Abstract

Background: To evaluate whether migraine is associated with an increased risk of developing central serous chorioretinopathy (CSCR) using a large real-world electronic medical record dataset.

Design: Retrospective cohort study.

Participants and controls: Adults aged 18-40 years with a diagnosis of migraine were identified from the TriNetX Global Collaborative Network and compared to individuals without migraine or other headache syndromes who underwent routine medical examinations. Patients with a history of CSCR or corticosteroid use within one month before or after the index migraine diagnosis were excluded. All patients had at least one year of follow-up data after the index date.

Methods: Propensity score matching (PSM) was performed in a 1:1 ratio to balance baseline demographic and clinical characteristics. Time-to-event analysis was conducted using Cox proportional hazards models and log-rank tests. Subgroup and sensitivity analyses assessed the robustness of the association across sex, migraine subtypes (with and without aura), age range, follow-up duration, corticosteroid exposure, and an alternative multivariable Cox regression in the full cohorts. Additional analyses evaluated the effect of migraine-related medications on CSCR risk.

Main outcome measures: Incident CSCR within five years of the index date. The E-value was calculated to evaluate robustness against unmeasured confounding.

Results: After matching, 413,663 patients with balanced baseline characteristics remained per cohort. During a five-year follow-up, migraine patients demonstrated a higher risk of CSCR compared with controls (1.43 vs. 0.51 per 10,000), with an HR of 2.74 (95% CI, 1.66-4.51). The calculated E-value was 4.92 for the observed HR and 2.71 for the lower confidence bound. The association remained consistent across all subgroups and sensitivity analyses. Positive and negative control outcomes further supported the specificity and internal validity of the associations. Analysis of medication exposure showed no significant differences in CSCR risk across nonsteroidal anti-inflammatory drugs, beta-blockers, topiramate, valproate, and antidepressants.

Conclusions: Migraine was associated with a significantly increased risk of incident CSCR in a large, real-world cohort. The consistency of this association across multiple analyses suggests a potential shared pathophysiologic mechanism warranting further investigation.