OBJECTIVES:  To explore the frequency of injection and non-injection visits for ranibizumab and aflibercept during the first, second and third years of follow-up for the treatment of neovascular age-related macular degeneration (nAMD) in a real-world setting. METHODS: Claims from the IMS Integrated Data Warehouse were used to identify treatment naïve nAMD patients with ≥1 claim for ranibizumab or aflibercept between 01 November 2011 and 31 December 2012 with a minimum 1080 days of follow up. Patients were considered to be treatment-naïve if they had received no anti-VEGF treatment for >180 days prior to the index date. Mean (± standard deviation (SD)) number of injections received within 360/720/1080 days of treatment initiation were calculated and compared for patients initiating therapy with ranibizumab or aflibercept. No adjustments for bilateral treatment were performed. RESULTS: Patient characteristics at the index date were similar for patients receiving ranibizumab (N=2,224) or aflibercept (N=1,290) (median age, 81 years; % female, 65%). In patients continuously treated with the index agent in the first, second and third year, the mean (±SD) number of injections in each year was comparable: 4.5±3.0 vs 5.1±2.8, 2.0±2.8 vs 3.1±3.1, and 1.6±2.8 vs 2.7±3.2 for ranibizumab and aflibercept, respectively. Mean (±SD) number of non-injection visits in each year for ranibizumab and aflibercept treated patients was also similar; 3.1±3.1 vs 2.2±2.6, 2.5±2.9 vs 2.0±2.5, and 2.0±2.7 vs 1.8±2.2, respectively. CONCLUSIONS: This study represents a long-term follow-up of the patient cohort included in the 1 year treatment patterns study by Ferreira et al. (2015). Injection frequency results in the first year are consistent with those observed in the original study. These results further suggest that in routine clinical practice ranibizumab and aflibercept are used in a similar manner during each of the first three years of treatment for nAMD.