Abstract
Purpose: :
Intravitreal injections (IVT) of Bevacizumab (Bev) (off-label) and Ranibizumab (Ran) are used to treat neovascular age-related macular degeneration (AMD). However, the safety of Bev used for IVT has not been fully established. Current clinical trials of Bev vs. Ran for AMD are not powered to evaluate some infrequent but important safety outcomes. This Medicare claims database analysis compares the relative safety of Ran and Bev for treating neovascular AMD.
Methods: :
Medicare claims data were obtained for all beneficiaries with 1+ neovascular AMD claim between 2005-2009. Bev & Ran treatment were defined based on their respective drug codes paired with an IVT. Beneficiaries receiving more than one treatment type were excluded. To allow for equal follow up time, the primary analysis was limited to beneficiaries initiating treatment in 2008-9. Participants were censored at first IVT steroid use. Propensity scores were used to help balance the groups on baseline characteristics. Systemic and ocular event rates were calculated. Hazard ratios (HR) comparing Bev with Ran were calculated using Cox proportional hazard models.
Results: :
The primary analysis cohort includes 77,886 beneficiaries (46%Ran). HRs adjusted for baseline comorbidities, demographics and socio-economic status proxies showed an 11% higher risk in overall mortality (HR: 1.11; 99% CI: 1.01-1.23) and a 57% higher risk of hemorrhagic cerebrovascular accident (CVA) in the Bev group (HR: 1.57; 99% CI: 1.04-2.37). No statistically significant differences were seen in risk of myocardial infarction or ischemic CVA. Compared to patients treated with Ran, patients treated with Bev were 19% less likely to have newly diagnosed ocular hypertension/glaucoma (HR 0.81; 99% CI: 0.71-0.93), but 80% more likely to have ocular inflammation (HR: 1.8; 99% CI: 1.2-2.8), and 11% more likely to have cataract surgery following AMD treatment (HR: 1.11; 99%CI: 1.01-1.23). Differences in overall mortality and hemorrhagic CVA were attenuated in secondary analyses that included use of Bev or Ran based on unclassified drug codes and data back to 2006.
Conclusions: :
Data from this Medicare claims analysis suggest differences in the safety profile of Bev vs Ran. However, this study is limited by incomplete information on some important confounding factors, e.g., smoking, lipid and blood pressure levels, which would further clarify the relative safety of these treatments in wet AMD.
Keywords: age-related macular degeneration • clinical (human) or epidemiologic studies: outcomes/complications