sleep and inhibiting the respiratory drive (2). In lower doses,... euphoriant effect predominates (5). The toxicity of GHB is

Transcription

sleep and inhibiting the respiratory drive (2). In lower doses,... euphoriant effect predominates (5). The toxicity of GHB is
Aronsen_jatsept10.qxd:JATLynneTemplate
8/9/10
4:49 PM
Page 1
Journal of Analytical Toxicology, Vol. 34, September 2010
Urinary Concentrations of Gamma-Hydroxybutyric Acid
and Related Compounds in Pregnancy
Guttorm Raknes, Lena Aronsen*, and Ole Martin Fuskevåg
Department of Clinical Pharmacology, University Hospital of North Norway Trust, NO-9038 Tromsø, Norway
Abstract
Endogenous production complicates the interpretation when
gamma-hydroxybutyric acid (GHB) is measured in urine for
forensic purposes. We performed a cross-sectional study to test the
hypothesis that pregnant women have higher levels of urinary GHB
than non-pregnant controls, and thus increased risk of falsepositive GHB tests. GHB, gamma-butyrolactone (GBL) and betahydroxybutyric acid (BHB) concentrations in urine from 66
pregnant women and 69 non-pregnant controls were analyzed by
liquid chromatography–tandem mass spectrometry (LC–MS–MS).
The mean GHB, GBL, and BHB concentrations were 0.36, 0.34 and
1.92 mg/L in the pregnant women, and 0.24, 0.08 and 0.40 mg/L
in the control group. The pregnant women had significantly higher
levels of GHB (1.5-fold), GBL (4.3-fold), and BHB (4.8-fold).
Creatinine-adjusted GHB concentrations were similar in both
groups. Pregnant women have higher urinary levels of GHB, GBL,
and BHB. In LC–MS–MS assays not distinguishing between GHB
and BHB, there is a significantly increased risk of false-positive
GHB tests in pregnant women. This false-positive rate can be
reduced by correcting for creatinine concentration, by using GHBspecific assays or by introducing higher interpretative cut-off levels
for pregnant women in assays that do not discriminate between
GHB and GBL or BHB.
Introduction
Gamma-hydroxybutyric acid (GHB) occurs naturally in all
mammals, but its function remains unknown. Being both a
precursor and a metabolite of gamma-aminobutyric acid
(GABA) (1), it is anticipated that GHB acts as a central nervous
system (CNS) neuromodulator, mediating its effects through
GABAB and GHB-specific receptors, or by affecting dopamine
transmission (2).
GHB has been used recreationally since it was first marketed in 1960. It is labeled as an illegal drug in most countries,
but also used as a legal drug (Xyrem) in patients with narcolepsy (1–4). In high doses, GHB inhibits the CNS, inducing
* Author to whom correspondence should be addressed. Email: [email protected].
394
sleep and inhibiting the respiratory drive (2). In lower doses, its
euphoriant effect predominates (5). The toxicity of GHB is
generally considered to be low, but several GHB-related fatalities have been reported (6–9). Pregnant women with a history
of drug abuse are often subjected to urine analysis to reveal intake of substances that could damage the fetus. The presence
of endogenous GHB in urine makes an interpretative cut-off
necessary to differentiate physiological GHB concentrations
from concentrations seen with exogenous ingestion. Although
10 mg/L is the most commonly proposed cut-off (2,10–14),
levels ranging from 2 to 20 mg/L have been suggested (3).
Few studies have taken urine dilution into consideration, but
in one study on 50 females, 323 µg GHB/mmol creatinine was
indicated as an upper limit for normal GHB concentration, and
1000 µg GHB/mmol creatinine was proposed as an interpretative cut-off (3). In another study, no statistical difference was
found between non-normalized and creatinine-normalized
urine GHB results in samples collected during one week from
eight persons (10).
When measuring GHB in urine, several complicating factors
should be addressed. Firstly, endogenous gamma-butyrolactone (GBL) is converted into GHB, and vice versa, in a pH-dependent process, and several GHB assays involve a step converting all GHB into GBL before analysis (2,10,11,15).
Secondly, there is a naturally occurring ketone body, beta-hydroxybutyric acid (BHB), which is an isomer of GHB (14).
Liquid chromatography–tandem mass spectrometry (LC–MS–
MS) is increasingly being applied for GHB analysis. BHB and
GHB have identical molecular weights (MW) and fragment
into identical daughter ions under LC–MS–MS conditions,
making them indistinguishable in an MS. If this is not taken
into account, by separating the isomers chromatographically
using a column that can retain small polar organic molecules,
for example, the measured GHB concentrations are, in reality,
BHB and GHB combined. We have reason to believe that such
methods have been applied even for forensic purposes. There
also exists an alpha-isomer, alpha-hydroxybutyric acid (AHB) of
GHB, which is elevated in diabetics (14). Thirdly, there is firm
evidence that GHB is produced in vitro, and unless the urine
samples are stored at low temperatures shortly after voiding,
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Aronsen_jatsept10.qxd:JATLynneTemplate
8/9/10
4:49 PM
Page 2
Journal of Analytical Toxicology, Vol. 34, September 2010
the GHB concentration might show a large increase (15–17).
1,4-Butanediol (1,4-BD) is a chemical precursor of GHB that
sometimes is detected in GHB abusers (2,5). One study detected 1,4-BD in postmortem human brain tissue, indicating
that 1,4-BD might be endogenously produced (18).
We have experienced that urine samples from pregnant
women tend to have higher GHB concentrations than samples
from other persons with no assumed GHB intake. If this is true,
pregnant women will have a higher risk of false-positive tests.
It is important to identify this, as it may affect treatment and
could lead to serious legal consequences. To our knowledge,
urine GHB concentrations in pregnant and non-pregnant have
not previously been systematically compared.
Our main objective was to test the hypothesis that pregnant women have higher urinary concentrations of GHB, GBL,
and/or BHB than non-pregnant controls (male and female).
Furthermore, we wanted to examine the effect of duration of
pregnancy and use of hormonal contraceptives on the endogenous levels of these compounds.
Materials and Methods
We performed a cross-sectional study comparing the GHB,
GBL, and BHB concentrations in pregnant women and nonpregnant controls. The pregnant women were recruited from
the midwife center of Tromsø municipality in northern
Norway, from October 2007 to March 2008. The non-pregnant
controls were nursing, dentistry, and medical students at
Tromsø University College and University of Tromsø, recruited
from August through November 2007. The study was approved
by the local ethics committee.
Midwives were instructed to ask pregnant women in their
second or third trimester attending routine appointments
whether they wanted to participate in the study. The women
were given questionnaires, consent forms, and urine sampling
tubes that they were to bring along at the next appointment if
they agreed to participate. The non-pregnant controls were
recruited by informing students about the study in connection
with a lecture. The urine samples, questionnaires, and consent
forms from the volunteers were collected shortly after the lecture. Both males and non-pregnant females were included because the cut-off values for analysis of drugs of abuse generally
are not gender-specific. As females were in the majority among
all included students, substantial power for subgroup analysis
of female controls was expected. Urine samples were registered and frozen at –80°C as soon as possible after voiding. All
samples were screened for amphetamines, benzodiazepines,
cannabinoids, cocaine, and opiates by immunoassay (CEDIA
techniques, Microgenics, Fremont, CA). Participants with samples containing any of these drugs were excluded. The urine
samples from female controls were tested for human chorionic
gonadotropin (hCG) (ClearView hCG, Inverness Medical, Bedford, U.K.) to exclude any unknown pregnancies. Written consent was required from all participants.
The main outcomes were GHB, GBL, and BHB concentrations in urine, both unadjusted and creatinine adjusted. There
is considerable intraindividual variation in diuresis and thus in
the concentration of renally excreted substances. This can be
accounted for by normalizing the measured concentrations
to creatinine concentrations, assuming a constant rate of creatinine excretion (19). The combined concentrations of
GHB+GBL and GHB+BHB were included as outcome variables to make the results comparable to studies not differentiating GHB from GBL or BHB. The concentration of GHB+GBL
(in mg/L) was calculated from the sum of molar concentrations
of GHB and GBL using the MW of GHB. The MWs of the isomers GHB and BHB are identical, whereas GBL has a lower
MW. 1,4-BD was also analyzed, but not included as an endpoint
because we expected to find it in negligible concentrations.
Secondary outcomes included GHB, GBL, and BHB levels in
pregnant women compared to female and male controls separately. Correlations between GHB concentrations and duration of pregnancy and use of hormonal contraceptives were
also examined. Post hoc, we decided to examine the false-positive rate in pregnancy and non-pregnancy at different cut-off
levels.
Analytical methods and chemicals
Analytical-grade formic acid and HPLC-grade methanol were
supplied by Merck (Darmstadt, Germany). GHB sodium salt
and GHB-d6 in methanol (1 mg/mL) were from Lipomed (Arlesheim, Switzerland). GBL-d6 (1 mg/mL in acetonitrile) was
purchased from Cerilliant (Round Rock, TX) and 1,4-BD was
from Riedel-de Haën (Seetze, Germany). GBL was from Fluka
(Steinheim, Germany) and BHB was from Aldrich (Steinheim,
Germany). Water was obtained from a Milli-Q purification
system (Millipore, Bedford, MA). A mixed stock solution with
GHB, GBL, BHB, and 1,4-BD was prepared by adding the compounds to synthetic urine to a final concentration of 104 mg/L.
Synthetic urine was prepared according to the method of Kark
et al. (20) with some modifications. Standards were prepared by
dilution of the stock solution with synthetic urine at the following concentrations: 6.66, 3.33, 1.66, 0.83, 0.42, 0.21, and
0.105 mg/L. Quality control (QC) samples at 5.2 and 0.52 were
prepared in the same manner. The samples, standard solutions, and QCs were stored at –80°C before analysis. Internal
standard solution was prepared by adding GBL-d6 and GHB-d6
to Milli-Q water to a final concentration of 8 mg/L.
Analyses of GHB, BHB, 1,4-BD, and GBL were performed in
a similar manner as previously described by Wood et al. (21) on
a Waters Acquity UPLC (Waters, Milford, MA) interfaced to
Waters Micromass Quattro Premier XE tandem-quadrupole
MS (Waters, Manchester, U.K.). The system was controlled by
MassLynx™ version 4.1. The chromatography was performed
on a 2.1- × 100-mm Waters Acquity HSS T3 (C18) column
maintained at 50°C. The mobile phase consisted of 8%
methanol in 0.5 g/L aqueous formic acid with a flow rate of 0.4
mL/min (isocratic). The method was validated and showed
good linearity from 0.1 to 83 mg/L (r2 > 0.997) for GHB, and
from 0.1 to 30 mg/L (r2 > 0.992) for BHB, GBL, and 1,4-BD.
Limits of detection (LOD) and quantification (LOQ) for GHB
and 1,4-BD were found to be 0.04 and 0.07 mg/L, respectively.
For GBL and BHB, the LOD and LOQ were 0.06 and 0.1 mg/L,
respectively. Accuracy was 101.1% with a relative standard de-
395
Aronsen_jatsept10.qxd:JATLynneTemplate
8/9/10
4:49 PM
Page 3
Journal of Analytical Toxicology, Vol. 34, September 2010
viation (RSD) of 6.6% at 5–33 mg/L, and intermediary precision varied from 3.2 to 5.6% RSD for GHB. The creatinine
concentrations were determined by using a method based on
the Jaffe reaction (Microgenics DRI Creatinine-Detect Test,
Microgenics and Olympus AU400) (22).
number of days to term predicted by ultrasound in the pregnant women.
Several studies have examined determinants for physiological GHB concentrations (2,3,4,12,14). Considerable intraindividual and interindividual variations in urine GHB
concentrations have been reported (4,10,12). Some authors
Possible confounders
reported gender differences (10), whereas others did not find
Potential confounders that were recorded include age, body
such differences (4), but these studies were limited by low
mass index (BMI), smoking status, time since last alcohol insample sizes. In a large study, no difference in median GHB
take, use of hormonal contraceptives by female controls, and
concentration was found between males (n = 130) and females (n = 77) (2). There may also be
Table I. Descriptive Data
differences in endogenous GHB concentration related to ethnicity (4,10), but
Pregnant Women
Controls
Female Controls Male Controls
another study (n = 207) did not report
(n = 66)
(n = 69)
(n = 52)
(n = 17)
such differences (2). In one study (n =
50), urinary GHB concentrations deAge
30.7
24.7*
24.7*
24.9*
creased with age in women, but this age
Weight (kg)
76.2
68.3*
64.0*
81.3*,†
effect was compensated for when correcting for creatinine concentrations (3).
Height (cm)
166
170*
167
183*,†
Another study found no significant age
BMI‡ (kg/m2)
27.6
23.3*
23.0*
24.2*
differences (2). Factors such as diet or
diabetes mellitus do not seem to influ†
Urine creatinine
9.84
7.86*
6.86*
10.92
ence GHB in urine (2,12,14). Another
concentration (mmol/L)
small study (n = 15) found no significant
Urine pH
6.33
6.62*
6.67*
6.48
effect of smoking or drinking alcohol on
GHB concentrations (4).
* p < 0.05 compared to pregnant women (independent sample t-test).
p < 0.05 compared to female controls (independent sample t-test).
We expected the non-pregnant controls
BMI: body mass index.
to be younger and have a lower BMI than
the pregnant women. Smoking, alcohol
consumption, and the use of therapeutic
Table II. Concentrations of GHB, GBL, BHB, GHB+GBL, and GHB+BHB (mg/L)
drugs and oral contraceptives are potenin Pregnant Women and Controls
tial confounding factors that are likely to
be less frequent in the pregnant populaPregnant Women
Controls
Difference for Mean
tion. Age and height are the only identi(n = 66)
(n = 69)
(95% Confidence Interval)
fied possible confounders that are unreGHB
Mean
0.36
0.24
0.12
lated to pregnancy.
†
‡
GBL
BHB
GHB+GBL
GHB+BHB
396
Median
0.32
0.20
Range
0–1.06
0–1.27
Mean
0.34
0.08
Median
0.21
0.03
Range
0–2.32
0–1.79
Mean
1.92
0.40
Median
0.90
0.16
Range
0–14.07
0–8.79
Mean
0.70
0.32
Median
0.54
0.22
Range
0–3.14
0–2.24
Mean
2.28
0.71
Median
1.43
0.36
Range
0–14.89
0–9.24
(0.04–0.19)
p = 0.002
0.26
(0.14–0.39)
p < 0.001
1.53
(0.85–2.20)
p < 0.001
0.38
(0.22–0.54)
p < 0.001
1.64
(0.94–2.34)
p < 0.001
Study size
In 50 previous samples analyzed at our
laboratory, the average creatinine-adjusted GHB concentration was 170 µg
GHB/mmol creatinine. The standard deviation was 152 µg GHB/mmol creatinine.
To detect a difference of 85 µg GHB/mmol
creatinine (50%), with a significance level
of 0.05 and a power of 0.9, we needed 67
participants in each group. These samples
were analyzed with an assay not differentiating GHB from BHB, and the concentrations used in the study size calculation
were higher than what we expected from
the more refined assay used in the crosssectional study. We aimed to include at
least 50 female controls, achieving a
power of 0.8 for subgroup analysis of female controls. The false-positive rate for
GHB was examined at four different previously proposed cut-off levels.
Aronsen_jatsept10.qxd:JATLynneTemplate
8/9/10
4:49 PM
Page 4
Journal of Analytical Toxicology, Vol. 34, September 2010
Statistical methods
Independent sample t-tests were used to compare the difference of means in unadjusted and creatinine-adjusted GHB,
GBL, and BHB concentrations with a 95% confidence interval
(CI) between pregnant women and all controls and by controls
stratified by gender. The difference of means with a 95% CI was
also used to compare unadjusted and creatinine-adjusted GHB,
GBL, and BHB concentrations between female users and nonusers of hormonal contraceptives in the control group. Simple
linear regression was used to examine correlations of adjusted
or unadjusted GHB concentrations to duration of pregnancy.
The covariates age, height, and urinary pH were controlled by
multivariable linear regression. We used odds ratios (OR) to examine the false-positive rate in pregnancy and non-pregnancy
at different cut-off levels. Statistical tests were performed using
Statistical Package for the Social Sciences (SPSS) version 14.0.
Results
Of the 68 pregnant women who wanted to participate in the
study, 66 were found eligible for analysis. One was excluded because of a missing questionnaire and one because of a missing
consent form. Of the 72 non-pregnant controls who wanted to
participate, 69 were found eligible for analysis. Two were excluded because of positive pregnancy tests and one because of
a positive benzodiazepine test.
Descriptive data of the study participants
are given in Table I. The pregnant
Table III. Creatinine-Adjusted Concentrations (µg/mmol creatinine) of GHB,
women
had a mean age of 30.6 years comGBL, BHB, GHB+GBL, and GHB+BHB in Pregnant Women and Controls
pared to 24.7 years in the control group.
Other recognized baseline differences in
Pregnant Women
Controls
Difference for Mean
weight, height, smoking and drinking
(n = 66)
(n = 69)
(95% Confidence Interval)
habits, and use of hormonal contracepGHB
Mean
39.2
41.5
–2.2
tives could be attributed to pregnancy
(–12.2–7.7)
status or gender.
Median
35.2
27.2
GBL
BHB
GHB+GBL
GHB+BHB
Range
0–90.2
Mean
40.8
11.4
Median
21.2
3.0
Range
0–335.3
0–256.5
p = 0.7
0–196.9
Mean
234.2
65.2
Median
109.8
27.2
Range
0–1832.9
0–1259.4
Mean
80.0
52.8
Median
60.5
33.5
Range
0–379.0
0–320.7
Mean
273.4
106.7
Median
169.1
61.9
Range
0–1876.7
0–1323.5
Main results
29.4
(12.8–45.9)
p = 0.001
169.0
(77.7–260.3)
p < 0.001
27.1
(6.4–47.9)
p = 0.01
156.3
(72.7–260.7)
p = 0.001
Table IV. False-Positive Rates and Odds Ratios for False-Positive Tests in
Pregnancy at Different Cut-Off Levels for GHB+BHB
Cut-Off
Level
(mg/L)
Pregnant Women
(n = 66)
(False-positive rate %)
Controls
(n = 69)
(False-positive rate %)
Odds Ratio for False-Positive
in Pregnant Women
(95% Confidence interval)
2
20 (30.3)
3 (4.3)
9.6 (2.7–34.1)
5
6 (11.0)
1 (1.4)
6.8 (0.8–58.1)
10
2 (3.0)
0
∞
20
0
0
–
GHB, GBL, and BHB urine concentrations in pregnant women and all controls
are given in Table II, and creatinine-adjusted GHB, GBL, and BHB concentrations are presented in Table III. All unadjusted concentrations were significantly
higher in the pregnant women than in
the control group. When adjusting for
creatinine, there was no difference in
GHB concentrations, but GBL, BHB,
GHB+GBL, and GHB+BHB were significantly higher in the pregnant women.
When stratifying the non-pregnant group
by gender (Figures 1 and 2), the female,
but not the male group had significantly
lower mean unadjusted GHB concentrations than the pregnant women. There
was no difference in creatinine-adjusted
GHB concentrations between pregnant
women or male or female controls. Both
male and female controls had significantly
lower crude and creatinine-adjusted GBL,
BHB and GHB+BHB concentrations than
the pregnant women. GHB+GBL corrected for creatinine was significantly
lower in male but not female controls
compared to the pregnant women. One
sample, from a female control, had extremely high creatinine-adjusted concentrations of all substances. Creatinineadjusted GHB in this person was more
than two times higher than the highest
observed in the pregnant population. Post
397
Aronsen_jatsept10.qxd:JATLynneTemplate
8/9/10
4:49 PM
Page 5
Journal of Analytical Toxicology, Vol. 34, September 2010
hoc analysis excluding this person did not significantly change
our main findings. 1,4-BD was not detected in any of the
samples.
Our study confirms the hypothesis that pregnant women
have higher average urine levels of not only GHB (1.5-fold), but
also GBL (4.3-fold) and BHB (4.8-fold). However, when adjusting for creatinine, the GHB concentration is similar in
pregnant women and the control group. GBL and BHB remain elevated among the pregnant women even after cor-
recting for creatinine.
Fortunately, most published GHB assays have no interference from BHB, and for these methods endogenous GHB or
GHB+GBL levels in pregnancy are expected to be well below
the recommended cut-off limit. However, our results show
that LC–MS–MS assays that measure the sum of GHB and
BHB have an increased risk of false-positive tests in pregnant
women.
One important limitation of this study is the difficulty in
defining a control group where pregnancy is the only different
characteristic. Even in prospective studies with fertile women
being their own controls before or after pregnancy, there will
be bias due to pregnancy-induced behavioral changes. These
give rise to potential confounding factors such as different
smoking or drinking habits. Our control group consisted of
students that were not matched to the pregnant women. There
were indeed significant differences in age, the pregnant women
being older, but according to earlier reports, this should not affect the creatinine-adjusted concentrations (3), and in one
comprehensive study on variations of urinary endogenous
GHB, there was no age effect in the age range of our participants (2).
The observed GHB concentrations in both the pregnant and
the non-pregnant population were somewhat lower than what
has been found by others (3,10,12–14). One possible explanation is that in vitro production of GHB was minimized by
freezing the samples at –80°C shortly after voiding. In vitro
production of GHB in urine is well known, and is dependent on
time to analysis and storage temperature (15,16). Our assay
measures GHB and GBL separately, whereas other methods involve a step in which GHB is converted to GBL before analysis.
When using the sum of GHB and GBL, our range of concentrations in the non-pregnant group (0–2.24 mg/L) was similar
to LeBeau et al. (2) (0–2.70 mg/L). Our results will also be
lower than results from assays not differentiating between
GHB and BHB. Finally, there could be unknown common environmental factors among our study participants that could
influence the levels of GHB, GBL, and/or BHB.
The collection of urine samples from the pregnant women
and non-pregnant controls was performed at different loca-
Figure 1. Mean urine concentrations ± 95% confidence interval of GHB,
GBL, BHB, GHB+GBL, and GHB+BHB in pregnant women (P) and female
(F) and male (M) controls.
Figure 2. Mean creatinine-adjusted urine concentrations ± 95% confidence interval of GHB, GBL, BHB, GHB+GBL, and GHB+BHB in pregnant
women (P) and female (F) and male (M) controls.
Other analysis
Multivariable linear regression revealed no association between unadjusted or creatinine-adjusted GHB, GBL or BHB
concentration and age, height, or urine pH (data not shown).
All pregnant women were in their second or third trimester,
and the mean length of pregnancy was 221.1 days (range 162–
277 days). The duration of pregnancy did not affect the GHB
concentrations (data not shown). There was no difference in
unadjusted or creatinine-adjusted GHB, GBL or BHB concentrations between the genders (Figures 1 and 2) or between female users and non-users of hormonal contraceptives in the
control group (data not shown).
All measured GHB and GHB+GBL concentrations were
below the standard 10 mg/L interpretative cut-off level, but in
two (3.0%) samples from the pregnant group, the GHB+BHB
concentrations were above this limit. Table IV shows the falsepositive rates for GHB+BHB concentrations at different previously suggested cut-offs (2,10–14). The odds ratios for falsepositive GHB+BHB tests among pregnant women were
significantly higher at low cut-off values. None of the samples
had creatinine-adjusted GHB or GHB+GBL concentrations
above the suggested cut-off at 1000 µg/mmol creatinine, but
creatinine-adjusted GHB+BHB gave false-positive rates at 6.1%
among the pregnant women and 1.4% in the control group
(OR: 4.4, 95% CI: 0.5–40.3).
Discussion
398
Aronsen_jatsept10.qxd:JATLynneTemplate
8/9/10
4:49 PM
Page 6
Journal of Analytical Toxicology, Vol. 34, September 2010
tions. Longer time from voiding until the urine sample was
frozen among the pregnant women could be one explanation
for the higher GHB concentration in these samples. The exact
time between voiding and freezing was not recorded, but all
collectors were instructed that the urine samples should be
frozen as soon as possible after voiding.
To our knowledge, this is the first study designed to examine
physiological GHB, GBL, and BHB concentrations in urine
from pregnant women. One study (n = 207) reported that the
GHB concentrations were within the normal range (0.12–0.72
mg/L, mean 0.36 mg/L, median 0.34 mg/L), but this study only
included six women who were either pregnant or had recently
given birth (2). Our results should encourage further studies to
examine possible explanations for elevated concentrations of
GHB, GBL, and BHB in pregnancy, such as increased production or increased renal excretion compared to other endogenous substances. Both maternal factors, such as latent diabetes mellitus in which BHB is shown to be elevated (14), or
fetal production of GHB, GBL, and BHB could be underlying
causes.
Our results imply that concentrations of endogenous GHB,
GBL, and BHB in urine are indeed elevated in pregnant women
in the second and third trimesters, although no significant
differences were found for creatinine-adjusted GHB concentrations. The combined concentrations of GHB and BHB were
above the 10 mg/L interpretative cut-off in two pregnant
women. In order to minimize the risk of false-positive results
when testing urine from pregnant women for legal purposes,
we suggest that only methods specifically identifying GHB
should be applied, a higher cut-off should be considered, and
that creatinine-adjusted concentrations should be used. The robustness of our main results when controlling for potential
confounding factors indicates that our findings also apply to
pregnant women in their third and late second trimesters in
general.
Acknowledgments
We thank Gunnbjørg Andreassen at the midwife centre of
Tromsø for coordinating recruitment of the pregnant women,
Anita Seljelund for validating the LC–MS–MS method, and
Inger Carlson for excellent technical assistance.
The study was entirely funded by the Department of Clinical
Pharmacology, University Hospital of North Norway Trust.
References
1. 1. S. Abanades, M. Farré, M. Segura, S. Pichini, D. Barral, R. Pacifici, M. Pellegrini, F. Fonseca, K. Langohr, and R. de la Torre.
Gamma-hydroxybutyrate (GHB) in humans: pharmacodynamics
and pharmacokinetics. Ann. N.Y. Acad. Sci. 1074: 559–576
(2006).
2. M.A. LeBeau, M.A. Montgomery, C. Morris-Kukoski, J.E. Schaff,
A. Deakin, and B. Levine. A comprehensive study on the variations
in urinary concentrations of endogenous gamma-hydroxybutyrate
(GHB). J. Anal. Toxicol. 30: 98–105 (2006).
3. C.E. Crookes, M.C. Faulds, A.R. Forrest, and J.H. Galloway. A reference range for endogenous gamma-hydroxybutyrate in urine by
gas chromatography–mass spectrometry. J. Anal. Toxicol. 28: 644–
649 (2004).
4. F. Moriya, H. Nishimura, J. Furumiya, and Y. Hashimoto. Effects of
drinking and smoking on endogenous levels of urinary gamma-hydroxybutyric acid, a preliminary study. Leg. Med. (Tokyo) 8: 231–
234 (2006).
5. A. Kankaanpaa, R. Liukkonen, and K. Ariniemi. Determination of
gamma-hydroxybutyrate (GHB) and its precursors in blood and
urine samples: a salting-out approach. Forensic Sci. Int. 170: 133–
138 (2007).
6. P. Kintz, M. Villain, A.L. Pelissier, V. Cirimele, and G. Leonetti. Unusually high concentrations in a fatal GHB case. J. Anal. Toxicol.
29: 582–585 (2005).
7. S. Mazarr-Proo and S. Kerrigan. Distribution of GHB in tissues and
fluids following a fatal overdose. J. Anal. Toxicol. 29: 398–400
(2005).
8. S.D. Ferrara, L. Tedeschi, G. Frison, and A. Rossi. Fatality due to
gamma-hydroxybutyric acid (GHB) and heroin intoxication.
J. Forensic Sci. 40: 501–504 (1995).
9. F.C. Kugelberg, A. Holmgren, A. Eklund, and A.W. Jones. Forensic
toxicology findings in deaths involving gamma-hydroxybutyrate.
Int. J. Legal Med. 124(1): 1–6 (2010).
10. M.A. LeBeau, R.H. Christenson, B. Levine, W.D. Darwin, and
M.A. Huestis. Intra- and interindividual variations in urinary concentrations of endogenous gamma-hydroxybutyrate. J. Anal. Toxicol. 26: 340–346 (2002).
11. D.T. Yeatman and K. Reid. A study of urinary endogenous gammahydroxybutyrate (GHB) levels. J. Anal. Toxicol. 27: 40–42 (2003).
12. S.P. Elliott. Gamma hydroxybutyric acid (GHB) concentrations in
humans and factors affecting endogenous production. Forensic
Sci. Int. 133: 9–16 (2003).
13. A.A. Elian. Determination of endogenous gamma-hydroxybutyric
acid (GHB) levels in antemortem urine and blood. Forensic Sci.
Int. 128: 120–122 (2002).
14. N. Shima, A. Miki, T. Kamata, M. Katagi, and H. Tsuchihashi.
Urinary endogenous concentrations of GHB and its isomers in
healthy humans and diabetics. Forensic Sci. Int. 149: 171–179
(2005).
15. M.A. LeBeau, M.A. Montgomery, C. Morris-Kukoski, J.E. Schaff,
and A. Deakin. Further evidence of in vitro production of gammahydroxybutyrate (GHB) in urine samples. Forensic Sci. Int. 169:
152–156 (2007).
16. S. Kerrigan. In vitro production of gamma-hydroxybutyrate in antemortem urine samples. J. Anal. Toxicol. 26: 571–574 (2002).
17. M.A. LeBeau, M.L. Miller, and B. Levine. Effect of storage temperature on endogenous GHB levels in urine. Forensic Sci. Int.
119: 161–167 (2001).
18. S.A. Barker, O.C. Snead, F. Poldrugo, C.C. Liu, F.P. Fish, and
R.L. Settine. Identification and quantification of 1,4-butanediol in
mammalian tissues: an alternative biosynthetic pathway for
gamma-hudroxybutyric acid. Biochem. Pharmacol. 34: 1849–
1852 (1985).
19. R.D. Perrone, N.E. Madias, and A.S. Levey. Serum creatinine as an
index of renal function: new insights into old concepts. Clin.
Chem. 38: 1933–1953 (1992).
20. R.M. Kark, J.R. Lawrence, R.C. Muehrcke, C.L. Pirani, V.E. Pollack,
and H. Silva. Primer of Urinalysis, 2nd ed. Hoeber Medical Division, Harper and Row, New York, NY, 1964.
21. M. Wood, M. Laloup, N. Samyn, M.R. Morris, E.A. de Bruijn,
R.A. Maes, M.S. Young, V. Maes, and G. De Boeck. Simultaneous
analysis of gamma-hydroxybutyric acid and its precursors in urine
using liquid chromatography–tandem mass spectrometry. J. Chromatogr. A 1056: 83–90 (2004).
22. S. Narayanan and H.D. Appleton. Creatinine: a review. Clin.
Chem. 26: 1119–1126 (1980).
Manuscript received October 26, 2009;
revision received February 10, 2010.
399