Twelve RCTs were included in the review. Eleven trials provided data on all-cause mortality (4,326 patients). The adequacy of randomisation was not reported. Blinding of outcomes was reported in four trials. Intention-to-treat data were reported in ten trials. Completeness of follow-up ranged from 94 to 100%; follow-up duration ranged from one to six years, although eight trials reported one year only.
A statistically significant increase in late mortality was reported following off-pump compared with on-pump CABG (RR 1.37, 95% CI 1.04 to 1.81; 11 trials; no statistically significant heterogeneity). Fixed-effect and random-effects models produced similar results.
Sensitivity analysis revealed that exclusion of the largest trial (with over 2,000 patients) resulted in a non-statistically significant outcome. Duration of follow-up was not significantly associated with all-cause mortality.
There was no evidence of publication bias.