Abstract 13030: Causes of Cardiovascular and Non-cardiovascular Mortality in the Ischemia Trial

Abstract

Background: In the ISCHEMIA (International Study of Comparative Health Effectiveness with Medical and Invasive Approaches) trial, all-cause mortality was similar in patients with stable ischemic heart disease (SIHD) randomized to invasive (INV) and conservative (CON) management strategies. This analysis details specific causes of cardiovascular (CV) and non-CV mortality by treatment group. Methods: In ISCHEMIA, 289 deaths occurred after a median follow-up of 3.2 years; 145 (5.6%) in INV and 144 (5.6%) in CON (HR 1.05, CI 0.83-1.32). Deaths were adjudicated by an independent Clinical Events Committee as CV, non-CV with or without a CV contributor or undetermined. The protocol defined CV death as deaths from CV causes, non-CV causes with CV contributor, and cause undetermined; non-CV death was defined as death from non-CV causes without a CV contributor. Multivariable analyses were used to identify factors associated with cause-specific death. Results: CV death was similar between groups [INV 92 (3.6%), CON 111 (4.3%); HR 0.87 (CI 0.66, 1.15)], but INV had more non-CV death [INV 53 (2.0%), CON 33 (1.3%); HR 1.63 (CI 1.06, 2.52)]; fewer undetermined deaths [INV 12 (0.5%) and CON 26 (1.0%); HR 0.48 (0.24, 0.95)] and more malignancy deaths [INV 41 (1.6%), CON 20 (0.8%); HR 2.11 (1.24, 3.61)]. In multivariable analysis, risk factors associated with CV death were age [HR 1.42 (CI 1.19-1.70) per 10-year increase], diabetes [HR 1.39 (CI 1.03-1.87)], history of heart failure [HR 1.96 (CI 1.33-2.91)], and eGFR [HR 1.18 (CI 1.11-1.26) per 5-ml/min decrease below 80ml/min]. Factors associated with non-CV death were age [HR 2.31 (CI 1.75-3.03) per 10-year increase] and randomization to INV [HR 1.76 (CI 1.13-2.75)]. Conclusions: In ISCHEMIA, all-cause mortality was similar for the INV and CON strategies. Excess non-CV deaths in INV with a higher number of deaths from malignancy but a higher number of undetermined deaths in CON requires further evaluation.