Identification of Testosterone and Testosterone Esters in Human Hair P. Kintz*, V. Cirimele, T. Jeanneau, and B. Ludes Institut de M~decine L~gale, 11 rue Humann, 67 Strasbourg, France Abstract In t 974, steroids were added to the list of doping agents banned by the International Olympic Committee because of their effects on the performance of the athletes. Testosterone and its esters promote the development of secondary male sexual characteristics and accelerate muscle growth. The mandatory test to detect testosterone abuse is to measure the ratio of testosterone to epitestosterone in the urine. However, because athletes can adjust their dosage to stay within the range permitted, there is a risk of test evasion. Therefore, we developed two original procedures to determine testosterone and its esters in human hair. First, testosterone was investigated in hair obtained from 26 control subjects. After decontamination with dichloromethane, 1 mg of hair was incubated in 1M NaOH in the presence of 1 ng of testosterone-d3. After neutralization, the extract was purified using solid-phase extraction with Isolute Ct8 columns followed by liquid-liquid extraction with pentane. After silylation, testosterone was analyzed by gas chromatography-mass spectrometry. Concentrations were in the range 1.2 to 11.4 pg/mg with a mean value of 3.8 pg/mg. To distinguish exogenous abuse from endogenous levels, the incorporation of testosterone esters into hair was investigated. Preparation involved methanolic incubation to avoid the cleavage of the esters. In a panel of eight esters, it was possible to identify testosterone propionate, testosterone enanthate, and testosterone decanoate in the hair of two bodybuuders and one weight lifter. This new technology may find useful applications in anabolic abuse control. Introduction In 1974, steroids were added to the list of doping agents banned by the International Olympic Committee because of their effects on the performance of the athletes and consequently on the results of competitions and, above all, because of their adverse effects on the athletes' health. Anabolic steroids are synthetic derivatives of testosterone, the male sex hormone secreted by the testes and the adrenal glands, and include many closely related ring structures and their esters. These compounds have two different effects on the body: they "Author to whom correspondence should be addressed promote the development of secondary male sexual characteristics (androgenic effects) and accelerate muscle growth (anabolic effects). Athletes use steroids because it has been claimed that they increase lean body mass, increase strength, increase aggressiveness, and lead to a shorter recovery time between workouts (1,2). Long-term effects such as severe cardiovascular side-effects or liver diseases and fatalities have been reported in young steroid abusers (3--6). Apart from competitive athletes, the largest group of users are recreational bodybuilders who take anabolic steroids for cosmetic reasons. Because of the limited availability of approved products, a black market for anabolic steroids has developed with an associated increase in trafficking. In a recent case, this laboratory was involved in a medicolegal case in which 25 tablets and 251 ampules of various steroids were discovered in the luggage of two tourists (7). The testing standard for anabolic steroids for doping control is gas chromatography coupled to mass spectrometry (GC-MS) conducted on a urine sample performed in an accredited laboratory. In such cases, exogenous abuse has to be distinguished from endogenous levels of identical substances such as testosterone. Scientific approaches for this differentiation (i.e., quantitation of testosterone/epitestosterone concentrations ratio or carbon isotope ratio discrimination) are always controversial because of common biological deviations. However, problems with testosterone abuse continue to proliferate. Athletes inject themselves with testosterone, then take their urine sample to the lab and have the testosterone/epitestosterone ratio measured. In this fashion, they can adjust their dosage to stay within the range permitted by the International Olympic Committtee (IOC). Ideally, unequivocal confirmation of testosterone administration could be achieved by direct detection in the body of the actual drug applied. Testosterone is commonly administered as different 17 [3-hydroxy esters. Nearly complete in vivo hydrolysis of these testosterone esters by esterases gives rise to the active testosterone. Nevertheless, minute amounts of unchanged esters remain in the body, and it was proposed in 1995 to test for testosterone esters in plasma (8). Liquid chromatography-tandem mass spectrometry was recently proposed for testing for testosterone esters in plasma (9). Despite some promising results, the possi- 352 Reproduction (photo~opying) of editorial content of this journal is prohibited without publisher's permission.
bility of using blood for the confirmation of testosterone abuse has not been fully explored. Hair specimens have been used for 2 years in toxicology to document drug exposure in various forensic, occupational, and clinical situations (1). The major practical advantage of hair testing compared to urine testing for drugs is its larger surveillance window, weeks to months, depending on the length of the hair shaft, versus 2-4 days for most xenobiotics, with the exception of anabolic steroids in an ester preparation. For practical purposes, the two tests complement each other. Urinalysis provides shortterm information on an individual's drug use, whereas long-term histories are accessible through hair analysis. Analysis of urine specimens cannot distinguish between chronic use or single exposure, but hair analysis can offer this distinction. Its greatest use, however, may be in identifying users who abstain before the test or try to beat the test by diluting their urine samples, as this will not alter the concentration in hair. Urine does not indicate the frequency of drug intake in subjects who might deliberately abstain for several days before screening. Doping Table I. Selected Ion (m/z) and Retention Times for Each Analyte Analytes Retention time (min) Ions (m/z) Testosterone 9.8 417-421 - 432 Testosterone-d 3 9.6 42-424 - 435. Testosterone acetate 9.36 28-387 - 42 Testosterone propionate 9.8 28-41 - 416 Testosterone isocaproate 11.35 28-443 - 458 Testosterone enanthate 12.45 28-457 - 472 Testosterone benzoate 14.61 28-449 - 464 Testosterone cypionate 15.1 28-469 - 484 Testosterone decanoate 16.26 28-499 - 514 Testosterone phenylpropionate 17.17 28-477 - 492 1- ~ 6-8, 7 ~ r2=.9962 "i / 3 during training and abstinence during the competition should therefore be detected, as athletes cannot evade the test. However, there is a need for procedures sensitive enough to document the negative results. The international literature is very deficient in papers dealing with the identification of anabolic steroids in human hair. Most of the reports focused on animals. Stanozolol in rats (12), nandrolone in guinea pigs (13), and testosterone in cattle (14) were reported. Only Gleixner and Meyer (15) reported a testosterone mean concentration (3.8 +.5 pg/mg) in 13 men. However, the growing interest of scientists in the detection of anabolic drugs was observed in late 1998 (16,17). The intent of this work was to establish a quantitative range of the physiological concentrations of testosterone in human hair and to document testosterone abuse through the qualitative identification of testosterone esters in human hair. Experimental Human hair samples Full-length hair samples to be tested for testosterone were taken at the surface of the skin from the vertex of 26 male volunteers. Subjects were aged from 16 to 63 years. None were professional athletes, and all denied the use of doping agents, regardless of the pharmacological class. In all cases, the urine testosterone/ epitestosterone ratio was lower than 2.. Hair specimens were also collected from two bodybuilders and one weight lifter suspected of anabolic steroid trafficking by the police. All hair samples were stored in tubes at room temperature. Chemicals Dichloromethane, pentane, and methanol were HPLC grade (Merck, Darmstadt, Germany). All other chemicals were of analytical grade and provided by Merck. Testosterone, testosterone-d3, testosterone acetate, testosterone propionate, testosterone isocaproate, testosterone enanthate, testosterone benzoate, testosterone cypionate, testosterone decanoate, and testosterone phenylpropionate were purchased from Sigma (Saint-Quentin Fallavier, France). N- Methyl-N-trimethylsilyltrifluoroacetamide (MSTFA), 2-mercaptoethanol, and ammonium iodide (NH4I) were purchased from Fluka (Saint- Quentin Fallavier, France). Isolute C18 columns were purchased from Touzart et Matignon (Courtaboeuf, France). 2 1 1 2 3 4 5 6 7 8 9 1 Testosterone concentration (pg/mg) Figure t. Calibration curve obtained by analyzing a female hair specimen spiked with testosterone at concentrations between 1 and 1 pg/mg. Sample extraction for testosterone Before analysis, samples were decontaminated twice in 5 ml of methylene chloride, for 2 min, at room temperature. A 4-cm segment from root was used for the analysis. One hundred milligrams of hair was incubated for 15 min at 95~ in i ml 1M NaOH in the presence of 1 ng of testosterone-d3 (prepared in methanol) used as internal standard. After cooling, 353
the homogenate was neutralized with I ml 1M HCI, and 2 ml of.2m phosphate buffer (ph 7.) were added. The Isolute C18 columns were conditioned with 3 ml of methanol, followed by 2 ml of deionized water. After sample addition, the columns were washed with 2 x 1 ml of deionized water. After column drying, the analytes were eluted by the addition of 2 x.75 ml of methanol. The eluant was evaporated to dwness under nitrogen flow at 4~ and the residue was reconstitued in 1 ml of.2m phosphate buffer (ph 7.). A further purification step was achieved by addition of 1 mg of Na~.CO3PNaHCO 3 (1:1, way) Table II. Testosterone Concentrations in 26 Hair Specimens Subject Age Testosterone (pg/mg) 1 63 4,1 2 55 6.4 3 48 2, 4 4 3.2 5 38 1.1 6 38 6.6 7 37 1.3 8 36 2.7 9 35 4.1 1 35 2.2 11 35 9. 12 33 11.4 13 33 2.9 14 33 1.5 15 3 3.4 16 29 2.5 17 26 2.1 18 26 3.8 19 25 3. 2 24 3.4 21 23 3.5 22 2 1.2 23 2 2.3 24 2 3.3 25 2 1.5 26 16 2.3 18!' 8: i'11," 19. 8 Ion 432. (431.7 to 432.7): ISOIO3.D Ion 435. {434.7 to 435.7): 1513.D testosterone: 3.8 pg/mg and 2 ml of pentane. After agitation and centrifugation, the organic phase was removed and evaporated to dryness. The residue was derivatized by 5 IJL MSTFA/NH4I/2-mercaptoethanol (1:2:5, v/v/v) for 2 min at 6~ Sample extraction for testosterone esters After decontamination, 1 mg of hair was cut into small pieces of about 1 mm and incubated overnight at 5~ in 5 ml methanol in presence of 2 ng testosterone-d3, used to determine the retention index. After evaporation of the methanol and reconstitution in phosphate buffer, the analytes were submitted to the Isolute C18 columns, as previously described. After evaporation of the eluant,.5 ml of Soerensen phosphate buffer (ph 7.6), 5 IJL of3m KOH, and 4 ml hexane/ethylacetate (7:3, v/v) were added to the residue. After agitation, centrifugation, and evaporation of the organic phase, the drugs were derivatizated as previously described. GC--MS procedure A 4-1JL aliquot of the derivatized extract was injected into the column of a Hewlett Packard (Palo Alto, CA) 689 series GC via a Hewlett Packard (7673) autosampler. The flow of carrier gas (helium, purity grade N 55) through the column (HP5-MS capillary column, 5% phenyl-95% methylsiloxane, 3 m x.25-ram i.d.,.25-ram film thickness) was 1. mljmin. The injector temperature was 27~ and splitless injection was employed with a split valve off-time of 1. rain using the pulsed mode. The column oven temperature was programmed to rise from an initial temperature of 1~ (held I min) to 295~ at 3~ and hold for the final 6 rain. The detector was a Hewlett Packard 5973 operated in the electron impact mode. The electron multipler voltage was set at 6 V above the EI-tune voltage. Testosterone method validation A standard calibration curve was obtained by adding.1 (1 pg/mg),.5 (5 pg/mg), 1. (1 pg/mg), 2. (2 pg/mg), 5. (5 pg/mg), and 1. ng (1 pg/mg) of testosterone to 1 mg of hair from a female member of the laboratory personnel. This hair was tested previously and found to be under the limit of detection for testosterone. Recovery and within-run precision for testosterone were determined using a strand of hair that was previously pulverized in a ball mill for homogenicity. The detection limit was evaluated with decreasing concentrations of testosterone spiked in hair from a female until a response equivalent to three times the background noise was observed. 6 4 i 2 i Results and Discussion Tim 9.3 9.35 9,4 9.T~ 9.5 9.55 Figure 2. Selected ion chromatogram for testosterone (m/z 432) and testosterone-d3 (m/z 435). Concentration was 3.8 pg/mg. Table I shows the ions monitored for each analyre and for the deuterated internal standard and the retention times. Responses for testosterone were linear in the 354
Table III. Detection Limits of the Eight Testosterone Esters 9 8 7 6 5 4 3 2 Esters Limit of detection (pg/mg) Acetate 2 Propionate 2 Isocaproate 1 Enanthate 1 Benzoate 2 Cypionate 2 Decanoate 5 Phenylpropionate 2 9.3Ion 416. (415.7 CO 416.7): 2411.D testosterone propionate ~ '9.'o' io!oo '~o:2d io14o... ~ - ' ~ ~me Figure 3. Selected ion chromatogram for testosterone propionate obtained after extraction of the hair of a bodybuilder. testosterone enanthate Tlme Figure 4. Selected ion chromatogram for testosterone enanthate obtained after extraction of the hair of a bodybuilder. 2. lsooo. zoooo~ Ion 514. (513.7 CO $14.7): 1212.D 16 26 testosterone decanoate Tlme Figure 5. Selected ion chromatogram for testosterone deconoate obtained after extraction of the hair of a weight lifter. range of 1 to 1 pg/mg with a correlation coefficient of.9981 (Figure 1). The within-run precision was 6.6%, as determined by analyzing eight pulverized replicates of 1 mg of hair obtained from the same subject and found to contain testosterone at 3.8 pg/mg. The same specimen was used to determine the extraction recovery (n = 3), which was evaluated to be 92.2%. The limit of detection, using a 1-rag hair sample, was.5 pg/mg, and the limit of quantitation was set at 1. pg/mg. Extensive chromatographic procedures (two purification steps by solid-phase and liquid-liquid extraction combined with injection of 4 I~L through the column in pulsed mode) were analytical prerequisites for successful identification of testosterone in hair because of the low target concentrations. Under the chromatographic conditions used, there was no interference with the analytes by any extractable endogenous materials present in hair. According to Segura et al. (18), the derivatizing agents were found suitable and stable for at least one week. Figure 2 shows a typical chromatogram using ions rn/z 432 and 435. The testosterone concentration was 3.8 pg/mg. Twenty-six specimens obtained from male volunteers were analyzed. Individual results are pre- sented in Table II. All specimens tested positive for testosterone at concentrations ranging from 1.2 to 11.4 pg/mg with a mean value of 3.8 pg/mg. The latter concentration corresponds to the mean value that was reported by Gleixner and Meyer (13). The concentrations detected were very low. In hair, testosterone concentrations are in the range of some picograms per milligram, whereas cocaine, amphetamines, or opiates are generally found in the range of several nanograms per milligram (9). In the hair of two bodybuilders, this laboratory identified testosterone at 46 and 71 pg/mg (7). Unlike testosterone in urine, the interpretation of concentration findings can be difficult and critical. The range between physiological concentrations and those found in heavy abusers like bodybuilders seems to be rather small. Therefore, in complement of testosterone determination, the identification of unique testosterone esters was investigated. This characterization should enable unambiguous identification of testosterone abuse, because the esters are certainly exogenous substances contrary to the endogenous testosterone. Chromatographic parameters of eight testosterone esters were established based on retention times and monitored ions (Table I). Esters involved were testosterone acetate, testosterone propionate, testosterone isocaproate, testosterone enanthate, testosterone benzoate, testosterone cypionate, testosterone decanoate, and testosterone phenylpropionate. All of the trimethylsilyl derivatives show a prominent molecular ion which is the base peak. These molecular ions appear in an area of the 355
mass spectrum generally characterized by a low biological background. The M-15 ion (corresponding to the loss of CH3) is always present and adds selectivity to the molecular ion. The separation of the eight compounds appears to be satisfactory with no risk of coeluting peak. The presence of testosterone esters was evaluated in a control group of males (n = 19) and females (n = 11). In all cases, no corresponding signal was observed, and the response was always under the limit of detection (LOD). Subjects under therapeutic medication of testosterone esters were not studied because of the lack of specimen. Figures 3-5 are typical chromatograms that demonstrate the power of hair analysis in documenting testosterone abuse. In three medicolegal cases, it was possible to identify testosterone propionate, testosterone enanthate, and testosterone decanoate, clearly indicating exposure to these esters. In this study, we did not attempt to quantitate these esters because their identification is enough to document exogenous testosterone administration. The LODs of the esters are indicated in Table III. LOD were established with decreasing concentrations of testosterone esters spiked in hair, until a response equivalent to three times the background noise was observed and found typically in the range of 2-1 pg/mg. Further investigations are necessary to determine if these concentrations are low enough in all cases of testosterone abuse. In conclusion, it is difficult to establish an absolute and definitive analytical method to distinguish exogenous testosterone from the natural physiological hormone. Some interesting steps have been made in this direction by the detection of testosterone esters in plasma or by isotope ratio measurement. The determination of testosterone esters in hair by mass spectrometry should allow a definitive unambiguous confirmation of the administration of exogenous testosterone. References 1. S. Plymate and K. Friedl. Anabolic steroids and muscle strength. Ann. Intern. Med. 116:27 (1992). 2. S.B. Karch, Ed. The Pathology of Drug Abuse. CRC Press, Boca Raton, FL, 1996, pp 49~-29. 3. R. McNutt, G. Ferenchick, R Kirlin, and N. Hamlin. Acute myocardial infraction in a 22-year-old world class weight lifter using anabolic steroids. Am. J. Cardiol. 62:164 (1988). 4. G. Ferenchick and S. Adehan. Myocardial infarction associated with anabolic steroids use in a previously healthy 37-year-old weight lifter. Am. Heart J. 124:57-58 (1992). 5. B. Madea, W. Grellner, F. Musshoff, and R. Dettmeyer. Medico-legal aspects of doping. J. Clin. Forensic Med. 5:1-7 (1998). 6. R. Hausmann, S. Hammer, and P. Betz. Performance enhancing drugs (doping agents) and sudden death--a case report and review of the literature. Int. J. Leg. Med. 111:261-264 (1998). 7. P. Kintz, V. Cirimele, H. Sachs, T. Jeanneau, and B. Ludes. Testing for anabolic steroids in hair from 2 bodybuilders. Forensic Sci. Int. 11: 29-216 (1999). 8. X. de la Torre, J. Segura, A. Polettini, and M. Montagna. Detection of testosterone esters in human plasma. J. Mass Spectrom. 3: 1393-144 (1995). 9. C.H.L. Shackleton, H. Chung, X. de la Torte, and J. Segura. Electrospray mass spectrometry of testosterone esters: potential for use in doping control. Steroids 62:523-529 (1997). 1. P. Kintz, Ed. Drug Testing in Hair. CRC Press, Boca Raton, FL, 1996. 11. P. Kintz. Hair testing and doping control in sport. Toxicol. Lett. 12: 19-113 (1998). 12. K.M. H~ld, D.E. Wilkins, D.J. Crouch, D.E. Rollins, and R.A. Maes. Detection of stanozolol in hair by negative, chemical ionization mass spectrometry. J. Anal. Toxicol. 2:345-349 (1996). 13. J. Segura. Possibilities of ELISA methodologies for hair analysis. Proceedings of the 1995 International Conference and Workshop on Hair Analysis in Forensic Toxicology, Abu Dhabi, United Arab Emirates, 1995, pp 351-369. 14. A. Gleixner and H.H.D. Meyer. Detection of estradiol and testosterone in hair of cattle by HPLC/EIA. Fresenius J. Anal. Chem. 357: 1198-121 (1997). 15. A. Gleixner and H.H.D. Meyer. Methods to detect ananabolics in hair: use for food hygiene and doping control. Int. Lab. July: 2-23 (1998). 16. D. Thieme, ]. Grosse, H. Sachs, and R.K. Muel]er. Detection of several anabolic steroids of abuse in human hair. Proceedings of the 16th Cologne Workshop ~n Dope Analysis, Koln, Germany, 1999, pp 9-29. 17. D. Thieme, J. Grnsse, R.K. Mueller, and H. Sachs. Proof of steroids abuse by detection of t(,stoster()ne esters in hair. Pr~;ceedin~4s of the SOFT-TIAFT 1998 meeting4, AJbuquerque, NM, 1998, in press. 18. J. Segura, R. W, ntura, and C. Jurado. Derivatization procedures for gas chromatography mass spectrometric determination of xenobiotics in biological samples, with special attention to drugs of abuse and doping agents. J. Chromato~r. B 713:61-9 (1998). Manuscript received October 19, 1998; revision received [anuary 12, 1999. 356