A 2026 study using the most comprehensive steroid profiling method yet applied to cannabis research found that young male cannabis users had significantly higher testosterone, DHT, and androstenedione than non-users. However, the cross-sectional design means we cannot determine whether cannabis causes these hormonal differences or whether men with naturally higher androgens are simply more likely to use cannabis.

With approximately 150 million cannabis users worldwide and legalization accelerating across jurisdictions, the question of whether cannabis use alters male sex hormones carries real clinical weight. Young men frequently ask clinicians whether cannabis affects their testosterone, fertility, or reproductive health. Until now, most research offered only a narrow testosterone measurement. This study profiles 70 steroid compounds in biomarker-confirmed cannabis users, providing the most granular hormonal picture to date and identifying potential pathway-specific effects that reshape how we should frame this clinical conversation.

The relationship between cannabis use and male sex hormones has been investigated for over fifty years, with early studies in the 1970s suggesting that cannabis lowered testosterone and more recent large cohorts pointing in the opposite direction. This study represents the most analytically ambitious attempt to resolve the question, drawing on a pre-existing Swiss military cohort with biomarker-confirmed cannabis exposure and applying liquid chromatography-tandem mass spectrometry (LC-MS/MS) to profile 70 steroid compounds across the gonadal, adrenal, and progesterone biosynthetic pathways. The rationale for such broad profiling is that testosterone alone cannot capture the complexity of steroidogenesis, and that pathway-level analysis may reveal whether cannabinoids act selectively on testicular versus adrenal hormone production.

The key findings were that bioactive gonadal androgens, specifically testosterone (p=0.002), androstenedione (p=0.008), and dihydrotestosterone (p=0.029), were significantly higher in cannabis users, while adrenal androgens including C11-oxygenated compounds, cortisol, and cortisone showed no significant differences. Two progesterone metabolites were markedly elevated, and one of these, 5-beta-dihydroprogesterone, showed a dose-response relationship with cannabis use intensity. The primary limitations are the cross-sectional design, the small sample of 47 individuals per group, the absence of gonadotropin measurements, and the lack of explicit correction for multiple comparisons across the 70-compound profile. The authors appropriately state that prospective and mechanistic studies are needed to determine whether cannabis causes these hormonal differences.

Cannabis Use Is Consistently Associated With Higher Testosterone in Young Men, But We Still Cannot Say Why

Across multiple independent cohorts on two continents, young men who use cannabis keep showing up with higher testosterone than those who do not. The pattern is real. But after decades of research, we still cannot answer the most fundamental question: does cannabis raise testosterone, or do men with higher testosterone simply use more cannabis? This Swiss study brings genuinely new tools to the problem. Rather than measuring testosterone alone, the researchers profiled 70 steroid compounds using rigorously validated mass spectrometry and confirmed cannabis exposure with blood biomarkers rather than relying on self-report. That combination of analytical breadth and exposure verification is the strongest feature of this work, and it moves the field beyond a narrow, single-hormone lens. The finding that gonadal androgens (testosterone, androstenedione, and DHT) are elevated while adrenal markers remain unchanged provides the most specific hypothesis yet: that cannabinoids may act selectively on testicular steroidogenesis rather than causing generalized endocrine disruption. The dose-response signal for 5-beta-dihydroprogesterone, a progesterone metabolite that scaled with cannabis use intensity, further strengthens the biological plausibility of the association.

However, sophisticated measurement does not overcome a fundamental design constraint. This is a cross-sectional snapshot: blood drawn at a single point in time, without any knowledge of what these men’s hormones looked like before they started using cannabis. Finding that firefighters have higher rates of smoke inhalation does not tell you whether smoke causes firefighting. You need to know which came first. Similarly, without pre-exposure baselines or longitudinal follow-up, we cannot distinguish a pharmacological effect of cannabis from a behavioral predisposition driven by pre-existing hormonal profiles. The study also tested 70 steroid analytes without applying formal correction for multiple comparisons, which matters practically: if you flip a coin 70 times and count anything below one-in-twenty odds as meaningful, you expect about three or four “significant” results from chance alone. Several of the reported p-values hover near that threshold, making their durability uncertain. The absence of gonadotropin measurements (LH and FSH) is another critical gap; without them, the authors cannot distinguish whether the androgen elevation originates from pituitary stimulation or from direct effects on Leydig cells in the testes.

When patients ask me whether cannabis is affecting their hormones, I can now point to a consistent pattern across several well-conducted studies showing somewhat higher testosterone in cannabis-using young men. But I am honest about the limits of what that pattern means: we do not know the direction of the effect, we do not know the clinical significance of the magnitude, and we have no evidence linking these differences to fertility, cardiovascular, or other health outcomes. For colleagues, this study earns attention for its methodological rigor and its pathway-level analysis, but it should be treated as hypothesis-generating, not confirmatory. For policymakers, it underscores an urgent case for investing in prospective, longitudinal research on cannabis and reproductive endocrinology, rather than drawing premature conclusions from a single cross-sectional dataset. The pattern of findings is now sufficiently consistent that the association is unlikely to be artifactual. But consistency of association across observational studies does not substitute for a single well-designed prospective study with appropriate biological measures. Replication of a cross-sectional finding is not the same as longitudinal confirmation.

This study sits at a critical juncture in the cannabis-and-hormones research arc. Early studies from the 1970s reported lower testosterone in cannabis users, while the most recent and largest cross-sectional cohorts from Denmark and the United States have consistently shown the opposite. The present work confirms and extends the newer directional findings with superior analytical methods and biomarker-verified exposure classification. It does not, however, resolve the fundamental question of causality that has persisted across all of these studies. What it does contribute is the first pathway-level evidence suggesting selectivity of the gonadal versus adrenal steroidogenic response, a finding that should shape the design of the next generation of research.

From a pharmacological standpoint, clinicians should note that the study population was not using medical cannabis products with defined cannabinoid concentrations, and no information was available on concomitant substance use, dietary patterns, or exercise habits that could influence steroid metabolism. The absence of clinical outcome data means there is no basis for reassuring or alarming patients about the health consequences of these hormonal differences. For clinicians counseling young men about cannabis, the most appropriate recommendation remains to acknowledge that hormonal associations have been reported in multiple studies, to note that the clinical significance of these associations is not yet established, and to encourage patients to discuss hormonal health openly as part of comprehensive primary care rather than relying on population-level findings to predict individual risk.

This is an original cross-sectional observational study published in a peer-reviewed Nature Portfolio journal, positioned in the lower-to-middle tier of the evidence hierarchy. Cross-sectional designs generate associations at a single time point and cannot establish causality, temporality, or directionality. The single most important inference constraint is that elevated androgens in cannabis users may reflect pre-existing biological differences rather than any pharmacological effect of cannabis.

This study confirms and extends findings from a large Danish military cohort (Gundersen et al.) that reported approximately 7% higher testosterone in cannabis users, and from a US cross-sectional study that found elevated testosterone in THC-positive participants regardless of use frequency. It contradicts the earlier Kolodny et al. (1974) finding of decreased testosterone, though that study used a far smaller sample and different analytical methods. The novelty here is the extended steroid profiling that reveals pathway selectivity, a dimension absent from all prior work. However, no prospective or longitudinal study has yet tested the directionality of the cannabis-androgen association, meaning this entire body of literature, despite its consistency, remains at the associational level of evidence.

The most consequential analytical choice was conducting univariate statistical tests across 70 steroid analytes without applying a formal false discovery rate correction. Had a Benjamini-Hochberg or Bonferroni adjustment been applied, some of the associations at the margins of significance, such as DHT (p=0.029) and 17-alpha-hydroxyprogesterone (p=0.03), might not have survived correction. The core testosterone finding (p=0.002) would likely remain significant under most correction schemes, but the broader pathway-level narrative depends substantially on whether the progesterone metabolite and DHT findings hold up. Additionally, the OPLS-DA multivariate model yielded a low predictive ability (Q-squared of 0.19), suggesting that steroid profiles alone discriminate cannabis users from controls only modestly, which tempers the claim of a distinct steroidomic signature.

The most likely overinterpretation is that cannabis causes elevated testosterone in young men. This study shows a statistical association between cannabis use and higher gonadal androgens, but its cross-sectional design means the direction of the relationship is entirely unknown. It is biologically plausible that men with pre-existing higher androgen levels are more inclined toward risk-taking behaviors, including cannabis use, a form of reverse causality that this design cannot detect or exclude. Equally important, higher testosterone is not inherently beneficial. The paper provides no data on fertility outcomes, cardiovascular risk, or any clinical endpoint, so interpreting the findings as evidence of a health benefit or a health risk would exceed what the evidence supports.

This study contributes the most analytically comprehensive steroid profile yet applied to biomarker-confirmed cannabis exposure in young men, confirming the consistent association between cannabis use and elevated gonadal androgens. It does not establish causality, directionality, or clinical significance of the observed hormonal differences. For clinical practice, these findings reinforce the importance of discussing hormonal health with cannabis-using patients while awaiting the prospective, longitudinal evidence needed to determine whether the association reflects a genuine pharmacological effect.

Does cannabis increase testosterone?

Multiple studies, including this one, have found that young men who use cannabis tend to have somewhat higher testosterone levels than non-users. However, none of these studies can tell us whether cannabis causes the increase. It is equally possible that men with naturally higher testosterone are more likely to use cannabis. Until prospective studies track hormone levels before and after cannabis use, we cannot make a definitive causal statement.

Should I be concerned about cannabis affecting my hormones?

The hormonal differences observed in this study, while statistically significant, have not been linked to any specific health outcome, whether positive or negative. The study did not measure fertility, sexual function, cardiovascular health, or any clinical endpoint. If you have concerns about how cannabis might be affecting your health, discuss them with your physician, who can evaluate your individual hormonal profile in context.

Do these findings apply to women or older adults?

No. This study was conducted exclusively in young Swiss men aged 18 to 23. The findings cannot be generalized to women, older men, or individuals with different health conditions. Steroid metabolism differs substantially between sexes and across age groups, so separate research would be needed to understand cannabis’s hormonal associations in other populations.

What makes this study different from previous cannabis and testosterone research?

This study is distinguished by two features: it confirmed cannabis exposure using blood biomarkers (THC and THC-COOH in serum) rather than relying on self-report, and it measured 70 different steroid compounds rather than just testosterone. This broader profiling allowed the researchers to observe that gonadal androgens were elevated while adrenal hormones were not, a level of pathway specificity not available in prior work.

Have thoughts on this? Share it: