Migraine with aura has been associated with increased risk of ischemic and hemorrhagic stroke. Prior studies have shown a further increase in risk in women using combined hormonal contraceptives (CHCs). This has led to guidelines recommending against use of CHCs in this population. We sought to assess whether the risk of stroke is associated with the dose of estrogen and whether there is evidence of synergism between migraine and CHCs. We also sought to assess whether an interaction effect exists between migraine and CHCs.
We searched PubMed, the Cochrane Library, and EMBASE from inception through January 2016 for relevant English-language studies of adults, of any design. We included studies that examined exposure to CHCs and reported outcomes of ischemic or hemorrhagic stroke. Data extraction and assessment of study quality were conducted independently by reviewer pairs and quality was assessed with the GRADE and Newcastle Ottawa scales.
Of 2480 records, 15 studies met inclusion criteria and six provided odds ratios for the relevant population. The point estimates for the odds ratios for ischemic stroke in women with migraine who used CHCs with any dose of estrogen ranged from 2.08 to 16.9. Studies were generally small and confidence intervals were wide. No studies reported odds ratios for stroke risk as a function of estrogen dose in women with migraine, largely due to insufficient sample sizes. No interaction effect between migraine and CHCs was seen in the seven studies that assessed this. One study differentiated risk by presence or absence of migraine aura and found an increased risk in the migraine with aura population (OR 6.1; CI 3.1 to 12.1 in migraine with aura vs 1.8; CI 1.1 to 2.9 in the migraine without aura group). Studies generally had high Newcastle Ottawa scores and low GRADE levels of evidence. No studies met all three supplementary quality criteria (assessed migraine subtype, used International Classification of Headache Disorders criteria for diagnosis of migraine, and stratified risk by estrogen dose).
This systematic review shows a lack of good quality studies assessing risk of stroke associated with low dose estrogen use in women with migraine. Further study in this area is needed. The available evidence is consistent with an additive increase in stroke risk with CHC use in women with migraine with aura. Since the absolute risk of stroke is low even in the presence of these risk factors, use of CHCs in women who have migraine with aura should be based on an individualized assessment of harms and benefits.