In order to analyse the relationship between grade of tubal damage and pregnancy rate following in-vitro fertilization (IVF) treatment, 50 consecutive couples with isolated tubal infertility who underwent IVF without previous reconstructive surgery were studied. The couples had tried to conceive for 2.5-5.6 years (mean 3.9 years) and had gone through a complete investigation for infertility. All women fulfilled the standard criteria for normal ovulatory cycles. All males had normal spermiograms. Tubal function was in each case evaluated by hysterosalpingography and laparoscopy. Grade of tubal damage was classified on a scale of I-IV. Patients with milder tubal damage (groups I-II) showed considerably higher take-home baby rate per started cycle (48%) than patients with severe tubal damage (group IV; 6%) (P < 0.001) even though patients in group IV underwent more treatment cycles than patients in groups I-II (2.7 versus 1.8; P < 0.01). In the first treatment cycle nine out of 12 patients in groups I-II became pregnant as compared with only one out of 16 in group IV (P < 0.01). As indicators of lower ovarian response, the women in group IV needed a higher number of gonadotrophin ampoules per cycle (P < 0.001) and exhibited lower pre-ovum retrieval oestradiol concentrations (P < 0.001). Oocyte retrievals per started cycle (I-II: 100%; III: 98%; IV: 84%) did not differ significantly between the groups, neither did pre-embryo replacements (I-II: 95%; III: 80%; IV: 73%) nor pre-embryos per replacement (I-II: 3.0; III: 2.9; IV: 2.8). The present results concerning the role of tubal damage grade for the outcome of IVF treatment should be considered in patient counselling. Further studies are needed to elucidate the mechanisms by which the extent of tubal damage may impair ovarian function and reduce IVF outcome.