Exercise-based cardiac rehabilitation improves exercise capacity and health-related quality of life in people with atrial fibrillation: a systematic review and meta-analysis of randomised and non-randomised trials

Abstract

Objective: The aim of this study was to undertake a contemporary review of the impact of exercise-based cardiac rehabilitation (CR) targeted at patients with atrial fibrillation (AF). Methods: We conducted searches of PubMED, EMBASE and the Cochrane Library of Controlled Trials (up until 30 November 2017) using key terms related to exercise-based CR and AF. Randomised and non-randomised controlled trials were included if they compared the effects of an exercise-based CR intervention to a no exercise or usual care control group. Meta-analyses of outcomes were conducted where appropriate. Results: The nine randomised trials included 959 (483 exercise-based CR vs 476 controls) patients with various types of AF. Compared with control, pooled analysis showed no difference in all-cause mortality (risk ratio (RR) 1.08, 95% CI 0.77 to 1.53, p=0.64) following exercise-based CR. However, there were improvements in health-related quality of life (mean SF-36 mental component score (MCS): 4.00, 95% CI 0.26 to 7.74; p=0.04 and mean SF-36 physical component score: 1.82, 95% CI 0.06 to 3.59; p=0.04) and exercise capacity (mean peak VO₂: 1.59ml/kg/min, 95% CI 0.11 to 3.08; p=0.04; mean 6 min walk test: 46.9 m, 95% CI 26.4 to 67.4; p<0.001) with exercise-based CR. Improvements were also seen in AF symptom burden and markers of cardiac function. Conclusions: Exercise capacity, cardiac function, symptom burden and health-related quality of life were improved with exercise-based CR in the short term (up to 6 months) targeted at patients with AF. However, high-quality multi-centre randomised trials are needed to clarify the impact of exercise-based CR on key patient and health system outcomes (including health-related quality of life, mortality, hospitalisation and costs) and how these effects may vary across AF subtypes.