Oral contraceptives and venous thromboembolism: a case-control study designed to minimize detection bias.

BACKGROUND Previous epidemiologic studies of venous thromboembolism and oral contraceptive use are susceptible to bias in the detection of venous thromboembolic events. This case-control study uses a unique design to minimize the influence of detection bias. METHODS Nonpredisposed women younger than the age of 40 years who underwent pulmonary angiography, lower extremity venography, or lower extremity duplex Doppler sonography at a large urban hospital were classified into a case group or control group based on results of their diagnostic studies. Medical records were reviewed for a history of current oral contraceptive use. RESULTS Fifty-seven women met the study criteria during the 11-year study period. Seven of 9 women in the case group and 17 of 48 women in the control group were currently using oral contraceptives (odds ratio 6.38; 95 percent confidence limits 1.19, 34.2). CONCLUSIONS The association previously noted between venous thromboembolism and oral contraceptive use is not due to bias in the detection of venous thromboembolic events.