We studied the role of transvaginal ultrasonography and clinical factors in the diagnosis of pelvic adhesions in a population of 139 consecutive pre-menopausal non-pregnant women submitted to diagnostic and/or operative laparoscopy between February 1995 and November 1996. All patients underwent transvaginal ultrasonography and were interviewed within 2 days of their laparoscopy. The ultrasonographic impressions were then compared with the laparoscopic diagnosis. Patients were classified as having tuboperitoneal abnormalities if evidence of fimbrial, peritubal and/or peri-ovarian adhesions was encountered during surgery. The overall agreement between the ultrasound test result and the surgical findings was calculated using the kappa index. The adhesion of the ovary to the uterus, as evaluated by transvaginal ultrasonography, is most accurate in diagnosing pelvic adhesions (kappa = 0.5) in comparison with the other ultrasonographic findings and clinical parameters. According to the logistic regression equation that was obtained, the probability of the presence of pelvic adhesions varied between a minimum of 12% for patients with no risk factors to a maximum of 93% for patients with three risk factors (previous pelvic surgery and transvaginal ultrasound findings of blurring of the margins of the ovary and adhesion of the ovary to the uterus).