Migraine in transient global amnesia: a meta-analysis of observational studies.

BACKGROUND

PURPOSE: Although many studies have investigated the relationship between transient global amnesia (TGA) and migraine, to date, no meta-analysis has confirmed the existence and size of their association.

METHODOLOGY: Literature search involved MEDLINE, EMBASE, CENTRAL and PsycINFO. Observational controlled studies including TGA patients (Caplan, Hodges and Warlow) were retrieved. Quality evaluation was based on the Newcastle-Ottawa scale. The prevalence of migraine was compared in TGA patients vs. healthy controls (HC), as well as in TGA against TIA individuals. Data from case-control, cross-sectional and cohort studies were pooled separately.

RESULTS

Literature search yielded 1178 articles, 12 of which were included in the present meta-analysis. Results from case-control (ten), cohort (one) and cross-sectional (one) studies were compatible with an association between TGA and migraine. The nationwide inpatient cross-sectional study was of lesser value due to its inpatient orientation. The high-quality, population-based, retrospective cohort (158,301 participants per group) determined a higher relative-risk (RR) of TGA for migraine vs. non-migraine individuals [RR = 2.48, 95%confidence-interval (95% CI) = (1.32, 4.87)]. Sensitivity testing based on stricter diagnostic criteria strengthened the estimated association [RR = 3.84, 95% CI = (1.57, 9.38)]. Additionally, pooled data from eight case-control studies (700 TGA, 746 HC) yielded similar results [Odds-Ratio, OR = 2.51, 95% CI = (1.85, 3.41)], with the association mainly driven by the three high-quality studies, rather than the five articles of moderate quality. Finally, pooled findings from four case-control studies of moderate-quality revealed a higher prevalence of migraine among TGA compared to TIA patients [OR = 1.82, 95% CI = (1.22, 2.73)].

CONCLUSIONS

A significant association between TGA and migraine was established. The underlying connecting mechanism remains undetermined, yet.