Causes and predictors of long-term mortality in patients referred to a cardiac rehabilitation program

Abstract Introduction Cardiac rehabilitation (CR) programs have proven to reduce short-term mortality in patients with ischemic heart disease (IHD) and hospitalization rates in patients with heart failure (HF) (1-3). However, there is insufficient evidence about long-term mortality (>24-36 months) for patients receiving treatments based on recent clinical practice recommendations. Purpose To assess the long-term prognosis and predictors of mortality in a heterogenous group of patients referred to cardiac rehabilitation, and to obtain a detailed overview of death causes. Methods An inferential analysis of consecutive patients referred for exercise-based CR at a single site between January 2016 and January 2021 was performed. Included subjects were classified attending to four indications for referral: IHD, HF, valvular heart disease (VHD) and cancer therapy-related cardiovascular toxicity (CTR-CVT). Patients who did not initiate the CR program or were referred for other indications were excluded. The primary endpoint was all cause death. The secondary endpoint was cardiovascular (CV) death. A survival analysis was performed using Kaplan Meier estimates for each referral indication and a Cox proportional hazards model was fitted to investigate mortality predictors. The initiation of the exercise program was considered the inclusion date, and patients were followed until they met the primary endpoint or censored by January 2024. Results A total of 1180 patients were included in the analysis (81% male, mean age 61 ± 11 years), with a median follow-up of 5.3 years. Most of the patients had been referred due to IHD (87%), while the proportion for HF, VHD, and CTR-CVT was 9.9%, 1.8% and 1.4%, respectively. Seventy-eight patients (6.6%) died during follow-up, with a Kaplan-Meier estimated survival of 94.5% at 5 years (95% CI 93.2%-95.9%). Nineteen patients (24.4%) died due to a cardiovascular event, mostly sudden cardiac death (47%). The predominant non-CV causes of death were infectious disease (36%) and cancer (34%). A multivariate Cox model analysis (Table 1) showed an increased risk of all cause death for patients referred due to HF (HR 2.95, 95% CI 1.33 - 6.55), VHD (HR 5.55, 95% CI 1.19 - 25.9) and CTR-CVT (22.47, 95% IC 7.32 - 68.95) when compared to IHD (Figure 1). Other significant predictors of mortality were LVEF <40%, male sex, age, diabetes, peripheral vascular disease, and multivessel coronary disease. Only LVEF <40%, VHD, CTR-CVT and peripheral vascular disease were independent predictors of CV death. Conclusion Patients who initiated an exercise-based CR program had a low risk of all cause death at 5 years. Indications for CR other than IHD were associated with higher mortality rates after adjusting for confounder variables, likely due to a selection bias of patients in advanced stages of the disease.Table 1.Cox models for all-cause deathFigure 1.KM plot for all-cause death