Analysis of Gamma-Hydroxybutyrate (GHB)in Urine by Gas Chromatography-Mass Spectrometry Rachel R. McCusker, Helen Paget-Wilkes, Chris W. Chronister, and Bruce A. Goldberger* Forensic Toxicology Laboratory, Department of Pathology, Immunology, and Laboratory Medicine, University of Florida College of Medicine, P O. Box 100275, Gainesville, Florida 32610-0275 Mahmoud A. EISohly EISohly Laboratories, Inc., 5 Industrial Park Drive, Oxford, Mississippi 38655 Abstract A simple method for the direct analysis of gamma-hydroxybutyrate (GHB) from human urine is described. The method uses solid-phase extraction, liquid-liquid extraction, and silyl-derivatization, then gas chromatographic--mass spectrometric analysis using GHB-d6 as the internal standard. The method was linear from 5 to 500 rag/l, and coefficients of variation were less than 10%. Twenty-six urine specimens previously analyzed by an existing method were analyzed and yielded GHB concentrations ranging from 0 to 6100 rag/l; the results correlated between the two methods. Compared with existing methods, the method described here is superior because it is specific to GHB and can discriminate between GHB and gammabutyrolactone. Introduction Gamma-hydroxybutyrate (4-hydroxybutyrate, GHB) (Figure 1) is a naturally occuring metabolite for gamma-aminobutyric acid found in the central nervous system and peripheral tissues. GHB is metabolized by oxidative enzymes and has a short plasma half- ]ife (0.3-1.0 h). Less than 5% of an oral dose of GHB is excreted unchanged in the urine, and GHB is not detectable in urine after 12 h (1-3). GHB was first synthesized in 1961 for anesthetic purposes; however, because of its adverse and unpredictable effects, its use was discontinued. In the 1980s, GHB was sold in health food stores as an aid for weight control and muscle building. Currently, the U.S. Food and Drug Administration (FDA) classifies GIIB as an unapproved drug except for investigational use in the treatment of narcolepsy. In addition, the U.S. Drug Enforcement Administration is presently reviewing the status of GHB to determine if it should become a controlled substance. Common street names attributed to GHB include Georgia Home Boy, Grievous Bodily Harm, Liquid Ecstasy, and Scoop. During the 1990s, GHB has become a popular drug of abuse that has been commonly associated with sexual assaults ("date rape"). As a result, there has been an increase in requests for * Author to whom correspondence should be addressed. GHB analysis. Unfortunately, the analysis of GHB in biological fluids has been difficult because, in part, of the limited amount of reference literature. To date, the most common analytical techniques used to detect GHB in biological fluids are high-pressure liquid chromatography, gas chromatography (GC), and gas chromatography-mass spectrometry (GC-MS). In 1991, the FDA developed an assay for GHB in nonbiological laboratory submissions; the assay included trimethylsilyl derivatization followed by GC-flame ionization detection (5). Because GHB is a small polar molecule that is not easily isolated, the present analytical methods used to extract GHB in biological fluids typically involve the conversion of GHB to gamma-butyrolactone (GBL) by acid catalysis (3,6,7). After liquid-liquid extraction, some methods were designed to detect GBL (3), and others hydrolyzed GBL and detected derivatized GHB (8). GHB has also been isolated from rat brain tissue and analyzed using GC-MS (9,10). Because of the recent status of GHB, it is increasingly important that a laboratory technique differentiate GHB and GBL. Recently, United Chemical Technologies, Inc. (UCT, Bristol, PA) developed a solid-phase extraction method that permits the direct isolation of GHB in urine without the formation of its associated lactone, GBL (11). In this study, 26 urine specimens were tested for GHB using the UCT method. GHB O HO~,~OH GHB-diTMS CH 3 O I II CH3 H3C--Si-O~ A ~ I I v v -O Si_CH 3 CH3 I CH3 Figure I. Chemical structures of gamma-hydroxybutyrate and gammahydroxybutyrate-ditms. Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 301
.... I Journal of Analytical Toxicology, Vol. 23, September 1999 8 i- ll es ~C ::i T =i / Avelllge ~1:1,161 I~ ~211 ll~n.: GI~UNIEX.D _i [ m I / -- I '" /t ~. '= ; [.,o,,= I '",='" I, " 9 ~ "~ m i~ iiol~o+~"i~o ~i~ ifo iloi~o~o ~o zlo ~io""~io " nvz Figure 2. Full-scan mass sl~:trum of gamma-hydroxybutyrate-ditms. 000084 dnc T=~176... li~_. > 0 3. 4.03 '2/50' ' 3.100.... 3.50' ' '4 JO0.... 4.~50.... 5.'00 ' Tgt m/z 233.2 Q1 m/z 234.2 Q2 m/z 235.2 GHB ~und... - ~und... Iuncl~'~ci A i Iunclanci i 3.174 i.3.174,3.4 3., l l ~ ; 3~4 l i,i i 1 3.~4, Ti 24 m ~1me-->3.24 3.'95 i ~1me-->3.'24 3.95! Tgt m/z 239.2 Q1 m/z 240.2 Q2 m/z 241.2 GHB-D6 @bundah~ ~b~ndance... ~undance ~undance 20000 2000 20000 ' 3.20 0,, ; 0 -----~J~! 0..-~'~-- i 0. ~ Time-->3.26 3.97 me-->3.26 3.97 ~ime-->3.26 3.'97 I LTime-->3.'26 3.97 J 5.lo~ 400000 i JI 3.76 4.38 :I 200000 3.21 I ' ~ O, '..., JI ' ':" 5 "~ime- - > 2 ~50 ~ O 3.:50 4.00 4.50. I rl Tgt m/z 233.2 Q1 m/z 234.2 Q2 m/z 235.2 GHB " ~:~'a~.~e 3 6 'and... 36 1 ~8~e 3-6~ i00000 20000 10000 _. ~... 3._ ~... ~_~... B I ~... 3.'24'3'.,9~'i~, _ :.... ~. ~ 9. ~.... _~. Tgt m/z 239.2 Q1 m/z 240.2 Q2 m/z 241.2 GHB-D6 d:)unctahce - ~bundanc~ ~bundance ~1~b-undance --- 200003F41 /] ~000~13p4~ 3188!/ 2000~/3V4! / 20000!3{4 Figure 3. A, Chromatogram of a 25 mg/l gamma-hydroxybutyrate standard prepared in phosphate buffer. B, An authentic urine specimen containing gamma-hydroxybutyrate at a concentration of 29 mg/i.. The retention time of GHB was approximately 3.7 min. Materials and Methods Specimens Urine specimens were obtained from ElSohly Laboratories (reference laboratory) under a public service program that provides analysis of urine specimens collected by law enforcement personnel, at rape crisis centers, and in emergency rooms where sexual assault victims were treated. The specimens were submitted to the reference laboratory for drugs-of-abuse testing that included GC-MS analysis of GHB. The reference laboratory procedure for GHB uses acid catalysis followed by chloroform extraction and GC-MS analysis of GBL. The internal standard was gamma-valerolactone. Chemicals Fisher Scientific (Fairlawn, NJ) certified ACSgrade methanol (CH3OH), hexane, ammonium hydroxide (NH4OH), potassium hydroxide, and potassium phosphate monobasic were used. N,N- Dimethylformamide (DMF) was purchased from Sigma Chemical Co. (St. Louis, MO). GHB, GBL, and GHB-d6 were purchased from Radian International (Austin, TX). Buffer (ph 6.0 0.1) was prepared from 0.1M potassium phosphate and 5M potassium hydroxide. Bis(trimethyl-silyl)- trifluoroacetamide (BSTFA) with 1% trimethylchlorosilane (TMCS) was purchased from Pierce Chemical Co. (Rockford, IL). Deionized water was prepared using a Millipore purification system. The extration columns were CLEAN SCREEN ZSGHB020 (United Chemical Technologies, Inc.). Standard and control preparation All drug standards and the internal standard solution (GHB-d6) were prepared in methanol. The standard curve, consisting of a minimum of five concentrations, was prepared in blank phosphate buffer ranging from 5 to 500 mg/l. The batches included two control samples prepared in urine and two control samples prepared in phosphate buffer at concentrations of 25 and 100 mg/l. The batches also included two negative control samples consisting of freshly voided urine and phosphate buffer. Extraction In culture tubes, 200 IJL of urine, 100 pl of internal standard (1.0 IJg, GHB-d6), and 100 pl of phosphate buffer were combined. The mixtures were vortex mixed for approximately 10 s. The CLEAN SCREEN GHB extraction columns were conditioned by adding 3 ml of methanol, 3 ml deionized water, and 0.5 ml phosphate buffer (ph 6.0) and aspirating after each addition without 302
drying out the sorbent. The urine specimens were decanted into the columns and aspirated into culture tubes. The original culture tubes were washed with I ml of CH3OI-I/NH4OH (99:1), and the wash solutions were decanted into the columns. The columns were aspirated into the culture tubes that contained the first aspirate. The culture tubes were removed, and the extracts were dried under a stream of nitrogen at 60~ in a Zymark TurboVap LV Evaporator (Hopkinton, MA). DMF (200 IJL) and I ml hexane saturated with DMF were added to the dried residue. The samples were mixed by inversion for 5 min, then centrifuged at 2000 rpm for 5 rain. The top hexane layer was discarded, and the lower DMF layer was transferred to a clean culture tube. The samples were evaporated to dryness at 50~ under a stream of nitrogen. Ethyl acetate and BSTFA with 1% TMCS (100 pl each) were added to the residue. The extracts were vortex mixed, heated for 5 min at 60~ and transferred to autosampler vials for analysis by GC-MS. Instrumentation GC-MS analyses were performed using a Hewlett-Packard 5890 series II plus GC and an HP 6890 series automatic injector interfaced to an HP 5972 series mass selective detector (Little Falls, DE). A cross-linked methyl siloxane capillary column was used (Hewlett-Packard HP-1MS, 12 m x 0.25 ram, 0.25-1Jm film thickness) with ultra-high-purity-grade helium as the carrier gas (constant flow rate, 1.60 ml/min). Injections (1 pl) were made in the splitless mode. The GC temperature settings were as follows: injector port, 260~ transfer line 290~ initial column temperature, 65~ hold time, 0.50 min; temperature ramp, 15~ to 105~ then 25~ to 300~ The total run time was 10.97 rain. The mass selective detector was run in the selected ion montoring mode (SIM). The following ions were monitored at a dwell time of 20 ms (underlined ions were used for quantitation): GHB-diTMS: m/z 233, 234, and 235 and GHB-do-diTMS: m/z ~ 240, and 241. The electron multiplier was operated at 100 ev above the autotune value. GHB was identified based on comparison of retention time and ion ratios with the corresponding values of standards analyzed in the same batch. of GHB in urine was based upon internal standardization using GHB*d6. The full scan mass spectrum of GHB (Figure 2) identifies the following ions in order of abundance: m/z 147, 233, 117, 158, 148, 149, 143, 133, 204, 234, and 235. Because urea also forms a ditms derivative, it has ions similar to the ditms deriva- Table I. Results of Gamma-Hydroxybutyrate (GHB) Analysis Performed by the University of Florida Forensic Toxicology Laboratory and the Reference Laboratory Specimen University of Florida Reference Laboratory number GHB results (mg/l) GHB results (mg/l) 1 ND* ND 2 ND ND 3 ND ND 4 ND ND 5 9.76 3.71 6 t ND 4.49 7 9.71 6.13 8 6100 5142 9 547 665 10 6.04 3.28 11 188 251 12 1590 2205 13 435 556 14 920 1110 15 3310 27O6 16 2.28 3.90 17 29.33 25.3 18 14.97 9.81 19 5.59 3.33 20 650 605 21 14.98 7.67 22 6.13 3.27 23 10.46 3.O7 24 10.06 18.7 25 59.96 47.8 26 11.82 7.23 * NO = none detected. Specimen 6 contains more than 95% GBL. ~ 7500 g m -r (:1 50O0 0 0 m,.i 2500 8 c y = 0.856x + 69.441 R== 0.9748 Results and Discussion Urine specimens were extracted by the UCT method, derivatized, and analyzed by GC-MS for ditms derivative of GHB (Figure 1). Quantitation m 0 2500 5000 75'00 UF Laboratory- GHB (mg/l) Figure 4. Correlation plot of gamma-hydroxybutyrate determinations performed by the University of Florida Forensic Toxicology Laboratory and the reference laboratory. 303
tive of GHB, notably m/z 147, 148, and 149 (11). Therefore, the less abundant ions, m/z 233, 234, and 235, were used for the SIM analysis. Figure 3 illustrates characteristic SIM spectra obtained using this method. Figure 3A is a 25 mg/l GHB extracted standard prepared in phosphate buffer, and Figure 3B is an authentic urine specimen containing GHB at a concentration of 29 mg/l. The retention time of GHB was approximately 3.7 rain. Analysis of GHB standards revealed a range of linearity of 5-500 mg/l in most batches. Chromatographic and/or detector saturation in standards and specimens was occasionally observed at higher concentrations. When overload occurred, the derivatized extracts were diluted with an appropriate volume of ethyl acetate and re-injected. All control values were within 5% of the target concentration, and between-run precision data of the assay using control samples revealed coefficients of variation of less than 10% at both concentrations in phosphate buffer and urine. The mean concentrations (standard deviation) of GHB in phosphate buffer were 25.7 (1.49) and 99.0 (5.04) mg/ml. The mean concentrations (standard deviation) of GHB in blank urine were 26.0 (1.39) and 101.1 (9.35) mg/ml. Limit of sensitivity and recovery studies were not conducted because detection of GHB in the concentration range described is easily accomplished by GC-MS SIM analysis. In addition, concentrations below the lowest standard (5 rag/l) may be attributed to endogenous GHB. The baseline GHB concentrations in urine in the current study were below 5 mg/l. Recommendation for a cutoff value that indicates GHB ingestion rather than endogenous presence is not possible at this point, and further study is needed. Table I and Figure 4 compare the GHB results obtained by the University of Florida Forensic Toxicology Laboratory with those obtained by the reference laboratory. The results demonstrate good agreement between the methods, with 14% bias, positive intercept of 69 rag/l, and correlation coefficient of 0.9748. The deviation could be attributed to the reference laboratory procedure because specimens with GHB concentrations outside of the linear range of the assay were not diluted to obtain a concentration within the range of linearity (5-200 rag/l). In order to investigate the in vitro conversion of GBL to GHB using the University of Florida method, four samples were spiked with varying concentrations of GBL (25-500 rag/l) and subjected to GHB analysis. GHB was detected at concentrations of less than 0.1% of the fortified GBL concentrations. In addition, GHB was not detected in specimen 6 of the comparison study by the University of Florida method, but when tested by the reference laboratory, GHB was present at a concentration of 4.4 mg/l. Upon further testing without acid catalysis by the reference laboratory, the specimen was found to contain more than 95% GBL. This result further supports the specificity of the University of Florida procedure, which only identifies the presence of GHB. We recently applied the University of Florida method to the analysis of GHB in postmortem urine specimens. The range of GHB concentrations in these specimens was 5 mg/l to greater than 500 mg/l. The results of these analyses were used by the Medical Examiner to corroborate previously reported histories of GHB use. In their study of GHB concentrations in ante- and postmortem blood and urine, Fieler et al. (12) detected GHB in postmortem blood obtained from cases not known to be GHB related, whereas GHB was absent in blood from living persons. However, GHB was not detected (< 1 rag/l) in the postmortem urine specimens. This finding suggests that GHB is a product of postmortem decomposition; thus, it was concluded that urinary GHB determinations were more reliable than blood GHB analyses. Conclusions This method is useful for the simple and rapid identification and quantitation of GHB in urine. The results of the current study were compared with data from a laboratory that routinely analyzes GHB in urine, and the results were similar. The major advantages of this method include (1) small sample volume, (2) rapid extraction using technologies readily available to toxicology laboratories, (3) detection of higher molecular weight ions compared with previous methods, and (4) high specificity and sensitivity. Unfortunately, this method is not readily applicable to the analysis of GHB in blood. Finally, the analysis of GBL by this method should be straightforward, and further investigation may lead to an assay that permits the simultaneous analysis of GHB and GBL. Acknowledgments The authors would like to thank United Chemical Technologies for providing the GHB extraction columns for the study. References 1. R.C. Baselt. Gamma-hydroxybutyrate. In Disposition of Toxic Drugs and Chemicals in Man, 5 th ed. Chemical Toxicology Institute, Foster City, CA, 2000. 2. J.E. Dyer. 7-Hydroxybutyrate: a health-food product producing coma and seizure-like activity. Am. J. Emerg. Med. 9:321-324 (1991). 3. S.D. Ferrara, L. Tedeschi, G. Frison, F. Castagna, L. Gallimberti, R. Giorgetti, G.L. Gessa, and P. Palatini. Therapeutic gammahydroxybutyric acid monitoring in plasma and urine by gas chromatography-mass spectrometry. J. Pharm. Biomed. Anal. 11: 483-487 (1993). 4. J.D. Ropero-Miller and B.A. Goldberger. Recreational drugs: current trends in the 90s. Clin. Lab. Med. 11]: 727-746 (1998). 5. R.H. Johnson and J.L. Bussey. Laboratory information bulletin no. 3532: Assay procedure for the sodium salt of gamma hydroxybutyric acid. United States Food and Drug Administration, 1991. 6. J.D. Doherty, O.C. Snead, and R.H. Roth. A sensitive method for quantitation of 7-hydroxybutyric acid and 7-butyrolactone in brain by electron capture gas chromatography. Anal. Biochem. 69" 268-277 (1975). 7. T.B. Vree, E. van der Kleijn, and H.J. Knop. Rapid determination of 4-hydroxybutyric acid (Gamma OH) and 2-propyl pentanonate (Depakine) in human plasma by means of gas-liquid chromatography. J. Chromatogr. 121: 150-152 (1976). 304
8. K.M. Gibson, S. Aramaki, L. Sweetman, W.L. Nyhan, D.C. DeVivo, A.K. Hodson, and C. Jakobs. Stable isotope dilution analysis of 4-hydroxybutyric acid: an accurate method for quantification in physiological fluids and the prenatal diagnosis of 4-hydroxybutyric aciduria. Biomed. Environ. Mass Spectrom. 19:89-93 (1990). 9. M. Eli and F. Cattabeni. Endogenous y-hydroxybutyrate in rat brain areas: postmortem changes and effects of drugs interfering with y-aminobutyric acid metabolism. J. Neurochem. 41" 524-530 (1983). 10. J.D. Ehrhardt, Ph. Vayer, and M. Maitre. A rapid and sensitive method for the determination of y-hydroxybutyric acid and trans- l- hydroxycrotonic acid in rat brain tissue by gas chromatography/ mass spectometry with negative ion detection. Biomed. Environ. Mass Spectom. 15:521-524 (1988). 11. C.J. Kitchen and T. August. Product application: a solid phase extraction method for the determination of gamma-hydroxybutyrate (GHB) in urine without conversion to gamma-butyrolactone (GBL). United Chemical Technologies, Inc., Bristol, PA, 1999. 12. E.L. Fieler, D.E. Coleman, and R.C. Baselt. t-hydroxybutyrate concentrations in pre- and postmortem blood and urine. Clin. Chem. 44:692 (1998). Manuscript received June 7, 1999; revision received July 26, 1999. 305