Association of BMI with overall survival in patients with mCRC who received chemotherapy versus EGFR and VEGF-targeted therapies

Abstract

Although a raised body mass index (BMI) is associated with increased risk of colorectal cancer (CRC) and recurrence after adjuvant treatment, data in the metastatic setting is limited. We compared overall survival (OS) across BMI groups for metastatic CRC, and specifically examined the effect of BMI within the group of patients treated with targeted therapies (TT). Retrospective data were obtained from the South Australian Registry for mCRC from February 2006 to October 2012. The BMI at first treatment was grouped as underweight <18.5 kg/m(2) , Normal = 18.5 to <25 kg/m(2) , Overweight = 25 to <30 kg/m(2) , Obese I = 30 to <35 kg/m(2) , Obese II ≥35 kg/m(2) . Of 1174 patients, 42 were underweight, 462 overweight, 175 Obese I, and 77 Obese II. The OS was shorter for patients who were underweight and overweight compared to normal (OS 13.7 and 22.3 vs. 24.1 months, respectively, hazard ratio [HR] 2.21 and 1.23). The adjusted median OS was longer for normal versus overweight or obese I patients receiving chemotherapy + targeted therapy (35.7 vs 25.1 or 22.8 months, HR 1.59 and 1.63, respectively) with no difference in OS for chemotherapy alone. On breakdown by type of targeted therapy, overweight and obese I patients had a poorer outcome with Bevacizumab. The BMI is predictive of a poorer outcome for underweight and overweight patients in the whole population. Of those receiving chemotherapy and targeted therapy, BMI is an independent predictor for OS for overweight and obese I patients, specifically for those treated with Bevacizumab. Patients who are overweight or obese (group I) may be a target group for lifestyle and nutrition advice to improve OS with TT.

Keywords: BMI; Bevacizumab; colorectal cancer; metastatic; targeted therapies.

Figure 1

Kaplan–Meier survival curves (A) and multivariate Cox proportional hazard model (hazard ratio [HR] and and 95% CI) for overall survival by BMI for the whole population (B). Model was adjusted by potential confounders: age, sex, synchronous disease, >1 met site, number of lines of chemotherapy and number of lines of antibody.

Figure 2

Kaplan–Meier survival curves (A), and Cox proportional hazard model (hazard ratio [HR], 95% CI and median overall survival [OS]) for overall survival for patients receiving chemotherapy alone versus chemotherapy and TT (B).

Figure 3

Kaplan–Meier overall survival curves (A) and multivariate Cox proportional hazard model (Hazard ratio [HR] and and 95% CI) for overall survival by BMI for patients receiving any TT (B). Model was adjusted by potential confounders: age, sex, synchronous disease, >1 met site, number of lines of chemotherapy and number of lines of TT.