Between November 1988 and December 1993, 100 patients with a common, unruptured ectopic pregnancy were treated with 1 mg/kg injection of intratubal methotrexate under transvaginal sonographic control. Patients were not excluded from this series on the basis of the size of the adnexal mass, the term of ectopic pregnancy or initial beta-human chorionic gonadotrophin (HCG) concentrations. Patients were excluded following uncertain diagnosis, signs of a ruptured ectopic pregnancy, or a significant haemoperitoneum on ultrasound scans. The mean age of the patients was 29.5 years (range 20-41). The mean gestational age and initial HCG concentration were 7.5 weeks (5-11) and 11,614 mIU/ml (192-105,000 respectively). Of the 100 patients, 22 (22%) had an ectopic pregnancy with active cardiac activity. Complete resolution was obtained in 78 out of these 100 ectopic pregnancies. Of these, 66 patients (85%) needed only one intratubal methotrexate injection, and 12 patients (15%) required a second i.m. methotrexate injection of 1 mg/kg. In this study, local treatment with one single intratubal methotrexate injection was successful in only 66% of patients. The mean resolution time for reduction of beta-HCG concentrations was 23.5 days (range 7-40). There was no statistically significant correlation between initial beta-HCG concentrations and outcomes after methotrexate treatment of ectopic pregnancy in our study. Where embryonal heart beats were observed, the success rate of the procedure was 40.9% (nine out of 22 cases). In the absence of cardiac activity, or when ultrasound examination showed no embryo, the success rate achieved was 84.6% (66 out of 78 cases) (P < 0.01). In all, 34 patients were considered to be incompletely cured after only one intratubal methotrexate injection: 12 patients required a second i.m. injection, a stagnation of beta-HCG concentrations was observed in 15 patients, abdominal pain occurred in six patients, and one patient suffered tubal rupture with haemoperitoneum. A total of 22 patients required secondary surgical management (salpingectomy). No biochemical or clinical side-effects of methotrexate treatment occurred. Tubal alteration ascribable to methotrexate injection occurred in one patient in our study. Out of 75 patients in this series who wished to conceive, 21 (28%) became pregnant within 1 year with the following outcomes: 11 pregnancies at term, three miscarriages, one induced abortion and six recurrent ectopic pregnancies (four occurred on the same side). Our findings suggest that treatment of common unruptured ectopic pregnancy without prior selection of patients, by a single intratubal methotrexate administration was associated with a 66% success rate. This was dependent only on the presence of embryonal heart beats and there was no correlation between the success rate and initial beta-HCG concentrations. Successful outcome after methotrexate administration for ectopic pregnancy could be perfected by way of an improved selection of patients based on inactive embryonal hearts and absence of a visualized embryo.