Venous thromboembolic disease and combined oral contraceptives: results of international multicentre case-control study. World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception.
Researchers compared data on 1143 women aged 20-44 who had deep vein thrombosis (DVT) and/or pulmonary embolism (PE) who were hospitalized at one of 21 centers in 17 countries with data on 2998 age-matched controls to examine the relationship between idiopathic venous thromboembolism (VTE) and use of combined oral contraceptives (OCs). OC use was linked to a significantly increased risk of a first VTE event (odds ratio [OR] = 4.15 in Europe and 3.25 in developing countries). In Europe, the only confounding variable identified was a history of hypertension in pregnancy (HIP). No confounder was identified in developing countries. Risk estimates tended to be higher for DVT than for PE. The risk of VTE increased within four months of beginning OCs. Duration of current episode of OC use did not affect increased risk. The increased risk of VTE disappeared within three months of discontinuing OCs. Age, history of hypertension (except for HIP), and smoking did not affect relative risk of VTE among OC users. As body mass index (BMI) increased, so did the risk of VTE ; p < 0.001), regardless of OC use. Among OC users, a BMI greater than 25 kg/sq. m increased the risk of VTE (OR = 7.01 in Europe and 9.44 in developing countries). In Europe, OC users with a history of HIP faced an increased risk of VTE (OR = 9.24 vs. 4.04 for OC users with no history of HIP). The risks associated with the use of OCs containing a third generation progestogen were greater than those with first generation progestogens (e.g., among combined OCs containing <50 mcg estrogen, OR = 7.36 vs. 3.37 in Europe, OR = 12.23 vs. 0 in developing countries). Among OCs containing first and second generation progestogens, the risk tended to be lower in those OCs containing less than 50 mcg estrogen. These findings demonstrate an association between OC use and VTE. The overall risk estimates were lower than those in the majority of earlier studies of non-fatal idiopathic VTE, however.