Section 1: General

Abstract

Background: Randomized trials have reported a doubling of the risk of congestive heart failure (CHF) among colorectal cancer (CRC) patients treated with bevacizumab (BEV, Avastin(registered trademark)), a humanized monoclonal antibody against vas- cular endothelial growth factors. However, the cardiovascular risk attributable to BEV has not been established in general clinical practice, where patients are typically older and more comorbidity than RCT participants. Objectives: To evaluate BEV use and its effects on the risk of CHF and cardiomyopathy (CM) in CRC patients aged 65+ with chemotherapy. Methods: Using the U.S. SEER-Medicare database, we identified all incident CRC patients diagnosed in 2005 who received chemotherapy with or without BEV and followed through 2007. We defined outcomes as the first observed CHF or CM diagnosis after chemotherapy initiation using diagnostic codes from outpatient and inpatient claims. We first conducted a propensity score analysis of BEV use in this cohort. Second, we used Cox proportional hazard (PH) models applying the inverse-propensity treatment weighted (IPTW) method to calculate hazard ratios (HR) for the risk of CHF or CM with Bev. We modeled BEV and other chemotherapies as time-dependent covariates. Results:We observed 188 CHF or CM events in 3,832 CRC patients (median age=75, 50% males). Younger age and having fewer comorbidities were significantly associated with receipt of BEV. BEV use increased the risk of CHF or CM (HR=2.32, 95% C.I. =1.70, 3.16). The increased risk was strong among patients age 75+ (HR=3.65, 95% C.I.=2.55, 5.23) and those with prior cardiac conditions (HR=3.00, 95% C.I.=2.07, 4.35). BEV was not associated with an increased risk of CHF or CM among patients aged 65-74 years or in the absence of prior cardiac disease. Conclusions: The risk of CHF or CM with BEV was higher in this elderly, general oncology population than reported in randomized trials. These risks are substantially greater in patients over age 75 and those with prior history of cardiac disease. This suggests the need for more cautious selection of patients for BEV in general practice to avoid these harms.