BACKGROUND: Guidelines recommend exercise-based cardiac rehabilitation (EBCR) for patients with heart failure (HF). However, established research has not investigated the longer-term outcomes including mortality and hospitalisation in light of the contemporary management of HF.

METHODS: This was a systematic review including a meta-analysis of EBCR on all-cause mortality, hospital admission, and standardised exercise capacity using four separate exercise tests in patients with heart failure over a minimum follow-up of six months from January 1999-January 2013. Electronic searches were performed in the databases: Medline, CENTRAL, EMBASE, CINAHL, and PsycINFO constrained to randomised controlled trials (RCTs).

RESULTS: A total of 46 separate RCTs qualified for the meta-analysis, which employed conventional methods for binary and continuous data. The relative risk (RR) ratio for hospital admission (12 studies) was significantly reduced (RR ratio 0.65; 95% confidence interval (CI) 0.50-0.84; p = 0.001), but mortality (21 studies) was not (RR ratio 0.88; 95% CI 0.77-1.02; p = 0.08). The standardised exercise capacity (26 studies) showed a standardised mean difference (SMD) in favour of the exercise group as compared with the controls (SMD 0.98, 95% CI 0.59-1.37; p < 0.001). Women and elderly people were less frequently enrolled in the RCTs independent of the outcomes. Heterogeneity was moderate to high in the analysis of hospital admission and the standardised exercise capacity demonstrated through skewedness in their funnel plots.

CONCLUSIONS: EBCR in patients with HF is associated with significant improvements in exercise capacity and hospital admission over a minimum of six months follow-up, but not in all-cause mortality.