Twenty-three patients with laparoscopically diagnosed endometriosis and pelvic pain were allocated randomly to treatment with cyproterone acetate 27 mg plus ethinyl estradiol 0.035 mg/day (11 patients) or danazol 600 mg/day (12 patients). All women received treatment for 6 months, except for one in the cyproterone group who suspended treatment for nonmedical reasons and was excluded from analysis of the results. The clinical condition and pain symptoms were monitored in all patients for 1 year after treatment suspension. The intensity of pelvic pain at diagnosis, during treatment, and at follow-up was evaluated by a multidimensional verbal score and an analogue scale. At the end of treatment, a repeat laparoscopy was performed in those patients who agreed (four in the cyproterone group, five in the danazol group); the results showed a partial regression of endometriotic lesions in both groups, with no significant differences between them. Dysmenorrhea disappeared in all patients during treatment. At 6 months after suspension, dysmenorrhea recurred in 66% of the cyproterone group and 58% of the danazol group, and at 1 year in 89 and 92%, respectively. Intermenstrual pelvic pain improved markedly during treatment in both groups; 6 months after treatment withdrawal it was present in four cyproterone subjects and four danazol group patients, whereas after 1 year, only one woman in the danazol group did not have this symptom. Deep dyspareunia was less affected by treatment, and 6 months later had recurred in all the women.(ABSTRACT TRUNCATED AT 250 WORDS)