Determination of gamma-hydroxybutyrate (GHB) in biological

Forensic Science Journal
2006;5:41-54
Available online at:fsjournal.cpu.edu.tw
FORENSIC SCIENCE
JOURNAL SINCE 2002
Determination of gamma-hydroxybutyrate (GHB) in
biological specimens by simultaneous extraction and
chemical derivatization followed by GC-MS
1,
Sheng-Meng Wang, * Ph.D.; Yun-Seng Giang,1 Ph.D.; Min-Jen Lu,1 B.S.; Tsung-Li Kuo,2 Ph.D.
1Department
of Forensic Science, Central Police University, Taoyuan33304, Taiwan, ROC
2, Graduate School of Forensic Medicine, National Taiwan University, Taiwan, ROC
Received: September 26, 2006 /Received in revised form: October 19, 2006 /Accepted: October 23, 2006
Abstract
Urine and chicken liver fortified with gamma-hydroxybutyrate (GHB) were pretreated with in-situ liquid-liquid
extraction/chemical derivatization (LLE-ChD) or in-situ solid-phase extraction/chemical derivatization (SPE-ChD)
followed by gas chromatography-mass spectrometry (GC-MS). GHB as its N,O-bis(trimethylsilyl)-trifluoroacetamide
(BSTFA) derivative was recovered from urine in 23.7 % through the LLE-ChD procedure, in contrast to 60.7 % via
the SPE-ChD counterpart. In the selective ion monitoring (SIM) mode, the qualifier ions selected for GHB-BSTFA
were m/z 233, 234 and 235, with m/z 233 being the quantifier ion. Meanwhile, m/z 239, 240 and 241 were the qualifier
ions for GHB-d6-BSTFA, and m/z 239 the quantifier ion. These ions have made all the cross contributions between
the analyte and the corresponding internal standard (IS) smaller than 3.0 %. The calibration curves plotted for GHB
in urine is linear within 500－10000 ng/mL, with correlation coefficients typically exceeding 0.999. The limits of
detection and quantitation obtained via LLE approach are 300 and 500 ng/mL, respectively, and those via SPE approach,
100 and 200 ng/g, respectively. For urine samples fortified with 4000 and 8000 ng/mL of GHB, up to 32.6 % of the
GHB decomposed after a period of fifteen days. In acidic media, 23.8 % of γ-butyrolactone (GBL) on average were
hydrolyzed into GHB, whereas 11.8 % of GHB were converted to GBL.
Keywords: Gamma-hydroxybutyrate (GHB), γ-Butyrolactone, 1,4-Butanediol, Club drug, Gas chromatography-mass
spectrometry (GC-MS), Urine drug testing, Drug urinalysis
Introduction
Following the footsteps of such “club drugs” as
MDMA, FM2, ketamine, etc., the newly emerging gammahydroxybutyrate (IUPAC name: γ-hydroxybutyric acid;
common name: GHB) has also been increasingly abused
in a similar fashion among teenagers and young adults
over the past couple years. It has often been used for
criminal offenses, mostly as a date-rape drug.
GHB is a substance naturally present in trace
* Corresponding author, e-mail:[email protected]
amount within mammal species. Properties of
neurotransmitter or neuromodulator are generally given
to this substance [1]. GHB was first synthesized in 1961
for anaesthetic use [2]. In Europe it has more often
been used in the treatments of alcoholism [3-6], opiate
withdrawal syndrome [7,8], and narcolepsy [9-11], while
in the States it is just being evaluated for treating the
latter.
Oral administration of GHB by the abuser is
followed by rapid gastric and/or intestinal absorption
42
Forensic Science Journal 2006; Vol. 5, No. 1
of the drug, with its half-life being ca. 0.3－1.0 h. The
climax of GHB level in urine for a 100-mg/kg oral dose
is around 110 mg/dL and is usually reached no later than
4 h after the ingestion [12,13]. Regardless of the dosage,
no more than 5 % of the orally ingested GHB will be
detected intact in urine at any time, no urinary intact
GHB at all will be detected at 12 h after the ingestion,
and nearly all of the metabolites will be gone from the
urine within the first 24 hours [1]. In contrast, the peak
of blood GHB for a 25-mg/kg oral dose comes as high
as 8 mg/dL yet as early as 0.5 h after the ingestion [14].
With this dosage most GHB user feel dizzy and sleepy.
Thus, although the climaxes of GHB in both blood and
urine shift toward later periods as the dosage increases
[15], the fact is that the window of detection of GHB is
very short in these fluids and its presence is very difficult
to prove after a rape case [1,16]. The metabolism of
GHB is illustrated in Fig. 1.
Fig. 1 Metabolism of GHB
GHB and its metabolites are currently not one
of the NIDA-5 standardly tested for in the basic drug
test, nor are they included in the extended drug tests
[17]. This along with the short detection period and
the inherent cross-interferences with the analysis has
made GHB test much more expensive to give than the
basic test. Nevertheless, in recent years, the tremendous
increase in the abuse of GHB has prompted worldwide
forensic toxicology laboratories to incorporate its
analysis into their routine screening procedures. So far
as the analysis of GHB is concerned, although GHB
is a naturally occurring constituent of human body,
the average level of endogenous GHB is reportedly
so low (ranging 0.9－3.5 µg/mL with a mean of 1.65
µg/mL) as to be negligible compared to the proposed
1000-µg/mL cutoff concentration [18]. However,
two organic solvents structurally similar or related to
GHB, i.e., γ-butyrolactone (GBL) and 1,4-butanediol
(1,4-BD; simply denoted BD in this paper), if involved
in the sample, may complicate or interfere with the
analysis. On the one hand, once absorbed into human
body these two solvents are readily converted into
GHB under the catalysis of alcohol dehydrogenase and
lactonase as is shown in Fig. 2. On the other hand, the
acidic conditions employed in some LLE, SPE, solidphase micro-extraction (SPME), and/or derivatization
procedures or even during the course of instrumental
analysis may well transform in vitro the GHB into
GHB urinalysis　43
GBL through an intramolecular esterification reaction
[19-26]. It is therefore paramount that the quantification
of GHB can be differentiated from those of GBL and
BD, or kept free from the interferences of the latter
two. Beside the improvements that have been made on
the above stated sample preparation methods [22-26],
isotope dilution GC-MS protocols employing GHB-d6
as IS [21,24] as well as capillary electrophoresis (CE)
and high performance liquid chromatography (HPLC)
free from degradation of GHB to GBL commonly
occurring on the GC injection port or column have also
been reported [27,28]. Since GC-MS as a confirmatory
test has proved most evidential and worldwide adopted
for drug testing [22,29,30], the present study is devoted
to the development of an analytical scheme comprising
a compatible sample preparation procedure, optimal
GC-MS parameters/conditions, and quality data
interpretation.
Fig. 2 In-vivo transformation of 1,4-BD into GHB, and in-vivo interconversion between GHB and GBL.
Experimental
Materials
Authentic GHB (1 mg/mL in methanol), GBL
(1 mg/mL in methanol), and GHB-d6 (100 µg/mL in
methanol) were purchased from Cerilliant Co., USA.
They were each diluted tenfold with methanol to serve as
the respective stock solutions, i.e., 100 µg/mL for GHB
and GBL, and 10 µg/mL for GHB-d6. The derivatizing
agents, N,O-bis(trimethylsilyl)-trifluoroacetamide
(BSTFA) fortified with 1% trimethylchlorosilane, was
ordered from Aldrich Chemical Co., USA. The SPE
cartridge, CLEAN SCREENTM ZSGHB020, containing
200 mg of CSGHB sorbent in a 10-mL column was
obtained from United Chemical Technologies, Inc. All
the other chemicals/solvents used in this study were in
analytical or reagent grade, were from local chemical
suppliers, and were directly used without further
purification.
Sample collection and spiking
44
Forensic Science Journal 2006; Vol. 5, No. 1
Urine previously certified negative for all drugs
was contributed by a voluntary student from the Central
Police University. Drugs-free chicken liver was bought
at a local supermarket in Taipei and was homogenized
with physiological saline prior to spiking. To simulate
the actual situation in Taiwan, all spiking steps were
performed while the blank urine and chicken liver
had not yet been refrigerated, but all specimens to be
analyzed were preserved at 4℃ prior to the analysis.
The serial calibrators used for plotting the method
calibration curve were each prepared by adding an
appropriate amount (10, 20, 50, 100, 200 µL for urine;
5, 10, 25, 50, 100 µL for chicken liver) of 100-µg/mL
GHB solution to a screw-cap topped test tube containing
2 mL of blank urine or saline-homogenized chicken liver
(containing 1 g of liver tissue), and 100 µL of 10-µg/mL
GHB-d6 solution. (This would make the corresponding
urinary spikes contain 500, 1000, 2500, 5000, and
10000 ng/mL, respectively, of GHB and 500 ng/mL
of GHB-d6 and make the corresponding chicken liver
spikes represent 500, 1000, 2500, 5000, and 10000 ng/g,
respectively, of GHB and 500 ng/g of GHB-d6.) This
study did not involve case work. If a real-case specimen
is to be analyzed for GHB, simply 100 µL of 10-µg/mL
GHB-d6 needs to be added at this starting step.
a screw-cap topped test tube was added 100 µL of pH
6 phosphate buffer. After 3 min of vortex, the mixture
was passed (under reduced pressure, 20 mmHg) through
an SPE cartridge that had been preconditioned in
succession with 3 mL of methanol, 3 mL of D.I. water,
and 0.5 mL of pH 6 phosphate buffer. The foregoing
sample test tube was thoroughly rinsed with 1 mL of
freshly prepared CH3OH/NH4OH (99:1,v/v) and the
filtration under reduced pressure was repeated. The two
portions of filtrate were combined in another screwcap topped test tube, purged at 60℃ with nitrogen gas
to dryness, and then cooled to ambient temperature.
The residues were re-dissolved with 200 µL of N,Ndimethylformamide (DMF) and 2 mL of n-hexane. After
10 min of vortex and 5 min of centrifugation (3000 rpm),
the lower layer was transferred to a screw-cap topped
concentration tube and purged at 60℃ with nitrogen gas
to dryness. A 100-µL portion of BSTFA was added. The
mixture was incubated with vortex at 70℃ for 20 min,
cooled to ambient temperature, and then purged with
nitrogen gas to dryness. Bring the volume to 100 µL
with EA. A 1-µL aliquot of the reconstituted solution
was injected for the GC-MS analysis.
General procedure of simultaneous LLE and BSTFAChD
The GC-MS analyses were carried out using a
Hewlett-Packard HP-5890 Series II gas chromatograph
coupled to an HP-5971 Series mass selective detector
(MSD). The GC column used was an HP-5 MS capillary
column (25 m * 0.2 mm I.D., 0.33μm film thickness).
The GC was operated in the splitless mode (i.e., purge
off) when performing injection with the aid of an
HP-7673 autosampler, but 1 min later the purge valve
was turned on. The injector temperature was 250℃.
For the analysis of BSTFA-derivatized GHB, the column
temperature was programmed from 60 to 100℃ at 15℃/
min, then from 100 to 250℃ at 25℃/min, with the initial
temperature held for 2 min and final temperature held
for 5 min. Helium of 99.999 % purity was used as the
carrier gas at a flow-rate of 1 mL/min. Effluents from
the GC column was transferred via a transfer line held at
280℃ to a 70-eV electron impact (EI) ionization source
held at 180℃.
The instrument was first operated in full-scan
mode where relevant total ion current (GC-EIMS TIC)
chromatograms and mass spectra were acquired to look
into the fragmentation nature (with the aid of a software
named “High ChemTM Mass Frontier Version 1.0”) of
the previously unexplored BSTFA-derivatized analytes,
To 2 mL of urine or saline-homogenized chicken
liver tissue fortified with GHB and GHB-d6 (see
“sample collection and spiking” subsection) in
a screw-cap topped test tube were added 0.2 mL of
0.5 N HCl and 6 mL of ethyl acetate (EA). After 10
min of vortex (LLE), the mixture was subjected to
centrifugation at 3000 rpm for 5 min. The supernatant
was transferred to a screw-cap topped concentration
tube and purged with nitrogen gas to dryness. A
100-µL portion of BSTFA was added. The mixture was
incubated with vortex at 70℃ for 20 min, cooled to
ambient temperature, and then purged with nitrogen gas
(Model N-EVAPTM112 equipped with twenty four Teflon
tubes, Organomation Associates Inc., USA) to dryness.
Bring the volume to 100 µL with EA. A 1-µL aliquot of
the reconstituted solution was injected for the GC-MS
analysis.
Simultaneous SPE and BSTFA-ChD
To 2 mL of urine fortified with GHB and GHB-d6
(see “sample collection and spiking” subsection) in
GC-MS analysis
GHB urinalysis　45
and then in selected ion monitoring mode (GC-EIMS
SIM) to further evaluate the qualifier and quantifier ions
and run the formal analysis. The calibration curves
were produced by plotting the quantifier-ion-abundance
ratio (analyte : IS) obtained from the SIM measurement
against the concentration of the analyte in the fortified
samples. Each peak-area ratio used was the mean of
triplicate analyses.
Results and discussion
Mass chromatography
Possessing two protic moieties on each of their
molecules, GHB, GHB-d6, and BD were readily diderivatized with BSTFA (denoted di-TMS-GHB, diTMS-GHB-d6, and di-TMS-BD, respectively, where
“TMS” stands for a “trimethylsilyl group”). In
Fig. 3
the GC-EIMS TIC chromatogram shown in Fig. 3, while
di-TMS-GHB and di-TMS-BD are well separated, diTMS-GHB and di-TMS-GHB-d6 are also sufficiently
resolved in the sense of analyte-IS quantifier ion pair
cross contribution which is crucial to the accuracy
of quantitation (see the subsection that follows). In
contrast, the intramolecularly dehydrated GBL was
neither capable of BSTFA-ChD nor co-detectable
with the other three compounds in the GHB-oriented
analysis although GBL and GHB were more or less
inter-convertible under the acidic conditions adopted
for sample preparation. Thus, the GC-EIMS TIC
chromatogram obtained (with the same instrumental
parameters as those for Fig. 3) upon the direct injection
of 1 µL of underivatized GBL in ethyl acetate is shown
separately in Fig. 4.
The isotope dilution GC-EIMS TIC chromatogram obtained for a urinary specimen containing 1,4-BD,
GHB-d6, GHB, and GBL after the LLE/BSTFA-ChD pretreatment. GBL was not co-detected in this GHBoriented analysis.
46
Forensic Science Journal 2006; Vol. 5, No. 1
Fig. 4
The GC-EIMS TIC chromatogram obtained upon the direct injection of 1 µL of GBL in ethyl acetate. The
instrumental conditions/parameters are the same as those for Fig. 3.
Selection of qualifier and quantifier ions
As the analyses of the chromatographically noninterfering GBL and 1,4-BD are not our primary concern
and their deuterium-labeled ISs are not commercially
available, we shall focus the discussion on the GCEIMS analysis of GHB and GHB-d6. Judging from
our previous experience in analyzing amphetamines
[31] and ketamine [32] as well as the relative mass
difference between di-TMS-GHB and its deuteriumlabeled IS, satisfactory GC resolution between the latter
two compounds might need the use of an IS labeled
with more than eight or nine deuterium atoms. Being
more available, however, the d6-labeled IS was used
throughout the present study. This along with the
demand of rapid analysis in forensic practice did lead to
superficially inadequate GC separation between di-TMSGHB and di-TMS-GHB-d6 (Fig. 3). The appreciable
overlap between the analyte and IS peaks caused their
respective fragment ions to be mutually mixed up.
Fortunately, through the more appropriate yet easier
selection of qualifier/quantifier ions described below, the
essential or effective resolution and hence the accuracy
and precision for the to-be-routine GC-EIMS SIM
analyses can still be secured.
Shown in Fig. 5 are the mass spectra of GBL, diTMS-1,4-BD, di-TMS-GHB, and di-TMS-GHB-d6
acquired in the full-scan mode. Those fragment ions
which looked more intense and characteristic made good
candidates of the respective qualifier and/or quantifier
ions, and were further subjected to a detailed “analyteIS ion-pair cross-contribution evaluation” based on the
total analysis (incorporating LLE-ChD and GC-EIMS
SIM) of a working solution containing 5000 ng/mL
each of GHB and GHB-d6 in methanol. Displayed in
Table 1 are the relevant cross contributions calculated
for the considered analyte-IS ion pairs. In the light of
“minimized cross contribution,” the qualifier ions
decided for the GC-EIMS SIM analysis of di-TMS-GHB
were m/z 233, 234, and 235, and those for di-TMSGHB-d6 were m/z 239, 240, and 241, with their cross
contributions typically below 3 %. Of these ions, the
most abundant m/z 233 and 239 were decided to be the
quantifier ions for di-TMS-GHB and di-TMS-GHB-d6,
respectively. Since the base peak ion (m/z 147; assigned
to be [TMS－O－DMS]＋, where “DMS” stands for
a “dimethylsilyl group”) was common to di-TMSGHB and di-TMS-GHB-d6, it should be considered as
neither a qualifier nor the quantifier ion.
GHB urinalysis　47
Fig. 5
Mass spectra of (a) GBL, (b) di-TMS-1,4-BD, (c)di-TMS-GHB, and (d) di-TMS-GHB-d6
48
Forensic Science Journal 2006; Vol. 5, No. 1
Table 1 Relative intensities and cross contributionsa of ions with potential for designating the analyte and the IS
Analyte (di-TMS-GHB)
(Full-scan)
(SIM) % intensity
Ion (m/z)
Relative
contributed by diintensity (%)
TMS-GHB-d6
117
15.0
8.34
Ion (m/z)
120
IS (di-TMS-GHB-d6)
(Full-scan)
(SIM) % intensity
Relative
contributed by diintensity (%)
TMS-GHB
5.46
2.20
b
147
100
NC
147
100
NCb
204
6.57
1.84
206
4.70
11.1
29.9
0.65
233c,d
27.9
1.72
239e,f
234d
5.55
1.85
240f
6.12
0.68
3.14
241f
2.84
0.73
235d
2.52
a
Relative intensities are based on full-scan data and expressed in percentage, while analogs’contributions (crosscontributions) are derived from SIM data and expressed also in percentage.
b Not calculated.
c Decided quantifier ion for the analyte.
d Selected qualifier ion for the analyte.
e Decided quantifier ion for the IS.
f Selected qualifier ion for the IS.
Quantitation
As the accuracy and precision for the measurement
of drug level is highly crucial to the following legal
interpretation of the results, every quantifier-ionabundance ratio used for plotting the calibration curves
was the mean of triplicate analyses. Summarized in
Table 2 are the relevant five-point calibration equations,
linearity ranges, linear correlation coefficients (r2),
and method limits of detection and quantitation
(LODs/LOQs; previously defined [32,33]) for the
determination of GHB in urine and chicken liver by the
two protocols under comparison. The SPE protocol
afforded significantly lower LODs/LOQs than did the
LLE protocol. Although there have been no officially
adopted cutoff concentrations for GHBs, the LODs/
LOQs achieved hereby sufficiently meet the previously
proposed 10-µg/mL cutoff for routine forensic GHB
urinalysis, while that cutoff far overpasses the foregoing
endogenous GHB level [18].
Table 2 Calibration equations, linearity ranges, linear correlation coefficients (r2), and method limits of detection and
quantitation (LODs/LOQs) for the determination of GHB in urine and chicken liver by simultaneous extraction
and BSTFA-ChD followed by GC-EIMS SIM
Linearity range
(ng/mL for urine;
Calibration equation
ng/g for chicken
liver)
(a) GC-EIMS SIM instrumental calibration a
y=0.0023x-0.081
500－10000
Linear
correlation
coefficient
(r2)
LOD
(ng/mL for
urine; ng/g for
chicken liver)
LOQ
(ng/mL for
urine; ng/g for
chicken liver)
0.999
(b) Method calibration for simultaneous LLE and BSTFA-ChD followed by GC-EIMS SIM
Urine: y=0.0021x+0.7059
500－10000
0.999
400
500
Chicken liver: y=0.0023x+0.5331
500－10000
0.999
400
500
(c) Method calibration for simultaneous SPE and BSTFA-ChD followed by GC-EIMS SIM
Chicken liver: y=0.0021x-0.0224
a
500－10000
0.999
Standard GHB unextracted and quantitative BSTFA-ChD presumed.
100
200
GHB urinalysis　49
Analyte recoveries indicative of method performances
A five-pointed instrumental calibration curve (i.e.,
standard GHB unextracted and quantitative BSTFA-ChD
presumed) was established in advance with the equation
being y = 0.0023x-0.081 (Table 2). Five levels (within
the working range, 500－10000 ng/mL) of urinary GHB
spikes were then subjected respectively to the above
described simultaneous extraction and BSTFA-ChD
followed by GC-EIMS SIM analysis. The recoveries
of GHB were obtained by dividing the hereby regressed
concentration of GHB (actually recovered in the form
of di-TMS-GHB) by the corresponding concentration
originally spiked. Thus, two series of recoveries of GHB
from urine achieved via LLE and SPE, respectively, are
displayed in Table 3. The SPE process with recoveries
ranging 44.0－84.5 % and the mean being 60.7 % has
proved much more effective than the LLE process with
recoveries ranging 20.8－26.7 % and the mean being
23.7 %. The foregoing higher LODs/LOQs for the
LLE protocol may be largely attributed to its poorer
recoveries.
Table 3 Recoveries of GHB from urine achieved via LLE and SPE a
Level of GHB (ng/mL) b
Recovery (％）
LLE
SPE
500
22.37
84.47
1000
20.77
66.07
2500
23.36
43.96
5000
25.32
46.42
10000
26.69
62.69
Mean = 23.70
Mean = 60.72
a
b
Quantitative ChD presumed.
All tested levels are within the LLE and SPE’
s linearity range, 500－10000 ng/mL.
Influences of temperature and time length on the
efficiency of BSTFA-derivatization of GHB.
Peak area ratio (Analyte:IS)
The temperature (70℃) and time length (20 min)
for the BSTFA-derivatization of GHB described in the
experimental section was not adopted until they turned
out to be the optimal. Fixed amounts of GHB and
GHB-d6 in urine (2 mL × 5000 ng/mL) were subjected
to the foregoing LLE-ChD procedure with the ChD being
carried out at 60, 70, 80, and 90℃, and for 5, 10, and
20 min as variants. The resulting di-TMS-derivatives
were submitted to the GC-EIMS SIM analysis, and the
relative ChD efficiencies in terms of analyte-to-IS ratio
are portrayed in Fig. 6. Apparently, 70℃ and 20 min
were the optimal conditions and were used throughout
the study.
Fig. 6 Influences of temperature and time length on the efficiency of BSTFA-derivatization of GHB
50
Forensic Science Journal 2006; Vol. 5, No. 1
Comparison among five solvents with respect to their
efficiencies in extracting GHB from urine.
the relative LLE efficiencies in terms of analyte-to-IS
ratio are contrasted in Table 4. Although chloroform/
isopropanol (80:20, v/v) is superior to the other four in
doing the job, the second most efficient solvent, EA, was
chosen for use throughout the study due to environmental
protection consideration and its better compatibility with
GC-MS analysis.
A fixed amount (6 mL) of EA, chloroform/
isopropanol (80:20, v/v), n-haxane, acetonitrile, and
methylene chloride as the extraction solvent was
respectively applied to the foregoing LLE-ChD of 2 mL
of 5000-ng/mL GHB. The resulting di-TMS-derivatives
were submitted to the GC-EIMS SIM analysis, and
Table 4 Efficiencies achieved for the LLE of GHBa from urine using five solventsb
Solvent
Acetonitrile
Efficiency
(in terms of analyte-to-IS peak-area ratio)
0
Ethyl acetate
2.55
n-Hexane
0
Chloroform/isopropanol (80:20, v/v)
4.92
Methylene chloride
0
a
b
GHB used was 5000 ng/mL × 2 mL.
The amount of solvent used for each extraction was 6 mL.
Optimization of LLE efficiency by pH adjustment
The optimization process was performed in two
steps. First, various normalities (12, 6, 3, 1, 0.5, and 0.1
N) of hydrochloric acid were respectively applied to the
foregoing LLE-ChD procedure while keeping the other
parameters unchanged. The resulting di-TMS-derivatives
were submitted to the GC-EIMS SIM analysis, and the
relative LLE efficiencies in terms of analyte-to-IS ratio
are profiled in Fig. 7. The efficiency increases with
decreasing HCl normality until a maximum is reached
at 0.5 N, where the efficiency begins to turn down
drastically to that resulting from 0.1 N HCl. Afterward
various amounts (100, 200, 300, 400, 500, and 600 µL)
of the optimal (0.5 N) HCl were respectively applied to
the same LLE-ChD procedure while keeping the other
parameters invariant. The resulting di-TMS-derivatives
were submitted to the GC-EIMS SIM analysis again,
and the relative LLE efficiencies in terms of analyte-toIS ratio are profiled in Fig. 8. The optimal amount of
0.5 N HCl for the LLE turned out to be 200 µL, which is
equivalent to pH 1-2.
Fig. 7 Variation of efficiency with pH in the LLE of GHB from urine.
Peak area ratio (Analyte;IS)
GHB urinalysis　51
Fig. 8 Variation of efficiency with the amount of 0.5 N HCl used in the LLE of GHB from urine.
In vitro interconversion between GHB and GBL
Putting aside the previously mentioned negligible
phenomenon of endogenous GHB, this subsection deals
with the potential problem caused by the also foregoing
interconversion between GHB and GBL, as 0.5 N HCl
was routinely used for our LLE-ChD procedure.
Shown in Table 5 are the estimated percentages of
GBL in urine (5000 ng/mL×2 mL, initially) hydrolyzed
into GHB after going through the LLE-ChD procedure
employing various normalities (0.1, 0.5, 1, 3, 6, and 12
N) of 200-µL HCl. The percentage of GBL hydrolysis is
directly dependent upon the amount of GHB recovered
after the LLE-ChD process. Although 0.1 N HCl led
to the smallest percentage of GBL hydrolysis (3 %),
that was attributed mainly to the above shown (Fig. 7)
lowest LLE-ChD efficiency for GHB in urine under
0.1 N HCl. The truly smallest and most desirable GBL
hydrolysis percentage should be the one estimated under
0.5 N HCl, i.e., an average of 23.8 % based on triplicate
analyses, with the RSD being 2.9 %. The higher the HCl
normality, the larger the GBL hydrolysis percentage.
The largest GBL hydrolysis percentage was 47.5 %
estimated for 12 N HCl. Since normally an individual
should not have ingested any GBL and his endogenous
GBL should also be negligibly low, we would usually
ignore the error caused by GBL hydrolysis for the
quantification of GHB.
Table 5 Percentages of GBL in urine (5000 ng/mL×2 mL, initially) hydrolyzed into GHB after going through the
LLE-ChD procedure employing various normalities of 200-µL HCl
HCl normality
Final GHB level (ng/mL)
% GBL hydrolyzed into GHB
0.1
164
Trial 1 1222
Trial 2 1188
Trial 3 1152
3.3
24.5
23.8
23.1
Mean = 1187
RSD = 2.9 ﹪
Mean = 23.8
1256
1447
1468
2374
25.1
28.9
29.4
47.5
0.5
1
3
6
12
52
Forensic Science Journal 2006; Vol. 5, No. 1
Table 6 presents the estimated percentages of GHB
in urine (5000 ng/mL×2 mL, initially) that converted
into GBL after going through the LLE-ChD procedure
employing various normalities (0.1, 0.5, 1, 3, 6, and 12
N) of 200-µL HCl. The percentage of GHB esterification
is proportional to the difference between the two GHB
levels before and after the LLE-ChD process. Once
again the percentage of GHB esterification estimated
for 0.1 N HCl is unreliable due to the too poor GHB
recovery after the LLE-ChD process. The percentage
estimated for 0.5 HCl (average of triplicate analyses,
11.8 %; RSD, 3.5 %) is smaller than those for 6 and
12 N HCl (25.3 %), comparable with that for 3 N HCl,
but larger than that for 1 N HCl (2.2 %). Probably
due to non-quantitative LLE recovery and ChD yield,
the percentages estimated for the corresponding GHB
esterification and GBL hydrolysis, apparently, will not
correlate well with the theoretical equilibrium constant
for GHB-GBL interconversion (unavailable, though).
However, since the extent of GHB esterification is
already significant under 0.5 N HCl, if it were not for
the reason stated in the following subsection, it would be
suggested that for routine GHB-case samples a relative
percentage of 10 be added to the GHB outcome to
compensate for the possible negative error.
Table 6 Percentages of GHB in urine (5000 ng/mL×2 mL, initially) converted into GBL after going through the
LLE-ChD procedure employing various normalities of 200-µL HCl
HCl normality
Final GHB level (ng/mL)
% GHB converted into GBLa
0.1
42
Trial 1 4407
Trial 2 4570
Trial 3 4258
Mean = 4412
99.2
11.9
8.6
14.8
Mean = 11.8
0.5
1
3
6
12
a
RSD = 3.5 ﹪
4890
4456
4189
3734
2.2
10.9
16.2
25.3
Estimated from the difference between the two GHB levels before and after the LLE-ChD process.
Stability of GHB in urine
Keeping the above stated in vivo and in vitro
instability of GHB in mind, we conducted an experiment
to further understand the durableness of GHB urine
specimens. Displayed in Table 7 are the percentages of
GHB in urine (4000 and 8000 ng/mL×2 mL, initially)
that converted automatically into GBL after 2, 3, and
15 days of preservation at 4℃. For the 4000-ng/mL
specimens, no GHB esterification was discerned
after 2 or 3 days of storage, but 32.6 % of GHB was
converted into GBL after 15 days of storage. For the
8000-ng/mL specimens, 5.9, 7.4, and 29.5 % of the
GHB was transformed into GBL after 2, 3, and 15
days, respectively, of preservation. That is, the higher
the tested GHB level and the longer the sample stored,
the more appreciable the GHB loss. Thus, depending
on the firstly detected GHB level, an appropriate
compensation needs to be made for the storage-loss of
GHB. Theoretically, this addition to GHB level should
have incorporated the contribution from the foregoing
“relative 10 %”. Unfortunately, relevant data obtained
hereby seem to correlate poorly. More detailed and
precise evaluation and calibration are in need. At this
time, it is still safe to suggest that for routine GHB-case
samples a relative percentage of 10 be added to the GHB
outcome to compensate for the possible negative error.
GHB urinalysis　53
Conclusions
The results presented in this report demonstrated
that simultaneous extraction and BSTFA-ChD followed
by GC-EIMS SIM is a promising protocol for the
determination of GHB in urine and other biological
specimens. While the GC-EIMS SIM counterpart leaves
no problem with the identification and quantification
of GHB, the SPE protocol provides higher extraction
efficiency and, hence, lower LOD/LOQ than the LLE
protocol. However, the LLE approach is better suited
for the exploration of many basic yet unestablished
analytical properties of the newly targeted GHB, such
as solvent and pH effects on the extraction efficiency,
possible source of analytical errors, and sample
durableness. For either of the proposed protocols to be
more efficient, accurate, and precise, we shall undertake
more detailed evaluation and calibration with respect to
the above stated properties.
Acknowledgment
Financial support by the National Bureau of
Controlled Drugs, Department of Health, Republic of
China under Grant No. DOH91-NNB-1011 is gratefully
acknowledged.
References
1. Kintz P, Cirimele V, Jamey C, Ludes, B. Testing for
GHB in hair by GC/MS/MS after a single exposure
application to document sexual assault. J Forens Sci
2003; 8(1):195-200 .
2. L a b o r i t H, J o u a n y J M, G e r a l d J, F a b i a n i F.
Generalities concernant l’etude experimentale de
l’emploi clinique du gamma hydroxybutyrate de
Na. Aggressologie 1960; :407.
3. Addolorato G, Castelli E, Stefanini GF, Casella G,
Caputo F, Marsigli L, Bernardi M, Gasbarrini G. An
open multicentric study evaluating 4-hydroxybutyric
acid sodium salt in the medium-term treatment of
179 alcohol dependent subjects. Alcohol Alcoholism
1996; 1(4):341-5.
4. Agabio R, Colombo G, Loche A, Lobina C, Pani
ML, Reali R, Gessa GL. Gamma-hydroxybutyric
acid reducing effect on ethanol intake: evidence
in favour of a substitution mechanism. Alcohol
Alcoholism 1998; 3(5):465-74.
5. Galloway GP, Frederick SL, Staggers FE Jr,
Gonzoles M, Stalcup SA, Smith DE. Gammahydroxybutyrate: an emerging drug of abuse that
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
causes physical dependence. Addicition 1997;
2(1):89-96.
Poldrugo F, Addolorato G. The role of gammahydroxybutyric acid in the treatment of alcoholism:
from animal to clinical studies. Alcohol Alcoholism
1999; 4(1):15-24.
Gallimberti L, Cibin M, Pagnin P, Sabbion R,
P i r a s t u P P, F e r r a r a S D, G e s s a G L. G a m m ahydroxybutyric acid for the treatment of opiate
withdrawal syndrome. Neuropsychopharmacol
1993; (1):77-81.
Rosen MI, Pearsall HR, Woods SW, Kosten TR.
Effects of gamma-hydroxybutyric acid in opioiddependent patients. J Subst Abuse Treat 1997;
4(2):149-54.
Mamelak M, Scharf MB, Woods M. Treatment of
narcolepsy with gamma-hydroxybutyrate. A review
of clinical and sleep laboratory findings. Sleep
1986; (1):285-9.
S c h a r f M B, L a i A A, B r a n i g a n B, S t o v e r R,
Berkowitz DB. Pharmacokinetics of gammahydroxybutyrate in narcoleptic patients. Sleep
1998; 1(5):507-14.
Scrima L, Hartman PG, Johnson FH Jr, Thomas
E E, a n d H i l l e r F C. T h e e ff e c t s o f g a m m ahydroxybutyrate on the sleep of narcolepsy patients:
a double-blind study. Sleep 1990; 3(6):479-90.
Baselt R. Gamma-hydroxybutyrate in disposition of
toxic drugs and chemicals in man, 5th Ed. Chemical
Toxicology Institute, Foster City, CA, 2000.
Anderson D. Gamma hydroxybutyrate analysis
and interpretation. Calif Assoc Toxicol Proc 2001;
9(1):43-48.
P a l a t i n i P, Te d e s c h i L, F r i s o n G. D o s edependent absorption and elimination of gammahydroxybutyric acid in healthy volunteers. Eur J
Clin Pharm 1993; 5:353-356.
Ferrara SD, Zotti S, Tedeschi L, Frison G, Castagna F,
Gallimerti L, Gessa GL, Palaini P. Pharmacokinetics
of gamma-hydroxybutyric acid in alcohol dependent
patients after single and repeated oral dose. Br J
Clin Pharmacol 1992; 4(3):231-5.
Verstraete AG. Detection times of drugs of abuse
in blood, urine, and oral fluid. Therap Drug
Monitor 2004; 6(2):200-205. ＜http://www.drugmonitoring.com＞
Erowid. GHB and drug tests. The vaults of Erowid,
19 July 2006. ＜http://www.erowid.org/chemicals/
ghb/ghb_testing.shtml＞
54
Forensic Science Journal 2006; Vol. 5, No. 1
18. Ye a t m a n D T, R e i d K. A S t u d y o f U r i n a r y
Endogenous Gamma-Hydroxybutyrate (GHB)
Levels.J Anal Toxicol 2003; 7(1):40-42.
19. Ferrara SD, Tedeschi L, Frison G, Castagna F,
Gallimberti L, Giorgetti R, Gessa GL, Palatini P.
Therapeutic gamma-hydroxybutyric acid monitoring
in plasma and urine by gas chromatographymass spectrometry. J Pharm Biomed Anal 1993;
1:483-487.
20. Doherty JD, Snead OC, Roth RH. A sensitive
method for quantitation of gamma-hydroxybutyric
acid and gamma-butyrolactone in brain by electron
capture gas chromatography. Anal Biochem 1975;
9:268-277.
21. Vree TB, Kleijn E ven der, Knop HJ. Rapid
determination of 4-hydroxybutyric acid (Gamma
OH) and 2-propyl pentanonate (Depakine) in human
plasma by means of gas-liquid chromatography. J
Chromatogr 1976; 21:150-152.
22. L e B e a u M A, M o n t g o m e r y M A, M i l l e r M L,
Burmeister SG. Analysis of biofluids for gammahydroxybutyrate and gamma-butyrolactone by
headspace GC-FID and GC-MS. J Anal Toxicol
2000; 4:421-428.
23. Couper FJ, Logan BK. Determination of gammahydroxybutyrate in biological specimenss by gas
chromatography-mass spectrometry. J Anal Toxicol
2000; 4:1-7.
24. M c C u s k e r R R, P a g e t-Wi l k e s H, C h r o n i s t e r
C W, G o l d b e rg e r B A. A n a l y s i s o f g a m m ahydroxybutyrate in urine by gas chromatography/
mass spectrometry. J Anal Toxicol 1999; 3:301-305.
25. Ciolino LA, Mesmer MZ, Satzger RD, McCauley
AC, Mohrhaus AS. The chemical interconversion of
GHB and GBL: forensic issues and implications. J
Forensic Sci 2001; 6(6):1315-1323.
26. Blair S, Song M, Hall B, Brodbelt J. Determination
of gamma-hydroxybutyrate in water and human
u r i n e b y s o l i d p h a s e m i c r o e x t r a c t i o n-g a s
chromatography/quadrupole ion trap spectrometry.
J Forensic Sci 2001; 6(3):688-693.
27. Mesmer MZ, Satzger RD. Determination of
gamma-hydroxybutyrate and gamma-butyrolactone
b y H P L C/U V-V I S s p e c t r o m e t r y a n d H P L C/
thermospray mass spectrometry. J Forensic Sci
1998; 3:3:489-492.
28. De Vriendt CA, ven Sassenbroeck DK, Rosseel
MT, van de Velde EJ, Verstraete AG, Heyden YV,
Belpaire FM. Development and validation of a
29.
30.
31.
32.
33.
high-performance liquid chromatographic method
for the determination of gamma-hydroxybutyric acid
in rat plasama. J chromatogr 2001; 52:85-90.
E l i M, C a t t a b e n i F. E n d o g e n o u s g a m m ahydroxybutyrate in rat brain areas: postmortem
changes and effects of drugs interfering with
gamma-aminobutyric acid metabolism. J neurochem
1983; 1:524-530.
E h r h a r d t J D, Va y e r P h, M a i t r e M. A r a p i d
and sensitive method for the determination of
gamma-hydroxybutyric acid and trans-gammahydroxycrotonic acid in rat brain tissue by gas
chromatography/ mass spectrometry with negative
ion detection. Biomed Environ Mass Spec 1988;
5:512-524.
Wang S-M, Giang Y-S, Ling Y-C. Simultaneous
supercritical fluid extraction and chemical
derivatization for the gas chromatographic-isotope
dilution mass spectrometric determination of
amphetamine and methamphetamine in urine. J
Chromatogr B 2001; 59:17-26.
Chou S-L, Yang M-H, Ling Y-C, Giang Y-S. Gas
chromatography-isotope dilution mass spectrometry
preceded by liquid-liquid extraction and chemical
derivatization for the determination of ketamine
and norketamine in urine. J Chromatogr B 2004;
99:37-50.
W a n g S - M , W a n g T- C , G i a n g Y- S .
Simultaneous determination of amphetamine
and methamphetamine enantiomers in urine
by simultaneous liquid-liquid extraction and
diastereomeric derivatization followed by gas
chromatography-isotope dilution mass spectrometry.
J Chromatogr B 2005; 816:31-143.