A Prospective, Randomized, Controlled Trial of Autologous Platelet-Rich

FEATURE
A Prospective, Randomized, Controlled
Trial of Autologous Platelet-Rich
Plasma Gel for the Treatment of
Diabetic Foot Ulcers
Vickie R. Driver, DPM, MS, FACFAS; Jason Hanft, DPM, FACFAS; Carelyn P. Fylling, RN, MSN; Judy M. Beriou, RN,
MHA; and the AutoloGel™ Diabetic Foot Ulcer Study Group
Nonhealing diabetic foot ulcers are a common cause of amputation. Emerging cellular therapies such as platelet-rich plasma gel provide
ulcer management options to avoid loss of limb. The purpose of this prospective, randomized, controlled, blinded, multicenter clinical study
was to evaluate the safety and efficacy of autologous platelet-rich plasma gel for the treatment of nonhealing diabetic foot ulcers. One hundred, twenty-nine (129) patients were screened; 72 completed a 7-day screening period and met the study inclusion criteria.Patients were
randomized into two groups — the standard care with platelet-rich plasma gel or control (saline gel) dressing group — and evaluated
biweekly for 12 weeks or until healing. Healing was confirmed 1 week following closure and monitored for another 11 weeks. An independent audit led to the exclusion of 32 patients from the final per-protocol analysis because of protocol violations and failure to complete treatment. In the 40 wounds per-protocol group, 13 out of 19 (68.4%) of the platelet-rich plasma gel and nine out of 21 (42.9%) of
the control wounds healed. After adjusting for wound size outliers (n = 5), significantly more platelet-rich plasma gel (13 out of 16, 81.3%)
than control gel (eight out of 19, 42.1%) treated wounds healed (P = 0.036, Fisher’s exact test). Kaplan-Meier time-to-healing also was
significantly different between groups (log-rank, P = 0.0177). No treatment-related serious adverse events were reported and bovine
thrombin used in the preparation of PRP did not cause Factor V inhibition. When used with good standards of care, the majority of nonhealing diabetic foot ulcers treated with autologous platelet-rich plasma gel can be expected to heal.
KEYWORDS: platelet-derived growth factors; platelet releasate; platelet-rich plasma; PRP; platelet gel; diabetic foot ulcers; Factor V;
bovine thrombin, autologous growth factors
Ostomy/Wound Management 2006;52(6):68–87
ore than 20.8 million persons in the US have
diabetes mellitus; 2002 data estimates from
the Cen ters for Disease Con trol and
Prevention indicate that 82,000 lower limb amputations
M
were performed in pers ons with diabetes.1 Characteri s tic
pathological changes attri buted to autonomic and sensory
neuropathy, often combi n ed with vascular disease, lead to a
high-risk situati on for the person with diabetes.2,3 Pers ons
Dr. Driver is Director, Clinical Research, Center for Lower Extremity Ambulatory Research, Dr. William M. Scholl College of
Podiatric Medicine, Rosalind Franklin University of Medicine; and Chief, National Center for Limb Preser vation, Advocate Lutheran
General Hospital, Niles, Ill. Dr. Hanft is Dire c t o r, Doctors Research Network, South Miami, Fla. Ms. Fylling is Vice President,
Professional Ser vices, and Ms. Beriou is Senior Research and Clinical Coordinator, Cytomedix, Inc. Please address correspondence to: Carelyn Fylling, Cytomedix, Inc., 416 Hunge r f o rd Drive, Rockville, MD 20850; email: [email protected].
Carelyn Fylling and Judy Beriou disclose they are employees and/or stockholders of Cytomedix, Inc.
68
OstomyWound Management
who have had such pathology and ex peri ence trauma or
published.7-21 In vi tro research has shown that platel ets
infection are at high risk for devel oping ulcera ti on of the
contain com ponents and properties for wound healing22;
4
foot or ankle. Pecora ro et al documen ted the causal pathlikewi s e , plasma contains fibrin matri x .23
way of an amput a ti on and found that in 81% of cases
In 2001, Margolis24 p u bl i s h ed a retro s pective study
faulty wound healing con tri buted to amput a ti on .
analyzing the treatment re sults of 26,599 patients wi t h
Healthcare practi tion ers should utilize wound tre a tments
diabetic neu ropathic foot ulcers who had been tre a ted
that can redu ce the ra te of f a u l ty wound healing; thus,
with an auto l ogous platel et releasate. The re sults su ggest
preven ting amputations.
that platel et rel e a s a teprovi ded with standardized care was
According to the Am erican Di a betes As s ociation, m ore
more ef fective than standard care alone.
than 60% of n on traumatic lower-limb amputations occur
The purpose of the current stu dy was to determine the
in people with diabetes; the ra te of amputation for people
safety and ef fectiveness of tre a ting diabetic foot ulcers
with diabetes is 10 times high er than for people wi t h o ut
with PRP gel versus a con trol tre a tm ent (normal saline
d i a betes; Mexican Am ericans are 1.8 times as likely, n on gel ) . The primary objective of the 12-week stu dy was to
Hispanic Blacks are 2.7 times as likely, and Am erican
compare the safety and inciden ce of com p l ete wound cl oIndians are three to four times as likely to ex peri ence
sure bet ween PRP gel- and con tro l - treated wounds at the
lower-limb amputations. Amputation ra tes are 1.4 to 2.7
end of the stu dy. Secon d a ry obj ectives inclu ded com p a rtimes high er in men than women with diabetes.5 Fryk berg
ing the ra te of wound healing during the 12-week study
6
et al cites a 1998 stu dy of 67,000 diabetes-rel a ted lower
and inciden ce of wound rec i d ivism among healed ulcers
ex trem i ty amput a ti on (LEA) and a similar stu dy that
during a 3-month fo ll ow-up peri od . Sa fety va ri a bl e s
resulted in a total of 984,000 hospital days, e ach length of
inclu ded adverse events, s erious adverse events, and cl i n istay avera ging 15 days. Nonhealing diabetic foot ulcers
cal labora tory tests.
and the re su l ting po ten tial amputations pre s ent sign i f iStudy Design and Methods
cant costs to the healthcare system and redu ce patient
qu a l i ty of life.
This pro s pective, ra n domized, con tro ll ed , do u bl e The goal of diabetic foot ulcer treatment is to obtain
blinded, mu l ti cen ter trial was condu cted under the US
wound cl o su re as exped i ti o u s ly as possible. Accepted therFood and Drug Ad m i n i s tration (FDA) Inve s ti ga ti onal
apeutic obj ectives and standards of c a re for diabetic foot
Device Exem pti on (IDE) reg u l a ti ons. Constella Clinical
ulcers inclu de wound debri dement,
pressu re rel i ef in the wound area,
Ostomy Wound Management 2006;52(6):68–87
a ppropri a te wound managem en t
KEY POINTS
( eg, moist wound healing), infection
• While some foot ulcers in persons with diabetes mellitus will heal within a reasonm a n a gement, isch emia manageable period of time when optimal care is provided, others will not; thus, increasing
m en t , medical managem ent of
the risk of complications such as amputation.
com orbidities, and su r gical man• To test the safety and efficacy of autologous platelet-rich plasma (PRP) gel in the
agem ent as needed .6 Emer ging celtreatment of chronic nonhealing diabetic foot ulcers, researchers randomly
lular therapies su ch as platelet - ri ch
assigned persons with nonhealing diabetic foot ulcers to a 12-week treatment periplasma (PRP) can have an adjuncod of offloading and control or offloading and PRP gel.
tive role in a standardized , qu a l i ty
• No short-term or long-term (24-week) safety concerns were observed but a variety
tre a tm ent plan.
of protocol violations reduced the power of the study to detect statistically significant differences.
P l a tel et releasates, i n cluding
• In the total group, 68% of PRP gel and 43% of control gel treated wounds healed.
mu l tiple growth factors, have been
When standardized for baseline size, the proportion of wounds healed was signifiused to treat wounds since 1985. In
cantly different (81% versus 42%).
vivo pro s pective con tro ll ed studies
• The results of this controlled study suggest that, as part of an overall program of
as well as retro s pective and co s t
optimal care, PRP gel is more effective than non-PRP gel. The results also confirm
ef fectiveness studies documenting
the appropriateness of using moisture-retentive control dressings in all chronic
the ef fect of this thera py have been
wound studies.
June 2006 Vol. 52 Issue 6
69
Informatics (Du rh a m , NC) served as the Cl i n i c a l
Research Organization (CRO) to implem ent and monitor
the tri a l , gather the data into a cen tral database, and audit
the data. A data safety monitoring board provi ded eva lua ti ons of the safety - rel a ted events thro u gh o ut the tre a tment phase of the study. Independent stati s ticians were
con tracted to initi a lly power the stu dy, devel op the statistical plan, and analy ze the safety and ef fectiveness data.
Fourteen (14) investi ga tive sites from ac ross the co u ntry participated in the stu dy. Sites inclu ded wound care
physicians’ and podiatrists’ offices, o utpatient wound care
cen ters, a univers i ty - b a s ed co ll ege of pod i a tric medicine
clinic, Vetera n’s Ad m i n i s tration wound care clinics, and
an Army hospital limb pre s erva ti on progra m . E ach site
obt a i n ed IRB approval to condu ct the stu dy.
Study el i gi b i l i ty. Pers ons with type 1 or type 2 diabetes
bet ween the ages of 18 and 95 with an ulcer of at least 4wee k s’ durati on were el i gi ble for the stu dy if t h ey met
ad d i ti onal inclusion / exclu s i on cri teria: h em oglobin A1C
<12; index foot ulcer located on the plantar, medial, or
lateral aspect of the foot (including all toe su rf aces); and
wound area (length x width) measurem ent bet ween 0.5
cm2 and 20 cm2, inclusive. Wounds located under a
Charcot deform i ty had to be free of ac ute ch a n ges and
must have under gone appropriate stru ctu ral consolidation. The index ulcer had to be cl i n i c a lly noninfected
( a l t h o u gha culture was obtained, infecti on was diagn o s ed
thro u gh clinical signs and sym ptoms ra t h er than cultu re
results25) and full - t h i ckness wi t h o ut ex po su re of bone,
mu s cl e , ligaments, or ten dons (Un iversity of Tex a s
Tre a tm en t - Ba s ed Di a betic Foot Classification System :
Grade 1 A26).
The protocol requ i red that post debridem ent the ulcer
would be free of necro tic debri s , forei gn bodies, sinus
tract s , tu n n el i n g, and undermining; compri s ed of healthy
vascularized tissue; and at least 4 cm from any ad d i ti onal
wound. Addition a lly, the limb had to have adequ a te perfusion as shown by examination and non-inva s ive va s c ular te s ting ankle brachial index (ABI) and toe brachial
index (TBI). Women of childbe a ring age could not be
pregnant or lact a ti n g ; both men and women had to be
wi lling to use a med i c a lly accepted form of bi rth con trol
throughout the trial and for 6 months fo ll owi n g. Pa ti ent
h i s tory, physical examinati on (including a SemmesWei n s tein monofilament test for neu ropathy), and bl ood
for baseline labora tory studies were obtained.
70
OstomyWound Management
Approved , inform ed , s i gn ed consent stipulating that
the patient was able to com p ly with all spec i f i ed care and
visit requ i rem ents was sec u red from the patient, caregiver, or legal repre s en t a tive before stu dy en ro ll m ent. The
Investi ga tor documen ted reasonable ex pect a ti on that the
patient was medically stable and capable of com p l eting
the stu dy. Stu dy exclu s i on cri teria are listed in Table 1.
After meeting all initial inclusion cri teria and signing the
informed consent, a ll pati ents completed a 7-day screening-peri od. This included initial excision / debri dem ent,
baseline wound measurements and evaluation, and application of the control saline gel to the wound. For standardization, s h a rp debri dement guidelines were provi ded as
p a rt of the protocol. Patients, a family member, or other
designated parties were provi ded supplies and instru cted to
change the dressings once midway thro u gh the screening
peri od . The patient also was requ i red to use a fixed anklefoot orthoses that could be removed for the dressing
change and at night. Crutches or a walker were used for
added safety. Screening data were captured on case report
forms (CRFs) for data analysis. Pati ents whose wounds
reduced in area by >50% du ring the screening peri od were
not ra n domized to treatment and discontinued from any
furt h er study participati on because they appe a red to be
able to heal wi t h o ut more advanced interven ti on .
Random i z a ti on and blinding procedu res. The ra ndom i z a ti on sch edule was el ectron i c a lly gen era ted,
bl ocked per investi ga ti onal cen ter, and provi ded to the site
by the con tract research organization (CRO). E ach el i gible stu dy participant was assign ed to one of two tre a tm en t
groups, PRP or control, and received the next available
consecutive ra n dom i z a ti on nu m ber. E ach site had one
designated “unbl i n ded” pers on to treat the patient (also
blinded) and maintain doc u m ents in a sec u re priva te area
to maintain blinding of the investi ga tor, investi ga tive site
staff, patient, s ponsor, and CRO staff and mon i tor. Th i s
pers on did not parti c i p a te in any other aspect of the
patient’s care. The bl i n ded investi ga tors and staff measured the wounds; perform ed all tests, assessments, and
debri dement; and determ i n ed wound closure. A stra tegic a lly placed dra pe prohibited the pati ent from seei n g
wh i ch treatment was app l i ed to the wound. Bl ood was
drawn from both the treatment and con trol pati ents to
maintain blinding.
PRP prepara ti on process. The PRP separa ti on system
utilized in the stu dy is a newer gen era ti on , point-of - c a re
TABLE 1
STUDY EXCLUSION CRITERIA
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Patient currently enrolled in another investigational device or
d rug trial or previously enrolled (within last 30 days) in
investigative research of a device or pharmaceutical agent
Ulcer decreased ≥50% in area during 7-day screening
period
Ulcer is due to non-diabetic etiology
Patient’s blood vessels are non-compressible for ABI testing
Evidence of gangrene in ulcer or on any part of the foot
Patient has radiographic evidence consistent with diagnosis
of acute Charcot foot
Patient is currently receiving or has received radiation or
chemotherapy within 3 months of randomization
Topical, oral, or IV antibiotic/antimicrobial agents or medications have been used within 2 days (48 hours) of
randomization
Patient has received growth factor therapy (eg, autologous
platelet-rich plasma gel, becaplermin, bilayered cell therapy,
d e rmal substitute, extracellular matrix) within 7 days of randomization.
Screening serum albumin level <2.5 g/dL
Screening hemoglobin <10.5 mg/dL
Screening platelet count < 100 x 109/L
Patient is undergoing renal dialysis, has known immune
insufficiency, known abnormal platelet activation disorders —
ie, gray platelet syndrome, liver disease, active cancer
(except remote basal cell of the skin), eating/nutritional,
hematologic, collagen vascular disease, rheumatic disease,
or bleeding disorders
History of peripheral vascular repair within the 30 days of
randomization
Patient has known or suspected osteomyelitis
Surgical correction (other than debridement) required for
ulcer to heal
Index ulcer has exposed tendons, ligaments, muscle, or
bone
Patient is known to have a psychological, developmental,
physical, emotional, or social disorder, or any other situation
that may interfere with compliance with study requirements
and/or healing of the ulcer
History of alcohol or drug abuse within the last year prior to
randomization
Patient has inadequate venous access for blood draw
Patient has a religious or cultural conflict with the use of
platelet gel treatment
system for processing autologous platel ets and plasma to
be used for the tre a tm ent of nonhealing wounds. This system is compri s ed of two components: a small , port a bl e
cen tri f u ge to separate whole bl ood into PRP and a convenience kit that inclu des items for the bl ood draw, processing, and PRP gel application. The platel et - ri chplasma
gel (AutoloGel™, Cytomedix, Inc, Rock vi ll e , Md) was used
to treat pati ents in the treatment gro u p. Wounds in the
con trol group were tre a ted with a saline gel (Normlgel®,
Mölnlycke Health Care, Norcross, Ga). Either PRP gel or
saline gel was app l i ed to the prep a red wound bed .
The first step of the PRP sep a ra ti on process inclu ded
performing a ven i p u n cture to draw <20 mL of bl ood ,
depending on the wound size, from the patient. The bl ood
was spun in a small , portable cen tri f u ge for 1.5 minutes to
s ep a ra te the PRP from the whole bl ood. The PRP was
ex tracted into a syri n ge wh ere re a gents were ad ded to
activa te the platel ets and plasma as well as to achieve
proper gel con s i s tency (gel con s i s tency was usu a lly
a t t a i n ed within 15 to 30 seconds); the gel then was immediately app l i ed to the wound. A contact layer dressing was
applied over the gel . A foam dressing (non - a b s orben t
s i de) was placed over the con t act dressing layer so the PRP
gel was not absorbed. This was covered with the
a b s orbent side of a foam dressing (to absorb any leaking
wound exudates) and sec u red . For protection, b a rrier
cream was placed on intact skin su rrounding the wound.
For pati ents ra n domized to the control group, normal
saline gel was applied to the wound following wound bed
prep a ra tion. Similar to PRP gel application, a contact layer
dressing was applied over the saline gel, followed by the
non-absorbent side of a foam dressing, and covered with
the absorbent side of a foam dressing before being secured.
Clinical evalu a ti ons and procedures. Wounds were
a s s e s s ed and measu red (length, wi d t h , and depth using a
metric tape measu re at each visit. The measurem ents and
o t h er wound va ri a bles including undermining or tu n n eling, ch a racteri s tics of wound exudates (ie, presence, co l or,
amount, and odor), necro tic tissue, and gra nu l a ti on tissue
were documented .27 Re sults rel evant to ad d i ti onal va riables/findings wi ll be published in a sep a rate paper.
Care and management efforts provi ded at each treatment visit included cleansing and assessing the wound and
obtaining vital signs and an interim wound history, including informati on regarding adverse events, concomitant
medications, nutri ti on and wei gh t - be a ring status, and
June 2006 Vol. 52 Issue 6
71
TABLE 2
LABORATORY STUDIES’ SCHEDULE
week later but asked to continue wearing the offloading orthosis walker. At this
Laboratory Parameter
Baseline/ q2 Weeks During 6 Weeks 12 Weeks 24 Weeks
visit, if the wound had
Screening
Treatment
reopen ed, the patient was
re-entered into the study at
Complete blood count
X
X
X
X
the same timeline (coinChemistry 7 panel,
X
As needed
X
X
ciding with the retu rn vi s i t
serum albumin
HgbA1C
X
X
for con ti nu ed care) and
Partial thromboplastin
X
X
X (if during X (if during
continued until the wound
time, prothrombin
treatment) treatment)
ei t h er healed or until wee k
time, thrombin time
13, visit 1 without healing.
If the wound stayed healed
Antibody for Factor V
X
X
X
X
after the 1-week interval,
Fibrinogen
X
Antigenic fibrinogen
PRN positive
the patient en tered the folfibrinogen
low-up phase and return ed
after 3, 7, and 11 weeks.
o t h er aspects of c a re since the last vi s i t . A fac i l i tyde s i gn ee
During this 3-month follow-up phase, the healed wound
perform ed ph l ebotomy; the unbl i n ded person perform ed
was evaluated for breakdown and the patient was qu eri ed
all subsequ ent gel processing. The principal investi ga tor,
regarding adverse events.
who did not ob s erve treatment procedu res, d i rected
At the end of the 12-week treatment peri od , unhealed
wound care provi s i on du ring the care visit.
wounds were treated per physician protocol. The patient
The need for con s i s tency in produ ct app l i c a ti on and
was discharged from the site for follow-up at a fac i l i ty of
maintenance of the blinding process dict a ted that dresshis/her choice. All participants were asked to return for
ings were app l i ed on ly at the Investi ga tor’s site except for
final Factor V testing at week 24 post-ra n domizati on date.
the provi s i on of a on e - time dressing change at home
End-of-study occ u rred at com p l eti on of the week 24 clinishould circumstances prevent clinic atten d a n ce . Pa ti ents
cal lab evaluation, wi t h d rawal of patient consent, or death
retu rn ed twi ce wee k ly at 3- or 4-day interva l s ; procedu re s
of the patient. Patient procedures/instru cti ons were repe a tand processes de s c ri bed were perform ed at each visit for a
ed, including a complete history and physical examination
maximum of 12 weeks. Tre a tm ent con ti nued until the
with testing of pedal pulses, vascular testing (ABI and TBI),
wound healed , the 12-week treatment phase was comSemmes-Weinstein monofilament testi n g, l a bora tory tests,
plete, or patient study participation was term i n a ted by the
dressing change as per the Investigator’s order, and edu c aInvesti ga tor, s ponsor, or because the pati ent withdrew
ti on with a discussion of healthcare options (see Figure 1).
consent or failed to retu rn for visits.
Statistical Analysis
To eva luate safety, clinical labora tory tests were conducted throughout the stu dy to determine the impact of
Healing rate. The power of the stu dy was determ i n ed
tre a tm ent interven ti ons (see Ta ble 2). Ex pected or unexb a s ed on two data sources. The PRP gel healing ra te was
pected adverse events (AE) that occ u rred du ring the
determ i n ed from the healing ra te in an unpubl i s h ed diaco u rse of the stu dy, wh et h er observed by the Inve s ti ga tor
betic foot ulcer retro s pective stu dy28 and feed b ack from
or by the patient, were reported in det a i l . The Inve s ti ga tor
clinicians who had used the PRP gel .29 The control group
mon i tored the patient for AEs or lab abnorm a l i ties until
healing ra te was based on a met a - a n a lysis of healing ra tes
the para m eter retu rn ed to normal or it was determ i n ed
in the con trol groups of 10 pro s pective studies.30
that fo ll ow-up was no lon ger necessary.
Sample size. To determine the sample size, the ex pectEnd-of-trea tm ent visit procedu res. When the
ed proporti on of patients with com p l etely cl o s edwounds
Investi ga tor pron o u n ced the wound cl o s ed — ie, 100%
was determ i n ed to be 0.60 ("t) in the PRP gel arm and
epithelialized — the patient was sch edu l ed for a visit 1
0.20 ("c) in the con trol arm . To calculate the sample size,
72
OstomyWound Management
Assessed for Eligibility
(n = 129)
h e a l ed within each group of
i nvestiga tive sites and all
groups com bi n ed .
Because of varying
Excluded
en ro ll m ent at each site
Patient withdrawal
(bet ween one and 14
Failure to meet inclusion/exclusion criteria (n = 57)
patients), sites were gro u ped
for
analysis
purposes
according
to
provi
der setIntent to Treat
ting and dem ogra ph i c s
Randomization (n = 72)
( Groups I to V). Due to the
va rying natu re of investi gative sites and their ability to
enro ll pati en t s , sites were
grouped into five categories:
Randomized to Control Group
Randomized to Treatment Arm
te aching fac i l i ties, army
(n = 32)
(n = 40)
facility, physicians in priva te
practi ce (two site s ) , and
ambulatory carecl i n i c s . The
goal was to en ro ll ei ght to 21
Independent Site Audit Documentation
patients in each gro u p.
Patients Excluded Due to: Failure to Complete Treatment
Re sults from the indepen d(n = 8); Protocol Violations (n = 24)
ent audit el i m i n a ted all
patients in one group, leaving four groups for per proTotal Per Protocol Treatment Group
Total Per Protocol Control Group
tocol (PP) analysis.
(n = 19)
(n = 21)
Odds rati o / confiden ce
l evel . The odds ra tio and
Healed Wounds Entered 3-Month Follow-up
Healed Wounds Entered 3-Month Follow-up
95% con f i dence interval for
(n = 13)
(n = 9)
e ach group were calculated
for the proporti on of
Figure 1. Trial profile.
patients
with
healed
wo u n d s . The Mantelthe fo ll owing nu ll hypothesis (H0) and the altern a tive
Haen s zel (M-H) combined odds ra tio (com bining over
hypothesis (HA) were formu l a ted :
groups) along with the 95% con f i dence interval was calH0: "t = "c versus HA: "t ≠ "c .
culated . In additi on , the M-H test for hom ogen ei ty of
A Fisher’s exact test with a 0.050, two-sided signifiodds ra tios of groups was perform ed . The M-H method
cance level would have 80% power to detect the differwas used to test the hypothesis wh et h er the M-H comence between PRP gel proportion, "t, of 0.60 and conbined odds ra tio was one.
trol proportion, "c, of 0.20 when the sample size in each
Ad d i ti onal variables. Ot h er efficacy va ri a bles were 1)
arm of treatment is 27 (ie, a total of 54 patients). The
percent change in wound area at en d - of - s tu dy vi s i t
sample size was increased from 54 to 72 patients to
( E O S V) from baseline (BL); 2) percent change in wound
accommodate drop-outs.
vo lume at EOSV from BL; 3) area cl o su re ra te per day at
The primary ef f i c acy va ri a ble was the proporti on of
EOSV; and 4) vo lume cl o su re ra te per day at EOSV. These
patients with a healed wound. Fisher’s ex act test was
ef f i c acy va ri a bles were of con ti nuous type . For each of the
applied to compare the two treatments for proportions
va riables, the two treatm ents were com p a red using
June 2006 Vol. 52 Issue 6
73
Stu den t’s t- test. The stati s tical tests were perform ed using
sof t w a re pack a ge STATA, Release 8.2 (Stata Corpora ti on,
Co ll ege Station, Tex ) .
Kaplan-Mei er. In ad d i ti on, the Kaplan-Mei er31,32 produ ct-limit met h od was used to analy ze time to healing of
the PP wounds and the majority of wound sizes dataset .
Kaplan-Mei er functi ons or curves of the PRP gel and control groups were obtained for each dataset . The log-rank
test was used to test the hypothesis that the Kaplan-Mei er
healing functions are the same ac ross the two treatments.
Labora to ry safety. To evalu a te clinical labora tory safety,
observed va lues at each visit and ch a n ges from baseline at
post-baseline visits and en d point were su m m a ri zed
de s c ri ptively (nu m ber (n), mean, standard deviati on
(SD), minimum, median, maximum) for each treatment
group. Re sults bet ween treatments at the en d point were
com p a red using non - p a ra m etric analysis of va ri a n ce
tech n i ques (Wilcoxon Rank Sum Test, two - s i ded).
Shift analyses. The nu m ber and percent of patients from
each treatment group that shifted in and out of normal
ra n ge from baseline to en d point were calculated for each
laboratory variable. To compare treatments for important
shift changes, the number of patients whose labora tory
re sults shifted from NORMAL or LOW at baseline to
HIGH at endpoint and those shifting from NORMAL or
HIGH at baseline to LOW at en d point were compared in
separate 2 x 2 tables and differen ces were tested for statistical significanceusing two - s i ded Fisher’s ex act tests. Patients
whose labora tory results did not shift from baseline ra n ge
or who were NORMAL at en d point were analyzed toget her with those that had out of ra n ge shifts via ex act tests to
compare treatment groups (Exact test for row [R] x column [C] tables). With these three sep a ra te analyses, va riables can be eva luated either for single directi on shifts of
interest or shifts in either direction, when both HIGH and
LOW devi a ti ons may be of significance.
Results
Initially, 129 pati ents provi ded Inform ed Con s en t
forms and parti c i p a ted in active screening (see Figure 1).
Of these patients, 57 (44%) were dropped from the study
due to redu ction in the wound size of ≥ 50% du ring the
7-day screening period or for failure to meet the inclusion / exclusion criteria. Ul ti m a tely, 72 pati ents were
en ro ll ed , each patient having one wound (index ulcer )
designated for study inclusion.
74
OstomyWound Management
TABLE 3
INTENT TO TREAT GROUP: PATIENT DEMOGRAPHIC AND BASELINE WOUND VARIABLES
Variable
Patient
Age (years)
Hgb A1C
Wound
Area (cm2)
Volume (cm3)
N
PRP Gel (n = 19)
Mean SD
Min
Max
N
Mean
Control (n= 21)
SD
Min
Max
P value
40
37
56.4
8.1
10.2
1.8
31.0
5.5
75.0
13.1
32
30
57.5
8.0
9.1
1.8
45.0
5.0
86.0
11.5
NS*
NS
40
40
4.0
1.7
5.3
4.1
0.4
0.1
24.0
24.8
32
32
3.2
0.9
3.5
1.2
0.5
0.1
15.8
5.4
NS
NS
Treatment
Characteristics
Total
PRP Gel (n = 40)†
No. of patients
Percent
Sex
Male
Female
Race
Caucasian
Hispanic
Black
Other, specify
Foot
Right
Left
Location
Toe
Heel
*
†
Control (n= 32)†
No. of patients
Percent
No. of patients
Percent
32
8
80.00
20.00
27
5
84.38
15.63
59
13
81.94
18.06
26
8
5
1
65.00
20.00
12.50
2.50
18
9
3
2
56.25
28.13
9.38
6.25
44
17
8
3
61.11
23.61
11.11
4.17
23
17
57.50
42.50
18
14
56.25
43.75
41
31
56.94
43.06
13
18
32.50
45.00
14
10
43.75
31.25
27
28
37.50
38.89
NS = not significant (P >0.05)
Demographic / ulcer location variables not statistically significantly different between groups
In the intent-to-treat (ITT) population, the mean and
standard deviations (SD) for age, HgbA1C, wound area,
and volume in the two treatments were not significantly
different, but the wound volume in the PRP gel group
was significantly more variable than in the control group
(SDs 4.1 versus 1.2, P <0.0001) (see Table 3). Ulcer location information was missing for nine patients (three in
PRP gel group and six in the control group). No significant differences in patient demographics, wound distribution, or ulcer location were observed between the two
treatment groups (additional data to be analyzed in a
future publication). For purposes of the ITT analyses,
the ITT population compri s ed all active patients wh o
com p l eted the stu dy as well as those who were lost to fo llow-up, f a i l edto com p l ete the treatment, or had protocol
violations. In the ITT group, 13 out of 40 patients
(32.5%) in the PRP gel and nine out of 32 patients
(28.1%) in the control group had co mpletely healed
wounds after 12 weeks (P = 0.79). Because the results of
the ITT analyses did not seem to reflect previous clinical
outcomes, the study sponsor commissioned an independent audit to ensure study compliance with Good
Clinical Practices (GCP) at the investigative sites.
During the audit, patient source documents, Case
Report Forms (CRF), and other stu dy source documents
were reviewed . Five obj ective cri teria were devel oped
against wh i ch all audited patient records were evalu a ted.
The pro tocol vi o l a ti ons that caused exclusion of patients
inclu ded use of the wrong cen tri f u ge (causing the patient
not to receive the ri ght tre a tm ent — ie, PRP gel ) ; lack of
June 2006 Vol. 52 Issue 6
75
source documentati on to su pport case report form en tri e s ; and inclu s i on of
patients and/or wounds that did not meet the inclu s i on / exclusion cri teri a .
The predeterm i n ed stati s tical plan iden ti f i ed that analysis would be perform ed on patients who com p l eted treatment; thus, patients with early termination due to re a s ons unrel a ted to the index wound and patients lost to
fo ll ow-up also were exclu ded from the PP analysis.
Site audits revealed that 32 out of 72 patients (44%) had protocol vi o l ations or did not meet the cri teria for participation throughout the co u rse of
the study. Of the 32 exclu ded patients, 24 (75%) had protocol vi o l a ti ons and
ei ght (25%) failed to com p l ete treatment. The protocol vi o l a ti ons appe a red
to affect outcomes; thus, the PP dataset , at audit com p l etion, became the primary dataset for analysis — 19 patients were in the PRP gel group and 21
p a ti ents were in the con trol group. These pati ent outcomes ref l ect
patients/wounds treated PP.
In the PP group, on ly the proporti on of Caucasian versus non-Caucasian
p a rticipants was sign i f i c a n t ly different (P = 0.02). The proporti on of
Caucasians was sign i f i c a n t ly high er in the PRP gel group. No stati s ti c a lly significant differen ce bet ween the PRP gel group and the con trol group rel a ted
to age , HgbA1C, wound area, wound volume, s ex , or wound loc a ti on were
observed . This is the same as the ITT group (see Table 4). (Additional differen ces between baseline findings for other wound va ri a bles wi ll be ad d re s s ed
in a later publication.)
Efficacy outcomes. In the PP dataset , 13 of 19 (68.4%) pati ents in PRP gel
and nine out of 21 (42.9%) pati ents in the con trol group healed (P = 0.125,
two-sided Fisher ’s ex act test). The 95% CI for the percent proportion of completely healed wounds was 47.5% to 89.3% and 21.7% to 64.0% for PRP gel
and control groups, re s pectively. The Kaplan-Mei er median time to com p l ete
closure was 45 days for PRP gel compared to 85 days for con trol (log-rank
test, P = 0.126) (see Figure 2).
Although the inclusion/exclusion criteria included wounds with an area
range at least 0.5 cm2 to no larger than 20 cm2, size frequency distributions
showed that the majority (35 out of 40, 88%) of wound sizes were in the
range of both ≤7.0 cm 2 in area and ≤2.0 cm 3 in volume and five (three
patients/wounds in the PRP and two in the control group) were outliers.
The mean area of the outliers was 10.53 cm2 (SD 8.9)and 14.63 cm2 (SD 1.6)
for the PRP and control group, respectively. Because the size range of this
group correlates with the average wound size in multiple published diabetic foot ulcer studies,33 the analyses were repeated using ulcers in this size
range only. This subset of the PP dataset will be referred to as the majority
wounds group.
Wh en standardized for size, (mean area PRP = 2.01 cm2 (SD 1.3) and control 2.43 cm2 (SD 1.6), the proportion of com p l etely healed wounds was 13
out of 16 (81.3%) and ei ght out of 19 (42.1%) in PRP gel and con trol treatment groups, re s pectively (P = 0.036, Fisher’s ex act test). The 95% CI for the
percent proporti ons of com p l etely healed wounds was 62.1% to 100% for the
PRP and 19.9% to 64.3% for the control group.
76
OstomyWound Management
Figure 2. Per protocol patient group (n = 40) Kaplan-Meier time-tohealing curves.
Test = log-rank; Chi square = 2.34;
Degrees of freedom = 1; P value = 0.126
Figure 3. Majority wound patient group (n = 35) Kaplan-Meier timeto-healing curves.
Test = log-rank; Chi square = 5.62
Degrees of freedom = 1; P value = 0.0177
TABLE 4
PER PROTOCOL PATIENT DEMOGRAPHICS AND WOUND CHARACTERISTICS AT BASELINE
Variable
Patient
Age (years)
Hgb A1C
Wound
Area (cm2)
Volume (cm3)
N
PRP Gel (n = 19)
Mean SD
Min
Max
N
Mean
Control (n= 21)
SD
Min
Max
P value
19
18
58.3
7.8
9.7
1.5
43.0
5.5
75.0
11.1
21
20
55.9
8.1
8.1
1.8
45.0
5.0
78.0
11.4
NS*
NS
19
19
3.4
0.9
4.5
1.3
0.8
0.1
20.0
4.7
21
21
3.6
1.0
4.0
1.4
0.5
0.1
15.8
5.4
NS
NS
Treatment
Characteristics
Total
PRP Gel (n = 19)
No. of patients
Percent
Sex
Male
Female
Race
Caucasian
Hispanic
Black
Other, specify
Foot
Right
Left
Location
Toe
Heel
*
†
Control (n= 21)
No. of patients
Percent
No. of patients
Percent
16
3
84.21
15.79
16
5
76.19
23.81
32
8
80.00
20.00
15
4
78.95†
21.05
9
8
3
1
42.86†
38.10
14.29
4.76
24
12
3
1
60.00
30.00
7.50
2.50
14
5
73.68
26.32
10
11
47.62
52.38
24
16
60.00
40.00
6
10
31.58
52.63
9
6
42.86
28.57
15
16
37.50
40.00
NS = not significant (P >0.05)
P = 0.02, two-sided Fisher’s Exact Test
June 2006 Vol. 52 Issue 6
77
The Kaplan-Mei er curves of proporti on for com p l etely healed wounds
over time for the majori ty wound dataset showed that the two curves started
s ep a ra ting from each other on abo ut day 28 (see Figure 3), l og - rank test, P
value = 0.018.
When evaluating the range of healing outcomes in the four investigative
site groups groupings, wide variations in healing outcomes between the site
groups were observed. For the PP dataset, the site group percent of complete healing proportion varied from 50% to 100% for PRP gel-treated
wounds and from 25% to 67% for control-treated wounds. For the majority wound dataset, the site group percent of complete healing proportion
varied from 60% to 100% and from 25% to 60% in the PRP gel- and control-treated wounds, respectively. A trend for increased wound healing in
the PRP group compared to the control group also was observed. Despite
the between-site group variations, healing outcomes in each treatment
group were consistent and similar in both the total dataset of the per protocol and majority wound groups.
Rate of healing. In the PP dataset , the avera ge wound area cl o su re ra te per
day was 0.051 cm2 for the PRP gel group versus 0.054 cm2 for the con trol
group. In the majori ty wound dataset, the wound area closure per day was
0.042 cm2 for the PRP gel group and 0.043 cm2 for the con trol group; these
differen ces were not stati s ti c a lly significant.
In the PP dataset , wounds in the PRP gel group healed after a mean of 42.9
days (SD 18.3) compared to 47.4 days (SD 22.0) for wounds in the con trol
group. While the nu m ber of days to healing was the same in the majori ty
wound group (mean 42.9 and 42.8 days), 81.3% of PRP gel - treated wounds
and 42.1% of con trol gel - tre a ted wounds healed du ring that time.
Follow-up. Of the 40 patients in the PP dataset , 22 with healed wounds
participated in the 12-week fo ll ow-up phase; of those, one in the PRP gel
group had a wound that reopen ed. None of the control-tre a ted patients’
wounds re-open ed ; this differen ce was not stati s ti c a lly significant.
Safety outcomes: adverse events (AE). An AE in a clinical stu dy patient
who has been administered an investi ga ti onal agent is any unu sual med i c a l
occurrence that has appe a red or wors en ed after the start of stu dy wh et h er or
not the occurren ce was rel a ted to the use of the investi ga ti onal produ ct .
Adverse events were captured for any clinical abnorm a l i ties that appe a red or
wors en ed bet ween the patient’s start of the 7-day screening peri od and 30
days after receiving the last dose of stu dy treatment.
Of the 127 adverse events, five occ u rred in two pati ents before randomizati on du ring the 7-day screening peri od. Of the remaining 122 adverse
events occ u rring after random i z a ti on, 60 (49%) were in the PRP gel group
and 62 (51%) in the con trol group. Of these, two were iden ti f i ed as definitely rel a ted to the treatment: one case of con t act derm a ti tis occ u rred in a PRP
gel tre a ted wound and one instance of macera ti on occ u rred in a con trol
tre a ted wound.
Safety outcomes: serious adverse events (SAE). An SAE is an adverse
event that meets any of the fo ll owing outcome criteri a :
78
OstomyWound Management
TABLE 5
REPORTED SERIOUS ADVERSE EVENTS (N = 23 EVENTS)
PRP Gel Group (6 events; 5 patients)
Medra Term
Myocardial infarction
Congestive heart failure
Pneumonia
Pneumonia
Pneumonia; osteomyelitis
During Treatment
Post Treatment
Event
X
Death
Hospitalization
Hospitalization
Hospitalization
Hospitalization
X
X
X
X
Severity as deter- Relationship
mined by
to device
Investigator
Severe
Unrelated
Mild
Unlikely
Moderate
Unlikely
Mild
Unrelated
Mild
Unrelated
Control Group (17 events; 7 patients)
Localized infection
Chest pain (gallstones)
Infected left foot; left foot ulcer
Bacterial arthritis/
encephalopathy
Gangrene, anemia, renal failure,
cardio-respiratory arrest
Diabetic foot, cellulitis,
osteomyelitis
Cellulitis, arthritis bacterial,
atrioventricular block,
elevated blood glucose
X
X
X
X
Hospitalization
Hospitalization
Hospitalization
Hospitalization
Severe
Mild
Mild
Moderate/severe
Death
Severe
Unlikely
Unlikely
Unlikely
Unlikely
Unlikely
Unlikely
X
Hospitalization
Severe
Unlikely
X
Hospitalization
Mild/severe
Unrelated
X
• is fatal
• is life-thre a tening — ie, the patient was, in the view
of the Inve s ti ga tor, at immediate risk of death from
the re action as it occ u rred ; however, it does not
inclu de a re action that, had it occ u rred in a more
serious form , m i ght have caused de a t h
• requires or prolon gs inpatient hospitalization
• re sults
in
significant
or
pers i s ten t
disability/incapacity
• is a con genital anomaly or bi rth defect
• is an important medical event, based on appropriate medical judgment, that may jeopardize the
patient or require the patient to seek medical or
surgical intervention to prevent one of the other
outcomes above.
Of the 122 adverse events after ra n domization, 23 were
cl a s s i f i ed as serious adverse events; six occurred in the
PRP gel group and 17 in the control group. All seri o u s
adverse events were unlikely or unrel a ted to devi ce usage
as def i n ed by the investi ga tors (see Table 5).
Clinical labora tory re sults. To analy ze the safety of
tre a ting patients with the PRP gel , labora tory tests were
condu cted according to the study’s predeterm i n ed ti m eframe. Because a pro s pective trial on the use of PRP gel in
patients with diabetes had not been condu cted before,
questions were ra i s ed wh et h er the bl ood draws, use of
bovine thrombin, or the tre a tm ent itsel f would have a systemic ef fect on persons with diabetes. These concern s
determ i n ed the clinical labora tory tests that were condu cted du ring the trial (see Ta ble 6).
Of the 72 parti c i p a ting pati ents, 56 (78%) retu rn ed for
the day 168 labora tory tests. No statisti c a lly or clinically
s i gnificant differen ces were noted bet ween the PRP gel
and con trol from baseline to en d point labora tory shifts in
hemato l ogy, cl o t ting factors , and Factor V tests. Al t h o u gh
no stati s ti c a lly significant differen ce was noted bet ween
the PRP gel and control in rel a tion to shifts of cl o t ting factors , a shift (increase) was noted in PT and PTT re sults in
both tre a tm ent groups. No clinically important changes
in cl o t ting factors that would cause concern abo ut the
June 2006 Vol. 52 Issue 6
79
ef fect of the PRP gel or control on Factor V
activi ty were found du ring an indepen dent
monitor review of the medical records,
including concomitant med i c a tions.
No clinical or statisti c a lly significant differen ces were noted in ch emistry test results
bet ween the PRP gel and con trol from baseline to en d point for sod ium, potassium,
ch l ori de , bi c a rbonate, creatinine, or albumin. A statistically significant differen ce
was observed bet ween treatments in the
ch a n ge from baseline for BUN; the BUN of
the PRP gel - tre a ted pati ents decre a s ed
while the BUN of con trol pati ents increased
(no explanation was determ i n ed ).
Serum glucose or HbA1C re sults showed
that more patients shifted to high at en dpoint in the PRP gel compared to the control group. These differen ces were not statistically significant or cl i n i c a lly meaningful; t h ey su ggested that more patients in the
PRP gel group had uncon tro ll ed diabetes.
These safety re sults doc u m ent the minimal occ u rren ce of adverse events. No serious adverse events were attri but a ble to PRP
gel and minimal were attri but a ble to labora tory shifts. These ef fects were comparable
with the con trol group.
Discussion
This is the first reported pro s pective,
randomized , bl i n ded, con tro ll ed trial in the
US on the use of PRP for the tre a tm ent of
d i a betic foot ulcers. In this FDA- a pproved
study, s ome of their requ i rem ents (certain
inclu s i on / exclusion cri teri a , blinding system , ch oi ce of con trol treatment, exten s ive
labora tory tests, and doc u m entati on )
ad ded to the ri gor and complex i ty of the
study design, wh i ch in tu rn caused some
difficulty en ro lling patients.
This stu dy com prised two levels of
screening: pre- and active screen i n g.
Initi a lly, an investi ga tor eva luated wh et h er a
patient was a po ten tial candidate for the
study. Approximately 650 patients were
80
OstomyWound Management
TABLE 6
CHANGES IN CLINICAL LABORATORY RESULTS
Laboratory Test
Hematology
Hemoglobin (G/DL)
Hematocrit (%)
Platelet count (10^3/UL)
White blood cells (10^3/UL)
Chemistry
Albumin (G/DL)
Bicarbonate (MEQ/L)
Blood urea nitrogen (MG/DL)
Chloride (MEQ/L)
Creatinine (MG/DL)
Glucose, serum (MG/DL)
Potassium (MEQ/L)
Sodium (MEQ/L)
PRP Gel (n = 40)
Baseline (mean) Endpoint (mean)
13.7
13.4
40.6
40.1
264
280
8.0
7.9
Control (n = 32)
Baseline (mean) Endpoint (mean)
13.1
12.8
39.2
38.5
263
262
7.8
8.0
P value*
0.740
0.644
0.076
0.877
3.8
23.9
21.1
101.5
1.1
187.2
4.6
138.0
3.7
23.7
19.6
102.2
1.1
202.7
4.5
137.9
3.7
23.9
20.7
102.1
1.1
175.6
4.4
137.8
3.6
23.4
23.1
102.4
1.1
211.5
4.5
137.3
0.4693
0.6674
0.0405
0.5721
0.4143
0.4045
0.2343
0.6172
8.0
8.5
8.0
8.0
0.1232
Factor V activity (%)
105.2
104.8
101.0
103.1
0.6113
Clotting Factors
PT (seconds)
PTT (seconds)
TT (seconds)
TT Human (seconds)
12.8
29.7
12.4
11.3
13.4
31.9
12.9
11.7
13.0
30.2
12.8
11.3
13.2
30.3
12.4
11.1
0.8545
0.3738
0.4315
0.2196
Hemoglobin A1C (%HB)
*
Wilcoxon rank sum test, two-sided
pre - s c reened — ie, revi ewed by the investi ga tor for po s s ible inclu s i on — to sec u re the 129 active screen i n g
patients. Active screening com pri s ed baseline wound
assessment, physical exam, labora tory tests, wound culture, vascular tests, wound excision/debri dement, and the
7-day screening period for patients meeting the requ i remen t s. From the 129 actively screen ed patients, 72 were
randomized and 40 met stu dy protocol requ i rements (the
clinical data audit prompted exclusion of 32 patients,
dropping the PP nu m ber to 40 patients). Thus, ultimately, on ly 6% of patients who participated in the pre - s c reening process were en ro ll ed .
During the analysis of the PP gro u p, f requency distri buti on dem on s tra ted that the majori ty of t h e
wounds (35 out of 40) random i zed into the stu dy met
the cri teria of wound area ≤7.0 cm2 and vo lume ≤2.0
cm3. The remaining five larger wounds had areas of 9
cm2 to 20 cm2 and re sults of various studies su ggest
that a wound size of <7.0 cm2 is most com m on.30-32
Efficacy va ri a bles also were analy zed for the su b s et of
“majority wo u n d .”
Average baseline area in the majority wound group
was similar to that reported in a tissue-engineered product study (n = 208: mean wound area 2.97 – 3.1 cm2),34
another tissue-engineered product (n = 15 wounds; efficacy noted in wounds < 6 cm2),35 two recombinant
growth factor studies (n = 118 patients and 132 wounds
respectively; mean area 5.5 cm2 and 2.6 cm2),36,37 and one
retrospective study of diabetic foot ulcers (n = 26,599
patients of which 5,320 wounds averaged 1.53 cm2, 5,320
wounds averaged 1.84 cm2, and 5,319 wounds averaged
4.41 cm2 in area24). In the largest study published to
date,24 60% of patients had wounds that matched the
majority group in this PRP ge l study, increasing the
potential external validity of the current study results.
In the majority wound group, PRP gel - tre a ted wounds
were sign i f i c a n t ly more likely to heal than con trol-tre a ted
wounds even though healing ra tes in the control group
June 2006 Vol. 52 Issue 6
81
were high er (42% healed
after 12 weeks) than most
con trol group healing ra te s
reported in other studies30
( s ee Figure 4). Spec i f i c a lly,
re sults of a met a - a n a lysis of
healing outcomes in the
con trol arm of o t h er diabetic foot pro s pective studies
su ggest that 24% can be
ex pected to heal after 12
weeks of providing good
care.30 Most studies use wet to - m oist ga u ze saline ga u ze
dressings, the recogn i zed
standard treatment by the
m edical com mu n i ty5 as a
con trol dressing. However,
wet - to - m oist
dressings
requ i re dressing ch a n ges Figure 4. Comparison of diabetic foot ulcer prospective trial control outcomes.30
three to four times daily by
the pati ent or caregiver and may not provi de an optimal
s m a ll frequ ent ex po sure to it. To evalu a te this po tential
moist environ m ent for healing. Some studies have shown
i m p act , Factor V tests occ u rred at 6-week intervals for all
that hydroco ll oid dre s s i n gs may be more ef fective than
patients regardless of random i z a ti on arm and at 24 wee k s
wet - to - m oist for the tre a tm ent of certain types of ulcers.38
post ra n domization. None of the pati ents dem on s tra ted
The power calculati ons were based on reported
any Factor V inhibiti on thro u gh o ut the stu dy and into the
results of control treatments in other studies and a samfo ll ow-up. This is the first time that this has been doc uple size of 27 in each treatment arm. The better-thanmented in a pro s pective study.
expected control healing rates and large number of proLabora tory tests indicated that the majori ty of patients
tocol violations caused underpowering of the PP and the
had el eva ted bl ood glu cose throughout the study peri od .
majority wound groups yet some statistically significant
Eleva ted glu cose levels have been documented to redu ce
differen ces were ob s erved, su ggesting that larger
healing40,41; h owever, in this stu dy, the majori ty of the
between-group differences could be expected in studies
wounds healed .
using a larger sample size.
Numerous articles have been published regarding the
The FDA’s con cern abo ut the effects of frequ en t
use of PRP in the surgical setting; notably, the orthopet h o u gh small amount of bl ood co ll ecti on (30 mL or less
dic, plastic surgery, and dental field.42-44 Some of the PRP
each visit) on health and safety as patients underwen t
separation systems used require specialized technicians
periodic hemato l ogy and other labora tory tests thro u ghto perform the necessar y procedures. This is the first
out the trial was ad d re s s ed . Test re sults documented that
published study of an autologous PRP separation system
these frequ ent bl ood draws did not redu ce the hem ogl oto heal wounds that can be used by health professionals
bin, hematocrit, or platel et co u n t . Because bovine thromwithin a traditional healthcare setting. A small, compact,
bin was used in the processing to activa te the PRP and evipoint-of-care system such as the one described in the
den ce exists of Factor V leading to bl eeding disorders in
study makes this technology available to multiple care
33
patients ex po s ed to large doses of bovine thrombin, conproviders, including physician office, hospital unit, outcerns were ra i s ed as to wh et h er patients would similarly
patient clinic, long-term care facility, and home health
devel op Factor V antibodies to the bovine thrombin wi t h
care staff.
82
OstomyWound Management
Conclusion
The results of this study show that PRP gel is safe for
use in the treatment of nonhealing diabetic foot ulcers.
In the most common size of diabetic foot ulcers (≤7.0
cm2 in area and ≤2.0 cm 3 in volume), PRP gel-treated
wounds are also significantly more likely to heal than
control gel treated wounds. Treating wounds with PRP
or saline gel resulted in healing in appr oximately 6
weeks, but in the most common wound sizes, almost
twice as many PRP treated wounds healed in that timeframe. The number of adverse events was minimal; no
adverse events were serious. Further, the study demonstrated that bovine thrombin used in the preparation of
PRP does not cause Factor V inhibition; thus, it does not
cause coagulopathy. In addition, withdrawal of a small
amount of blood twice weekly did not affec t patient
hemoglobin, hematocrit, or platelet count. Clinically
meaningful shifts in laboratory values studied from
baseline to endpoint were not obser ved. This type of
PRP system could be utilized by healthcare providers to
treat diabetic foot ulcers in multiple settings. Using PRP
gel to treat diabetic foot ulc ers may not only enhance
healing, but it also may prevent lower extremity amputations caused by nonhealing wounds.
Implicati ons for future re s e a rch include implem en ting
a trial de s i gn that would permit greater subj ect en ro llment on a larger sample size to va l i d a te these results. Th i s
trial would not need to re - evaluate some of the major
qu e s tions answered in this trial, su ch as Factor V inhibition, i m p act on the pati ent’s hem a to l ogy and other cl i n ical labora tory outcomes, and impact of a con trol that had
not perform ed in a prospective trial previ o u s ly. Di a beti c
foot ulcers with ch a ll en ging presentati ons (ie, mild to
m odera te vascular disease, ex po s ed ten don or bone,
patient hyper glycemia, and/or inadequ a te nutri tional status) could be studied to determine wh et h er PRP gel could
assist in healing in these comprom i s edscen a ri o s . In ad d ition, studies to determine wh et h er this novel thera py is
synergistic with other adva n ced wound care modalities
could be conducted . - OWM
Neal Mozen, DPM; Jeffrey Pa ge , DPM; Al ex a n der
Rey zelman, DPM; Thomas Ro u k i s , DPM; Roger
Schech ter, MD; Th omas Seren a , MD; Ji n sup Son g, DPM;
and Robert Snyder, DPM.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Acknowledgments
The authors acknowl edge the AutoloGel™ Di a beti c
Foot Ulcer Group: Margaret Do u cet te, DO ; Mich ael
Dellacorte, DPM; Robert Fryk berg, DPM; G a briel
Ha l peri n , DPM; Robert Kirsner, MD; Mi ch ael Mi ll er, DO ;
13.
14.
Centers for Disease Control and Prevention. Na ti onal diabetes fact sheet: gen eral information and nati onal esti m a tes
on diabetes in the United States. Atlanta, Ga: United States
Dep a rtm ent of Health and Human Services. Center for
Disease Control and Prevention;2005.
Am erican Di a betes Association. Consensus developm ent
conference on diabetic foot wound care. Diabetes Care.
1999;22(8):1354–1360.
Mani R, Fa l a n ga V, Shearman CP, Sa n deman D. Chronic
Wound Healing: Clinical Measurem ent and Basic Science.
New York, NY: WB Sa u n ders;1999:17–23.
Pecora ro RE, Reiber GE, Burgess EM. Pathways to diabeti c
limb amputation. Basis for preven ti on. Di a betes Care.
1990;13(5):513–521.
American Diabetes Association. Available at: www.diabetes.or g.Accessed Janu a ry 25, 2006.
Fryk berg RG, Armstrong DG , Giu rini J, et al. Diabetic Foot
Di sorders: A Clinical Practi ce Guideline. Park Ridge, Ill :
Am erican Coll ege of Foot and Ankle Surgeons;2000.
Ho ll oway GA, Steed DL, DeMarco JM, et al. A ra n domized
controll ed dose response trial of activa ted platel et su pernatant, topical CT-102 in chronic, nonhealing, d i a beti c
wounds. WOUNDS. 1993;5:160–168.
Steed DL, Gloslen JB, Ho ll oway GA, et al. CT-102 activa ted
p l a tel et su pernatant topical: a ra n domized, prospective,
do u ble blind trial in healing of chronic diabetic foot ulcers.
Di a betes Care. 1992;15:1598–1604.
Atri SC, Mi s tr J, Bisht D, et al. Use of homologous platel et
factors in achieving total healing of recalcitrant skin ulcers.
Su rgery. 1990;108:508–512.
Knighton DR, Ciresi K, Fiegel VD, et al. Stimulation of
repair in chronic nonhealing cut a n eous ulcers: a pro s pectively ra n dom i zed blinded trial using platel et - derived
wound healing formula.
Surg Gy necol Obstet.
1990;170:56–60.
Fylling CP, Baxter CR, Bennett KG, et al. Multi-center cl i nical revi ew: using auto l ogous platel et gel for the treatment
of d i a betic plantar wounds. Di a bete s. 2001;50 (su pp l
2):229A.
Glover JL, Weinga rten MS, Buchbinder DS, et al. A fouryear multicenter retrospective stu dy of topical mu l tiple
growth factor thera py for chronic wounds. Adv Wound
Care. 1997;10(1):27–32.
Keyser JE. Di a betic wound healing and limb salva ge in an
outpatient wound care program. South Med J.
1993;86:311–317.
Fylling CP. A five center study of diabetic wound healing in
June 2006 Vol. 52 Issue 6
83
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
comprehensive wound managem ent programs. Diabetes.
1992; 41(suppl A):82A.
Knighton DR, Fylling CP, Fiegel VD, et al. Amputation prevention in an independently revi ewed at-risk diabetic population using a compreh ensive wound care protocol. Am er
J Su rg. 1990;160:466–472.
Fylling CP, Knighton DR, G ordinier RH. The use of a comprehensive wound care protocol including topical growth
factor therapy in treatment of d i a betic neuropathic ulcers.
In: Ward J, Goto Y. Diabetic Neuropathy. New York, NY:
John Wi l ey & Sons;1990:567–578.
Doucet te MM, Fylling CP, Knighton DR. Amputation prevention in a high - risk population through a comprehensive
wound healing protocol. Arch Phys Med R ehabil.
1989;70:780–785.
Knighton DR, Fylling CP, Do u cette MM. Wound healing
and amputation in a high-risk diabetic population.
WOUNDS. 1989;2:107–114.
Knighton DR, Ciresi KF, Fiegel VD, et al. Classification and
tre a tm ent of chronic nonhealing wounds: su ccessful treatment with auto l ogous platel et-derived wound healing factors (PDWHF). Ann Surg. 1986;3:322–330.
Bentkover JF, Champion AH. Econ omic eva luation of alternative methods of tre a tm ent for diabetic foot ulcer pati ents:
cost effectiveness of platel et releasate and wound care clinics. WOUNDS. 1993;5:207–215.
Fylling CP, McKeown PC. Cost and healing ef f i c acy of treating diabetic foot ulcers in a comprehensive wound managem ent program with growth factor thera py. Diabetes.
1990;39(suppl A):18A.
Bennett NT, Schultz GS. Growth factors and wound healing. Pa rt II. Role in chronic wound healing. Am J Surg.
1993;166(1):74–81.
Mosesson MW. Fibri n ogen and fibrin structu re and functions. J T h romb Haemost. 2005;3(8):1894–1904.
Ma r golis DJ, Kantor J, Santana J, et al. Effectiveness of
platel et releasate for the tre a tm ent of diabetic neuropathic
foot ulcers. Diabetes Care. 2001;24:483–488.
Lipsky BA, Berendt AR, Desery H. Di a gnosis and treatment
of d i a betic foot infecti ons. Guidelines for Diabetic Foot
Infecti ons. Clin Infe ct Dis. 2004:39:895–910.
Armstrong DG , Lavery LA, Harkless LB.Validation of a diabetic classification system; the contri bution of dept h , infection, and ischemia to risk of amputation. Diabetes Care.
1998;21(5):855–859.
Cooper DM. Assessment, measurement, and evaluation:
their pivotal roles in wound healing. In: Bryant RA. Acute
and Chronic Wou n d s. St. Louis, Mo: Mosby;2000:61.
Fylling CP, Ba x ter CR, Bennett KG, et al. Multi-center clinical revi ew: using auto l ogous platel et gel for the treatment
of diabetic plantar wounds. Poster and oral presentation at
the American Di a betes Association Annual Meeting and
Scientific Sessions. P h i l adelphia, Pa. June 22–26, 2001.
Beriou J, Fylling CP, Justice P. A retrospective revi ew of healing progress of 102 chronic wounds du ring the first 28 days
84
OstomyWound Management
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
of treatment with auto l ogous platel et rich plasma gel.
Po s ter presen t a ti on at: the Sym po s ium on Advanced
Wound Care and Medical Research Forum. Orlando, Fla.
May 5– 7, 2004.
Margolis DJ, Kantor J, Berlin J. Healing of diabetic neuropathic foot ulcers receiving standard treatment. Diabetes
Care. 1999;22:692–695.
Lee ET. St a ti s tical Met h ods for Su rvival Data Analys i s.
Belmont, Calif: Lifetime Learning Publications,
Wadsworth, Inc;1980.
Kalbflei s chJD, Pren ti ce RL.The Stati s tical Analysis of Failure
Time Data. New York, NY: John Wi l ey and Sons;1980.
Margolis DJ, Gelfand JM, Hoffs t ad O, Berlin J. Surroga te
endpoints for the tre a tm ent of d i a betic neuropathic foot
u l cers. Di a betes Care. 2003;26:1696–1700.
Veves A, Falanga V, Armstrong DG , et al. Graftskin, a
human skin equivalent, is effective in the management of
noninfected neuropathic diabetic foot ulcers. Diabetes Care.
2001;24(2):290–295.
Lipkin S, Chaikof E, Isseroff Z, et al. Effectiveness of bilayered cellular matrix in healing of neuropathic diabetic foot
u l cers: Results of a multicenter pilot trial. WOUNDS.
2003;15(7):230–236.
S teed DL. Clinical eva lu a ti on of recombinant human
p l a telet-derived growth factor for the tre a tm ent of lower
extrem i ty diabetic ulcers. J Va sc Su rg. 1995;21:71–81.
Wieman TJ, Smiell JM, Su Y. Efficacy and safety of a topical gel formulation of recombinant human plateletderived growth factor – BB (becaplermin) in patients
with ch ronic neu ropathic ulcers. Di a betes Care.
1998;21(5):822–827.
Bradl ey M, Cu llum N, Nelson EA, et al. System a tic revi ews
of wound care management: dressings and topical agen t s
used in the healing of chronic wounds. Health Technol
Assess. 1999;3(17 Part 2).
Winterbottom N, Kuo JM, Nguyen K, et al. Anti genic
responses to bovine throm bin ex po sure du ring surgery: a
pro s pective stu dy of 309 pati ents. J Applied Res Clin Exper
T h erapeu ti cs. 2002;2(1):20–29.
Jeffcoate WJ, Pri ce P, Harding KG, et al. Wound healing and
tre a tm ents for people with diabetic foot ulcers. Diabetes
Metab Res Rev. 2004;20(su ppl 1):S78–S89.
Ma rs ton WA, Derm a graft Di a betic Foot Ul cer Stu dy
Group. Risk factors associated with healing chronic diabetic foot ulcers: the importance of hyper glycemia. Ostomy
Wound Manage. 2006; 52(3):26–28,30,32.
Kassolis, JD, Reynolds, Ma. Eva luation of the adjunctive
benefits of platel et-rich plasma in subantral sinus augm entation. J Craniof a cial Surg. 2005;16(2):280–287.
Grageda E. Platel et - ri ch plasma and bone graft materials: a
review and a standardized research protocol. Im plant Den t.
2004;13(4):301–309.
Weibrich G, Hansen T, Kl eis W, et al. Effect of platel et concentra tion in platel et - ri ch plasma on peri-implant bone
regenera tion. Bo n e. 2004;34(4):665–671.