Studies of the associations between oral contraceptives and cardiovascular disease are limited by the extreme rarity of these problems among young women. There are no randomized, controlled trials, and the large, prospective cohort studies only have data regarding older oral contraceptive formulations that are now little used. Historical cohort studies that use record-linkage techniques to analyze data from computerized databases represent a new approach to assessing oral contraceptive use and thrombosis. The largest new studies, with the most sophisticated analyses, use the case-control design. A surprising result of the new studies was a difference in risk of thrombosis according to progestin type. Because these are observational rather than randomized studies, clinical factors influence the choice of oral contraceptive and may bias the study results. Controversy about the surprising results has stimulated additional analyses and critical reviews in an attempt to explain the associations.
PIP: This article discusses methodologic aspects of recent studies on oral contraceptive (OC) use and thrombosis. Because of the extreme rarity of cardiovascular disease among young women, no studies of cardiovascular outcomes have employed a randomized, controlled trial design. Rather, all knowledge of OC risks comes from observational studies. Although case-control studies have suggested a difference in thrombosis risk according to progestin type, clinical factors that influence the choice of OC may bias the results. Historical cohort studies based on record-linkage techniques to analyze data from computerized databases represent a promising new approach for assessing the thrombosis risk of OCs. Overall, epidemiologic studies are less useful for subtle, weak associations where the relative risk is close to 1--the case with the most recent round of studies on the OC-thrombosis association. A balanced assessment of the differential values of various combination OCs would require that all the risks and benefits be considered on a formulation-by-formulation basis.