Natural cycle frozen embryo transfer reduces obstetric complications: an inverse probability of treatment-weighted analysis.

OBJECTIVE: To compare obstetric complications between hormone replacement treatment (HRT) and natural cycle (NC) frozen embryo transfer (FET) using inverse probability of treatment weighting (IPTW) with generalized estimating equations (GEE), and to evaluate the impact of an institutional policy change actively adopting NC on clinical outcomes. DESIGN: Retrospective cohort study. SUBJECTS: Patients undergoing FET at a single center between September 2015 and September 2023. After exclusions, 11,873 cycles (HRT: 5,115; NC: 6,758) from 4,216 unique patients were analyzed. EXPOSURE: Endometrial preparation using HRT or NC, and a progressive shift in treatment policy from HRT-predominant to NC-predominant protocols over the study period. MAIN OUTCOME MEASURES: Live birth rate, miscarriage rate, and obstetric complications including cesarean section, hypertensive disorders of pregnancy (HDP), placenta accreta spectrum (PAS), and postpartum hemorrhage (PPH) in singleton live births. RESULTS: In the IPTW-GEE doubly robust analysis, the HRT group showed a significantly higher miscarriage rate (adjusted relative risk [aRR] 1.26, 95% confidence interval [CI] 1.14-1.41) and lower live birth rate (aRR 0.92, 95% CI 0.87-0.99) compared with NC. Among singleton live births, HRT was associated with significantly higher risks of HDP (aRR 1.56, 95% CI 1.21-2.02), PAS (aRR 5.13, 95% CI 3.25-8.12), and cesarean section (aRR 1.38, 95% CI 1.26-1.53). PPH ≥1,000 mL in vaginal deliveries was also significantly higher in the HRT group (aRR 2.74, 95% CI 2.08-3.63). Sensitivity analyses excluding low-dose aspirin users and restricting to matched luteal phase support confirmed these findings. In the policy analysis, each 10% increase in the annual NC proportion was associated with significant reductions in PAS (aRR 0.80, 95% CI 0.74-0.88), HDP (aRR 0.90, 95% CI 0.84-0.96), cesarean section (aRR 0.95, 95% CI 0.93-0.98), PPH ≥1,000 mL (aRR 0.92, 95% CI 0.86-0.99), and LGA (aRR 0.92, 95% CI 0.87-0.97) after adjusting for covariates. CONCLUSION: NC is associated with significantly lower risks of miscarriage and major obstetric complications compared with HRT in singleton deliveries. The institutional policy shift toward NC was associated with significant reductions in adverse outcomes, supporting a shift toward treatment strategies prioritizing NC.