Recent concern about the safety of combined oral contraceptives (OCs) with third-generation progestagens prompted an examination of data from a population-based case-control study (Leiden Thrombophilia Study). We compared the risk of deep-vein thrombosis (DVT) during use of the newest OCs, containing a third-generation progestagen, with the risk of "older" products. We also investigated the influence of family history of thrombosis, previous pregnancy, age, and the thrombogenic factor V Leiden mutation. We selected 126 women with DVT and 159 controls aged 15-49 (mean age 34.9) and premenopausal and found, as compared with non-users, the highest age-adjusted relative risks to be that for an OC containing desogestrel and 30 micrograms ethinyloestradiol (relative risk [RR] 8.7, 95% CI 3.9-19.3). We found lower relative risks for all other types of OC, ranging from 2.2 to 3.8. In a direct comparison, users of the desogestrel-containing oral contraceptive had a 2.5-fold higher risk (95% CI 1.2-5.2) than users of all other OC types combined. The relative risk for the desogestrel-containing OC was similar among women with and without a family history--ie, preferential prescription because of family history cannot explain our findings. Nor could the excess risk be explained by previous pregnancy, and it was highest in the youngest age categories, where we would expect most new users. The age-adjusted RR for the desogestrel-containing contraceptive was 9.2 (3.9-21.4) among non-carriers of the factor V Leiden mutation and 6.0 (1.9-19.0) among carriers of the mutation. This latter risk is superimposed on the 8-fold increased risk of venous thrombosis for carriers of the factor V Leiden mutation. The risk of carriers using the desogestrel-containing OC as compared with noncarrier non-users will therefore be increased almost 50-fold. Use of low-dose OCs with a third-generation progestagen carries a higher risk of DVT than the previous generation of OCs. The absolute risk of DVT associated with these OCs seems to be especially high among carriers of the factor V Leiden mutation and among women with a family history of thrombosis. However, the higher risk associated with OC with a third-generation progestagen compared with previous generations was also present in women without factor V Leiden and with no family history.