A retrospective cohort study was set up to evaluate the effectiveness of conservative and radical surgery for tubal pregnancy towards subsequent fertility. Consecutive patients undergoing conservative or radical surgery for tubal pregnancy between January 1990 and August 1993 in two university hospitals were included in the study. Outcome measures were spontaneous intrauterine pregnancy (IUP) and repeat ectopic pregnancy (EP). Of the 135 patients analysed, 56 underwent conservative surgery and 79 underwent radical surgery. Patients treated with conservative surgery achieved a higher 3-year cumulative pregnancy rate than those treated radically (P < 0.001, log-rank test). In patients treated conservatively, there was only one spontaneous IUP in the period between 18 months and 3 years after the tubal pregnancy. In contrast, patients treated radically continued to conceive in this period. Multivariate analysis showed a fecundity rate ratio (FRR) of 1.9 [95% confidence interval (CI): 0.91 to 3.8] for IUP after conservative surgery in the first 18 months of follow-up. In patients with a history of bilateral tubal disease the FRR was 3.1 (95% CI: 0.76 to 12), whereas in patients without a history of bilateral tubal disease the FRR was 1.4 (95% CI: 0.13 to 16). The FRR for repeat EP was 2.4 (95% CI: 0.57 to 11). Our data indicate a beneficial effect of conservative surgery towards subsequent fertility that was not, however, statistically significant in the multivariate analysis. In view of these inconclusive data and the importance of this major health problem, randomized studies are required to assess whether conservative surgery really improves the fertility prospects of patients with tubal pregnancy.