Objective: To determine the prevalence of histologically proven endometriosis in a subset of infertile women.

Design: Retrospective case series with electronic file search and multivariable logistic regression analysis.

Setting: Tertiary academic fertility center.

Patient(s): Two hundred twenty-one infertile women without previous surgical diagnosis for infertility with regular cycles (variation, 21-35 days) whose partners have a normal semen analysis.

Intervention(s): Diagnostic laparoscopy and, if necessary, operative laparoscopy with CO(2) laser excision.

Main outcome measurement(s): The prevalence of endometriosis and of fertility-reducing nonendometriotic tubal and/or uterine pathology.

Result(s): The prevalence of endometriosis was 47% (104/221), including stage I (39%, 41/104), stage II (24%, 25/104), stage III (14%, 15/104), and stage IV (23%, 23/104) endometriosis, and was comparable in patients with (54%, 61/113) and without (40%, 43/108) pelvic pain. The prevalence of fertility-reducing nonendometriotic tubal and/or uterine pathology was 29% in all patients (15% in women with and 40% in women without endometriosis). A multivariate logistic regression model including pain, ultrasound data, age, duration of infertility, and type of fertility was not or not sufficiently reliable for the prediction of endometriosis according to the revised American Fertility Society (rAFS) classifications I-II and rAFS III-IV, respectively.

Conclusion(s): Reproductive surgery is indicated in infertile women belonging to the study population, regardless of pain symptoms or transvaginal ultrasound results, since half of them have endometriosis and 40% of those without endometriosis have fertility-reducing pelvic pathology.