The antiprogesterone mifepristone in combination with a suitable prostaglandin provides an effective method for induction of abortion in early pregnancy up to 63 days of gestation. The combination of 600 mg mifepristone followed by 1 mg of gemeprost vaginal pessary 48 h later is one of the standard regimens in practice, which is registered in several countries in Europe. In 1995, we reduced the doses for both mifepristone and gemeprost to 200 mg and 0.5 mg respectively, as this was shown to decrease significantly the incidence of side effects whilst maintaining a high efficacy. In this article, we report our experience with this regimen in routine clinical practice by analysing 3161 consecutive medical abortions retrospectively. Twelve case notes (0.4%) were not available, and for 310 (9.8%) women, the outcome was not known with certainty as they did not return for their follow up visit. Of the remaining 2839 women, 2732 (96.2%) had a complete abortion following their treatment. One-hundred-two (3.6%) women required an evacuation of the uterus: for incomplete abortion in 63 (2.2%) and ongoing pregnancy in 39 (1.4%). Three women had to undergo surgery for ectopic pregnancies. The surgical intervention rate was significantly higher at gestation of >49 days compared to < or = 49 days (5.7% vs. 2.6%, p = 0.002) and at >56 days than among those at < or = 56 days (6.7% vs. 3.1%; p <0.001). However, for incomplete abortion a significant increase was only seen at gestation >49 days compared to < or = 49 days (3% vs. 1.6%, p = 0.017). The incidence of ongoing pregnancies increased significantly only after 56 days of gestation compared to < or = 56 days (3.8% vs. 0.9%; p <0.001). Parity was related to the outcome with parous women having significantly more incomplete/ongoing abortions compared to nulliparous women (5.4% vs. 2.0%; p <0.001), although parous women did present earlier in pregnancy for termination than nulliparous women (p = 0.01). The incidence of complications was low: 165 (5.8%) women were given antibiotics for presumed genital infection and severe haemorrhage occurred in 11 (0.4%) women, of whom only two required blood transfusion. In summary, the recommended regimen with the reduced doses of mifepristone and gemeprost is highly effective, meeting the anticipated efficacy with a complete abortion rate of >95%. We have concluded from the data that gestation and parity are strong predictors for clinicians to anticipate the probability of a successful medical termination of pregnancy.