Fourteen RCTs were included in the review (n=832). None of the trials were double blind. Only four reported blinding outcome assessors. Out of a possible quality score of 5, seven studies scored 1, five scored 2 and two scored 3.
Overall, exercise training significantly improved ejection fraction (weighted mean difference was 1.83; 95% CI: 0.45 to 3.21; 14 trials). Significant heterogeneity was observed in this analysis (I2=49.2%). When grouped by type of exercise, aerobic training alone significantly improved ejection fraction (weighted mean difference was 2.59; 95% CI: 1.44 to 3.74; nine trials), but strength training alone (one trial) or in combination with aerobic training (four trials) did not. There was no significant heterogeneity in the analyses where studies were pooled in these subgroups.
Overall, exercise training was associated with a significant reduction in end diastolic volume (weighted mean difference was -9.75 mL, 95% CI: -16.64 to -2.86; seven trials) and end systolic volume (weighted mean difference was -12.31 mL, 95% CI: -17.12 to -7.49; seven trials) as was aerobic training alone for end diastolic volume (weighted mean difference -11.94 mL, 95% CI: -19.95 to -3.02; five trials) and end systolic volume (weighted mean difference was -12.87 mL, 95% CI: -17.80 to -7.93; five trials), but not strength training in combination with aerobic training (two trials). None of these analyses were reported to show significant heterogeneity between studies.
Results for sensitivity analyses were reported. There was no evidence of publication bias.