ABSTRACT
RESEARCH QUESTION: How does use of a levonorgestrel-releasing intrauterine system (LNG-IUS) in infertile women with endometrial hyperplasia without atypia affect endometrial hyperplasia regression and pregnancy rates compared with oral medroxyprogesterone acetate (MPA)? DESIGN: This prospective cohort study included 215 infertile women with an indication for assisted reproductive technology (ART) and a diagnosis of endometrial hyperplasia without atypia. Endometrial hyperplasia was diagnosed by hysteroscopic endometrial biopsy. At the time of first- and second-line treatment, patients were offered therapy with either oral MPA 10 mg daily or LNG-IUS. Follow-up biopsies were scheduled after 90 days' treatment. After endometrial hyperplasia regression, patients were admitted to IVF/intracytoplasmic sperm injection (ICSI) cycles. RESULTS: Baseline characteristics and confounders including age at diagnosis, body mass index and duration of infertility did not differ between LNG-IUS users and control participants and were accounted for using propensity score weighting. Endometrial hyperplasia regression rate at first follow-up was higher in the LNG-IUS group than the oral progestins group (28/28, 100% and 110/187, 58.8%; P < 0.001), while that after second-line treatment was comparable between the two groups (89/91, 97.8% and 122/124, 98.4%; P = 0.22). Clinical pregnancy rate, miscarriage rate and cumulative live birth rate following ART in patients ever receiving LNG-IUS were similar to those of patients receiving only MPA (34% versus 39.5%, 22.6% versus 34.7% and 26.4% versus 25.8%). CONCLUSION: Endometrial hyperplasia regression is greater in women receiving LNG-IUS compared with oral MPA, while live birth rates following ART are comparable between the two groups. The use of LNG-IUS does not jeopardize the chances of pregnancy in women seeking fertility treatment.