The time-dependent alterations in the serum concentrations of ethinyl estradiol, gestodene, and 3-keto-desogestrel during treatment with 30-mu-g of ethinyl estradiol + 75-mu-g of gestodene or 30-mu-g of ethinyl estradiol + 150-mu-g of desogestrel were investigated during 12 months. The levels of gestodene and 3-keto-desogestrel increased between days 1 and 21 of each cycle, reaching maximal levels during the third and sixth cycles. The serum concentrations of gestodene were fourfold to fivefold higher than those of 3-keto-desogestrel. The ethinyl estradiol levels increased significantly between days 1 and 10 during each cycle and were significantly higher by 70% during intake of ethinyl estradiol/gestodene compared with ethinyl estradiol/desogestrel, although the dose was identical. Intake of gestodene, in addition to 35-mu-g of ethinyl estradiol + 2 mg of cyproterone acetate, caused a rise in ethinyl estradiol levels. During treatment with ethinyl estradiol/gestodene and an additional 150-mu-g of levonorgestrel, there was a continuous increase in gestodene levels, although sex hormone-binding globulin level did not change. During treatment with 30 or 35-mu-g of ethinyl estradiol and 75-mu-g of gestodene, 150-mu-g of desogestrel, or 2 mg of cyproterone acetate, there were large intraindividual and interindividual variations in the steroid levels and ratios of estrogen: progestogen levels. There was no correlation with the occurrence of intermenstrual bleedings. It is concluded that ethinyl estradiol and nortestosterone derivatives may inhibit steroid-metabolizing enzymes in the liver, which results in a rise in the serum levels of contraceptive steroids. The cause of the large intraindividual variations is as yet unknown, but it is probably from changes in steroid metabolism.