Baseline serum AMH‑to‑FSH ratio, menstrual pattern, and risk of intra‑uterine adhesion recurrence: a prospective multi‑state Markov model study

BACKGROUND: Intra-uterine adhesion (IUA) frequently recurs after hysteroscopic adhesiolysis, undermining menstruation and fertility. We aimed to determine whether baseline serum anti‑Müllerian hormone (AMH)‑to-follicle‑stimulating hormone (FSH) ratio and menstrual pattern independently and jointly predict IUA severity transitions after adhesiolysis using a multi‑state Markov model. METHODS: We conducted a single-center prospective study of 210 women (18-45 years) undergoing first adhesiolysis for moderate-to-severe IUA. Baseline serum AMH and FSH were measured and menstrual pattern was classified as normal, hypomenorrhea, or amenorrhea. Hysteroscopy at 3, 6, 12, and 24 months defined four severity states (none, mild, moderate, severe) and two absorbing outcomes (pregnancy ≥ 12 weeks, repeat surgery). Continuous-time multi-state Markov models estimated transition-specific adjusted hazard ratios (HRs) with 95% confidence intervals (CIs), controlling for age, body-mass index (BMI), baseline adhesion severity, and adjuvant therapy. Internal validation used 200-bootstrap c-index and calibration. RESULTS: Of 210 participants, 83 women (39.5%) had any recurrence and 28 (13.3%) progressed to severe adhesions. Pregnancy ≥ 12 weeks occurred in 54 (25.7%) and repeat surgery in 36 (17.1%). Each one-standard-deviation decrease in AMH/FSH increased hazards from mild to moderate (adjusted HR 1.81, 95% CI 1.29-2.54) and from moderate to severe (HR 2.15, 95% CI 1.41-3.28), with a non-significant trend from none to mild (HR 1.20, 95% CI 0.95-1.52). Amenorrhea versus normal was associated with higher hazards for none to mild (HR 1.88, 95% CI 1.23-2.99) and mild to moderate (HR 1.55, 95% CI 1.10-2.18). An interaction suggested that amenorrhea amplified the adverse effect of a low AMH/FSH ratio on progression to severe disease (HR 1.39, 95% CI 1.03-1.88). The model showed good performance, with bootstrap‑corrected C‑index 0.78 (95% CI 0.74-0.82), optimism‑corrected C‑index 0.76, and optimism‑corrected calibration slope 0.94. CONCLUSIONS: Baseline endocrine status summarized by the AMH/FSH ratio and menstrual pattern independently and jointly predict IUA dynamics after adhesiolysis. A biomarker-based, multi-state risk tool may help personalize severity-anchored surveillance and re-intervention after adhesiolysis.