Abstract
INTRODUCTION: Migraine is an established risk factor for cerebral ischemic stroke, with an especially robust association in patients with migraine with aura. However, it is not known if migraine is a risk factor for retinal stroke (central or branch retinal artery occlusion; CRAO or BRAO). METHODS: We performed a retrospective, observational, cohort study using population-based data from the State Inpatient Databases and State Emergency Department Databases from New York (2006-2015), California (2003-2011), and Florida (2006-2015) to determine the association between hospital-documented migraine and retinal stroke. The primary exposure was hospital-documented migraine (ascertained from admission or emergency department diagnosis codes). The primary endpoint was time to hospital-documented CRAO (ICD-9-CM code 362.31 in the primary diagnosis position) and secondary endpoints included time to BRAO and any retinal artery occlusion (RAO). Cause-specific hazard models were used to model the association between migraine and subsequent CRAO. RESULTS: Of 39,835,024 patients included in the study, 1109,140 had migraine documented during our two year ascertainment window. Patients with migraine were younger (40.2±15.2 vs. 46.9±19.8, standardized difference (SD) 0.38), more likely to be female (81.4 % vs. 54.7 %, SD 0.6), and had a lower burden of atrial fibrillation (4.5 % vs. 6.9 %, SD 0.1), chronic kidney disease (1.9 % vs. 3.6 %, SD 0.2), and congestive cardiac failure (2.7 % vs. 5.1 %, SD 0.12). Migraine was not associated with CRAO in the primary diagnostic position (adjusted hazard rate (aHR) 1.15 (95 % CI: 0.79-1.67). However, it was associated with CRAO in any diagnostic position (aHR 1.39 (95 % CI: 1.08-1.78). As positive controls, we replicated previously established associations of migraine with cerebral ischemic stroke (aHR 1.35 (95 % CI: 1.32-1.38) and embolic ischemic stroke (aHR 1.15 (95 % CI: 1.08-1.22). CONCLUSIONS: In a large, nationally-representative, claims-based study of patients from 3 regions in the United States (US), we did not find an adjusted association between migraine and a primary discharge diagnosis of CRAO. Our hypothesis-generating finding that migraine was associated with CRAO when using a broader definition sets the stage for future work leveraging both outpatient and pharmacy based claims to further explore this finding.