Risk of venous thromboembolism with oral contraceptives.

E1278 CMAJ, December 13, 2011, 183(18) © 2011 Canadian Medical Association or its licensors The lifetime incidence of venous thromboembolism for both men and women is estimated at 1 per 1000. Increasing our understanding of the precipitating risk factors, as well as optimizing the prevention of this condition in high-risk populations, could help to reduce the burden of venous thromboembolism. A retrospective cohort study by Gronich and colleagues challenges us to weigh the thrombotic risk of various oral contraceptives. Older age, admission to hospital, surgery, cancer and other comorbidities dramatically increase the rate of venous thromboembolism. Conversely, effective thromboprophylaxis is known to substantially reduce the incidence. The difficulty of ensuring good compliance with thromboprophylaxis remains a major barrier to reducing the adverse outcomes associated with venous thromboembolism. The absolute rate of venous thromboembolism in young women is low, but the use of combined oral contraceptives increases this rate threeto five-fold, with an even higher rate in the presence of associated risk factors such as thrombophilia. Gronich and colleagues focus our attention on women 12 to 50 years of age taking hormonal contraceptives and attempt to quantify the relative difference in risk of venous thromboembolism attributable to the type of oral contraceptive prescribed. The reported incidence of venous thrombo embolism in users of oral contraceptives is about 0.06 per 100 pill-years, significantly lower than the rate of 0.2 per 100 years at risk during pregnancy and the postpartum period. The high contraceptive efficacy of oral contraceptives, despite their thrombotic risk, should be weighed against the risks associated with less effective contraception methods, including the potential thrombotic consequences of an unwanted pregnancy. The aim of numerous previous studies has been to evaluate the differential thrombotic risk of third-generation (containing desogestrel, gestodene or norgestimate) and secondgeneration (containing norgestrel or levonor gestrel) oral contraceptives. Although not completely consistent, the overall findings have substantiated a relatively lesser thrombotic risk with the use of second-generation contraceptives, albeit with heterogeneity among the studies. Drospirenone-containing oral contraceptives, which have become available more recently, have a mineralocorticoid-derived progestational component. The study by Gronich and colleagues adds further evidence of a higher relative risk of venous thromboembolism among women taking this type of oral contraceptive, rela tive to the alternatives of either thirdor second-generation oral contraceptives. Although earlier publications did not detect any such relative difference in thrombotic risk, carefully designed evaluations performed more recently have found evidence of such a risk. The industry-funded phase IV European postmarketing surveillance study collected data by mailed questionnaires from women for whom an oral contraceptive had been newly prescribed. In that study, all subgroups of pill users (categorized by additional risk factors or by drug prescribed) had a similar incidence of venous thromboembolism, which, according to the study design, excluded the possibility of a twofold higher incidence of venous thromboembolism in drospirenone users and provided evidence of non-inferiority. Subsequently, a large case–control study that evaluated a substantially larger database (in terms of women-years of pill use) more convincingly demonstrated a difference in thrombotic risk. The authors of that study compared incidence rates of idiopathic, nonfatal venous thrombo embolism between drospirenone users and levo norgestrel users, a comparison that emphasized the subset of events most likely influenced by the type of oral contraceptive used. The increased Risk of venous thromboembolism with oral contraceptives