First Author: S.Kelly UK
Co Author(s): A. Ferreira R. Hashmonay J. Lu E. Souied 0 0 0 0 0 0 0 0 0
Back to previous
Ranibizumab and aflibercept are currently licensed for the treatment of neovascular age-related macular degeneration (nAMD). These treatments are in widespread use and have proven efficacy. However, a low but consistent rate of endophthalmitis has been observed following such intravitreal injections. Recently, emerging real-world evidence suggests a potential difference in endophthalmitis rates between these anti-vascular endothelial growth factor (anti-VEGF) therapies. These two medications have different molecular structures and distinct mechanisms of action. Given the critical clinical significance of endophthalmitis, we examined endophthalmitis occurrence following ranibizumab or aflibercept injection for the treatment of nAMD in a retrospective US claims database analysis.
Because of the low frequency of endophthalmitis in clinical care following injections, large numbers of patients are required to accurately assess such occurrence rates. In this study, injection-level data were obtained from IMS Health’s Integrated Data Warehouse, a US physician-recorded database of medical claims.
Injections of ranibizumab or aflibercept between 1 November 2011 and 31 August 2013 administered to adults with a concomitant International Classification of Disease (ICD-9) diagnosis code for nAMD were identified. Each injection was followed for 30 days or until one of the following: endophthalmitis (as indicated by an ICD-9 endophthalmitis code) occurred; a second anti-VEGF injection was given; or the study period ended. Unadjusted event rates (ERs) and relative risks (RR) of endophthalmitis were determined along with 95% confidence intervals (95% CIs) and were compared using a Poisson regression. Adjusted relative risks accounting for the use of repeated observations in individual patients were also estimated using general estimating equations. Multiple sensitivity analyses were conducted to evaluate the robustness of the primary results.
A total of 253 647 ranibizumab injections and 179 147 aflibercept injections were eligible for analysis. The total number of injections included in the present analysis represented ranibizumab injections in 54 551 patients and aflibercept injections in 39 389 patients. Patient characteristics for both groups were similar among patients receiving ranibizumab vs aflibercept for median age at the time of their index injection (82.0 years vs 82.0 years) and proportion of patients who were female (63.3% vs 61.9%). Most endophthalmitis events occurred within 5 days of injection (including 83% of post-ranibizumab events and 88% of post-aflibercept events). The mean time from injection to the first endophthalmitis claim was 3.1 days for ranibizumab and 3.0 days for aflibercept. The number of endophthalmitis events was 162 with ranibizumab treatment (ER [95% CI]: 0.64 [0.54–0.74] events per 1000 injections) and 189 events with aflibercept treatment (1.06 [0.91–1.21] events per 1000 injections) (p<0.0001); this result censored injections and events after a patient’s first endophthalmitis claim, to avoid counting repeated treatment for one event as separate events. The adjusted RR of the occurrence of endophthalmitis with aflibercept compared with ranibizumab was calculated as 1.65 (95% CI: 1.34–2.04).
To our knowledge, this real-world, injection-level database analysis provides the first direct comparison of the occurrence of endophthalmitis following ranibizumab and aflibercept injections for the treatment of nAMD in routine clinical practice. Results from this claims database analysis involving over 400 000 injections suggest that an endophthalmitis event would be observed every 1561 (95% CI: 1353–1845) injections with ranibizumab and every 945 (95% CI: 827–1102) injections with aflibercept, and that the risk of endophthalmitis is 65% higher following treatment with aflibercept compared with the risk following ranibizumab in ophthalmology care in the USA. Importantly, all sensitivity analysis undertaken also supported these differences. However, data from this retrospective analysis should be interpreted with caution, because of the inherent limitations of this type of study and limited understanding of mechanisms to explain the apparent difference in endophthalmitis risk with aflibercept. Additional studies will be required to further explore the differences in risk of endophthalmitis identified by this study and the implications for clinical practice.