Age-stratified hormonal profiles in adolescent and young adult PCOS/PMOS: higher mid-adolescent AMH largely...

Polycystic ovary syndrome (PCOS) - for which a global multistakeholder consensus has recently proposed the name polyendocrine metabolic ovarian syndrome (PMOS) - shows phenotypic variation across adolescence that is poorly characterised in Turkish populations, and because puberty raises both body mass index (BMI) and reproductive hormones in parallel, age and BMI effects are potentially confounded. Within a single-centre retrospective cohort of clinician-assigned PCOS/PMOS (January 2024-February 2026), we described how hormonal, anthropometric and clinical profiles differ across age and whether these differences are explained by the parallel BMI gradient; because the cohort has no healthy controls, all analyses describe intra-PCOS variation and cannot establish PCOS-specificity. Of 1,228 patients with a clinician-assigned PCOS diagnosis among 11,324 adolescent visits, 878 met inclusion criteria (94.5% fulfilled ≥2 Rotterdam criteria post hoc; 40.8% of those ≤19 years met the stricter 2023 ESHRE adolescent criteria). Early-follicular-phase samples (cycle days 2-5; progesterone withdrawal in anovulatory patients) were analysed across four age groups (≤15, n=100; 16-17, n=217; 18-19, n=416; ≥20, n=145), with eight pre-specified sensitivity analyses. Overweight/obesity rose across age groups (35.0%→66.9%) but BMI correlated only weakly with the primary hormonal markers. Oligomenorrhoea/amenorrhoea predominated (78.0%); hirsutism (mFG ≥8) affected 13.0%. AMH (ϵ²=0.203, 95% CI 0.160-0.256), LH/FSH ratio (ϵ²=0.098), total testosterone (ϵ²=0.064) and DHEA-S (ϵ²=0.094) all reached their highest median values in mid-adolescents (16-17 years; all p<0.001). The higher mid-adolescent AMH values were the most robust signal, persisting after BMI-adjusted multivariable median regression and within both the strict Phenotype A and the non-hyperandrogenic Phenotype D (ϵ²=0.197) subsets. Total testosterone was the only significant correlate of hirsutism (OR 1.070 per ng/dL, 95% CI 1.055-1.086; AUROC 0.818, optimism-corrected 0.815) after age and BMI adjustment; because SHBG and free androgen index were unavailable, this reflects the strongest correlate within the available biochemical panel rather than a complete androgen assessment. Oestradiol showed the largest age-related effect overall (ϵ²=0.390), consistent with progressive pubertal oestrogenisation. In this cross-sectional, control-free cohort, BMI did not materially explain the age-related hormonal variation; findings are hypothesis-generating regarding pubertal (HPO-axis) mechanisms and do not demonstrate within-individual maturation or PCOS-specificity. The P5-P95 distributions represent intra-PCOS benchmarks, not population norms.