A hydrophilic interaction liquid chromatography electrospray tandem mass

spectrometry method for the simultaneous determination of g-hydroxybutyrate

and its precursors in forensic whole blood
Lambert K. Srensen*, Jrgen B. Hasselstrm
Section for Toxicology and Drug Analysis, Department of Forensic Medicine, Aarhus University, Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark
1. Introduction
The drug g-hydroxybutyric acid (GHB) is a short-chain
hydroxylated fatty acid that is endogenously present in various
mammalian tissues, blood and urine. It is a powerful central
nervous system depressant and was used in legal medicine as an
anaesthetic agent prior to being phased out because of its very
steep dose-response curve and various side effects. Today, GHB is
mainly used for the treatment of narcolepsy and less frequently
during the treatment of alcohol dependence [1]. However, GHB is
also abused for recreational purposes and is available as both a
colourless, odourless liquid and as a solid material on the illicit
market. GHB related fatalities have been reported [2,3] and the
drug has been found in both the body uids and hair from
suspected victims of drug-facilitated sexual assaults [4,5].
Although GHB is a controlled drug, the pharmacologic effects of
GHB may be obtained indirectly from the intake of the
commercially available and legal chemicals g-butyrolactone
(GBL) and 1,4-butanediol (1,4-BD). Both are used as industrial
solvents and they may be constituents in various cleaners or used
during the synthesis of other chemicals (e.g. 1,4-BD is used during
the production of tetrahydrofuran). Following the direct intake of
GBL, the substance is rapidly converted to GHB by the lactonase
enzymes present in the blood of mammals. However, GBL can also
be easily converted to GHB by treatment with lye in a simple non-
industrial (kitchen) process. In humans, 1,4-BD is converted to GHB
through the combined actions of alcohol and aldehyde dehydro-
genases. GHB is rapidly and extensively metabolised in the body.
Approximately 1% is excreted unchanged in urine [6].
In forensic toxicology, body uids are monitored for sub-
stances that may have been abused, for instance, by vehicle
drivers. However, testing for GHB in biological matrices is
challenging because of its polar properties, its equilibrium state
with GBL favouring GBL at lowpH, and the complete conversion of
GLB to GHB at high pH values. Determination of GHB in biological
uids is typically performed by gas chromatography mass
spectrometry (GCMS) after derivatisation of the substance, such
as silylation. The use of LCMS/MS allows for omission of the
derivatisation step, but poor retention on analytical columns is
often observed. Methods based on LCMS/MS of the native
Forensic Science International 222 (2012) 352359
A R T I C L E I N F O
Article history:
Received 20 April 2012
Received in revised form 17 July 2012
Accepted 28 July 2012
Available online 20 August 2012
Keywords:
Gamma-hydroxybutyrate (GHB)
Gamma-butyrolactone (GBL)
1,4-Butanediol
Whole blood
LCMS/MS
A B S T R A C T
A liquid-chromatographytandem-mass-spectrometry method using pneumatically assisted electro-
spray ionisation (LCESI-MS/MS) was developed for the simultaneous determination of g-hydro-
xybutyric acid (GHB), g-butyrolactone (GBL) and 1,4-butanediol (1,4-BD) in human ante-mortem and
post-mortem whole blood. The blood proteins were precipitated using a mixture of methanol and
acetonitrile, and the extract was cleaned-up by passage through a polymeric strong cation exchange
sorbent. Separation of the analytes and their structural isomers was obtained using a column with a
zwitterionic stationary phase. Matrix-matched calibrants, combined with isotope dilution, were used for
quantitative analysis. GHB was determined in both positive and negative ion modes. The relative intra-
laboratory reproducibility standard deviations were better than 10% and 6% for blood samples at
concentrations of 2 mg/L and 20150 mg/L, respectively. The mean true extraction recoveries were 80%
for GHB and greater than 90% for GBL and 1,4-BD at concentration levels of 2050 mg/L. The limits of
detection were approximately 0.5 mg/L for GHB and GBL, and 0.02 mg/L for 1,4-BD in ante-mortem
blood. The corresponding lower limits of quantication were less than 1 mg/L for GHB and GBL, and less
than 0.1 mg/L for 1,4-BD. GBL was unstable in whole blood freshly preserved with a sodium uoride
oxalate mixture, but the stability could be improved signicantly by preservation with a sodiumuoride
citrate EDTA mixture.
2012 Elsevier Ireland Ltd. All rights reserved.
* Corresponding author. Tel.: +45 87167500; fax: +45 86125995.
E-mail address: lks@forensic.au.dk (L.K. Srensen).
Contents lists available at SciVerse ScienceDirect
Forensic Science International
j ou r nal h o mepage: w ww. el sevi er . co m/ l oc at e/ f o r sc i i nt
0379-0738/$ see front matter 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.forsciint.2012.07.017
molecule have been published for the determination of GHB in
plasma/serum, [7,8] and for the simultaneous determination of
GHB, GBL and 1,4-BD in whole blood and urine [9,10]. These
methods use reverse phase chromatography for the separation of
the components and acidied mobile phases to obtain some
retention of GHB, which restricts the electrospray ionisation (ESI)
to be performed in the positive ion mode where only a single
abundant product ion is produced. To improve the chro-
matographic and retention properties, derivatisation has also
been applied in the determination of GHB in urine and serum
using LCMS [11]. However, the chromatographic separation of
small polar molecules can often be directly improved through the
use of hydrophilic interaction liquid chromatography (HILIC)
instead of classical reverse phase chromatography. The HILIC
conditions allow the chromatography to be performed under
neutral conditions.
The present LCMS/MS method is based on the HILIC analysis of
extracts that have been cleaned-up using cation-exchange solid
phase extraction (SPE) for adsorption of interfering substances. The
method was validated as a robust technique, suitable for the
determination of GHB, GBL and 1,4-BD in both ante-mortem and
post-mortem whole blood samples. Several abundant transition
products of GHB were obtained for the proper identication of the
substance.
2. Materials and methods
2.1. Chemicals
a-Hydroxybutyric acid (AHB) sodium salt, b-hydroxybutyric acid (BHB) sodium
salt, GHB, GBL, 1,4-BD and 2,3-butanediol (2,3-BD) were obtained from Sigma
Aldrich (Schnelldorf, Germany). The isotope analogues, GHB-D6 and GBL-D6, were
obtained from Cerilliant (Round Rock, Texas) and 1,4-BD-D4 was obtained from
Cambridge Isotope Laboratories (Andover, MA). Whole blood samples, used for
calibration, were obtained from the Blood Bank, Aarhus University Hospital (Skejby,
Denmark). Formic acid, ammonium acetate and sodium dihydrogen phosphate
were purchased from Merck (Darmstadt, Germany). Acetonitrile (MeCN) and
methanol (MeOH) of LCMS grade were purchased from SigmaAldrich. Water was
puried using a Direct-Q 3 apparatus (Millipore, Bedford, MA).
2.2. Samples
Samples of ante-mortem and post-mortem whole blood, used for validation of
the method, were obtained from the Department of Forensic Medicine, Aarhus
University. Ante-mortem blood samples that were used in the validation study
were preserved in Venosafe VF-053SFC32 tubes that contained 6.8 mg of sodium
uoride (NaF) and 15.7 mg of citrate-EDTA buffer ingredients (FC mixture) for a
3 mL draw volume of blood (Terumo Europe, Leuven, Belgium). Post-mortem blood
samples were preserved with 200 mg of NaF per 30 mL of blood. Venosafe VF-
109SFX07 tubes (Terumo Europe) that contained 100 mg of NaF and 22.5 mg of
potassium oxalate (FO mixture) for a 9-mL draw volume of blood and Venosafe VF-
052SDK tubes (Terumo Europe) that contained 3.9 mg K
2
EDTA for a 2-mL draw
volume of blood were included in a stability study on GHB and its precursors in
ante-mortem whole blood.
2.3. Standards
Separate stock solutions containing 1 mg/mL of the active substances were
prepared in MeOH and stored at 20 28C. The combined standard solutions for the
fortication of the samples and preparation of the calibrants were prepared by diluting
the stock solutions with MeOH. An internal standard solution (IS) containing 0.1 mg/
mL of the deuterated analogues of GHB, GBL and 1,4-BD was prepared in MeOH.
2.4. Equipment
The liquid chromatography system was a Waters Acquity UPLC system that
consisted of a binary pump, an autosampler with a 10-mL sample loop
thermostated at 7 2 8C and a column oven thermostated at 30 2 8C (Waters,
Milford, MA). The mass spectrometer was a Waters Xevo TQMS triple-quadrupole
instrument with an ESI ion source and a programmable infusion pump. The
separation was performed using a SeQuant ZIC HILIC (5 mm, 200 A

, 2.1 mm
I.D. 100 mm) column (Merck SeQuant, Umea , Sweden). Solid phase extraction
(SPE) was performed on a 3-mL Strata-X-C cartridge containing 60 mg of a polymeric
strong cation exchange (SCX) sorbent (Phenomenex, Torrance, CA). A VacMaster-20
vacuum manifold (Biotage, Uppsala, Sweden) was used during the SPE procedure.
Disposable 2-mL polypropylene Safe-Lock tubes (Eppendorf, Hamburg, Germany)
were used for the extractions. Autosampler vials made of glass (Mikrolab, Aarhus,
Denmark) were used for storage of the nal extracts. Other equipment used included
pipettes (Biohit, Helsinki, Finland) and a Heraeus Biofuge Pico (Thermo Scientic,
Langenselbold, Germany).
2.5. Extraction and clean-up
A 200-mL volume of sample was transferred to a disposable 2 mL centrifuge tube.
Then, 100 mL of MeOH and 100 mL the IS were added, and the tube contents were
mixed gently. Shortly thereafter, a 600-mL volume of MeCN was added and the tube
was immediately closed and vigorously vortex-mixed for a few seconds. After a
standing time of 510 min, the mixture was centrifuged at 10,000 g for 5 min. A
volume of 600 mL of the supernatant was mixed with 250 mL of water and was then
passed through a CX SPE cartridge with a ow rate of max 0.5 mL/min. The cartridge
was previously conditioned sequentially with 1 mL of MeOH, 1 mL of water, 1 mL of
a 1 M sodium dihydrogen phosphate solution and 1 mL of water and then dried
under full vacuum suction for few seconds. The eluate was collected, and 200 mL of
the eluate was mixed with 600 mL of MeCN in an autosampler vial. The total sample
dilution factor was 28.
2.6. Calibration
The calibrants for both the ante-mortem and post-mortem samples were based
on blank donor blood from single persons preserved with FC mixture. The samples
were treated using the same procedure with the exception that the 100 mL of MeOH
was replaced by 100 mL of the mixed standards of the drug substances. Sample
concentrations were obtained at 5, 50, 100, 150 and 200 mg/L of GHB, and 2.5, 25,
50, 75 and 100 mg/L of GBL and 1,4-BD in the original blood sample. In addition, a
blank sample (a processed matrix sample without analyte and without IS) and a
zero sample (a processed matrix sample with IS) were included to verify the
absence of detectable concentrations of the analytes. The calibration curves were
created from weighted (1/x) linear regression analysis of the IS-normalised peak
areas (analyte area/IS area) and were forced through the origin.
2.7. LCMS/MS analysis
The sample extracts were maintained at 7 2 8C prior to analysis. A 10-mL
volume was injected onto a SeQuant ZIC HILIC column running 5% mobile phase A
(1 mM ammonium acetate) and 95% mobile phase B (MeCN). The mobile phase was
changed through a linear gradient to 50% A and 50% B over 4 min. Then, the
gradient was changed to 95% A over 0.2 min. Five minutes after injection, the
gradient was returned to 5% A over 0.5 min, and the column was equilibrated for
4.5 min before the next injection resulting in a total runtime of 10 min. The column
ow rate was 200 mL/min and the column temperature was maintained at
30 2 8C. During the interval of 0.52.5 min, a reagent containing 0.1% formic acid
was infused post column with a owrate of 100 mL/min. The eluent was diverted to
waste during the time interval of 49 min after injection by using a post-column
switch. The source and desolvation temperatures were set at 150 8C and 600 8C,
respectively, and the cone and desolvation nitrogen gas ows were set at 50 L/h and
800 L/h, respectively. The mass spectrometer was operated in positive ion mode
with a probe voltage of 3 kV and in negative ion mode with a probe voltage of
2.5 kV. The dwell time was 5064 ms depending on the number of ion transitions
processed during the same time period. At least 12 data points were obtained
across the peaks. Selected reaction monitoring (SRM) was applied using the
conditions shown in Table 1. Argon was used for collision-induced dissociation
(CID). The data acquisition and processing were performed using MassLynx 4.1
(Waters).
2.8. Limits of detection and quantication
The limits of detection (LODs) were determined using a random selection of 20
different samples of ante-mortem whole blood and 20 different samples of post-
mortem blood. The samples were spiked prior to extraction with the individual
substances in order to obtain concentrations that were approximately three to ve
times the signal/noise ratio. The LODs were calculated as 2 t
0.95
SD
B
(t
0.95
= 1.645), where SD
B
is the standard deviation of the results obtained from
the spiked samples. The lower limits of quantication (LLOQs) were calculated as 10
times the SD
B
of the quantier ions.
2.9. Precision, trueness and recovery
The repeatability standard deviation (SD
r
) (i.e., the variability of independent
analytical results obtained by the same operator using the same apparatus under
the same conditions on the same test sample and in a short interval of time) and the
intra-laboratory reproducibility standard deviation (SD
R,intra-lab
) (i.e., the variability
of independent analytical results obtained on the same test sample in the same
laboratory by different operators on different days) were determined on blank
control samples of ante-mortem whole blood spiked to levels of 2, 20 and 75 mg/L
of GBL and 1,4-BD, and 2, 20 and 150 mg/L of GHB and from post-mortem samples
with an endogenous content of GHB. Duplicate analyses were performed on eight
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 353
different days. The repeatability and intra-laboratory reproducibility parameters
were calculated in accordance with ISO standard 57252 [12].
The matrix effects (including ion-suppression and ion-enhancement effects) were
investigated on 20 blank samples of both ante-mortem and post-mortem blood that
were spiked after SPE to a level that was equivalent to 20 mg/L in the original samples.
They were analysed in attenuating order along with the pure standards at the same
concentration level; the matrix effect from each sample was calculated from the peak
areas (As) without IS correction using the closest standards in the series;
matrix effect

A
pure standard
A
fortified sample
100 =

A
pure standard
. The true extrac-
tion recoveries were determined from 20 samples of each type that were spiked to a
level of 20 mg/L. The standards used for the determination of the true recoveries were
the sameblank samples that were spiked after SPE. Finally, the trueness of the method,
which is the closeness in agreement between the average value obtained from a large
series of test results and the accepted reference value, was determined from 20
different blank samples of both ante-mortem and post-mortem blood that had each
been spiked with GBL, 1,4-BD and GHB to a level of 20 mg/L with the exception of GHB
in post-mortem blood that was spiked with 50 mg/L. These samples were different
from the samples used in the investigation of ion-suppression effects.
3. Results and discussion
3.1. Precursor ions and transition products
Electrospray ionisation was applied in positive ion mode
(ESI(+)) for GBL and 1,4-BD, and the dominant Q1 ions were the
protonated molecular ions ([M + H]
+
). Signicant abundances of
two product ions were obtained for both substances (Table 1).
Electrospray ionisation of GHB was possible in both positive and
negative ion mode. Three abundant transition products were
obtained in ESI() (Table 1). However, only one product ion of
signicant abundance was obtained in ESI(+). Effective ionisation
of GBL required acidic conditions, whereas the ionisation of 1,4-
BD was only slightly improved in the presence of acids. The
optimal acidic conditions for GBL were obtained by post column
infusion of 0.1% formic acid. Without this infusion, the abundance
of the GBL product ions was reduced by approximately 30 times.
However, the acidic conditions were not optimal for the ionisation
of GHB in ESI() as the responses were reduced almost 8 times.
Therefore the infusion was stopped before the GHB eluted from
the column (more details on the chromatography are given in
Section 3.3). In ESI(+), the ionisation of GHB was not signicantly
inuenced by the presence of acids. Because of the lack of more
than one abundant product ion in ESI(+), an in source conversion
of GHB to GBL has beenusedfor the qualicationof the presence of
GHB in ESI(+) mode [9]. This in source conversion of GHB to GBL
was also observed in the present study but it was not considered
suitable for qualication because of the relative low conversion
factor (6%).
The relative abundances of the transition products were
constant in the calibrated range, i.e. the average RSDs of the ion
ratios across the calibration points in the precision study were less
than 3% for all analytes. The RSDs of the ion ratios obtained from
ante-mortem and post-mortem samples (20 of each) were 8% for
GBL, 3% for 1,4-DB and 2% for GHB at a concentration of 20 mg/L.
3.2. Extraction and clean-up
The blood samples were extracted using a mixture of MeOH and
MeCN. The methanol was added to obtain a disperse precipitate of
the protein-containing blood components. Because of the hydro-
philic properties of the analytes and the lack of ionic functional
groups, except in the case of GHB, it was not possible to perform a
selective clean-up using SPE for all of the analytes. However, it was
possible to remove a signicant portion of the dissolved matrix
components by eluting the extract through a polymeric SCX
sorbent. Sodium was used as counter-ion to avoid conversion of
GHB to GBL. If this clean-up was not applied, skewed peaks for GHB
and a noisier baseline were in some cases observed.
The stability of the blood-based calibrants at 7 2 8C and
20 2 8C was tested daily over a period of ve days. The calibrants
were extracted according to the procedure with the exception that
the internal standard was not added to the blood sample at the
beginning but instead was added to subsamples of the extracts on the
day of the LCMS/MS analysis. The slopes of the calibration curves for
GLB and 1,4-BD did not change signicantly during the ve days of
storage at 7 8C. However, GHB was not stable in solution. The slope of
the calibration curve was reduced by approximately 50% after storing
for ve days at 7 8C. No signicant changes in the slopes were
observed during this storage period at 20 8C. GHB-D6 showed
similar instability as GHB at 7 8C, and when the IS was added prior to
extraction, there were no signicant changes in the slopes of the
calibration curves observed during ve days of storage.
3.3. Chromatography
Using reverse-phase LC, the typical elution order is GHB, 1,4-BD
followed by GBL. As the substances are very hydrophilic, only
limited possibilities are available to control the retention times
(RT), especially for GHB. By using the HILIC technique, the elution
order is reversed, which makes it easier to optimise the RT of GHB
by adjusting the composition of the mobile phase. It was possible
to obtain almost baseline separation between GBL, 2,3-BD, 1,4-BD,
AHB, BHB and GHB within a 4 min elution period (Fig. 1). In ESI()
mode, AHB, but not BHB, produced a signicant response at the
same transitions as GHB. Conversely, in ESI(+) mode, BHB, but not
AHB, produced a signicant response at the same transitions as
GHB. The HILIC technique required neutral elution conditions for
optimal separation of the substances. Neutral elution conditions
are also optimal in preventing interconversion of GBL and GHB, but
are not optimal for the ionisation of GBL. Therefore, 0.1% formic
acid was infused in the uid stream between the column and the
ESI probe. No detectable carry-over was observed when the
samples that had been spiked with 500 mg/L of the individual
substances and blank control samples were analysed in attenuated
order. A typical chromatogram of the blank blood samples is shown
in Fig. 2.
Table 1
Mass spectrometric conditions and relative retention times (RRTs). The bold ions were used for quantication. The underlined ions were used as qualiers. In matrix extracts,
the response of the ESI() transition m/z 103/85 was approximately half of the response of the ESI(+) transition m/z 105/87.
Substance ESI mode Transition Cone voltage (V) Collision energy (eV) Relative abundance RRT
Q1 (m/z) Q3 (m/z)
GBL + 87 45/43 20 10/8 100/65 0.427
1,4-BD + 91 73/55 22 6/10 100/47 0.568
GHB 103 85/57/101 20 8/12/12 100/93/32 0.998
GHB + 105 87/69 12 7/8 100/2 0.998
GBL-D6 + 93 49 20 10 0.428
1,4-BD-D4 + 95 77 20 6 0.570
GHB-D6 109 90 20 8 1.000
GHB-D6 + 111 93 20 7 1.000
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 354
The gradient was changed to 95% A after elution of the analytes
in order to clean the column and to maintain the selectivity
constant from injection to injection. According to the column
manufacturer, it should be possible to improve the selectivity of
the column by ushing with an aqueous salt solution between
injections. We veried the manufacturers recommendation by
ushing the column with a 100 mM NH
4
Ac solution. In general,
however, it was not necessary to do this when the described
gradient was used.
3.4. Matrix effects and quantication
The matrix effects of the blood samples on the ionisation of GLB
and 1,4-BD were close to zero (Table 2). For the more polar GHB, an
ion suppression of approximately 30% was observed in ESI(+).
The same level of ion suppression was reported by Johansen
and Windberg [9]. However, the ion suppression was much higher
in ESI(). This was investigated in more detail through the
continuous infusion of GHB at a concentration of 6 mg/L using a
Time
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
2.05
1.86
2.05
1.53
3.56
1.54
3.56
Time
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
3.55
1.23
1.35 1.85 1.62 2.21 2.332.41
3.38
3.55
3.13
3.55
3.39
m/z 91 > 73
ESI+
m/z 91 > 55
ESI+
m/z 87 > 45
ESE+
m/z 87 > 43
ESE+
1,4-BD
1,4-BD
2,3-BD
2,3-BD
GBL
GBL
In source produced GBL
In source produced GBL
AHB
AHB
GHB
GHB
GHB
BHB
BHB
m/z 103 > 57
ESI-
m/z 103 > 59
ESI-
m/z 103 > 85
ESI-
m/z 105 > 87
ESI+
Fig. 1. Chromatograms of the quantier and qualier product ions in the extract of ante-mortem whole blood that was spiked with 50 mg/L of GBL, 1,4-BD, 2,3-BD, AHB, BHB
and GHB. A chromatogram of the transition m/z 103/59 for BHB is included to show the retention time of this substance.
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 355
ow rate of 50 mL/min after injecting an extract of a blank blood
sample (Fig. 3). Basically, the ESI() transition m/z 103/85 was
more sensitive than the ESI(+) transition m/z 105/87, but at the RT
of GHB, signicant matrix effects caused a reduction of the signal in
ESI(). The gradient itself also caused a reduction in the signal, but
the effect of the matrix induced an almost 80% reduction. In spite of
this reduction, ESI() tended to result in better LODs and precision
gures than ESI(+). The insignicant effects of the ion suppression
on method accuracy may arise from a combination of the
following: (a) the RT of GHB was in the relatively at region of
the suppression curve and the suppression was mostly constant
from sample to sample and (b) an efcient correction could be
obtained by using GHB-D6 as an internal standard because the RTs
of GHB and GHB-D6 were very close, only differing by 0.01 min.
The matrix effect on GHB could be reduced to approximately
40% in ESI() by an additional clean-up in which GHB was adsorbed
to a silica-based strong anion exchange (SAX) sorbent (data not
shown). However, the matrix effect was unchanged in ESI(+).
Time
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.84
1.46 1.30
1.20
1.57
1.62
1.87
1.91
2.33
2.02
2.13 2.42
2.47
2.65
1.85
1.80
1.18
1.24
1.42
1.33
1.57
1.64
1.87
2.44
2.19 1.90
2.04
2.25
2.54
1.23
1.53 1.29
2.24
1.59 2.17
1.79
1.86
1.94
2.35
2.54
2.66
1.44
1.24
1.31 1.85
1.54
1.59
2.19
1.99 2.37
2.44 2.54
Time
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80
%
0
100
3.42
3.37
3.31
2.77
2.93 3.15
3.02
3.19
3.70
3.49
3.62
3.97
3.75 3.92
3.13
3.14
2.95
2.93
3.58
3.29
3.39
3.71
3.92
m/z 91 > 73
ESI+
m/z 91 > 55
ESI+
m/z 87 > 45
ESE+
m/z 87 > 43
ESE+
m/z 105 > 87
ESI+
m/z 103 > 101
ESI-
m/z 103 > 85
ESI-
m/z 103 > 57
ESI-
RT of 1,4-BD
RT of 1,4-BD
RT of GBL
RT of GBL
RT of GHB
RT of GHB
RT of GHB
RT of GHB
Fig. 2. Chromatograms of the quantier and qualier product ions in the extract of a typical blank ante-mortem whole blood sample. The expected retention times of GBL, 1,4-
BD and GHB are marked with arrows.
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 356
3.5. Method performance parameters
The mean true recoveries of GLB and 1,4-BD were greater than
90% for both ante-mortem and post-mortem blood (Table 2). The
mean true recovery of GHB was approximately 80%. The trueness
of the method, as determined from 20 additional samples each of
ante-mortem and post-mortem blood, was close to 100% (Table 2).
The LODs of the quantier and qualier ions were approximately
0.5 mg/L for GHB and GBL, and 0.02 mg/L for 1,4-BD in ante-
mortem blood (Table 3). The corresponding LLOQs were less than
1 mg/mL for GHB and GBL, and less than 0.1 mg/L for 1,4-BD.
Different cut-off levels have been proposed by different investi-
gators due to an endogenous content of GHB in blood. In practice,
cut-off levels in the range of 1020 mg/L are often used, although a
cut-off level of 4 mg/L has recently been proposed for ante-mortem
blood [13]. In post-mortem blood, the LODs were estimated to be
less than 2 mg/L for GBL and less than 0.1 mg/L for 1,4-BD (Table 3).
The LOD for GHB could not be precisely determined due to the high
endogenous content of this substance. The mean endogenous level
of the tested post-mortem samples was 17 mg/L (range 134 mg/
L), which is consistent with the results reported by Elliott [14] and
within the major distribution reported by Kintz et al. [15]. The RSD
r
and RSD
R,intra-lab
values obtained in the precision study were below
10% at a concentration of 2 mg/L and below 6% at a concentration
level of 20150 mg/L for ante-mortem samples (Table 4). The
RSD
R,intra-lab
for GHB determined from post-mortem samples
containing 1420 mg/L endogenous GHB was less than 9% in
ESI(), but somewhat higher in ESI(+). That is considered
acceptable when compared to the rule of Horwitz [16]. According
to the Horwitz equation, RSD
R
values of 14% and 8% between
laboratories would be acceptable at concentration levels of 2 and
75 mg/L respectively. From the standard analyses of quality control
samples over a three-month period, RSD
R,intra-lab
values of less than
4% were obtained for all substances in ante-mortem blood spiked
to a level of 20 mg/L.
As shown earlier, this method is selective for the structural
isomers 2,3-BD, AHB and BHB that may be present in blood
samples. From a list of approximately 1000 substances, covering
therapeutic drugs having a licit medical purpose and drugs used
illicitly, that are frequently or occasionally analysed by our
institution, this method was also evaluated for other potential
interfering substances, e.g., substances differing by less than one
mass unit from the analytes. There were no substances with
potential interference found.
The calibration curves were created using weighted regression
analysis. The R
2
values obtained in the precision study from eight
independent test series and analysed within a three week period
were generally better than 0.996 (overall mean 0.9990, SD 0.0008)
for the transitions used in the quantication. The calibration curves
of 1,4-BD, GHB ESI() and GHB ESI(+) could be considered to be
linear with slopes of 0.062, 0.042 and 0.052 L/mg respectively
(independent variable: blood concentration; dependent variable:
peak area ratio of analyte and internal standard). However, for GBL
a minor but statistically signicant improvement in the standard
Table 2
Matrix effects, true recoveries and method trueness obtained from single determinations of spiked ante-mortem blood (AMB) and post-mortem blood (PMB) samples (n = 20
for each sample type). Matrix-matched standards used in the calculation of the recovery were the same samples spiked after SPE.
Substance Spiked concentration
a
mg/L Matrix effect mean (SD),
%
True recovery mean (SD), % Trueness mean (SD), %
AMB PMB AMB PMB AMB PMB
GBL 20/20 4 (4) 5 (7) 102 (8) 102 (9) 98 (3) 98 (7)
1,4-BD 20/20 4 (5) 2 (6) 96 (7) 92 (7) 103 (3) 100 (5)
GHB ESI() 20/50 79 (1) 79 (1) 84 (6) 76 (5) 99 (4) 101 (7)
GHB ESI(+) 20/50 35 (4) 24 (7) 80 (4) 80 (6) 103 (3) 106 (7)
a
Ante-mortem/post-mortem.
Time
3.80 3.60 3.40 3.20 3.00 2.80 2.60 2.40 2.20 2.00 1.80 1.60 1.40 1.20
%
0
100
3.34
Dotted line:
Reduction in response caused
by the LC gradient in ESI(-)
RT of GHB
ESI(-) m/z 103/105
ESI(+) m/z 105/107
Fig. 3. Injection of an extracted blank whole blood sample during post-column infusion (50 mL/min) of a 6 mg/L GHB standard prepared in water.
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 357
error of the residuals was obtained by inclusion of a quadratic term
in the calibration model. The coefcients of the linear and the
quadratic terms were 0.038 and 0.00002, respectively.
3.6. Stability of GBL, 1,4-BD and GHB in blood samples
The stability of GBL, 1,4-BD and GHB in the blank blood bank
samples that were freshly preserved with FO and FC mixtures,
which resulted in sample pHs of ca. 7.4 and 5.9, respectively, was
tested at storage temperatures of 20 2 8C for 7 days, and 5 2 8C
and 20 2 8C for 3 weeks. Samples were individually spiked with a
single substance at a concentration of 15 mg/L. Both 1,4-BD and GHB
were stable at 20 2 8C for at least 7 days, and at 5 2 8C and
20 2 8C for at least 3 weeks when the samples were preserved
with FO or FC mixtures. However, GBL was very unstable in non-
preserved blood and in blood freshly preserved with the FO mixture.
In less than 30 min most of the GBL was converted into GHB at
ambient temperature. The GBL was more stable in blood that had
been preserved with FC mixture, but in the investigated cases,
approximately half of the GBL was converted to GBL within 7 days
when the samples were stored at ambient temperature (Fig. 4). At
5 8C, approximately 20% and 40% of the GBL was converted after 7
days and 3 weeks, respectively. At 208C, the GBL was stable for at
least 3 weeks. Based on observations, the rate of conversion is sample
dependent. The conversion is probably due to the activity of lactonase
in blood samples, as the conversion was pH dependent and was
almost stopped by the addition of 100 mL MeOH to 200 mL blood.
Furthermore, observations indicated that the rate of conversion in
samples preserved prior to spiking declined by increasing the time lag
between preservation and spiking. When blood collection vessels
solely containing EDTA were used for sample storage, the conversion
of GBL to GHB was also inhibited, but to a lesser extent than with the
FC mixture-containing vessels. According to Fishbein and Bessman
[17], the lactonase activity is inhibited by EDTA but not efciently
inhibited by NaF. These properties should be considered during the
development of sampling strategies when the analysed concentra-
tions of GBL and GHB should resemble the chemical composition of
the sample at the sampling time. Because of the higher molecular
weight of GHB (MW = 104) compared to GBL (MW = 86), the mass
increase in GHB concentration was greater than the mass decrease in
GBL concentration.
In a study reported by LeBeay et al. [18], elevated GHB
concentrations (mean 8.7 mg/L) were observed in blood samples
stored for 636 months at 20 8C in tubes containing a mixture
Table 3
Limits of detection determined in ante-mortem blood (AMB) and post-mortem blood (PMB) samples (n = 20 for each sample type).
Substance Q3 product ion, m/z Spiked
concentration
a
, mg/L
Measured concentration, mean (SD), mg/L LOD, mg/L
AMB PMB AMB PMB
GBL 45 0.5/1.0 0.73 (0.074) 2.10 (0.46) 0.3 1.5
43 0.5/1.0 0.76 (0.11) 2.19 (0.54) 0.4 1.8
45 0/0 0.30 (0.067) 1.34 (0.60)
43 0/0 0.35 (0.072) 1.32 (0.60)
1,4-BD 73 0.05/0.1 0.045 (0.0055) 0.087 (0.012) 0.02 0.04
55 0.05/0.1 0.047 (0.0054) 0.074 (0.0082) 0.02 0.03
73 0/0 0.005 (0.0014) 0.006 (0.010)
55 0/0 0.005 (0.0023) 0.007 (0.014)
GHB ESI() 101 0.5/ 0.49 (0.18) 0.6
85 0.5/ 0.40 (0.087) 0.3
57 0.5/ 0.54 (0.14) 0.5
101 0/0 0.095 (0.10) 16 (7.9)
85 0/0 0.080 (0.058) 17 (8.7)
57 0/0 0.084 (0.14) 17 (8.5)
GHB ESI(+) 87 0.5/ 0.40 (0.20) 0.7
87 0/0 0.10 (0.10) 16 (5.0)
a
Ante-mortem/post-mortem.
Table 4
Average method precision estimated at different drug concentration levels in ante-
mortem blood (AMB) and post-mortem blood (PMB).
Substance Matrix Spiked
concentration,
mg/L
RSD
r
a
, % RSD
R,intra-lab
b
, %
GBL AMB 2 9.2 9.2
AMB 20 4.2 4.2
AMB 75 3.7 4.2
1,4-BD AMB 2 2.7 4.8
AMB 20 5.0 5.0
AMB 75 4.1 5.6
GHB ESI() AMB 2 7.1 8.6
AMB 20 3.3 3.7
AMB 150 4.4 5.3
PMB 14
c
6.3 8.8
PMB 20
c
4.8 6.8
GHB ESI(+) AMB 2 5.5 12
AMB 20 6.1 6.1
AMB 150 4.8 5.1
PMB 14
c
6.5 14
PMB 20
c
3.4 12
a
Relative standard deviation of repeatability.
b
Relative standard deviation of intra-laboratory reproducibility.
c
Endogenous content of GHB.
-1
1
3
5
7
9
11
13
15
17
19
25 20 15 10 5 0
Storage time, days
C
o
n
c
.
,

m
g
/
L
Fig. 4. Stability of 15 mg/L GLB in an ante-mortem whole blood sample that was
stored at 20 8C (*), 5 8C (~) and 20 8C (&) in Venosafe tubes containing a
uoride-citrate-EDTA additive. The conversion of GLB into GHB at 20 8C (*), 5 8C
(~) and 20 8C (&) was monitored by measuring the GHB content.
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 358
of trisodium citrate, citric acid and dextrose when compared
to tubes containing EDTA. In the present stability study, we
used a NaF-citrate-EDTA additive for preservation of the
blood samples. No formation of GHB was observed after 3
weeks of storage at 5 2 8C or after 4 months of storage at
20 2 8C.
3.7. Application of the method
The developed method has been applied to clinical and autopsy
cases with suspected intake of Fantasy. In a clinical case, a blood
concentration of GHB of 120 mg/L was found in a person suspected
of driving under the inuence of drugs. In two recent cases of
intoxication that were treated successfully at the local care units,
GHB concentrations of 97 and 104 mg/L were detected in venous
whole blood. Unfortunately we do not have information on the
time lag between intake and sampling.
In an autopsy case a 25-year-old male was found unconscious
on a lawn, and the person was declared dead after 45 min of
unsuccessful resuscitation. According to the investigation the
deceased was with three other males and a trivial struggle
occurred prior to his death. Both autopsy and toxicological
investigation were performed. Apart from minor lesions most
likely due to the struggle, the cause of death could not be
determined during the autopsy. The toxicological analysis
revealed a high GHB concentration of 235 mg/L in the peripheral
whole blood, which can result in hypotension, bradycardia,
respiratory depression, seizures and lapsing into a coma [19]. A
GBL concentration of 6 mg/L was also detected in the blood
sample. The effect of GHB may have been increased by the
ethanol concentration of 1.75 per mille. The GHB concentration
of the stomach content was 250 mg/L. The urine, which was
analysed by a separate method [20] contained 2840 mg/L of
GHB. A small bottle containing minor residual material of an
unknown origin was seized at the scene. The material was
dissolved in 2 mL of water and analysed using the current
method. The liquid contained 100,000 mg/L of GHB, 680 mg/mL
of GBL and 32 mg/L of 1,4-BD.
4. Conclusion
A rapid and selective analytical method using a simple
extraction and clean-up procedure, which avoids derivatisation
and critical steps such as sample treatment under acidic or basic
conditions and solvent evaporation, was developed. The chroma-
tography, which is based on the HILIC technique, improved control
of the retention of GHB as compared to classical reverse phase
chromatography. The ruggedness and reproducibility of the
method were strengthened through the use of stable isotope
labelled internal standards.
References
[1] S.C. Sweetman, Martindale: The Complete Drug Reference, 35th ed., Pharmaceu-
tical Press, London, 2007, p. 1618.
[2] K.S. Kalasinsky, M.M. Dixon, B.S.G.A. Schmunk, S.J.K. Kish, Blood, brain, and hair
GHB concentrations following fatal ingestion, J. Forensic Sci. 46 (2001) 728730.
[3] D. Lenz, M.A. Rothschild, L. Kro ner, Intoxications due to ingestion of g-butyr-
olactone: Organ distribution of g-hydroxybutyric acid and g-butyrolactone, Ther.
Drug Monit. 30 (2008) 755761.
[4] M.A. ElSohly, S.J. Salamone, Prevalence of drugs used in cases of alleged sexual
assault, J. Anal. Toxicol. 23 (1999) 141146.
[5] R. Rossi, M. Lancia, C. Gambelunghe, A. Oliva, N. Fucci, Identication of GHB and
morphine in hair in a case of drug-facilitated sexuel assault, Forensic Sci. Int. 186
(2009) e9e11.
[6] R. Brenneisen, M.A. ElSohly, T.P. Murphy, J. Passarelli, S. Russmann, S.J. Salamone,
D.E. Watson, Pharmacokinetics and excretion of gamma-hydroxybutyrate (GHB)
in healthy subjects, J. Anal. Toxicol. 28 (2004) 625630.
[7] H.-L. Fung, E. Haas, J. Raybon, J. Xu, S.-M. Fung, Liquid chromatographic-mass
spectrometric determination of endogenous g-hydroxybutyrate concentrations
in rat brain regions and plasma, J. Chromatogr. B 807 (2004) 287291.
[8] S. Dresen, J. Kempf, W. Weinmann, Prevalence of g-hydroxybutyrate (GHB) in
serum samples of amphetamine, metamphetamine and ecstasy impaired drivers,
Forensic Sci. Int. 173 (2007) 112116.
[9] S.S. Johansen, C.N. Windberg, Simultaneous determination of g-hydroxybutyrate
(GHB) and its analogues (GBL, 1.4-BD, GVL) in whole blood and urine by liquid
chromatography coupled to tandemmass spectrometry, J. Anal. Toxicol. 35 (2011)
814.
[10] S.R. Dahl, K.M. Olsen, D.H. Strand, Determination of gamma-hydroxybutyrate
(GHB), beta-hydroxybutyrate (BHB), pregabalin, 1,4-butane-diol (1,4BD) and
gamma-butyrolactone (GBL) in whole blood and urine samples by UPLCMSMS,
J. Chromatogr. B 885886 (2012) 37.
[11] E. Kaufmann, A. Alt, Determination of GHB in urine and serum by LC/MS using a
simple one-step derivative, Forensic Sci. Int. 168 (2007) 133137.
[12] International Organization for Standardization, ISO Standard 5725-2, Accuracy
(trueness and precision) of measurement methods and results Part 2: Basic
method for the determination of repeatability and reproducibility of a standard
measurement method. Geneva, Switzerland, 1994.
[13] H. Andresen, N. Sprys, A. Schmoldt, A. Mueller, S. Iwersen-Bergmann, Gamma-
hydroxybutyrate in urine and serum: additional data supporting current cut-off
recommendations, Forensic Sci. Int. 200 (2010) 9399.
[14] S.P. Elliott, Further evidence for the presence of GHBinpostmortembiological uid:
Implications for the interpretation of ndings, J. Anal. Toxicol. 28 (2004) 2026.
[15] P. Kintz, M. Villain, V. Cirimele, B. Ludes, GHB in postmortemtoxicology discrim-
ination between endogenous production from exposure using multiple speci-
mens, Forensic Sci. Int. 143 (2004) 177181.
[16] W. Horwitz, Evaluation of analytical methods used for regulation of foods and
drugs, Anal. Chem. 54 (1982) 67A76A.
[17] W.N. Fishbein, S.P. Bessman, Purication and properties of an enzyme in human
blood and rat liver microsomes catalyzing the formation and hydrolysis of
g-lactones, J. Biol. Chem. 241 (1966) 48354841.
[18] M.A. LeBeay, M.A. Montgomery, R.A. Jufer, M.L. Miller, Elevated GHB in citrate-
buffered blood, J. Anal. Toxicol. 24 (2000) 383384.
[19] R.C. Baselt, Disposition of Toxic Drugs and Chemicals in Man, 9th ed., Biomedical
Publications, California, 2011, pp. 744747.
[20] L.K. Srensen, J.B. Hasselstrm, Determination of g-hydroxybutyrate in urine by
an HILICMS/MS method generating multiple transition products, in: The 50th
Annual Meeting of The International Association of Forensic Toxicologists (TIAFT),
Japan, 2012, Available at http://pure.au.dk/ws/les/45766526/tiaft_lks_2012.pdf
(Aarhus University web site).
L.K. Srensen, J.B. Hasselstrm / Forensic Science International 222 (2012) 352359 359