Postoperative maintenance levonorgestrel-releasing intrauterine system and endometrioma recurrence: a randomized controlled study.

BACKGROUND According to 3 randomized trials, the levonorgestrel-releasing intrauterine system significantly reduced recurrent endometriosis-related pelvic pain at postoperative year 1. Only a few studies have evaluated the long-term effectiveness of the device for preventing endometrioma recurrence, and the effects of a levonorgestrel-releasing intrauterine system as a maintenance therapy remain unclear. OBJECTIVE The objective of the study was to evaluate whether a maintenance levonorgestrel-releasing intrauterine system is effective for preventing postoperative endometrioma recurrence. STUDY DESIGN From May 2011 through March 2012, a randomized controlled trial including 80 patients with endometriomas undergoing laparoscopic cystectomy followed by six cycles of gonadotropin-releasing hormone agonist treatment was conducted. After surgery, the patients were randomized to groups that did or did not receive a levonorgestrel-releasing intrauterine system (intervention group, n = 40, vs control group, n = 40). The primary outcome was endometrioma recurrence 30 months after surgery. The secondary outcomes included dysmenorrhea, CA125 levels, noncyclic pelvic pain, and side effects. RESULTS Endometrioma recurrence at 30 months did not significantly differ between the 2 groups (the intervention group, 10 of 40, 25% vs the control group 15 of 40, 37.5%; hazard ratio, 0.60, 95% confidence interval, 0.27-1.33, P = .209). The intervention group exhibited a lower dysmenorrhea recurrence rate, with an estimated hazard ratio of 0.32 (95% confidence interval, 0.12-0.83, P = .019). Over a 30 month follow-up, the intervention group exhibited a greater reduction in dysmenorrhea as assessed with a visual analog scale score (mean ± SD, 60.8 ± 25.5 vs 38.7 ± 25.9, P < .001, 95% confidence interval, 10.7-33.5), noncyclic pelvic pain visual analog scale score (39.1 ± 10.9 vs 30.1 ± 14.7, P = .014, 95% confidence interval, 1.9-16.1), and CA125 (median [interquartile range], -32.1 [-59.1 to 14.9], vs -15.6 [-33.0 to 5.0], P = .001) compared with the control group. The number-needed-to-treat benefit for dysmenorrhea recurrence at 30 months was 5. The number of recurrent cases requiring further surgical or hormone treatment in the intervention group (1 of 40, 2.5%, 95% confidence interval, -2.3% to 7.3%) was significantly lower than that in the control group (8 of 40, 20%, 95% confidence interval, 7.6-32.4%; P = .031). CONCLUSION Long-term maintenance therapy using a levonorgestrel-releasing intrauterine system is not effective for preventing endometrioma recurrence.