OBJECTIVE: To assess the risk of developing cerebral thromboembolism among pregnant women and among fertile women with hypertension, migraine, diabetes, and previous thrombotic disease, and to investigate the interaction of these risk factors with the use of oral contraceptives. DESIGN: A retrospective case-control study. SETTING: All gynaecological, medical, neurological, and neurosurgical departments in Danish hospitals. SUBJECTS: Seven hundred and ninety-four women in Denmark aged 15 to 44 who suffered a cerebral thromboembolic attack during the period 1985 to 1989 and 1588 age-matched, randomly selected controls. RESPONSE: Of the 692 case and 1584 control questionnaires sent out, 590 (85.1%) and 1396 (88.1%), respectively, were returned. Of the 590 cases, nine had had cerebral thrombosis before 1980, 15 refused to participate, 44 had a revised diagnosis (primarily multiple sclerosis) and 25 had an unreliable diagnosis, leaving 497 with a reliable cerebral thromboembolic diagnosis. Among the 1396 controls, 26 either refused to participate, were mentally handicapped, lived abroad or returned an uncompleted questionnaire, leaving 1370 controls included in the study. RESULTS: After multivariate analysis, pregnancy implied an odds ratio (OR) for a cerebral thromboembolic attack of 1.3 (nonsignificant), diabetes an OR of 5.4 (P < 0.001), hypertension an OR of 3.1 (P < 0.001) and migraine an OR of 2.8 (P < 0.01). Women with previous non-cerebral thrombotic disease had an OR for cerebral thrombo-embolism of 5.3 (P < 0.001). Women with other predisposing medical diseases had an OR of 8.3 (P < 0.001). These ORs were identical among users and non-users of combined oral contraceptives. CONCLUSION: In this study pregnancy implied a non-significant elevated odds ratio of 1.3 for cerebral thromboembolism whereas diabetes, hypertension, migraine and past thromboembolic events increased the risk of cerebral thromboembolism significantly. Women with these increased thrombotic risks should use oestrogen-containing oral contraceptives only after careful considerations of the risks, if at all.
Øjvind Lidegaard (1995) studied this question.