Abstract [from journal]
Purpose: To determine if female hormonal therapy (FHT) increases the incidence of non-infectious uveitis.
Design: Retrospective cohort study PARTICIPANTS: FHT-exposed women and matched unexposed women enrolled in a national insurance plan.
Methods: Estimation of non-infectious uveitis incidence used multivariable Cox proportional hazards regression. To account for differences between the exposed and unexposed cohorts, a propensity score for being prescribed FHT was created using logistic regression, and inverse probability of treatment weighting was performed.
Main Outcomes and Measures: Incidence of non-infectious uveitis. For the primary outcome, incident non-infectious uveitis was defined as a new diagnosis code for non-infectious uveitis followed by a second instance of a non-infectious uveitis code within 120 days. For the alternative outcome definition, a corticosteroid prescription or code for an ocular corticosteroid injection within 120 days of the uveitis diagnosis code was used instead of the second uveitis diagnosis code.
Results: There were 217,653 FHT-exposed women and 928,408 unexposed women. For the primary outcome, the hazard ratio (HR) for incident non-infectious uveitis was not significantly different between the FHT and unexposed cohorts [HR=0.99, 95% Confidence Interval (CI): 0.83-1.17, P=0.87]. With the alternative outcome definition, the FHT cohort was more likely to develop uveitis (HR=1.21, 95% CI: 1.04-1.41, P=0.01). When examined by anatomic subtype, for anterior uveitis there was a greater likelihood of incident uveitis in the exposed cohort (HR=1.23, 95% CI: 1.05-1.45, P=0.01) for the alternative outcome definition but not for the primary outcome. With age stratification, exposed women ≥ 45 years old at time of FHT prescription were more likely to develop uveitis (HR=1.23, 95% CI: 1.03-1.47, P=0.03) for the alternative outcome definition. A similar HR (1.22) was seen for women ≤ 44 years old at time of prescription, but this association did not meet statistical significance (P=0.20).
Conclusions: Exposure to FHT increases the rate of incident non-infectious uveitis when uveitis is defined based on both diagnostic codes and documentation of corticosteroid treatment.