Real-World Treatment Patterns in a Population With Neovascular AMD Treated With Anti-VEGF Agents

BACKGROUND AND OBJECTIVES: To characterize mean number of injections, injection type, and injection frequency during the first year of treatment; assess factors significantly related to injection interval; and identify predictive factors related to patient outcomes. PATIENTS AND METHODS: A retrospective, non-comparative, nonrandomized cohort study of ocular treatment with intravitreal injections of bevacizumab, ranibizumab, and aflibercept in eyes with neovascular age-related macular degeneration (nAMD). Data from January 1, 2012, through March 31, 2018, were systematically extracted from the electronic medical record system at Cole Eye Institute. Eligible patients had three or more injections within the first 12 months of treatment and received no prior injections. RESULTS: Patients received an average 8.12 +/- 2.45 injections, and 45% of patients received injections at an interval of 8 weeks or less (<= q8 weeks), 33% received injections at 8 to 12 weeks (q8-12 weeks), and 22% received injections at greater than 12 weeks (>q12 weeks). Age (P =.007) and initial central subfield thickness (CST) (P =.043) had statistically significant trend relationships (P =.007) with injection interval, whereas younger patients and patients with higher CST measurements tended to have shorter injection intervals. Injection interval was a significant predictor of visual acuity (VA) and CST. Patients receiving injections at q8-12 weeks were more likely to have better VA outcomes than patients with injection intervals at =q8 weeks (odds ratio [OR] = 1.66 [1.16, 2.37]; P =.005). Patients receiving injections at >q12 weeks did not show a significant improvement in VA (P =.06) and were more likely to have worse CST outcomes than patients receiving injections at =q8 weeks (OR = 1.95 [1.17, 3.26]; P =.011). CONCLUSION: A significant portion of patients receive injections at an interval longer than every 8 weeks. Age and baseline CST had a significant relationship with injection interval. Injection interval was a significant predictor of VA and CST at 1 year. Patients with an injection interval of >q12 weeks tended to have less VA improvement and CST reduction compared to the =q8 weeks and q8-12 weeks groups. These findings suggest an extended injection interval >q12 weeks may be at the expense of potential VA improvement.