A randomized double-blind study of two combined oral contraceptives and two progestogen-only oral contraceptives was conducted using the same protocol at WHO Collaborating Centres for Clinical Research in Human Reproduction in Bombay and Ljubljana of the 518 women admitted to the trial, 123 received mestranol 50 micrograms + norethisterone 1mg (MES 50 + NET 1); 137 received ethinyl estradiol 30 micrograms + levonorgestrel 150 micrograms (EE 30 + LNG 150); 130 received norethisterone 350 micrograms/NET 350); and 128 received levonorgestrel 30 micrograms (LNG 30). At one year, between 52.6 and 61.0 percent of those recruited had discontinued oral contraceptive use for all reasons, and by two years, between 70.5 and 76.5 percent had discontinued the treatment. These rates did not differ between the four treatment groups. However, discontinuation rates for all medical reasons at one and two years, and at two years pregnancy rates and discontinuation rates for bleeding disturbances, were significantly lower in the EE/LNG preparation. The groups receiving the MES/NET, LNG and NET had similar pregnancy rates, discontinuation rates for all medical reasons and all bleeding disturbances. There were two ectopic pregnancies among the 22 pregnancies in the progestogen-only groups. Discontinuation because of headache, dizziness and other central nervous system symptoms were significantly more common in those receiving MES/NET compared to EE/LNG. In contrast, discontinuation for gastro-intestinal disturbances were significantly higher in the EE/LNG combined preparation. Bleeding disturbances in the first few cycles tended to be higher in NET than in the LNG group. The data suggest that greater consideration be given to the benefits and risks of including progestogen-only oral contraceptives in the family planning programmes of some countries.