Impact of body mass index on survival and serious adverse events in advanced non-small cell lung cancer treated with bevacizumab: a meta-analysis of randomized clinical trials

Abstract

Abstract Purpose Lung cancer accounts for 28% of all cancer deaths, more deaths than any other cancer in the United States. The influence of body composition has been evaluated in several studies, specifically, the influence of obesity on lung cancer survival. Outcomes have been mixed, with some studies demonstrating a paradoxical beneficial effect in early lung cancer where survival is improved in obese patients. The study aim was to evaluate the impact of obesity on overall survival (OS), progression free survival (PFS), and occurrence of serious adverse events (SAE) in clinical trials evaluating bevacizumab for advanced non-small cell lung cancer (NSCLC). Methods We performed a post hoc analysis combining available individual level data from bevacizumab randomized clinical trials available through the Clinical Study Data Request database. The primary outcome measured in our analysis was the influence of bevacizumab on OS stratified by body mass index (BMI). In addition to OS, both PFS and the occurrence of SAE requiring therapy interruption were evaluated. All endpoints were evaluated in patients who were obese (BMI ≥30.0\u2009kg/m2) compared with non-obese (BMI <30.0\u2009kg/m2). As a sensitivity analysis, endpoints were also evaluated in patients who were overweight (BMI ≥25.0\u2009kg/m2) compared with non-overweight (BMI <25.0\u2009kg/m2). In addition to analysis of each individual study, a meta-analysis was performed in order to calculate pooled hazard ratios (HR). Hazard ratios for both OS and PFS were calculated using multivariable Cox proportional hazards models. Odds ratios for SAE were calculated using multivariable logistic regression. The validity of the regression models was tested using a log-log plot and overall fit using the goodness of fit test. Results After adjusting for covariates using a Cox proportional hazards model and combining the resulting adjusted hazard ratios using meta-analysis, there was no significant difference between obese and non-obese groups for OS or PFS. In addition, when treatment discontinuation due to an adverse event was assessed, none of the trials showed a significant difference between the obese and non-obese groups. Conclusion In this analysis of clinical trial data, obesity was not associated with worse survival versus non-obese individuals in advanced NSCLC. In addition, serious adverse events were similar between patients with and without obesity.