Thyroid Hormone Abuse Among Elite Athletes

Abstract

Context

Thyroid hormone (TH) abuse for performance enhancement in sport remains controversial and it is not prohibited in sports under the World Anti-Doping Code. However, the prevalence of TH usage in athletes is not known.

Objective

We investigated TH use among Australian athletes undergoing antidoping tests for competition in World Anti-Doping Agency (WADA)–compliant sports by measuring TH in serum and surveying mandatory doping control form (DCF) declarations by athletes of all drugs used in the week prior to the antidoping test.

Methods

Serum thyroxine (T4), triiodothyronine (T3), and reverse T3 were measured by liquid chromatography–mass spectrometry and serum thyrotropin, free T4, and free T3 by immunoassays in 498 frozen serum samples from antidoping tests together with a separate set of 509 DCFs.

Results

Two athletes had biochemical thyrotoxicosis giving a prevalence of 4 per 1000 athletes (upper 95% confidence limit [CL] 16). Similarly, only 2 of 509 DCFs declared usage of T4 and none for T3, also giving a prevalence of 4 (upper 95% CL 16) per 1000 athletes. These estimates were consistent with DCF analyses from international competitions and lower than the estimated T4 prescription rates in the age-matched Australian population.

Conclusion

There is minimal evidence for TH abuse among Australian athletes being tested for competing in WADA-compliant sports.

The use of thyroid hormones (THs), thyroxine (T4) or triiodothyronine (T3), for doping to enhance performance in elite sports has long been controversial [1-4]. Despite claims of TH abuse in sport raising the issue of their prohibition [5], THs have not been prohibited in sports by World Anti-Doping Agency (WADA) [1]. Under the World Anti-Doping Code [6], substances and methods can be prohibited in sports if they meet 2 of 3 equally weighted criteria—whether substances or methods are (1) performance enhancing, (2) harmful to athletes, and (3) contrary to the spirit of sport. They can then be added to the annually updated WADA Prohibited List [7]. Therapeutic use exemptions can be approved by antidoping organizations for valid therapeutic use of banned substances [8, 9]. The overall prevalence of doping in elite sports is very difficult to estimate [10] due to numerous methodological challenges [11] to identifying breaches of antidoping rules, which remain highly confidential, covert practices as professional athletes aim to avoid career-ending sanctions if antidoping rule violations are proven. Overall estimates remain <5% with a median estimate of <10% [11]; lower estimates come from biological testing of an athlete's samples whereas higher estimates arise from athlete questionnaire surveys that embody strong assumptions on honest responses and knowledge of antidoping rules [10, 11].

Available evidence suggests that in euthyroid individuals use of high (ie, supraphysiological) TH doses would create thyrotoxicosis, with suppressed serum thyrotropin (TSH) reflecting the sensitive TH sensing with negative hypothalamic–pituitary feedback, a state highly likely to be detrimental to sport performance due to loss of skeletal [12-21] and cardiac [22, 23] muscle mass, leading to reduced exercise capacity with either natural [24, 25] or experimental T4-induced [26] or T3-induced [27] thyrotoxicosis, notwithstanding anecdotal claims of enhancement that lack objective verification [1]. Furthermore, TH abuse can produce severe, dangerous thyrotoxicosis [28-30]. Nevertheless, sports medicine doctors and administrators receive anecdotal reports of nonprescribed TH use by athletes, especially those competing in weight-classified sports (weightlifting, combat sports) aiming to “make-weight” at formal precompetition weigh-in, and remain concerned about the prevalence, safety, and doping uses of THs. Yet, the prevalence of TH abuse among athletes undergoing antidoping tests for competition in WADA-compliant sports has not been studied systematically. Hence the present study aimed to use a multimodality approach to estimate the prevalence of TH use among such athletes. This consisted of investigating circulating TH concentrations in randomly selected, frozen stored serum samples leftover from routine antidoping tests as well as the prevalence of TH use reported in a separate random sample of doping control forms (DCFs), which must be completed with every antidoping test. Both these estimated TH usage rates were then compared with national age- and sex-specific prescription rates for TH in a young population comparable with the athletes undergoing antidoping tests.

Materials and Methods

Serum Samples

A random selection of frozen serum samples (n = 500) left over from WADA-compliant antidoping tests obtained between 2009 and 2020 (72 from 2009-10, 92 from 2011, 98 from 2012, 100 from 2013, 91 from 2019, and 45 from 2020) and stored frozen at −80 °C at the Australian WADA-accredited antidoping laboratory (Australian Sports Drug Testing Laboratory, National Measurement Institute, Ryde, Sydney, NSW, Australia) were provided for TH assays. Antidoping serum samples are immediately split into A and B aliquots. The A samples are thawed once for immediate testing. The B samples are retained to verify a positive A sample when its findings are classified as an adverse analytical finding (<1% of A samples). Serum samples in this study were from long-term stored A and B samples and had therefore been unthawed or thawed once, although the proportion of once-thawed A and mostly unthawed B samples was not known. At the time of the antidoping test, athletes are asked for approval for use of their de-identified samples in unspecified future research (95% provided consent, Speers personal communication) and only samples with approval were included in this study.

Doping Control Forms

Athletes undergoing an antidoping test must complete a DCF in which they declare all substances used in the previous 7 days prior to the test. A random selection of DCFs from 2021-2 were reviewed for the substances declared and categorized according to the Olympic and Paralympic Formulary [31] using simplified broader drug categories.

The serum samples and DCFs were from different populations of athletes. Approval for use of stored serum samples and DCFs from consenting athletes was provided by Sports Integrity Australia. For privacy reasons, additional information was restricted to age, sex, and sport for each serum sample and DCF.

Liquid Chromatography–Mass Spectrometry Analysis

Details of the liquid chromatography–mass spectrometry (LCMS) measurements including validation and sensitivity are provided in elsewhere [32]. Briefly, serum T4, T3, and reverse T3 (rT3) were measured by LCMS using an in-house–developed method with standards (L-T4, catalog # T-073; 3,3′,5-T3 #T-074; 3,3′,5′-rT3 #T-075) and internal standards (L-T4-¹³C6 #T-076; 3,3′,5-L-T3-¹³C6 #T-077; 3,3′,5′-L-rT3-¹³C6 #T-078) from Sigma-Aldrich. All chemicals were obtained from Sigma-Aldrich including LCMS grade acetonitrile (Merck #1.00030.2500), methanol (Thermo-Fisher LCMS grade #M/4062/17), formic acid (Thermo-Fisher, LCMS grade #AC10760050) with reagent grade citric acid (#251275), ascorbic acid (#A92902), and dithiothreitol (#D0632). The limits of detection and limits of quantitation, defined according to the FDA [33], were 0.05 and 0.125 ng/mL for T4 and 0.01 and 0.125 ng/mL for both T3 and rT3. Over 5 levels of assay standards for each analyte, the extraction efficiency ranged between 87.1% and 100.7% and the matrix recovery ranged between 88.1% and 113.3%. Reproducibility according to the coefficient of variation over 4 levels of standard (0.5, 5, 15, and 60 ng/mL for T4 and 0.05, 0.5, 1.5, and 6 ng/mL for T3 and rT3) was between 96% and 107% for between day (20 replicates) and 94% to 107% for within day (10 replicates).

Immunoassays

Serum TSH (RRID:AB2756377), “free” T4 (FT4, RRID:AB2893401) and “free” T3 (FT3, RRID:2827365) were measured by Elecsys electrochemiluminescence immunoassay on a Roche Cobas autoanalyzer system according to the manufacturer’s instructions. Serum TSH was traceable to the second International Reference Preparation World Health Organization Reference Standard 80/558, and FT4 and FT3 were reported to be indirectly standardized to equilibrium dialysis by the kit manufacturer. Expected reference ranges (95%) were 0.290 to 4.20 mIU/L for TSH, 12 to 22 pmol/L for FT4, and 3.1 to 6.8 pmol/L for FT3 according to the manufacturer. The limits of detection and limits of quantitation were both 0.005 mIU/L for TSH, 0.5 and 1.3 pmol/L for FT4, and 0.6 and 1.5 pmol/L for FT3, respectively. Reproducibility was routinely assessed using Bio-Rad Liquidchek Immunoassay plus controls at 3 levels (low, mid, high range) spanning the assay working range. For TSH the coefficient of variation was between 1.6% and 1.9% (at mean 0.95, 6.36, and 34.6 mIU/L concentrations), for FT4 between 2.9% and 3.6% (at mean 14.4, 35.5 and 60.3 pmol/L), and between 2.1% and 4.2% for FT3 (at mean 3.5, 10.9 and 20.8 pmol/L).

Thyroid Hormone Prescribing

The prevalence of TH prescribing over the period 2012-2020 was estimated by the number of prescriptions recorded by the Pharmaceutical Benefits Scheme, the component of the Australian national health insurance which subsidizes prescription costs of approved medicines. Data were provided for men and women aged 15-44 years and classified according to whether they were for prescription of T4 or T3. Using additional data, the age- and sex-specific population prevalence of T4 usage was estimated from the age- and sex-specific population from the 2021 Australian census and an estimate of 2 prescriptions per year per person on T4 treatment group based on a prescription supplying 200 tablets.

Data Analysis

Data were analyzed by descriptive methods, the unpaired t-test, and analysis of variance and covariance, as appropriate, using NCSS 2022 software (NCSS, Kaysville, UT, USA). The distribution of TH measurements was evaluated by the Shapiro–Wilks test for normality and optimal normalizing transformations were determined by Box–Cox analysis. As a result, for analysis serum TSH, T4, T3, and rT3 were log₁₀ transformed whereas FT4 and FT3 were not transformed. The empirical 95% confidence limits for each TH were determined from the data by nonparametric estimation of the 2.5% and 97.5% percentiles of the distribution. The confidence interval of a proportion (with continuity correction) was estimated as described [34].

Results

Two of 500 serum samples (184 females, 316 males) were unusable due to severe hemolysis so the analysis included 498 samples from 27 sporting discipline groups with the most frequent comprising football codes 127, aquatic sports (swimming) 76, cycling 75, weightlifting 66, and athletics 38. The mean age of athletes providing samples was 25.6 ± 5.3 (median 25 years, interquartile range 22-29 years). Reference ranges for THs are tabulated (Table 1).

Serum TSH Immunoassay

The empirical distribution of the serum TSH in this population of athletes was significantly higher than the manufacturer's recommended reference range (Table 1, Fig. 1). Median serum TSH did not differ significantly between male and female athletes (Fig. 1) nor across the years of sampling (Fig. 2).

Using the empirical reference range, TH concentrations are tabulated for those with low (Table 2) or high (Table 3) serum TSH concentration. Among 12 with a low TSH (<0.68 mIU/L), 2 had biochemical thyrotoxicosis and the remainder had no other out of range TH concentrations. Among those with a high TSH, 1 had a definite biochemical hypothyroidism and 2 others had possible hypothyroidism, with 1 also having a high T4 concentration.

Using the expected reference ranges, there were 4 with low TSH, with all having normal serum T4 and rT3. One man had a high FT4 and FT3, and another had a high FT4 but normal FT3 (Table 2). Of the 8 athletes with serum TSH between the lower limits of expected (0.28 mIU/L) and empirical (0.68 mIU/L) reference ranges (Table 2), none had any out-of-range TH concentrations.

Among the 55 with higher serum TSH concentrations than the expected reference range (data not shown), 4 had low and 2 had high T4 concentrations, 5 had high T3 and 1 low T3, with 1 having both T4 and T3 high and another with both low T4 and T3. Four had high and 1 low FT4, and 5 had high and 3 had low FT3. Two had both high FT4 and FT3.

In concert, these data provide a point estimate of prevalence of overt biochemical thyrotoxicosis (low serum TSH with high serum T4 or FT4) of 4 per 1000 athletes with the 95% upper confidence limit of 16 per 1000.

Serum T4, T3, and rT3 (LCMS)

Females had higher median concentrations of serum T4 (100 vs 93 nmol/L, P < .001) and T3 (1.45 vs 1.37 nmol/L, P = .037) but similar median rT3 (0.21 vs 0.21 nmol/L) (Fig. 1). All 3 THs displayed significantly progressive decrease in concentration over time and the ratio of T3 to T4 decreased progressively over time (Fig. 2).

Serum T4 was high (>152 nmol/L) in 11 athletes. None had a low serum TSH, 1 had a high serum TSH and T3 (8.9 mIU/L, 2.5 nmol/L), 2 others had high T3, 1 had a high FT4, and 1 had a high FT3 and 1 a low FT3. Serum T4 was low (<53.7 nmol/L) in 12 athletes. One had a high serum TSH with a low FT4 (47.1 mIU/L, 10.8 pmol/L), with 5 having a low T3, 1 having a low FT4, and all had normal FT3 and rT3.

Serum T3 was high (>2.25 nmol/L) in 12 athletes. One man had a very low serum TSH (0.002 mIU/L), with high serum T3 (3.15 nmol/L), FT4 (29.2 pmol/L), and FT3 (12.0 pmol/L). One had an isolated high serum TSH (10.8 mIU/L) and another had an isolated high serum T4 (209 nmol/L). Serum T3 was low (<0.87 nmol/L) in 12 athletes. Among those with a low serum T3, none had elevated serum TSH, and 5 had a low T4 with all having normal serum FT4, FT3, and rT3.

Serum FT4 and FT3 (Immunoassay)

The empirical distributions of FT4 and FT3 were minimally higher than the manufacturer's expected reference ranges. Females had lower serum FT4 (16.0 vs 17.0 pmol/L, P < .001) and FT3 (5.1 vs 5.5 pmol/L, P < .001) than males (Fig. 1). Serum FT3, but not FT4, decreased significantly over time (Fig. 2). The serum FT3 to FT4 ratio changed significantly but not monotonically over time.

Using the empirical reference range, serum FT4 was increased (22.7 pmol/L) in 10 and decreased (<12.8 pmol/L) in 11 athletes. Among those with high FT4, 3 had a low serum TSH with 1 having a high T3 and FT3 but normal serum T4 and rT3. The other 2 had normal serum T4, T3, FT3, and rT3.

Among those with decreased serum FT4, 3 had a high serum TSH (47.1, 16.5, 10.3 mIU/L) with the remainder having normal serum TSH. One had a high serum T4 (153 nmol/L) and FT4 (24.8 pmol/L) with normal serum T3 and FT3, and the remainder had normal serum T4, T3, FT4, and FT3.

Serum FT3 was increased (>7.0 pmol/L) in 12 and decreased (<3.68 pmol/L) in 11 samples. Among those with increased FT3, 1 had a low TSH, none had a high T4, 3 had high T3, and 3 had high FT4. Among those with low FT3, 2 had a high serum TSH, 2 had high T4, and 2 had low FT4, with all having normal T3 and rT3.

Doping Control Forms Analysis

DCFs from 509 antidoping tests were analyzed for declared substances comprising 197 females and 312 males representing 16 sports disciplines, including football 175, athletics 80, cycling 68, basketball 50, aquatics 46, cricket 25, and rowing 25. Athletes completing DCF samples had a mean age of 26.0 ± 5.3 years (median 25 years, interquartile range 22-29 years). There was a total of 779 declarations from 450 athletes reporting 1 or more substances used, consisting of reported use by each individual of 6 (1), 5 (1), 4 (11), 3 (61), 2 (165), 1 (211), and 0 (59) substances.

There were 2 declarations of T4 use and none for T3 use, providing a prevalence estimate for T4 usage of 4 (upper 95% confidence limit 16) per 1000 athletes. Other substances declared included 342 (67%) over-the-counter diet supplements (including nutraceuticals, vitamins, and iron), 207 (41%) analgesics, 44 (7%) psychotropics (anxiety, depression, sleep, epilepsy, migraine), 41 (8%) asthma/cough, 39 (8%) antihistamine/antiallergy, 37 (7% overall, 18% females) oral contraceptives/estrogens, 21 (4%) antibiotics, 17 (4%) gastrointestinal (antacid, laxative, antiemetic, diarrhea), 8 (2%) topicals, 2 (0.4%) cardiovascular, and 18 (4%) miscellaneous.

Thyroid Hormone Prescribing

Over the period 2012-2020, the overall national population rate of T4 prescription was consistently and significantly higher for females than males (Supplementary Material #2, top left [32]). Over that time period there was a gradual rise by about 50% from 100 000 to 150 000 prescriptions per annum in females while male rates rose from 16 000 to 24 000 per annum.

Over the same period, the age-group specific prescribing of T4 increased in females from 32 000 prescriptions per annum at age 15-19 years to 360 000 prescriptions per annum at age 404-44 years in females (Supplementary Material #2, bottom left [32]). The corresponding male age group–specific rates rose from 10 750 at 15-19 years to 64 000 at age 40-44 years per annum (Supplementary Material #2, top right [32]).

Using the national sex- and age-specific population at the median age of the athletes in the 2 samples within this study, the estimated national prevalence of T4 prescriptions (per 1000) at the age of 25-26 years was 10 for men, 65 for women, and 37 for all persons.

The overall proportion of T3 prescription (prevalent rate) remained at about 1% of T4 prescriptions over the time period 2012-20, although the proportion of new (incident) T3 prescriptions rose gradually from 0.3% to 0.7% of corresponding T4 prescription rates over the same period (Supplementary Material #2, bottom right) [32].

Discussion

The present study indicates minimal evidence for TH abuse among Australian athletes undergoing antidoping tests to compete in WADA-compliant sports over the period from 2012 to 2020. This interpretation is based on 2 separate random samples of around 500 Australian athletes undergoing antidoping tests, representing complementary approaches to determining the prevalence of TH use. One approach was to measure serum concentrations of THs (TSH, T4, T3, rT3, FT4, FT3) in frozen sera left over from WADA-mandated antidoping tests, and the second was to analyze a separate set of DCF declarations of drugs used by the athlete in the week prior to a mandatory antidoping test. In both approaches the prevalence of TH use was congruent and low, 4 per 1000 athletes with an upper 95% confidence limit of 16 per 1000. Those prevalence estimates are substantially lower than the age-specific T4 prescription rate (37 per 1000) among Australians of similar age to the athletes studied. The DCF-based estimate of declared T4 usage is also consistent with the overall estimate of 14 per 1000 drawn from DCF analysis of nearly 23 000 Olympic and European games athletes [1]. It is substantially lower than an estimate from Italian athletes between 2017 and 2021 [35].

To estimate the prevalence of TH abuse from TH measurements, we assume that high (supraphysiological) TH doses would increase net circulating TH action thereby suppressing serum TSH, forming a sensitive generic indicator of biochemical thyrotoxicosis. In euthyroid individuals, lower, subreplacement T4 doses titrated to not fully suppress serum TSH have no detectable effects on body composition or muscular function [36]. Using the criterion of suppressed serum TSH with increased serum T4, according to the empirical or expected reference ranges, the prevalence of biochemical hyperthyroidism was lower than the age-specific rate of T4 prescription among Australians of comparable age to the athletes. Crucially, however, reflecting the likely difficulties of solely laboratory-based surveillance for TH abuse, no clinical evaluation was possible due to the study's privacy restrictions. Evaluating such biochemical thyrotoxicosis would require additional evidence to evaluate the possibility of TH doping. Notably a thyroid uptake scan would be required to distinguish endogenous thyrotoxicosis from natural causes such as Graves disease or toxic thyroid nodule(s) requiring medical management, from exogenous ingestion of T4 or T3 or of thyroiditis [37]. A low thyroid uptake could be due to exogenous T4 or T3 ingestion but could also be due to thyroiditis, for which additional tests (thyroglobulin, inflammatory and immune biomarkers) would be required. This is analogous to the requirement for evaluating a positive human chorionic gonadotrophin (hCG) finding in a man's urine antidoping test to distinguish between an early testis cancer and hCG doping [38, 39].

Mild or subclinical thyrotoxicosis, defined as minimal or no symptoms with suppression of serum TSH with normal serum T4 (FT4) and T3 (FT3) concentrations [40], has a population prevalence of <1% in iodine-replete populations [41]. In our cohort, utilizing the expected reference range, 4 patients (4/498) had subnormal TSH with normal T3 and T4, which equates to ∼0.8%, in keeping with an iodine-replete population. Some reported practices of abusing TH involve short-term periods using T3 (Cytomel), which may be seen biochemically as subclinical T3 thyrotoxicosis, and therefore we cannot completely exclude subtle or intermittent use of exogenous TH in this small group where it is not sufficient to suppress serum TSH. The health status of subclinical thyrotoxicosis in elderly people remains contentious, with debate about its significance for cardiovascular (including atrial fibrillation and reduced exercise tolerance), bone, and mental health [40, 42]. Among healthy young individuals, mild or subclinical thyrotoxicosis is less likely to have such major adverse health consequences [43], but with nonspecific (indiscriminate) weight loss with reduced muscle as well as fat mass [13] leading to impaired skeletal [12-21] and cardiac [22, 23] muscle mass and function. This leads to reduced exercise capacity from natural [24] or experimental thyrotoxicosis due to T4 [26] or T3 [27] treatment, resulting in reduced maximal exercise tolerance from a hyperdynamic cardiac output [25]. These effects are likely to have detrimental effects on peak performance in power or endurance sports. In this population, most suppressed TSH concentrations have no clinical relevance and may simply reflect the unusually low fat mass and reduced leptin contribution to hypothalamic TSH-releasing factor and pituitary TSH secretion [44-47]. An interesting population for further study in this regard is the mild iatrogenic thyrotoxicosis induced by T4-induced suppression of serum TSH to prevent recurrence of thyroid cancer [48, 49].

One unexpected finding from TH measurements is that serum TSH concentrations among this young athlete population are biased upwards from those of the general population, as reflected in the manufacturer's expected reference range, which was based on a study of 516 healthy Europeans [50]. The empirical reference range for serum FT4 and FT3 was only minimally higher than the manufacturer’s recommended expected range, though the origins of those reference ranges are not disclosed. The significance of this finding is unclear. Australia's national state of mild iodine deficiency [51] was rectified by the introduction of mandatory iodine salt fortification in 2009, so slow resolution of residual effects of iodine deficiency may explain the progressive temporal changes in TH concentrations and relativities. Whether regular training may explain such small effects on TH concentrations cannot be excluded based on the variable effects reported of long-term exercise on serum TSH and TH concentrations [52, 53]. Prior preliminary investigations of TH measured by LCMS in athletes has been very limited [54]. Another consideration is the stability of serum THs during long-term frozen storage for over 10 years. However, there are few studies of TH measurements over long-term frozen storage, and none using LCMS measurements of serum T4 or T3. Based on the structural similarity of TH and sex steroids, it is notable that sex steroid LCMS measurement displays highly stable and reproducible measurement after frozen (−80 °C) storage for 10 years [55]. In 1 immunoassay study, serum TSH and FT4 were stable with small increases FT3 concentrations over up to 25 years of frozen (−25 °C) storage [56, 57]. Another study of long-term storage at −80 °C demonstrated inconsistent serum TSH, FT4, and FT3 measurements using different assays. Repeated freeze–thaw cycles did not significantly change serum TSH [58, 59] or serum T4, T3, rT3, and FT4 measurements by immunoassays [58]. For serum TSH immunoassays, most studies suggest stability of measured concentrations over prolonged storage or freeze–thaw cycles with only a single study reporting a reduction in serum TSH during frozen storage [60]. The latter may resemble the loss of immunoassayable hCG in prolonged frozen storage at −20 °C, but not −80 °C, due to dissociation of glycoprotein hormone subunits which hinders dual-site immunoassays with epitopes on different subunits [61].

Other aberrations of TH measurements were mostly isolated deviations of a single TH from the reference range with uncertain significance. In this young healthy population the potential benefits or harms of mild or subclinical thyrotoxicosis are speculative and unlikely [42]. Nevertheless, small, isolated increases in serum T4 or FT4 without TSH suppression might be consistent with ingestion of small or subreplacement T4 doses or of TH metabolites or analogs [62-64]; however, subsuppressive TSH doses are unlikely to have any beneficial or detrimental effects on athletic performance [36], whereas the effects of doses sufficient to suppress serum TSH would be thereby indirectly detectable. Similarly, despite evidence from the unregulated bodybuilding internet sites encouraging T3 use as a more potent and faster-acting TH [1], there was no evidence of T3 toxicosis such as a disproportionate increase in serum T3 or FT3 with or without suppressed serum TSH in this athlete population.

If THs were prohibited in sports, surveillance testing of THs would be required to detect TH abuse. Although TH function tests are adequate for clinical care in conjunction with medical history, physical examination, and thyroid uptake scans, laboratory antidoping tests primarily operate in isolation and require a medico-legal standard sufficient to potentially bar professional athletes from conducting their profession. In that context, the limited harmonization of TH immunoassays could create medico-legal problems if used alone. The pivotal serum TSH screening immunoassays, despite lacking a pure gravimetric standard, have achieved significant harmonization [65], although without universal acceptance [66-69]. By contrast, despite availability of a pure gravimetric T4 standard, the diversity and incongruence of commercial T4 immunoassays remains suboptimal even for clinical care [70, 71], with analogous or greater problems arising for T3 immunoassays [70, 72, 73]. Still more serious limitations arise for harmonization of FT4 and FT3 immunoassays, which inherently lack any certified reference standard and have achieved only limited standardization [73-76]. Nevertheless, the extensive quality control of the commercial TH immunoassays and in-house TH LCMS measurements using certified reference standards means that while the precise numbers may vary a little, the deduced patterns from this study are likely to remain the same with other comparable assays. The use of LCMS measurements for TH for which valid reference standards exist, may offer better accuracy [72, 73, 77] when operating within WADA laboratories with extensive LCMS expertise, but they remain not widely available in general clinical pathology laboratories.

Prevalence of TH misuse or abuse in the general community, outside WADA surveillance, may be higher in certain niches such as regular gym attendees and bodybuilders than among these athletes studied. Whereas policing under WADA regulations of drug use within elite sports is a strong deterrent, there is no regulatory hindrance to use of substances for image enhancement in bodybuilding and related drug-based hobbies outside WADA compliance, for which THs are readily available without prescription over the internet. TH abuse, including nonprescription use of T3, without critical reference to scientific data or medical supervision, is promoted unrestricted by some bodybuilding regimens operated by those marketing illicit supplies of drugs.

Limitations of this study include the privacy restriction, which limited relevant analytical clinical variables and prevented linkage of the sera to the DCFs. Nevertheless, the 2 separate random samples of the Australian athletes provide consistent estimates of the T4 usage rates, both lower than the age-specific T4 prescription rates in the general Australian population. Furthermore, despite internet encouragement to use T3 as the more potent TH, there was no evidence of selective T3 usage. Another limitation is that athletes undergoing antidoping tests can refuse to allow their samples to be used for unspecified future research because such research is often aimed to develop more sensitive antidoping detection tests that could hypothetically lead to retrospective disqualification for using banned substances. Although in theory refusal to consent could bias the findings, the actual rate of refusal was very low. Furthermore, there is no incentive to underreport TH usage because TH are not prohibited. Although DCFs survey only the last week prior to testing, even truthful completion may not include earlier use of substances; however, the unusually long half-life of T4 (1 week vs 1 day for T3) means that T4 doping is less likely to be missed by the 1-week time window of DCFs. Nevertheless, if the faster onset/offset of T3 was being exploited to “make-weight” in weight-classified sports, the maximum effect would be at the time of precompetition weigh-in, thereby within a 1 week recall of the DCFs and a suppressed serum TSH would be evident. As nonprescribed TH use by athletes aiming to improve performance is not based on sound science, the motivation for usage is likely to arise from social influencers such as coaches, team members, or friends, which could lead to isolated niches of TH abuse [35] that may be missed in surveys assuming a uniform patterns of usage.

We conclude that TH abuse among Australian athletes undergoing antidoping tests is minimal with the prevalence of biochemical thyrotoxicosis no higher than the population risks of thyroid disease or valid T4 prescribing rather than TH doping.

Acknowledgements

This study was funded by the Partnership for Clean Competition. The authors are grateful to Vanessa Agon, Janelle Grainger, and Catrin Goebel of the Australian Sports Drug Testing Laboratory (National Measurement Institute) for provision of serum samples and to Dr. Kerry Atkins, Director, Drug Utilisation Section (Office of Health Technology Assessment Policy Branch, Australian Government Department of Health) for access to thyroid hormone prescribing data.

Disclosures

D.J.H. and N.S. are members of the World Anti-Doping Agency's List and Laboratory Expert Groups, respectively. No other author has a relevant disclosure. Opinions expressed are those of the authors and do not represent the views of any organization.

Data Availability

This study data are not publicly available and subject to privacy restrictions. Summary data may be provided on reasonable request to the corresponding author.

References

Abbreviations

 
• CL

  confidence limit

 
• DCF

  doping control form

 
• hCG

  human chorionic gonadotrophin

 
• LCMS

  liquid chromatography–mass spectrometry

 
• TH

  thyroid hormone

 
• T3

  triiodothyronine

 
• T4

  thyroxine

 
• TSH

  thyrotropin

 
• WADA

  World Anti-Doping Agency