ADAS’s dataset from 19 years of work provides rare longitudinal insight, useful for examining gender patterns in anti-doping outcomes within a single national system and a whole range of different sport disciplines. The length of follow-up and completeness of the national testing pool across this entire period constitute a key methodological strength of this study, allowing us to examine relatively stable gender patterns despite multiple revisions of the WADA Code and Prohibited List. Our main result, which shows that male athletes are significantly more likely to commit an ADRV than female athletes, corresponds to global evidence showing lower female AAF/ADRV rates [28]. Moreover, our study’s results correspond to not only the overall results of other studies but also the magnitude of the male–female gap common to multi-country analyses, suggesting that upstream determinants (e.g., sport culture, risk perception, substance market structure) may be operating similarly across various contexts [28,29].

Empirical studies show striking discrepancies regarding incidents of ADRVs by sex, as reports at the national level provide evidence that AAFs in women are lower in comparison to men, particularly for anabolic agents. In contrast, some drugs (such as β2-agonists and diuretics) have been reported more often in female test-positive athletes [30]. Recent global figures align with this pattern, and among ADRVs arising from AAFs in 2022, 77% involved men and 23% women [31]. Integration of the doping data further reinforces these conclusions and provides evidence that, while anabolic agents are the most detected drug in a broad cross-section of sports, their occurrence differs substantially between the genders and different sports [32]. A significant difference in the proportion of monitored substance classes between sexes was reported by one study. The reported use of substances (generally) from classes S1 (Anabolic Agents), S4 (Hormone and Metabolic Modulators), and S8 (Cannabinoids) was lower among female athletes than their male counterparts and higher from classes S3 (Beta-2 Agonists), S5 (Diuretics and Masking Agents), and S9 (Glucocorticoids). This study also found that, for two substance classes, women only used recombinant erythropoietin (S2), while men used more substances, including hGH and growth factors. In the S4 class, anti-estrogens were the only substances detected in women; however, men used aromatase inhibitors and metabolic modulators. Comparable class-specific patterns are reported in a seven-year national dataset (2013–2019), where women showed a higher relative detection of terbutaline (S3) and furosemide (S5), while men showed clustered use of S1 combined with S4 agents and broader S2 profiles; the same study also documents that women accounted for about 22% of all controls over that period, underscoring the need to report sex-disaggregated denominators [33].

In addition to differences in overall prevalence rates, the distribution of prohibited substance classes exhibits notable gender-specific patterns. Among female athletes, anabolic agents (S1) prevail but are heavily concentrated in bodybuilding, which emphasizes muscularity and leanness as scoreable standards. This focus is supported by recent literature that has linked sport types emphasizing physique to increase steroid exposure among female athletes [30]. Our national data therefore suggest that for Serbian women, ADRVs are primarily linked to physique-oriented and weight-class sports, while for men, ADRVs also encompass team and strength/power sports where anabolic agents and stimulants are used to support the training load and explosiveness, and, in some cases, for recreational use (e.g., cannabinoids). This supports the idea that performance requirements and body ideals in sports steer athletes toward certain substances. Also, men are more likely to test positive for S6 (23.3%) and S8 (21.3%), whilst women show higher rates of detection of diuretics/masking agents (S5) and hormone/metabolic modulators (S4). Data collected from international surveillance has also indicated that men are more likely to be found with substances linked with acute performance enhancement or recreational use, whereas females are commonly found with drugs linked to body-mass control and endocrine function [34,35]. These class-specific patterns suggest that ADRVs represent a spectrum of behaviors rather than a single phenomenon. The findings in our study show similarities but also notable differences when comparing with reports from Western European settings. Consistent with some studies, male athletes show a predominance of S1 violations with additional clustering of stimulants and cannabinoids [30,33,34,35]. However, unlike some Western European datasets where S3 (β2-agonists) and S5 (diuretics) dominate among female test-positive athletes, Serbian female ADRVs remain heavily anchored in S1, with diuretics and hormone modulators in a secondary position. This suggests that, in our context, female doping behaviors may be more tightly coupled to bodybuilding and strength-aesthetic sports than has been described in some larger Western European programs.

These sex differences may be related to the interaction of biologic, sociocultural, and sports-specific factors. The observed gender gap is likely shaped by at least three sets of external factors that are themselves gender-biased: (i) sport participation and testing exposure, (ii) sport-specific performance demands and body ideals, and (iii) broader sociocultural expectations about masculinity, femininity, and risk-taking. Consider androgenic-anabolic agents with strong sexual dimorphism in their pharmacodynamics, and adverse event profiles (including virilization risk, which may discourage use in some women), but larger hypertrophic effects for athletes with less muscle mass at a given dose. These might paradoxically elevate demand in regard to the female physique depending on the sport and if there is a heavy emphasis on the visible hypertrophic/optical appearance of necessarily muscular bodies [36]. In social terms, some cultural ideals of masculinity/femininity and styles of coaching may work to encourage or restrain risk-taking in relation to doping within a certain environment. Sport discipline prejudices are also a significant factor. The pressures on weight-class and aesthetic sports, driven by high expectations, increase athletes’ attraction to S4/S5 agents, whereas in power/speed sports, the anabolic trajectory dominates, making S1 substances more dominant; meanwhile, as this combination input is exaggerated across these types by a gendered participation gap, additional burdening constraints are increased or distorted. Economic rewards and exposure play a role as well, since the larger the number of men in professionalized leagues, the more commercially attractive and viewed the sport becomes. This also matters when it comes to discovery dynamics and therefore experimentation decisions. Taken together, these factors help to explain why, even when the test volume is considered, Serbian male athletes exhibit a substantially higher ADRV rate than female athletes, and why the substance-class profiles differ across sexes.

When examining the “sport type” of doping-positive athletes, the sport type emerged as a statistically significant (yet modest) predictor, since team sport violations were less likely to be associated with female than male athlete offenders, such as individual sport offences (Nagelkerke R²= 0.094), showing that only a small portion of the variance was explained by the discipline. The regression model explains very little and itself suggests that future studies need more multivariate analyses. However, this tendency is in general agreement with large anti-doping databases and survey research, which shows that men are more likely to use or report using performance-enhancing agents than women both at a general level and also nationally [14]. Moreover, research has revealed moderate significant effects of the sex on the type of sport and doping interaction, where male athletes in both team and individual sports used performance-enhancing substances at higher rates than female ones, and where male players had higher scores for their likelihood of doping (e.g., basketball and handball) [37]. When all party-level variables are taken together, these findings imply that the sport type on its own contributes relatively little to the prediction, but the gendered pattern we found in doping-positive cases matches with previous findings in the literature describing male athletes—especially those who engage in team sports—as a more risk-laden group for doping.

Such findings have practical implications for prevention and testing. Education must be both sport- and gender-inclusive in practice. For women in physique- or weight-class-restricted sports, such curricula should be dedicated to educating on the potential health and detection ramifications of S1 use; the endocrine and renal effects of S4/S5 strategies; as well as using evidence-based non-pharmacological approaches to body periodization. Men: An additional focus is needed on team and combat sports where there is a high likelihood of stimulant abuse, polydrug effects with pseudoephedrine, under the influence of fatigue, and spillover risks from social use of cannabis or synthetic cannabinoids, where anti-doping rules do not align with the civil law. The testing plan can involve gendered and sport-specific risk indicators without sliding into discriminatory practice, such as targeted windows around known weight-cut periods, in-season peaks in stimulant temptation, or off-season cycles in physique sports, all while maintaining proportionality and respect for athletes’ rights. Those working as supporting staff for athletes should be trained to identify gender-specific red flags and provide safe alternatives.

Despite its benefits, this study has several limitations. It is a retrospective, quantitative cross-sectional study of previously reported ADAS data. While it can map trends and associations, the nature of our study is such that it cannot resolve the causal pathways linking motives, contexts, and substance choices. Across 19 years, multiple moving parts changed: the WADA Prohibited List changed and analytical technologies improved. Each shift modifies both the true underlying prevalence and the probability of detection, making trend analysis sometimes challenging. The ABP (hematological and steroidal modules) further reweighted detection towards certain classes and usage patterns, potentially affecting gender comparisons if ABP penetration differed by sport where one gender predominates. Moreover, different ADRVs may vary by gender and sport but are not decomposed here. Procedural elements and behavioural factors could introduce differential verification probabilities.

The sport composition of the sample is important: If women are overrepresented in sports where S1 use is at once more advantageous and surveilled (e.g., bodybuilding competitions with high surveillance), proportions by class might reflect sport ecology rather than gender per se. There is also the potential for age and an athlete’s career stage to be confounding variables, with differences in debut age and peak performance windows by sport (and sex), where younger cohorts might have different risk appetites or supplement practices. Access to medical oversight and the quality of nutritional counseling may also vary by sex and sport, which would impact both the decision-making quality and TUE use for valid medical conditions that overlap with S4 agents (e.g., selective estrogen receptor modulators following injury).

Equity and ethical concerns are very important. There is a clear need for gender-responsive prevention efforts to reduce incentives and chances for the use of doping substances. A “one-size-fits-all” message is not appropriate in this context, but there is also a need to be cautious in transmitting the appropriate message so that gender stereotypes are not being reinforced. For example, regarding “weight-control agents” as health-endangering rather than an existential failure can have more of an effective appeal among women participating in aesthetic and weight-class sports, whereas interventions targeting coping/anxiety and sleep management can reduce stimulant and cannabis temptations for men without further stigmatizing help-seeking. It is essential to guarantee that, in male and female programs, sanctioning, access to education, and medical supervision are equivalent between genders, in order to prevent structural drivers of inequality.

For policy and practice, three courses of action are identified. First, incorporate mixed methods into normal surveillance: coupling sport-stratified quantitative tracking of these outcomes with confidential interviews and validated psychometrics on motives, risk perception, and team norms can reveal decision pathways otherwise opaque in laboratory data. Second, further develop risk-based test planning with dynamic indicators: competition schedules, known weight-cut durations, cycles of training camp periods, and transfer/contract windows where performance anxiety skyrockets. Third, fund ASP capacity-building to limit accidental violations, especially S4/S5 exposure from inadequately vetted supplements or off-label prescribing; structured checklists and decision aids tailored for sport and sex can minimize mistakes.

Subsequent research should move beyond the current focus of between-gender heterogeneity to examine within-gender heterogeneity. The clustering of female S1 findings in one sport indicates that the findings should not be overgeneralized; micro-epidemics can prevail within a competitive circuit. Associating the ADRV timing with training mesocycles, travel loads, and injury episodes may uncover windows of increased vulnerability. Longitudinal cohorts that follow athletes with year-over-year observation for junior-to-senior transition may help to disentangle changes by sex in norms and incentives. From a methodological standpoint, reconciling denominators and accounting for list/technology changes would improve inference around actual behavioral trends. Theme-wise, we need to look further at the motives: cultural scripts around female/masculine identity and compliance, body image pressures, and financial insecurity may well mediate perceived benefits/harms in a way not covered by the biological rationale alone. Third, the inclusion of evaluation components in context-specific programming can provide causal evidence of which messages change intentions and by which delivery mechanisms men and women are best reached.

Finally, this study of ADAS 19-year records points to a stable gender gap in ADRVs. This gap is consistent with international studies and can be plausibly explained by intertwined biological, social, and sport-ecological factors. Such complexity directs is to adopt a precise prevention approach rather than generic one, and that approach must be aimed at protecting health and fairness while reducing unintended harm.