IMPORTANCE - Mortality is a common outcome in trials comparing percutaneous coronary intervention (PCI) with coronary artery bypass grafting (CABG). Controversy exists regarding whether all-cause mortality or cardiac mortality is preferred as a study end point, because noncardiac mortality should be unrelated to the treatment.

OBJECTIVE - To evaluate the difference in all-cause and cause-specific mortality in randomized clinical trials (RCTs) comparing PCI with CABG for the treatment of patients with coronary artery disease.

DATA SOURCES - MEDLINE (1946 to the present), Embase (1974 to the present), and the Cochrane Library (1992 to the present) databases were searched on November 24, 2019. Reference lists of included articles were also searched, and additional studies were included if appropriate.

STUDY SELECTION - Articles were considered for inclusion if they were in English, were RCTs comparing PCI with drug-eluting or bare-metal stents and CABG for the treatment of coronary artery disease, and reported mortality and/or cause-specific mortality. Trials of PCI involving angioplasty without stenting were excluded. For each included trial, the publication with the longest follow-up duration for each outcome was selected.

DATA EXTRACTION AND SYNTHESIS - For data extraction, all studies were reviewed by 2 independent investigators, and disagreements were resolved by a third investigator in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses guideline. Data were pooled using fixed- and random-effects models.

MAIN OUTCOMES AND MEASURES - The primary outcomes were all-cause and cause-specific (cardiac vs noncardiac) mortality. Subgroup analyses were performed for PCI trials using drug-eluting vs bare-metal stents and for trials involving patients with left main disease.

RESULTS - Twenty-three unique trials were included involving 13 620 unique patients (6829 undergoing PCI and 6791 undergoing CABG; men, 39.9%-99.0% of study populations; mean age range, 60.0-71.0 years). The weighted mean (SD) follow-up was 5.3 (3.6) years. Compared with CABG, PCI was associated with a higher rate of all-cause (incidence rate ratio, 1.17; 95% CI, 1.05-1.29) and cardiac (incidence rate ratio, 1.24; 95% CI, 1.05-1.45) mortality but also noncardiac mortality (incidence rate ratio, 1.19; 95% CI, 1.00-1.41).

CONCLUSIONS AND RELEVANCE - Percutaneous coronary intervention was associated with higher all-cause, cardiac, and noncardiac mortality compared with CABG at 5 years. The significantly higher noncardiac mortality associated with PCI suggests that even noncardiac deaths after PCI may be procedure related and supports the use of all-cause mortality as the end point for myocardial revascularization trials.