Impact of cardiac comorbidity on use and outcomes of adjuvant chemotherapy (ADJ) for colorectal cancer (CRC): A real-world population-based study.

Abstract

491 Background: Cardiac comorbidities such as myocardial infarction (MI) and congestive heart failure (CHF) may pose challenges in the treatment of CRC. As the population ages, cancer patients (pts) will be increasingly affected by cardiac comorbidities. We performed a population-based analysis of CRC to evaluate the prevalence of MI and CHF, use of ADJ, and survival outcomes. Methods: We evaluated 8601 pts diagnosed with resected stage 2 or 3 CRC from 2004 to 2015 in Alberta, Canada. Baseline patient, tumor, and treatment characteristics were compared between those with and without MI or CHF. Survival analysis was conducted using Kaplan-Meier methods and Cox regression models. Results: In total, 506 (5.9%) patients (pts) had MI and 440 (5.1%) had CHF. CRC patients with prior MI or CHF were older (median 76 and 79 years, respectively) and had worse Charlson Comorbidity Index (median CCI 2 for both) than those without cardiac comorbidities (median age 67 and CCI 0) (p < 0.001). Only 24% and 15% of pts with a MI or CHF history, respectively, received ADJ when compared to their counterparts (52% and 53%, respectively, p < 0.001). Among those who received ADJ (N = 3409), an oxaliplatin-based regimen was used in 26% of MI pts versus 42% of those without MI (p = 0.002), and in 31% of CHF pts versus 42% of those without CHF. Kaplan-Meier analysis revealed significantly worse overall survival (OS) in pts with prior MI (9.1 vs 4.3 years, p < 0.001) or CHF (9.2 vs. 2.7 years, p < 0.001) when compared to those without. However, cancer-specific survival (CSS) was not statistically different with or without MI (p = 0.348) and with or without CHF (p = 0.611). In Cox regression that adjusted for use of ADJ, MI was no longer a significant predictor of OS (HR = 1.01, 95% confidence interval (CI) 0.88-1.15), but CHF remained significant (HR 0.65, 95% CI 0.57-0.74). Neither MI nor CHF were predictors of CSS (HR 1.09, 95% CI 0.98-1.33, and HR 0.94, 95% CI 0.77-1.15). Conclusions: CRC pts with MI or CHF experienced lower use of ADJ and worse OS, but no difference in CSS was observed. ADJ-treated pts with prior MI appeared to benefit while worse outcomes in pts with prior CHF appear to be driven by non-cancer related causes.