Application of testosterone to epitestosterone ratio to horse urine - a complimentary approach to detect the administrations of testosterone and its pro-drugs in Thoroughbred geldings

Marjaana Viljantoa,e, James Scartha, Pamela Hincksa, Lynn Hillyerb, Adam Cawleyc, Craig Suannc, Glenys Nobled, Christopher J. Walkere, Andrew T. Kicmane and Mark C. Parkine

a LGC, Newmarket Road, Fordham, Cambridgeshire, CB7 5WW, UK b British Horseracing Authority, 75 High Holborn, London, WC1V 6LS, UKc Racing NSW, Level 11, 51 Druitt Street, Sydney, NSW, Australia, 2000d School of Animal and Veterinary Sciences, Charles Sturt University, Locked Bag 588, Wagga Wagga, NSW, Australia, 2678e Drug Control Centre, Analytical and Environmental Sciences Research Divisions, King’s College London, 150 Stamford Street, London, SE1 9NH, UK

ABSTRACT

Detection of testosterone and/or its pro-drugs in the gelding is currently regulated by the application of an international threshold for urinary testosterone of 20 ng/mL. The use of steroid ratios may provide a useful supplementary approach to aid in differentiating between the administration of these steroids and unusual physiological conditions that may result in atypically high testosterone concentrations. In the current study, an ultra high performance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS) method was developed to quantify testosterone (T) and epitestosterone (E). The method was used to analyse 200 post-race urine samples from geldings in order to generate the ratios for the reference population. Following statistical analysis of the data, an upper limit of 5 for T:E ratio in geldings is proposed. Samples collected from 15 geldings with atypical urinary testosterone concentrations (> 15 ng/mL) but otherwise normal steroid profile, had T:E ratios within those observed for the reference population. The applicability of an upper T:E ratio to detect an administration was demonstrated by the analysis of a selection of incurred samples from testosterone propionate, dehydroepiandrosterone (DHEA) and a mixture of DHEA and pregnenolone (Equi-Bolic®) administrations. These produced testosterone concentrations above the threshold of 20 ng/mL, but also T:E ratios above the proposed limit of 5. In conclusion, consideration of the T:E ratio appears to be a valuable complimentary aid to evaluate whether an atypical testosterone concentration could be caused by a natural biological outlier as opposed to the administration of these steroids.

Keywords

Endogenous steroids, testosterone:epitestosterone ratio, gelding, urine, threshold.

1. INTRODUCTION

Detection of the administration of anabolic-androgenic steroids, which are identical in structure to those produced endogenously in the horse, can be challenging. Currently, the

123456

789

1011121314

15

16171819202122232425262728293031323334

35

36

37

3839

detection of doping with testosterone and its pro-drugs is regulated through the use of international thresholds in urine and plasma [1] and by the detection of synthetic steroid esters in plasma and hair [2]–[5].

The current international urinary threshold of 20 ng/mL for free and conjugated testosterone in geldings was based on a population study conducted in Hong Kong for post-race (n=110) and training (n=382) urine samples [6]. The proposed threshold was supported by international post-race population data from the UK (n=105) and France (n=117) [7] and a further population study by Hong Kong (n=2012) [8]. This threshold has been supported to be appropriate for regulatory purposes by the results of studies involving the intramuscular (IM) administrations of 500 mg of testosterone [9], a mix of testosterone esters (100 mg of testosterone propionate, 200 mg of testosterone isocaproate and 200 mg of testosterone phenylpropionate) [6] and oral administrations of 1 mg/kg of testosterone pro-drugs; dehydroepiandrosterone (DHEA) and androstenedione [10].

In recent years, an apparent increase in the incidence of atypically high testosterone concentrations in Thoroughbred gelding post-race urine samples has been observed in the UK and Australia. A number of these samples breached the testosterone threshold of 20 ng/mL, but the overall steroid profile was not indicative of testosterone administration since it should not cause increases in epitestosterone and DHEA concentrations. Consideration of two other causative factors that could have elevated urinary testosterone concentrations were also ruled out, these being dehydration [11] or cryptorchidism, where the detection of estr-5(10)-ene-3β,17α-diol would be expected [12]. As a result these samples were categorised as ‘atypical’ and other explanations for their presence were explored. A reasonable hypothesis is that the atypical results have arisen due to considerable adrenal stimulation markedly raising the urinary testosterone production rate, with these geldings having experienced more physiological ‘stress’ compared to others.

The adrenal cortex is likely to be the main contributor to testosterone and epitestosterone production in the gelding. However, the proportion of testosterone and/or epitestosterone that arise from peripheral conversion of precursor steroids secreted as opposed to direct secretion is currently unknown. Artificial stimulation with synthetic adrenocorticotrophic hormone (tetracosactide) results in an increase in the urinary concentration of testosterone, and also DHEA and the isomers of androst-5-ene-3,17-diols [13]. Natural causes may also disturb the pituitary-adrenal axis, such as physiological stress caused by transportation, exercise and disease. Increased plasma/saliva hydrocortisone (cortisol) concentrations have been observed in horses transported for short (< 300 km) [14] and long (< 1370 km) distances [15], [16], and in horses affected by diseases, such as laminitis and acute abdominal syndrome [17]. The increase in adrenal steroidogenesis could also result in increased testosterone and epitestosterone concentrations in horses, especially since exercise induced stress has been shown to have an augmentative effect on urinary/plasma hydrocortisone concentrations [18]–[20], as well as on plasma testosterone concentrations [21]. Even though this evidence points to the aetiology of atypical urinary concentrations to be due to unusually raised adrenal

404142

43444546474849505152

535455565758596061626364

656667686970717273747576777879

steroidogenesis, the possibility of a rare polymorphism(s) in the metabolism of testosterone cannot be discounted.

Steroid ratios have been used historically to monitor the use of testosterone and its pro-drugs in geldings and mares/fillies. Prior to the current concentration threshold of 20 ng/mL in geldings, a testosterone to DHEA (T:DHEA) ratio greater than 5 was proposed for both geldings and mares/fillies [22]. This was established using a population of 153 geldings and 65 mares/fillies, and shown to be exceeded by the IM administration of 50 mg of testosterone phenylpropionate. Furthermore, a T:E ratio of 12 was adopted as a threshold for the detection of testosterone in mares/fillies [7]. This threshold was based on data collected from the analysis of 294 post-race urine samples and also following the IM dose of 1,000 mg of testosterone hexahydrobenzoate [23]. The current international concentration based threshold of 55 ng/mL in fillies replaced the T:E ratio of 12 in 2000 following a study of 1531 animals in six international laboratories [24].

Historically, the detection and quantification of low concentrations of endogenous anabolic steroids (low ng/mL to pg/mL amounts) has been analytically challenging, such as that for testosterone and epitestosterone in the urine from the gelding. However, more sensitive liquid chromatography tandem mass spectrometry (LC-MS/MS) methods have been developed since the aforementioned studies were conducted using gas chromatography mass spectrometry (GC-MS). Chemical derivatisation can also be used to further improve the electrospray ionisation of neutral steroids, pertinent to the analysis of testosterone and epitestosterone in equine plasma and urine [2], [25]–[27].

The aim of the current study was to determine whether a urinary T:E ratio could be used as a supplementary approach to help differentiate atypical testosterone concentrations in geldings from scenarios involving administration of testosterone or its prodrugs. This required a development of a sensitive and robust analytical method incorporating LC-MS/MS and chemical derivatisation for the quantitative analysis of testosterone and epitestosterone in equine urine. Extraction of a total fraction (free and conjugated) was required since testosterone is mainly excreted as a sulphate conjugate and epitestosterone as a glucuronide conjugate in the horse [28]. Furthermore, the current internationally approved threshold of 20 ng/mL testosterone in gelding urine is based on determination of the total fraction [1], so defining the reference range of T:E values using the same approach would then allow for meaningful use of the resulting data in relation to routine anti-doping processes. The resulting population data would be used to statistically determine an upper limit for a T:E ratio using a normal post-race sample size of 200 geldings. The proposed upper T:E ratio limit was applied and evaluated for its ability to detect the administrations of testosterone propionate, DHEA and a mixture of DHEA and pregnenolone (Equi-Bolic®) and to determine whether the observed atypical samples could be biological outliers.

2. EXPERIMENTAL

2.1 Reagents and standards

8081

8283848586878889909192

93949596979899

100

101102103104105106107108109110111112113114115116

118

119

Methoxylamine hydrochloride, sodium phosphate, acetyl chloride and pancreatin (8 x USP specifications) were obtained from Sigma-Aldrich (UK) and methanol, ethyl acetate, sodium hydroxide and potassium dihydrogen orthophosphate were from Fisher Scientific Ltd (UK). Laboratory water was purified using a Triple Red Duo Water system (Triple Red Laboratory Technology, Bucks, UK).

Testosterone sulphate, epitestosterone sulphate, 16,16,17α-d3-testosterone sulphate and 2,2,4,4-d4-androsterone glucuronide were purchased from the Australian Government National Measurement Institute (Australia).

Stock solutions containing individual compounds at a concentration of 1 mg/mL were prepared in methanol, and stored at – 20 °C. Mixed stock solutions for analytes and deuterated internal markers were prepared in methanol at a concentration of 10 µg/mL, and they were subsequently diluted with methanol in order to obtain separate spiking solutions at appropriate concentrations.

2.2 Equine urine samples

2.2.1 Population samples

The reference population consisted of 200 anonymised urine samples collected post-race from Thoroughbred geldings at different race meetings in the UK during January (n=63), April (n=65) and September (n=72) 2015 as part of the anti-doping programme carried out by the British Horseracing Authority (BHA). Samples were stored at – 20 ºC prior to analysis.

A further 15 urine samples collected post-race over a longer time frame from Thoroughbred geldings known to have atypically high concentrations of testosterone (≥15 ng/mL), but notably otherwise having an unremarkable steroid profile, were also analysed and compared with the reference population.

2.2.2 Administration samples

A selection of pre- and post-administration urine samples were analysed following the administrations of testosterone propionate, DHEA and a mix of DHEA and pregnenolone to Thoroughbred geldings. The UK studies were conducted under Licence from the Home Office according to the framework of the Animal Scientific Procedures Act (1986) with statutory ethical review, whilst the Australian study was conducted following the approval of Charles Sturt University (CSU) Animal Care and Ethics Committee.

Testosterone propionate (Testoprop®, 50 mg/mL by Jurox) was administered intramuscularly in five consecutive weekly doses of 50 mg to two Thoroughbred geldings at the BHA’s Centre for Racehorse Studies (CRS), UK. The horses were 7 and 10 years old, and their bodyweights were 482 and 469 kg, respectively. Urine samples were collected pre-administration and at 6 and 173 hours following the final dose for horse 1 and horse 2, respectively. The samples were stored at -20 ºC prior to analysis.

120121122123124

125126127

128129130131132

133

134

135136137138

139140141142

143

144145146147148149

150151152153154155

DHEA (by BOVA Compounding Pharmacy) was administered at a dose of 1 mg/kg in a capsule via naso-gastric tube to six Thoroughbred geldings at CSU School of Animal and Veterinary Sciences, Australia. The ages of the animals ranged from 2 to 15 years old, and the mean bodyweight was 575.3 ± 60.1 kg. Urine samples were collected pre-administration and up to 96 hours post-administration for each horse. The samples were stored at -20 ºC at Racing NSW, Australia and they were delivered in dry ice to LGC, UK for further analysis.

A supplement containing a mix of DHEA and pregnenolone (Equi-Bolic, purportedly 500 mg/15mL each) was administered orally in five consecutive daily doses of 500 mg to two Thoroughbred geldings at the BHA’s CRS. The animals were 8 and 9 years old, and they weighed 510 and 528 kg, respectively. Urine samples were collected pre-administration and 7 and 4 hours following one dose for horse 1 and horse 2, respectively. The samples were stored at -20 ºC prior to analysis.

2.3 Sample extraction

1.5 mL aliquots of equine urine were diluted with 3.6 mL of phosphate buffer (1 M, pH 6.3) containing d3-testosterone sulphate at a concentration of 28.2 ng/mL and 100 µL of pancreatin solution. Aliquots were also spiked with 75 µL of d4-androsterone glucuronide at a concentration of 1 µg/mL. Solid phase extraction (SPE) was performed using Waters Oasis®

WCX cartridges (3 mL, 60 mg), a mixed mode sorbent consisting of a polymeric reversed-phase incorporating a weak ion exchanger. This phase was used in the clean-up of neutral steroids to remove potential cationic interferences, whilst the alkaline washes were used to remove potential anionic interferences. The phase was conditioned with 2 mL of methanol followed by 2 mL of purified water prior to sample loading. Cartridges were washed with 2 mL of sodium phosphate buffer (0.25 M, pH 8.3) followed by 2 mL of purified water prior to eluting with 3 mL of methanol. Aliquots were dried and subsequently hydrolysed by adding 0.5 mL of anhydrous methanolic hydrochloride (1 M) and incubating at 60 ºC for 15 minutes [29]. The reaction was quenched with 3 mL of sodium phosphate buffer (0.25 M, pH 8.3) prior to another SPE step. Again, WCX cartridges were conditioned with 2 mL of methanol followed by 2 mL of purified water prior to sample loading and they were washed with 2 mL of sodium hydroxide (0.1 M) and 2 mL of purified water. Elution was carried out with 3 mL of ethyl acetate and aliquots were dried prior to reconstitution with 200 µL of anhydrous ethyl acetate. 25 µL portions of samples were transferred to glass HPLC vials and they were dried. To enhance analytical sensitivity, a derivatisation step was then performed, converting the oxo function on ketonic steroids, such as testosterone and epitestosterone to their methyloxime derivatives [2], [25]. 200 µL of methoxylamine hydrochloride solution in 80 % methanol in purified water (0.1 M) was added to each vial and then incubated at 80 ºC for an hour. 2 µL of each extract was injected into the LC-MS/MS system.

2.4 LC-MS/MS analysis

Sample analysis was performed on a LC-MS/MS system consisting of a Waters Acquity I-Class UPLC coupled with a Waters Xevo TQ-S triple quadrupole mass spectrometer in

156157158159160161

162163164165166167

168

169170171172173174175176177178179180181182183184185186187188189190191

192

193194

positive electrospray ionisations mode at a capillary voltage of 2.0 kV, a source temperature of 150 °C and a desolvation gas temperature at 550 °C. Collision gas was argon at a flow rate of 0.15 mL/min. Analysis was carried out in selected reaction mode (SRM), and selected transitions, cone voltages and collision energies are shown in Table 1. Data was processed using TargetLynx software. The peak areas of the analytes and internal marker were recorded and used to calculate testosterone and epitestosterone concentrations and their ratios.

Chromatographic separation was achieved on an Acquity BEH C18 (100 mm x 2.1 mm, 1.7 µm) reversed phase UPLC column using 0.1 % formic acid in methanol (A) and 0.1 % formic acid in water (B) as mobile phases. Gradient was operated at 60 ºC and at a flow rate of 0.4 mL/min. It was started at 20 % organic for 0.5 minutes followed by increases to 60 % A at 1 minute, 80 % A at 6 minutes and 99.9 % at 6.3 minutes. 99.9 % A was held for 1 minute before resuming the initial conditions and re-equilibrating for 1 minute. A total run time was 8.5 minutes.

2.5 Method validation

A matrix ‘standard addition’ approach was used to prepare calibration lines and quality control (QC) samples for each precision and accuracy batches, since both testosterone and epitestosterone are endogenous to all genders in horses [11], [30]. Pooled gelding urine was spiked at known concentrations of the analytes and the slope and intercept of the generated calibration line were used to calculate their endogenous concentrations in pooled urine. Subsequently, the calibration line and QC samples were adjusted to account for that endogenous concentration.

The method was quantitatively validated using measures of linearity, intra- and inter-batch precision and accuracy, specificity, selectivity and sensitivity (adhering to unpublished European Horserace Scientific Liaison Committee quantitative method validation guidelines). Three separate precision and accuracy batches were extracted and analysed and each of them contained a calibration line in duplicate at concentrations of endogenous (E) only, E+2, E+5, E+10, E+20, E+50, E+100, E+200 and E+500 ng/mL and six replicates of QC samples at concentrations of E, E+2, E+20, E+200 and E+400 ng/mL. Sample dilution was also investigated by spiking pooled gelding urine in six replicates at a concentration of 2 µg/mL and performing 1-in-10 dilution with pooled gelding urine to enable its quantification in the validated range.

Cross-talk between testosterone, epitestosterone and d3-testosterone were investigated by spiking pooled urine in duplicate with each standard and monitoring the changes in peak abundances at the respective retention times of the other analytes. Furthermore, the selectivity of the method was assured for each analyte through the monitoring of quantifier and qualifier ion ratios for both endogenous and augmented analyte concentrations.

2.6 Sample analysis

195196197198199200

201202203204205206207

208

209210211212213214215

216217218219220221222223224225

226227228229230

231

Following the validation, the method was used to analyse 200 post-race gelding urine samples, 15 gelding urine samples with atypical testosterone concentrations and a selection of administration samples. These samples were analysed alongside a calibration line and QC samples in duplicate at concentrations of E+3 (low QC), E+250 (medium QC) and E+400 ng/mL (high QC). A matrix ‘standard addition’ approach was again used to calculate the endogenous concentrations for the calibrators, whilst the endogenous concentrations generated in the validation were used for the QC samples to ensure acceptable inter-batch performance.

2.7 Statistical analysis of the population data

Statistical analysis of the data was performed using SPSS Statistics software (version 22). The Kolmogorov-Smirnov test for normality was used to determine whether the population data were Gaussian.

3. RESULTS AND DISCUSSION

3.1 Method validation

Chromatographic separation was achieved for testosterone and epitestosterone (Figure 1). Production of methoxylamine derivatives resulted in formation of two geometric isomers for both testosterone and epitestosterone due to their 3-keto-Δ4 structure [31]. Ideally from a computation point of view, either no separation or alternatively complete separation of the syn and anti isomers of the methoxylamine derivatives of testosterone and epitestosterone is desirable, even though the latter is unlikely to be ever achievable given the number of theoretical plates associated with the UPLC column employed. Whereas the two forms of derivatised epitestosterone were not resolved by liquid chromatography using the selected gradient, there was partial resolution for that of derivatised testosterone. Even so, satisfactory results for method validation were achieved for both steroids, as described below, based on the use of peak area data.

Method selectivity was assured by monitoring the ion ratios of quantifier (m/z 318>126) and qualifier (m/z 318>138) MS/MS transitions for both testosterone and epitestosterone. These product ions (Figure 2) are in agreement to the fragmentation patterns previously observed for oxime derivatives [25]. No cross-talk was observed for the urine samples spiked at a concentration of 400 ng/mL between testosterone, epitestosterone and d3-testosterone.

Method selectivity was further demonstrated by the analysis of 18 non-spiked urine samples from different genders (colt, gelding and mare/filly) and no interfering matrix peaks were observed at the respective retention times of the analytes using the quantifying transition (m/z 318>126). Both testosterone and epitestosterone were detected in all samples analysed due to their endogenous nature.

232233234235236237238239

240

241242243

244

245

246247248249250251252253254255256

257258259260261

262263264265266

The generated calibration lines (weighted 1/x) indicated linear regression in the range from the endogenous concentration to 500 ng/mL for both analytes with the values for coefficient of determination (r2) greater than 0.99.

Precision and accuracy were measured in terms of coefficient of variation (%CV) and relative error (%RE). The method was shown to be accurate and reproducible with low inter-batch variability for precision (within ± 10.6%) and accuracy (within ±4.0 %) for both analytes at all QC concentrations (Table 2). QC samples diluted 1-in-10 also produced accurate results with relative error less than 7.0 %.

Method sensitivity was determined in terms of lower limit of quantification (LLOQ) and limit of detection (LOD). LLOQ values were determined to be the endogenous concentrations of control urine for both testosterone (1.6 ng/mL) and epitestosterone (1.4 ng/mL). LODs were calculated by extrapolating the mean signal-to-noise (S/N) of the endogenous peaks to 3, which equated to 0.15 ng/mL for both analytes.

3.2 Sample analysis

3.2.1 Analysis of QC samples

Analysis of QC samples (Table 2) showed continuously acceptable inter-batch performance with %CV less than 9.1 % and %RE less than 7.3 %. Thus, minimal variability in the results was observed between the different batches extracted and analysed.

3.2.2 Analysis of the reference population samples

Steroid concentrations and ratios measured following the analysis of 200 post-race Thoroughbred gelding urine samples, which formed a reference population, are shown in Table 3.

Testosterone concentrations in the reference population were similar to those observed in a previously published study for 105 post-race gelding urine samples in the UK giving a mean concentration of 2.9 ng/mL with a maximum concentration of 13.0 ng/mL [7].

Endogenous epitestosterone concentrations observed in the current study were generally higher than those of testosterone, which is in agreement with a previously published study with mean testosterone concentration of 2.5 ± 2.2 ng/mL and a mean epitestosterone concentration of 4.3 ± 4.2 ng/mL for 47 geldings [32].

3.2.3 Analysis of the atypical samples

The 15 urine samples previously analysed to have atypically testosterone concentrations (above 15 ng/mL) were separated and treated as an atypical distinct population for out-group comparison purposes since these were collected over a protracted timeframe relative to the

267268269

270271272273274

275276277278279

280

281

282283284

285

286287288

289290291

292293294295

296

297298299

reference population. Testosterone concentrations measured for these atypical samples were higher than those measured for the reference population (Table 4, Figure 3). The urinary threshold of 20 ng/mL for testosterone in geldings was exceeded in eight samples analysed. Epitestosterone concentrations were similar or slightly higher compared to those of testosterone.

The calculated T:E ratios for the atypical samples were well within the range of values observed for the reference population despite the significant differences in the measured testosterone concentrations (Figure 3). This may suggest that the whole adrenal pathway was stimulated in the case of the atypical samples and an upper limit for T:E ratio could be used to distinguish these samples from the administrations of testosterone and DHEA.

3.3 Statistical analysis

Generated T:E ratios were assessed for normal (Gaussian) distribution using the Kolmogorov-Smirnov test for normality, where the distribution is regarded normal if p>0.05. Untransformed T:E ratios did not follow a normal distribution with p<0.01 (Figure 4a), and thus logarithm, cube root and square root transformations were evaluated.

The most appropriate transformation for T:E ratio was the logarithm transformation (p=0.09), when all the testosterone and epitestosterone concentrations observed for the reference population (n=200) were included in the statistical analysis (Figure 4b). Following the RSC Analytical Method Committee recommendations all the sub-LLOQ values for testosterone and epitestosterone were included uncensored to avoid biasing the population data [33]. A peak was detected for all the sub-LLOQ values, thus there being no need for replacement of zeros with a small value to enable logarithmic transformation. The upper limit of 4.9 was produced for the T:E ratio with a risk of 1 in 10,000 of a normal outlier exceeding the value (Table 5). However, as this proposed upper limit for T:E would only be applied to those samples where the testosterone concentration was greater than the international threshold of 20 ng/mL for testosterone, the combined probability of above limit T:E and testosterone concentration occurring by chance would be very small indeed, i.e. considerably lower than 1 in 10,000.

3.4 Application to analysis of administration samples

Steroid concentrations and ratios measured for the pre- and post-dose samples following the administrations of testosterone propionate (Testoprop®, IM, 5 x 50 mg, weekly), DHEA (oral, 1 mg/kg) and a mix of DHEA and pregnenolone (Equi-Bolic, oral, 500 mg) are shown in Table 6. The post-dose results in Table 6 were chosen on the basis that they represented the peak concentration (Cmax) time point for testosterone following each administration. Data was acquired for the full time course in each case, but is not presented herein to avoid the potential misuse of detection times.

300301302303304

305306307308309

310

311312313314

315316317318319320321322323324325326327

328

329330331332333334335

Sample analysis showed that testosterone concentrations exceeded the international urine threshold following the administration of each preparation. This is in agreement with the previously published studies, which have shown significant increases in testosterone concentrations for the administrations of testosterone esters [7,8], DHEA alone [10] and in combination with pregnenolone [34].

Epitestosterone concentrations were not increased following the administration of testosterone propionate, although a small increase in epitestosterone concentrations was observed following dosing with DHEA and Equibolic. This suggests that the exogenous testosterone is not converted to epitestosterone, which is supported by the previous studies following the administrations of 14C-testosterone [26] and testosterone esters [23]. Exogenous DHEA and androstenedione have been previously shown to convert to epitestosterone [10] but this occurs to a lesser extent than conversion to testosterone.

The T:E ratio exceeded the proposed upper limit of 5 following the administrations of testosterone propionate, DHEA and a mix of DHEA and pregnenolone. Importantly, the upper T:E limit was exceeded at all time points that the international threshold of 20 ng/mL was also breached; suggesting that a T:E assessment would serve as a useful adjunct. These administrations can be excluded as a cause of elevated testosterone concentrations observed in the atypical samples where T:E ratios were within the reference range.

The T:DHEA ratio is another approach previously presented in the literature [22]. However, this approach is only believed to be applicable to detect testosterone administration. In theory, this could be used as a supplementary approach to the current testosterone threshold and the proposed T:E ratio to detect the administration of epitestosterone as a masking agent.

Further work will be carried out to determine the applicability of the T:E ratio following the administration of synthetic adrenocorticotrophic hormone.

4. CONCLUSIONS

Statistical analysis of the post-race population of 200 Thoroughbred geldings produced an upper limit of 4.9 for the T:E ratio with a risk of 1 in 10,000 of a normal sample exceeding it. An upper limit of 5 is proposed to be adopted and has been shown to detect the administration of testosterone propionate, DHEA and a mix of DHEA and pregnenolone. Since the T:E ratios observed for the atypical samples were below the proposed upper T:E ratio, it can be concluded that elevated urinary testosterone concentrations were caused by the stimulation of the whole adrenal pathway rather than the administration of testosterone and/or its prodrugs.

ACKNOWLEDGEMENTS

Funding for the DHEA administration was provided by Racing Australia, through its Chief Executive Mr Peter McGauran. Staff and students at the CRS, UK, and at CSU School of

336337338339340

341342343344345346347

348349350351352353

354355356357

358359

360

361362363364365366367368

369

370

371372

Animal and Veterinary Sciences, Australia, are also acknowledged for their care and sampling of the horses involved in this work.

REFERENCES

[1] International Federation of horseracing Authorities 2015 International Agreement on breeding, racing and wagering. http://www.ifhaonline.org/resources/2015Agreement.pdf.

[2] B. P. Gray, P. Teale, C. M. Pearce. Analysis of methyloxime derivatives of intact esters of testosterone and boldenone in equine plasma using ultra high performance liquid chromatography tandem mass spectrometry. Drug Test. Anal. 2011, 3(4), 206.

[3] Y. You, C. E. Uboh, L. R. Soma, F. Guan, X. Li, Y. Liu, J. A. Rudy, J. Chen, D. Tsang, Simultaneous separation and determination of 16 testosterone and nandrolone esters in equine plasma using ultra high performance liquid chromatography–tandem mass spectrometry for doping control. J. Chromatogr. A. 2011, 1218, 3982.

[4] Z. Kaabia, G. Dervilly-Pinel, F. Hanganu, N. Cesbron, E. Bichon, M. A. Popot, Y. Bonnaire, B. Le Bizec. Ultra high performance liquid chromatography/tandem mass spectrometry based identification of steroid esters in serum and plasma: An efficient strategy to detect natural steroids abuse in breeding and racing animals. J. Chromatogr. A. 2013, 1284, 126.

[5] B. P. Gray, M. Viljanto, J. Bright, C. Pearce, S. Maynard. Investigations into the feasibility of routine ultra high performance liquid chromatography–tandem mass spectrometry analysis of equine hair samples for detecting the misuse of anabolic steroids, anabolic steroid esters and related compounds. Anal. Chim. Acta 2013, 787, 163.

[6] P. W. Tang, W. C. Law, D. L. Crone. A urinary threshold for testosterone in geldings. in Proceedings of the 11th International Conference of Racing Analysts and Veterinarians, (Eds: D.E. Auer, E. Houghton). R&W Publications, Newmarket, 1996, pp. 68–72.

[7] Y. Bonnaire, L. Dehennin, P. Plou, P. L. Toutain, A. Ginn, L. Grainger, R. Armstrong, M. C. Dumasia, A. Nedderman, E. Houghton, P. W. Tang, W. C. Law, D. L. Crone. The determination of a threshold value for testosterone: approaches to the problem. in Proceedings of the 11th International Conference of Racing Analysts and Veterinarians, (Eds: D. E. Auer, E. Houghton). R&W Publications, Newmarket, 1996, pp. 56–61.

[8] P. W. Tang, W. C. Ip, W. C. Law, T. S. M. Wan, D. L. Crone. Further studies on the testosterone threshold for geldings. in Proceedings of the 12th International Conference of Racing Analysts and Veterinarians, (Eds: B. Laviolette, M. R. Koupai-Abyazani). R&W Publications, Newmarket, 1998, pp. 107–112.

[9] E. N. M. Ho, W. H. Kwok, D. K. K. Leung, C. M. Riggs, G. Sidlow, B. D. Stewart, A. S. Y. Wong, T. S. M. Wan. Control of the misuse of testosterone in castrated horses based on an international threshold in plasma. Drug Test. Anal. 2015, 7(5), 414.

[10] L. Dehennin, Y. Bonnaire, P. Plou. Human nutritional supplements in the horse. Dehydroepiandrosterone versus androstenedione: comparative effects on the androgen profile and consequences for doping analysis. J. Anal. Toxicol. 2001, 25, 685.

[11] A. Clarke, J. P. Scarth, J. Hands, P. Teale, A. C. Mill, R. Macarthur, J. Kay. Detection of nandrolone and boldenone abuse in the ovine by GC–MS–MS. Chromatographia 2011, 73(11–12), 1155.

[12] D. K. K. Leung, F. P. W. Tang, T. S. M. Wan, J. K. Y. Wong. Identification of cryptorchidism in horses by analysing urine samples with gas chromatography/mass spectrometry. Vet. J. 2011, 187(1), 60.

[13] E. Houghton, A. Ginn, P. Teale, M. C. Dumasia, L. Grainger. ACTH administration to the gelding: effect on urinary androgens. in Proceedings of the 7th International Conference of Racing Analysts and Veterinarians, (Eds: T. Tobin, J. Blake, M. Potter, T. Wood). University of Kentucky Department Veterinary, Kentucky, 1990, pp. 507–511.

[14] E. Fazio, P. Medica, V. Aronica, L. Grasso, A. Ferlazzo. Circulating β-endorphin, adrenocorticotrophic hormone and cortisol levels of stallions before and after short road transport: stress effect of different distances. Acta Vet. Scand. 2008, 50(6), 1.

373374

375

376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423

[15] C. L. Stull, A. V. Rodiek. Physiological responses of horses to 24 hours of transportation using a commercial van during summer conditions. J. Anim. Sci. 2000, 78, 1458.

[16] A. Schmidt, S. Biau, E. Mostl, M. Becker-Birck, B. Morillon, J. Aurich, J.-M. Faure, C. Aurich. Changes in cortisol release and heart rate variability in sport horses during long-distance road transport. Domest. Anim. Endocrinol. 2010, 38, 1790.

[17] I. Ayala, N. F. Martos, G. Silvan, C. Gutierrez-Panizo, J. G. Clavel, J. C. Illera. Cortisol, adrenocorticotropic hormone, serotonin, adrenaline and noradrenaline serum concentrations in relation to disease and stress in the horse. Res. Vet. Sci. 2012, 93, 103.

[18] C. H. . Irvine, R. J. Huun, S. L. Alexander. The cortisol concentration in post-race urine of racehorses, and factors affecting it. in Proceedings of the 7th International Conference of Racing Analysts and Veterinarians, (Eds: T. Tobin, J. Blake, M. Potter, T. Wood). University of Kentucky Department Veterinary, Kentucky, 1990, pp. 101–105.

[19] P. L. Toutain, V. Lassourd, M. A. Popot, V. Laroute, M. Alvinerie, Y. Bonnaire. Urinary cortisol excretion in the resting and exercising horse. Equine Vet. J. 1995, 18, 457.

[20] M. Marc, N. Parvizi, F. Ellendorff, E. Kallweit, F. Elsaesser. Plasma cortisol and ACTH concentrations in the warmblood horse in response to a standardized treadmill exercise test as physiological markers for evaluation of training status. J. Anim. Sci. 2000, 78, 1936.

[21] L. C. Golland, D. L. Evans, G. M. Stone, C. M. Tyler-McGowan, D. R. Hodgson, R. J. Rose. Maximal exercise transiently disrupts hormonal secretory patterns in standardbred geldings. Equine Vet. J. Suppl. 1999, 30,581.

[22] E. Houghton. Threshold values - current and future problems. in Proceedings of the 10th International Conference of Racing Analysts and Veterinarians, (Eds: P. Kallings, U. Bondesson, E. Houghton). R&W Publications, Newmarket, 1994, pp. 297–302.

[23] Y. Bonnaire, L. Dehennin, P. Plou, P. L. Toutain. Testosterone administration to mares: criteria for detection of testosterone abuse by analysis of metabolites in plasma and urine. J. Anal. Toxicol. 1995, 19, 175.

[24] Y. Bonnaire, L. Dehennin, M. A. Popot, P. Plou. Testosterone in mares and fillies: a new threshold. in Proceedings of the 13th International Conference of Racing Analysts and Veterinarians, (Eds: R. B. Williams, E. Houghton, J. F. Wade). R&W Publications, Newmarket, 2000, pp. 60–63.

[25] S. Liu, J. Sjovall, W. J. Griffiths. Analysis of oxosteroids by nano-electrospray mass spectrometry of their oximes. Rapid Commun. Mass Spectrom. 2000, 14, 390.

[26] C. H. L. Shackleton, H. Chuang, J. Kim, X. de la Torre, J. Segura. Electrospray mass spectrometry of testosterone esters: potential for use in doping control. Steroids 1997, 62, 523.

[27] C. H. F. Wong, D. K. K. Leung, F. P. W. Tang, J. K. Y. Wong, N. H. Yu, T. S. M. Wan. Rapid screening of anabolic steroids in horse urine with ultra-high-performance liquid chromatography/tandem mass spectrometry after chemical derivatisation. J. Chromatogr. A 2012, 1232, 257.

[28] M. C. Dumasia, E. Houghton. Studies related to the metabolism of anabolic steroids in the horse: the indentification of some 16-oxygenated metabolites of testosterone and a study of the phase II metabolism. Xenobiotica 1981, 11(5), 323.

[29] P. W. Tang, D. L. Crone. A new method for hydrolyzing sulfate and glucuronyl conjugates of steroids. Anal. Biochem. 1989, 182, 289.

[30] J. P. Scarth, A. Clarke, J. Hands, P. Teale, R. Macarthur, J. Kay. Validation of a Quantitative Multi-Residue Urinary Assay for the Detection of Androgen, Oestrogen and Progestagen Abuse in the Bovine. Chromatographia 2010, 71, 241.

[31] M. G. Horning, A. M. Moss, E. C. Horning. Formation and gas-liquid chromatographic behavior of isometric steroid ketone methoxime derivatives. Anal. Biochem. 1968, 22, 284.

[32] A. Decloedt, L. Bailly-Chouriberry, J. Vanden Bussche, P. Garcia, M.-A. Popot, Y. Bonnaire, L. Vanhaecke. A validated UHPLC-MS/MS method to quantify low levels of anabolic-androgenic steroids naturally present in urine of untreated horses. Anal. Bioanal. Chem. 2015, 407(15), 4385.

[33] What should be done with results below the detection limit? Mentioning the unmentionable. AMC Technical Brief , Royal Society of Chemistry 2001.

424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477

[34] H. K. Knych, R. M. Arthur, S. D. Stanley, D. S. McKemie. Disposition and metabolic profile of the weak androgen dehydroepiandrosterone (DHEA) following administration as part of a nutritional supplement to exercised horses. Drug Test. Anal. 2015, 7(1), 39.

Table 1 MS/MS parameters for testosterone, epitestosterone and d3-testosterone

Analyte

Ion transition for quantification (m/z)

Collision energy (eV)

Ion transition for qualification (m/z)

Collision energy (eV)

Cone voltage (V)

Testosterone 318.2 > 126.1 25 318.2 > 138.1 25 50

Epitestosterone 318.4 > 126.1 30 318.4 > 138.1 30 82

D3-testosterone 321.4 > 126.2 30 n/a n/a 80

Table 2 Validation results for inter-batch precision (%CV) and accuracy (%RE)

Analyte QC sample ID (ng/mL)

Nominal conc.

(ng/mL)*

Measured mean conc.

(ng/mL)%CV %RE

Testosterone

Validation batches

E (endogenous)E + 2

E + 20E + 200E + 400

Dilution QC

1.63.6

21.6201.6401.62002

1.63.721.5

209.7403.42139**

5.24.32.72.12.41.2

0.72.6-0.44.00.57.0

Sample analysis

batches***

E + 3E + 250E + 400

4.6251.6401.6

4.9262.1421.4

5.23.32.7

7.34.24.9

Epitestosterone

Validation batches

EE + 2

E + 20E + 200E + 400

Dilution QC

1.43.4

21.4201.4401.42001

1.33.421.6

201.5407.42132**

10.67.22.63.53.12.8

-3.10.41.20.11.56.6

Sample analysis

batches***

E + 3E + 250E + 400

4.4251.4401.4

4.5255.3405.3

9.16.75.2

2.71.51.0

*Nominal values were calculated using the mean of the determined endogenous QC concentrations (n=18) in three precision and accuracy batches.**Concentrations for dilution QC have been adjusted for 1-in-10 dilution *** Measured mean concentrations were calculated for seven sample analysis batches with each QC sample in duplicate (n=14)

478479480481482

483

484

485486

487

488

489490491492493

494

Table 3 Steroid concentrations and ratios observed for the reference population (n=200)

Testosterone concentration

(ng/mL)

Epitestosterone concentration

(ng/mL)T:E ratio

Mean 2.71 5.92 0.55Median 1.70 4.10 0.48

SD 2.52 5.58 0.39Min 0.2 0.3 0.09Max 14.0 44.9 3.07

n 200 200 200n below LLOQ 85 23 87

Table 4 Steroid concentrations and ratios observed for 15 earlier collected urine samples, categorised as ‘atypical’, following the screening analysis.

Sample ID Testosterone concentration (ng/ml)

Epitestosterone concentration (ng/ml)

T:E ratio

1 23.7 27.6 0.92 14.8 18.7 0.83 16.1 11.1 1.54 18.2 16.5 1.15 13.8 15.3 0.96 22.3 35.9 0.67 16.0 23.9 0.78 27.6 25.3 1.19 13.9 11.9 1.210 22.2 17.1 1.311 37.8 77.2 0.512 40.0 33.5 1.213 24.0 22.1 1.114 25.7 32.1 0.815 16.0 9.4 1.7

Table 5 A calculated tentative upper limit of T:E ratios in gelding (n=200) with a 1 in 10,000 probability of exceeding the concentration by chance

No transformation

Log10

transformationMean 0.55 -0.35

SD 0.39 0.28

Mean+3.72xSD 2.00 0.69

Anti-log of transformed mean n/a 0.45Anti-log of transformed

mean+3.72xSD n/a 4.91

495496

497

498499500

501

502503

504

505

Table 6 Steroid concentrations and T:E ratios for pre- and post-dose samples following the administrations of testosterone propionate, DHEA and Equi-Bolic to Thoroughbred horses. The post-dose samples represent the Cmax time point for testosterone following each administration. Data was acquired for the full time course in each case, but is not presented herein to avoid the potential misuse of detection times. The proposed T:E ratio of 5 was exceeded at all times that the international testosterone threshold of 20 ng/ml was exceeded; thus showing that the proposed approach is fit-for-purpose.

Administration study

Sample ID

Sample collection

Testosterone concentration

(ng/ml)

Epitestosterone concentration

(ng/ml)T:E ratio

DHEA

Horse 1 pre-dose 0.2 0.2 1.0post-dose 95.1 16.5 5.8

Horse 2 pre-dose 0.4 4.9 0.1post-dose 529.7 33.7 15.7

Horse 3 pre-dose 0.2 0.2 1.0post-dose 217.6 41.6 5.2

Horse 4 pre-dose 0.2 1.2 0.2post-dose 265.8 27.0 9.8

Horse 5 pre-dose 0.3 3.9 0.1post-dose 622.1 34.6 18.0

Horse 6 pre-dose 0.3 0.8 0.4post-dose 163.6 8.3 19.7

Testosterone propionate

Horse 1 pre-dose 1.6 1.2 1.3post-dose 368.4 2.3 160.2

Horse 2 pre-dose 0.2 0.4 0.5post-dose 266.8 2.9 92.0

Equi-BolicHorse 1 pre-dose 0.8 0.8 1.0

post-dose 357.2 11.8 30.3

Horse 2 pre-dose 0.9 0.8 1.1post-dose 924.1 21.8 42.4

506

507

508

509

510

511

512

513514515516517518519520521

522

523