The effects of three different desogestrel-containing combined oral contraceptive preparations on lipid metabolism were compared with those of two levonorgestrel preparations. The following preparations were studied: (1) monophasic desogestrel (150/30), (2) monophasic desogestrel (150/20, containing 20 micrograms of ethinyl estradiol instead of 30 micrograms of ethinyl estradiol, (3) biphasic desogestrel, (4) monophasic levonorgestrel (150/30), and (5) triphasic levonorgestrel. The effects of these preparations were assessed on high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, apolipoprotein A-I, apolipoprotein B, the ratio of high-density lipoprotein cholesterol to low-density lipoprotein cholesterol, and the ratio of apolipoprotein A-I to apolipoprotein B after 3 months of treatment, and the percentage changes with regard to pretreatment were calculated. The monophasic desogestrel (150/30) and biphasic desogestrel preparations induced higher high-density lipoprotein cholesterol and apolipoprotein A-I levels than did their levonorgestrel-containing counterparts. Low-density lipoprotein cholesterol levels were increased in monophasic levonorgestrel and clearly decreased in the lowest ethinyl estradiol-containing monophasic desogestrel (150/20) and biphasic desogestrel preparations. Apolipoprotein B increased in all preparations. The antiatherogenic indexes (ratios of high-density lipoprotein cholesterol to low-density lipoprotein cholesterol and apolipoprotein A-I to apolipoprotein B were higher for monophasic desogestrel (150/30) and biphasic desogestrel than for comparable levonorgestrel-containing preparations. The differences seen between the desogestrel and levonorgestrel preparations can best be explained by the lower intrinsic androgenicity of 3-keto-desogestrel (active metabolite of desogestrel) than that of levonorgestrel.