Background: The association between oral contraceptive (OC) use and acute myocardial infarction (AMI) was established in studies from northern Europe and the USA, which took place during the 1960s and 1970s. Few data are available to quantify the risk worldwide of AMI associated with use of OCs introduced since those early studies. This hospital-based case-control study examined the association between a first AMI and current OC use in women from Africa, Asia, Europe, and Latin America (21 centres).
Methods: Cases were women aged 20-44 years who had definite or possible AMI (classified by history, electrocardiographic, and cardiac-enzyme criteria), who were admitted to hospital, and who survived for at least 24 h. Up to three hospital controls matched by 5-year age-band were recruited for each of the 368 cases (941 controls). All participants were interviewed while in hospital with the same questionnaire, which included information on medical and personal history, lifetime contraceptive use, and blood-pressure screening before the most recent episode of OC use. Odds ratios compared the risk of AMI in current OC users and in non-users (past users and never-users combined).
Findings: The overall odds ratio for AMI was 5.01 (95% CI 2.54-9.90) in Europe and 4.78 (2.52-9.07) in the non-European (developing) countries; however, these risk estimates reflect the frequent coexistence of other risk factors among OC users who have AMI. Very few AMIs were identified among women who had no cardiovascular risk factors and who reported that their blood pressure had been checked before OC use; odds ratios associated with OC use in such women were not increased in either Europe or the developing countries. Among OC users who smoked ten or more cigarettes per day, the odds ratios in Europe and in the developing countries were over 20. Similarly, among OC users with a history of hypertension (during pregnancy or at any other time), odds ratios were at least ten in both groups of countries. No consistent association between odds ratios for AMI and age of OC users or oestrogen dose was apparent in either group of countries. No significant increase in odds ratios was apparent with increasing duration of OC use among current users, and odds ratios were not significantly increased in women who had stopped using OCs, even after long exposure. The study had insufficient power to examine whether progestagen dose or type had any effect on AMI risk.
Interpretation: Current use of combined OCs is associated with an increased risk of AMI among women with known cardiovascular risk factors and among those who have not been effectively screened, particularly for blood pressure. AMI is extremely rare in younger (< 35 years) non-smoking women who use OCs, and the estimated excess risk of AMI in such women in the European centres is about 3 per 10(6) woman-years. The risk is likely to be even lower if blood pressure is screened before, and presumably during, OC use. Only among older women who smoke is the degree of excess risk associated with OCs substantial (about 400 per 10(6) woman-years).