Oncologic and Reproductive Outcomes After Fertility-Sparing Treatments for Endometrial Hyperplasia with Atypia: A Systematic Review and Meta-Analysis

<b>Background</b>: Atypical endometrial hyperplasia is a precursor of endometrial carcinoma, increasingly diagnosed in reproductive-aged women. Standard hysterectomy may constitute overtreatment. Fertility-sparing approaches, oral or local progestins, hysteroscopic resection, and combined regimens are widely used, though evidence largely derives from early-stage carcinoma. <b>Methods</b>: CENTRAL, EMBASE, Scopus, Web of Science, and PubMed were searched from inception to 13 April 2025. Eligible studies included premenopausal women with AEH treated with oral progestins, levonorgestrel intrauterine devices (LNG-IUDs), hysteroscopic resection, or combination regimens (e.g., oral progestins + metformin, LNG-IUD + GnRH analogues). Random-effects meta-analyses with multilevel modeling were applied. Risk of bias was assessed using RoB-2 and ROBINS-I; certainty of evidence was graded with GRADE-PRO. <b>Results</b>: Forty-nine studies (2313 women) were included. The pooled complete response (CR) rate was 85% (95% CI 80-89%). LNG-IUDs resulted in 88% CR and oral progestins in 80%. Combination therapies achieved up to 95% CR, though data were limited. Hysteroscopic resection combined with hormonal therapy reported high CR rates (96-97%) but relied on small, heterogeneous cohorts with uncertain reproducibility. The pooled recurrence rate was 19% (95% CI, 13-25%), lower with LNG-IUDs (14%) compared with oral progestins (22%). No response occurred in 14% overall and was lower with LNG-IUDs (13%) than oral progestins (19%). Among women attempting conception, the pooled pregnancy rate was 41%, and the live birth rate was 30%. The mean time to achieve CR was 5.6 months, whereas the mean time to recurrence was approximately 33 months. <b>Conclusions</b>: LNG-IUDs provide the most effective and well-supported fertility-sparing treatment for AEH, ensuring high remission with low recurrence. Despite favorable numerical outcomes, the evidence for hysteroscopic resection is limited and inconsistent, not supporting its routine use. Combination regimens may improve response but require confirmation in larger studies.