Coffee drinkers had more testosterone in circulation

In men, each additional cup per day was associated with 0.29 nmol/L more total testosterone and 0.57 nmol/L more sex hormone-binding globulin. Bioavailable testosterone rose by 0.08 nmol/L. Free testosterone fell by 0.01 nmol/L, and the free androgen index, which estimates how much testosterone is circulating unbound, fell by 0.36.

Compared as groups rather than per cup, the direction held. Moderate and high consumers had significantly higher total testosterone than non-consumers, and high consumers higher again than low and moderate ones. The associations survived adjustment for BMI, education, smoking, physical activity and alcohol.

SHBG is the protein that binds testosterone and holds it out of circulation, and it rose alongside the total. More testosterone is present, and more of it is spoken for. Bioavailable testosterone, which covers the free fraction plus the loosely albumin-bound portion, still rose slightly. Only the free fraction and the ratio-based free androgen index moved the other way.

"What stood out in our findings was a distinct hormonal signature that didn't disappear even after we took into account BMI and lifestyle factors, with several of these associations differing between men and women," said lead author Luca Verroest, a doctoral researcher at Oulu.

In women without polycystic ovary syndrome, the picture was narrower. SHBG rose by 1.27 nmol/L per cup per day while free testosterone and the free androgen index fell, with no significant association for total or bioavailable testosterone. Women with PCOS, or with PCOS status unrecorded, were excluded from the hormone analysis.

BMI, waist circumference and hip circumference were similar across the intake groups. Body composition was not. Higher intake tracked with lower body fat percentage and fat mass, a smaller visceral fat area and greater skeletal muscle mass, and the pattern held after participants at BMI extremes were excluded.

Metabolic markers pointed the same way, more clearly in men: lower fasting insulin, lower insulin at 60 and 120 minutes during an oral glucose tolerance test, lower modelled insulin resistance and higher insulin sensitivity. Both sexes showed lower branched-chain amino acids, which earlier work has associated with insulin resistance and type 2 diabetes risk when chronically elevated. In women the associations were fewer, mainly a lower waist-to-hip ratio, lower high-sensitivity CRP and lower fasting insulin.

The study is cross-sectional, intake was self-reported, and only 70 participants drank no coffee at all. The cohort is a single homogeneous population in a country averaging roughly 11.8 kg of coffee per person per year. Most correlations were small, and reverse causation and confounding cannot be excluded.

The authors describe these as population-level endocrine patterns, not evidence that an extra cup produces a meaningful hormonal change in any individual. Animal work is under way, with human intervention studies the longer-term aim.