Phenotype-specific progression of insulin resistance and hyperandrogenism in polycystic ovary syndrome: A 15-year longitudinal HOMA-IR and FAI trajectory analysis

초록

What is known already: Polycystic ovary syndrome is a heterogeneous disorder with substantial variability in metabolic risk across phenotypes. Insulin resistance and hyperandrogenism are central pathophysiological features and are strongly influenced by adiposity and fat distribution. Cross-sectional studies demonstrate worse metabolic profiles in obese PCOS, but long-term data on within-phenotype trajectories are limited. Body composition, particularly visceral adiposity, independently predicts cardiometabolic disease; however, extended longitudinal studies integrating insulin resistance, androgen indices, and body composition across PCOS phenotypes are scarce. This knowledge gap constrains timing and intensity of phenotype-specific preventive interventions. Participants/materials, setting, methods: Participants were women aged 16–49 years with PCOS attending a tertiary reproductive endocrinology clinic. Measurements included fasting and postprandial glucose and insulin, HOMA-IR, free androgen index, total and free testosterone, DHEA-S, AMH, LH, FSH, lipids, and HbA1c. Body composition was assessed by bioelectrical impedance, including fat mass, lean mass, visceral fat area, and body-fat percentage. Trajectories by BMI phenotype were analysed using repeated-measures ANOVA, correlation analyses, and multivariable linear regression. Main results and the role of chance: Over 15 years, the overweight/obese PCOS phenotype showed significant deterioration in insulin resistance, with mean HOMA-IR increasing from 2.84±0.45 at baseline to 3.68±0.61 at final follow-up (29.6% relative increase, p < 0.001). FAI in this phenotype rose from 8.2±3.1 to 12.4±4.7 (51.2% increase, p < 0.001), and the prevalence of FAI >10 increased from 33% to 67% (p < 0.001). Free testosterone increased from 1.02±0.36 to 1.68±0.51 pg/mL (p = 0.008). In contrast, the normal-weight phenotype (HOMA-IR 2.18±0.38 to 2.29±0.39, p = 0.42; FAI 4.1±2.1 to 4.8±2.4, p = 0.18) and the lean phenotype (HOMA-IR 1.95±0.35 to 2.05±0.37, p = 0.61; FAI 3.2±1.8 to 3.6±1.9, p = 0.51) showed no statistically significant metabolic change. Body-composition analysis demonstrated that obese women had a marked increase in fat mass from 27.4±6.2 kg to 35.2±8.1 kg (p < 0.001), visceral fat area from 98.3±24.5 cm² to 154.3±38.7 cm² (p < 0.001), and body-fat percentage from 38.2±5.1% to 45.3±6.2% (p < 0.001), with a modest decline in lean mass from 45.3±4.8 kg to 43.6±5.1 kg (p = 0.03). Normal-weight and lean phenotypes exhibited minimal body-composition change (all p > 0.05). HOMA-IR change correlated positively with fat-mass gain (r = 0.58, p < 0.001), and visceral adiposity independently predicted HOMA-IR progression (β = 0.041, p = 0.002), indicating that the observed trajectories are unlikely to be explained by chance alone. Wider implications of the findings: These findings support early, intensive insulin-sensitising and weight-management strategies in obese PCOS to prevent long-term metabolic deterioration, while lean and normal-weight phenotypes may benefit from more targeted, non-metabolic approaches. Integrating serial body-composition assessment with HOMA-IR and FAI can refine phenotype-based risk stratification and guide personalised PCOS management algorithms.