Corticosteroid-Related Adverse Events Systematically Increase with Corticosteroid Dose in Noninfectious Intermediate, Posterior, or Panuveitis

Purpose: Chronic use of corticosteroids for the treatment of uveitis has been linked with drug-associated toxicity and adverse events (AEs). This study examines the association between corticosteroid dosage and incidence rates of corticosteroid-related AEs. Design: A post hoc analysis of the VISUAL-1 and VISUAL-2 placebo-controlled clinical trials. Participants: The clinical trials consisted of adults with active (VISUAL-1) and inactive (VISUAL-2) noninfectious intermediate, posterior, and panuveitis. Patients were randomized to receive adalimumab or placebo and underwent a protocol-defined mandatory taper to discontinue their oral corticosteroids. Methods: Adverse event data were collected at each visit and included an assessment of the corticosteroid relationship by the investigator. A longitudinal Poisson regression model was estimated controlling for timedependent corticosteroid dose, age, sex, prior oral corticosteroid dose, prior topical corticosteroid use, and concomitant immunosuppressive drug use. Only patients randomized to placebo were considered. Main Outcome Measures: The primary outcome measure was the frequency of AEs. Results: The incidence rates of corticosteroid-related AEs among placebo patients during the prednisone treatment period in VISUAL-1 was statistically higher than after discontinuation (454.2 per 100 patient-years [PY] vs. 36.1 per 100 PY, incident rate ratio = 12.6, P < 0.001). Incidence rate ratios among VISUAL-2 patients were similarly high (317.5 per 100 PY vs. 41.1 per 100 PY, incident rate ratio = 7.7, P < 0.001). Based on the Poisson multivariate longitudinal Generalized Estimating Equation (GEE) model, each 10 mg increase in prednisone dose is associated with a 1.5-and 2.6-fold increase (P < 0.001 and P < 0.001) in the rate of corticosteroid-related AEs in VISUAL-1 and VISUAL-2, respectively. This implies in turn that a patient with active uveitis taking 60 mg/day of prednisone will experience, on average, an additional 10.1 (95% confidence interval (CI), 6.3-14.5; P < 0.001) corticosteroid-related AEs per year compared with a patient taking 10 mg/day, whereas a patient with inactive uveitis taking 35 mg/day of prednisone will experience, on average, an additional 23.5 (95% CI, 7.6-52.7; P = 0.05) corticosteroid-related AEs per year compared with a patient taking 10 mg/day. Conclusions: Evidence from VISUAL-1 and VISUAL-2 suggests that the incidence rates of corticosteroidrelated AEs increase systematically with corticosteroid dose. (C) 2017 by the American Academy of Ophthalmology.