An increased risk of venous thrombosis has been demonstrated in women receiving oral contraceptives (OCs). This risk has been primarily associated with the oestrogen content, but recent studies showed that the progestogen may also plays a role. A higher risk was found with the so-called third-generation (desogestrel, gestodene) as compared with the second-generation progestogens (levonorgestrel). The risk was approximately two-fold. These unexpected results have been the subject of many debates, and bias-such as selection bias-has been suggested. The existence of bias cannot be completely excluded, but the thrombotic risk seems however to be slightly higher with the third-generation progestins. Haemostatic changes have been observed during OC intake. Both coagulation and fibrinolytic activity are increased: the beneficial profibrinolytic effect may counterbalance the deleterious procoagulant effect. This may explain the absolute risk of venous thromboembolism is low during OC treatments. Some women who have pre-existing haemostatic abnormalities such as deficiency in antithrombin or activated protein C resistance with factor V Leiden, may be at a higher risk. Theoretically, this could be due to an increased coagulation or to a lack of increased fibrinolysis as compared with second-generation progestogens. The only difference presently reported with third-generation OCs is a decreased sensitivity to activated protein C, possibly resulting in hypercoagualibility of greater magnitude. The selection bias suggested in epidemiological studies may also exist for the later study, as women taking third- or second-generation OCs were not randomized. The possible increased risk related to third-generation OCs should not change the known general contra-indication. Practical guidelines are proposed for women with personal or family history of venous thromboembolism, and for those with a congenital cause of thrombophilia.

Key words: haemostasis/oral contraceptives/third-generation progestogens/thrombosis