement between the NICaS and the TD is
in very close proximity to the FDA standard bioequivalence.
About the validity of correcting coefficients
To appreciate the competence of the correcting factors, it is suffic to compare the CO results
of Frinerman’s and Bonjer’s formulae (figure 3(A) and (B)). As stated earlier, the only
difference between these two equations is the addition of Frinerman’s correcting factors to
Bonjer’s formula. Ri, for example, which is the corrected R0, may reach twice the value of
the measured R0. Kw may increase the measured R by up to 45%. The HF may either
reduce or increase the value of R by only slight to moderate degrees. Thus, it can be
discerned that when using Bonjer’s equation alone (figur 3(B)) the 2SD limits of agreement
are approximately ±2.4 l min−1, and when Frinerman’s variables are added (figur 3(A)) the
2SD limits of agreement are ±1.0 l min−1. There is a 20% disparity between the NICaS and
the TD, and a 45% improvement of Bonjer’s reinforced performance.
Similarly, annexing the Frinerman correcting coefficient to the TIC dZ/dt × T
formula dramatically improved the calculated CO results, from 2SD limits of agreement of
73 of 158.
824
G Cotter et al
approximately ±1.7 l min−1 (table 1, figur 3(C)) to 2SD limits of agreement of approximately
±1.25 l min−1 (table 1, figur 3(D)).
It has been repeatedly shown by Hoffer et al (1970), Lukaski et al (1986), Organ et al
(1994), Ward et al (2000) and others that the body is not a homogeneous conductor, and its
basal resistance, and consequently the spatial distribution of conductivities, is a function of
the body composition. It is also recognized that the two main factors which determine the
impedance variabilities are the percentages of body water and fat. As shown here, by using the
RIC approach, the variabilities of age, sex, height and weight can be quantitated and translated
to the effective correcting coefficients
About the validity of the wrist–ankle electrodes
The most significan advantage of RIC in comparison to TIC is the use of the peripheral
impedance signal for calculation of the SV. According to Kauppinen et al (2000), 75% of the
peripheral (RIC) impedance waveform is borne by the systolic blood volume pulsations of
the arterial vasculature of the upper and lower limbs, and the remaining 25% arrives from the
trunk (thorax). Still, the sole source of the peripheral signal is generated by the blood volume
pulse of the arterial vasculature.
It must be borne in mind that the aorta and its peripheral ramification comprise a single
anatomophysiological structure, and the pulse waveform, which evolves throughout its length,
occurs almost at the same time. This is possible only because the velocity of the pulse wave is
so rapid that the arterial expansion is completed before the termination of the left ventricular
contraction (Guyton and Hall 2000). The mean value of the pulse wave velocity (PWV) from
the thorax to the calves is approximately 7–10 m s−1 (Guyton and Hall 2000).7 Similarly, the
impedance volume pulse travels at approximately the same rate as reported by Risacher et al
(1995).
Thus, in accordance with the existing knowledge of the cardiac role in the formation of
the pulse (McVeigh et al 2002), the RIC peripheral volumetric signal is borne throughout the
length of the arterial tree beginning with the left ventricular stroke volume ejection.
In contrast, the thoracic (TIC) waveform is generated by multiple sources, including
the aorta, lungs, atria, vena cava and artifacts due to heart movements (Wang et al 2001,
Kauppinen et al 1998, Wtorek 2000). In normal people, some studies have shown that TIC
gives reasonably reliable CO results (Raaijmakers et al 1999), but in the presence of cardiac
conditions, distortions of the TIC waveforms were already observed in the earliest days of the
emergence of this technology (Kubicek et al 1974). This is contrary to RIC, where we see
very few changes, if any, in the waveform shape incurred by the different cardiac conditions.
The immunity of the peripheral impedance waveform to the distorting influenc of cardiac
conditions is reflecte by the results of the present study. The average disparity between TICCO and TD-CO measured in cardiac patients is 44%. In the present trial, the average disparity
between TD-CO and ICG-CO, which was calculated by the combination of the TIC algorithm
and a peripheral impedance waveform, was 34%. This 23% higher accuracy, which is obtained
by the standard TIC algorithm, can be attributed only to one factor, the reliance for calculation
of the SV on the peripheral, rather than the thoracic, impedance signal.
The fact that the RIC region consists of only part of the whole body but can be used
to calculate the CO of the whole body is explained by the electrophysiological relationships
which exist between RIC and ICGWB. In the ICGWB technique, electrodes are placed on all
four limbs, and yet, despite the fact that the head which consumes 13% of the CO is not
7
74 of 158.
Significan differences exist in these values, depending on the elasticity of the arterial vasculature.
825
included in the electrical field the reliability of the CO results of this technology has been
validated (Tischenko 1973, Koobi et al 1999).
The common electrophysiological denominator of RIC and ICGWB is in that the same
basal R of a body that is measured by RIC is twice the value measured by ICGWB. As
measured in this series, and by and large as measured by others (Kauppinen et al 2000,
Lukaski et al 1986, Organ et al 1994), the basal R of the region between the wrist and the
ankle is approximately 450 .8 This, when the value of the basal R of ICGWB is in the range
of 200–250 (Tischenko 1973, Lamberts et al 1984, Kauppinen et al 2000). The 2:1 ratio of
the basal R in RIC versus ICGWB is attributed to the fact that in ICGWB the two upper and the
two lower limbs are measured in parallel. Consequently, the cross-sectional area of the limbs
is twice as large in ICGWB, resulting in the reduction of the electrical resistance to one-half.
Thus, it is the same physiological mechanism which facilitates the appropriate measurement
of the total CO by ICGWB, which also holds for the exceptionally accurate RIC results.
Conclusions
The present CO results measured by the NICaS device indicate that, with regard to accuracy
of measuring the CO, the RIC technology outperforms any other ICG technology, being twice
as accurate as TIC.
The peripheral systolic impedance changes are more reliable than the TIC impedance
changes for calculating the cardiac stroke volume.
The main disadvantage of the RIC technology is that in approximately 15% of the patients
who need the test, there are exclusion criteria which preclude the use of the NICaS, a problem
which is also shared by other impedance technologies.
References
Bland J M and Altman D G 1986 Statistical methods for assessing agreement between two methods of clinical
measurement Lancet 1 307–10
Bonjer F H 1950 Circulatieonderzoek door Impedantiemeting Groningen, Drukkerij I, Oppenheim NV
Bonjer F H, Van Den Berg J W and Direden M N J 1952 The origin of the variations of body impedance occurring
during the cardiac cycle Circulation 6 415–20
Cohen A J, Arnaudov D, Zabeeda D, Shultheis L, Lashinger J and Schachner A 1998 Non-invasive measurement of
cardiac output during coronary artery bypass grafting Eur. J. Cardiothorac. Surg. 14 64–9
Cotter G, Moshkovitz Y, Kaluski E, Cohen A J, Miller H, Goor D A and Vered Z 2004 Accurate, noninvasive
continuous monitoring of cardiac output by whole-body electrical bioimpedance Chest 125 1431–40
Critchley L A H, Leung D H Y and Short T G 1996 Abdominal surgery alters the calibration of bioimpedance cardiac
output Int. J. Clin. Monit. Comput. 13 1–8
Drazner M, Thompson B, Rosenberg P, Kaiser P A, James M S N, Boehrer D, Baldwin B J, Dries D L and
Yancy C W 2002 Comparison of impedance cardiography with invasive hemodynamic measurements in patients
with heart failure secondary to ischemic or nonischemic cardiomyopathy Am. J. Cardiol. 89 993–5
Faes Th J C, Raaijmakers J H, Meijer H G, Goovaerts H G and Heethaar R M 1999 Towards a theoretical understanding
of stroke volume estimation with impedance cardiography Ann. New York Acad. Sci. 873 128–34
Frank O 1926 Die Theorie der Pulswellen Z. Biol. 85 95–180
Guidance for Industry Statistical Approach to Establishing Bioequivalence 2001 US Department of Health and Human
Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), BP
Guyton A C and Hall J E 2000 Textbook of Medical Physiology 10th edn (Philadelphia, PA: WB Saunders)
8
It should be borne in mind that in females the electrical resistance between the wrist and the ankle is higher than
in males by 80–100 (Organ et al 1994, Lukaski et al 1986, Hoffer et al 1970). Furthermore, there are differences
in the basal R values of different ethnic populations (Ward et al 2000), where different antropometric characteristics
exist. In the present series, the average basal R of the postoperative cases was lower by 100 compared to the
non-operative cardiac patients, and this is attributed to the over-hydration induced during the surgery.
75 of 158.
826
G Cotter et al
Hamwi G T 1964 Therapy Changing Dietary Concepts in Diabetes Mellitus: Diagnosis and Treatment
ed T S Danowski (New York: American Diabetes Association) pp 73–8
Handelsman H 1991 Public health service reassessment: measuring cardiac output by electrical bioimpedance Health
Technology Assessment Report 6 (US Dept Health and Human Services, Public Agency for Health Care Policy
and Research) pp 1–13
Hoffer E C, Meador C K and Simpson D C 1970 A relationship between whole body impedance and total body water
volume Ann. New York Acad. Sci. 170 452–61
Kauppinen P K, Hyttinen J A and Malmivuo J A 1998 Sensitivity distributions of impedance cardiography using band
and spot electrodes analyzed by a three-dimensional computer model Am. Biomed. Eng. 26 694–702
Kauppinen P K, Koobi T, Hyttinen J and Malmivuo J 2000 Segmental composition of whole-body impedance
cardiogram estimated by computer simulations and clinical experiment Clin. Physiol. 20 106–13
Kim D W, Baker L E, Pearce J A and Kim W K 1988 Origins of the impedance change in impedance cardiography
by a three-dimensional f nite element model IEEE Trans. Biomed. Eng. 35 12
Koobi T, Kaukinen S and Kauppinen P 2001 Comparison of methods of cardiac output measurements Crit. Care
Med. 29 1092 (Letter to the Editor)
Koobi T, Kaukinen S and Turjanmaa V M 1999 Cardiac output can be reliably measured noninvasively after coronary
artery bypass grafting operation Crit. Care Med. 27 2206–11
Kubicek W G, Karnegis J N, Patterson R P, Witsoe D A and Mattson R H 1966 Development and evaluation of an
impedance cardiac output system Aerospace Med. 37 1208–2
Kubicek W G, Kottke F J, Ramos M U, Patterson R P, Witsol D A, Labree J W, Remole W, Layman T E, Schoening H
and Garamela J T 1974 The Minnesota impedance cardiograph: theory and applications Biomed. Eng. 9 410–6
Lamberts R, Visser K R and Zijlstra W G 1984 Impedance Cardiography (Assen, The Netherlands: Van Gorkum)
pp 21–94
Leslie S J, McKee S, Newby D E, Webb D J and Denvir M A 2004 Non-invasive measurement of cardiac output in
patients with chronic heart failure Blood Pressure Monit. 9 277–80
Lukaski H, Bolonchuk W W, Hall C B and Siders W A 1986 Validation of tetrapolar bioelectrical impedance method
to assess human body composition J. Appl. Physiol. 60 1327–32
McVeigh G E, Hamilton P K and Morgan D R 2002 Evaluation of mechanical arterial properties: clinical, experimental
and therapeutic aspects Clin. Sci. 102 51–67
Nyboer J 1950 Electrical impedance plethysmography: a physical and physiologic approach to peripheral vascular
study Circulation 2 811–21
Organ L W, Bradham G B, Gore D T and Lozier S L 1994 Segmental bioelectrical impedance analysis: theory and
application of a new technique J. Appl. Physiol. 77 98–112
Patterson R P 1965 Cardiac output determinations using impedance plethysmography MSc Thesis University of
Minnesota
Patterson R P 1985 Sources of thoracic cardiogenic electrical impedance signal as determined by a model Med. Biol.
Eng. Comput. 23 411–7
Patterson R 1989 Body flui determinations using multiple impedance measurements IEEE Eng. Med. Biol. 8 16–9
Patterson R P, Kubicek W G, Kinnen E, Witsoe D A and Noren G 1964 Development of an electrical impedance
plethysmography system to monitor cardiac output Proc. 1st Annual Rocky Mt Bioeng. Symp. pp 56–71
Raaijmakers E, Faes Th J C, Scholten R J P M, Goovaerts H G and Heethaar R 1999 A meta-analysis of published
studies concerning the validity of thoracic impedance cardiography Ann. New York Acad. Sci. 873 121–34
Risacher F, Jossinet J and Schmitt M 1995 Comparison of global and local pulse wave velocity in man measured by
multichannel impedance plethysmography Proc. 9th Int. Conf. on Electrical Bioimpedance
Sageman W S, Riffenburgh R H and Spiess B D 2002 Equivalence of bioimpedance and thermodilution in measuring
cardiac index after cardiac surgery J. Cardiothorac. Vasc. Anesth. 16 8–14
Thomasset A 1962 Bioelectrical properties of tissue impedance measurements Lyon Med. 207 107–18
Tischenko M I 1973 Estimation of stroke volume by integral rheogram of the human body Sechenov Physiol. J. 59
1216–24 (in Russian)
Torre-Amione G et al 2004 Non-invasive measurement of cardiac output by whole-body electrical bioimpedance in
patients treated for acute heart failure: a prospective, double-blind comparison with thermodilution J. Am. Coll.
Cardiol. 5 209A (abstract)
Van De Water J M, Miller T W, Vogel R L, Mount B E and Dalton M L 2003 Impedance cardiography—the next
vital sign technology? Chest 123 2028–33
Wang Y, Haynor D R and Kim Y 2001 A f nite-element study of the effects of electrode position on the measured
impedance change in impedance cardiography IEEE Trans. Biomed. Eng. 48 1390–401
Wang L and Patterson L 1995 Multiple sources of the impedance cardiogram based on 3D f nite difference human
thorax models IEEE Trans. Biomed. Eng. 42 141–8
76 of 158.
827
Ward L C et al 2000 Association between ethnicity, body mass index, and bioelectrical impedance. Implications for
the population specificit of prediction equations Ann. New York Acad. Sci. 904 199–204
Wotton M J, Thomas B J, Cornish B H and Ward L C 2000 Comparison of whole body and segmental bioimpedance
methodologies for estimating total body water Ann. New York Acad. Sci. 904 181–6
Wtorek J 2000 Relations between components of impedance cardiogram analyzed by means of f nite element model
and sensitivity theorem Ann. Biomed. Eng. 28 1352–61
77 of 158.
Value of Noninvasive Hemodynamics to Achieve Blood Pressure Control in
Hypertensive Subjects
Ronald D. Smith, Pavel Levy, Carlos M. Ferrario and for the Consideration of
Noninvasive Hemodynamic Monitoring to Target Reduction of Blood Pressure Levels
Study Group
Hypertension 2006;47;771-777; originally published online Mar 6, 2006;
DOI: 10.1161/01.HYP.0000209642.11448.e0
Hypertension is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX
72514
Copyright © 2006 American Heart Association. All rights reserved. Print ISSN: 0194-911X. Online
ISSN: 1524-4563
The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://hyper.ahajournals.org/cgi/content/full/47/4/771
Subscriptions: Information about subscribing to Hypertension is online at
http://hyper.ahajournals.org/subscriptions/
Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax:
410-528-8550. E-mail:
[email protected]
Reprints: Information about reprints can be found online at
http://www.lww.com/reprints
78 of 158.
Downloaded from hyper.ahajournals.org by on June 11, 2011
Value of Noninvasive Hemodynamics to Achieve Blood
Pressure Control in Hypertensive Subjects
Ronald D. Smith, Pavel Levy, Carlos M. Ferrario; for the Consideration of Noninvasive
Hemodynamic Monitoring to Target Reduction of Blood Pressure Levels Study Group
Abstract—Abnormal hemodynamics play a central role in the development and perpetuation of high blood pressure. We
hypothesized that hypertension therapy guided by noninvasive hemodynamics with impedance cardiography could aid
primary care physicians in reducing blood pressure more effectively. Uncontrolled hypertensive patients on 1 to 3
medications were randomized by 3:2 ratio to either a standard arm or hemodynamic arm that used impedance
cardiography (BioZ, CardioDynamics). Each patient completed 5 study visits with a 2-week washout period followed
by 3 months of treatment. A total of 164 patients from 11 centers completed the study, 95 in the standard arm and 69
in the hemodynamic arm. At baseline and after washout, there were no differences between arms in number of
medications or demographic, blood pressure, or hemodynamic characteristics. Systolic blood pressure reductions in the
hemodynamic arm were greater from baseline (19 mm Hg versus 11 mm Hg; P⬍0.01) and after washout (25 mm Hg
versus 19 mm Hg; P⬍0.05). Diastolic blood pressure reductions were also greater in the hemodynamic arm from
baseline (12 mm Hg versus 5 mm Hg; P⬍0.001) and after washout (17 mm Hg versus 10 mm Hg; P⬍0.001). The
hemodynamic arm achieved goal blood pressure (⬍140/90 mm Hg) more frequently (77% versus 57%; P⬍0.01) and
a more aggressive blood pressure level (⬍130/85 mm Hg) more frequently (55% versus 27%; P⬍0.0001). These study
results indicate that antihypertensive therapy guided by impedance cardiography in uncontrolled hypertensive patients
on ⱖ1 medications is more effective than standard care. (Hypertension. 2006;47:771-777.)
Key Words: hemodynamics 䡲 cardiac output 䡲 vascular resistance 䡲 hypertension, arterial
䡲 hypertension, essential 䡲 blood pressure
A
pproximately 65 million people in the United States1
and 1 billion people worldwide2 have hypertension; it
is the most common reason adults visit US physicians.3
Hypertension is a major public health concern, because it
significantly increases risk of coronary artery disease,
heart failure, renal disease, and stroke.4 In spite of major
public health and medical education efforts and availability
of effective antihypertensive agents, blood pressure (BP) control rates in the United States remain low, with only 31% of
hypertensives and 54% of those actively treated and taking
medications achieving BP ⬍140/90 mm Hg.5 Why is BP control
such an elusive goal? The reasons are numerous and complex.
However, inadequate pharmacological treatment remains the
most common cause of uncontrolled BP in actively treated
patients.6
Hypertension is a hemodynamic-related disorder. BP rises
as the result of increased systemic vascular resistance (SVR),
cardiac output (CO), fluid volume, or a combination of these
factors.7,8 Consequently, antihypertensive agents lower BP by
reducing SVR, CO, fluid volume, or combinations thereof.9
Previous authors hypothesized that hemodynamic information could help tailor therapy and subsequently improve BP
control.10 Invasive procedures for hemodynamic profiling are
not warranted in outpatient clinics, and noninvasive procedures, such as echocardiography, are costly and operator
dependent.11
Impedance cardiography (ICG) has emerged as a reliable
noninvasive method to measure hemodynamics in physician
offices. In a randomized, controlled trial, ICG-guided treatment improved BP control rates in resistant hypertension
treated by hypertension specialists.12 We hypothesized that
ICG-guided treatment could aid physicians in reducing BP
more effectively than standard care in a population of
uncontrolled hypertensive patients receiving 1 to 3 medications in a primary care setting.
Methods
Eligibility
Male and female patients (age range, 18 to 75 years) were eligible if
they had a diagnosis of essential hypertension and were currently on
Received August 10, 2005; first decision August 29, 2005; revision accepted December 12, 2005.
From the Hypertension and Vascular Disease Center, Wake Forest University School of Medicine, Winston-Salem, NC.
R.D.S. and C.M.F. received honorarium from CardioDynamics for consultation and speaker fees. Site investigators received study support from
CardioDynamics.
Correspondence to Carlos Ferrario, Hypertension and Vascular Disease Center, Wake Forest University School of Medicine, Medical Center Blvd,
Winston-Salem, NC 27157-1032. E-mail [email protected]
© 2006 American Heart Association, Inc.
Hypertension is available at http://www.hypertensionaha.org
79 of 158.
DOI: 10.1161/01.HYP.0000209642.11448.e0
771
Downloaded from hyper.ahajournals.org
by on June 11, 2011
772
Hypertension
April 2006
Figure 1. Study design comparing efficacy of standard arm vs hemodynamic arm treatment of high BP.
1 to 3 antihypertensive medications with systolic BP of 140 to
179 mm Hg and/or diastolic BP of 90 to 109 mm Hg. Patients were
excluded if they were on ⬎3 antihypertensive agents; had abnormal
laboratory findings; or had history of heart failure, left ventricular
ejection fraction ⬍40%, atrial fibrillation, severe valvular disease,
cerebrovascular event within 3 months, severe renal disease, nephrotic syndrome, or cirrhosis. Patients in whom ICG might be
subject to technical limitations were excluded (height ⬍47 or ⬎75
inches or weight ⬍66 or ⬎341 pounds, presence of activated
minute-ventilation pacemaker, known hypersensitivity to sensor gel
or adhesive, or skin lesion interfering with sensor placement).
Patients who were enrolled and subsequently found to have not met
the inclusion/exclusion criteria were terminated and excluded from
analysis. Therefore, this was not an intention-to-treat analysis. The
study was approved by an independent institutional review board,
which adheres to the Declaration of Helsinki and US Code of Federal
Regulations. These hypertensive outpatients provided written informed consent and had study procedures consistent with the
protocol (no. 20021400).
Hemodynamic Evaluation Assignment
Eligible patients (N⫽164) underwent a 2-week washout period at
which time all of the antihypertensive medications were discontinued
according to the manufacturer’s recommendations. After screening
and medication washout, each patient had 3 monthly office visits
(Figure 1). After the 2-week washout period, patients meeting
inclusion/exclusion criteria were randomized in a 3:2 ratio to the
standard arm (n⫽95) or ICG-aided hemodynamic arm (n⫽69) using
a central telephone service and stratified by site. All of the physicians
were educated on the hemodynamic treatment strategy illustrated in
Figure 2.
Procedures
BP determinations were made in the seated position using the
oscillometric technique. ICG data were collected by trained technicians at each visit in all of the patients, but ICG findings were not
revealed in the standard arm to treating physicians or patients. ICG
was performed with patients in the supine position, resting for 5
minutes before measurement (BioZ ICG Monitor, CardioDynamics).
ICG involves the measurement of thoracic impedance through
placement of 4 dual sensors, 2 on the neck and 2 on the chest.
Electrical impedance changes are digitally processed to calculate
CO, SVR, and thoracic fluid content (TFC).13 CO and SVR are
normalized for body size by indexing to each patient’s body surface area to obtain cardiac index (CI) and SVR index (SVRI). TFC
is the inverse of baseline chest impedance, and any changes in TFC
are directly proportional to total fluid (intravascular and extravascular) changes.14 TFC has different normal ranges for each gender that
are displayed and printed for reference. The reproducibility of this
ICG device in stable outpatients has been established,15 and accuracy
Figure 2. Suggested treatment strategy for hemodynamic arm; BB indicates ␤ blocker; CAA, central acting agent; and VD, vasodilator.
80 of 158.
Smith et al
has been validated versus invasive methods in patients with various
cardiovascular disorders.16,17
Outcome Measures
Physician investigators were instructed that the treatment goal was to
reduce systolic and diastolic BP as low as they believed would be
beneficial to the patient and to achieve sustained BP ⬍140/
90 mm Hg. The primary study end points were reductions in systolic
and diastolic BP from baseline and post-washout visit. Additional study end points were achievement of: (1) goal BP ⬍140/
90 mm Hg, (2) more aggressive BP of ⬍130/85 mm Hg, and (3) BP
of ⬍140/90 mm Hg with normal values of CI and SVRI. Normal
range for CI was defined as 2.5 to 4.2 L/min per m2 and for SVRI as
1680 to 2580 dyne⫻s⫻m2/cm5. Isolated systolic hypertension was
defined as systolic BP ⬎140 mm Hg and diastolic BP ⬍90 mm Hg.
Noninvasive Hemodynamics and BP Control
enrollment criteria (BP ⬎140/90 mm Hg at screening) and
were removed as protocol violations. No reported adverse
events (minor or serious) were attributable to ICG.
There were no differences in the number of antihypertensive medications, patient demographic, clinical, BP, or ICG
variables at baseline or after washout (Table 1). At baseline,
there were no differences in the percentage of patients in the
standard versus hemodynamic arm on 1 (42% versus 45%;
P⬎0.05), 2 (48% versus 44%; P⬎0.05), or 3 (6% versus
10%; P⬎0.05) medications. Baseline medication usage in the
TABLE 1.
Patient Characteristics
Variable
Standard
Care
(n⫽95)
Hemodynamic
Care
(n⫽69)
P
Value
Age, y
54.5⫾9.4
55.2⫾9.2
ns
Body mass index, kg/m2
30.2⫾6.3
30.8⫾5.1
ns
51 (53.4)
38 (55.1)
ns
Interventions
After randomization, therapy was initiated in all of the patients at the
post-washout visit, 2 weeks after screening. Physician investigators
prescribed medications consistent with published guidelines, their
usual practice patterns, and patient clinical characteristics. In the
hemodynamic arm, the treating physician was also encouraged to use
a hemodynamic treatment strategy to guide therapeutic decisions
about pharmacological agents and dosing (Figure 2).
Physicians could share ICG information with patients in the
hemodynamic arm, and patients in both arms received education on
the importance of medication compliance, which was reinforced with
a nurse telephone call midway between each study visit. Compliance
was assessed at each visit by asking patients to estimate the
percentage of prescribed pills they had taken over the previous
month. Patients were considered compliant with the prescribed
protocol if pill count was ⬎85% over the prior month.
Data from case report forms were collected by study coordinators
and entered into a locked database. Statistical analysis was performed with SAS statistical analysis software, version 8.2. Continuous variables are expressed as mean⫾SD and categorical variables
as n (%). Differences in continuous variables between treatment
groups were examined by the Student t test and by ANOVA and in
discrete variables using Fisher’s exact tests. Subgroup analysis was
performed in subjects with isolated systolic hypertension, age ⱖ55
years, and those receiving a thiazide diuretic. Additional evaluation
of age-specific results was performed by a 2-way ANOVA for
achievement of BP end points, in which treatment arm and dichotomized age (ⱖ55 years) were included in the model. In combination
agents, each class and dosage was counted separately for analysis.
Equivalency of defined daily doses for each class of medication was
calculated using World Health Organization criteria.18 Medication
changes were evaluated in visits where such changes affected BP end
points (visits 2 versus 1, 3 versus 2, and 4 versus 3). Medication class
and dose were compared with the prior study visit, with any change
in class or dose counted separately. Sample size was powered using
5 mm Hg as the detectable difference between treatment groups with
a type I error of 5% and type II error of 20%. The expected
heterogeneity in treatment approach for patients in the standard arm
was offset by the greater number of patients randomized to the
standard arm. Although this approach increased the probability that
the standard arm results would reflect actual practice patterns, it
required a larger sample size to power the study.
Results
Eleven primary care centers participated in the study between
November 2002 and November 2004. Of 262 patients
screened, 184 were randomized. A total of 164 patients (95 in
the standard arm and 69 in the hemodynamic arm) completed
the study and were analyzed. There were 20 early terminations, including 2 who withdrew and 18 who were randomized but were subsequently found not to have met BP
81 of 158.
773
Men
Ethnicity
White, non-Hispanic
75 (79.0)
53 (76.8)
ns
White, Hispanic
7 (7.4)
5 (7.3)
ns
Black
8 (8.4)
6 (8.7)
ns
Asian
3 (3.2)
3 (4.4)
ns
Type II diabetes mellitus
4 (4.2)
3 (4.4)
ns
Ischemic heart disease
2 (2.1)
5 (7.3)
ns
14 (14.7)
12 (17.4)
ns
147⫾9
History
Hyperlipidemia
Baseline BP and
hemodynamics
Systolic BP, mm Hg
148⫾12
ns
Diastolic BP, mm Hg
87⫾10
89⫾8
ns
Heart rate, bpm
75⫾12
74⫾13
ns
Cardiac index, L/min/m2
2.8⫾0.5
2.9⫾0.6
ns
Systemic vascular
resistance index,
dyne⫻s⫻m2/cm5
2933⫾576
2956⫾605
ns
Thoracic fluid content,
/kOhm
28.6⫾4.9
28.0⫾4.8
ns
46 (48.4)
31 (44.9)
ns
Systolic BP, mm Hg
156⫾13
155⫾13
ns
Diastolic BP, mm Hg
92⫾9
94⫾9
ns
Isolated systolic hypertension
at baseline
Post-washout BP and
hemodynamics
Heart rate, bpm
79⫾12
78⫾14
ns
2.9⫾0.5
2.9⫾0.5
ns
Systemic vascular
resistance index,
dyne⫻s⫻m2/cm5
3083⫾630
3122⫾672
ns
Thoracic fluid content,
/kOhm
29.1⫾5.0
28.4⫾4.3
ns
1.7⫾0.8
1.7⫾0.7
ns
Medications
Total antihypertensive
medications
Categorical variables expressed as n (%), continuous variables as
mean⫾SD; ns indicates not significant.
774
Hypertension
April 2006
standard versus hemodynamic arm was as follows: ␣ blockers
(2.1% versus 1.4%; P⬎0.05), angiotensin converting enzyme
inhibitors (ACEI; 53.7% versus 47.8%; P⬎0.05), angiotensin
II receptor blockers (ARB; 14.7% versus 29.0%; P⬍0.05), ␤
blockers (23.2% versus 13.0%; P⬎0.05), calcium channel
blockers (CCB; 33.7% versus 39.1%; P⬎0.05), central acting
agents (0% versus 1.4%; P⬎0.05), diuretics (31.6% versus
26.1%; P⬎0.05), and other vasodilators (0% versus 0%;
P⬎0.05).
BP and ICG values at the final visit and their differences from
baseline and post-washout visits are shown in Table 2. Systolic
BP reductions were greater in the hemodynamic arm from
baseline (19⫾17 versus 11⫾18 mm Hg; P⬍0.01) and postwashout (25⫾18 versus 19⫾17 mm Hg; P⬍0.05). Diastolic
BP reductions were also greater in the hemodynamic arm
from baseline (12⫾11 versus 5⫾12 mm Hg; P⬍0.001) and
post-washout (17⫾12 versus 10⫾11 mm Hg; P⬍0.001).
Final BP was lower in the hemodynamic arm (129/76⫾14/11
versus 136/82⫾15/10 mm Hg; P⬍0.01). Figure 3 demonstrates that goal BP (⬍140/90 mm Hg) was achieved more
frequently in the hemodynamic arm (77% versus 57%;
P⬍0.01), and the more aggressive BP (⬍130/85 mm Hg) was
also achieved more often (55% versus 27%; P⬍0.0001).
Patients with isolated systolic hypertension in the hemodynamic arm (n⫽31) had greater systolic BP reductions from
TABLE 2.
Final BP and Hemodynamic Values
Standard
Care
(n⫽95)
Hemodynamic
Care
(n⫽69)
P
Value
136⫾15
129⫾14
⬍0.01
⌬ baseline to final
⫺11⫾18
⫺19⫾17
⬍0.01
⌬ post-washout to final
⫺19⫾17
⫺25⫾18
⬍0.05
82⫾10
76⫾11
⬍0.01
Variable
Systolic BP, mm Hg
Final
Diastolic BP, mm Hg
Final
⫺5⫾12
⫺12⫾11
⬍0.001
⫺10⫾11
⫺17⫾12
⬍0.001
77⫾13
76⫾11
ns
Heart rate, bpm
Final
1⫾12
2⫾13
ns
⫺2⫾13
⫺2⫾13
ns
Final
2.9⫾0.5
2.9⫾0.5
ns
0.1⫾0.5
0.0⫾0.5
ns
0.0⫾0.5
0.0⫾0.5
ns
index, dyne⫻s⫻m2/cm5
2714⫾619
2523⫾581
⬍0.05
⫺219⫾667
⫺433⫾660
⬍0.05
⫺369⫾642
⫺599⫾738
⬍0.05
Final
Thoracic fluid content, /kOhm
Final
27.8⫾4.1
28.2⫾4.9
ns
⫺0.8⫾3.6
0.1⫾3.0
ns
⫺1.2⫾3.3
⫺0.2⫾2.7
Variables expressed as mean⫾SD; ns indicates not significant.
82 of 158.
⬍0.05
Figure 3. Target BP achievement; with 95% CI; *P⬍0.01 vs
standard care, †P⬍0.0001 vs standard care.
baseline (22⫾16 versus 11⫾17 mm Hg; P⬍0.01) and postwashout (28⫾16 versus 18⫾16 mm Hg; P⬍0.05) than those
in the standard arm (n⫽46). Patients ⱖ55 years in the
hemodynamic arm (n⫽33) had greater systolic BP reductions
compared with the standard arm (n⫽51) from baseline (21⫾17
versus 11⫾20 mm Hg; P⬍0.05) and trended greater from
post-washout (26⫾20 versus 21⫾19 mm Hg; P⬎0.05). Diastolic BP reductions were also greater in those ⱖ55 years
in the hemodynamic arm from baseline (13⫾11 versus
4⫾12 mm Hg; P⬍0.001) and post-washout (16⫾11 versus
10⫾12 mm Hg; P⬍0.05). In patients ⱖ55 years, goal BP
(⬍140/90 mm Hg) was achieved more frequently in the hemodynamic arm (76% versus 53%; P⬍0.05), and the more aggressive BP (⬍130/85 mm Hg) was also achieved more often (58%
versus 27%; P⬍0.01). ANOVA also indicated that age ⱖ55
years had no effect on study end points (P⬎0.05).
SVRI was reduced to a greater extent in the hemodynamic
arm than in the standard arm from baseline and post-washout.
There were no significant differences between arms at the
final visit for heart rate, CI, or TFC. However, the standard
arm had a small but significant reduction in TFC from
post-washout to final. The percentage of patients achieving
normal hemodynamic values defined as simultaneously normal values of BP, CI, and SVRI was 52% in the hemodynamic arm and 29% in the standard arm (P⬍0.01). Patients in
either arm who achieved BP ⬍130/85 mm Hg had lower
SVRI (2646⫾592 versus 2855⫾606 dyne⫻s⫻m2/cm5; P⬍0.05)
and lower CI (2.7⫾0.5 versus 2.9⫾0.5 L/min/m2; P⬍0.05) than
those who did not achieve BP ⬍130/85 mm Hg. Patients in
the hemodynamic arm who achieved BP ⬍130/85 mm Hg
trended toward lower SVRI (2446⫾580 versus 2573⫾612
dyne⫻s⫻m2/cm5; P⬎0.05) and higher CI (2.8⫾0.5 versus
2.6⫾0.5 L/min/m2; P⬎0.05) than those in the standard care
arm who achieved BP ⬍130/85 mm Hg.
In the visit after medication washout, patients in the
hemodynamic arm were more likely to be prescribed an
ACEI, ARB, or CCB (92.8% versus 80.0%; P⬍0.05). Over
Smith et al
the course of the study, patients in the hemodynamic arm
were more likely to be prescribed an ACEI, ARB, or CCB
when their SVRI was high, per the hemodynamic treatment
strategy (78.3% versus 67.1%; P⬍0.05). However, there
were no differences in the other 2 treatments encouraged by
the hemodynamic treatment strategy, ␤ blocker use based on
high CI, or in diuretic use when TFC did not decrease in
response to diuretic initiation or increase. Patients in the
hemodynamic arm were more likely to avoid ␤ blocker use or
to have their ␤ blocker reduced in the presence of low or
normal CI (85.4% versus 77.0%; P⬍0.05) as the hemodynamic strategy suggested. Direct vasodilators were not used,
and, therefore, changes in vasodilator use in the presence of
normal SVRI were not evaluated. Table 3 lists all of the
medications at the final visit. Patients in the standard arm
were on 2.0⫾0.8 medications compared with 2.1⫾0.9 for the
hemodynamic arm (P⬎0.05). In the hemodynamic arm, ARB
use was higher (46.4% versus 30.5%; P⬍0.05), and ACEI
use was similar (49.3% versus 53.7%; P⬎0.05). However,
the percentage of patients in the hemodynamic arm who were
prescribed either an ACEI or ARB was not significantly
different (87.0% versus 76.8%; P⬎0.05). There were no
differences in the percentage of patients in the hemodynamic
care on 1 (25% versus 26%), 2 (48% versus 53%), 3 (19%
versus 15%), 4 (9% versus 5%), or 5 (0% versus 1%)
medications at the final visit (P⬎0.05 for all). There were a
greater number of medication dose increases in the standard
versus hemodynamic arm (3.6⫾1.3 versus 3.0⫾1.2; P⬍0.001),
as well as a greater number of dose decreases (2.7⫾1.3 versus
1.7⫾1.0; P⬍0.001). Medication class changes in the standard
and hemodynamic arm were similar in both class initiation
(1.0⫾0.9 versus 1.1⫾0.9; P⬎0.05) and removal (0.8⫾0.8 versus 0.7⫾0.8; P⬎0.05).
Thiazide diuretic use at baseline was similar in the standard
versus hemodynamic arm (28.4% versus 24.6%; P⬎0.05). A
similar proportion of patients were prescribed thiazide diuretics at some point during the trial in both the standard and
TABLE 3.
Final Antihypertensive Medications
Antihypertensive
Medication
Standard
Care
(n⫽95)
Hemodynamic
Care
(n⫽69)
P
Value
No. at final visit
2.0⫾0.8
2.1⫾0.9
ns
1 (1.0)
1 (1.4)
ns
ACEI
51 (53.7)
34 (49.3)
ns
ARB
29 (30.5)
32 (46.4)
⬍0.05
␣ Blocker
␤ Blocker
18 (19.0)
6 (8.7)
ns
Calcium channel blocker,
dihydropyridine
36 (37.9)
28 (40.6)
ns
Calcium channel blocker,
nondihydropyridine
6 (6.3)
7 (10.1)
ns
Central acting agent
0 (0.0)
1 (1.4)
ns
ns
Diuretic, thiazide
32 (33.7)
24 (34.8)
Diuretic, loop
1 (1.1)
0 (0.0)
ns
Diuretic, potassium sparing
6 (6.3)
3 (4.3)
ns
Vasodilator
0 (0.0)
0 (0.0)
ns
Categorical variables expressed as n (%); ns indicates not significant.
83 of 158.
775
hemodynamic arms (44.2% versus 40.2%; P⬎0.05), and use
was similar at the final visit (33.7% versus 34.8%; P⬎0.05).
Medication doses were not different between arms except that
patients in the standard arm were on higher doses of thiazide
diuretics (18.9⫾8.3 versus 13.0⫾2.6 mg/day; P⬍0.01).
There were no differences in the hemodynamic arm in the
dosing of ACEIs (19.1 versus 19.1 mg/day; P⬎0.05), ARBs
(93.9 versus 87.0 mg/day; P⬎0.05), ␤ blockers (65.6 versus
80.9 mg/day; P⬎0.05), or CCBs (7.9 versus 7.9 mg/day;
P⬎0.05). The greater mean dose of thiazide diuretics was
because of a higher percentage of patients taking ⱖ25 mg/day
versus 12.5 mg/day in the standard arm (40.1% versus 8.3%;
P⬍0.05). When the study end points were analyzed only for
patients on a thiazide diuretic in the final visit, patients in
hemodynamic arm had greater decreases in systolic BP from
baseline (26⫾19 versus 8⫾17 mm Hg; P⬍0.001) and postwashout (36⫾17 versus 21⫾20 mm Hg; P⬍0.01) and greater
decreases in diastolic BP from baseline (16⫾11 versus
3⫾14 mm Hg; P⬍0.001) and post-washout (20⫾12 versus
11⫾13 mm Hg; P⬍0.01). There were no differences in
patient-reported compliance between the standard and hemodynamic arm in visit 3 (96.8% versus 97.1%; P⬎0.05), 4
(96.8% versus 98.6%; P⬎0.05), or 5 (100% versus 100%;
P⬎0.05) or when these visits were combined (97.9% versus
98.6%; P⬎0.05).
Discussion
Our results demonstrate that ICG-guided antihypertensive
treatment was more effective in reducing BP than standard
therapy and empiric selection of antihypertensive medications. Patients in the 2 arms of our study were not significantly different at baseline, and each patient underwent a
medication washout period to additionally equalize the 2
groups. The 57% BP control rate in the standard arm was
substantial and compared favorably to BP control rates of
long durations in large antihypertensive trials.19,20 However,
the 77% BP control rate in the hemodynamic arm was even
more impressive with an 8/7 mm Hg greater BP reduction
from baseline and a 6/7 mm Hg greater BP reduction from
post-washout. As a result, patients in the hemodynamic arm
achieved goal BP of ⬍140/90 mm Hg 35% more often (77%
versus 57%) and the more aggressive level of BP control
(⬍130/85 mm Hg) 104% more often (55 versus 27%) than
those in the standard arm. The hemodynamic arm maintained
superiority in 3 key subgroups: patients who were older, on
thiazide diuretics, or had isolated systolic hypertension.
Why did the hemodynamic arm achieve greater reductions
in BP and higher BP control rates than the standard arm? The
fundamental difference between the two arms was that patient
treatment in the hemodynamic arm was individualized and
targeted at the hemodynamic abnormality associated with the
elevated BP. This approach led to greater reductions in SVRI
in the hemodynamic arm, which allowed greater decreases in
both systolic and diastolic BP. The mechanistic and hemodynamically based improvement in BP was also demonstrated in patients achieving BP ⬍130/85 mm Hg through
significantly lower SVRI and higher CI in both arms. In
theory, the larger drop in SVRI and BP levels in the
hemodynamic arm could have occurred through use of more
776
Hypertension
April 2006
medications, more effective medications, greater dosing intensity, more effective combination therapy, or better patient
compliance. Our study allowed full discretion by the physician in choosing the agents, and a multitude of classes and
doses within classes were used. The study was not powered to
find small disparities in medication use, and most medication
differences did not reach statistical significance.
On the other hand, some differences are worth noting.
Patients in the standard arm were more likely to experience
both increases and decreases in their medication doses,
whereas medication class changes were not different between
arms. This result might have been expected, because treatment in the standard arm followed guidelines and usual
practice patterns in which a stepped approach to therapy
contributes to a “trial-and-error” method of determining
whether agents and doses are working. In the hemodynamic
arm, the initial selection of antihypertensive medications
appears to have been influenced by the hemodynamic data,
because these patients were more likely to be prescribed a
vasodilating agent to reduce SVRI. Additionally, the hemodynamic treatment strategy influenced medication use when
SVRI was considered high, because patients in the hemodynamic arm were more likely to have received an ACEI, ARB,
or CCB, as was suggested. The hemodynamic treatment
strategy did not influence the prescription of ␤ blockers in the
presence of high CI or in diuretic use in response to TFC
changes. However, ␤-blocker use was lower or reduced in the
presence of low or normal CI in the hemodynamic arm.
Although the final number of antihypertensive medications
given to patients in both arms of the study was similar,
patients in the hemodynamic arm were more likely to be
prescribed an ARB. However, when ACEI and ARB use was
combined into a single category (renin–angiotensin–aldosterone system inhibitors), the hemodynamic arm only trended
toward greater use at the final visit (87.0% versus 76.8%).
Thiazide diuretic use increased during the study but was
lower than in some pharmacological trials and what hypertension guidelines currently suggest. However, the percentage of patients in both arms who were prescribed a thiazide
diuretic at the final visit was very similar to the 35.6% usage
that was reported in recent analysis of over 25 000 hypertensive patients.21 The lower use of diuretics and ␤ blockers also
follows the previously recognized physician preference for
other antihypertensive agents.22 Some might hypothesize that
greater BP reductions could have been achieved in the
standard arm if diuretics were used more frequently. However, when patients taking a thiazide diuretic were examined
as a subgroup, the hemodynamic arm maintained its superiority. Additionally, although the higher doses of thiazide
diuretics in the standard arm may have contributed to a
greater drop in TFC from the post-washout visit, they did not
lead to better BP control.
Our study was not intended to evaluate whether a particular
antihypertensive agent was more effective at reducing BP
than another. Rather, it was designed to determine whether
providing hemodynamic data to the physician and patient
could more effectively reduce BP. Whether hemodynamic
data led to a more tailored approach to selection and monitoring of antihypertensive agents or by other factors, it
84 of 158.
resulted in greater reduction in BP and SVRI and better BP
control. Physicians cannot adequately estimate hemodynamics from routine clinical examination or BP measurements,23
because at similar levels of BP, SVR and CO can vary widely.
Therefore, the addition of accurate, noninvasive, and readily
obtainable hemodynamic measurements is clinically relevant.
Importantly, the current study also showed that patients in
the hemodynamic arm were almost twice as likely to achieve
BP control with normalization of both CI and SVRI. Improvements in vascular resistance may result in greater
benefits in reducing cardiovascular risk than improvement in
BP alone,24 and differences in SVRI at the same BP may
explain poorer prognosis for men versus women25 and black
versus nonblack patients.26 Hemodynamics are also known to
change with age. In older subjects, decreased arterial compliance and CI lead to increased SVRI, arterial BP, and pulse
pressure.27 In spite of the expected differences in the hemodynamics of older patients, this study demonstrated that
hemodynamically driven, individualized therapy was similarly effective regardless of age or existence of isolated
systolic hypertension.
The use of ICG to achieve greater BP control offers the
potential for better short-term use of healthcare resources. In
addition, the long-term benefits of even small levels of BP
reduction are well known. A sustained BP reduction of
4/3 mm Hg is expected to reduce stroke risk 23%, coronary
heart disease events 15%, heart failure 16%, and overall
mortality 14%.28 Accordingly, a recent meta-analysis of
major hypertension trials reveals that an antihypertensive
agent is judged favorably when it produces mean BP improvements versus placebo of only 3 or 4 mm Hg or versus
another antihypertensive agent of only 1 or 2 mm Hg.29
Previously, ICG has been used to profile hemodynamic
variability across BP values30 and to identify left ventricular
dysfunction.31 Changes in ICG parameters have demonstrated
the hemodynamic effect of antihypertensive agents32,33 and
dietary sodium.34 ICG-guided therapy has shown benefit in a
case series,35 observational study,36 and a randomized trial in
resistant hypertensive patients.12 In the randomized trial,
ICG-guided therapy resulted in better final BP and greater BP
control. Similar to our study, that study showed no differences in the number of medications between arms. In contrast
to our study with lower diuretic doses and fewer medication
changes in the hemodynamic arm, resistant hypertension
patients receiving ICG-guided therapy had higher diuretic
doses and more medication changes. The differences between
the studies might be expected because of the difference in
patients (severe hypertension on more medications versus
milder hypertension on fewer medications) and setting (specialist versus generalist). However, in both studies, ICGguided therapy led to more effective treatment as evidenced
by better BP outcomes.
The conclusions of this study may be limited to its duration
of 3 months. However, in pharmacological trials, short-term
reductions in BP are typically sustained over longer periods.37
Another limitation may be in our use of patient-reported
medication compliance. Without using automatic counting
procedures, our goal was to educate both arms equally and to
reinforce patient compliance with follow-up phone calls.
Smith et al
Lastly, treatment differences in the hemodynamic arm do not
imply superiority of one medication over another, because the
study was not designed to evaluate this question.
18.
Perspectives
The results of this study indicate that ICG-guided antihypertensive therapy in uncontrolled hypertensive patients on 1 to
3 antihypertensive medications is more effective than standard care. This was evident by greater reductions in systolic
and diastolic BP and by achieving a better level of BP control.
Our study showed that, in clinical practice, inclusion of ICG
hemodynamic assessment may improve BP control rates in
patients who are not controlled on initial therapy.
19.
20.
Acknowledgments
Consideration of Noninvasive Hemodynamic Monitoring to Target
Reduction of Blood Pressure Levels (CONTROL) Site Investigators:
Fred E. Abbo, Douglas C. Beatty, Donald M. Brandon, Milan L.
Brandon, Anthony V. Dallas, Jr, Lawrence Dinenberg, Mazhar El
Amir, Nigar Enayat, Neil W. Hirschenbein, Dorothy Lebeau, Bernard A. Michlin, John Millspaugh, Nell Nestor, Melissa Noble,
Robin F. Spiering, Joseph Taylor, and Allen R. Walker. Statistical
support was provided by Gerard Smits and the study was sponsored
by CardioDynamics (San Diego, CA).
21.
22.
23.
24.
References
1. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella EJ, Sorlie P. The
burden of adult hypertension in the United States 1999 to 2000: a rising
tide. Hypertension. 2004;44:398 – 404.
2. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J.
Global burden of hypertension: analysis of worldwide data. Lancet. 2005;
365:217–223.
3. Chery DK, Woodwell DA. National Ambulatory Care Medical Survey:
2000 summary. Advance Data. 2002;328:1–32.
4. American Heart Association. Heart Disease and Stroke Statistics —2004
Update. Dallas, Texas: American Heart Association; 2003.
5. Wang Y, Wang QJ. The prevalence of prehypertension and hypertension
among US adults according to the new joint national committee guidelines:
new challenges of the old problem. Arch Intern Med. 2004;164:2126–2134.
6. Yakovlevitch M, Black HR. Resistant hypertension in a tertiary care
clinic. Arch Intern Med. 1991;151:1786 –1792.
7. Ferrario CM, Page IH. Current views concerning cardiac output in the
genesis of experimental hypertension. Circ Res. 1978;43:821– 831.
8. Davidson RC, Ahmad S. Hemodynamic profiles in essential and secondary hypertension. In: Izzo JL, Black HR, eds. Hypertension Primer.
3rd ed. Dallas, TX: Council on High Blood Pressure Research, American
Heart Association; 2003:349 –351.
9. Sullivan JM, Schoeneberger TE, Ratts ET, Palmer ET, Samaha JK,
Mance CJ, Muirhead EE. Short-term therapy of severe hypertension.
Hemodynamic correlates of the antihypertensive response in man. Arch
Intern Med. 1979;139:1233–1239.
10. Egan B, Schmouder R. The importance of hemodynamic considerations
in essential hypertension. Am Heart J. 1988;116:594 –599.
11. Northridge DB, Findlay IN, Wilson J, Henderson E, Dargie HJ. Noninvasive determination of cardiac output by Doppler echocardiography
and electrical bioimpedance. Br Heart J. 1990;63:93–97.
12. Taler SJ, Textor SC, Augustine JE. Resistant hypertension: Comparing hemodynamic management to specialist care. Hypertension. 2002;39:982–988.
13. Ventura HO, Taler SJ, Strobeck JE. Hypertension as a hemodynamic
disease: the role of impedance cardiography in diagnostic, prognostic, and
therapeutic decision making. Am J Hypertens. 2005;18:26S– 43S.
14. Van De Water JM, Mount BE, Chandra KM, Mitchell BP, Woodruff TA,
Dalton ML. TFC (thoracic fluid content): a new parameter for assessment
of changes in chest fluid volume. Am Surg. 2005;71:81– 86.
15. Treister N, Wagner K, Jansen PR. Reproducibility of impedance cardiography parameters in outpatients with clinically stable coronary artery
disease. Am J Hypertens. 2005;18:44S–50S.
16. Van De Water JM, Miller TW. Impedance cardiography: the next vital
sign technology? Chest. 2003;123:2028 –2033.
17. Yung G, Fedullo P, Kinninger K, Johnson W, Channick R. Comparison
of impedance cardiography to direct fick and thermodilution cardiac
85 of 158.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
777
output determination in pulmonary arterial hypertension. Congest Heart
Fail. 2004;10(2 suppl 2):7–10.
Defined Daily Dose (DDD) developed by the World Health Organization
for use in drug utilization studies. WHO Collaborating Centre for Drug
Statistics Methodology. Anatomical Therapeutic Chemical (ATC) Classification Index With Defined Daily Doses (DDDs). Oslo, Norway: WHO
Collaborating Centre; 2000.
Dahlof B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U,
Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm
LH, Nieminen MS, Omvik P, Oparil S, Wedel H; LIFE Study Group.
Cardiovascular morbidity and mortality in the Losartan Intervention for
Endpoint reduction in hypertension study (LIFE): a randomised trial
against atenolol. Lancet. 2002;359:995–1003.
ALLHAT Officers and Coordinators for the ALLHAT Collaborative
Research Group. The Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial. Major outcomes in moderately hypercholesterolemic, hypertensive patients randomized to pravastatin vs usual care:
The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart
Attack Trial (ALLHAT-LLT). JAMA. 2002;288:2998 –3007.
Singh H, Johnson ML. Prescribing patterns of diuretics in multi-drug
antihypertensive regimens. J Clin Hypertens. 2005;7:81– 87.
Ubel PA, Jepson C, Asch DA. Misperceptions about ␤-blockers and
diuretics: a national survey of primary care physicians. J Gen Intern Med.
2003;18:1060 –1061.
Connors AF, Dawson NV, Shaw PK, Montenegro HD, Nara AR, Martin
L. Hemodynamic status in critically ill patients with and without acute
heart disease. Chest. 1990;98:1200 –1206.
Fagard RH, Pardaens K, Staessen JA, Thijs L. Prognostic value of
invasive hemodynamic measurements at rest and during exercise in
hypertensive men. Hypertension. 1996;28:31–36.
Messerli FH, Garavaglia GE, Schmieder RE, Sundgaard-Riise K, Nunez
BD, Amodeo C. Disparate cardiovascular findings in men and women
with essential hypertension. Ann Intern Med. 1987;107:158 –161.
Hinderliter A, Blumenthal J, Waugh R, Chilukuri M, Sherwood A. Ethnic
differences in left ventricular structure: Relations to hemodynamics and
diurnal blood pressure variation. Am J Hypertens. 2004;17:43– 49.
Alfie J, Galarza C, Waisman G. Noninvasive hemodynamic assessment of
the effect of mean arterial pressure on the amplitude of pulse pressure.
Am J Hypertens. 2005;18:60S– 64S.
Blood Pressure Lowering Treatment Trialists Collaboration. Effects of
different blood-pressure-lowering regimens on major cardiovascular
events: results of prospectively-designed overviews of randomized trials.
Lancet. 2003;362:1527–1545.
Williams B. Recent hypertension trials: implications and controversies.
J Am Coll Cardiol. 2005;45:813– 827.
Abdelhammed AI, Smith RD, Levy P, Smits GJ, Ferrario CM. Noninvasive hemodynamic profiles in hypertensive subjects. Am J Hypertens.
2005;18:51S–59S.
Bhalla V, Isakson S, Bhalla MA, Lin JP, Clopton P, Gardetto N, Maisel
AS. Diagnostic ability of B-type natriuretic peptide and impedance cardiography: testing to identify left ventricular dysfunction in hypertensive
patients. Am J Hypertens. 2005;18:73S– 81S.
Mitchell A, Buhrmann S, Saez AO, Rushentsova U, Schafers RF, Philipp
T, Nurnberger J. Clonidine lowers blood pressure by reducing vascular
resistance and cardiac output in young, healthy males. Cardiovasc Drugs
Ther. 2005;19:49 –55.
Breithaupt-Grogler K, Gerhardt G, Lehmann G, Notter T, Belz GG.
Blood pressure and aortic elastic properties–verapamil SR/trandolapril
compared to a metoprolol/hydrochlorothiazide combination therapy. Int
J Clin Pharmacol Ther. 1998;36:425– 431.
Ashida T, Nishioeda Y, Kimura G, Kojima S, Kawamura M, Imanishi M,
Abe H, Kawano Y, Yoshimi H, Yoshida K, Kuramochi M, Omae T.
Effects of salt, prostaglandin, and captopril on vascular responsiveness in
essential hypertension. Amer J Hypertens. 1989;2:640 – 642.
Sharman DL, Gomes CP, Rutherford JP. Improvement in blood pressure
control with impedance cardiograph-guided pharmacologic decision
making. Congest Heart Fail. 2004;10:54 –58.
Sramek BB, Tichy JA, Hojerova M, Cervenka V. Normohemodynamic
goal-oriented antihypertensive therapy improves the outcome (abstract).
Amer J Hypertens. 1996;9:141A.
Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, Hua
T, Laragh J, McInnes GT, Mitchell L, Plat F, Schork A, Smith B,
Zanchetti A. VALUE trial group. Outcomes in hypertensive patients at
high cardiovascular risk treated with regimens based on valsartan or
amlodipine: the VALUE randomised trial. Lancet. 2004;363:2022–2031.
The European Journal of Heart Failure 8 (2006) 136 – 140
www.elsevier.com/locate/heafai
Non-invasive measurement of cardiac output by whole-body
bio-impedance during dobutamine stress echocardiography: Clinical
implications in patients with left ventricular dysfunction and ischaemia
Marina Leitman a, Edgar Sucher a, Edo Kaluski a, Ruth Wolf a, Eli Peleg a, Yaron Moshkovitz b,
Olga Milo-Cotter a, Zvi Vered a, Gad Cotter c,*
b
a
Cardiology Department, Assaf-Harofeh Medical Center, Israel
Cardiac Surgical Department, Ramat-Marpe-Assuta Medical Center, Petah Tikva, Sackler School of Medicine, Tel-Aviv University, Israel
c
Duke Clinical Research Institute, PO Box 17969, Durham NC, 27715, USA
Received 20 December 2004; accepted 30 June 2005
Available online 29 September 2005
Abstract
Objectives: To compare non-invasive determination of cardiac index (CI) by whole body electrical bioimpedance using the NICaS apparatus
and Doppler echocardiography, and the role of cardiac power index (Cpi) and total peripheral resistance index (TPRi) calculation during
dobutamine stress echocardiography (DSE).
Subjects and methods: We enrolled 60 consecutive patients undergoing DSE. Patients were prospectively divided into 3 groups: Group 1
(n = 20): normal DSE (control). Group 2 (n = 20): EF < 40% without significant ischaemia. Group 3 (n = 20): patients with significant
ischaemia on DSE. Measurements of CI were performed at the end of each stage of DSE by both echocardiographic left ventricular
outflow track flow and the NICaS apparatus, using whole-body bio-impedance. MAP was measured simultaneously and TPRi and Cpi
were calculated.
Results: The correlation between non-invasive CI as determined by NICaS and echocardiography was 0.81, although Echocardiographic
readings of CI were higher during administration of higher doses of dobutamine. Lower EF correlated with lower Cpi, especially stress
induced Cpi. Hence, patients with reduced EF (group 2) had a blunted increase in Cpi during stress. Patients with ischaemia (group 3) had a
blunted increase in Cpi as well as a decrease in Cpi and increase in TPRi during the last stages of DSE.
Conclusion: Measurement of CI by NICaS correlated well with Doppler derived CI. The calculation of Cpi and TPRi changes during
dobutamine stress may provide important clinical information.
D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.
Keywords: Cardiac index; Cardiac power; Bioimpedance
1. Background and aims
Cardiac power index (Cpi) is the product of simultaneously measured cardiac index (CI) and mean arterial
blood pressure (MAP). Cpi increases — cardiac power
reserve during stress [1,2] or dobutamine administration
[3] was shown in previous studies to be an important
* Corresponding author. Duke Clinical Research Institute, 2400 Pratt
Street, Durham, 27715 NC, USA.
E-mail address: [email protected] (G. Cotter).
measure of systolic cardiac contractile reserve, better than
VO 2 and echocardiographic ejection fraction (EF).
Recently, Samejima et al. [4] demonstrated that CO
increase during stress using non-invasive CO determination with bioimpedance was correlated with stress induced
dyspnea.
The use of hemodynamic measures such as increase in
vascular resistance for the detection of ischaemia was
suggested almost a decade ago [5]. However, this research
avenue has not been pursued due to the lack of simple noninvasive devices for CO measurement. Recently Weiss et al.
1388-9842/$ - see front matter D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.ejheart.2005.06.006
86 of 158.
M. Leitman et al. / The European Journal of Heart Failure 8 (2006) 136 – 140
Table 1
Baseline characteristics of patients in the three groups
2. Patients and methods
Normal
DSE
Ischaemia
by DSE
EF < 40%
20
59 T 10
25%
50%
25%
35%
30%
30%
20
59 T 17
55%
75%
45%
55%
35%
10%
20
54 T 20
90%
80%
50%
40%
30%
10%
0.6
0.003
0.12
0.46
0.6
0.7
0.3
Medical therapy
Aspirin
Beta-blockers
ACEi/AT II antagonists
Diuretics
Statins
75%
45%
30%
10%
45%
95%
75%
55%
45%
80%
85%
80%
90%
90%
75%
0.17
0.03
0.04
0.03
0.1
Echocardiographic findings
Left ventricular EF (%)
LVH
57 T 6
40%
50 T 11
50%
28 T 7
45%
<0.001
0.78
N
Age (years)
Prior MI
Hypertension
Diabetes mellitus
Hyperlipidaemia
Prior smoking
Family history of ischaemic
heart disease
137
P value
EF: Ejection fraction, MI: Myocardial infarction, ACEi: Angiotensin
converting enzyme inhibitors, ATII antagonists — angiotensin II antagonists, LVH: Left ventricular hypertrophy.
We enrolled 60 consecutive patients undergoing standard DSE using incremental dobutamine infusion from 10
to 40 Ag/min and atropine up to 1 mg as required to reach
the pre-determined target heart rate. Patients were recruited
in our outpatient clinic during a once weekly session. All
consecutive patients attending the clinic during that day for
the purpose of DSE were considered for the study. Patients
were divided into 3 groups: Group 1: Control. Normal
DSE, including baseline EF >40% and no significant
ischaemia, Group 2: LV systolic dysfunction as determined
by baseline echocardiographic EF < 40% without significant ischaemia and Group 3: Significant ischaemia as
determined by improvement of contractility during lowdose dobutamine infusion followed by decreased contractility during high dose dobutamine infusion in at least one
non-infarcted myocardial segment.
Exclusion criteria were inability to achieve good
echocardiographic visualization, significant hypotensive or
hypertensive reactions or tachy or bradyarrhythmias during
dobutamine infusion and inability to reach the pre-determined heart rate.
2.1. Study protocol
87 of 158.
DSE was performed according to a standard protocol by
2 DSE teams. NICaS CI was measured by one NICaS
operator. CI was determined by both echocardiographic left
ventricular outflow track (LVOT) diameter and flow
velocity as well as the NICaS apparatus. Operators
measuring CI by one method were blinded to the result
of the other method throughout the examination. Thereafter, based on NICaS determined CI, we calculated Cpi
and total peripheral vascular resistance (TPRi) for each of
the above mentioned time-points. NICaS and Doppler
determined CI were not calculated during dobutamine
administration in 3 patients (5%) in whom it was judged by
10
9
R=0.81
8
7
NICas CI
[6] demonstrated that in patients with significant ischaemia
during stress, CO increase by bio-impedance is lower than
in patients without ischaemia.
The NICaS apparatus uses whole body bio-impedance and
the Tsoglin –Frinerman formula for non-invasive determination of CI [7]. In short, a small electrical current is
transferred from the left wrist to the right foot, and the
impedance to its transit is detected (termed whole-body bioimpedance). The instantaneous change in bio-impedance has
previously been shown to be related to the pulsatile changes
in the volume of the great arteries. The Tsoglin – Frinerman
formula uses this change in bio-impedance (DR) as well as
population based constants correcting for age, sex, weight
and body composition (electrolytes, haematocrit and changes
in baseline bio-impedance) to calculate the stroke volume
(SV). Thereafter by electrocardiographically measuring the
pulse rate it calculates cardiac output and CI. In a few recently
published studies [7 –9], NICaS measurements of CI in
patients with various cardiac conditions showed good
reproducibility and correlated well with thermodilution
(R = 0.8 – 0.9), with no bias and precision of approximately
0.6 L/min/M2.
The aim of the present study was two fold: first, to
compare CI measurements by NICaS and Doppler echocardiography over a wide range of values during dobutamine
stimulation and, secondly, to determine whether the noninvasive continuous measurement of CI and MAP and
calculation of Cpi and TPRi changes during dobutamine
stress could be used for diagnosis of significant left
ventricular (LV) dysfunction or myocardial ischaemia as
determined by dobutamine stress echocardiography (DSE).
6
5
4
3
2
1
0
0
2
4
6
8
10
Echo CI
Fig. 1. Correlation between Doppler echocardiography CI and NICaS CI
measurements. Horizontal axis — CI measured by echocardiography, L/
min/M2, Vertical axis: CI measured by NICaS, L/min/M2.
Difference in CI (Echo Doppler, NICaS)
138
3
2
1
0
-1
-2
-3
-4
-5
0
1
2
3
4
5
6
7
Mean CI (Echo Doppler; NICaS)
8
9
Fig. 2. Difference of CI measurements by Doppler echocardiography and NICaS vs. their mean (Bland and Altman analysis).
the operator that LVOT obstruction occurred due to
dobutamine administration.
2.2. Study end-points
(1) The correlation between NICaS and Doppler echocardiography derived CI and (2) the absolute and relative
changes in NICaS determined Cpi and TPRi during dobutamine stress in the 3 groups.
2.3. Statistical methods
All data is reported based on pre-determined group
allocations. To compare the baseline characteristics of the
groups we used the Student’s t-test to compare continuous
variables and the chi-square test to compare categorical
variables. Since the cardiac index and cardiac power
measurements were not normally distributed we used the
Spearman-rank test for comparison of NICaS and echocardiographically determined CI and for comparison of
resting EF and Cpi during the different stages of DSE.
Since both NICaS and Doppler echocardiography are not
regarded as gold standard for CI determination, when
comparing the two methods we used the Bland and
Altman [10] recommendations and for each dobutamine
stage, as well as for the whole cohort, we determined bias
(mean difference between the 2 methods) and the limits of
agreement (precision) calculated as 2 SD of the bias.
Analysis of variance with repeated measurements (time *
group) was used to compare changes in Cpi and TPRi
over time in the different groups. All P values < 0.05
were considered significant.
3. Results
Sixty consecutive patients where enrolled in the 3
prospectively defined groups. The baseline characteristics
of the three groups are presented in Table 1. As expected,
patients differed with respect to baseline echocardiographic EF, age and severity of background diseases.
The correlation between CI as determined by echocardiographic LVOT area and mean flow velocity and the NICaS
apparatus was R = 0.81 (Fig. 1). The Bland – Altman
distribution of CI measurements is depicted in Fig. 2. The
correlation was better for CI measurements at baseline and
during the infusion of dobutamine at doses of up to 20 Ag/
min than for CI determinations during administration of
dobutamine at a rate of 30 and 40 Ag/min (Table 2, Fig. 2),
due to increasing bias (i.e., CI measurements by Doppler
echocardiography tending to be higher) and lower precision.
Therefore, the CI increase in the different stages of DSE was
similar by both techniques (Fig. 3). Again, a slight tendency
was observed for higher increase in the Doppler echocardiography determined CI during the last phase of dobutamine infusion.
Table 2
CI measurements obtained by Doppler echocardiography and NICaS during the different DSE stages
Dobutamine infusion
rate
Mean NICaS-CI
Mean echo-Doppler CI
Spearman Rank
Correlation
Baseline
10 Ag/min
20 Ag/min
30 Ag/min
40 Ag/min
All measurements
2.9 T 0.6
3.4 T 0.9
4.1 T1.1
4.7 T 1.4
4.7 T 1.2
3.9 T 1.3
2.8 T 0.6
3.4 T 1.1
4.2 T 1.5
4.9 T 1.5
5.2 T 1.4
4.1 T1.5
0.77
0.81
0.87
0.81
0.62
0.81
88 of 158.
Bias
0.06
0.04
0.16
0.24
0.49
0.17
Precision
0.39
0.71
0.8
1.16
1.18
0.9
5.5
1.4
5
1.2
Cpi (watt/M2)
CI (Liter/min/M2)
4.5
4
3.5
3
2.5
0
10
20
30
Dobutamine dose (mcg/min)
NICaS CI
40
Echo-Doppler CI
Fig. 3. Trend changes in CI during dobutamine infusion: Doppler
echocardiography vs. NICaS CI.
We have observed a significant correlation between EF and
Cpi based both on echocardiographically determined CI and
on NICAS-CI (Table 3). Interestingly, and consistent with
previous studies [11], this correlation was better for stress
Cpi than for rest Cpi.
During dobutamine infusion, NICaS determined Cpi
increase was significantly different in the three groups
(Fig. 4). Cpi increase was smallest in the group of patients
with LV dysfunction followed by the group of patients
with significant ischaemia and highest in patients with
normal DSE ( p = 0.002 comparing LV dysfunction with
normal DSE, p = 0.03 comparing patients with ischaemia
with patients with normal DSE and p = 0.02 comparing
patients with ischaemia to patients with LV dysfunction).
Baseline TPRi was significantly higher at baseline in the
LV systolic dysfunction group as compared to patients with
normal DSE (3120 T 1020 vs. 2450 T 940 dynes s M2,
p = 0.04) however, during dobutamine stress it decreased
steeply in all 3 groups, to a similar degree.
A significant difference was observed in Cpi and TPRi
changes during the last phase of dobutamine infusion in
patients in the significant ischaemia group. As compared to
patients in the control as well as the systolic LV dysfunction
group, patients who were found by DSE to have significant
ischaemia had during the last phase of DES a significant
decrease in Cpi ( 0.16 T 0.15 vs. + 0.1 T 0.15 W/M2,
Table 3
Correlations between resting EF and Cpi by NICaS and echocardiography
during the different DSE stages
Dobutamine
infusion rate
Spearman Rank
Correlation between
resting EF and
echo-Doppler Cpi at
different DSE stages
Spearman Rank Correlation
between resting EF and
NICaS-Cpi at different
DSE stages
Baseline
10 Ag/min
20 Ag/min
30 Ag/min
40 Ag/min
All measurements
0.44
0.56
0.61
0.58
0.59
0.57
0.44
0.39
0.54
0.53
0.45
0.52
89 of 158.
139
1
0.8
0.6
0.4
0.2
0
0
10
20
30
Dobutamine dose(mcg/min.)
Normal DSE
Ischaemia
40
EF<40%
Fig. 4. Changes in Cpi during dobutamine infusion in the 3 groups.
p = 0.0002) and increase in TPRi. (+ 279 T 636 vs. 59 T
169 dynes s M2, P = 0.022).
No significant adverse events were recorded during the
dobutamine stress test.
4. Discussion
The results of the present study demonstrate that CI
determination by whole-body bio-impedance using the
NICaS device is correlated well with CI determination by
Doppler echocardiography. However, during infusion of
higher doses of dobutamine ( 30 Ag/min), the correlation
became less accurate, mainly due to significant bias, i.e.
Doppler echocardiography CI measurements tended to be
significantly higher than NICaS readings.
In the overall cohort we found a correlation between
severity of LV dysfunction by resting EF and Cpi at rest and
especially Cpi during the dobutamine stress; i.e., the lower
the EF the lower the rest and peak DSE Cpi. Hence, in the
group of patients with reduced baseline EF (group 2), Cpi
increase during exercise was significantly blunted. This
finding is substantiated by the results of previous studies
showing that lower Cpi increase during stress (lower cardiac
power reserve) is correlated with poor outcome — although
this correlation was superior to the correlation of EF and
outcome. In the present study, the small number of patients
enrolled did not allow for outcome analysis, however, it is
possible that accurate non-invasive CI determination and
calculation of Cpi reserve by whole body bio-impedance
during stress may become a useful predictor of outcome in
patients with reduced left ventricular function.
The results of the present cohort, in concordance with
previous studies [5,6] show that during significant ischaemia, cardiac power tends to decrease and vascular resistance
increases. Again, such non invasive calculation may enable
an additional important indication for significant ischaemia
in addition to conventional signs on DSE.
Importantly, all haemodynamic data was obtained in the
present study using a simple non-invasive device. Although
140
the size of the present cohort does not allow for far-reaching
conclusions, if these results are reproduced by additional,
larger studies, cardiac power and vascular resistance
changes could be used for non-invasive detection of left
ventricular dysfunction and ischaemia during simple stress
tests such as electrocardiographic exercise stress test or
mental stress test. Moreover, serial changes in these
measures may be useful for improving detection of
ischaemia in patients at home using telemedicine.
[2]
[3]
[4]
4.1. Study limitations
[5]
The study included a small, select group of patients
referred for DSE due to various symptoms. Hence, the
results require further confirmation in a larger study
including a non-selected group of outpatients.
[6]
5. Conclusion
[7]
The results of the present study suggest that a simple
stress test using dobutamine infusion and non-invasive
determination of CI by the NICaSi 2001 apparatus and
calculation of Cpi and TPRi can be used for easy out-patient
screening of patients for systolic LV dysfunction and
myocardial ischaemia.
References
[1] Williams SG, Cooke GA, Wright DJ, Parsons WJ, Riley RL, Marshall
P, et al. Peak exercise cardiac power output. A direct indicator of
90 of 158.
[8]
[9]
[10]
[11]
cardiac function strongly predictive of prognosis in chronic heart
failure. Eur Heart J 2001;22:1496 – 503.
Cohen-Solal A, Tabet JY, Logeart D, Bourgoin P, Tokmakova M,
Dahan M. A non-invasively determined surrogate of cardiac power
(Fcirculatory power_) at peak exercise is a powerful prognostic factor
in chronic heart failure. Eur Heart J 2002;23:806 – 14.
Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with severe
heart failure. J Am Coll Cardiol 1997;29:422 – 8.
Samejima H, Omiya K, Uno M, Inoue K, Tamura M, Itoh K, et al.
Relationship between impaired chronotropic response, cardiac output
during exercise, and exercise tolerance in patients with chronic heart
failure. Jpn Heart J 2003;44:515 – 25.
Mohr R, Rath S, Meir O, Smolinsky A, Har-Zahav Y, Neufeld HN,
et al. Changes in systemic vascular resistance detected by the arterial
resistometer: preliminary report of a new method tested during
percutaneous transluminal coronary angioplasty. Circulation 1986;
74:780 – 5.
Weiss SJ, Ernst AA, Godorov G, Diercks DB, Jergenson J, Kirk JD.
Bioimpedance-derived differences in cardiac physiology during
exercise stress testing in low-risk chest pain patients. South Med J
2003;96:1121 – 7.
Cohen AJ, Arnaudov D, Zabeeda D, Schultheis L, Lashinger J,
coronary artery bypass grafting. Eur J Cardiothorac Surg 1998;14:
64 – 9.
Cotter G, Moshkovitz Y, Kaluski E, Cohen AJ, Miller M, Goor DA,
et al. Accurate, non-invasive continuous monitoring of cardiac output
by whole body electrical bio-impedance. Chest 2004;125:1431 – 40.
Torre-Amione G, Cotter G, Kaluski E, Salah A, Milo O, Richter C,
et al. Non-invasive measurement of cardiac output by whole-body
electrical bioimpedance in patients treated for acute heart failure: a
prospective, double blind comparison with thermodilution. J Am
Coll Cardiol 2004;43:109 [abstract].
Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986;1:307 – 10.
Cotter G, Williams SG, Vered Z, Tan LB. The role of cardiac power in
heart failure. Curr Opin Cardiol 2003;18:215 – 22.
CLINICAL INVESTIGATIONS
Impact of Impedance Cardiography on
Diagnosis and Therapy of Emergent
Dyspnea: The ED-IMPACT Trial
W. Frank Peacock, MD, Richard L. Summers, MD, Jody Vogel, MD, Charles E. Emerman, MD
Abstract
Background: Dyspnea is one of the most common emergency department (ED) symptoms, but early diagnosis and treatment are challenging because of multiple potential causes. Impedance cardiography (ICG) is
a noninvasive method to measure hemodynamics that may assist in early ED decision making.
Objectives: To determine the rate of change in working diagnosis and initial treatment plan by adding ICG
data during the course of ED clinical evaluation of elder patients presenting with dyspnea.
Methods: The authors studied a convenience sample of dyspneic patients 65 years and older who were
presenting to the EDs of two urban academic centers. The attending emergency physician was initially
blinded to the ICG data, which was collected by research staff not involved in patient care. At initial ED
presentation, after history and physical but before central lab or radiograph data were returned, the attending ED physician completed a case report form documenting diagnosis and treatment plan. The physician then was shown the ICG data and the same information was again recorded. Pre- and post-ICG
differences were analyzed.
Results: Eighty-nine patients were enrolled, with a mean age of 74.8 7.0 years; 52 (58%) were African
American, 42 (47%) were male. Congestive heart failure and chronic obstructive pulmonary disease
were the most common final diagnoses, occurring in 43 (48%), and 20 (22%), respectively. ICG data
changed the working diagnosis in 12 (13%; 95% CI = 7% to 22%) and medications administered in 35
(39%; 95% CI = 29% to 50%).
Conclusions: Impedance cardiography data result in significant changes in ED physician diagnosis and
therapeutic plan during the evaluation of dyspneic patients 65 years and older.
ACADEMIC EMERGENCY MEDICINE 2006; 13:365–371 ª 2006 by the Society for Academic Emergency
Medicine
Keywords: dyspnea, hemodynamics, impedance cardiography, bioimpedance, cardiac output, systemic
vascular resistance, noninvasive
yspnea is one of the most common emergency
department (ED) symptoms in older patients.1
Various conditions, including heart failure,
chronic obstructive pulmonary disease, pneumonia, pulmonary embolus, and acute coronary syndromes, may
occur alone or in combination in a given patient, adding
D
uncertainty to diagnosis and treatment. In patients with
both cardiac and pulmonary disease, the initial assessment and therapy in the ED are challenging.
Because cardiovascular disease, specifically decompensated heart failure (HF), is a relatively common cause
of dyspnea in elders, an assessment of hemodynamics,
From the Department of Emergency Medicine, Cleveland Clinic
(WFP), Cleveland, OH; Department of Emergency Medicine,
University of Mississippi (RLS), Jackson, MS; Wayne State
University (JV), Detroit, MI; and Department of Emergency
Medicine, Case Western Reserve University (CEE), Cleveland, OH.
Received September 6, 2005; revision received November 7,
2005; accepted November 15, 2005.
Supported by GE Medical Systems (Milwaukee, WI), which
provided devices and disposables for this study, and by CardioDynamics (San Diego, CA), which provided a study grant for
support of research assistants. In addition, CardioDynamics
participated in the creation of the educational process for the participating physicians before study commencement. The sponsor
had no role in data collection or statistical analyses, and the manuscript is the sole responsibility of the authors. W.F.P. and R.L.S.
received honoraria in 2003 for speaking for CardioDynamics.
Presented as a moderated poster at the American College of
Emergency Physicians Research Forum, October 2003.
Address for correspondence and reprints: W. Frank Peacock,
MD, The Cleveland Clinic Foundation, Department of Emergency Medicine, Desk E-19, 9500 Euclid Avenue, Cleveland,
OH 44195. Fax: 216-445-4552; e-mail: [email protected].
ª 2006 by the Society for Academic Emergency Medicine
doi: 10.1197/j.aem.2005.11.078
91 of 158.
ISSN 1069-6563
PII ISSN 1069-6563583
365
366
Peacock et al.
including cardiac output, systemic vascular resistance,
and fluid status, may provide important information
and aid decision making beyond what is possible from
history and physical examination alone. Unfortunately,
hemodynamic parameters cannot be accurately determined by patient history or physical examination.2–5 Until
recently, hemodynamic data could only be obtained by
pulmonary artery catheterization. Because this invasive
procedure is not practical in the ED, physicians typically
are left to make diagnosis and treatment decisions without reliable information about a patient’s hemodynamic
status.
Noninvasive hemodynamic monitoring by impedance
cardiography (ICG) has been used in more than four
million patients. Cardiac output (CO) by ICG has been
shown to correlate well with CO obtained by invasive
methods in hospitalized patient populations with correlation coefficients for CO by ICG and thermodilution ranging from 0.76 to 0.89.6–10 ICG also has been used as an
alternative to invasive monitoring in the critical care
setting.11 In the ED setting, ICG has been studied for the
differential diagnosis of dyspnea12–14 and the identification
of pulmonary edema15,16 and provides prognostic information about hospitalization costs and length of stay.17
ICG results are available within a few minutes, allowing
more rapid patient evaluation than that afforded by
radiographic or laboratory studies.
Given the high rate of morbidity, mortality, and hospital readmissions for patients with dyspnea and acute decompensated HF, there is an urgent need to examine
technologies that could lead to improvements in care in
the ED. The present study examines an aspect of therapeutic efficacy18 as it relates to ICG and the acutely dyspneic emergency patient, and not simply the performance
of ICG as a testing modality. Put into context, previous
studies of commonly used ED tools, such as pulse oximetry19,20 and B-type natriuretic peptide (BNP) testing,21
suggest that a 5% to 11% rate of change in diagnosis,
or 10% rate of change in therapy, is clinically relevant.
The effect of ICG-derived hemodynamics on diagnosis
and treatment of dyspnea in the ED is not yet known.
The purpose of this study was to determine the rate of
change in diagnosis and therapy resulting from the availability of ICG data during the initial evaluation of older
ED patients presenting with dyspnea.
IMPEDANCE CARDIOGRAPHY IN EMERGENT DYSPNEA
by the ED physician. Because ICG is not currently recommended (per U.S. Food and Drug Administration guidelines) for the diagnosis of acute coronary syndromes,
including acute myocardial infarction (MI), patients were
excluded if electrocardiogram (ECG) or serum markers
were positive for acute MI. Additional exclusions included the following: if ICG monitoring was not possible
because of inability to place electrodes, if the patient’s
weight was greater than 341 pounds, or if the patient
had an activated minute ventilation pacemaker (which
uses an impedance signal). Also, although severe aortic
regurgitation that could give a falsely elevated ICG CO
is rare and generally evident on ED evaluation, we excluded those with aortic regurgitation by past history,
and those with the typical diastolic murmur. Last, because
the treatment and disposition actions for patients needing
immediate intubation and mechanical ventilation are generally well defined from the emergency physician point
of view, and because it was our intent to study the
diagnostically most challenging patients, those requiring
urgent intubation and mechanical ventilation upon
presentation to the ED were excluded.
Study Protocol
Project coordinators screened candidates and an independent research nurse, not involved in the diagnosis
or treatment of the patient, obtained hemodynamic
data. Hemodynamic data were collected by using the
BioZ ICG monitor (CardioDynamics, San Diego, CA), as
has been described elsewhere.22 ICG data are obtained
by the following technique: four dual sensors (each sensor consisting of two electrodes) are placed on the
patient, as shown in Figure 1, on opposite sides of the
neck at a level between the ears and shoulders and on either side of the chest in the mid-axillary line at the level of
the xiphoid process. The outer electrodes in each sensor
transmit a low-amplitude, high-frequency current (2.5
mA, 70 kHZ), and the inner electrodes detect thoracic
voltage changes. Changes in voltage are used to calculate
changes in impedance (Z). Baseline, static impedance is
indicative of chest fluid volume, and dynamic impedance
METHODS
Study Design
This was a prospective study of dyspneic patients that
was designed to determine the frequency of change in
the ED physician’s initial diagnosis and therapeutic plan
after physician access to noninvasive ICG hemodynamic
data. The study was approved by each hospital’s institutional review board. All patients gave informed consent
before enrollment in the study.
Study Setting and Population
The setting was two large urban academic EDs with
experience in using noninvasive hemodynamic monitoring. A convenience sample was obtained from patients
age 65 years or older who were presenting with a chief
complaint of dyspnea or symptoms of HF, as determined
92 of 158.
Figure 1. Front view of impedance cardiography method.
ACAD EMERG MED
April 2006, Vol. 13, No. 4
www.aemj.org
is affected by aortic blood volume and velocity. Beatto-beat changes in thoracic impedance are processed to
calculate blood flow per heartbeat (stroke volume) and
per minute (cardiac output). By using standard equations,
other hemodynamic parameters, such as systemic vascular resistance, are calculated. The reciprocal of baseline
thoracic impedance can provide an index of intrathoracic
fluid and is termed thoracic fluid content (TFC). TFC has
been used to identify intravascular and extravascular
fluid changes23,24 and to titrate diuretic therapy.25
Before study initiation, participating physicians received instruction regarding the interpretation of the
hemodynamic values obtained by the ICG device.
Attending physicians, all of whom were board-certified
or board-eligible in emergency medicine, were given a
description of ICG technology and hemodynamic parameters provided on the ICG report, including definitions
and normal values for cardiac index (CI), resistance, thoracic fluid content, and measures reflecting left ventricular performance. This was performed at departmental
grand rounds and at the monthly attending-physician
staff meeting. Additional information was disseminated
in hardcopy by mailing and was duplicated by e-mail.
The pathophysiology of HF and hemodynamic findings
most suggestive of dyspnea caused by decompensated
HF (reduced CI, elevated systemic vascular resistance,
and increased TFC) were described. The expected effects
of various medications on hemodynamic parameters
were discussed, including use of diuretics, vasodilators,
and drugs affecting contractility. Additionally, physicians
were provided personal reference cards for use at their
discretion that detailed normative values for all ICG
data. Copies of the data card were also kept fixed to the
ICG device. These data were also shown at the time of
ICG unblinding. For any given parameter, ICG data are
presented as a bar indicating the normal human range.
The average result and the currently measured data point
then are indicated on this bar, such that variations from
normal are readily apparent.
All staff involved with patient care were blinded to the
ICG data until after the initial history and physical examination by the attending physician. After the initial history and physical exam, but before initiation of therapy
(other than supplemental oxygen), and before obtaining
any central laboratory or radiographic data, the attending physician completed a case report form indicating
his or her working diagnoses and short-term therapeutic
plans. The physician was then immediately shown the
ICG hemodynamic data and was asked to complete the
case report form again, this time with consideration of
the ICG data. All patient care then proceeded according
to usual ED routine. Blood tests, including electrolytes,
blood urea nitrogen (BUN), serum creatinine (Cr), and
BNP levels, were obtained in the majority of cases.
Although these data were not mandated as part of the
protocol, they were used in most cases to determine final
ED diagnosis.
Measures
The two primary endpoints were 1) the rates of change
in working diagnosis and 2) medical therapy after the
addition of ICG data to the physician’s initial clinical
assessment and therapeutic plan. In the cases in which
93 of 158.
367
the diagnosis changed on the basis of ICG data, a comparison to the final diagnosis was made to determine
whether the pre- or post-ICG diagnosis was more consistent with the final ED diagnosis. The final ED primary diagnosis was defined as the principal diagnosis at the end
of the ED visit after all diagnostic testing was completed
and reviewed by the ED physician responsible for disposition. A change in therapy was defined as the addition or
subtraction of a drug or procedure. Changing the dose of
a previously ordered drug was not considered a therapeutic change. Adverse events were defined as cardiac
arrest, intubation for respiratory failure, urgent cardioversion, or blood transfusion.
Data Analysis
The size of the study was prospectively determined on the
basis of the number needed to detect a 5% rate of change
in diagnosis or therapy. Given an alpha of 0.05 and a beta
of 0.20, a sample size of 100 was needed to detect a statistically significant change. Data were analyzed by an independent statistician using SAS Software (Cary, NC).
Demographic data are reported descriptively. Continuous variables are reported as mean standard deviation
(SD). Rates of change were calculated by dividing the
number of patients in whom diagnosis or therapeutic
plan changed by the total number of patients and were
reported as percentages. An analysis of variance was
performed to assess for differences among vital signs
and ICG parameters in the final diagnosis categories.
RESULTS
Eighty-nine patients, cared for by 31 ED staff physicians,
were enrolled from December 2001 through July 2003
and are included in the analysis. No adverse event,
defined as cardiac arrest, intubation, cardioversion, or
blood transfusion, occurred during the course of the
ED observation during this study.
The patient characteristics and vital signs are summarized in Table 1. The mean (SD) age of the subjects
was 74.8 (7.0) years. Fifty-eight percent of the patients
were African American, and 61% had a history of HF,
Table 1
Patient Characteristics and Vital Signs
Patient Characteristic
Value
Male, n (%)
Ethnic background, n (%)
African American
White
Other
Medical history, n (%)
History of heart failure
History of chronic obstructive
pulmonary disease or asthma
History of heart failure and COPD
Vital signs (mean SD)
Temperature (ºC)
Respiratory rate (minÿ1)
Heart rate (minÿ1)
Systolic blood pressure (mm Hg)
42 (47)
N = 89.
52 (58)
36 (41)
1 (1)
43 (48)
34 (38)
11 (13)
36.5 0.7
22.2 5.1
84.6 18.8
145.6 29.3
368
Peacock et al.
Table 2
Hemodynamic Values at Presentation
Hemodynamic Parameter
Thoracic fluid content
(kOhmÿ1)
Systemic vascular resistance (dyne sec cmÿ5)
Cardiac index (L/min/m2)
Stroke index (mL/m2)
Mean SD
34.2 10.4
1,685 579
2.6 0.6
31.8 10.1
Normal Range
30–50 (male)
21–37 (female)
742–1,378
2.5–4.2
35–65
N = 89.
including 13% with a history of both HF and chronic lung
disease. The prevalence of chronic lung disease, including asthma, was 38%. The average respiratory rate was
22.2 (5.1) minÿ1 with systolic BP and heart rate of
145.5 (29.3) mm Hg and 84.6 (18.8) minÿ1, respectively.
The hemodynamic values for the population as a whole
are listed in Table 2.
Patients could be categorized by final primary diagnosis at the time of ED discharge or hospital admission into
three major groups: 1) HF (n = 43); 2) chronic obstructive
pulmonary disease (COPD; 20); and 3) ‘‘other’’ (26). The
other group included other cardiovascular and lung
conditions not included in the HF or COPD groups: atrial
fibrillation (n = 4), bronchitis (4), hypertension (2), pneumonia (2), pulmonary hypertension (2), anemia (1), influenza (1), lung cancer (1), palpitations (1), upper
respiratory infection (1), atypical chest pain (1), hypoxia
(1), intra-abdominal abscess (1), non-cardiac shortness
of breath (1), pulmonary fibrosis (1), vertigo (1), and dehydration (1).
Chest radiographs and ECG results were recorded by
the ED physician in 85 patients (96%). The various ECG
and radiographic findings are summarized in Table 3.
ECG findings were described as normal or nonspecific
in the vast majority (82%), and the chest radiograph
was normal or nondiagnostic in nearly half, with only
16% showing either HF or upper zone redistribution
consistent with pulmonary venous congestion.
Table 3
Electrocardiographic and Chest Radiographic Findings
Parameter
Electrocardiographic findings
Normal, nonspecific, or ‘‘nondiagnostic ECG’’
Left ventricular hypertrophy
Atrial fibrillation
Prior myocardial infarction
Left bundle branch block
Relevant chest radiography findings
Normal or no acute disease
Increased cardiothoracic ratio (>0.5)
or cardiomegaly
Pleural effusion
Pulmonary edema or ‘‘heart failure’’
Upper zone redistribution
Pulmonary infiltrate
Atelectasis
Hyperinflation
Diagnosis Changes
A summary of the rates of diagnosis and therapy change
that resulted from ICG data are presented in Figure 2.
ICG data changed the working diagnosis in 12 (13%;
95% CI = 7% to 22%). When diagnoses were categorized
as either cardiac or noncardiac, the post-ICG diagnosis
was the same as the final diagnosis in 8 of 12 patients
in whom ICG resulted in a change (67%, 95% CI = 35%
to 90%). Of the four patients in whom a change in diagnosis after ICG did not match the final ED diagnosis,
one who was ultimately diagnosed with a cardiac cause
of dyspnea had normal hemodynamic parameters, suggesting a pulmonary cause. In another patient, altered
hemodynamic parameters suggested cardiac dyspnea
that was later attributed to an exacerbation of COPD.
One patient with lung cancer had hemodynamic findings
consistent with diastolic HF. Finally, one patient who
initially was thought to have pulmonary dyspnea had
altered hemodynamic findings that were believed to be
nondiagnostic by the evaluating physician; that patient
was ultimately treated for fluid overload and discharged
home.
Results Grouped by Final Diagnosis
A summary of the patient vital signs and hemodynamic
characteristics grouped by final ED diagnosis is listed
in Table 4. No diagnosis group had vital sign data that
were significantly different from those of any other
group (p = 0.1332). Of the hemodynamic parameters,
cardiac index, systemic vascular resistance index, and
thoracic fluid content had one diagnosis group that
differed significantly from the other two (p < 0.02). HF
patients had greater amounts of lung water, as reflected
by a mean TFC (38.5 12.3 kOhmÿ1) that was significantly higher than that of the other two diagnosis groups
(30.0 6.17 and 30.4 5.6 for the COPD and other
groups, respectively). Patients with COPD had higher
CI (3.08 0.57 vs. 2.39 0.56 and 2.48 0.65) and lower
SVR (1,361 407 vs. 1,772 565 and 1,789 638) than did
patients in the HF or other groups, respectively.
Laboratory measurements, including electrolytes,
BUN, Cr, WBC, Hgb, and BNP were analyzed by
final diagnosis. Of the laboratory measurements, only
BNP, measured in 72 patients, exhibited a statistically
n (%)
70
15
11
8
3
(82)
(18)
(13)
(9)
(4)
42 (49)
13 (15)
8
8
6
5
2
1
(9)
(9)
(7)
(6)
(2)
(1)
For each finding group, n = 85. Because of multiple findings, total for ECG
diagnoses is greater than 85.
94 of 158.
Figure 2. Category rate of change from pre-ICG to post-ICG
(mean values with 95% confidence intervals).
ACAD EMERG MED
April 2006, Vol. 13, No. 4
www.aemj.org
Table 4
Initial Vital Signs and Selected Hemodynamic Parameters by
Final Diagnosis
Final Diagnosis
Parameter
Heart
Failure
(n = 43)
Temperature (ºC)
36.4
Systolic blood
149.5
pressure (mm Hg)
Diastolic blood
80.2
pressure (mm Hg)
84.1
Heart rate (minÿ1)
Respiration rate (minÿ1) 22.1
38.5
(kOhmÿ1)
1,772
Systemic vascular
resistance (dyne
sec cmÿ5)
2.39
Cardiac index
(L/min/m2)
Mean arterial
100.0
pressure (mm Hg)
29.7
Stroke index (mL/m2)
Chronic
Obstructive
Pulmonary
Disease
(n = 20)
75.7 13.0
78.5 21.2
18.0 88.2 19.4
5.7
23.8 3.8
12.3* 30.0 6.17
82.9 20.1
21.0 4.6
30.4 5.6
565 1,361 407* 1,789 638
0.56
3.08 0.57* 2.48 0.65
20.0
98.5 14.2 100.7 21.8
9.9
36.3 9.7
31.6 10.0
Data are mean SD.
* Different from other two categories, p < 0.02.
significant difference (p < 0.0001) among the three diagnosis groups. The HF group had a significantly higher
mean BNP level (940 pg/mL) than did the other diagnosis
groups (137 pg/mL, and 357 pg/mL for COPD and other
groups, respectively).
Treatment Changes
Changes in planned medication orders, occurring after
ICG information was revealed (and without other input
to the treating physician), are shown in Table 5. Thirtyfive patients (39%, 95% CI = 29% to 50%) had a total of
54 changes in the medication plan after initial assessment
and review of ICG data. Use of diuretics most often was
altered based on ICG findings, suggesting fluid overload
or cardiac cause of dyspnea. When looking at medication
changes by category of final diagnosis, there were 17
medication changes in the 43 patients with a final diagnosis of HF, 20 medication changes in the 20 patients with a
Table 5
Therapeutic Changes Post- vs. Pre-ICG Listed by Medication
Class
Medication Class
Diuretics
Nitroglycerin
Bronchodilators
Steroids
Antibiotics
Anticoagulants
Other
n (%) [95% CI]
12
4
11
6
6
5
10
(13) [7%–22%]
(4) [1%–11%]
(12) [8%–24%]
(7) [3%–14%]
(7) [3%–14%]
(6) [2%–13%]
(11) [6%–20%]
There were 54 total medication changes in 35 patients. Because of multiple therapeutic changes, total changes are greater than the number of
patients.
95 of 158.
final diagnosis of COPD, and 17 medication changes in
the 26 patients in the other group. Twenty-three of 54
(43%) medication changes that resulted from the availability of hemodynamic information were changes in
the use of diuretics or bronchodilators.
DISCUSSION
Other
(n = 26)
0.7
36.6 0.8
36.7 0.5
30.5 135.6 19.1 146.8 32.2
18.0
369
In cases of elder dyspneic patients who may require
urgent treatment, the ED physician must assess status,
formulate a working diagnosis, and institute therapy, in
many cases before all information is available. Hemodynamic information, which reflects the contribution of
the cardiovascular system to the current presentation,
may have an important impact on the process of care.
Our results demonstrate that knowledge of ICG data
leads to a change of working primary diagnosis in 13%
of elder patients presenting with dyspnea to the ED.
When changes in diagnosis were made, they were consistent with the final diagnosis at time of ED disposition
in two-thirds of cases. In addition to changes in diagnosis, ED physicians made medication changes on the basis
of ICG-derived hemodynamic information in 39% of
cases. Finally, unlike vital signs, which were similar across
the various diagnostic groups, hemodynamic data varied
based on causes of dyspnea. These findings are consistent
with the hypothesis that hemodynamic information is
relevant and actionable in the ongoing evaluation and
treatment of such patients.
Patients presenting with dyspnea are commonly at risk
for exacerbation of either cardiac or pulmonary disease.
Those with acute HF typically have reduced cardiac
output and elevated vascular resistance. Those with a
pulmonary or other noncardiac cause of their dyspnea
typically have normal cardiac output and hemodynamic
parameters. Because the accuracy and reproducibility
of ICG have been validated in a variety of patient populations and settings, it is not surprising that physicians
used this information to help guide diagnosis and treatment in dyspneic patients. Our finding of different values
of hemodynamic parameters among the diagnostic groups
is consistent with this paradigm. The relatively high rate
of change of diagnosis, when ICG-derived information
was revealed to treating physicians, suggests acceptance
of the technical and diagnostic accuracy efficacies of the
test. Not only does the information result in altered diagnosis, but the noninvasive hemodynamic data provided
by ICG was applied by the physicians to therapeutic decision making, an indication of therapeutic efficacy, as
defined by Pearl.18 Thus, our results support the potential
value of such information and support a practical role for
this technology in the ED assessment of such patients.
Recently, BNP testing has been shown to be a useful
bedside tool to aid in diagnosis of patients presenting
with shortness of breath.26 However, despite the availability of point-of-care laboratory testing, real-time diagnosis and treatment can be delayed. In fact, one large
trial of cardiac markers found that even with point-ofcare testing, the door-to-brain time (the time from ED
arrival until cardiac marker results are available for the
physician to act upon) exceeded one hour.27 ICG data
are available within several minutes. And, unlike the hemodynamic information obtained by a pulmonary artery
370
Peacock et al.
catheter, performance of ICG is noninvasive and can be
readily accomplished in the ED without specialized training and at minimal risk to the patient.
The magnitude of the changes in diagnosis and treatment resulting from ICG-derived hemodynamic data
can be compared with that from other technologies
that are currently the standard of care in most EDs. Historically, changes in therapy on the order of 5% to 11%
appear to define utility of testing in the ED. In one study,
Summers et al.19 reported that the ED physician assessment of patient severity of illness was changed by pulse
oximetry in 3% of cases. Kosowky et al.,21 evaluating
BNP testing in patients older than 40 years of age, found
that BNP data changed the diagnosis in 10%, and treatment in 11%, of cases. These trials suggest that the rate
of change that resulted from ICG use in the present study
would be clinically significant in the ED environment.
Moreover, ICG can be performed concurrently with
existing diagnostic and therapeutic strategies, such that
the information is incremental in the decision-making
process.
The changes in ED decision making from esophageal
Doppler results, a more invasive and less common form
of cardiac output measurement, have been studied elsewhere.28 Those investigators found a change in management decisions in 31% of cases. Our results show a
greater change in therapy alone, perhaps because of
the incremental information provided by systemic vascular resistance and thoracic fluid content parameters.
Although most ED physicians would not subject a patient
to esophageal monitoring to obtain hemodynamic measurements, it is likely that many would consider the
collection of ICG data, which requires little more time or
inconvenience than obtaining an electrocardiogram.
We did not measure the time required to obtain ICG
data; however, in routine use, these data can be obtained
in about 3 to 5 minutes and require 30 to 60 seconds
to interpret. Because ICG provides early and accurate
data, there is a potential for significant clinical impact
from its use. We did not specifically study the financial
effects of ICG in this study or how it might have affected
length of ED stay or hospital admission rate. However,
at a procedural cost for each test of less than $20 and
with the cost of a day in intensive care at more than
$1,000, the provision of ICG would be cost-effective
even if, for example, it reduced hospital length of stay
by only one day for every 50 patients monitored.
LIMITATIONS
We acknowledge several limitations. This study evaluated the effect of ICG on working diagnosis and initial
treatment plan before the results of chest radiograph,
ECG, or BNP level. Thus, it is impossible to gauge the relative importance of the information obtained from ICG
to that obtained by these other tests or to judge the additional contribution of ICG for cases in which the results
of other tests were available before performing ICG. Although blood work and various ancillary testing such as
chest radiography are part of the complete ED evaluation
of such patients, the results are generally not available
within the first few minutes of patient assessment. By design, this study evaluated ICG’s effect on working diag-
96 of 158.
nosis and therapy in a manner that would be consistent
with clinical practice in the ED, where patients presenting with dyspnea might be evaluated with ICG either
before or within minutes of the ED physician’s initial
assessment. Furthermore, as seen in our study, the findings of ECGs and chest radiographs are often normal or
nonspecific and may not provide significant diagnostic
certainty. Because ICG is not part of the diagnostic criteria for acute coronary syndromes, including acute MI, we
did exclude patients with evidence of myocardial necrosis from analysis. Therefore, the role of ICG in providing
possible clues in the evaluation of patients with dyspnea
as a manifestation of MI cannot be assessed by the present study.
Our study was also limited by the use of the final ED
diagnosis as the criterion standard for diagnostic categorization. Although it is possible that this diagnosis
was incorrect or incomplete in some patients, this represents the real-life diagnosis based on current evaluation
strategies during the patient’s ED visit. It is also possible
that a physician had the right diagnosis and treatment
plan before reviewing ICG results and that ICG data
resulted in inappropriate therapies. A larger prospective
outcome-based study will be required to determine the
potential for this to occur.
In our study, ICG data were available and likely contributed to the final ED diagnosis, thereby introducing
possible bias. However, the goal of this study was not
to assess technical accuracy of the technology, which
has been evaluated in previous studies. In contrast, this
study was designed to assess whether physicians would
incorporate early hemodynamic information into the
process of formulating an initial working diagnosis and
treatment plan. In addition, the study design does not allow us to draw conclusions about the sensitivity or specificity of ICG criteria, or to compare diagnostic accuracy
to other measures, such as BNP or chest radiography.
The accuracy of the post-ICG diagnosis based on these
hemodynamic criteria could only be verified by a more
standardized diagnostic approach including cardiac
imaging studies, blinded reviews of subsequent hospital
records with adjudication of discordant diagnoses, and
long-term follow-up, which were not within the scope
of the current study.
CONCLUSIONS
Knowledge of ICG data early in the ED evaluation of patients older than 65 years of age presenting with dyspnea
results in significant changes in diagnosis and treatment
plan. Whether changes in diagnosis, diagnostic certainty,
or therapy from ICG improve outcomes or are cost-effective will require a prospective, randomized clinical trial
with longer periods of clinical follow-up.
The authors thank Gerard Smits, PhD, for his statistical assistance.
References
1. McCraig LF, Burt CW. National hospital ambulatory
medical care survey: 2001 emergency department
summary. Report from Centers for Disease Control
ACAD EMERG MED
April 2006, Vol. 13, No. 4
www.aemj.org
and Prevention, National Center for Health Statistics.
Advance Data Vital Health Stat. 2003; 335:18.
2. Eisenberg PR, Jaffe AS, Schuster DP. Clinical evaluation compared to pulmonary artery catheterization in
the hemodynamic assessment of critically ill patients.
Crit Care Med. 1984; 12:549–53.
3. Speroff T, Connors AF Jr, Dawson NV. Lens model
analysis of hemodynamic status in the critically ill.
Med Decis Making. 1989; 9:243–52.
4. Neath SX, Lazio L, Guss DA. Utility of impedance cardiography to improve physician estimation of hemodynamic parameters in the emergency department.
Congest Heart Fail. 2005; 11:17–20.
5. Van De Water JM, Dalton ML, Parish DC, Vogel RL,
Beatty JC, Adeniyi SO. Cardiopulmonary assessment: is improvement needed? World J Surg. 2005;
29(Suppl 1):S95–8.
6. Sageman WS, Riffenburgh RH, Spiess BD. Equivalence of bioimpedance and thermodilution in measuring cardiac index after cardiac surgery. J
Cardiothorac Vasc Anesth. 2002; 16:8–14.
7. Drazner M, Thompson B, Rosenberg P, Yancy C.
Comparison of impedance cardiography with invasive hemodynamic measurements in patients with
heart failure secondary to ischemic or nonischemic
cardiomyopathy. Am J Cardiol. 2002; 89:993–5.
8. Van De Water JM, Miller TW. Impedance cardiography: the next vital sign technology? Chest. 2003; 123:
2028–33.
9. Yung GL, Fedullo PF, Kinninger K, Johnson FW,
Channick RN. Comparison of impedance cardiography to direct Fick and thermodilution cardiac output
determination in pulmonary arterial hypertension.
Congest Heart Fail. 2004; 10(2 Suppl 2):7–10.
10. Albert NM, Hail MD, Li J, Young JB. Equivalence of
bioimpedance and thermodilution in measuring cardiac output in hospitalized patients with advanced,
decompensated chronic heart failure. Am J Crit
Care. 2004; 3:469–79.
11. Silver M, Cianci P, Brennan S, Longeran-Thomas H,
Ahmad F. Evaluation of impedance cardiography as
an alternative to pulmonary artery catheterization
in critically ill patients. Congest Heart Fail. 2004; 10(2
Suppl 2):17–21.
12. Springfield C, Sebat F, Johnson D, Lengle S, Sebat C.
Utility of impedance cardiography to determine
cardiac vs. noncardiac cause of dyspnea in the
emergency department. Congest Heart Fail. 2004;
10(2 Suppl 2):14–6.
13. Han J, Lindsell C, Tsurov B, Storrow A. The clinical
utility of impedance cardiography in diagnosing congestive heart failure in dyspneic emergency department patients [abstract]. Acad Emerg Med. 2002; 9:
439–40.
14. Barcarse E, Kazanegra R, Chen A, Chiu A, Clopton
P, Maisel A. Combination of B-type natriuretic pep-
97 of 158.
371
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
tide levels and non-invasive hemodynamic parameters in diagnosing congestive heart failure in the
emergency department. Congest Heart Fail. 2004;
10:171–6.
Peacock WF, Albert NM, Kies P, White RD, Emerman
CL. Bioimpedance monitoring: better than chest
x-ray for predicting abnormal pulmonary fluid? Congest Heart Fail. 2000; 6(2):32–5.
Newman RB, Pierre H, Scardo J. Thoracic fluid conductivity in peripartum women with pulmonary
edema. Obstet Gynecol. 1999; 94:48–51.
Milzman D, Morrisey J, Pugh C, Napoli A, Gerace T,
Fernandez E. Occult perfusion deficits in heart failure
patients: identification through noninvasive central
hemodynamic monitoring [abstract]. Crit Care Med.
1999; 27:A88.
Pearl WS. A hierarchical outcomes approach to test
assessment. Ann Emerg Med. 1999; 33:77–84.
Summers R, Anders R, Woodward L, Jenkins A, Galli
R. Effect of routine pulse oximetry measurements on
ED triage classification. J Emerg Med. 1999; 16:5–7.
Mower WR, Sachs C, Nicklin EL, Safa P, Baraff LJ.
Effect of routine emergency department triage pulse
oximetry screening on medical management. Chest.
1995; 108:1297–302.
Kosowsky JM, Weiner C, Morrissey JH. Impact of
B-type natriuretic peptide testing on medical decision-making for older patients with dyspnea. Ann
Emerg Med. 2003; 42:S11.
Summers R, Schoemaker W, Peacock WF, Ander D,
Coleman T. Bench to bedside series: impedance cardiography (ICG). Acad Emerg Med. 2003; 10:669–80.
Luepker R, Michael JR, Warbasse JR. Transthoracic
electrical impedance: quantitative evaluation of a
noninvasive measure of thoracic fluid volume. Am
Heart J. 1973; 85:83–93.
Van de Water JM, Mount BE, Chandra KM, Mitchell
BP, Woodruff TA, Dalton ML. TFC (thoracic fluid
content): a new parameter for assessment of changes
in chest fluid volume. Am Surg. 2005; 71:81–6.
Taler SJ, Textor SC, Augustine JE. Resistant hypertension: comparing hemodynamic management to
specialist care. Hypertension. 2002; 39:982–8.
Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid
measurement of B-type natriuretic peptide in the
emergency diagnosis of heart failure. N Engl J Med.
2002; 347:161–7.
Peacock WF, Roe MT, Chen AY, et al. Vein-to-brain
time: an emergency department quality of care
marker for non-ST-segment elevation acute coronary syndrome [abstract]. Acad Emerg Med. 2004;
11:569.
Urrunaga J, Rivers E, Mullen M, et al. Hemodynamic
assessment of the critically ill: the clinician versus
esophageal Doppler monitoring (EDM) [abstract].
Acad Emerg Med. 2000; 7:587b.
AJH
2005; 18:26S– 43S
Featured Review Article
Hypertension as a Hemodynamic Disease: The
Role of Impedance Cardiography in Diagnostic,
Prognostic, and Therapeutic Decision Making
Hector O. Ventura, Sandra J. Taler, and John E. Strobeck
Hypertension is the most common cardiovascular disease,
affecting approximately 60 million Americans. Despite the
importance of this condition, only the minority of patients
are appropriately identifie and treated to reach recommended blood pressure (BP) goals. Although historically
define as an elevation of BP alone, hypertension is characterized by abnormalities of cardiac output, systemic
vascular resistance, and arterial compliance. These hemodynamic aspects of hypertension have implications for
diagnosis, risk stratification and treatment. Impedance
cardiography (ICG) has emerged as a unique and highly
accurate noninvasive tool that is used to assess hemody-
Key Words: Hypertension, hemodynamics, impedance
cardiography.
hen functioning properly, the cardiovascular
system provides normal blood flo to the various tissues of the body under normal arterial
blood pressure (BP). Historically, BP is the most commonly measured parameter of cardiovascular function. Hypertension–typically define by BP levels of 140/90 mm Hg
and higher—leads to increased rates of coronary artery
disease, heart failure, renal disease, and stroke. Therefore,
BP control is of paramount importance for both individual
and public health considerations.
Blood pressure by itself is an incomplete indicator of
the status of the cardiovascular system. Mean arterial
pressure (MAP) is the product of two hemodynamic components: cardiac output (CO), the flo of blood pumped
by the heart each minute; and systemic vascular resistance
(SVR), the force the left ventricle must overcome to expel
blood into the systemic vasculature, also called total peripheral resistance. Hypertension results from elevations
of CO, SVR, or both. Because “hemodynamics” literally
refers to blood flow–relate parameters of the arterial
system, CO and SVR are fundamental to obtaining greater
insight into the pathophysiology of hypertension, and they
can help to guide diagnostic, prognostic, and therapeutic
management decisions. Thus, the hemodynamic model of
W
hypertension has intrigued scientists and clinicians since
the early part of the last century1 and has been reviewed
extensively by leaders in the field 2–7
The hemodynamic components of BP, CO, and SVR,
and other related parameters such as arterial compliance
provide insight into mechanisms of hypertension8 and
have implications for management of patients with this
condition. Historically, most hemodynamic information
used in research has been obtained using invasive techniques, including arterial cannulation and placement of a
pulmonary artery catheter for the measurement of cardiac
output and determination of SVR. However, invasive procedures are not feasible in the routine care of patients with
hypertension. Echocardiography provided early noninvasive measurement of cardiac output, but it is costly and
highly operator dependent, and it is therefore impractical
for frequent serial measurements in the clinical setting.
Recent advancements in noninvasive hemodynamic monitoring with impedance cardiography (ICG) have been
achieved, elevating its role as a unique and valuable noninvasive tool for the assessment of hemodynamic status in
patients with hypertension.
This review describes the historical use of hemodynamics in hypertension and reveals the growing body of evi-
Received October 29, 2004. Accepted November 4, 2004.
From the Ochsner Clinic Foundation (HOV), New Orleans,Louisiana; Mayo Clinic College of Medicine (SJT), Rochester, Minnesota;
and Cabrini Medical Center (JES), New York, New York.
Address correspondence and reprint requests to Dr. Hector O.
Ventura, Ochsner Clinic Foundation, 1514 Jefferson Highway, Room
3E419, New Orleans, LA 70121; e-mail: [email protected]
0895-7061/05/$30.00
98 doi:10.1016/j.amjhyper.2004.11.002
of 158.
namic parameters. Measurement of the various hemodynamic components using ICG in those with hypertension
allows more complete characterization of the condition, a
greater ability to identify those at highest risk, and allows
more effectively targeted drug management. This article
reviews the importance of hemodynamic factors in hypertension and the evolving role of ICG technology in the
assessment and management of this important cardiovascular condition. Am J Hypertens 2005;18:26S– 43S
© 2005 American Journal of Hypertension, Ltd.
© 2005 by the American Journal of Hypertension, Ltd.
Published by Elsevier Inc.
AJH–February 2005–VOL. 18, NO. 2, Part 2
HYPERTENSION AS A HEMODYNAMIC DISEASE
Table 1. JNC 7 Classification of blood pressure for
adults ⱖ18 years of age
BP classification
Normal
Prehypertension
Stage 1 hypertension
Stage 2 hypertension
Systolic
BP
(mm Hg)
⬍120
120–139
140–159
ⱖ160
Diastolic
BP
(mm Hg)
and
or
or
or
⬍80
80–89
90–99
ⱖ100
JNC ⫽ Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.
dence using noninvasive monitoring of hemodynamics
with ICG. The specifi role of hemodynamics in diagnostic, prognostic, and therapeutic decision making in the
patient with hypertension is reviewed in detail.
Hypertension: Definition and
Clinical Presentation
Hypertension is most commonly define as a systolic BP
(SBP) of ⱖ140 mm Hg or a diastolic BP (DBP) of ⱖ90
mm Hg. In patients at high risk for complications from
elevated BP levels, such as those with diabetes or chronic
renal disease, lower levels of BP (eg, ⬍130/80 mm Hg)
are recommended. Between BP levels of 115/75 mm Hg
and 185/115 mm Hg, each 20 –mm Hg increase in SBP or
10 –mm Hg increase in DBP doubles the risk of a cardiovascular event.9 In recognition of this increase in risk from
levels as low as 115/75 mm Hg, the Seventh Report of the
Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7)9
recently identifie “prehypertension” (define as a BP of
120 to 139 / 80 to 89 mm Hg) as a significan potential
health problem associated with “increased risk for progression to hypertension.” The JNC 7 recommended lifestyle modificatio for the management of prehypertension.
Table 1 lists the stages of hypertension as define in the
JNC 7 classification
Secondary hypertension results from an identifiabl
cause such as renal, adrenal, or vascular pathology. In
contrast, 90% of patients have no specifi identifiabl
cause of their BP elevation10 and are thus diagnosed with
essential hypertension. Most patients with hypertension
are asymptomatic and do not show evidence of acute
pathologic changes; their clinical presentation has been
termed “benign,” although their long-term risk of cardiovascular complications is significantl greater than in normotensive persons without so-called benign hypertension.
Severe elevation of BP, associated with papilledema on
fundoscopic examination, is termed “malignant hypertension.” Similar levels of BP elevation with congestive heart
failure, anginal symptoms, or other evidence for accelerated end-organ injury (but without papilledema) are
termed “hypertensive urgencies” or “hypertensive emer-
99 of 158.
27S
gencies.” The etiologies of hypertension (essential versus
secondary) and the various clinical presentations (benign,
malignant, unspecified and with or without associated
co-morbidities) are reflecte in the World Health Organization’s International Classificatio of Diseases, Ninth
Revision (ICD-9), coding for hypertension (Table 2).
Hypertension: Magnitude of the
Problem
Hypertension affects up to 60 million Americans and as
many as 1 billion persons worldwide; and it is the most
common reason that patients in the United States visit their
physicians.9,11 The incidence of hypertension increases
significantl with advancing age (Fig. 1), such that a
normotensive adult in the United States 55 years of age
still has a 90% lifetime risk of developing hypertension.9
In fact, the most common group with hypertension is
comprised of elderly patients with systolic hypertension.12
Although controlling BP levels reduces the incidence of
stroke and other cardiovascular complications, BP control
in the US is well below stated goals. For an individual
patient, this may be due to the lack of recognition of the
condition, failure to institute effective treatment, or the result
of a suboptimal long-term medical regimen (Table 3). There
remains a substantial need for improvement in the effectiveness of hypertension treatment. As reported in JNC 7, only
34% of adults aged 18 to 74 years with hypertension have
achieved BP control, despite a published goal of 50%. In
the elderly population, BP control is even less successful:
fewer than 20% of treated patients 70 years or more of age
attain BP levels of ⬍140/90 mm Hg.9
Hypertension treatment commonly requires multiple
medications. “Refractory hypertension” has been define
as hypertension that is not controlled on two or more
antihypertensive medications.13 “Resistant hypertension”
has been define by some as BP readings of ⱖ140/90 mm
Hg “despite an optimal two-drug regimen that has had
adequate time to work (at least 1 month since last drug or
dosage adjustment).”14 As define by Gifford, resistant
hypertension is the failure to reach goal BP in patients who
are adhering to full doses of an appropriate three-drug
regimen that includes a diuretic.15 The JNC 7 recommends
a goal BP of ⬍140/90 mm Hg for the general population,
with the tighter goal of ⬍130/80 mm Hg for persons with
chronic renal disease or diabetes mellitus.
Hypertension substantially increases the incidence of
cardiovascular events, especially the risk of stroke. Wilking et al,16 in data from the Framingham study, reported on
the prognostic significanc of systolic hypertension. They
found that for men and women, the relative risk of cardiovascular disease event adjusted for age was approximately 2.5 times greater for persons with isolated systolic
hypertension compared with those with BP levels
⬍140/95 mm Hg. Lower BP levels are thus associated
with improved prognosis and decreased incidence of morbidity and mortality. From pooled data of more than 60
28S
Table 2. International Classification of Diseases codes for hypertension
Essential hypertension
401.0
401.1
401.9
Hypertensive heart disease
402.00
402.01
402.10
402.11
402.90
402.91
Hypertensive renal disease
403.00
403.01
403.10
403.11
403.90
403.91
Hypertensive heart and
renal disease
404.00
404.01
404.02
404.03
404.10
404.11
404.12
404.13
404.90
404.91
404.92
404.93
Secondary hypertension
405.01
405.09
405.11
405.09
405.91
405.99
Essential hypertension; Malignant
Benign essential hypertension
Unspecified essential hypertension
Hypertensive
Hypertensive
Hypertensive
Hypertensive
Hypertensive
Hypertensive
heart
heart
heart
heart
heart
heart
disease;
disease;
disease;
disease;
disease;
disease;
malignant; without congestive heart failure
malignant; with congestive heart failure
benign; without congestive heart failure
benign; with congestive heart failure
unspecified; without congestive heart failure
unspecified; with congestive heart failure
Hypertensive
Hypertensive
Hypertensive
Hypertensive
Hypertensive
Hypertensive
renal
renal
renal
renal
renal
renal
disease;
disease;
disease;
disease;
disease;
disease;
malignant; without mention of renal failure
malignant; with renal failure
benign; without mention of renal failure
benign; with renal failure
unspecified; without mention of renal failure
unspecified; with renal failure
Hypertensive heart and renal disease;
heart failure or renal failure
and renal failure
heart failure and renal failure
and renal failure
heart failure or renal failure
failure
failure and renal failure
Secondary
Secondary
Secondary
Secondary
Secondary
Secondary
hypertension;
hypertension;
hypertension;
hypertension;
hypertension;
hypertension;
prospective studies and 1 million patients, Lewington et al17
report that 10 –mm Hg reductions in systolic BP would be
expected to reduce stroke mortality by as much as 40%.
FIG. 1 Prevalence of hypertension by age.
100 of 158.
malignant; w/o mention of congestive
malignant; with renal failure
benign; w/o mention of congestive
benign; with renal failure
unspecified; w/o mention of congestive
unspecified; with congestive heart
unspecified; with renal failure
unspecified; with congestive heart
malignant; renovascular
malignant; other
benign; renovascular
benign; other
unspecified; renovascular
unspecified; other
Importantly, the authors note that even a 2–mm Hg reduction in systolic BP is associated with a 10% lower death
rate from stroke. These reductions in risk apply all the way
to BP levels of 115/75 mm Hg. Thus, the failure to lower
BP even modestly in patients with hypertension is responsible for a significan number of preventable cardiovascular events each year.
The financia implications of hypertension and hypertension management are substantial.18 The direct costs of
treating hypertension exceeded $37 billion in the year
2003, and additional costs due to loss of productivity were
more than $13 billion (Table 4). Of the ten medical conditions evaluated for their effects on absenteeism from
work and loss of productivity, hypertension was the most
29S
Table 3. Trends in awareness, treatment, and control of high blood pressure in adults with hypertension 18
to 74 years of age
NHANES
Awareness
Treatment
Control
II
(1976–1980)
III-1
(1988–1991)
III-2
(1991–1994)
1999–2000
51
31
10
73
55
29
68
54
27
70
59
34
NHANES ⫽ National Health and Nutrition Examination Survey.
All numbers expressed as percentages.
expensive— costing businesses an average of $392 per
eligible employee per year.19 Improving the efficienc and
effectiveness of drug management in the hypertensive
population would likely reduce these costs in addition to
decreasing morbidity and mortality associated with the
condition.
Hemodynamic Measurements
Using ICG
The historical use of BP without CO or SVR is, in part,
because it has been impractical to estimate or measure
these parameters in most clinical settings. The assessment
of the hemodynamic components of hypertension from
clinical evaluation alone is unreliable. Even in patients
with acute conditions such as those requiring the emergency department or patients with decompensated congestive heart failure (in whom hemodynamic derangements
are greater than those in patients with essential hypertension), clinicians are generally unable to estimate CO or
SVR with accuracy.20,21
Echocardiography has been used to measure cardiac
output and in some studies has demonstrated acceptable
correlation with invasive techniques.22 However, in a
comparison with ICG, echocardiography is considerably
more time consuming and technically demanding.23 In the
offic setting, CO is not generally reported by most physicians interpreting echocardiograms in clinical practice.
Thus, until recently, CO and SVR were commonly measured only in the intensive care setting or catheterization
laboratory setting using invasive means such as a pulmonary artery catheter.
In recent years, ICG has emerged as an accurate, safe,
and inexpensive tool with which to measure hemodynamic
parameters by noninvasive means. The procedure is most
commonly performed in the physician offic setting by
medical assistants or nurses, requiring about 5 min to
complete the test. Using four sets of paired sensors on the
neck and chest, ICG measures the instantaneous change of
an electrical signal across the thoracic cavity (Fig. 2). As
the changes of thoracic impedance during the cardiac
cycle are most dependent on the changes in the size and
the blood volume of the thoracic aorta, ICG is able to
calculate the amount of blood ejected from the left ventricle (that is, the stroke volume [SV]). The product of
heart rate (HR) and SV yields CO. In addition, ICGderived parameters related to the changes of thoracic impedance are indicative of aortic blood velocity and
acceleration, and they correlate with measures of inotropic
state and cardiac performance. As flui is the best conductor of the electrical signal through the chest (when
Table 4. Direct and indirect costs attributable to
hypertension
Type of cost
Direct costs
Inpatient
Professional services
Drugs and medical durables
Home health care
Total direct costs
Indirect costs of lost productivity
Related to morbidity
Related to mortality
Total indirect costs
Total costs
Cost (in $
billion)
8.7
9.2
17.8
1.5
37.2
7.0
6.1
13.1
50.3
Data are from Heart Disease and Stroke Statistics—2003 Update.
American Heart Association; 2002.
101 of 158.
FIG. 2 Measurement of impedance signal using four sets of paired
sensors. Sensors transmit and record electrical signal from which
multiple hemodynamic parameters are derived.
30S
FIG. 3 Representative simultaneous tracings of electrocardiogram
(ECG), thoracic impedance (⌬Z), and first-time derivative of impedance (dZ/dt), and dZ/dt waveforms. Fiducial points are identified on
the dZ/dt tracing from which various hemodynamic and parameters
and timing intervals are derived.
compared with bone, air and fat, in particular), the total
thoracic impedance is inversely related to an index of flui
termed the “thoracic flui content” (TFC). Finally, using a
simultaneous electrocardiographic recording, ICG measures the pre-ejection period and LV ejection time—timing intervals that relate to cardiac performance.
Representative simultaneous tracings of an electrocardiogram, change in thoracic impedance (⌬Z), and firs
derivative of impedance (dZ/dt) are shown in Fig. 3. From
the measured variables and from HR and mean BP determined by oscillometry, SVR and other parameters are
calculated and displayed. An ICG test report is shown in
Fig. 4. A more detailed description of selected parameters
is provided in Table 5.
tients’ disease progression, response to therapy, and
need for further intervention. Thermodilution, using a
pulmonary artery catheter, has traditionally been the
standard to which ICG has been compared. Van De
Water et al24 assessed the relative reproducibility of
ICG and thermodilution cardiac outputs in hospitalized
patients in whom a pulmonary artery catheter was
placed for hemodynamic monitoring after bypass surgery. Serial ICG measurements in a given patient
showed better reproducibility than serial CO measurements using thermodilution technique (Table 7). The
investigators concluded that current ICG technology has
advanced such that ICG provides “a level of agreement
that is equivalent to thermodilution.” Their finding
support the clinical utility of ICG for serial measurements in patients with cardiovascular disease.
In a stable group of patients in the outpatient setting,
Verhoeve et al30 demonstrated a high reproducibility of
measurements performed on the same day and appropriate
sensitivity for the physiologic variations expected from
day to day. The variation in the average of readings for
CO, SVR, and thoracic flui content (TFC) ranged between 3% and 7% for serial measurements 1 week apart.
Figure 5 illustrates the high degree of correlation between
stroke index measured on day 1 and then 1 week later in
96 patients who were clinically stable.
The ICG technique is widely applicable, and reliable
information can be obtained in minutes at virtually no
Validation of Current ICG
Technology
Placement of a pulmonary artery catheter is a costly procedure requiring special training and expertise; and it is
associated with risks of bleeding, infection, and damage to
vascular and other structures. Because of the risks inherent
in invasive methods for measuring hemodynamics, studies
comparing ICG to invasive techniques of hemodynamic
measurement are only available from populations with
significan underlying cardiovascular conditions or situations that justify the risks associated with pulmonary artery catheter placement. In such clinical settings and
patient populations, multiple studies have shown that current ICG technology, using advanced data processing and
modeling techniques, yields data that are significantl
more accurate than those obtained with prior generations
of ICG devices.24 Five additional validation studies of
ICG presented since 1998, 25–29 using refine ICG technology (BioZ ICG Monitor, CardioDynamics, San Diego,
CA), demonstrate the high correlation and accuracy available with ICG when compared with invasive techniques
(Table 6).
The ability to measure changes in hemodynamic
parameters reliably in a given patient is critically important from a clinical perspective, as the changes in
serial measurements are the basis for evaluating pa-
102 of 158.
FIG. 4 Hemodynamic status report.
31S
Table 5. Impedance cardiography variables
Impedance Cardiography
Variable
Units
Blood flow
Stroke volume
Stroke index
Cardiac output
Cardiac index
Resistance
Systemic vascular resistance index
Contractility
Velocity index
Measurement/Calculation
mL
mL/m2
L/min
L/min/m2
VI ⫻ LVET ⫻ VEPT (Z MARC algorithm)
SV/BSA
SV ⫻ HR
CO/BSA
dyne · sec · cm⫺5
dyne · sec · cm⫺5 · m2
[(MAP - CVP)/CO] ⫻ 80
[(MAP - CVP)/CI] ⫻ 80
/100/sec2
msec
msec
-
1000 ⫻ first-time derivative of ⌬Zmax/
baseline impedance
100 ⫻ second-time derivative of
⌬Zmax/baseline impedance
ECG Q wave to aortic valve opening
Aortic valve opening to closing
PEP/LVET
kg · m/m2
(MAP - PCWP) ⫻ CI ⫻ 0.0144
/kOhm
1000 ⫻ 1/baseline impedance
/1000/sec
Acceleration index
Pre ejection period
Left ventricular ejection time
Systolic time ratio
Cardiac work
Left cardiac work index
Fluid status
BSA ⫽ body surface area; cm ⫽ centimeter; CVP ⫽ central venous pressure (estimated value of 6 mmHg); ECG ⫽ electrocardiography; HR
⫽ heart rate; ICG ⫽ impedance cardiography; MAP ⫽ mean arterial pressure; PCWP ⫽ pulmonary capilary wedge pressure (estimated value
of 10 mmHg); R to R interval ⫽ 60/ heart rate: VEPT ⫽ volume of electrically participating tissue; Z MARC ⫽ impedance modulating aortic
compliance.
risk to the patient. However, ICG has some limitations
related to the technology and patient factors. Although
ICG equations have demonstrated accuracy over a wide
range of conditions and patient populations, ICG has
not been evaluated extensively in patients ⬍66 pounds
or ⬎342 pounds. Severe aortic insufficienc may affect
ICG reliability, but it has not been fully studied and
validated in such patients. In addition, a few models of
permanent pacemakers use impedance technology to
measure minute ventilation. If the minute ventilation
function is activated, the paced rate may increase because of ICG signals31; therefore, patients with such
pacemakers must have the minute ventilation sensor
function inactivated before ICG testing. In patients with
atrial fibrillatio or frequent premature ventricular contractions, marked irregularity in heart rhythm can affect
data collection and analysis of wave forms.
Hemodynamic Parameters in
Hypertension
Hypertension is the result of complex cardiac, renal, neurohormonal, and vascular mechanisms that are modulated
by both genetic and environmental factors.10,32 The interactions of these many factors result in endothelial dysfunction and hemodynamic derangements of arterial
compliance, CO, and SVR. As noted earlier, MAP is the
product of CO and SVR, and elevations of BP can result
Table 6. Validation studies of impedance cardiography (ICG)
Population
AuthorRef
25
Parameter
Comparison
r Value
Bias
Precision
HF in ICU
HF in catheterization
laboratory
Albert et al
CO
ICG-TD
0.89
0.08
1.38
Drazner et al26
CO
Post-CABG
Post-CABG
Pulmonary hypertension
Ziegler et al27
Sageman et al28
Van De Water et al24
Yung et al29
CO
CI
CO
CO
ICG-Fick
TD-Fick
ICG-TD
ICG-TD
ICG-TD
ICG-TD
ICG-Fick
TD-Fick
ICG-TD
0.73
0.81
0.76
0.89
0.92
0.81
0.84
0.89
0.80
0.74
0.75
0.03
⫺0.45
0.07
⫺0.17
⫺0.24
0.19
⫺0.43
1.1
0.95
1.1
1.2
0.40
1.09
0.87
0.76
1.01
CABG ⫽ coronary artery bypass surgery; CO ⫽ cardiac output; HF ⫽ heart failure; ICU ⫽ intensive care unit. TD ⫽ thermodilution;
103 of 158.
32S
Table 7. Reproducibility of serial measurements:
impedance cardiography (ICG) versus thermodilution (TD)
Comparison
Correlation
(r value)
SD (L/min)
TD 2 v TD 1
TD 3 v TD 2
TD 3 v TD 1
ICG 2 v ICG 1
ICG 3 v ICG 2
ICG 3 v ICG 1
0.83
0.84
0.83
0.97
0.98
0.97
1.02
1.01
1.07
0.44
0.39
0.43
Adapted from Van De Water et al.24
from elevation of either or both of these hemodynamic
parameters. Pulse pressure (PP), that is, the difference
between systolic BP and diastolic BP, is determined by SV
and total arterial compliance (TAC). Arterial compliance
is a complex parameter that is most closely approximated
using a complicated and sophisticated model (the threeelement Windkessel model) that incorporates the ratio of
the decay time constant to peripheral resistance.33,34 True
arterial compliance is thus tedious and time-consuming to
measure and is not clinically useful. However, studies
have shown that arterial compliance can be reliably estimated as the ratio of SV to PP.34,35 Relationships among
the hemodynamic parameters including PP, SV, MAP,
CO, and SVR are shown in Fig. 6. The hemodynamics of
hypertension have been studied for decades, and previously various aspects have been extensively reviewed.3,36 –39
Hemodynamics of Hypertension:
Diagnostic Considerations
Numerous studies using either invasive or noninvasive
techniques have demonstrated that there are distinct hemodynamic subsets among various groups of patients with
hypertension. Hemodynamic measurements allow the differentiation of patients with primarily elevated CO from
those in which elevated SVR (signifying a vasoconstricted
state) is the primary mechanism of their hypertension.
FIG. 5 Reproducibility of stroke index (SI) measurements made 1
week apart. Adapted from Verhoeve et al.30
104 of 158.
FIG. 6 Components of mean arterial pressure (MAP) and pulse
pressure (PP). CO ⫽ cardiac output; SV ⫽ stroke volume; SVR ⫽
systemic vascular resistance.
Moreover, hemodynamic measurements can elucidate the
relative contributions of SV and arterial compliance to
elevations in PP.
Invasive Hemodynamic and
Echocardiographic Studies
In the Tecumseh, Michigan study, 40 patients were studied
using echocardiographic techniques and investigators
found that 37% of patients with hypertension were “hyperkinetic,” as define by increased cardiac index, HR,
forearm blood flow and plasma norepinephrine levels.
The distribution of cardiac index in this population study
is shown in Fig. 7. The wide distribution of cardiac index
values in these patients provides corroboration that hypertension represents a heterogeneous mix of various hemodynamic subsets.
In general, aging is associated with decreases in CO
and increases in SVR, as shown in Fig. 8. In young adults,
hypertension may be more commonly associated with
increased CO, whereas in older adults it is more commonly associated with elevated SVR. Lund-Johansen41
found a change in hemodynamic pattern in patients with
borderline hypertension at 10 and 17 years of follow-up.
There was a significan and progressive decrease in CO
over time, associated with an increase in SVR.
Age-related changes in hemodynamic status, as evidenced by changes in arterial compliance, occur in patients
with hypertension even in the absence of changes in CO or
SVR. Slotwiner et al42 used echocardiographic estimates
of cardiac output to study hemodynamic parameters in 272
patients who were 25 to 80 years of age and had mild
hypertension. These investigators found that in their study
group, CO and SVR levels did not vary significantl with
age. However, vascular stiffness, as reflecte by the ratio
of PP to SV (the reciprocal of TAC) increased with age,
which is possibly the mechanism for increased rates of
cardiovascular events in elderly individuals. Others have
noted that arterial stiffness exerts deleterious effects due to
increases in central aortic pressure—another hemodynamic mechanism that is key in the pathophysiology of
hypertensive cardiovascular disease.43
FIG. 7 Distribution of cardiac index values in Tecumseh, Michigan
study. Cardiac index values for the hypertensive population show a
bimodal distribution. From Julius et al.40
Other factors besides age appear to predict general
trends in the hemodynamic parameters in hypertensive
populations. Hemodynamic parameters differ between hypertensive men and women. Messerli et al44 measured PP,
CO, and SVR using invasive techniques in 200 subjects.
Despite equal levels of arterial BP, women had signifi
cantly higher CO, PP, and lower SVR compared with men.
Isometric exercise was associated with an increase in
arterial pressure that was nearly 50% greater in men than
in women. The hemodynamic differences between men
and women were confine to premenopausal women, suggesting that estrogens play a significan role in the cardiovascular and hemodynamic responses in patients with
hypertension. The mechanisms of hypertension seen with
acute stressors, such as public speaking or mental arithmetic, also vary based on gender. Studies have shown that
men and postmenopausal women have a more significan
increase in SVR in response to acute stressors, whereas
premenopausal women exhibit a hypertensive response
that is due primarily to increases in CO. Some studies have
suggested that hypertension early in the course of diabetes
and with obesity are associated with increased CO and
relatively normal SVR.45 Others have shown that the
earliest hemodynamic abnormalities may be changes in
arterial compliance.46 To a significan degree, hypertension in patients on dialysis results from volume expansion
and can be associated with signs of sympathetic stimulation such as increased HR and SV.
33S
CO were lower and HR and SVR were higher in women.
Age-related changes included an increase in total thoracic
impedance, equivalent to a decrease in its reciprocal TFC
and consistent with decreasing cardiopulmonary volume
or muscle mass or both.
In a study comparing hemodynamic variables between
pre-menopausal and post-menopausal women, Hinderliter
et al48 showed that post-menopausal women had lower CO
and higher SVR for any given BP level compared with
pre-menopausal women. Importantly, these significan
changes in CO and SVR occurred without significan
changes in BP levels, suggesting that the hemodynamic
parameters underlying BP provide more information than
does MAP alone. This is also seen in data from a study by
Galarza et al,49 in which, despite relatively stable DBP
levels in patients from the third to seventh decades of life,
the investigators found significan increases in SVRI of
nearly 50% and decreases in cardiac index of 27%. Alfi
et al50 used impedance techniques to show that elevations
in the difference between SBP and DBP (pulse pressure
[PP]) occurred due to different hemodynamic mechanisms
in men ⬍30 years of age compared with those middle aged
and older. In younger men, increased PP was associated
with increases in stroke index, reflectin preserved hemodynamic load with normal arterial compliance. In contrast,
after age 50 years, men showed increases in PP associated
with decreases in stroke index, reflectin age-related decreases in arterial compliance. Thus, BP readings alone
did not reflec the underlying hemodynamic differences in
groups with presumably different cardiovascular risk despite similar levels of MAP and PP. Gender differences are
seen in impedance studies of the hypertensive response to
caffeine: men who show hypertensive responses to caffeine increase their SVR, whereas women primarily increase SV and CO.51 Yu et al52 studied hemodynamic
parameters in patients with different mood states. Findings
of correlation of CO and SVR— but not SBP, DBP, or
Impedance Cardiographic Studies
Impedance cardiography has been used to evaluate the
hemodynamic parameters in normotensive individuals at
different ages and in various hypertensive populations. In
a study of 640 normal subjects evaluated as renal transplant donors, Taler et al47 found that increasing age was
associated with increasing BP, increasing SVR, and decreasing CO due to decreased SV. Hemodynamic changes
with age were similar in men and women, although BP and
105 of 158.
FIG. 8 Age-related changes in cardiac output and peripheral resistance. With increasing age, peripheral resistance rises in the
hypertensive population and at higher levels it is associated with
decreasing cardiac output and ultimately congestive heart failure.
34S
Table 8. Improved predictive power of pulse pressure (PP) to stroke volume index ratio (SVIR) compared with pulse pressure alone*
PP
PP to SVIR
Hazard rate,
CV event
Hazard rate,
mortality
1.46 (0.91–2.35)†
1.79 (1.15–2.80)‡
1.65 (0.83–3.30)†
2.05 (1.15–3.65)‡
CV ⫽ cardiovascular.
Values and 95% confidence intervals for 1-SD increase in the
variable of interest.
Adapted from Fagard et al.56
* Adjusted for age, gender, mean arterial pressure, and heart
rate; † P ⫽ NS; ‡ P ⬍ .01.
MAP—with affective state is further evidence of the heterogeneity of hemodynamic subsets within various groups
clinically identified
Hinderliter53 reported that African American men and
women had increased SVR, decreased CO, and associated
LV remodeling compared with Caucasian men and women
despite similar BP readings. In normotensive African
Americans, Calhoun et al54 postulated that vasoconstrictor
responses seen with mental stress and cold presser testing
may contribute to elevated SVR and the development of
hypertension.
These and other studies using ICG technology show
that within any given population SVR, CO, and TAC show
significan variation. The heterogeneity of hemodynamic
finding within various cohorts is evidence that the specifi hemodynamic status of an individual patient cannot
reliably be predicted on the basis of age, gender, or ethnic
background. Moreover, that hemodynamic values cannot
be identifie by BP levels or clinical assessment alone
supports the need for measurement of hemodynamic parameters in individual patients with hypertension.
average of 16.2 years of follow-up. In this study, exercise
SVR— but not exercise BP—added prognostic value to
parameters measured at rest, suggesting that hemodynamic
variables other than BP might have greater prognostic
value.
A subsequent article56 reported the relationship between
another hemodynamic parameter, namely, the ratio of PP to
stroke index (PP-to-SVi ratio), and outcomes in patients
followed for an average of 16.5 years. In this study, the
PP-to-SVi ratio, that is, the reciprocal of TAC index, was
independently associated with cardiovascular events or death.
Each increase in PP-to-SVi ratio of 0.75 mm Hg/(mL/m2)
was associated with a 79% increase in the risk of a cardiovascular event (P ⫽ .01) and greater than double the risk of
all-cause mortality (P ⫽ .01). As shown in Table 8, the
increased hazard rates with PP-to-SVi ratio compared with
PP alone demonstrates the additional predictive value when
the hemodynamic parameter of flo (SVi) is added to the
pressure measurement alone.
De Simone et al57 studied the effects of TAC on cardiovascular events over a 10-year period. They found that
risks of fatal and total cardiovascular events were independently correlated with age, LV mass, and lower levels
of arterial compliance, define as decreasing values of the
measured ratio of echocardiographic SV to PP (SV/PP) to
that predicted from previously developed equations (%
SV/PP). Moreover, consistent with the results of Fagard et
al,56 the investigators found that systolic BP, mean BP, or
PP alone (without including the flow-relate hemodynamic parameter of SV) were not independent predictors
of prognosis. After adjustment for age and LVH there
remained an independent effect of % SV/PP on cardiovascular endpoints at 10-year follow-up (Fig. 9). These investigators found that hemodynamic parameters such as
arterial compliance and percent predicted arterial compliance correlated better with changes in cardiac structure (ie,
hypertrophy and remodeling) than did BP levels alone.
Prognostic Considerations
The underlying hemodynamic abnormalities in hypertension result in structural and functional changes in the
cardiovascular system that adversely affect prognosis, ie,
that increase risk of morbidity or mortality. Increases in
hemodynamic measures such as SVR and reductions in
arterial compliance provide prognostic information in addition to that obtained by BP measurements alone.
Invasive Hemodynamic and
Echocardiographic Studies
Elevated arterial BP is the result of increased arterial
stiffness and increased SVR. This results in increased LV
wall stress, the best measure of LV afterload. Using catheterization techniques, Fagard et al55 demonstrated that
SBP and SVR at rest and during exercise correlated with
the risk of cardiovascular events and total mortality at an
106 of 158.
FIG. 9 Event-free survival based on predicted ratio of stroke volume (SV) to pulse pressure (PP). Event-free survival is reduced in
hypertensive patients with reduced arterial compliance, as defined
as a low predicted SV/PP ratio. From De Simone et al.57
Gender has long been known to affect prognosis in
patients with hypertension. In 1913, Janeway58 published
the observation that women with hypertension tended to
have a better prognosis than men. More recent studies
have suggested that this difference may be related to
different hemodynamic substrates in men compared to
women with elevated BP levels. Messerli et al44 suggested
that the disparate prognosis between men and women
might be explained on the basis of differing hemodynamic
mechanisms: “For any level of arterial pressure, total
peripheral resistance (and therefore the risk of hypertensive cardiovascular disease) was lower in women than in
men.”
The mechanism of the adverse prognosis from hypertension is in part related to structural changes in the heart
that result from elevated wall stress. Prolonged increases
in wall stress lead to left ventricular (LV) structural
changes with relative increases in wall thickness, overall
LV mass or both.59 In concentric remodeling, there is a
relative increase in LV thickness without increase in overall LV mass. This structural change appears to be related
to increased pressure load but with relative decrease in
volume as evidenced by low normal CO (termed “volume
underload”). Concentric left ventricular hypertrophy
(LVH) is characterized by an increase in wall thickness
with increase in LV mass or mass index and also results
from pressure overload caused by long-term hypertension.
Eccentric hypertrophy is define as increased LV mass
index with preserved relative wall thickness and is associated with both pressure and volume overload. This pattern, a common result of the afterload and volume
excesses in long-standing severe aortic insufficiency results in spherical remodeling of the LV. Interestingly, in
this study of hypertensive individuals,59 both eccentric
hypertrophy and concentric remodeling were more common than the “classic” pattern of hypertensive heart disease, namely, concentric LVH.
Nonetheless, LVH is a powerful predictor of cardiovascular risk and is independently associated with mortality in
patients with coronary artery disease.60 – 63 For example,
Vakili et al63 reported on the pooled results of 20 published studies of LVH as define by electrocardiographic
or echocardiographic criteria. They demonstrated a
weighted mean relative risk of cardiovascular morbidity
from LVH of 2.3, independent of all covariates analyzed.
As reported by Ichkhan et al,64 LVH is associated with
abnormalities of ventricular repolarization and at least a
twofold increase in the risk of serious ventricular arrhythmias.
Hypertension results in abnormalities of endothelial
function, affecting hemodynamic factors such as arterial
compliance. Gomez-Cerezo et al65 demonstrated impaired
brachial artery flow–mediate dilation, a common test of
endothelial function, in patients with sustained or labile
hypertension. They found that flow-mediate dilation was
abnormal to a similar degree in patients with sustained
essential hypertension or “white-coat hypertension” but
107 of 158.
35S
was normal in individuals with normal BP levels. Others
have evaluated measures of arterial compliance (or, alternatively, arterial stiffness) using the measure of pulse
wave velocity.66,67
Additional structural changes occuring at the level of
the heart and blood vessel have prognostic significanc in
persons with hypertension, including vascular remodeling
with changes in lumen to wall thickness.68,69 Apoptosis, or
programmed cell death, contributes to the vascular
changes (ie, remodeling) in hypertension. Inflammatio
and fibrosi similarly contribute with the accumulation of
various components in the extracellular matrix such as
collagen and fibronectin Intengan and Schiffrin69 reviewed the factors that result in arterial remodeling and
altered hemodynamic parameters in patients with hypertension.
Importantly, studies have shown that treatment with
antihypertensive agents may result in regression of the
structural abnormalities caused by long-standing hypertension and may result in improved prognosis.70 –72 Mathew
et al,70 reporting on data from the Heart Outcomes Prevention Evaluation (HOPE) study, demonstrated that treatment with the angiotensin-converting enzyme (ACE)
inhibitor ramipril was associated with regression of LVH
by electrocardiographic criteria compared with placebo
control. That BP showed minimal difference between the
treatment and control groups is consistent with other studies demonstrating improvement in overall hemodynamics
(as shown by significan decreases in SVR and parallel
increases in CO) that are not evident from BP levels alone.
Ofil et al,73 in an echocardiographic substudy of the
Systolic Hypertension in the Elderly Program (SHEP),
demonstrated partial regression of LVH in patients treated
with a diuretic-based regimen for a minimum of 3 years. In
a meta-analysis of ⬎1000 patients with serial echocardiography during treatment of essential hypertension, Verdecchia et al74 demonstrated that patients whose LVH
regressed during treatment had significantl fewer cardiovascular events compared with those in whom LV mass
increased, consistent with the hypothesis that improvements in hemodynamics correlate with improved prognosis in patients with appropriately treated hypertension.
Impedance Cardiographic Studies
The ICG technique has been used to explore age-related
changes in hemodynamic variables and their correlation
with cardiovascular risk and the adverse prognosis. These
studies support previous finding that future risk in patients with hypertension may not be reflecte in BP levels
alone. Alfi et al50 demonstrated that despite similar elevations in PP, younger men had preserved stroke index
(and arterial compliance) compared with older men. They
concluded that preserved arterial compliance and cardiac
pump function may explain the lack of prognostic significance of elevated PP in younger men. These finding lend
further support to the value of the incremental information
36S
FIG. 10 Hemodynamic components of hypertension. BP ⫽ blood
pressure; CO ⫽ cardiac output; HR ⫽ heart rate; SV ⫽ stroke volume; SVR ⫽ systemic vascular resistance.
provided by the hemodynamic components of BP and PP
levels.
Hemodynamic differences have been demonstrated in
patients who have experienced complications of hypertension compared with those with hypertension alone. In a
study of hemodynamic status in hypertensive patients with
and without a history of stroke, Galarza et al75 found lower
cardiac index and higher SVR index in those with history
of stroke. These differences occurred in the absence of
differences in BP or antihypertensive treatment, providing
another example of the unreliability of BP to reflec the
severity of underlying hemodynamic abnormalities.
Therapeutic Considerations
Hypertension management includes hygienic measures
such as sodium restriction and weight loss; and, in most
cases, it requires the use of one or more antihypertensive
agents. Antihypertensive medications exert their BP-lowering effects by reductions in SVR or CO. Hemodynamic
effects can be used to classify antihypertensive agents,
predict the response to antihypertensive therapy, and guide
both the initiation and titration of these agents.76 –78
Just as interpretation and treatment of serum cholesterol
level improves when its components (HDL-cholesterol and
LDL-cholesterol) are measured, hypertension may be better
diagnosed and treated by examining its hemodynamic components (CO and SVR). As MAP is the product of CO and
SVR, elevated mean BP results from elevated CO, SVR, or
both. As shown in Fig. 10, CO is the product of HR and SV.
Stroke volume is determined in part by LV fillin (preload)
and contractile (inotropic) state. Hypertension can thus result
from increases in SVR (vasoconstriction), HR (hyperchronotropy), preload (hypervolemia), or contractility (hyperinotropy).
Invasive and Echocardiographic Studies
In a small group of men with severe hypertension, Sullivan
et al79 studied the relationship between baseline hemodynamic status and the response to various antihypertensive
108 of 158.
agents that were randomly selected. Patients with elevated
SVR responded with decreases in SVR, and those with
elevated CO had BP control associated with normalization
of CO.
Treatment targeted at the specifi hemodynamic cause
of hypertension has predictable and appropriate results.
Easterling et al80 studied noninvasive hemodynamic parameters using Doppler echocardiography in 19 pregnant
hypertensive women. Ten patients had elevated CO,
whereas nine patients had elevated SVR, demonstrating
hemodynamic heterogeneity within this apparently homogeneous population. Patients with elevated CO were
treated with a ␤-blocker (atenolol) and those with elevated
SVR were treated with hydralazine, a vasodilator targeted
at elevated SVR. Patients given hydralazine had dramatic
improvements in CO in association with decreases in
SVR; those given atenolol for elevated CO had improvement in BP and normalization of CO. The investigators
suggest that the failure of previous studies to show consistent results in the drug management of hypertension in
pregnancy may have resulted from treating a heterogeneous hemodynamic group with a single regimen. The
implication of their study is that hemodynamically guided
therapy would be expected to show more consistent results
in hemodynamically diverse populations.
Differential effects of antihypertensive medications on
hemodynamic variables may not be evident from changes
in BP alone. Resnick and Lester81 studied the effects of
various BP medications on arterial compliance in patients
referred to an outpatient practice specializing in hypertension. The changes in compliance of the large arteries
(capacitive compliance) and in smaller arteries (reflectiv
compliance) were evaluated during treatment with ACE
inhibitors, angiotensin-receptor blockers (ARBs), calcium
channel blockers (CCBs), and ␤-blockers. The researchers
found that despite similar changes in SBP, DBP, and PP
during treatment, there were improvements in arterial
compliance with ACE inhibitors, ARB, and CCB but not
with ␤-blockers. These researchers suggest that choosing
medications that have favorable effects on both BP and
arterial compliance “might further enhance the potential
clinical benefi of drug therapy in hypertension.” Similarly, Zusman78 reported that despite similar degrees of BP
reduction, the hemodynamic effects of the CCB nifedipine
were favorable when compared with the ␤-blocker atenolol, resulting in decreased SVR, increased CO and improved measures of LV contractility and diastolic
function. Others have shown significantl different hemodynamic effects between various ␤-blockers such as between metoprolol and carvedilol due to the ␣-adrenergic
blocking properties of the latter.82
Studies suggest that most patients require multiple
medications to achieve BP control. In the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT),83 a large randomized trial
comparing outcomes in patients treated with different
classes of antihypertensive agents, 90% of patients were
109 of 158.
2 or NC
2
2
1
2
1 or NC
SVR
CO
Stroke volume
Heart rate
Intravascular volume
LVH
ACE ⫽ angiotensin converting enzyme; CO ⫽ cardiac output; LVH ⫽ left ventricular hypertrophy; NC ⫽ no change; SVR ⫽ systemic vascular resistance; 1 ⫽ Increase, 2 ⫽ decrease.
Adapted from Houston.85
2 or NC
2 or NC
2 or NC
2 or NC
NC or 1
2
2
1
1
1
1
1
2
NC or 1
NC or 1
NC
NC or 1
2
2
NC
NC
2 or NC
2 or NC
2
2
1
1
NC
2
2
2
1 or NC
1 or NC
NC or 2
2
2
Diuretics
Effect
1
2
2
2
1 or NC
NC
ACE
inhibitors
Calcium
channel
blockers
The foregoing discussion has presented data on the role of
hemodynamic information for diagnostic, prognostic, and
therapeutic decision making for patients with hypertension. The use of ICG-derived hemodynamic information to
improve BP control requires accurate assessment of baseline hemodynamic state, creation of a therapeutic regimen
based on hemodynamic status, and timely measurement of
changes in various hemodynamic parameters in response
to therapy.86 Studies have shown that it is very difficult—i
not impossible—to make an accurate assessment of CO
and SVR at the bedside by physical examination
alone.20,21,87,88 Therefore, it is not likely to be possible to
use physical examination to reliably identify baseline hemodynamic subsets or changes in hemodynamic status so
as to optimize therapy.
Clinicians have used ICG in various patient care settings to assess its applicability in the assessment and
treatment of hypertension.89 –95 As noted earlier here and
in Fig. 7, there is significan hemodynamic heterogeneity
among individuals with hypertension, suggesting that BP
level alone is not adequate to categorize patients into
clinically meaningful subgroups. De Divitiis et al91 used
ICG to confir the presence of distinct hemodynamic
profile in patients with hypertension: 1) elevated CO in
association with normal or nearly normal SVR, and 2)
predominantly elevated SVR. Margulis et al96 evaluated
other hemodynamic parameters in untreated patients with
hypertension. They found impairment of cardiac performance with decreased indices of contractility and evidence
for increased thoracic flui content, suggesting increased
water content of the lungs or thoracic wall tissues.
Thoracic flui content, the reciprocal of total thoracic
Table 9. Selected hemodynamic effects of various antihypertensive agents by class
Rationale for an ICG-Guided
Approach to Antihypertensive
Therapy
Central
␣-agonists
␣blockers
Direct
vasodilators
␣- and ␤blockers
␤-blockers
with ISA
on treatment at time of pre-randomization visit, although
only 27% had adequate BP control. After 5 years of treatment, 66% had achieved levels of BP ⬍140/90 mm Hg. Of
the participants whose hyptertension was controlled, 63%
were on two or more medications, indicating that BP
control required combination therapy in the majority of
cases. Yakovlevitch and Black84 reviewed 436 charts to
identify 91 cases of resistant hypertension referred to their
hypertension clinic of a tertiary care center and evaluated
for possible causes of resistance, including medication
noncompliance, secondary causes of hypertension, drug
interactions, and the appropriateness of the medical regimen. They found that the most common cause of inadequate BP control in the 91 patients identifie with resistant
hypertension was an inappropriate medical regimen.
The hemodynamic responses of various classes of antihypertensive medications have been categorized in an
extensive review of hypertension by Houston,85 and selected hemodynamic effects are summarized in Table 9.
2
2 or NC
2 or NC
NC or 1
NC
1
␤
blockers
37S
38S
impedance, is strongly correlated with amount of flui in
the chest cavity, whether intravascular or extravascular. In
patients undergoing thoracentesis, Petersen et al97 demonstrated a strong correlation between the volume of pleural
flui removed and the change in total thoracic impedance
(correlation coefficient 0.97). In studies using lower-body
negative pressure to create pooling of venous blood in the
lower extremities, Ebert et al98 found a nearly perfect
linear correlation with changes in central venous pressure
and changes in thoracic impedance. Thus, TFC has been
used to monitor changes in flui volume and guide diuretic
therapy in patients with hypertension.
Linb et al89 reported that BP reductions resulted from
improvements in baseline hemodynamic abnormalities;
patients with elevated CO responded to targeted therapy
with a ␤-blocker (propranolol), whereas those with elevated SVR responded to treatment with the vasodilating
CCB (nifedipine). Mattar et al99 showed that an intensive
regimen of diet and exercise resulted in improvements in
hemodynamic parameters with substantial increases in CO
and decreases in SVR despite only modest changes in
MAP. Moreover, the investigators speculated that failure
of some hypertensive patients to show hemodynamic improvement on serial ICG measures resulted from the inappropriate choice of medications that were not targeted
toward the underlying hemodynamic abnormalities.
Hemodynamic parameters derived from ICG have been
used to evaluate the differential effects of medications in
patients with essential hypertension. In a study of the
effects of a cardioselective ␤-blocker compared with a
␤-blocker with intrinsic sympathomimetic activity, Toth et
al100 studied 57 patients randomized to treatment with
either atenolol or pindolol for 12 weeks. Pindolol therapy
was associated with a 12% decrease in SVR compared
with minimal change with atenolol. Atenolol-related improvement in BP resulted from decrease in HR and cardiac
index. Breithaupt-Grogler et al67 reported on the differential hemodynamic effects of combination therapy with
verapamil/trandolapril (Vera/Tran) compared with metoprolol/hydrochlorothiazide (Meto/HCTZ) in 26 patients
after 6 months of therapy. In addition to ICG-derived CO
and SVR, the authors measured carotidofemoral pulse
wave velocity as a measure of arterial stiffness. The combination of CCB and ACE inhibitor (Vera/Tran) reduced
diastolic BP to a greater degree than Meto/HCTZ and
lowered SVR by about 15% compared with minimal
change with the ␤-blocker/diuretic combination. Treatment with Meto/HCTZ was associated with a significan
reduction in CO compared with baseline, which was not
seen with Vera/Tran. However, pulse wave velocity decreased with Vera/Tran but not with Meto/HCTZ, suggesting an improvement in the elastic properties of the aorta
with the former drug regimen.
The ICG technique has been used to assess the hemodynamic effects of sodium restriction in a small group of
subjects with mild hypertension.101 During sodium restriction, ICG-derived measures of SV decreased in association
110 of 158.
with fall in diastolic BP. An increase in overall thoracic
impedance (the reciprocal of TFC) was consistent with a
decrease in extracellular flui volume. In addition, ICG
has been used to explore the mechanisms of responses to
ACE inhibitors and prostaglandin inhibitors in patients
who are either salt sensitive or salt insensitive.102 These
examples illustrate the use of ICG in assessing the mechanisms of BP elevation and the hemodynamic effects of
nonpharmacologic interventions in hypertension.
As noted above, antihypertensive medications ultimately act on one or more of the hemodynamic components that determine BP.79,80 Once the baseline
hemodynamic status is known, an appropriate medical
regimen can be designed based on the expected hemodynamic effects of various medications. However, individual
patients vary in their responses to antihypertensive drugs
such that the actual hemodynamic effects and side effects
cannot reliably be predicted. Therefore, empiric selection
of drug combinations based on their general hemodynamic
actions as a class may not be successful in managing a
specifi patient, even if the baseline hemodynamic status is
known. The ICG technique is unique in that it can provide
not only an accurate hemodynamic profil noninvasively
but can guide therapy toward a drug regimen that is most
appropriate for the specifi patient based on serial measurements. Periodic measurements of hemodynamic status
allow the physician to monitor therapy when results are
suboptimal or unexpected. It is for these reasons that ICG
has emerged as a valuable tool in the evaluation and
treatment of patients with hypertension.
Outcome Studies Using ICGGuided Therapy in Hypertension
The observation that hypertension is a hemodynamic disease implies that measurement of hemodynamic parameters can be used to guide medication selection, to titrate
dose, and to evaluate efficac of the medical regimen.
Several studies have used ICG to evaluate hemodynamic
parameters and demonstrated that ICG-guided therapy improves BP control. Taler et al103 randomized 104 patients
with hypertension uncontrolled on two or more drugs to a
3-month trial of ICG-guided therapy or standard therapy
directed by a hypertension specialist. In this study, BP
control (define as achieving BP ⬍140/90 mm Hg) occurred 70% more often in the ICG-guided group (Fig. 11).
Use of ICG resulted in greater reductions in SVR index
and more intensive use of diuretic therapy, guided by
levels of TFC. According to the study investigators, measurement of hemodynamic and impedance parameters was
more effective than clinical judgment alone in guiding
selection of antihypertensive therapy patients resistant to
empiric therapy.
Sharman et al104 studied a cohort of patients in the
primary care offic setting with drug-resistant hypertension, define as systolic BP ⱖ140 mm Hg or diastolic BP
ⱖ90 mm Hg during treatment with two antihypertensive
FIG. 11 Percentage of patients achieving blood pressure (BP)
control using impedance cardiography (ICG)– guided therapy
compared with non–ICG-guided therapy. Adapted from Taler et
al.103
medications. Patients were treated based on a published
ICG-guided treatment algorithm (Fig. 12) for an average
of 7 months. In this study, ICG resulted in BP control in
57.1% of patients who were not controlled before the use
of ICG-guided therapy. The average number of medications increased from 2.0 at time of entry to 2.5 ⫾ 0.7 at the
end of the study period. The observation that hemodynamic information derived from ICG resulted in BP control with two medications in some patients and three or
more in others is consistent with both higher intensity and
more appropriate medical regimens. The investigators
concluded that ICG is safe and cost-effective and could
assist community-based physicians in treating uncontrolled hypertension.
Sramek et al105 reported on a series of 322 patients
with hypertension uncontrolled despite previous therapy with two or more antihypertensive agents for periods of 2 years or more. The researchers directed the
management of hypertension at both control of BP and
improvement in underlying hemodynamic parameters
including CO and SVR. At baseline, 16% of subjects
had significantl reduced CO (ie, were considered hypodynamic) and approximately 19% were hyperdynamic. In this large series of patients treated using the
results of ICG evaluation, so-called normodynamic
goal-oriented therapy controlled BP in 203 subjects
(63%) within several weeks. The investigators highlight
the observation that ICG was able to identify medications that were optimal and specifi for the individual
patients, resulting in an approach superior to the conventional “trial-and-error” method.
39S
achieve adequate BP control for reasons other than the
responses to specifi medications. Common barriers to BP
control include lack of awareness of the condition, inability to make necessary dietary and other lifestyle modifi
cations, noncompliance with medications, complicating
factors such as drug interactions, secondary causes of
hypertension, and comorbidities such as kidney disease.
Testing with ICG using currently available equipment
may favorably affect each of these issues. Oscillometric
measurements of BP, as with the most widely used ICG
equipment, are more reliable and less operator dependent
than standard BP techniques. The accurate and reproducible measures of CO and SVR identify patients with abnormal hemodynamic states and may increase clinical
suspicion and diagnostic sensitivity for those with borderline or prehypertensive BP readings. The ICG reports are
useful teaching tools for patients and may provide motivation for the dietary and other lifestyle changes that assist
in BP control.
Changes in hemodynamic parameters may identify instances when patients stop their medications or when there
are complicating factors such as worsening renal function
or interactions with medications such as over-the-counter
nonsteroidal anti-inflammator drugs. As noted,103 an increase in one class of hypertensive agents may result in
compensatory flui retention, leading to an increase in
TFC as measured by ICG and the need for higher doses of
diuretics. Similarly, flui retention resulting from the renal
effects of anti-inflammator medications may be recognized by changes in TFC. Importantly, ICG-derived measures of cardiac performance, such as velocity index or
Additional Roles of ICG in Patients
With Hypertension
The evidence cited above supports the use of ICG-derived
hemodynamic information in guiding the selection, initiation, titration, and evaluation of antihypertensive medication. However, patients and their physicians fail to
111 of 158.
FIG. 12 Algorithm for hemodynamically guided therapy of hypertension. Adapted from Sharman et al105 and Taler et al.104 In the
latter study, postural changes in total body impedance (TBI), the
reciprocal of thoracic fluid content (TFC), were used as the criteria
for decisions reguarding fluid status.
40S
systolic time ratio, may be the initial signs of the development of LV dysfunction.106,107
Implications of Hemodynamics
and Future Considerations
In addition to improving the diagnosis and therapy of
hypertension, hemodynamic measurements provide insights into other aspects of cardiovascular function. For
example, studies have shown the importance of endothelial function in the development and progression of cardiovascular disease.108 Endothelial dysfunction, as
measured by reduced flow-mediate arterial dilation, is
associated with abnormal hemodynamic measures, including elevated SVR.109 In the HOPE study,110 an ACE
inhibitor—a drug that both lowers SVR and improves
endothelial function—reduced mortality from cardiovascular disease despite only minor effects on BP. Future
studies will likely examine the significanc of elevated
SVR and arterial compliance in individuals with hypertension or prehypertension and will correlate ICG-derived
hemodynamic parameters with other evolving markers of
increased cardiovascular risk such as C-reactive protein,
homocysteine level, and the metabolic syndrome.
The studies included in this supplement of the journal add
to the growing body of literature that supports the accuracy,
reliability, and clinical utility of ICG in diagnostic and prognostic assessment and therapeutic management of patients
with hypertension. The use of ICG has added significantl to
our understanding of hypertension as a disease with both
hemodynamic causes and hemodynamic consequences. Just
as congestive heart failure reflect abnormal flo or inappropriate ventricular fillin pressures, hypertension occurs when
there is abnormal flo or inappropriate vascular resistance or
compliance. When hypertension impairs LV performance
(either systolic or diastolic), heart failure ensues. Although
these conditions often co-exist, in many cases hypertension is
a step in a hemodynamic continuum that leads to further
hemodynamic derangement and heart failure. It is believed
that future studies will confir recent finding that hemodynamic measurements in individual patients will improve diagnosis, risk assessment, and treatment for these patients. It is
also possible that further exploration of the implications of
hypertension as a hemodynamic disease will lead to studies
demonstrating that earlier detection and treatment of the
hemodynamic components of hypertension may change the
natural history of this disease process.
References
1.
2.
3.
Wiggers CJ: Hemodynamic changes in hypertension. Am Heart J
1938;16:515–525.
Freis ED: Hemodynamics of hypertension. Physiol Rev 1960;40:
27–54.
Freis ED: Studies in hemodynamics and hypertension. Hypertension
2001;38:1–5.
112 of 158.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Lund-Johansen P: Newer thinking on the hemodynamics of hypertension. Curr Opin Cardiol 1994;9:505–511.
Julius S: Autonomic nervous system dysregulation in human hypertension. Am J Cardiol 1991;67:3B–7B.
Susic D, Frohlich ED: Hypertension and the heart. Curr Hypertens
Rep 2000;6:565–569.
Messerli FH, DeCarvalho JG, Christie B, Frohlich ED: Essential
hypertension in black and white subjects. Hemodynamic finding
and flui volume state. Am J Med 1979;67:27–31.
Ventura H, Messerli FH, Oigman W, Suarez DH, Dreslinski GR,
Dunn FG, Reisin E, Frohlich ED: Impaired systemic arterial compliance in borderline hypertension. Am Heart J 1984;108:132–136.
Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA,
Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, Roccella EJ:
The Seventh Report of the Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure. The
JNC 7 Report. J Am Med Assoc 2003;289:2560 –2572.
Oparil S, Zaman MA, Calhoun DA: Pathogenesis of hypertension.
Ann Intern Med 2003;139:761–776.
Schappert SM, Nelson C: National ambulatory medical care survey:
1995–96 summary. Vital and health statistics, series 13, no. 142.
Washington, DC: Government Printing Office November 1999.
DHHS publication no. (PHS) 2000-1713.
Hyman DJ, Pavlik VN: Characteristics of patients with uncontrolled
hypertension in the United States. N Engl J Med 2001;345:479 –
486.
Setaro JF, Black HR: Refractory hypertension. N Engl J Med
1992;327:543–547.
Oparil S, Calhoun DA: Managing the patient with hard-to-control
hypertension. Am Fam Phys 1998;57:1007–1014.
Gifford RW Jr. Resistant hypertension: introduction and definitions
Hypertension 1988;11(Suppl II):II-65–II-66.
Wilking SV, Belanger A, Kannel WB, D’Agostino RB, Steel K:
Determinants of isolated systolic hypertension. J Am Med Assoc
1988;260:3451–3455.
Lewington S, Clarke R, Qizilbash N, Peto R, Collins R: Agespecifi relevance of usual blood pressure to vascular mortality: a
meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 2002;360:1903–1913.
Elliott WJ: The economic impact of hypertension. J Clin Hypertens
2003;5(3 Suppl 2):3–13.
Goetzel RZ, Long SR, Ozminkowski RJ, Hawkins K, Wang S,
Lynch W: Health, absence, disability, and presenteeism cost estimates of certain physical and mental health conditions affecting U.S.
employers. J Occup Environ Med 2004;46:398 – 412.
Steingrub JS, Celoria G, Vickers-Lahti M, Teres D, Bria W: Therapeutic impact of pulmonary artery catheterization in a medical/
surgical ICU. Chest 1991;99:1451–1455.
Connors AF, Dawson NV, Shaw PK, Montenegro HD, Nara AR,
Martin L: Hemodynamic status in critically ill patients with and
without acute heart disease. Chest 1990;98:1200 –1206.
Hoit BD, Rashwan M, Watt C, Sahn DJ, Bhargava V: Calculating
cardiac output from transmitral volume flo using Doppler and
M-mode echocardiography. Am J Cardiol 1988;62:131–135.
Northridge DB, Findlay IN, Wilson J, Henderson E, Dargie HJ:
Non-invasive determination of cardiac output by Doppler echocardiography and electrical bioimpedance. Br Heart J 1990;63:93–97.
Van De Water JM, Miller TW: Impedance cardiography: the next
vital sign technology? Chest 2003;123:2028 –2033.
Albert N, Hail M, Li J, Young J: Equivalence of bioimpedance and
thermodilution in measuring cardiac output and index in patients
with advanced, decompensated chronic heart failure hospitalized in
critical care. Am J Crit Care 2004;13:469 – 479.
Drazner M, Thompson B, Rosenberg P, Kaiser PA, Boehrer JD,
Baldwin BJ, Dries DL, Yancy CW: Comparison of impedance
cardiography with invasive hemodynamic measurements in patients
with heart failure secondary to ischemic or nonischemic cardiomyopathy. Am J Cardiol 2002;89:993–995.
27. Ziegler D, Grotti L, Krucke G: Comparison of cardiac output
measurements by TEB vs. intermittent bolus thermodilution in
mechanical ventilated patients (Abstract). Chest 1999;116:281S.
28. Sageman WS, Riffenburgh RH, Spiess BD: Equivalence of bioimpedance and thermodilution in measuring cardiac index after cardiac
surgery. J Cardiothoracic Vasc Anesth 2002;16:8 –14.
29. Yung, G, Fedullo P, Kinninger K, Johnson W, Channick R: Comparison of impedance cardiography to direct Fick and thermodilution cardiac output determination in pulmonary arterial
hypertension. Congest Heart Fail 2004;10(Suppl 2):7–10.
30. Verhoeve PE, Cadwell CA, Tsadok S: Reproducibility of noninvasive bioimpedance measurements of cardiac function (Abstract).
J Card Fail 1998;4(Suppl):S3.
31. Aldrete AJ, Brown C, Daily J, Buerke V: Pacemaker malfunction
due to microcurrent injection from a bioimpedance noninvasive
cardiac output monitor. J Clin Monit 1995;11:131–133.
32. Frohlich ED: Local hemodynamic changes in hypertension. Insights
for therapeutic preservation of target organs. Hypertension 2001;38:
1388 –1394.
33. Segers P, Brimioulle S, Stergiopulos N, Westerhof N, Naeije R,
Maggiorini M, Verdonck P: Pulmonary arterial compliance in dogs
and pigs: the three-element Windkessel model revisited. Am J
Physiol 199;277:H725–H731.
34. Randall OS, Westerhof N, van den Bos GC, Alexander B: Reliability of stroke volume to pulse pressure ratio for estimating and
detecting changes in arterial compliance. J Hypertens 1986;4(Suppl
5):S293–S296.
35. Chemla D, Hebert J-L, Coirault C, Zamani K, Suard I, Colin P,
Lecarpentier Y: Total arterial compliance estimated by stroke volume-to-aortic pulse pressure ratio in humans. Am J Physiol 1998;
274: (Heart Circ Physiol 43) H500 –H505.
36. Egan B, Schmouder R: The importance of hemodynamic considerations in essential hypertension. Am Heart J 1988;116:594 –599.
37. Izzo J: Arterial stiffness and the systolic hypertension syndrome.
Curr Opin Cardiol 2004;19:341–352.
38. Guyton AC: The relationship of cardiac output and arterial pressure
control. Circulation 1981;64:1079 –1088.
39. Fouad-Tarazi F, Izzo JL: Hemodynamic profile and responses, in
Izzo JL, Black HR (eds): Hypertension Primer, : The Essentials of
High Blood Pressure. Council on High Blood Pressure Research
(American Heart Association), Dallas, 1999, pp 115–117. Second
Edition
40. Julius S, Krause L, Schork NJ, Mejia AD, Jones KA, van de Ven C,
Johnson EH, Sekkarie MA, Kjeldsen SE, Petrin J: Hyperkinetic
borderline hypertension in Tecumseh, Michigan. J Hypertens 1991;
9:77– 84.
41. Lund-Johansen P: Hemodynamic patterns in the natural history of
borderline hypertension. J Cardiovasc Pharm 1986;8(Suppl 5):S8 –
S14.
42. Slotwiner DJ, Devereux RB, Schwartz JE, Pickering TG, de Simone
G, Roman MJ: Relation of age to left ventricular function and
systemic hemodynamics in uncomplicated mild hypertension. Hypertension 2001;37:1404 –1409.
43. Franklin SS: Blood pressure and cardiovascular disease: what remains to be achieved? J Hypertens 2001;19(Suppl 3):S3–S8.
44. Messerli FH, Garavaglia GE, Schmieder RE, Sundgaard-Riise K,
Nunez BD, Amodeo C: Disparate cardiovascular finding in men
and women with essential hypertension. Ann Intern Med 1987;107:
158 –161.
45. Davidson RC, Ahmad S: Hemodynamic profile in essential and
secondary hypertension, in Izzo JL, Black HR (eds): Hypertension
Primer: The Essentials of High Blood Pressure. Council on High
Blood Pressure Research (American Heart Association), Dallas,
2003, pp 349 –351. Third Edition
46. Romney JS, Lewanczuk RZ: Vascular compliance is reduced in
the early stages of type 1 diabetes. Diabetes Care 2001;24:2102–
2106.
47. Taler S, Driscoll N, Tibor M, Sprau G, Augustine J, Larson T,
113 of 158.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
41S
Stegall M, Textor S: Changes in hemodynamic patterns with age
in normotensive subjects (Abstract). Am J Hypertens 2004;17:
373.
Hinderliter AL, Sherwood A, Blumenthal JA, Light KC, Girdler SS,
McFetridge J, Johnson K, Waugh R: Changes in hemodynamics and
left ventricular structure after menopause. Am J Cardiol 2002;89:
830 – 833.
Galarza CR, Alfi J, Waisman GD, Mayorga LM, Camera LA, del
Rio M, Vasvari F, Limansky R, Farias J, Tessler J, Camera MI:
Diastolic pressure underestimates age-related hemodynamic impairment. Hypertension 1997;30:809 – 816.
Alfi J, Gabriel D, Waisman G, Camera MI: Contribution of stroke
volume to the change in pulse pressure pattern with age. Hypertension 1999;34:808 – 812.
Hartley TR, Lovallo WR,Whitsett TL: Cardiovascular effects
of caffeine in men and women. Am J Cardiol 2004;93:
1022–1026.
Yu BH, Nelesen R, Ziegler MG, Dimsdale JE: Mood states and
impedance cardiography-derived hemodynamics. Ann Behav Med
2001;23:21–25.
Hinderliter A, Blumenthal J, Waugh R, Chilukuri M, Sherwood A:
Ethnic differences in left ventricular structure: relations to hemodynamics and diurnal blood pressure variation. Am J Hypertens 2004;
17:43– 49.
Calhoun DA, Mutinga ML, Collins AS, Wyss JM, Oparil S: Normotensive blacks have heightened sympathetic response to cold
pressor test. Hypertension 1993;22:801– 805.
Fagard RH, Pardaens K, Staessen JA, Thijs L: Prognostic value of
invasive hemodynamic measurements at rest and during exercise in
hypertensive men. Hypertension 1996;28:31–36.
Fagard RH, Pardaens K, Staessen JA, Thijs L: The pulse pressure-to-stroke index ratio predicts cardiovascular events and
death in uncomplicated hypertension. J Am Coll Cardiol 2001;
38:227–31.
de Simone G, Roman MJ, Koren MJ, Mensah GA, Ganau A,
Devereux RB: Stroke volume/pulse pressure ratio and cardiovascular risk in arterial hypertension. Hypertension 1999;33:800 –
805.
Janeway TC: A clinical study of hypertensive cardiovascular disease. Arch Intern Med 1913;12:755–798.
Ganau A, Devereux RB, Roman MJ, de Simone G, Pickering TG,
Saba PS, Vargiu P, Simongini I, Laragh JH: Patterns of left ventricular hypertrophy and geometric remodeling in essential hypertension. J Am Coll Cardiol 1992;19:1550 –1558.
Kahan T: The importance of left ventricular hypertrophy in human
hypertension. J Hypertens 1998;16(Suppl):S23–S29.
Verdecchia P, Schillaci G, Borgioni C, Ciucci A, Gattobigio R,
Zampi I, Santucci A, Santucci C, Reboldi G, Porcellati C: Prognostic value of left ventricular mass and geometry in systemic hypertension with left ventricular hypertrophy. Am J Cardiol 1996;78:
197–202.
Verdecchia P, Carini G, Circo A, Dovellini E, Giovannini E, Lombardo M, Solinas P, Gorini M, Maggioni AP: Left ventricular mass
and cardiovascular morbidity in essential hypertension: the MAVI
study. J Am Coll Cardiol 2001;38:1829 –1835.
Vakili BA, Okin PM, Devereux RB: Prognostic implications of left
ventricular hypertrophy. Am Heart J 2001;141:334 –341.
Ichkhan K, Molnar J, Somberg J: Relation of left ventricular mass
and QT dispersion in patients with systematic hypertension. Am J
Cardiol 1997;79:508 –511.
Gomez-Cerezo J, Rios Blanco JJ, Suarez Garcia I, Moreno Anaya P,
Garcia Raya P, Vazquez-Munoz E, Barbado Hernandez FJ: Noninvasive study of endothelial function in white coat hypertension.
Hypertension 2002;40:304 –309.
London GM, Cohn JN: Prognostic application of arterial stiffness:
task forces. Am J Hypertens 2002;15:754 –758.
42S
67. Breithaupt-Grogler K, Gerhardt G, Lehmann G, Notter T, Belz GG:
Blood pressure and aortic elastic properties—verapamil SR/trandolapril compared to a metoprolol/hydrochlorothiazide combination
therapy. Int J Clin Pharmacol Ther 1998;36:425– 431.
68. Glasser SP: Hypertension syndrome and cardiovascular events.
Postgrad Med 2001;110:29 –35.
69. Intengan HD, Schiffrin EL: Vascular remodeling in hypertension:
roles of apoptosis, inflammation and fibrosis Hypertension 2001;
38:581–587.
70. Mathew J, Sleight P, Lonn E, Johnstone D, Pogue J, Yi Q, Bosch J,
Sussex B, Probstfiel J, Yusuf S: Reduction of cardiovascular risk
by regression of electrocardiographic markers of left ventricular
hypertrophy by the angiotensin-converting enzyme inhibitor
ramipril. Circulation 2001;104:1615–1621.
71. Corea L, Bentivoglio M, Verdecchia P, Providenza M, Motolese M:
Left ventricular hypertrophy regression in hypertensive patients
treated with metoprolol. Int J Pharmacol Ther Toxicol 1984;22:365–
370.
72. Yoshitomi Y, Nishikimi T, Abe H, Nagata S, Kuramochi M, Matsuoka H, Omae T: Left ventricular systolic and diastolic function
and mass before and after antihypertensive treatment in patients
with essential hypertension. Hypertens Res 1997;20:23–28.
73. Ofil EO, Cohen JD, St Vrain JA, Pearson A, Martin TJ, Uy ND,
Castello R, Labovitz AJ: Effect of treatment of isolated systolic
hypertension on left ventricular mass. J Am Med Assoc 1998;279:
778.
74. Verdecchia P, Angeli F, Borgioni C, Gattobigio R, de Simone G,
Devereux RB, Porcellati C: Changes in cardiovascular risk by
reduction of left ventricular mass in hypertension: a meta-analysis.
Am J Hypertens 2003;16:895– 899.
75. Galarza C, Alfi J, Waisman G, Burgos M, Zaniello G, del Rio M,
Serra L, Toselli P, Vasvari F, Camera M: Severe systemic hemodynamic impairment in patients with stroke (Abstract). Am J Hypertens 1996;9(Suppl 1):172A.
76. Johns DW, Peach MJ: Factors that contribute to resistant forms of
hypertension. Pharmacological considerations. Hypertension 1988;
11(Suppl II):II-88 –II-95.
77. Tsukiyama H: Choice of antihypertensive agents in hemodynamic
aspects to match pathophysiology and pharmacology in essential
hypertension. Jpn J Med 1989;28:261–264.
78. Zusman RM: Left ventricular function in hypertension. Relevance
to the selection of antihypertensive therapy. Am J Hypertens 1989;
2:200S–206S.
79. Sullivan JM, Schoeneberger TE, Ratts ET, Palmer ET, Samaha JK,
Mance CJ, Muirhead EE: Short-term therapy of severe hypertension. Hemodynamic correlates of the antihypertensive response in
man. Arch Intern Med 1979;139:1233–1239.
80. Easterling TR, Benedetti TJ, Schmucker BC, Carlson KL: Antihypertensive therapy in pregnancy directed by noninvasive hemodynamic monitoring. Am J Perinatol 1989;6:86 – 89.
81. Resnick LM, Lester MH: Differential effects of antihypertensive
drug therapy on arterial compliance. Am J Hypertens 2002;15:
1096 –1100.
82. Weber K, Bohmeke T, van der Does R, Taylor SH: Hemodynamic
differences between metoprolol and carvedilol in hypertensive patients. Am J Hypertens 1998;11:614 – 617.
83. ALLHAT Coordinators and Officers Major outcomes in high-risk
hypertensive patients randomized to angiotensin-converting enzyme
inhibitor or calcium channel blocker vs. diuretic. The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial
(ALLHAT). J Am Med Assoc 2002;288:2981–2997.
84. Yakovlevitch M, Black HR: Resistant hypertension in a tertiary care
clinic. Arch Intern Med 1991;151:1786 –1792.
85. Houston MC: Hypertension strategies for therapeutic intervention and prevention of end-organ damage. Prim Care 1991;18:
713–753.
86. Taler SJ, Augustine J, Textor SC: A hemodynamic approach to
resistant hypertension. Congest Heart Fail 2000;6:90 –93.
114 of 158.
87. Bailey JM, Levy JH, Kopel MA: Relationship between clinical
evaluation of peripheral perfusion and global hemodynamics in
adults after cardiac surgery. Crit Care Med 1990;18:1353–
1356.
88. Tibby SM, Hatherill M, Marsh MJ, Murdoch IA: Clinicians’ abilities to estimate cardiac index in ventilated children and infants. Arch
Dis Child 1997;77:516 –518.
89. Linb G, Eisenberg BM: Noninvasive techniques for evaluation of
heart function and hemodynamics in arterial hypertension. Acta
Cardiol 1990;45:133–139.
90. Abdelhamed AI, Levy P, Smith KS, Wilkins J, Smith R, Ferrario
CM: Value of non-invasive hemodynamic measurements in hypertensive patients. Am J Hypertens 2002;15(Suppl 1):A89.
91. de Divitiis O, Di Somma S, Liguori V, Petitto M, Magnotta C,
Ausiello M, Natale N, Brignoli M, Galderisi M: Effort blood pressure control in the course of antihypertensive treatment. Am J Med
1989;87:(Suppl 3C)46S–56S.
92. Eliot RS: The dynamics of hypertension—an overview: present
practices, new possibilities, and new approaches. Am Heart J 1988;
116:583–589.
93. Taler SJ, Augustine J, Burnett JC, Textor SC: Hemodynamic and
volume changes during intensive treatment (Rx) for resistant hypertension (ResHTN) (Abstract). Am J Hypertens 2000;13:61A– 62A.
94. Lang M: Noninvasive hemodynamic-driven care management.
Achieving quality patient outcomes across a continuum of care.
Cardiovasc Dis Manage 2001;71–73.
95. Vuurmans T, Boper P, Koomans H: Effects of endothelin-1 and
endothelin-1 receptor blockade on cardiac output, aortic pressure,
and pule wave velocity in humans. Hypertension 2003;41:1253–
1258.
96. Margulis F, Dalton R, Killinger C, Pisarello J, Alvarez C: Decreased
myocardial contractility and increased thoracic flui content detected by impedance cardiography in untreated hypertensive patients
(Abstract). Am J Hypertens 1997;10(Suppl 1):55A.
97. Petersen JR, Jensen BV, Drabaek H, Viskum K, Mehlsen J: Electrical impedance measured changes in thoracic flui content during
thoracentesis. Clin Physiol 1994;14:459 – 466.
98. Ebert TJ, Smith JJ, Barney JA, Merrill DC, Smith GK: The use of
thoracic impedance for determining thoracic blood volume changes
in man. Aviat Space Environ Med 1986;57:49 –53.
99. Mattar JA, Salas CE, Bernstein DP, Lehr D, Bauer R: Hemodynamic changes after an intensive short-term exercise and nutrition
program in hypertensive and obese patients with and without coronary disease. Arq Bras Cardiol 1990;54:307–312.
100. Toth PD, Demeter RJ, Woods, JR, Nyhuis AW, Judy WV:
Comparison of the effects of pindolol and atenolol on hemodynamic function in systemic hypertension. Am J Cardiol 1988;62:
413– 418.
101. Koga Y, Gillum RF, Kubicek WG: An impedance cardiographic
study of the mechanism of blood pressure fall after moderate dietary
sodium restriction. Jpn Heart J 1985;26:197–207.
102. Ashida T, Nishioeda Y, Kimura G, Kojima S, Kawamura M,
Imanishi M, Abe H, Kawano Y, Yoshimi H, Yoshida K, Kuramochi M, Omae T: Effects of salt, prostaglandin, and captopril on
vascular responsiveness in essential hypertension. Am J Hypertens 1989;2:640 – 642.
103. Taler SJ, Textor SC, Augustine JE: Resistant hypertension: comparing hemodynamic management to specialist care. Hypertension
2002;39:982–988.
104. Sharman DL, Gomes CP, Rutherford JP: Improvement in blood
pressure control with impedance cardiograph-guided pharmacologic
decision making. Congest Heart Fail 2004;10:54 –58.
105. Sramek BB, Tichy JA, Hojerova M, Cervenka V: Normohemodynamic goal-oriented antihypertensive therapy improves the outcome. Am J Hypertens 1996;9:141A.
106. Parrott C, Burnham K, Quale C, Lewis D: Comparison of changes
in ejection fraction to changes in impedance cardiography cardiac
index and systolic time ratio. Congest Heart Fail 2004;10(2 Suppl
2):11–13.
107. Thomson B, Drazner M, Dries D, Yancy C: Is impedance cardiography-derived systolic time ratio a useful method to determine left
ventricular systolic dysfunction in heart failure (Abstract)? J Card
Failure 2004;10(Suppl 4):S38.
108. Ross R: Atherosclerosis—an inflammator disease. N Engl J Med
1999;340:115–126.
115 of 158.
43S
109. Panza JA: High-normal blood pressure—more “high” than “normal.”N Engl J Med 2001;345:1337–1340.
110. Dagenais GR, Yusuf S, Bourassa MG, Yi Q, Bosch J, Lonn EM,
Kouz S, Grover J: Effects of ramipril on coronary events in high-risk
persons: results of the Heart Outcomes Prevention Evaluation
Study. Circulation 2001;104:522–526.
111. Herrera-Acosta J, Perez-Grovas H, Fernandez M, Arriaga J: Enalapril in essential hypertension. Drugs 1985;30(Suppl 1):35– 46.
European Journal of Heart Failure, June 2004
Whole-Body Electrical Bio-Impendance is accurate in Non Invasive Determination of
Cardiac Output: A Thermodilution controlled, Prospective, Double Blind Evaluation.
1
2
2
2
3
Guillermo Torre-Amiot MD, Gad Cotter MD, Zvi Vered MD, Edo Kaluski MD, Karl Stang MD.
1
2
From Baylor College of Medicine, Houston, TX, USA ( ), Assaf-Harofeh Medical Center, Israel ( )
3
and Charite Campus, Berlin, Germany ( ).
TM
Background: The NICaS is a novel non-invasive apparatus based on whole body electrical
bio-impedance for simple non-invasive continuous CO determination.
Patients and Methods: Patients were recruited while randomized in a study evaluating the
efficacy of Tezosentan (a ET-A/B endothelin antagonist) in patients admitted due to acute heart
failure (CHF). Patients were randomized after having been hospitalized due to acute heart failure
2
with dyspnea at rest, CI < 2.5 L/min/m and PCWP ≥ 20 mmHg. Study Protocol: At baseline and
during treatment with study drug at the pre-specified time points of 0.5,1,2,3,4 and 6 hours from
TM
randomization CO was determined by both thermodilution and the NICaS 2001 apparatus. At
TM
each time point CO was determined by thermodilution and NICaS 2001 apparatus by a two
independed, blinded operators.
Results: Out of 130 patients enrolled, in 93 CO was measured simultaneously by both methods
at all the pre-determined time points. The overall Correlation between the two methods was
R=0.81 (Figure). Precision and bias were 0.010.6 L/min. There was a difference between the
2
two methods in cardiac output readings. When Mean CI (of both methods) was < 2 L/min/M CO
2
readings were statistically significantly lower by NICaS while when CI was >3 L/min/M , CO
readings were statistically significantly higher by NICaS. We have calculated the cardiac power
index (Cpi=CI* mean arterial pressure), and found that low Cpi (indicating reduced myocardial
contractile reserve) was related to higher recurrent CHF. However, Cpi based on NICaS CI
measurement (NICaS Cpi) was a better predictor of recurrent CHF then thermodilution Cpi (Th
Cpi), due to less accurate prediction in patients with high Cpi.
Conclusions: NICaS is a novel accurate non-invasive method for CO determination. The results
of the present study suggest that NICaS might be more accurate then thermodilution for CO
determination due to the tendency of thermodilution to under estimate CO when high and over
estimate it when low.
116 of 158.
Accurate, Noninvasive Continuous
Monitoring of Cardiac Output by Whole-Body
Electrical Bioimpedance
Gad Cotter, Yaron Moshkovitz, Edo Kaluski, Amram J. Cohen, Hilton
Miller, Daniel Goor and Zvi Vered
Chest 2004;125;1431-1440
DOI 10.1378/chest.125.4.1431
The online version of this article, along with updated information
and services can be found online on the World Wide Web at:
http://chestjournals.org/cgi/content/abstract/125/4/1431
CHEST is the official journal of the American College of Chest
Physicians. It has been published monthly since 1935. Copyright 2007
by the American College of Chest Physicians, 3300 Dundee Road,
Northbrook IL 60062. All rights reserved. No part of this article or PDF
may be reproduced or distributed without the prior written permission
of the copyright holder
(http://www.chestjournal.org/misc/reprints.shtml). ISSN: 0012-3692.
117 of 158.
Downloaded from chestjournals.org on April 4, 2007
Copyright © 2004 by American College of Chest Physicians
clinical investigations in critical care
Accurate, Noninvasive Continuous
Monitoring of Cardiac Output by WholeBody Electrical Bioimpedance*
Gad Cotter, MD; Yaron Moshkovitz, MD; Edo Kaluski, MD;
Amram J. Cohen, MD, FCCP; Hilton Miller, MD†; Daniel Goor, MD; and
Zvi Vered, MD
Study objectives: Cardiac output (CO) is measured but sparingly due to limitations in its measurement technique (ie, right-heart catheterization). Yet, in recent years it has been suggested that CO
may be of value in the diagnosis, risk stratification, and treatment titration of cardiac patients,
especially those with congestive heart failure (CHF). We examine the use of a new noninvasive,
continuous whole-body bioimpedance system (NICaS; NI Medical; Hod-Hasharon, Israel) for measuring CO. The aim of the present study was to test the validity of this noninvasive cardiac output
system/monitor (NICO) in a cohort of cardiac patients.
Design: Prospective, double-blind comparison of the NICO and thermodilution CO determinations.
Patients: We enrolled 122 patients in three different groups: during cardiac catheterization (n ⴝ 40);
before, during, and after coronary bypass surgery (n ⴝ 51); and while being treated for acute
congestive heart failure (CHF) exacerbation (n ⴝ 31).
Measurements and intervention: In all patients, CO measurements were obtained by two independent blinded operators. CO was measured by both techniques three times, and an average was
determined for each time point. CO was measured at one time point in patients undergoing coronary catheterization; before, during, and after bypass surgery in patients undergoing
coronary bypass surgery; and before and during vasodilator treatment in patients treated for acute
heart failure.
Results: Overall, 418 paired CO measurements were obtained. The overall correlation between the
NICO cardiac index (CI) and the thermodilution CI was r ⴝ 0.886, with a small bias (0.0009 ⴞ 0.684
L) [mean ⴞ 2 SD], and this finding was consistent within each group of patients. Thermodilution
readings were 15% higher than NICO when CI was < 1.5 L/min/m2, and 5% lower than NICO when
CI was > 3 L/min/m2. The NICO has also accurately detected CI changes during coronary bypass
operation and vasodilator administration for acute CHF.
Conclusion: The results of the present study indicate that whole-body bioimpedance CO measurements obtained by the NICO are accurate in rapid, noninvasive measurement and the follow-up of
CO in a wide range of cardiac clinical situations.
(CHEST 2004; 125:1431–1440)
Key words: cardiac function test; cardiac output; congestive heart failure
Abbreviations: CABG ⫽ coronary artery bypass graft; CHF ⫽ congestive heart failure; CI ⫽ cardiac index; CO ⫽ cardiac
output; Cpi ⫽ cardiac power index; ISDN ⫽ isosorbide-dinitrate; MAP ⫽ mean arterial BP; NICO ⫽ noninvasive cardiac
output system/monitor; SV ⫽ stroke volume; SVRi ⫽ systemic vascular resistance index; TEB ⫽ thoracic electrical bioimpedance; WBEB ⫽ whole-body electrical bioimpedance
of cardiac output (CO) and the calM easurement
culation of cardiac index (CI) has been used
selectively over the last 2 decades, mainly due to the
fact that CI measurement requires the invasive procedure of right-heart catheterization and placement of a
Swan-Ganz catheter (Baxter Healthcare; Irvine, CA).
www.chestjournal.org
118 of 158.
Hence, the experience with its application for monitoring and risk stratification of cardiac patients is limited.
In recent years, however, evidence has accumulated to the effect that CI measurement and the
calculation of systemic vascular resistance index
(SVRi) and cardiac power index (Cpi, the product of CI
CHEST / 125 / 4 / APRIL, 2004
1431
multiplied by mean arterial BP [MAP]) measured
simultaneously might be instrumental in the monitoring and risk stratification of cardiac patients, especially
those with acute and chronic congestive heart failure
(CHF) and patients admitted with cardiogenic shock.1
In cardiogenic shock, a recent analysis of the SHOCK
(SHould We Emergently Revascularize Occluded Coronaries in Cardiogenic ShocK) registry data has shown
Cpi to be the strongest independent predictor of
in-hospital mortality,2 while in acute heart failure Cpi
was found to be an important tool for diagnosis and risk
stratification.3,4 In patients with chronic CHF, a few
studies5–7 have shown that noninvasive Cpi reserve (the
increase in Cpi during exercise or dobutamine stress) is
the strongest predictor of outcome (a better oxygen
consumption and echocardiographic ejection fraction)
in such patients. Furthermore, changes in acute SVRi
may be useful for early detection of myocardial ischemia.8 Moreover, in two separate studies9 –11 examining
the efficacy of vasodilators in patients with acute CHF,
medication was found to be effective mainly in patients
who were submitted to hemodynamic monitoring, implying that perhaps in order to be efficacious, vasodilator treatment should be monitored attentively to
prevent overtreatment and undertreatment. In different studies,12,13 we have also demonstrated that careful
titration of vasodilator treatment administered for acute
heart failure and acute coronary syndromes is important to optimize its efficacy. In the present study,
we evaluated the accuracy of a novel method of CI
measurement (whole-body electrical bioimpedance
[WBEB]) in different cardiac clinical settings (during
cardiac catheterization and coronary artery bypass graft
[CABG] surgery, and for monitoring patients with
acute CHF) and over a wide range of CI values and
severity of left ventricular dysfunction.
Materials and Methods
Patient Populations
Group 1: Group 1 consisted of 40 patients with coronary artery
disease referred during March to July 1993 for left and right
*From the Cardiology Department (Drs. Cotter, Kaluski, and
Vered), Assaf–Harofeh Medical Center, Zerifin; the Cardiac
Surgery Department (Drs. Moshkovitz and Goor), Sackler School
of Medicine, Tel Aviv University, Tel Aviv; the Department of
Cardiology (Dr. Miller), Sourasky Medical Center, Tel-Aviv; and
the Department of Cardiac Surgery (Dr. Miller), the Edith
Wolfson Medical Center, Holon, Israel.
†Deceased.
Manuscript received March 19, 2003; revision accepted September 19, 2003.
Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail:
[email protected]).
Correspondence to: Gad Cotter, MD, Cardiology Department,
Assaf-Harofeh Medical Center, 70300, Zerifin, Israel; e-mail:
[email protected]
cardiac catheterization based on conventional clinical indications.
During the right-heart study, a pulmonary artery catheter was
introduced under fluoroscopy; at a single time point, a paired
measurement of CI by a noninvasive cardiac output system/
monitor (NICO) [NICaS; NI Medical; Hod-Hasharon, Israel]
and by thermodilution was performed.
Group 2: Group 2 included 51 patients undergoing CABG
operations. The first 15 patients were studied at Wolfson Medical
Center (Israel) during October to November 1994. The next 16
patients were studied at Johns Hopkins Medical Center during
April to May 1995. The remaining 20 patients were studied again
at Wolfson Medical Center during August to October 1995. A
balloon-guided, Swan-Ganz catheter was introduced after the
induction of anesthesia, and five paired NICO CI and thermodilution CI measurements were obtained at specific operative and
postoperative stages: immediately prior to the skin incision; after
sternotomy; after pericardiotomy; 10 min after weaning from the
pump; and immediately after arrival to the ICU. The results
obtained from the first 31 cases of this series have already been
published.14
Group 3: Group 3 consisted of 31 patients admitted during
September to December 2001 to the ICU of Assaf–Harofeh
Medical Center (Israel) because of an acute exacerbation of
CHF. Prior to admittance, they underwent a right-heart study in
the catheterization laboratory for the assessment of their cardiac
condition, and a Swan-Ganz catheter was inserted under fluoroscopy. CO measurement was begun on arrival to the ICU, where
three baseline measurements were obtained, 15 min apart. In 17
patients who required vasodilator therapy, four measurements
were obtained during the initiation and up-titration of IV isosorbide-dinitrate (ISDN) treatment. In the 14 patients who did not
require ISDN treatment, an additional (fourth) baseline measurement was obtained. All the studies were approved by the
review boards and the Helsinki committees of the various
hospitals. Consents for the studies were obtained from each
patient.
Measuring CI
Thermodilution: In all three study groups, a No. 7F Swan-Ganz
balloon flotation catheter was placed in the pulmonary artery. In
groups 1 and 2, the Swan-Ganz catheter was introduced in the
catheterization room, and in group 2 on the operating table,
following anesthesia. In the 16 patients of group 2 who were
studied at Johns Hopkins Medical Center, CO measures were
obtained by the 7010 Series Marquette (Marquette; Madison,
WI). In the remaining 35 patients of group 2 studied at Wolfson
Medical Center, and also in group 1 patients at the Tel Aviv
Medical Center (Israel), the Horizon 2000 (Mennen Medical;
Rohovot, Israel) was used. In group 3, the CO was measured by
Marquette 8000 Clinical Information Center 419897-015 (Marquette).
A volume of 10 mL of 5% dextrose at room temperature was
injected in all patients via the proximal port. Temperature
changes were measured via the distal port located in the pulmonary artery, ascertained by fluoroscopy, oxygen saturation, and
wedge pressure measurements. In all patients, three CI measurements were obtained; when a ⬎ 15% disparity occurred between
the two extreme measurements, two further injections, or more,
were administered until an average of three measurements within
the 15% range was obtained.
NICO WBEB Technology
When transmitting a small electrical current through the body,
an impedance to its transmission (restivity, R) is being measured.
1432
119 of 158.
Clinical Investigations in Critical Care
This resistivity is called bioimpedance. According to Kirchov’s
law, electric current passes through conduits of higher conductance (lowest resistivity). The resistivity of blood and plasma is
the lowest in the body (resistivity of blood is 150; plasma, 63;
cardiac muscle, 750; lungs, 1,275; and fat, 2,500 ohm/cm).15
Thus, when an alternating current of 20 to 100 kHz is applied to
the body, it is primarily distributed via the extracellular fluid and
the blood. The changes in the body resistivity (⌬R) over time
(milliseconds) are therefore related to the dynamic changes of
the blood and plasma volume. This pertains particularly to the
passage of the stroke volume (SV) from the left ventricle into
the aorta and its branches. However, in the capillaries and in the
venous system the blood volume is relatively constant, because
the flow in these vessels is not pulsatile. Consequently, each
systolic increase in the aortic blood volume is associated with a
proportional increase in the measurable conductance of the
whole body (Fig 1).16 Thus, for measuring the aortic SV by means
of its impedance change, Frinerman and Tsoglin developed the
following algorithm14:
SV ⫽
H2corr ⌬R ␣ ⫹ ␤
Hctcorr
⫻ Kel ⫻ Kweight ⫻ IB ⫻
⫻
Ksex, age
R
␤
in which the ⌬R/R is corrected for hematocrit (Hctcorr), electrolytes (Kel), body composition (K sex, age), weight (Kweight),
time characteristics (␣ ⫽ systolic time, ␤ ⫽ diastolic time), and
index balance (IB), which measures the body water composition.
To collect patient signals, the NICO uses proprietary electrodes arranged in a wrist-to-ankle configuration (Fig 2); in
certain conditions when this particular form is not applicable (as
with severe peripheral edema or severe peripheral vascular
disease), a wrist-to-wrist configuration is used. The precise
positioning of the NICO electrodes is not critical; an untrained
operator can make the attachments. An alternating current of 30
kHz, 1.4 mA is delivered through the two electrodes, and the
bioimpedance and its fluctuations over time are measured. In
addition, a standard three-lead ECG connection is made for
measuring the pulse rate. The other variables required for SV and
CI calculation (age, gender, weight, height, hematocrit, electrolytes) are introduced into the machine only at the start of
monitoring. For measuring the CO, the SV is multiplied by the
heart rate.
Although the idea of the WBEB was conceived and tested in
the Soviet Union in 1941,17 it remained dormant until recently. Meanwhile another bioimpedance approach for measuring the CO was initially introduced by Kubicek et al18 in the
United States in 1966, and further refined in 1974.19 His
method is called thoracic electrical bioimpedance (TEB), and
the tools of this approach are commercially available. Noted
here are dissimilarities in the operation of the two technologies. In the NICO, one electrode is applied to the wrist and
the other to the ankle; in TEB, a number of electrodes are
placed at the root of the neck and another set around the lower
part of the chest cage. In WBEB, the SV is measured by the
impedance variation (⌬Z or ⌬R) induced by the systolic
volume ejected into the aorta. In the original TEB formula of
1966,18 the SV measurement was based on a similar principle.
However, since when the electrodes are placed on the chest,
the ⌬Z wave could be hardly detected, Kubicek et al19 adopted
another principle, whereby the SV measurement is based on
the depiction of the aortic systolic dp/dt (instantaneous pressure change over time) for calculation of the systolic blood
flow into the great arteries.
Figure 1. Recordings of the interrelation between impedance variations and hemodynamic parameters, according to Djordjevich and Sadove.16
120 of 158.
CHEST / 125 / 4 / APRIL, 2004
1433
Figure 2. Wrist-to-ankle configuration of the electrodes in WBEB. I ⫽ electric current; V ⫽ electric
voltage.
Measuring CI by the NICO
Results
For each CI determination, three NICO measurements were
obtained. Since the CO results that are displayed on the scope are
updated every 20 s, for the determination of the CI an average
CO is derived from a 60-s monitoring.
Statistical Methods
Agreement between the CI values of the NICO and thermodilution was evaluated in three ways: the mean CI values of the
NICO and thermodilution were compared by a paired Student t
test; correlation between these values was evaluated by calculating the Pearson correlation coefficient and by applying a linear
regression model of the NICO on thermodilution; the differences
between the paired CI values of the NICO and thermodilution
were plotted against the average CI values of both methods,
instead of against thermodilution alone. This statistical method
was recommended by Bland and Altman20 for evaluating a new
device (NICaS) against an established method (thermodilution),
which has its own inaccuracies. Bias was defined as the mean
difference between the NICO CI and thermodilution CI values.
Limits of agreement (precision) were calculated as bias ⫾ 2 SD
of the differences between the NICO CI and thermodilution CI
values. All three analyses were carried for the whole sample and
for each specific clinical group: cardiac catheterization patients
(group 1), CABG patients (group 2), and CHF patients (group 3).
We have examined the differences between the CI determination by the NICO and thermodilution at different CI readings
by classifying the readings into four subgroups according to the
mean CI levels: CI ⬍ 1.5, CI ⱖ 1.5 and ⬍ 2, and CI ⱖ 2 and ⬍ 3,
and CI ⱖ 3. CI readings by the NICO and thermodilution in each
group were compared using the paired Student t test and
presented as mean and SDs.
The sensitivity of the two methods (NICO and thermodilution)
to a change in a specific medical condition was compared in the
CABG group at five operative and postoperative stages. A
variance analysis with repeated measures (type of method and
stage) was performed, followed by contrast analysis that compared successive stages. This analysis could be performed only for
patients with complete data at all the five stages. A similar
analysis was performed at seven time points in CHF patients who
had been treated with an IV vasodilatory drug. All statistical
analyses were performed using the SAS System for Windows
(version 8.01; SAS Institute; Cary, NC).
No significant differences between the means of
NICO CI and thermodilution CI in the three clinical
groups, as well as the whole cohort, were observed
(Table 1). A significant, high correlation was found
between the NICO CI and the thermodilution CI
measurements: 0.886 in the whole cohort, and 0.881,
0.902, and 0.851 in the catheterization, bypass surgery, and CHF groups, respectively. All correlation
coefficients were statistically significant (p ⬍ 0.0001).
The results of applying linear regression models to
the data (Table 2, Fig 3) demonstrate similar models
in the three clinical groups, as the intercepts and
slopes of the regression lines are not significantly
different (intercepts, p ⫽ 0.2398; slopes, p ⫽ 0.2310).
Figure 4 shows differences between CI values plotted against the average value of the two methods
with limits of agreement: two SDs from the mean
difference.
Significant differences between the NICO CI and
thermodilution CI were observed when comparing
the average CI of the four CI ranges (Table 3). When
CI is ⬍ 2 L/min/m2, the thermodilution CI readouts
Table 1—Comparison Between the Mean CI Values of
the NICO and Thermodilution in the Three Clinical
Groups and in the Whole Cohort*
Thermodilution
CI, L/min/m2
Group
No.
Mean
SD
Mean
SD
p Value
Whole sample
Catheterization
CABG
CHF
418
40
208
170
2.39
2.81
2.33
2.35
0.70
0.72
0.72
0.63
2.38
2.81
2.31
2.38
0.73
0.68
0.77
0.66
NS
NS
NS
NS
*NS ⫽ not significant.
1434
121 of 158.
NICO CI,
L/min/m2
Table 2—Linear Regression Analysis, Bias, and Precision in the Three Clinical Groups and in the Whole Cohort
Sample
Intercept
Slope
R2
Bias (Mean BetweenMethod Difference),
L/min
Whole
Catheterization
CABG
CHF
0.18
0.45
0.08
0.29
0.92
0.84
0.96
0.89
0.79
0.77
0.81
0.72
⫺ 0.0009
0.0040
⫺ 0.0247
0.0271
are significantly higher than the NICO CI; when CI
is ⬎ 3 L/min/ m2, the thermodilution CI is lower
than the NICO CI (Table 3). When CI was between
2 L/min/m2 and 3 L/min/m2, there was a slight
difference of only 3.28% between the two methods,
with a borderline significance (Table 3).
Twenty-five patients with CABG (subgroup A)
had complete information of the NICO CI and
thermodilution CI results at five operative and postoperative stages (Table 4, Fig 5, top). A further 26
CABG patients (subgroup B) had incomplete data
(⬍ 5 paired measurements for each patient), yielding
additional 80 paired measurements (Table 4). There
were time-related changes in the CI, and the NICO
and the thermodilution followed these changes
by producing similar results at each time point
(p ⫽ 0.0035 and p ⫽ 0.0058, respectively; Fig 5,
top). A contrast analysis, performed to compare the
CO in successive stages of the measurements, found
Precision (Mean ⫾ SD),
L/min
⫺ 0.6849 ⫾ 0.6831
⫺ 0.7134 ⫾ 0.6393
⫺ 0.6789 ⫾ 0.7331
⫺ 0.684 ⫾ 0.692
that the difference between stage 3 and stage 4 was
statistically significant according to the two measurements in subgroup A (NICO CI, p ⫽ 0.0135; thermodilution CI, p ⫽ 0.002; Fig 5, top). In 17 of the
CHF patients in group 3 who were treated with
an IV vasodilator agent, and in whom the CI was
measured simultaneously at seven time points during
the treatment, the total time trend was significant in
the NICO (p ⫽ 0.0056) but not significant in thermodilution (Fig 5, bottom).
Discussion
In recent years it has been suggested that CO and
MAP measurement and the calculation of CI, Cpi,
and SVRi might be instrumental in the diagnosis,
treatment titration, and risk stratification of cardiac
patients, especially those with CHF.7–10 However,
Figure 3. Plot of CI values measured by the NICO and thermodilution.
122 of 158.
CHEST / 125 / 4 / APRIL, 2004
1435
Figure 4. Differences between CI values (NICO vs thermodilution) plotted against the average values
of the two methods.
CI has been measured only during invasive rightheart catheterization, which requires intensive care
admission and may be associated with complications.21–23 Hence, CI was measured only rarely, and
in the sickest patients. Therefore, a simple, reliable,
noninvasive, and continuous method for CI measurement has become necessary in order to enable
its application to cardiac patients with different
degrees of medical severity and in diverse settings.
Currently there are four accepted methods for
noninvasive CO measurement. The Doppler echocardiogram obtained from the left ventricular outflow track and CO2 rebreathing techniques have
been shown to be accurate in measuring CI. But
these methods are limited by the requirement for
expensive equipment and specialized personnel for
their application and therefore are not simple to use,
and moreover do not enable continuous measurements. Thoracic bioimpedance has been used in the
last decade for continuous CO measurement. Judg-
ing by the literature,24 –26 as long as the heart function is intact, TEB can be useful for monitoring the
hemodynamic state in various clinical conditions
such as trauma, massive surgery, sepsis, etc.24 But
when it comes to monitoring and managing pathologic cardiac conditions TEB requires further improvement.24 –26
Thus far, eight groups of patients who submitted
to CO measurements by WBEB have been reported
in six published articles. Kedrov,17 who was the first,
compared the average CI measured by the WBEB
in 57 subjects with normal hearts in published results
of the Fick method, revealing 3.3 ⫾ 28% vs 3.31
L/min/m2 (range, 2.4 to 4.2 L/min/m2), respectively.
Tischenko27 compared the CI results measured by
WBEB in three groups of subjects with normal
hearts vs three standard methods. There were 31
cases vs acetylen (r ⫽ 0.84), 28 cases vs thermodilution (r ⫽ 0.95), and 28 cases vs Fick (r ⫽ 0.99).
Using a modified Tischenko algorithm vs thermodi-
Table 3—Differences Between the NICO CI and Thermodilution CI Within the CI Ranges
CI Ranges by NICO CI
Results,
No.
NICO CI,
Mean
Thermodilution
CI, Mean
Relative
Difference, %
Significance
⬍ 1.5
1.5 ⬍ CI ⬍ 2
2 ⬍ CI ⬍ 3
CI ⬎ 3
30
98
220
70
1.278 ⫾ 0.16
1.749 ⫾ 0.14
2.433 ⫾ 0.28
3.594 ⫾ 0.57
1.515 ⫾ 0.35
1.876 ⫾ 0.33
2.392 ⫾ 0.40
3.449 ⫾ 0.64
⫺ 12.99
⫺ 4.65
3.28
5.44
0.0002
⬍ 0.0001
0.0484
0.0045
1436
123 of 158.
Table 4 —Trend Follow-up of the NICO CI and Thermodilution CI During the Five Operative and
Postoperative Stages
NICO CI
Thermodilution CI
Subgroups A ⫹ B
(n ⫽ 51)
Subgroup A
(n ⫽ 25)
Subgroups A ⫹ B,
(n ⫽ 51)
Subgroup B,
(n ⫽ 26)
Method
Mean
SD
Mean
SD
Mean
SD
Mean
SD
1
2
3
4
5
2.28
2.09
2.17
2.72
2.56
0.87
0.58
0.62
0.79
0.74
2.25
2.09
2.26
2.54
2.46
0.83
0.57
0.69
0.81
0.85
2.22
2.14
2.15
2.80
2.61
0.69
0.57
0.68
0.74
0.83
2.22
2.14
2.22
2.61
2.53
0.67
0.52
0.70
0.77
0.87
lution, Koobi et al28 obtained simultaneous measurements in 74 patients with coronary disease, reaching
a bias between the two methods of 0.25 ⫾ 0.8 L/min
(SD), where the limits of agreement (2 SD) were
– 1.37 L/min and 1.897 L/min, respectively. Using
the NICaS apparatus, Cohen et al14 compared its
performance against thermodilution by measuring
the CO in patients undergoing CABG operations,
with a correlation of r ⫽ 0.91. Thus, this six-clinical
series, which included 274 subjects, revealed similar
correlation coefficients between the compared
methods, just as in the present series. Moreover, in
none of these publications did the authors express
reservations about the performance of the WBEB.
There were, nonetheless, two publications in
which no correlation was found between WBEB and
thermodilution; in both instances, the underlying
clinical conditions are listed in the exclusion criteria
of the NICO. Lamberts et al29 compared the original
Tischenko equation with dye dilution CO in 10
patients, 4 of whom had significant aortic regurgitation and 1 had coarctation of the aorta. The NICO
apparatus cannot measure the CO in such conditions.
Imhoff et al30 compared the NICO apparatus
against thermodilution in 22 postpancreatectomy or
esophagectomy patients. They were all managed
postoperatively by Swan-Ganz catheters for boosting
the oxygen delivery to 600 mL/min/m2 and the CI to
4.5 L/min/m2. Hence, in these patients the radical
surgical procedures were followed by massive intercompartmental volume shifts due to IV administration of up to 6 L per 24 h of crystalloids and plasma,
often accompanied by massive peripheral edema. In
such hemodynamic situations, the baseline impedance should properly become distorted, preventing
an accurate measurement of the systolic impedance
variation.
In the present study, the agreement between
NICO CI and thermodilution CI as tested by comparisons of the means is highly significant. The
similarity between the means in the entire cohort as
124 of 158.
well as in each diagnostic group, together with the
relatively large sample size, further endorses the
significance of the results.
The mean difference between the NICO and
thermodilution in the entire sample range was
0.0009 L/min (Table 2, Fig 4), ranging from 0.0040
to 0.0271 L/min in the three diagnostic groups. This
disparity is smaller than the level of accuracy of
thermodilution, which is defined to a 15% range.31
Linear regression applied to the data revealed that
the line slope was close to 1.00 in the entire sample
range and in each specific diagnostic group. There
were no significant differences between the slopes
and the intercepts of the three diagnostic groups.
This indicates that the relation between the NICO
and thermodilution is similar in all diagnostic groups.
Following the recommendations of Bland and Altman,20 the differences between the two measurements were plotted against their means. This plot
demonstrates that the range of differences is similar
along the different values of the average.
Although the main purpose of this work was to
compare the performance of the NICO vs thermodilution, following the suggestion in previous studies31–35 that thermodilution tends to overestimate CI
when low and underestimate it when high, we have
compared the NICO CI and thermodilution CI in
the different CI ranges. The results of this analysis
have shown that when the CI was ⬍ 2 L/min/m2, the
thermodilution results were higher than the NICO
results; when the CI was ⬎ 3 L/min/m2, the thermodilution results were slightly lower than the
NICO results (Table 3). As a consequence, the
differences of the hemodynamic responses to vasodilation therapy may be better depicted by the
NICO when compared with thermodilution (Fig 5,
bottom).
The link between a low CI and a higher thermodilution readout, and between a higher CI with a
low thermodilution readout, is also expressed in the
stability of the range of differences along the different values of the average (Fig 4). Furthermore, the
CHEST / 125 / 4 / APRIL, 2004
1437
Figure 5. Top: Comparisons between the NICO CI and thermodilution CI at five operative and
postoperative stages. Bottom: NICO CI and thermodilution CI responses to vasodilating titration
therapy in patients with acute CHF.
almost identical results of thermodilution and the
NICO observed in the CI ⬍ 2 to ⬍ 3 range (Table 3)
yield a mutual endorsement of the technical operations of the measuring devices by the two methods.
It should also be noted that the two methods switch
their interrelations at CI of approximately ⬍ 2.5
L/min/m2 (Fig 5, bottom), which is close to the lower
border.
Limitations of the NICO
Diseases of the Aorta and Aortic Valve: The NICO
measures the SV of the aorta and its derivatives
(including the peripheral arteries). For determining
the CO, SV is multiplied by the heart rate. Any
aberration in the hemodynamics of the aorta and its
derivatives may interfere with the proper function of
1438
125 of 158.
the NICO. In aortic insufficiency, the SV is pathologically increased. Coarctation and significant aneurysms of the aorta and also severe peripheral arterial
disease may produce inaccurate readouts of the SV.
Significant (⫹3/⫹4) Peripheral Edema: In significant edema, the baseline impedance of the body is
occasionally very low, currently interfering with the
accuracy of the results.
Shunts: Our experience with congenital heart
disease is limited, but we assume that since there is
no complete separation between the two circulations
the impedance technique may not be appropriate.
Restlessness: The impedance test requires a motionless condition; in addition, restlessness is associated with fluctuating activation of the sympathetic
system, resulting in an unstable hemodynamic state.
Arrhythmias: In most cases, there is no interference with the CO measurement, although when
associated with a heart rate ⬎ 150 beats/min or when
there is a severe disarray of the complexes (ECG and
SV), the results may become inaccurate, as may be
for any technique measuring CO.
Resections: Major abdominal and thoracic surgical
resections, especially those that include major rapid
shifts in fluid distribution between the intravascular
and extravascular space.
7
8
9
10
11
12
13
Conclusions
In spite of these limitations, the NICO apparatus
offers a simple, noninvasive, reliable, and continuous
measurement of CI in cardiac patients with particular emphasis on CHF. This measurement combined
with MAP measurement and the calculation of Cpi
and SVRi is destined to become a safe, simple, rapid,
noninvasive method for evaluating and treating primarily coronary patients sustaining chronic and acute
exacerbations of CHF.
14
15
16
17
18
References
1 Cotter G, Williams SG, Vered Z, et al. The role of cardiac
power in heart failure. Curr Opin Cardiol 2003; 18:215–222
2 Cotter G, Fincke R, Lowe A, et al. Hemodynamic parameters
in cardiogenic shock due to myocardial infarction: a report
from the SHOCK trial registry [abstract]. Circulation 2003;
108:539
3 Cotter G, Moshkovitz Y, Milovanov O, et al. Acute heart
failure: a novel approach to its pathogenesis and treatment.
Eur J Heart Fail 2002; 4:227–234
4 Cotter G, Moshkovitz Y, Kaluski E, et al. The role of cardiac
power and systemic vascular resistance in the pathophysiology
and diagnosis of patients with acute congestive heart failure.
Eur J Heart Fail 2003; 5:443– 451
5 Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with
severe heart failure. J Am Coll Cardiol 1997; 29:422– 428
6 Williams SG, Cooke GA, Wright DJ, et al. Peak exercise
126 of 158.
19
20
21
22
23
24
cardiac power output: a direct indicator of cardiac function
strongly predictive of prognosis in chronic heart failure. Eur
Heart J 2001; 22:1496 –1503
Cohen-Solal A, Tabet JY, Logeart D, et al. A non-invasively
determined surrogate of cardiac power (‘circulatory power’)
at peak exercise is a powerful prognostic factor in chronic
heart failure. Eur Heart J 2002; 23:806 – 814
Mohr R, Rath S, Meir O, et al. Changes in systemic vascular
resistance detected by the arterial resistometer: preliminary
report of a new method tested during percutaneous transluminal coronary angioplasty. Circulation 1986; 74:780 –785
Torre-Amione G, Young JB, Colucci ES, et al. Hemodynamic
and clinical effects of tezosentan, an intravenous dual endothelin receptor antagonist, in patients hospitalized for acute
decompensated heart failure J Am Coll Cardiol 2003; 42:140 –
147
Teerlink JR, Massie BM, Cleland JG, et al, for the Ritz 1
Investigators. A double-blind, parallel-group, multi-center,
placebo-controlled study to investigate the efficacy and safety
of tezosentan in reducing symptoms in patients with acute
decompensated heart failure (RITZ 1) [abstract]. Circulation
2001; 104:II-526
VMAC Investigators. Intravenous nesiritide vs nitroglycerin
for treatment of decompensated congestive heart failure: a
randomized trial. JAMA 2002; 287:1531–1540
Cotter G, Faibel H, Barash P, et al. High-dose nitrates in the
immediate management of unstable angina: optimal dosage,
route of administration, and therapeutic goals. Am J Emerg
Med 1998; 16:219 –224
Kaluski E, Kobrin I, Zimlichman R, et al. RITZ-5: randomized intravenous tezosentan (an endothelin ET-A/B antagonist) for the treatment of pulmonary edema: a prospective
randomized, multicenter, double-blind placebo controlled
study. J Am Coll Cardiol 2003; 41:204 –210
Cohen AJ, Arnaudov D, Zabeeda D, et al. Non-invasive
measurement of cardiac output during coronary artery bypass
grafting. Eur J Cardiothorac Surg 1998; 14:64 – 69
Baker LE. Principles of impedance technique. IEEE Eng
Med Biol 1989; 3:11–15
Djordjevich L, Sadove MS. Basic principles of electrohaemodynamics. J Biomed Eng 1981; 3:25–33
Kedrov AA. An attempt of the quantify assessment of the
central and peripheral circulation by electrometrical method.
Klin Med 1948; 26:32–51
Kubicek WG, Karnegis JN, Patterson RP, et al. Development
and evaluation of an impedance cardiac output system.
Aerosp Med 1966; 37:1208 –1212
Kubicek WG, Kottke FJ, Ramos MU, et al. The Minnesota
impedance cardiograph: theory and applications. Biomed
Eng 1974; 9:410 – 416
Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986; I:307–310
Robin ED, McCauley RF. Monitor wizards can be dangerous.
Chest 1998; 114:1151–1153
Dalen JE, Bone RC. Is it time to pull the pulmonary artery
catheter? JAMA 1996; 276:916 –918
Connors AF, Speroff T, Dawson NV, et al, for the SUPPORT
investigators. The effectiveness of right heart catheterization
in the initial care of critically ill patients. JAMA 1996;
276:889 – 897
Raaijmakers E, Faes ThJC, Scholten RJPM, et al. A metaanalysis of published studies concerning the validity of thoracic impedance cardiography. Ann NY Acad Sci 1999;
873:121–134
CHEST / 125 / 4 / APRIL, 2004
1439
25 Patterson RP, Witsoe DA, From A. Impedance stroke volume
compared with dye and electromagnetic flowmeter values
during drug-induced inotropic and vascular changes in dogs.
Ann N Y Acad Sci 1999; 873:143–148
26 Marik PE, Pendelton JE, Smith R. A comparison of hemodynamic parameters derived from transthoracic electrical
bioimpedance with those parameters obtained by thermodilution and ventricular angiography. Crit Care Med 1997;
25:1545–1550
27 Tischenko MI. Estimation of stroke volume by integral
rheogram of the human body [in Russian]. Sechenov Phisiological J 1973; 59:1216 –1224
28 Koobi T, Kaukinen S, Turjanmaa VM, et al. Whole-body
impedance cardiography in the measurement of cardiac
output. Crit Care Med 1997; 25:779 –785
29 Lamberts R, Visser KR, Zijlstra WG. Impedance cardiography. Assen, The Netherlands: Van Gorkum, 1984; 21–94
30 Imhoff M, Lehner JH, Lohelin D. Noninvasive whole-body
electrical bioimpedance cardiac output and invasive ther-
31
32
33
34
35
modilution cardiac output in high-risk surgical patients. Crit
Care Med 2000; 28:2812–2818
Davidson CJ, Fishman RF, Bonow RD. Cardiac catheterization. In: Branwald E, ed. Heart disease: a textbook of
cardiovascular medicine. 5th ed. Philadelphia, PA: W.B.
Saunders Company, 1997; 192
van Grondelle A, Ditchey RV, Groves BM, et al. Thermodilution method overestimates low cardiac output in humans.
Am J Physiol 1983; 245:690 – 692
Kohanna FH, Cunningham JN. Monitoring of cardiac output
by thermodilution after open-heart surgery. J Thorac Cardiovasc Surg 1977; 73:451– 457
Espersen K, Jensen EW, Rosenborg D, et al. Comparison of
cardiac output measurement techniques: thermodilution,
Doppler, CO2-rebreathing and direct Fick method. Acta
Anaesthesiol Scand 1995; 39:245–251
Nanas JN, Anastasiou-Nana MI, Sutton RB, et al. Comparison of Fick and dye cardiac outputs during rest and exercise
in 1,022 patients. Can J Cardiol 1986; 2:195–199
1440
127 of 158.
Accurate, Noninvasive Continuous Monitoring of Cardiac Output by
Whole-Body Electrical Bioimpedance
Gad Cotter, Yaron Moshkovitz, Edo Kaluski, Amram J. Cohen, Hilton Miller,
Daniel Goor and Zvi Vered
Chest 2004;125;1431-1440
DOI 10.1378/chest.125.4.1431
This information is current as of April 4, 2007
Updated Information
& Services
Updated information and services, including
high-resolution figures, can be found at:
http://chestjournals.org/cgi/content/full/125/4/1431
References
This article cites 31 articles, 10 of which you can access
for free at:
http://chestjournals.org/cgi/content/full/125/4/1431#BIBL
Citations
This article has been cited by 1 HighWire-hosted
articles:
http://chestjournals.org/cgi/content/full/125/4/1431
Permissions & Licensing
Information about reproducing this article in parts
(figures, tables) or in its entirety can be found online at:
http://chestjournals.org/misc/reprints.shtml
Reprints
Information about ordering reprints can be found online:
http://chestjournals.org/misc/reprints.shtml
Email alerting service
Receive free email alerts when new articles cite this
article sign up in the box at the top right corner of the
online article.
Images in PowerPoint format Figures that appear in CHEST articles can be
downloaded for teaching purposes in PowerPoint slide
format. See any online article figure for directions.
128 of 158.
The European Journal of Heart Failure 5 (2003) 443–451
The role of cardiac power and systemic vascular resistance in the
pathophysiology and diagnosis of patients with acute congestive heart
failure
Gad Cottera,*, Yaron Moshkovitzb, Edo Kaluskia, Olga Miloa, Ylia Nobikovc, Adam Schneeweissa,
Ricardo Krakovera, Zvi Vereda
a
Cardiology Department, Assaf-Harofeh Medical Center, 70300 Zerifin, Israel
b
Cardiac Surgery Department, Ramat-Marpe Hospital, Petah-Tikva, Israel
c
Biostatistical Department, Sheba Medical Center, Sackler School of Medicine, Tel-Aviv University, Tel-Hashomer, Israel
Received 16 August 2002; received in revised form 6 January 2003; accepted 21 January 2003
Objective: Conventional hemodynamic indexes (cardiac index (CI), and pulmonary capillary wedge pressure) are of limited
value in the diagnosis and treatment of patients with acute congestive heart failure (CHF). Patients and methods: We measured
CI, wedge pressure, right atrial pressure (RAP) and mean arterial blood pressure (MAP) in 89 consecutive patients admitted due
to acute CHF (exacerbated systolic CHF, ns56; hypertensive crisis, ns5; pulmonary edema, ns11; and cardiogenic shock, ns
17) and in two control groups. The two control groups were 11 patients with septic shock and 20 healthy volunteers. Systemic
vascular resistance index (SVRi) was calculated as SVRis(MAPyRAP)yCI. Cardiac contractility was estimated by the cardiac
power index (Cpi), calculated as CI=MAP. Results and discussion: We found that CI-2.7 lyminym2 and wedge pressure )12
mmHg are found consistently in patients with acute CHF. However, these measures often overlapped in patients with different
acute CHF syndromes, while Cpi and SVRi permitted more accurate differentiation. Cpi was low in patients with exacerbated
systolic CHF and extremely low in patients with cardiogenic shock, while SVRi was increased in patients with exacerbated
systolic CHF and extremely high in patients with pulmonary edema. By using a two-dimensional presentation of Cpi vs. SVRi
we found that these clinical syndromes can be accurately characterized hemodynamically. The paired measurements of each
clinical group segregated into a specific region on the CpiySVRi diagnostic graph, that could be mathematically defined by a
statistically significant line (Lambdas0.95). Therefore, measurement of SVRi and Cpi and their two-dimensional graphic
representation enables accurate hemodynamic diagnosis and follow-up of individual patients with acute CHF.
䊚 2003 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.
Keywords: Cardiac power; Vascular resistance; Acute congestive heart failure
1. Introduction
Acute congestive heart failure (CHF) is a common
disease, accounting for over 700 000 annual admissions
to hospitals in the USA alone. We have recently suggested that this disease can be divided into four major
clinical syndromes w1x: (1) pulmonary edema, (2) cardiogenic shock, (3) hypertensive (HTN) crisis and (4)
exacerbated systolic CHF. However, the diagnosis of
*Corresponding author. Tel.: q972-8-977-9778; fax: q972-8-9779779.
E-mail address: [email protected] (G. Cotter).
these clinical syndromes of acute CHF may be difficult,
due to an overlap in symptoms and signs among the
different syndromes as well as lack of objective criteria
for their diagnosis. For example, both cardiogenic shock
and pulmonary edema patients present with severe circulatory and respiratory distress and in both cases CI is
low and wedge pressure is high. However, these two
clinical syndromes have a completely different course
(mortality rates at 1-month in the SHOCK study w2x
were approximately 60%, compared with a mortality
rate of approximately 10% for pulmonary edema patients
during the first 30 days in the RITZ-5 study w3x). In
addition, the pathophysiology of the two clinical syn-
1388-9842/03/$ - see front matter 䊚 2003 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.
PII: S 1 3 8 8 - 9 8 4 2 Ž 0 3 . 0 0 1 0 0 - 4
129 of 158.
Downloaded from eurjhf.oxfordjournals.org by guest on June 14, 2011
Abstract
444
G. Cotter et al. / The European Journal of Heart Failure 5 (2003) 443–451
130 of 158.
2. Methods
2.1. Inclusion criteria
Hemodynamic data was obtained in all patients undergoing right heart catheterization who were diagnosed by
the usual clinical criteria as having acute CHF. We also
enrolled two control groups: these were 11 patients with
septic shock and 20 healthy volunteers.
2.2. Exclusion criteria
Significant valvular disease, significant brady- or
tachy-arrhythmias or renal failure (creatinine )2.5 mgy
dl).
2.3. Clinical diagnosis criteria
2.3.1. Exacerbated systolic CHF
Patients admitted due to signs and symptoms of
worsening CHF, who were in a stable clinical condition;
not fulfilling the criteria for cardiogenic shock, pulmonary edema and HTN crisis and who had EF-35% on
echocardiography. (The echocardiographic criteria were
used to ensure that the symptoms of dyspnea were
indeed due to acute CHF.)
2.3.2. Pulmonary edema
Patients admitted due to clinical symptoms and signs
of acute pulmonary congestion accompanied by findings
of lung edema on chest X-ray who had severe respiratory
distress accompanied by O2 saturation -90% in room
air by pulse oxymetery during the invasive
measurements.
2.3.3. Cardiogenic shock
Systolic blood pressure -100 mmHg for at least 1 h,
not responsive to percutaneous revascularization,
mechanical ventilation, intra-aortic balloon-pump
(IABP), IV fluid administration and dopamine of at
least 10 mgykgymin and accompanied by signs of end
organ hypoperfusion but not accompanied by fever
)388 or a systemic inflammatory syndrome.
2.3.4. HTN crisis
Patients with signs and symptoms of acute CHF
accompanied by high blood pressure (MAP)130
mmHg during invasive measurements); not fulfilling the
criteria for pulmonary edema.
2.3.5. Septic shock
Systolic blood pressure -100 mmHg accompanied
by fever )388, systemic inflammatory syndrome and
signs of end organ hypoperfusion for at least 3 h not
responsive to IV fluids and IV dopamine of at least 10
mgykgymin. No evidence of an acute cardiac event.
dromes is completely different and their treatment is
almost opposite. Moreover, even the diagnosis of acute
CHF is sometimes difficult, due to an overlap in signs
and symptoms with those of acute exacerbations of
obstructive or restrictive lung diseases and the occasional difficulty in differentiation between septic and cardiogenic shock.
Measurement of invasive hemodynamic variables,
including CI and pulmonary capillary wedge pressure
has been used in patients with acute decompensated and
chronic compensated CHF, as well as cardiogenic shock,
for more than two decades. However, despite extensive
experience and numerous studies, no specific diagnostic
criteria or accurate cut-off points have been determined
w2,4–6x. In some studies, trends and changes in CI and
wedge pressure have been used, however, no reproducible criteria for diagnosis and follow-up could be
established.
Cardiac power, an index of cardiac contractility, is
calculated based on the classical physical rule of fluids,
i.e. powersflow=pressure, hence cardiac power index
(Cpi) is the product of simultaneously measured mean
arterial blood pressure (MAP) and cardiovascular flow
(CI): CpisMAP=CI=0.0022 w1,7–9x. The units are
Wym2. Cpi has been used extensively during recent
years to evaluate patients with chronic and acute CHF.
In three separate studies, Marmor and Schneeweiss w7x,
Tan et al. w8x and Cohen-Solal et al. w9x have demonstrated that Cpi increases during exercise (cardiac power
reserve) and is the strongest predictor of outcome in
patients with chronic CHF, stronger than O2 consumption
and echocardiographic ejection fraction. We have previously demonstrated w1x that in patients with exacerbated systolic CHF, baseline Cpi at admission is the
strongest predictor of short- and long-term outcome. On
the other hand, the main event preceding recurrent
worsening heart failure was a steep increase in SVRi.
In a recent analysis, we have also found that Cpi at
baseline and during follow-up was the strongest predictor of outcome in a large cohort of cardiogenic shock
patients (unpublished data).
The main hypothesis of the present study was that in
patients with acute CHF, as Cpi decreases, SVRi should
concomitantly increase. Therefore, for each Cpi decrease
the SVRi increase may be adequate, too high or too low
and, thus, CpiySVRi coupling may characterize the
clinical-hemodynamic state. Therefore, in the present
study, we examined in a two-dimensional representation,
the relationship between changes in Cpi (pump work)
and SVRi (resistance or work load) in the four clinical
syndromes of acute CHF, (i.e. exacerbated systolic CHF,
pulmonary edema, cardiogenic shock and HTN crisis),
as well as in two control groups: (i.e. septic shock and
normal subjects).
2.4. Assessment of hemodynamic variables
2.5. Calculation of hemodynamic variables
Cpi was determined as MAP=CI=0.0022 and SVRi
was determined as (MAPyright atrial pressure (RAP)y
CI. As RAP was not measured in normal subjects, it
was estimated to be 10% of MAP w11x.
2.6. Echocardiographic evaluation
All patients underwent routine echocardiographic
evaluation after initial stabilization. This included visual
estimation of cardiac function, evaluation of valvular
function and gross estimation of signs of diastolic
dysfunction.
best cut-off point for the various parameters, in terms
of highest sensitivity and specificity.
2.8. CpiySVRi diagnostic graph
A classification rule was developed using second
order discriminant analysis. The normality of the distribution of Cpi and SVRi was examined by the Wilk–
Shapiro test. Due to the skewness of the data in some
groups, both variables (CPi and SVRi) were transformed
into log scale for better approximation to normality.
Since the number of patients with HT crisis was small
they were considered together with the exacerbated
systolic CHF group. The classification used two steps.
In the first step the rule separated three classes: septic
shock, cardiogenic shock and a combined group, which
included the normal controls, compensated CHF and
pulmonary edema patients (N–C–P). If, after the first
step the patient was defined as N–C–P, the second
classification was used to differentiate between the
normal, exacerbated systolic CHF and pulmonary edema
subgroups.
All calculations were performed by SAS 6.12 (SAS
Institute Inc., Cary, NC) using procedures FREQ,
MEANS, GLM, DISCRIM, GPLOT.
Alfa level: 5%.
3. Results
Eighty-nine consecutive patients admitted due to acute
CHF and LV dysfunction (exacerbated systolic CHF,
ns56; pulmonary edema, ns11; cardiogenic shock,
ns17; and HTN crisis, ns5) as well as 11 patients
with septic shock and 20 healthy volunteers were
enrolled in the study. Baseline characteristics of the
different patient groups are presented in Table 1. The
mean CI, wedge pressure, MAP, SVRi and Cpi according
to clinical diagnosis are presented in Table 2.
3.1. Hemodynamic variables
2.7. Statistical methods
The five clinical groups were compared with regard
to all parameters using a one-way analysis of variance
(ANOVA). Therafter, the Ryan–Einot–Gabriel–Welsch
Multiple Range Test was used for pair-wise comparisons
between the groups, while Dunnett’s t-test was used to
compare all groups to the healthy controls.
A one-sample t-test was performed to compare mean
wedge pressure in each group to the wedge pressure of
normal people (-12 mmHg).
In order to determine the usefulness of the hemodynamic parameters to discriminate between the clinical
syndromes, ROC curves, derived from a logistic regression model were applied to the data to determine the
131 of 158.
3.1.1. Cardiac index (Fig. 1)
The average values of CI were significantly lower in
patients with acute CHF and higher in patients with
septic shock. ROC analysis found a cut-off point of
CI-2.7 lyminym2 useful for the determination that a
patient had acute CHF (sensitivitys1, specificitys
0.99). However, values between 1.2 and 2.7 lyminym2
could be found in all patients with exacerbated systolic
CHF and HTN crisis as well as 73% of patients with
pulmonary edema and 47% of patients with cardiogenic
shock. Moreover, the mean CI of patients with pulmonary edema and cardiogenic shock was found to be
almost identical (1.4"0.4 vs. 1.35"0.7 lyminym2, Ps
ns).
Prior to enrolment in this study all patients gave
written informed consent. The study protocol was
approved by the local ethics review board. In all patients
the hemodynamic variables were obtained during right
heart catheterization using a Swan–Ganz catheter placed
under fluoroscopic guidance. All measurements were
obtained while patients were at least 30 s without IABP
while on the same treatment used at the time the clinical
diagnosis was made. Measurement of hemodynamic
variables was performed at least 6 h after the last intake
of an oral drug and 2 h after intravenous drug therapy.
CI was measured by thermodilution, using the mean
of at least three consecutive measurements within a
range of -15%. In normal subjects, right heart catheterization was not performed for ethical reasons. The
values used in this cohort were obtained by standard
noninvasive cuff blood pressure measurement and evaluation of CI by the FDA approved NICaS娃 2001, a
noninvasive continuous cardiac output monitor w10x.
Therefore, wedge pressure was not assessed in normal
subjects. Instead, we used standard values documented
in the literature w11x.
445
446
Table 1
Baseline characteristics of the five patient groups
Sex (male:female)
Age (years)
Weight (kg)
Body surface area (m2)
IHD (%)
Previous MI
EF (%)
Diabetes mellitus (%)
Current smokers (%)
Hypertension (%)
Hyperlipidemia (%)
Baseline creatinine
Medications for CHF
Diuretic (%)
Digoxin (%)
ACE inhibitoryAII blocker (%)
Beta-blocker (%)
Nitrate (%)
Normal
volunters
Septic
shock
Exacerbated
systolic CHF
Pulmonary
edema
Cardiogenic
shock
12:8
60"8
79"14
1.92"0.22
0
0
55"3
20
60
50
65
7:4
55"11
77"10
1.91"0.23
18
9
46"9
73
55
45
73
135"75
51:10
69"10
72"8
1.88"0.21
79
62
27"5
66
36
56
66
124"55
5:6
73"12
70"9
1.81"0.24
73
55
41"10
66
44
88
66
144"81
11:6
67"11
80"14
1.92"0.24
100
18
24"6
53
53
71
53
110"47
0
0
0
0
0
0
0
9
9
0
82
33
95
62
41
91
45
91
55
55
6
0
29
24
0
septic shock. The analysis was based on normal values
reported in the literature (-12 mmHg w8x) (Ps0.001).
However, the overlap of wedge pressure in the different
acute CHF groups was extensive. Values between 12
and 38 mmHg were found in 82, 64, 76 and 18% of
patients with exacerbated systolic CHF, pulmonary edema, cardiogenic shock and septic shock, respectively.
3.1.2. Mean arterial blood pressure
By virtue of their clinical definition, the average
values of MAP were higher in patients with HTN crisis
and lower in those with septic and cardiogenic shock.
However, large areas of overlap were found between
pulmonary edema, HTN crisis and exacerbated systolic
CHF (MAP)100 mmHg) and between exacerbated
systolic CHF, cardiogenic shock and septic shock
(MAP-100 mmHg).
3.1.4. Cardiac power index (Fig. 3)
Compared with the normal controls, the mean values
of Cpi were low in patients with exacerbated systolic
CHF and pulmonary edema and extremely low in
3.1.3. Pulmonary capillary wedge pressure (Fig. 2)
The mean wedge pressure was significantly higher in
patients with acute CHF and lower in patients with
Table 2
Baseline distribution of the various hemodynamic parameters in the six diagnosis groups presented as means and S.D.
N
SVRi
Cpi
Wedge
MAP
CI
Exacerbated
systolic CHF
Pulmonary
edema
Cardiogenic
shock
HTN
crisis
Septic
shock
56
44.9"8.0
0.47"0.13
25.5"7.2
101"18
2.06"0.33
11
88.2"16.7
0.4"0.13
32.7"8.6
131.4"12.7
1.37"0.32
17
55.6"31.1
0.22"0.08
23.3"6.5
72.2"11.3
1.42"0.64
5
54.3"3.2
0.75"0.04
28.5"4.5
150"10.5
2.24"0.37
11
11.8"1.1
0.8"0.13
11.4"7.7
68.2"5.4
5.2"0.5
Normal
20
25.2"4.1
0.62"0.08
–
87.9"8.85
3.2"0.36
The results of the ANOVA for comparisons between exacerbated systolic CHF (CHF), pulmonary edema
(edema) and cardiogenic shock (shock) patients
Parameter
P-value for overall group
P-value for paired comparisons (Ryan–Einot–Gabrigroups
comparison (ANOVA)
el–Welsch multiple range test)
CHF-edema
CHF-shock
Edema-shock
CI
Wedge
SVRi
CPi
0.0001
0.0037
0.0001
0.0001
q
q
q
q
q
q
q
q
q
N, number of patients; SVRi, systemic vascular resistance index (wood m2 ); Cpi, cardiac power index (mmHg lyminym2 ); wedge, pulmonary
capillary wedge pressure (mmHg); MAP, mean arterial blood pressure (mmHg); CI: cardiac index (lyminym2 ). q, Significant group difference.
132 of 158.
IHD, ischemic heart disease; MI, myocardial infarction; EF, ejection fraction; CHF, congestive heart failure.
Fig. 1. CI (lyminym2) Box-plots (median and 25–75% percentile
range) in patients with the different syndromes of acute CHF and
patients with septic shock and normal controls.
3.1.5. Systemic vascular resistance index (Fig. 4)
Average values of SVRi were significantly higher in
all patients with exacerbated systolic CHF, HTN crisis
and extremely high in patients with pulmonary edema,
but were lower in patients with septic shock. SVRi was
found to be instrumental in the diagnosis of pulmonary
edema. All patients with this clinical syndrome had
SVRi)67 wood m2, while SVRi values in all other
patient groups, as well as in normal subjects, were
significantly lower than this value.
Fig. 2. Pulmonary capillary wedge pressure (mmHg) Box-plots
(median and 25–75% percentile range) in patients with the different
syndromes of acute CHF and patients with septic shock and normal
controls.
133 of 158.
Fig. 3. Cpi (Wym2) Box-plots (median and 25–75% percentile range)
in patients with the different syndromes of acute CHF and patients
with septic shock and normal controls.
3.2. CpiySVRi diagnostic graph (Fig. 5)
Since the number of patients with HTN crisis was
small they were included in the exacerbated systolic
CHF group. Distributions of SVRi and Cpi were highly
skewed. The normality of the distribution of Cpi and
SVRi was assessed by the Wilk–Shapiro test. The results
showed that Cpi and SVRi distribution was not normal
for the normal volunteers (Ps0.03 for Cpi), the exacerbated systolic CHF patients(Ps0.007 for Cpi and Ps
0.04 for SVRi) and the cardiogenic shock patients(Ps
0.016 for SVRi).
However, log(SVRi) and log(CPi) were normally
distributed (Psns for both Cpi and SVRi in all patient
groups). Therefore, for the analysis we used only log
Fig. 4. Systemic vascular resistance index (SVRi) (wood m2) Boxplots (median and 25–75% percentile range) in patients with the different syndromes of acute CHF and patients with septic shock and
normal controls.
patients with cardiogenic shock. However, some overlap
was encountered among the five groups.
447
448
numbers of observations in two of the groups prevented
us from using more flexible kernel functions.
Due to large variability of variances of the parameters
in the five groups, we could not suppose equal covariance matrices in the groups. (The test of homogeneity
of within covariance matrices gives Ps0.)
3.2.1. Classification rule
The calculations leading to the classification rule and
CpiySVRi diagnostic graph are given in Appendix A.
3.2.2. Classification results
The results of the application of the classification rule
to the sample are presented in Table 4.
Fig. 5. Diagnostic graph for classification of the hemodynamic status
of patients with different syndromes of acute CHF.
Table 3
Number of observations classified into the correct clinical group using log(Cpi) or log(SVRi) only
Group
Cardiogenic
shock
Exacerbated
systolic CHF
Normal
(a) Number of observations classified into appropriate groups: classification using log(CPi) only
Cardiogenic Shock
13
4
0
Exacerbated systolic CHF
1
44
14
Normal
0
9
8
Pulmonary edema
1
9
1
Septic shock
0
0
3
Pulmonary
edema
Septic
shock
Total
0
0
0
0
0
0
2
3
0
8
17
61
20
11
11
2
0
0
9
0
0
0
0
0
11
17
61
20
11
11
(b) Number of observations classified into appropriate groups using log(SVRi) only
Cardiogenic shock
Normal
Pulmonary edema
Septic shock
2
0
0
2
0
12
58
3
0
0
1
3
17
0
0
Table 4
Number of observations classified into the correct clinical group using log(SVRi) and log(CPi)
Group
Cardiogenic
shock
Exacerbated
systolic CHF
Normal
Pulmonary
edema
Septic
shock
Total
Cardiogenic shock
Normal
Pulmonary edema
Septic shock
15
0
0
2
0
2
60
0
0
0
0
1
20
0
0
0
0
0
11
0
0
0
0
0
11
17
61
20
11
11
134 of 158.
values. The distribution of the two log parameters (Cpi
and SVRi) was different among the five groups, however, none enabled separation of the groups (Table 3).
These data suggested that separation may be possible
using two-dimensional discriminant analysis. We used
classical discriminant analysis for normal distributions
with unequal covariance matrices because the small
3.2.3. Performance of the classification rule
The performance of a diagnostic procedure with only
two possible results and two classes of patients is usually
expressed using measures like positive (negative) predictive value w12x or diagnostic odds ratio w13x. For
more complex tests with many outcomes and many
classes of patients the overall performance may be
expressed through the difference between proportion of
erroneously classified patients with and without using
the test. This measure is usually called Lambda asymmetric (RNC), where R (rows) is the true group and C
(column) is the group where the patient was classified.
For our data Lambda (RNC)s0.95 (S.D. (Lambda)s
0.03) which corresponds to three errors of classification
according to the classification rule, instead of 59 errors
of classification according to the prior probabilities.
4. Discussion
4.1. Classical hemodynamic monitoring
CI is the most popular parameter used in invasive
hemodynamic monitoring of patients with acute CHF.
However, the results of the present study as well as
previous ones w2,4–6x show that CI measurements are
not sufficient for the diagnosis and treatment titration in
patients with acute CHF. This might be explained by
the fact that CI is actually a measure of cardiovascular
flow. Hence, CI (flow) is determined by both cardiac
contractility and vascular resistance and, therefore, may
change dramatically when Cpi decreases but also with
even mild changes in SVRi. Pulmonary capillary wedge
pressure is the second most popular hemodynamic variable used in hemodynamic monitoring, since it represents the hydraulic pressure transmitted backwards to
the pulmonary circulation, and hence, is an important
determinant of pulmonary edema. However, wedge pressure cannot be used for the exact diagnosis of the
different clinical syndromes of acute CHF, due to the
extent of overlapping values between patients with
exacerbated systolic CHF, HTN crisis and even cardiogenic shock.
4.2. Cpi and SVRi and their role in patients with acute
CHF
Cpi as measured in the present study w1,8,9x is a
simplified version of a previously described method of
measuring cardiac contractility w7x. This value is derived
from the entire cardiac cycle (instead of instantaneous
measurements) and is the product of the mean pressure
and flow. Cpi has been shown to be the best predictor
135 of 158.
of outcome in chronic CHF patients w7–9x, exacerbated
systolic CHF w1x and cardiogenic shock (unpublished
data).
In the present study, we found that in patients with
exacerbated systolic CHF either Cpi was decreased or
SVRi was increased or both changes occurred. In a
previous study, we described the sequence of events
leading to acute heart failure w1x. In most patients an
acute CHF event starts with a progressive decrease in
cardiac contractility and power (Cpi). Thereafter, as Cpi
decreases, neurohormonal vascular control mechanisms
are activated and SVRi increases w1,17x. This increase
is a very important protective mechanism for two
reasons:
1. The increase in SVRi in the face of decreased
contractility maintains blood pressure and the perfusion of vital organs.
2. This afterload increase (while within certain limits)
may improve contractility (possibly through the
Gregg phenomenon w18x), which may account for the
‘normal’ Cpi we observed in some patients with
echocardiographically
demonstrated
systolic
dysfunction.
However, SVRi increase in response to Cpi decrease
is not uniform. It can be appropriate (thus, leading to a
compensated state), inappropriately low (thus, leading
to low blood pressure, forward hypoperfusion and cardiogenic shock) or inappropriately high (thus, inducing
an extreme afterload mismatch leading to pulmonary
edema).
Indeed, in the present study, in patients who were
clinically diagnosed as cardiogenic shock, Cpi was found
to be extremely low, however, SVRi was only slightly
increased. This imbalance between very low Cpi and
inadequate increase in SVRi probably resulted in low
blood pressure and decreased perfusion pressure of vital
organs including the heart. This decrease in coronary
perfusion might lead to decreased contractility inducing
a vicious cycle of low contractility, low SVRi and
reduced perfusion. For this reason, in a previous study
w15x we treated patients with cardiogenic shock, by
short-term administration of a peripheral vasoconstrictor
(L-NMMA) with good clinical response.
On the other hand, in patients diagnosed as pulmonary
edema, despite what appears to be a similar clinical
presentation (pulmonary congestion, clammy extremities, low CI and high wedge pressure), the pathophysiological findings as well as the treatment, are the
complete opposite. In patients with pulmonary edema,
we measured Cpi values similar to those in exacerbated
systolic CHF, however, SVRi was markedly increased.
These findings are collaborated by the study by Gandhi
et al. w19x showing a dramatic increase in blood pressure
in patients with pulmonary edema. We hypothesize that
this increase in SVRi might be an inappropriate
During the last decade the pathogenesis of CHF has
become clearer. The emphasis on cardiac performance
as the sole pathogenic mechanism of CHF has changed
to a more comprehensive understanding of the importance of the interaction between cardiac contractility,
neurohormonal and inflammatory control mechanisms
and vascular resistance. We have recently studied the
treatment of the acute CHF syndromes of pulmonary
edema and cardiogenic shock and have shown that
treatment modalities with significant vascular effect are
effective in improving the outcome of these patients
w14–16x. These findings substantiated our theory that
the SVRi reaction to the decrease in Cpi determines the
hemodynamic condition and clinical syndrome of
patients with acute CHF.
449
450
response, related to neurohormonal, endothelial and
perhaps inflammatory activation w20,21x. This remarkable increase in SVRi induces an afterload mismatch,
reducing CI and increasing intracardiac pressures,
LVEDP and wedge pressure, resulting in the severe
congestive symptoms of pulmonary edema. Therefore,
as previously suggested w14,16,22,23x, vasodilator treatment is effective in the treatment of pulmonary edema.
4.3. Two-dimensional graphic representation of Cpiy
SVRi and its use in the treatment of cardiogenic shock
and pulmonary edema
4.4. Limitations
The results of the present study are based on a
relatively small number of patients, and therefore, need
confirmation by prospectively evaluating a larger group
of patients with acute CHF. Also, the measurements in
the present study were performed by thermodilution,
which has an inherent 10–15% deviation in measuring
cardiac output. Finally, cardiac power calculations were
performed using whole-cycle measurements (cardiac
output and MAP). Although we recognize that the
cardiovascular system operates in a pulsatile manner,
cardiac power calculated according to the methodology
used in the present study has previously been shown to
be a useful measure of cardiac contractility and contrac-
136 of 158.
Appendix A: Classification rule
Given a patient with measured values of SVRi and
Cpi, the classification may be performed either (A)
through special calculations or (B) using the ‘Graph for
classification of CHF patients (CpiySVRi graph)’.
(A) Classification using calculations
Step 1. Calculate three values v1, v2, v3 according to
the formulas below.
v1sLCPi2=21.54q2=LCPi=LSVRi=10.61
qLSVRi2=59.44yLCPi=305.24yLSVRi
=417.70q1408.89
v2sLCPi2=10.12q2=LCPi=LSVRi=5.67yLSVRi2
=4.99yLCPi=135.81yLSVRi=90.11q482.61
v3sLCPi2=7.29qLCPi=LSVRi=2.57qLSVRi2
=4.09yLCPi=97.41yLSVRi=58.22q368.16
Classify the patient
into the group ‘Septic shock’, if v1 is the smallest value
into the group ‘Cardiogenic Shock’, if v2 is the smallest
value
if v3 is the smallest value go to step 2
Step 2. Calculate three values v4, v5, v6 according to
the formula below.
v4sLCPi2=6.45y2=LCPi=LSVRi=0.45q
LSVRi2=16.01yLCPi=65.16yLSVRi=116.53q
391.67
v5sLCPi2=17.75q2=LCPi=LSVRi=26.56q
LSVRi2=54.27yLCPi=420.26ySVRi=758.55q
2775.78
v6sLCPi2=32.95q2=LCPi=LSVRi=3.09q
LSVRi2=19.72yLCPi=390.74yLSVRi=161.49q
1355.57
Classify the patient
into the group ‘Exacerbated Systolic CHF’, if v4 is the
smallest value among v4, v5, v6 and LSVRi-log(67)
into the group ‘Pulmonary Edema’, if v5 is the smallest
value among v4, v5, v6 and LSVRi)log(67)
into the group ‘Normal’, if v6 is the smallest value
among v4, v5, v6
The value of SVRis67 was used to separate patients
with exacerbated systolic CHF from patients with pulmonary edema since the group of ‘pulmonary edema’
was rather small and by classifying these patients
according to the usual rule we did not receive a
separating line for Cpi measures)250 Wym2.
In the present study, we found that when plotting on
a two-dimensional graph the results of Cpi and SVRi
for individual patients, each clinical group of patients
could be segregated into a specific area on the graph
which could be bound by a mathematically defined line
(Fig. 5). This graph enables exact clinical diagnosis of
most (95%) patients with exacerbated systolic CHF,
pulmonary edema, HTN crisis, cardiogenic shock and
septic shock. Of course, the boundaries on the graph are
somewhat arbitrary, since the definitions of the syndromes are as used by the medical community, and
therefore, arbitrary. However, this two-dimensional representation enables a better understanding of the pathophysiology of the different syndromes of acute CHF.
We believe that this new and simple diagnostic tool
may become useful for the initial evaluation of acutely
decompensated patients, while the clinical diagnosis has
not yet been established and initiation of appropriate
disease-specific treatment is crucial. This might become
even more important with the advent of new devices
that accurately measure CI noninvasively. The combination of noninvasive MAP and CI measurement, Cpi
and SVRi calculation and the two-dimensional Cpiy
SVRi graph could enable improved diagnosis of patients
even in paramedic units and in emergency rooms.
Furthermore, this method may become an important
tool for monitoring the patients’ response to treatment.
tility reserve in chronic and acute CHF as well as in
cardiogenic shock.
(B) Classification using the diagnostic graph
Put the point (CPi, SVRi) on the diagnostic graph
Fig. 5 or point (LCPi, LSVRi) and classify the patient
according to the area of the graph, where the point is
located.
References
137 of 158.
w10x Cohen JA, Arnaudov D, Zabeeda D, Schlthes L, Lashinger J,
Schachner A. Non-invasive measurement of cardiac output
during coronary artery bypass grafting. Eur J Card Thoracic
Surg 1998;14:64 –9.
w11x Lange RA, Hillis LD. Cardiac catheterization and hemodynamic assessment. In: Topol EJ, Textbook of Cardiovascular
Medicine. page 1653.
w12x Sasieni PD. Statistical analysis of the performance of diagnostic
tests. Cytopathology 1999;10:73 –8 (invited review).
w13x Lijmer JG, Willen Mol B, Heisterkamp S, et al. Empirical
evidence of design related bias in studies of diagnostic tests. J
Am Med Assoc 1999;282:1061 –6.
w14x Cotter G, Metzkor E, Kaluski E, et al. Randomised trial of
high-dose isosorbide dinitrate plus low-dose furosemide versus
high-dose furosemide plus low-dose isosorbide dinitrate in
severe pulmonary oedema. Lancet 1998;351:389 –93.
w15x Cotter G, Kaluski E, Blatt A, et al. L-NMMA (a nitric oxide
synthase inhibitor) is effective in the treatment of cardiogenic
shock. Circulation 2000;101:1358 –61.
w16x Sharon A, Shpirer I, Kaluski E, et al. High-dose intravenous
isosorbide-dinitrate is safer and better than Bi-PAP ventilation
combined with conventional treatment for severe pulmonary
edema. J Am Coll Cardiol 2000;36:832 –7.
w17x Mohr R, Rath S, Meir O, et al. Changes in systemic vascular
resistance detected by the arterial resistometer: preliminary
report of a new method tested during percutaneous transluminal
coronary angioplasty. Circulation 1986;74:780 –5.
w18x Karunanithi MK, Young JA, Kalnins W, Kesteven S, Dip G,
Feneley MP. Response of the intact canine left ventricle to
increased afterload and increased coronary perfusion pressure
in the presence of coronary flow autoregulation. Circulation
1999;100:1562 –8.
w19x Gandhi SK, Powers JC, Nomier AM, et al. The pathogenesis
of acute pulmonary edema associated with hypertension. New
Engl J Med 2001;344:17 –22.
w20x Bank AJ, Lee PC, Kubo SH. Endothelial dysfunction in patients
with heart failure: relationship to disease severity. J Card Fail
2000;6:29 –36.
w21x Milo O, Cotter G, Kaluski E, Blatt A, Krakover R, Vered Z,
Hershkovitz A. Inflammatory and neurohormonal activation in
cardiogenic pulmonary edema: Implications on the pathogenesis and outcome of acute ischemic versus non-ischemic acute
heart failure. Am J Cardiol (In Print).
w22x Northridge D. Furosamide or nitrates for acute heart failure?
Lancet 1996;347:667 –8.
w23x Cotter G, Kaluski E, Moshkovitz Y, Milovanov O, Krakover
R, Vered Z. Pulmonary edema: new insights on pathogenesis
and treatment. Curr Opin Card 2001;16:159 –63.
w1x Cotter G, Moshkovitz Y, Milovanov O, et al. Acute congestive
heart failure: a novel approach to its pathogenesis and treatment. Eur J Heart Fail 2002;4:227 –34 (review).
w2x Menon V, White H, LeJemtel T, Webb JG, Sleeper LA,
Hochman JS. The clinical profile of patients with suspected
cardiogenic shock due to predominant left ventricular failure:
a report from the SHOCK Trial Registry. Should we emergently
revascularize occluded coronaries in cardiogenic shock? J Am
Coll Cardiol 2000;36:1071 –6.
w3x Kaluski E, Kobrin I, Zimlichman R, et al. RITZ-5: Randomized
Intravenous Tezosentan (an Endothelin ET-AyB Antagonist)
for the treatment of pulmonary edema. A prospective randomized, multicenter, double-blind placebo controlled study. J Am
Coll Cardiol 2003;41:204–10.
w4x Morley D, Brozena SC. Assessing risk by hemodynamic profile
in patients awaiting cardiac transplantation. Am J Cardiol
1994;73:379 –83.
w5x Anguita M, Arizon JM, Bueno G, et al. Clinical and hemodynamic predictors of survival in patients aged -65 years with
severe congestive heart failure secondary to ischemic or nonischemic dilated cardiomyopathy. Am J Cardiol 1993;72:413 –
7.
w6x Levine TB, Levine AB, Goldberg D, Narins B, Goldstein S,
Lesch M. Reversal of end-stage heart failure is predicted by
long-term therapeutic response rather than initial hemodynamic
and neurohormonal profile. J Heart Lung Transplant
1996;15:297 –303.
w7x Marmor A, Schneeweiss A. Prognostic value of noninvasively
obtained left ventricular contractile reserve in patients with
severe heart failure. J Am Coll Cardiol 1997;29:422 –8.
w8x Williams SG, Cooke GA, Wright DJ, et al. Peak exercise
cardiac power output; a direct indicator of cardiac function
strongly predictive of prognosis in chronic heart failure. Eur
Heart J 2001;22:1496 –503.
w9x Cohen-Solal A, Tabet JY, Logeart D, Bourgoin P, Tokmakova
M, Dahan M. A non-invasively determined surrogate of cardiac
power (‘circulatory power’) at peak exercise is a powerful
prognostic factor in chronic heart failure. Eur Heart J
2002;23:806 –14.
451
www.lejacq.com
ID:3405
EDITORIAL
Beyond the Four Quadrants: The Critical and Emerging Role
of Impedance Cardiography in Heart Failure
H
Adrenergic System
X
60%
Ejection Fraction
eart failure (HF) is a disorder
characterized by hemodynamic
abnormalities including a reduction in
the heart’s ability to deliver oxygenated blood to the body. HF is also associated with important neurohormonal
abnormalities, including activation
of the renin-angiotensin-aldosterone
and sympathetic nervous systems and
their resulting effects on the heart
and vascular endothelium. Our understanding of the neurohormonal role
in the progression of HF has greatly
improved in the past 10 years,1 and
many of the therapies that significantly improve the symptoms and prognosis of patients with HF now target the
underlying neurohormonal abnormalities. As shown in Figure 1, neurohormonal activation can lead to progression of hemodynamic abnormalities
resulting in reduced cardiac output
(CO); increased filling pressures; and
ultimately worsening symptoms of
fatigue, dyspnea, and decreased exercise tolerance. Although the neurohormonal mechanisms may cause progression of the disease process, nearly
all medications used in HF treatment
have demonstrable effects on hemodynamics. Current acute HF treatment is aimed directly at stabilizing
and improving a patient’s short-term
hemodynamic condition; chronic HF
treatments can alter short-term and
improve long-term hemodynamics.
Specific hemodynamic measurements such as CO and systemic vascu-
Low
Perfusion
at Rest
NO
YES
20%
Hemodynamics
Reduced cardiac
output, increased
filling pressures
NO
X RAS
Secondary
Damage
Time
Congestion at Rest
Symptoms
Fatigue, dyspnea,
decreased exercise
tolerance
Figure 1. Neurohormonal activation and
resultant hemodynamic and symptom
changes. RAS=renin-angiotensin-aldosterone system
lar resistance are generally obtained for
only the most critically ill HF patients,
in large part due to the risk, discomfort, and cost of invasive procedures
such as pulmonary artery catheterization.2 Nonetheless, understanding
and measuring the factors that affect
CO are central to the assessment,
prognosis, and treatment of patients
with HF. The four determinants of
CO are the rate of the pump (heart
rate), the volume of blood available to
pump (preload), the pumping strength
(contractility), and the force the heart
must overcome to pump (afterload,
generally approximated by systemic
vascular resistance). Symptoms—
physical findings like vital signs—and
YES
A
B
Warm & Dry
Warm & Wet
(Low Profile)
(Complex)
L
Cold & Dry
C
Cold & Wet
Possible Evidence of Low Perfusion:
Narrow pulse pressure, cool extremities, sleepy/obtunded,
hypotension with ACE inhibitor, low serum sodium, renal/hepatic
dysfunction
Signs/Symptoms of
Congestion:
Orthopnea/PND JV distension, Hepatomegaly,
Edema, Rales, Abdominal-jugular reflex
Figure 2. Clinical profiles in heart failure.
PND=paroxysmal nocturnal dyspnea;
JV=jugular vein; ACE=angiotensinconverting enzyme. Adapted from J Am
Coll Cardiol. 2003;41(10):1797–1804.5
laboratory findings such as blood tests
and chest radiographs are imprecise
measures of hemodynamic function.
Unfortunately, they are the only data
many clinicians have at their disposal
when making important decisions in
the care of patients with HF.
The direct cost of treating HF is estimated to be $56 billion per year in the
United States3 and the number of HF
patients in this country may reach 10
million by 2010.4 A significant portion
of the cost of HF care is the high cost of
hospitalizations for patients with acute
decompensation. Through careful surveillance of patients with chronic HF
using improved methods for measuring
hemodynamic and neurohormonal
John E. Strobeck, MD, PhD;1 Marc A. Silver, MD,2 Co-Editors in Chief
From the Heart-Lung Center, Hawthorne, NJ;1 and the Department of Medicine and the Heart Failure Institute, Advocate Christ
Medical Center, Oak Lawn, IL2
Address for correspondence: Marc A. Silver, MD, Advocate Christ Medical Center, 4440 West 95th Street, Suite 428 South,
Oak Lawn, IL 60453-2600
editorial: emerging role of ICG in HF
march . april 2004 . supplement 2
1
138 of 158.
CHF-MarchApril-Suppl-04.indd 1
3/30/04 9:40:41 AM
Figure 3. Front view of impedance
cardiography method
status, primary care physicians and cardiologists may be able to intervene in a
timely manner and prevent acute episodes leading to hospitalization, major
morbidity, or death.
Warner-Stevenson5 has developed
and popularized the concept of categorizing HF patients by hemodynamic subset based on perfusion with CO (warm
vs. cold) and congestion with pulmonary artery wedge pressure (wet vs. dry).
The four quadrants, representing the
four hemodynamic classes, are shown
in Figure 2. Studies have suggested that
these profiles provide a useful framework
to risk stratify patients with HF, predict outcomes, and identify therapeutic
Figure 4. Lateral view of impedance
cardiography method
options. However, this framework is based
on invasive pulmonary artery catheterization, with its requisite risk and cost, or on
physical examination and patient history,
which have been shown to lack sensitivity and specificity, even in the hands of
experienced clinicians.6 HF management
using hemodynamic subsets could be
substantially improved by the existence of
more objective data with which to classify
patients and evaluate the effectiveness of
subsequent pharmacologic and implantable interventions.
Impedance cardiography (ICG) is
a noninvasive method of determining
hemodynamic status. In the past, studies
questioned the reliability of ICG technology,7,8 leading some to conclude that the
technology did not have value in clinical
decision making. However, refinements
in signal processing and CO algorithms
have greatly improved the reliability of
ICG technology. The latest generation
of ICG devices (BioZ ICG Monitor,
CardioDynamics, San Diego, CA; and
BioZ ICG Module, GE Medical Systems
Information Technologies, Milwaukee,
WI) are both highly reproducible and
accurate in a number of clinical settings,
including HF.9–11 A recent search of the
literature failed to show a single citation
since US Food and Drug Administration
510(k) clearances of these particular
devices that suggests they are not valid
for clinical applications.
ICG is a form of plethysmography
that utilizes changes in thoracic electrical impedance to estimate changes in
blood volume in the aorta and changes
in fluid volume in the thorax. As shown
in Figures 3 and 4, the ICG procedure
involves the placement of four dual
Table I. Impedance Cardiography Parameters
PARAMETER
ABBREVIATION
MEASUREMENT/CALCULATION
UNITS
Flow
Stroke volume
SV
VI × LVET × VEPT (Z MARC algorithm)
mL
Stroke index
SI
SV/body surface area
mL/m2
Cardiac output
CO
SV × heart rate
L/min
Cardiac index
CI
CO/body surface area
L/min/m2
SVR
([MAP – CVP]/CO) × 80
dyne × s × cm-5
Systemic vascular resistance index
SVRI
([MAP – CVP]/CI) × 80
dyne × s × cm-5 × m2
Preejection period
PEP
ECG Q wave to aortic valve opening
ms
Left ventricular ejection time
LVET
Aortic valve opening to closing
ms
Resistance
Contractility
Systolic time ratio
STR
PEP/LVET
No units
Velocity index
VI
/1000/s
Acceleration index
ACI
First time derivativemax/baseline impedance
Left cardiac work index
LCWI
(MAP – PCWP) × CI × 0.0144
kg × m/m2
1/baseline impedance
/kOhm
Second time derivativemax/baseline impedance
/100/s2
Fluid Status
TFC
VEPT=volume of electrically participating tissue; Z MARC=impedance modulating aortic compliance; CVP=central venous pressure (estimated value
of 6 mm Hg); MAP=mean arterial pressure; ECG=electrocardiogram; PCWP=pulmonary capillary wedge pressure (estimated value of 10 mm Hg)
2
139 of 158.
3/30/04 9:40:42 AM
sensors on a patient’s neck and chest.
A low-amplitude, high-frequency alternating current is delivered from the
four outer sensors and the four inner
sensors detect instantaneous changes
in voltage. As suggested by Ohm’s law,
when a constant current is applied to
the thorax, the changes in voltage are
directly proportional to the changes in
measured impedance. The overall thoracic impedance, called base impedance
(Z0) is the sum of the impedances of the
components of the thorax, including fat,
cardiac and skeletal muscle, lung and
vascular tissue, bone, and air. Changes
from Z0 occur due to changes in lung
volumes with respiration and changes
in the volume and velocity of blood in
the great vessels during systole and diastole. The rapidly changing component
of chest impedance (∆Z) is filtered to
remove the respiratory variation, leaving
the impedance changes due to ventricular ejection. Figure 5 details the elements
contributing to Z0 and ∆Z, and Figure 6
illustrates how the first derivative of the
impedance waveform (∆Z/∆t) is used
with an electrocardiogram to determine
the beginning of electrical systole, aortic
valve opening, maximal deflection of the
∆Z/∆t waveform, and the closing of the
aortic valve. From these fiducial points,
a variety of measured and calculated
parameters (Table I) are continuously
displayed on the ICG device screen for
monitoring purposes, or in a printed
report for review (Figure 7).
The hemodynamic parameters
derived from ICG can aid in the
diagnostic and prognostic evaluation
of patients with HF. Using ICG, a
clinician is able to evaluate direct or
indirect measures of each of the four
major determinants of CO (preload,
afterload, contractility, and heart
rate). Figure 8 is a conceptual diagram of CO and its determinants,
ICG parameters associated with the
determinants, and the effects of
pharmacologic agent classes on each
determinant. Due to greater acceptance of ICG in clinical and research
settings, clinicians are now able to
use ICG-derived hemodynamic data
to help decide when to initiate and
Table II. Summary of Impedance Cardiography Applications in Heart Failure
APPLICATION
DESCRIPTION
Assessment and
diagnostic
Establish baseline hemodynamics
Trend changes to gauge level of hemodynamic decompensation
Determine whether symptoms are due to hemodynamic deterioration
Aid in differentiation of systolic vs. diastolic dysfunction
Prognostic
Emergency department values predictive of length of stay and
hospital charges
Improvements associated with improving NYHA class, quality-of-life
measures
Abnormal values associated with mortality
Determine stability for initiation and up-titration of β-blocker and
ACE-inhibitor therapy
Assist in selection of drug agents and dosing
Measure response to adjustments in therapy
Determine need and optimal selection/dosing of IV therapy
(dobutamine, milrinone, nesiritide)
Optimize LVAD settings and wean patients from LVAD support
Determine optimal pacemaker settings in patients with AV sequential
pacemakers
Detect hemodynamic changes due to compensation, medication, and
diet compliance
Provide an adjunct to post-transplant myocardial biopsies
Treatment
NYHA=New York Heart Association; ACE=angiotensin-converting enzyme; LVAD=left
ventricular assist device
Adapted from Yancy C, Abraham W. Noninvasive hemodynamic monitoring in heart
failure: utilization of impedance cardiography. Congest Heart Fail. 2003:9(5):241–250.
Blood volume
Venous return
Posture
Heart rate
Preload
Exercise
Intrinsic
contractility
Contractility
Stroke
volume
Myocardial
compliance
Afterload
Arterial blood
pressure
Aortic
compliance
Age
Pulmonary
artery
compliance
Pulmonary
artery
compliance
∆Z
(Systole)
Body anatomy/composition
Cardiogenic pulmonary edema
Noncardiogenic pulmonary
edema
Specific
resistance
of thorax
Z0
Emphysema
Hematocrit
Orientation of erythrocytes
Specific
resistance
of blood
Figure 5. Contributing elements to thoracic impedance. Z0=baseline impedance;
∆Z=change in impedance
Adapted from Osypka MJ, Bernstein DP. Electrophysiologic principles and theory
of stroke volume determination by thoracic electrical bioimpedance. AACN Clin
Issues. 1999;10(3):385–399.
titrate these types of medications. A
summary of applications of ICG in HF
is presented in Table II, demonstrating
its broad clinical applicability.
In this supplement to Congestive
Heart Failure, we seek to further define
the role of ICG through a series of
original contributions. The study by
3
140 of 158.
3/30/04 9:40:43 AM
Q =Start of ventricular
depolarization
ECG
∆Z
B =Opening of pulmonic
and aortic valves
C =Maximal deflection
of the ∆Z/∆ t
X =Closure of aortic
valve
Y =Closure of pulmonic
valve
∆Z/∆t
O =Mitral opening/
rapid ventricular
filling
Figure 6. Fiducial points derived from electrocardiogram (ECG) and impedance
waveforms. ∆Z=change in impedance; ∆Z/∆t=first derivative of the impedance
waveform; PEP=preejection period; LVET=left ventricular ejection time
Figure 7. Impedance cardiography hemodynamic status report (BioZ ICG Monitor,
CardioDynamics, San Diego, CA)
Yung et al. (p. 7) validates the accuracy of ICG in patients with pulmonary
hypertension by comparing ICG to both
direct Fick method and thermodilution
4
CO. In doing so, the authors demonstrate the potential hazard of using
thermodilution as the only reference
standard for CO measurement. Parrott
et al. (p. 11) compare changes in ejection fraction by echocardiography to
changes in ICG parameters in established HF patients. Their findings demonstrate the ability of ICG to simply
and cost-effectively identify changes in
ventricular function. While pulmonary
artery catheterization in patients with
HF has been criticized and is largely
unproven by clinical trial, an estimated
2 million such catheters are sold worldwide each year.12 Springfield et al. (p.
14) illustrate the role of ICG in the
differential diagnosis of patients with
dyspnea. Although B-type natriuretic
peptide testing has gained wide attention recently as an aid to diagnose
HF in the emergency department,13
ICG may also have a diagnostic role
and provides additional value because
of its ability to identify appropriate
therapeutic options and monitor the
response to therapy in real time. Silver
et al. (page 17) report on the ability of ICG to replace pulmonary artery
catheterization, which has tremendous
cost implications for hospitals caring
for such patients. Vijayaraghavan et al.
(page 22) demonstrate the prognostic
role of ICG in patients with chronic
HF, and show strong association of ICG
changes to changes in functional status
and quality-of-life measures. Summers
et al. (page 28) provide a series of
case reports that illustrate ICG’s practical role in the initiation and titration of neurohormonal agents and their
patient-specific hemodynamic effects.
This compilation of studies adds to
the growing body of data supporting
the role of ICG in the management of
patients with HF. Within a year, the
results of two multicenter trials studying
key roles for ICG should be available:
PRospective Evaluation and identification of Decompensation by Impedance
Cardiography Test (PREDICT), conducted in patients with chronic HF;
and the BioImpedance cardioGraphy
(BIG) substudy of the Evaluation
Study of Congestive Heart Failure
and Pulmonary Artery Catheterization
Effectiveness (ESCAPE).14 PREDICT
specifically addresses the ability of
ICG-derived hemodynamic data to
141 of 158.
3/30/04 9:40:45 AM
identify patients at risk for death, hospitalization, or emergency department
visit. The BIG substudy will evaluate
the diagnostic and prognostic role of
ICG in both arms of a randomized,
controlled trial in pulmonary artery
catheter–hemodynamic-guided management of patients admitted with an
acute episode of HF.
There is now a compelling body of
literature that demonstrates the validity
of ICG using the most current technology. More and more studies have shown
the value of ICG in clinical settings
in addition to HF, including dyspnea,15
hypertension,16 and atrioventricular
sequential pacemakers.17 The studies
presented in this issue of Congestive
Heart Failure further define the role of
this valuable, noninvasive technology in
clinical medicine. It is likely that these
and other studies of ICG in HF will be
used to refine our understanding and
ability to assess patients and predict
prognosis, expanding on the concept of
the four quadrants presented in Figure
2. The impact of adding ICG hemodynamic data to the four quadrants
is depicted in Figure 9. Knowledge of
stroke index, cardiac index, systemic
vascular resistance index, and changes in fluid with thoracic fluid content
would likely provide more quantitative,
objective, and sensitive measurements
of hemodynamic factors, and has significant implications for the management of
patients with HF.
Incorporating this model of assessment into a proposed therapeutic algorithm is shown in Figure 10. Ideally,
a baseline measurement of ICG in
addition to other standard clinical
variables would be collected and utilized in combination to more precisely
assess a patient’s perfusion, congestion, and vasoactive status. This assessment would lead to a categorization
of the patient’s absolute or relative
change in hemodynamic profile, facilitating assessment of short-term risk for
adverse HF-related events. The change
in hemodynamic status and assessment
REFERENCES
1 Packer M. How should physicians view heart
failure? The philosophical and physiological
Cardiac output
(CO/CI)
Stroke
volume
(SV/SI)
Heart rate
(HR)
(-) Neg. Chronotropes
(+) Pos. Chronotropes
Preload
(∆TFC)
Afterload
(SVR/SVRI)
Contractility
(ACI, VI, PEP,
LVET, STR)
(-) Diuretics
(+) Volume Expanders
(-) Vasodilators
(+) Vasoconstrictors
(-) Negative Inotropes
(+) Positive Inotropes
Figure 8. Pharmacologic agent effect on cardiac output determinants and impedance cardiography parameters. CO=cardiac output; CI=cardiac index; HR=heart
rate; SV=stroke volume; SI=stroke index; ∆TFC=change in thoracic fluid
content; SVR=systemic vascular resistance; SVRI=systemic vascular resistance
index; VI=velocity index; PEP=preejection period; ACI=acceleration index;
LVET=left ventricular ejection time; STR=systolic time ratio
Congestion at Rest
NO
YES
TFC
NO
CI
YES
CI
Low
Perfusion
at Rest
TFC
A
B
Warm & Dry
Warm & Wet
(Low Profile)
(Complex)
L
Cold & Dry
C
Cold & Wet
SVRI
SVRI
Role of impedance cardiography
1.
2.
3.
4.
Establish baseline cardiac index (CI), systemic vascular resistance index (SVRI) and thoracic fluid content (TFC)
Monitor changes in CI to objectify perfusion assessment
Monitor changes in TFC to objectify congestion assessment
Monitor changes in SVRI to objectify vasoactive status
Figure 9. Model for clinical profiles in heart failure utilizing impedance cardiography hemodynamic measurements
of higher risk may lead to increased
clinical surveillance or a decision to
intervene to prevent a negative patient
outcome. In addition, ICG parameters
may aid in the assessment of a stable,
low-risk hemodynamic profile toward
the initiation and up-titration of neurohormonal agents that are often underprescribed but are known to improve
event-free survival.
Note: This supplement to Congestive
Heart Failure contains articles dealing
with ICG. Readers are reminded that
positive statements about the clinical
utility of ICG, and the BioZ ICG Monitor
in particular, are solely the opinions of
the authors and do not represent an official endorsement by Congestive Heart
Failure, its Editors or Editorial Board, or
the Heart Failure Society of America.
evolution of three conceptual models of the
disease. Am J Cardiol. 1993;71(9):3C–11C.
2 ACC/AHA guidelines for the evaluation and
management of chronic heart failure in the
5
142 of 158.
3/30/04 9:40:47 AM
ASSESSMENT
HEMODYNAMIC
PROFILE
PREDICTED
SHORT TERM
RISK
THERAPY
OPTIONS
COLD & WET
Lower SI / CI
Higher TFC
Baseline Evaluation
History & Physical
Lab tests
Diagnostic tests
ICG hemodynamics
+
Current Visit Evaluation
History & Physical
Lab tests
Diagnostic tests
ICG hemodynamics
Increase or add diuretic
Higher Risk
Need to decrease afterload?
Increase or add vasodilating agent
Higher SVRI
WARM & WET
Increase or add diuretic
Higher SI / CI
Higher TFC
Moderate Risk
Need to decrease afterload?
Higher SVRI?
COLD & DRY
Lower SI / CI
Lower TFC
Moderate Risk
Higher SVRI
Is patient currently taking neurohormonal
antagonists?
WARM & DRY
Higher SI / CI
Lower TFC
Lo w e r Risk
Higher SVRI?
1. ACEI or ARB
2. Beta blocker
3. Aldosterone inhibitor
If no, initiate
If yes, uptitrate toward target dose
Figure 10. Therapeutic algorithm for incorporating impedance cardiography (ICG) parameters into clinical assessment of
heart failure. SI=stroke index; CI=cardiac index; TFC=thoracic fluid content; SVRI=systemic vascular resistance index;
ACEI=angiotensin-converting enzyme inhibitor; ARB=angiotensin-receptor blocker
3
4
5
6
7
6
adult: executive summary. A report of the
American College of Cardiology/American
Heart Association Task Force on Practice. J
Am Coll Cardiol. 2001;38(7):2101–2113.
O’Connell JB. The economic burden of heart failure. Clin Cardiol. 2000;23(3 suppl):III6–III10.
Heart Disease and Stroke Statistics—2003
Update. Dallas, TX: American Heart
Association; 2002.
Nohria A, Tsang SW, Fang JC, et al. Clinical
assessment identifies hemodynamic profiles
that predict outcomes in patients admitted with heart failure. J Am Coll Cardiol.
2003;41(10):1797–1804.
Shah MR, Hasselblad V, Stinnett SS, et al.
Hemodynamic profiles of advanced heart
failure: association with clinical characteristics and long-term outcomes. J Card Fail.
2001;7(2):105–113.
Sageman WS, Amundson DE. Thoracic electrical bioimpedance measurement of cardiac
output in postaortocoronary bypass patients.
Crit Care Med. 1993;21(8):1139–1142.
8 Marik PE, Pendelton JE, Smith R. A comparison of hemodynamic parameters derived from
transthoracic electrical bioimpedance with
those parameters obtained by thermodilution
and ventricular angiography. Crit Care Med.
1997;25(9):1545–1550.
9 Greenberg BH, Hermann DD, Pranulis MF,
et al. Reproducibility of impedance cardiography hemodynamic measures in clinically
stable heart failure patients. Congest Heart
Fail. 2000;6(2):74–80.
10 Drazner M, Thompson B, Rosenberg P, et
al. Comparison of impedance cardiography
with invasive hemodynamic measurements
in patients with heart failure secondary to
ischemic or nonischemic cardiomyopathy. Am
J Cardiol. 2002;89(8):993–995.
11 Kazuhiko N, Kawasaki M, Kunihiko T.
Usefulness of thoracic electrical bioimpedance
cardiography: noninvasive monitoring of cardiac output. J Card Fail. 2003;9(suppl):S170.
12 Ginosar Y, Sprung CL. The Swan-Ganz catheter: twenty-five years of monitoring. Crit
Care Clin. 1996;12:771–776.
13 Mueller C, Scholer A, Laule-Kilian K, et al.
Use of B-type natriuretic peptide in the evaluation and management of acute dyspnea. N
Engl J Med. 2004;350:647–654.
14 Shah MR, O’Connor CM, Sopko G, et
al. Evaluation Study of Congestive Heart
Failure and Pulmonary Artery Catheterization
Effectiveness (ESCAPE): design and rationale.
Am Heart J. 2001;141(4):528–535.
15 Peacock F, Summers R, Emerman C. Emergent
dyspnea impedance cardiography-aided
assessment changes therapy: The ED IMPACT
trial. Ann Emerg Med. 2003;42(4):S82.
16 Taler SJ, Textor SC, Augustine JE. Resistant
hypertension: comparing hemodynamic management to specialist care. Hypertension.
2002;39:982–988.
17 Santos JF, Parreira L, Madeira J, et al.
Noninvasive hemodynamic monitorization for
AV interval optimization in patients with ventricular resynchronization therapy. Rev Port
Cardiol. 2003;22(9):1091–1098.
143 of 158.
3/30/04 9:40:49 AM
European Journal of Cardio-thoracic Surgery 14 (1998) 64–69
Non-invasive measurement of cardiac output during
coronary artery bypass grafting
Amram J. Cohen a , c ,*, Dimitri Arnaudov b , c, Deeb Zabeeda b , c,
Lex Schultheis d, John Lashinger e, Arie Schachner a , c
a
Department of Cardiovascular Surgery, The Edith Wolfson Medical Center, Holon, 58100 Israel
b
Department of Anesthesiology, The Edith Wolfson Medical Center, Holon, 58100 Israel
c
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
d
Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Medical Center, Baltimore, MD, USA
e
Department of Cardiovascular Surgery, Johns Hopkins Medical Center, Baltimore, MD, USA
Received 18 November 1997; revised version received 23 March 1998; accepted 15 April 1998
Abstract
Objective: A new device, using whole body bioresistance measurements and a new equation for calculating stroke volume has been
developed. Using this equation, an attempt was made to correlate whole body bioresistance cardiac output with thermodilution cardiac
output in patients undergoing coronary artery bypass grafting. Methods: Thirty-one adults undergoing elective coronary artery bypass
grafting were studied prospectively. Simultaneous paired cardiac output measurements by whole body bioresistance and thermodilution
were made at fiv time points during coronary artery bypass grafting: in anesthetized patients before incision (T1), after sternotomy (T2),
after opening the pericardium (T3), ten min post bypass (T4), and in the intensive care unit (T5). The patients had a mean of three
thermodilution cardiac outputs compared with a mean of three bioimpedance measurements at each time point. The bias and precision
between the methods were calculated. Results: There was good correlation between bioresistance cardiac output (nCO) and thermodilution
cardiac output (ThCO) measurements in both groups for all recorded times. The patients’ mean ThCO and nCO, as well as bias and
precision between methods were calculated. Mean ThCO ranged between 4.14 and 5.06 l/min; mean nCO ranged between 4.12 and 4.97 l/
min. Bias calculations ranged between −0.072 and 0.104 l/min. Precision (2 SD) calculations ranged between 0.873 and 1.228 l/min for
95% confidenc intervals. Pearson’s correlation ranged from 0.919 to 0.938. Conclusions: Cardiac output measured with the new device
correlates well with the thermodilution measurements of cardiac output during and immediately following coronary artery bypass grafting.
The overall agreement between the two methods was good. The new device is an accurate non-invasive method of measuring cardiac output
during coronary artery bypass grafting.  1998 Elsevier Science B.V. All rights reserved
Keywords: Cardiac output; Cardiac surgery; Thermodilution; Electrical bioimpedance; Bioresistance; Hemodynamics
1. Introduction
Hemodynamic monitoring continues to be an integral part
of peri- and postoperative care for patients undergoing cardiac surgery. Cardiac output (CO) is an important parameter
in these measurements. To date, the clinical tool used to
measure CO is the Swan–Ganz catheter using a thermodilution technique. However, the procedure is invasive, expensive, and may lead to complications [1,2].
* Corresponding author. Tel.: +972 3 5028723; fax: +972 3 5028735.
1010-7940/98/$19.00  1998 Elsevier Science B.V. All rights reserved
PII S1010-7940 (98 )0 0135-3
144 of 158.
A number of attempts have been made to determine CO
in a non-invasive manner [3–6]. Using thoracic electrical
bioimpedance techniques (TEB), moderate success has been
achieved in some clinical settings [3,4,6–8]. The technique
has been unsuccessful when applied to patients undergoing
cardiac surgery [9–11]. A new non-invasive cardiac system
device has been developed to utilize whole body bioresistance in a semi-empirical formula to determine the CO.
Accuracy of the CO measurement using this device was
established comparing thermodilution CO (ThCO) for
patients undergoing right and left heart catheterization
[12]. The purpose of this investigation was to compare the
A.J. Cohen et al. / European Journal of Cardio-thoracic Surgery 14 (1998) 64–69
new technology for measuring cardiac output with thermodilution measured CO in patients undergoing coronary
artery bypass grafting (CABG).
2. Patients and methods
2.1. Patients
Thirty-one patients undergoing elective, primary CABG
were prospectively studied; 15 at Wolfson Medical Center
and 16 at Johns Hopkins University School of Medicine.
Patients with aortic valve insufficien y, mitral and tricuspid
regurgitation or cardiac shunt were not included. The
CABG was similar in both hospitals. All patients underwent
a median sternotomy. Cardiopulmonary bypass was established using an aortic and single venous cannula. Diastolic
arrest was achieved with cold sanguineous potassium cardioplegia in both antegrade and retrograde fashion. The internal mammary artery was used to graft the left anterior
coronary artery (LAD) in all cases, and saphenous vein
grafts were used to bypass the other obstructed vessels.
Proximal anastomoses were performed using a partial
cross clamp. During the study, no patient had atrial flutte
nor atrial fibrillation
2.2. Protocol
The study protocol was approved by Edith Wolfson and
Johns Hopkins Institutional Review Boards. For each
patient demographic and clinical data was tabulated. In
each patient, CO measurements were taken at fiv different
time intervals during the procedure; after induction of
anesthesia but before incision (T1), after sternotomy (T2),
after creation of a pericardial pocket (T3), ten min after
weaning from bypass (T4), and immediately after arrival
in the intensive care unit (T5).
At each time interval, three adequate thermodilution measurements were made. All measurements were made at end
expiration during the respiratory cycle. Thirty-one patients
underwent 155 thermodilution CO measurements. Nineteen
were excluded because of 10% variation between the measurements. Simultaneously, three bioresistance measurements were taken at each interval, and their average was
considered the bioresistance CO.
Each bioresistance measurement took 20 s. Since these
measurements were continuous, there was no time between
measurements such that to achieve an average bioresistance
CO measurement took 1 min.
2.3. Thermodilution method
At both hospitals seven French true size thermodilution
catheters (Baxter Healthcare, Irvine, CA) were inserted in
the operating room after induction of anesthesia. The
proper location of the thermodilution catheter was con-
145 of 158.
65
f rmed by hemodynamic measurements and by postoperative chest roentgenograms. The pulmonary artery
catheter was connected to a thermodilution cardiac output
monitor (‘Horizon 2000’, Mennen Medical, Israel was used
at Wolfson Medical Center, and 7010 Series Marquette,
Madison, WI at Johns Hopkins Institutions). Ten milliliters
of room temperature 5% dextrose solution was injected
manually at end-expiration by an experienced anesthesiologist who was unaware of the bioresistance CO
results.
2.4. Whole body resistance method
NICaSy 2001, a non-invasive cardiac system device
(NICaSy 2001, Teledyne-NIM, LLC, North Andover,
MA, 510K, certificat number K942227), was utilized for
measuring bioresistance CO. Two proprietary designed,
NICaSy disposable electrodes (510K, certificat number
K972002) were applied to each wrist, and were then connected to the non-invasive cardiac system device. The system passed an AC current 30 kHz and 1.4 mA through the
whole body. It then measured the resistive portion of the
bioimpedance. Since the electrodes were placed on the distal aspect of the extremities, the current passed through all
major arteries and veins of the systemic circulation. As
such, the system measured the bioresistance of the systemic
circulation and allowed calculation of the stroke volume.
In 1992, Tsoglin and Frinerman and developed a new
semi-empirical formula for calculating the stroke volume
(SV) [13].
SV =
Hctcorr
H 2 DR
a+b
× kel × Kweight × IB × corr
×
Ksex, age
R
b
where Kel is a coefficien related to blood electrolytes,
Kweight is a ratio of a patient’s weight to ideal weight, IB
is the index balance equal to the ratio of the measured
extracellular flui (as measured from the baseline impedance) to the expected extracellular flui volume, HCTcorr
is a hematocrit correction factor proportional to measured
hematocrit, Ksex,age is a coefficien depending upon a
patient’s sex and age, H is a patient’s height, DR is the
incremental change of the resistive portion of the bioimpedance, R is the baseline whole body bioresistance and
a + b/b is the ratio of the sum of the systole time plus
diastole time divided by the diastole time derived from
the fin structure of the varying portion of the bioresistance.
Using whole body resistance measurements and the
above formula, SV was calculated and converted to CO.
2.4.1. Statistical analysis
The correlation between methods at each time point was
evaluated by Pearson’s correlation coeff cient and simple
linear regression. The degree of agreement between methods at each time point was evaluated by calculation of the
bias (mean between-method difference) and precision
(mean bias ±2 SD) [14].
66
Table 1
Patient demographic and clinical data
Males (N)
Females (N)
Average age (years)
Minimum age (years)
Maximum age (years)
Average weight (kg)
Minimum weight (kg)
Maximum weight (kg)
Average height (cm)
Minimum height (cm)
Maximum height (cm)
Temperature (°C)a
Number of vessels bypassed
Diabetes (%)
Hypertension (%)
a
Wolfson Medical
Center
Johns Hopkins
Medical Center
11
4
63.26 ± 9.94
46
79
73.13 ± 14.74
42
102
166.3 ± 10.4
140
183
26.8 ± 0.77
3.33 ± 0.96
46.67
53.33
13
3
64.2 ± 9.37
51
81
87.14 ± 13.29
59
108
176.92 ± 8.96
160
185
25.56 ± 0.403
2.55 ± 0.73
18.75
56.25
Lowest temperature during the operation.
3. Results
Thirty-one patients were evaluated in the study. Demographic and clinical data are shown in Table 1. The agreement between the average CO measured by bioresistance
versus thermodilution is shown in Table 2. The plot of the
regression analysis and difference versus mean for all measurements at all times in the 31 patients is shown in Fig.
1a,b. There was good correlation between ThCO and nCO
ranges of cardiac output during all phases of CABG and the
immediate postoperative period.
4. Discussion
The initial attempt to obtain CO by measuring the stroke
volume (SV) through thoracic electrical bioimpedance
(TEB) was performed by Kubicek [15] at the National Aeronautics Space Agency (NASA) where he introduced the
equation that became the basis for bioimpedance cardiography:
SV = r((dZ =dt)) × ((L2 T))=(Z02 )
where SV is related to the resistivity of blood (r), dZ/dt is
the firs peak of the derivative of the bioimpedance curve, L
is the distance between the electrodes, Z0 is the mean time
averaged thoracic bioimpedance and T is the ventricular
ejection fraction.
The equation was modif ed by Bernstein [16] to:
SV = VEPT × ( dZ =dt)=Z0 )T
where VEPT is a coeff cient that represents the volume of
electrically participating tissue. Using this equation, limited
clinical success has been achieved in determining CO using
bioimpedance techniques. Correlation to thermodilution
techniques have been achieved in healthy volunteers,
patients undergoing operations without cardiopulmonary
bypass, patients undergoing procedures in the cardiac
catheterization laboratory, and small numbers of intensive
care unit patients [3,4,7,17–20].
This equation has been problematic due to diff culties in
accurately computing VEPT, and its application becomes
impractical in a patient undergoing rapid hemodynamic
changes. Furthermore, determination of VEPT is dependent
upon the accurate placement of electrodes, which is not
always possible during open heart surgery. In addition,
this equation still depends upon the firs derivative of the
bioimpedance curve, which is a rapidly fluctuatin factor.
As a result, attempts to correlate bioimpedance CO in the
patient with thermodilution techniques during CABG have
been unsuccessful [9,10].
In addition to practical problems, there is a conceptual
problem in applying TEB to measure stroke volume. TEB
measurements of the bioimpedance includes both systemic
and pulmonary circulations. The two circulations cannot be
separated by measurements and variation in proportion of
these circulation will lead to inaccurate measurements of SV
[21,22]. This will not happen in whole body bioresistance
measurements where the systemic circulation dominates the
measurement and reflect LV SV.
The physiological and physical basis of bioimpedance
has been studied by many authors [23–25]. However, the
variation between different tissues and their complex structures make it difficu t to model their electrical behavior. In
terms of bioresistance, the body may be divided into the
‘blood compartment’ and ‘tissue compartment’. In the
Table 2
Agreement of CO by NICaS and thermodilution measurements (n = 31)
Time
Mean
ThCO
Mean
nCO
R/R 2
Regression
slopea
SEE
y intercept
Bias (mean betweenmethod difference)
(l/min)
Precision mean ± 2
SD (l/min)
After anesthesia
After sternotomy
After pericardiotomy
Immediately after bypass
ICU admission
4.19
4.14
4.14
5.06
4.71
4.19
4.12
4.24
4.97
4.62
0.93/0.86
0.92/0.85
0.93/0.86
0.92/0.85
0.94/0.88
1.128
0.898
0.915
0.967
0.891
0.328
0.185
0.326
0.367
0.306
−0.529
0.404
0.456
0.324
0.424
0.0086
−0.019
0.104
−0.0483
−0.072
−1.113–1.131
−0.863–0.823
−1.015–1.223
−1.234–1.138
−1.156–1.012
SEE, standard error of the estimate; ICU, intensive care unit.
a
y, cardiac output by NICaS 2001 bioimpedance; x, cardiac output by thermodilution.
146 of 158.
‘blood compartment’, conductivity is high and therefore a
current introduced into the body will travel primarily in this
compartment. The resistance changes in this compartment
as the blood volume varies in the great arteries due to pulsatile flow In the ‘tissue compartment’, there is less signif-
67
icant current flo and constant resistance. In each
individual, the resistance in the ‘tissue compartment’ is
determined by the patient’s height, lean body weight (muscle) to fat ratio, sex, age, body build, extracellular flui
volume, and electrolytes. In the ‘blood compartment’, the
Fig. 1. (a) Linear regression analysis comparing bioresistance measured cardiac output with thermodilution cardiac output for all averaged measurements in
the study. (b) Mean difference between bioresistance measured cardiac output and thermodilution cardiac output for all averaged measurements in the study.
147 of 158.
68
resistance is determined by hematocrit, electrolytes and the
pulsatile blood volume changes within the system. To measure SV, the pulsatile changes within the blood compartment
and baseline bioresistance must be measured accurately and
all other factors must be accounted for with appropriate
correction factors. The measurement of whole body bioresistance allows for the accurate measurement of pulsatile
changes and baseline resistance, and the proposed correction factors allow for the accurate calculation of the systemic bioresistance with which accurate LV SV can be
calculated.
The method utilized in this study has been proven accurate in patients undergoing catheterization [12]. It was also
shown to be accurate in a pilot study in patients undergoing CABG [26]. Compared with previous attempts to
correlate bioresistance with SV, the equation has major
advantages.
1. The equation does not depend on rapidly fluct ating
time derivatives of bioimpedance.
2. The equation uses empirically derived coeffici nts that
are obtained from laboratory values and simultaneously
measured bioimpedance values instead of the artifi ial
and difficu t to approximate VEPT.
3. The precise placement of the electrodes is not a critical
factor.
4. The NICaSy electrode arrangement is optimal to measure the left ventricular SV.
5. The respiration has almost no effect on the NICaSy CO
measurements in real time.
The new methodology showed good correlation between
thermodilution and bioresistance CO during all phases of
CABG and the immediate postoperative period in two independent hospital populations. Correlation was good in all
ranges, including low cardiac output. This fact is important
since patients undergoing CABG frequently have low cardiac output. This was even true in the immediate post cardiopulmonary bypass period where changes in the patients’
volume status, temperature, blood electrolytes and hemodynamics are changing rapidly. Such a correlation would be
impossible using previous techniques.
The new device has certain limitations. First, it cannot
measure CO while using diathermy. Second, the device is
sensitive to movement so that the patient cannot be manipulated while the CO is being measured. These two factors
require that the operation stop for the 20–30 s required to
measure CO. Finally, the device measures SV and multiplies
it by heart rate to measure CO. Arrythmias in which the
heart rate varies signific ntly will result in a non-representative measurement of CO.
In summary, a new device using a semi-empirical equation relating SV to changes in the resistive portion of the
patients’ bioresistance has been developed. Information
about specif c patient data, blood components and body
habitus inserted into the analytical software has signific ntly
improved the ability to calculate SV from the bioresistance
148 of 158.
data. The device allows accurate, easy to obtain, non-invasive CO during cardiac surgery. Within the above stated
limitations, we have confir ed the validity of the new bioresistance methodology in 31 patients who underwent CABG.
Bioresistance measured cardiac output correlated well with
ThCO during the CABG procedure and the immediate postoperative period.
Acknowledgements
This paper was prepared in consultation with Diklah
Geva who prepared the statistical calculations, and with
the technical assistance of Sally Esakov.
References
[1] Mermel LA, Maki DG. Infectious complications of Swan–Ganz
pulmonary artery catheters: pathogenesis, epidemiology, prevention,
and management. Am J Respir Crit Care Med 1994;149:1020–
1036.
[2] Choh JH, Khazei AH, Ihm HJ, Thatcher WC, Batty PR. Catheter
induced pulmonary arterial perforation during open heart surgery. J
Cardiovasc Surg (Torino) 1994;35:61–64.
[3] Denniston JC, Maher JT, Reeves JT, Cruz JC, Cymerman A, Grover
RF. Measurement of cardiac output by electrical impedance at rest
and during exercise. J Appl Physiol 1976;42:91–95.
[4] Smith SA, Russell AE, West MJ, Chalmers J. Automated non-invasive measurement of cardiac output: comparison of electrical bioimpedance and carbon dioxide rebreathing techniques. Br Heart J
1988;59:292–298.
[5] Reybrouck T, Fagard R. Assessment of cardiac output at rest and
during exercise by a carbon dioxide rebreathing method. Eur Heart J
1990;11 (suppl I ):21–25.
[6] Belardinelli R, Ciampani N, Costantini C, Blandini A, Purcaro A.
Comparison of impedance cardiography with thermodilution and
direct Fick methods for non-invasive measurement of stroke volume
and cardiac output during incremental exercise in patients with
ischemic cardiomyopathy. Am J Cardiol 1996;77:1293–1301.
[7] White SW, Quail AW, DeLeeuw PW, Traugott FM, Brown WJ,
Porges WL, Cottee DB. Impedance cardiography for cardiac output
measurement: an evaluation of accuracy and limitation. Eur Heart J
1990;11 (Suppl ):79–92.
[8] Teo KK, Hetherington MD, Haennel R, Greenwood PV, Rossall RE,
Kappagoda T. Cardiac output measured by impedance cardiography
during maximal exercise tests. Cardiovasc Res 1985;19:737–743.
[9] Sageman WS, Amundson DE. Thoracic electrical bioimpedance
measurement of cardiac output in postaortocoronary bypass
patients. Crit Care Med 1993;21 (8 ):1139–1142.
[10] Clarke DE, Raffi TA. Thoracic electrical bioimpedance measurement of cardiac output – not ready for prime time. Crit Care Med
1993;21 (8 ):1111–1112.
[11] Thomas AN, Ryan J, Doran BR, Pollard BJ. Bioimpedance versus
thermodilution cardiac output measurement: the Bomed NCCOM3
after coronary bypass surgery. Intensive Care Med 1991;17 (7):383–
386.
[12] Miller HI, Frinerman E, Tsoglin A, Frinerman A, Rosenschein U,
Keren G, Roth A, Laniado S. Non-invasive bioimpedance measurement of cardiac output: a validation study (Abstr). Second Conference of the Israel Heart Society, 1993:58.
[13] Tsoglin, A., Frinerman, E, N.I. Medical, Ltd. Non-invasive method
and device for collecting measurements representing body activity
[14]
[15]
[16]
[17]
[18]
[19]
and determining cardiorespiratory parameters of the human body
based upon the measurements collected. US Patent No. 5,469,859,
Nov. 28 1995.
Bland, J.M., Altman, D.G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
1986;Feb 8–1(8476):307–310.
Kubicek WG, Patterson RP, Witsoe DA. Impedance cardiography as
a non-invasive method of monitoring cardiac function and other
parameters of the cardiovascular system. Ann NY Acad Sci
1970;170:724–729.
Bernstein DP. A new stroke volume equation for thoracic electrical
bioimpedance: theory and rationale. Crit Care Med 1986;14:904–
909.
Donovan KD, Dobb GJ, Woods WP, Hockings BE. Comparison of
transthoracic electrical impedance and thermodilution methods for
measuring cardiac output. Crit Care Med 1986;14:1038–1044.
Salandin V, Zussa C, Risica G, Michielon P, Paccagnella A, Cipolotti
G, Simini G. Comparison of cardiac output estimates by thoracic
electrical bioimpedance, thermodilution and Flick methods. Crit
Care Med 1988;16:1157–1158.
Thomas SHL. Impedance cardiography using Sramek-Bernstein
method: accuracy and variability at rest and during exercise. Br J
Clin Pharmacacol 1992;34:467–476.
149 of 158.
69
[20] Wong DH, Onishi R, Tremper KK, Reeves C, Zaccari J, Wong AB,
Miller JB, Cordero V, Davidson J. Thoracic bioimpedance and Doppler cardiac output measurement: learning curve and interobserver
reproducibility. Crit Care Med 1989;17:1194–1198.
[21] Fuller HP. Evaluation of left ventricular function by impedance cardiography: a review. Prog Cardiovasc Dis 1994;36:267–273.
[22] Patterson RP. Sources of the thoracic cardiogenic electrical impedance signal as determined by a model. Med Biol Eng Comput
1985;23:41–417.
[23] Karnegis JN, Kubicek WG. Physiological correlates of the cardiac
thoracic impedance waveform. Am Heart J 1970;79:519–523.
[24] Sakamoto K, Muto K, Kanai H, Iizuka M. Problems of impedance
cardiography. Med Biol Eng Comput 1979;17:697–709.
[25] Paul E, Marik MD, Judy E. A comparison of hemodynamic parameters divided from transthoracic electrical bioimpedance with those
parameters obtained by thermodilution and ventricular angiography.
Crit Care Med 1997;25:1545–1550.
[26] Cohen A, Frinerman E, Katz M, Ezra S, Dotan A, Weissberg A,
Schachner A. Validity of bioimpedance hemodynamic measurements
during coronary artery bypass grafting (CABG) (Abstr). J Cardiovasc
Diag Proc 1996;13:57.
EVALUATION AND MANAGEMENT OF THE ELDERLY HOMEBOUND
PATIENT WITH CONGESTIVE HEART FAILURE
C. Gresham Bayne MD
BACKGROUND: The frail elderly patient population is characterized not only by age
demographics, but by the inadequacy of the office-based medical model to meet their
needs when they can no longer reach the physician’s office on a regular basis. In
addition, we are often called to see the new patient with presenting symptoms of heart
failure precipitated by one of many proximate causes.
Congestive Heart Failure (CHF) is the leading cause for hospitalization after the age of
65. There are one million admissions for CHF annually with 80% of the patients over 65
years of age. CHF is the most costly DRG in the United States with annual costs
exceeding $10 billion. The incidence and prevalence of CHF are rising in the population
with prevalence doubling each decade after age 45. The one year survival for Class III and
IV CHF elder patients is still only about 60%. The best objector predictor for survival
has been the cardiothoracic ratio on standard chest Xray.
There is a close and perhaps causative relationship between hypertension and CHF in the
elderly due to the complex neurohormonal reflexes involved (see below). As people age
the cardiovascular system normally increases its electrical impedance, has lowered betaadrenergic responsiveness, alters its myocardial energy metabolism, with the heart usually
showing impaired diastolic relaxation and compliance. The net effect is a marked
reduction in the cardiovascular reserve.
PRECIPITATING EVENTS:
These precipitating events must be evaluated first, before the patient’s failure can be
addressed. The most common of these is the failure to take their medications due to
financial, social, or cognitive problems. Despite the medication failure, the benefit might
be afforded us to switch them to the more acceptable geriatric approach to congestive
heart failure than has often been the case in the past. Briefly stated, the philosophy of
treatment has changed dramatically in past years from a diuretic-based ramp-up to the
primary use of ACE inhibitors and other afterload reducers. This modern approach will
be discussed more fully later.
Other precipitating factors common to our population include anemia, hyperthyroidism,
infection, new onset arrhythmias such as atrial fibrillation, change in diet, silent
myocardial infarction, worsening valvular disease, inappropriate use of the sodiumretaining NSAIDS, and environmental changes such as increased ambient humidity or
temperature, and stress. Obviously, these historical and clinical factors must be evaluated
in perspective.
The absence of a readily available medical records and EKG/Xray documentation
mandates the use of appropriate technology when patients present with symptoms
suggestive of failure. Even in non-acute settings, the ability to have local physicians copy
150 of 158.
and send forward medical records is so remote, it should not be anticipated that a chart is
forthcoming. In fact, many of the patients will be well-served by an entirely new
approach to their clinical symptoms, especially those on multiple medications.
Obviously, precipitating causes should be addressed before redirecting the primary
interest in care to the failure itself.
PHYSIOLOGY REVIEW:
The function of the heart is to pump arterialized blood forward under sufficient pressure
to meet the peripheral tissue metabolic needs. Given adequate oxygenation and a normal
hemoglobin, the cardiac output, defined as the pulse times the stroke volume, should
pump enough blood to prevent angina (ischemia to the heart), cold knee caps (ischemia to
the skin), oliguria or azotemia (ischemia to the kidney) and confusion (ischemia to the
brain). Thus, a quick iSTAT and pulse oximetry is required to rule out azotemia,
hypoxemia, and a metabolic acidosis demonstrated by the base deficit. Even a venous
sample with a normal base excess is adequate to prove the absence of a low perfusion
state.
NOTE that the above definition of inadequate cardiac output (CO) did NOT mention the
blood pressure at all! In fact, the only time hypotension is an emergency for bedbound
patients is when the diastolic pressure is below 60 mmHg and the patient experiences
angina or failure symptoms. Since coronary arterial blood flow is 85% during diastole,
the aortic root pressure is critical in determining myocardial perfusion. Thus, in the
emergency room, some patients in shock with angina find that dopamine is the best
analgesic!
To evaluate the failing heart, one needs to think in terms of myocardial efficiency, NOT
blood pressure. Efficiency of the heart is measure in terms of Oxygen Demand by the
heart as balanced against the Stroke Work Index. We will first address the Oxygen
Demand, which is determined by five factors, of which only three are clinically relevant:
the pulse, the mean arterial pressure and the wall tension index (as measured by the CXR
assessment of cardiomegaly). Myocardial oxygen demand is linearly increased as the
pulse increases over 100, dramatically increased by mean arterial pressure (usually
estimated by a value one-third the distance between the diastolic and systolic pressures),
and increases despite the same cardiac output with increases in the left ventricular enddiastolic volume (cardiomegaly).
Thus, the patient with a pulse of 110, a blood pressure of 140/100 and cardiomegaly on
CXR is demanding much more oxygen for their heart than a patient of normal heart size
and vital signs. When the presenting symptoms include angina, the approach to failure is
first to treat the increased demand by using nitrates or nifedipine to lower the MAP, slow
the pulse, and vasodilate the coronary arteries. Once stabilized, the failure may be
addressed.
151 of 158.
Stroke Work Index (SWI) is a complicated clinical measurement requiring the use of a
pulmonary arterial catheter, but can be thought of in clinical terms quite simply.
Remember, the heart does not have to create higher than normal pressures to increase its
cardiac output (CO) if the peripheral vascular resistance (PVR) is reduced. The SWI is
calculated basically by multiplying the cardiac output times the PVR. The human heart is
a very efficient flow generator, but a very inefficient pressure generator. Thus, oxygen
demand to increase cardiac output is efficiently handled, but requirements for
hypertension are disastrous. It is important to note that the failing heart may, due to its
inefficiencies and maladaptive neurohumoral reflexes, simply increase the blood pressure
without generating much increase in the CO. Thus, hypertension is the enemy of both
ischemic heart disease and congestive heart failure.
The physiological changes in the elderly which occur in the chronic state of heart failure
are numerous and complex. They include:
1. Increased renin levels, which increases the conversion of Angiotension I to
Angiotension II, the most potent vasoconstrictor known
2. Increased renin levels promoting the kidney to increase aldosterone leading to
increased retention of sodium and water
3. Increased adrenergic background tone by the adrenal cortex with elevated
circulating norepinephrine levels and inappropriate hypertension
4. Decreased levels of adenyl cyclase leading to decreased levels of cyclic AMP
which can lead to lower protein kinase, calcium entry levels and calcium reuptake by the myofibrils
The net result is a chronic “fight or flight” background level of physiologic stress which
ultimately leads to cardiac failure. I think of my elderly patients as on an endogenous
catecholamine drip producing more hypertension than forward flow. Since it is unlikely
that the patient will ever be “normal,” it is useful to think in terms of therapy to maximize
forward blood flow at the lowest physiologic cost.
TREATMENT PRINCIPLES:
To maximize the cardiac output, we typically intervene with the three major forces we
can easily effect with medication: Preload, Contractility, and Afterload.
Preload: The hypovolemic patient may be in shock or hypotensive simply because of
fluid depletion. This is often easily discerned by the history, physical and iSTAT testing,
but often problematic in the bedbound, demented patient. The hypervolemic patient is
often much harder to evaluate, since chronic rales, interstial lung disease, emphysematous
changes, obesity, patient intolerance to the exam, may all affect our confidence level.
Therefore, some simple in-home testing may be useful. The CXR showing cardiomegaly
is strong support for a chronic hypervolemic state and elevated left ventricular filling
pressures. The presence of LVH or left ventricular strain pattern in the EKG (R waves in
V5 and V6 add up to more that 25mv) is often associated with overload.
152 of 158.
A final test, done only in the normotensive or hyptertensive patient, is the so-called
“Chatterjee” test when one monitors the pulse before and after a single dose of sublingual
nitroglycerine. Since NTG predictably lowers the pulmonary capillary wedge pressure
(PCWP) (left ventricular filling pressure), the hypervolemic patient will NOT have a
tachycardic change in pulse with NTG. The normal or hypovolemic patient will increase
their pulse by 10% or more. Thus, the absence of tachycardia in response to NTG is de
facto evidence of increase PCWP and indicates the need for (temporary) diuresis.
The use of nitrates alone to control preload and subsequent CHF symptoms is
complicated by the tolerance that all patients develop for the longer acting versions. The
use of nitrates in the elderly is most often useful on the acute housecall when the patient
has suddenly decompensated and has hypertension in addition to the symptoms of CHF.
One must be cautious in the home when using nitrates to be able to place the patient in
recumbency before administration, as the “nitrate syncope” syndrome is much more likely
in our patients.
The use of diuretics has been the mainstay of treatment for both hypertension and failure
for some forty years, despite the fact that diuretics have yet to be shown in controlled
prospective clinical trials to increase survival rates. Currently, some 35-40% of all
seniors over the age of 65 are taking a diuretic in some form. However, cumulative
studies are now overwhelming in their condemnation of the indiscriminate application of
either loop or thiazide diuretics due to the following known side effects:
1. 60% of all toxic drug reactions are due to diuretics in the elderly
2. diuretics are the most common drug reaction requiring hospitalization
3. both types cause increased serum cholesterol, although subsequent cardiac
disease is unproven
4. both cause increased uric acid levels with subsequent tophaceous gout
5. both cause glucose intolerance in Type 2 prone patients, and the instigation of
Type 2 diabetes is NOT reversible by stopping the diuretic (bumex less than
lasix)
6. both cause deafness, even in chronic oral dosing regimes (bumex less than
lasix)
7. both cause decreased calcium absorption and worsen osteoporosis (loop worse
than thiazides)
8. both cause behavioral risks with incontinence in the female, urinary retention
in the male
9. both can lead to hyperkalemia in the elderly as well as hypokalemia
10. both can cause severe hyponatremia in the elderly
Thus, the use of diuretics is now considered second or third line therapy, after more
contemporary treatment for both hypertension and/or failure has failed.
Contractility: The state of inotropy of the heart in failure patients is compromised by
many of the maladaptive neurohormonal changes discussed above as well as the loss of
cardiac tissue from current ischemic heart disease and past infarcts. Unfortunately, the
153 of 158.
use of phosphodiesterase inhibitors in the elderly to increase inotropy has been clinically
unsuccessful and the use of adrenergic agents such as dopamine and dolbutamine requires
monitoring and infusion therapy along with the hassles of an intravenous. Early studies
on NY Stage IV CHF patients on home dobutrex drips have prolonged life but at a
significant cost on the quality of life.
What we have learned in the ICU from the two adrenergic agents in most common use
today is important. Whereas, dopamine causes a more potent increase in inotropy and
blood pressure, it has little or no effect on PCWP, increases ventricular irritability, and
requires much more care. Dolbutamine, on the other hand, not only has less toxicity at
the same cardiac output, but lowers the PCWP dramatically, thereby restoring the normal
cardiac anatomy and reducing myocardial oxygen demand.
To reduce the oxygen demand of the heart, the eighties and early nineties saw an increase
in the use of beta blockers even in failure patients. Unfortunately, the use of beta
blockers to control hypertension in the elderly involves much more toxicity, and (other
than post MI) has shown to increase mortality due to side effects, including an increase in
cholesterol for all age groups.
The use of digoxin remains controversial in patients without the need for rate control in
atrial fibrillation. Although digoxin is the oldest and most proven inotropic agent, and
the only one which does so while decreasing the wall tension index (cardiomegaly), its
toxicity and need for periodic serum measurements makes it of much more problematic in
the demented, homebound elderly CHF patient. A NEJM study in 1997 on Class II or III
CHF patients showed no change in mortality with digoxin therapy in addition to ACE
inhibition (see below), but a significant improvement in functional status and decrease in
hospital days. Most patients with an ejection fraction of over 45% will do well on
digoxin alone for control of symptoms of CHF. Obviously, the ability to measure directly
an improvement in cardiac output with digitalis would allow better patient selection in the
future.
Suffice to say, patients with cardiomegaly and/or a S3 gallop should have a trial on
digoxin alone before using multiple drug regimes. Finally, no one argues about the use of
digoxin to control rapid ventricular response in the patient with atrial fibrillation.
Afterload: If the SWI can be reduced by reducing the MAP, one might be able to both
increase the forward flow of blood (CO) as well as decrease the myocardial oxygen
demand. This would be the best of both worlds! Indeed, we now know that peripheral
vasodilatation has less morbidity and better survival rates with better functional status
than conventional beta blocker/diuretic regimes. The geriatric meetings are replete with
studies confirming the observation that the ACE inhibitors by themselves can often
control hypertension and should be used for patients with risk of CHF (such as LVH or
cardiomegaly on CXR), and especially diabetics even when they are asymptomatic.
154 of 158.
By using ACE inhibition at lower doses beginning at night, one can often avoid
orthostatic hypotension and preserve the peripheral vasomotor reflexes so important to
prevent side effects. Many patients can be converted from 2-4 older medications
(digoxin, lasix, KCL, tenormin) to a single dose of benzapril 10 mg hs. Titrating the
benzapril up to 80 mg daily may control the blood pressure while relieving the symptoms
of failure as well. The therapeutic effects must be evaluated by the patient’s subjective
reports of exercise tolerance or orthopnea, as well as objective weights. If CHF is to be
controlled on ACE inhibitors alone, the patient will lose weight on it.
In addition, the Cooperative North Scandinavian Enalapril Survival Study showed that
ACE inhibition improves symptoms and exercise tolerance, while affording significant
survival benefits among patient with severe CHF. These agents not only decrease
afterload, but preload as well, and deter activation of the neuroendocrine system leading
to reduced risk of hypertensive episodes, stress reduction, less fluid retention, and perhaps
less arrhythmogenicity.
Before reverting back to the diuretics which some patients simply must have, adding a
calcium channel blocker in angina patients or alpha receptor blockade may be useful to
improve inotropy by increasing coronary blood flow and further decreasing the peripheral
vascular resistance. In addition, the newer calcium-channel blockers such as amlodipine
have been shown to increase end-diastolic ventricular relaxation, promoting ventricular
filling and increasing stroke volume at the same filling pressures. Once the afterload is
under control, the SWI is decreased enough to usually require no further therapy
requirements for heart failure. Of course, if you cannot know the cardiac output, you
cannot calculate afterload, so using BP as a surrogate if fraught with undertreatment
biases. Should further therapy be required, a diuretic may be cautiously introduced since
their effect is additive to both ACE inhibition and calcium channel blockade.
What has been lacking in both the critical care and outpatient arenas is an ability to know
the systemic vascular resistance. Thus, we are still basing our clinical decisions on nonphysiological parameters: i.e., mean arterial pressure. Although MAP may be a useful
guideline for stroke prevention, many patients with borderline /low cardiac outputs have
normotension while their SVR is markedly elevated. We cannot know this without
measurement of the cardiac output, which heretofore has been invasive and expensive
with the pulmonary arterial catheter. The advent of non-invasive, inexpensive cardiac
impedance studies in the outpatient and other settings has opened up a dramatic
opportunity to not only titrate therapies toward maximum lowering of the SVR to
improve cardiac output, but also to titrate therapies both acutely and chronically toward
the best cardiac efficiency factors, since impedance can track the pre-ejection period and
minimize isovolumetric contraction time.
The Use of Cardiac Impedance in the Home-Bound Patient
Background: For over twenty-five years it has been possible to provide a non-invasive
monitoring system capable of measuring the trends in cardiac output, stroke volume,
155 of 158.
ejection fraction and total intra-thoracic fluid. Since the chest is essentially a box
containing a variable amount of non-conducting air in the lungs and airways, and a
variable amount of salt water in tissue and blood, measuring the resistance to the flow of
electricity through the chest should correlate to fluctuations in air and/or fluid.
In fact, this measurement, known as impedance, can easily be measured across the chest
by a series of paired electrodes placed on the neck and lateral chest wall. The static
values measured on an insensitive scale correlate to Total Fluid Content or TFC of the
chest and will change over time with things like pleural effusions, extravascular lung
water, and congestive heart failure. Using a very sensitive scale, one can see alterations
in impedance conforming to the pumping action of the heart during each systole. Since
the heart pumps blood out of the chest once the aortic valve opens, the velocity of this
pumping action can be calculated from the first derivative of impedance or dZ/dT. The
peak value of dZ/dT correlates closely with the stroke volume (SV) of each heartbeat.
Since the cardiac output is easily calculated by multiplying the SV time heart rate, one
can watch online as therapies affect the cardiac output.
In the 1960’s impedance research was done primarily in Israel where engineers were best
in signal acquisition technology or Russia, where cognitive strengths were focused on
complex algorithms to reduce the signal noise from motion or respiratory artifact.
Recently, a combination of these talents has come together to produce an FDA-approved
non-invasive, whole-body impedance measurement which has the distinct advantage of
not having to undress female patients (electrodes are placed on two extremities). There
continues to be controversy over which is “best,” trans-thoracic or whole body
impedance. In addition, there remains argument over which proprietary algorithm should
be used. Proponents of the Sramek Equation believe selecting the single best systole for
detailed examination and reporting is superior to averaging data from many systolic
waveforms during a one minute sample (for example) analyzed by the Kubicek Equation.
The second derivative of impedance varies with the acceleration of the blood as it exits
the left ventricle and correlates closely with the inotropic state of the heart. Since
electrodes can easily sense the EKG tracing of a patient, one can pinpoint the R wave and
measure the time between the initiation of systole and the time when the aortic valve
opens, sensed as the onset of the acceleration of blood or the second derivative of
impedance. This period is called the PEP or Pre-Injection Period, a time when the heart
is maximally consuming oxygen during systole, but doing no useful work, since no
bloodflow has yet occurred in the aorta. The longer the PEP, the more inefficient the
heart is for any given cardiac output. Examples of clinical conditions which prolong the
opening of the aortic valve or PEP are hypertension, myocardial dyskinesis, ventricular
aneurysm, and aortic stenosis. Finding the maximum cardiac output for the minimum
PEP has now become routine in setting time intervals for two channel pacemakers.
Since the closing of the aortic valve is also marked by an abrupt deceleration of blood
flow in the aorta, one can use impedance to mark the closure of the aortic valve. Thus,
the LVET, or Left Ventricular Ejection Time, when the left ventricle is actually pumping
156 of 158.
blood, can now be tracked with non-invasive monitoring. More importantly, it turns out
that clinical studies of patients with heart disease, including mitral valve disease and atrial
fibrillation, have shown an extremely close inverse correlation of the left ventricular
ejection fraction and the PEP/LVET ratio.
Clinical Significance in the Home Bound Patient:
Many patients over the age of 80 are not only homebound, but extremely demented and
frail. Not only do they have a high prevalence of congestive heart failure, the number one
cause of death in the elderly, but they cannot get to the cardiologist for appropriate
studies. Even if they could, they cannot tolerate a treadmill or other means of evaluating
their heart. For those that can make it to the cardiologists office, Medicare policy limits
payment for the gold standard of doppler ultrasound to annual usage in most cases. Thus,
the reality is that the homebound patient is usually not followed objectively in the
management of their cardiac function during therapeutic changes directed at blood
pressure or congestive heart failure.
It is clear now that afterload and CHF control are critical to maximizing both the lifespan
and the functional status of these patients. Unfortunately, we have not had access to
objective measurements for any of our therapies in the home. For instance, it is generally
agreed that ACE inhibition therapy should be maximized to a physiologic endpoint, but
daily weights, clinical exam, and pulse oximetry are late findings and too non-specific
with too many covariables to give us confidence.
The use of cardiac impedance as a non-invasive static and dynamic monitoring system to
document cardiac function in the home represents a break-through technology allowing
us to pursue better quality of care for these patients. Using the priniciple of transthoracic
bioimpedance, intensivists and surgeons have known for over a decade that cardiac output
and ejection fractions can be monitored with as much reliability as thermodilution SwanGanz catheter measurements. Not only is the data correlated with direct Fick
measurements as well as thermodilution, the bioimpedance measurements show less
inter-operator variability and higher reproducibility for trend analysis with a given patient.
In summary, the use of FDA-approved portable bioimpedance devices (paid by Medicare
at about $40/test), such as the (www.cardiodynamics.com ), the HemoSapiens
(www.hemosapiens.com) or NIMedical (www.ni-medical.com) offers another step
forward in low-cost, comprehensive testing and monitoring of patients outside of the
hospital environment: whether in the home, the outpatient clinic, or the physician’s
office. It is likely that the ease of use and higher physiological parameters measured will
make this monitoring a mainstay for both future cardiologists and housecall physicians,
although the high capital costs of machines and low per-test reimbursement have limited
growth in the fee-for-service sector.
157 of 158.
Examples of theoretical usage of non-invasive measurement of transthoracic impedance in the homebound patient:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
158 of 158.
Evaluation of the new patient in heart failure
Weaning CHF patients off their high-dose beta-blockers and calcium
channel/diuretic therapies
Titrating to the maximum dose of ACEI/ARB consistent with renal
impairment
Monitoring diuretic therapies in the bedbound patient or others that
cannot be weighed
Adjustment of pacemaker rate settings for maximum cardiac output in
marginal patients
Titrate hypertensive therapies to the point of maximum cardiac
efficiency
Evaluate the patient with frequent PVCs, AF or other arrhythmia
affecting output
Evaluate starting or stopping of digoxin therapy
Safety in rapid rehydration or transfusion therapy at home
Monitoring the terminal patient remotely to aid the family in planning,
etc.
Monitoring patient compliance with dietary restrictions or fluid
hydration overnight
Use in pharmcologic stress test for IHD evaluations in the bedbound
patient
Monitoring of pleural effusions over time or during thoracentesis
Screening for CHF in patients with poor histories and weakness
Evaluate Cyclosporine or EPO-induced hypertension
Short term stabilization of the patient in pulmonary edema

NICaS Study Review 2014

Transcription

Similar documents

Introduction to Electrophysiology

2009 Newsletter

Pericystic Fibrosis of a Cardiac Hydatid Cyst

Physiology of the vasovagal reflex

LifeForce August 2009 Newsletter copy

Cardiac Arrest Top10 Things to Know

Darren Emery AEE Iridium NEXT presentation March 2015

Cardiac Tool - Magnetic Resonance Development and Application

Nerves of the heart: a comprehensive review with a

Unspoken “Rules” of EMS - New York State Volunteer Ambulance