Epidemiologic data support the notion that first-generation high-dose oral contraceptives (containing >80 mu g of estrogen) increased the incidence of thromboembolic events. The quantitative interpretation of these data is difficult because results were often confounded by life-style factors and inadequate diagnostic procedures. With the introduction of modern low-dose combination oral contraceptives, the incidence of thromboembolic events decreased markedly. Although all combined oral contraceptives induce statistically significant changes in hemostatic factors, these changes are generally within normal ranges, and their clinical significance is questionable. Overall, increased activity in hemostatic mechanisms appears to remain in balance. Progestin-only formulations seem to affect hemostatic parameters to a much lesser degree, and their use has not led to an increased risk of thrombosis. Interindividual variations in pharmacokinetics and pharmacodynamics of contraceptive steroids are great and could tentatively explain why certain persons may be at an increased risk of thrombosis. Although most studies have looked at steady-state conditions during contraceptive steroid intake, it would seem prudent to investigate further the hemostatic system during a non-steady-state condition, such as that occurring during the first few days of the pill-free interval.