Cardiac mortality, adequate power, and objective inclusion of the entire evidence are key to accurately define the long-term effect of revascularisation vs. medical therapy alone in stable coronary syndromes

This commentary refers to ‘Cardiac mortality in patients randomised to elective coronary revascularisation plus medical therapy or medical therapy alone: a systematic review and meta-analysis’, by E.P. Navarese et al., https://doi.org/10.1093/eurheartj/ehab246; ‘When a meta-analysis equals a single large-scale trial with meaningful follow-up’, by E.P. Navarese et al., https://doi:10.1093/eurheartj/ehab460; and the discussion pieces ‘Cardiac death should be the primary endpoint for revascularisation trials and meta analyses’, by H.D. White, https://doi.org/10.1093/eurheartj/ehab676; ‘In the pool: dilution or drowning?’, by V. Dayan et al., https://doi.org/10.1093/eurheartj/ehab443. Meta-analyses are fundamental tools placed at the top of the pyramid of evidence-based medicine. A recent meta-analysis by us reports a significant 21% proportional risk reduction in cardiac mortality by revascularisation plus medical therapy (MT) vs. MT alone in clinically stable coronary artery disease (CAD) patients, with benefits related to follow-up duration and to prevention of spontaneous myocardial infarction (MI).1 Dr White compellingly explains why appropriate primary endpoint definition in revascularisation trials and meta-analyses should be cardiac (or cardiovascular) and not all-cause mortality.2 We share Dr White’s arguments. The paradigmatic ISCHEMIA trial and many other elected cardiovascular (not all-cause) mortality as component of their primary endpoint. In our analysis, cardiac mortality was the prespecified primary endpoint (preregistered in PROSPERO’s database as CRD42021225598) to avoid masking potential efficacy signals by noncardiac deaths.1 We agree with Dr White on the competing risk of noncardiac modes of death that accrue over time, impacting longer more than shorter follow-up.2 In our analysis, the mean follow-up was 5.7 years with cardiac deaths representing 39% of all deaths.1 Trends in causes of death up to five years after percutaneous coronary intervention have shown proportional increases of noncardiac compared to cardiac deaths over the decades (regardless of patient setting, age, single or multivessel disease),3 diluting specific effects of interventions on cardiac death if only all-cause mortality is considered. In our analysis, the effect of revascularisation vs. none on all-cause mortality was consistent with that on cardiac death, but—as expected—attenuated.1 For these reasons, all-cause mortality, as advocated in a recent not externally peer-reviewed,4 would be misleading to study the longterm effects of revascularisation vs. medical therapy in CAD patients. Adequate statistical power (e.g. sample size and follow-up) are mandatory for precise cardiac mortality estimates. Single stable CAD revascularisation trials are all underpowered for cardiac mortality as individual endpoint. Similarly, to provide meaningful conclusions, meta-analyses need adequate power. Subjective and partial inclusion of evidence into meta-analyses represents a gross methodological error strongly discouraged by international guidelines.5 Compared to our analysis,1 the abovementioned recent article chose to meta-analyse trials from 2007 (considered a benchmark year for effective MT), discarding 11 trials (nine before and two after 2007), 44% of the entire evidence and 30,855 person-years, producing follow-up contraction to 3.5 years and neutral cardiac mortality results for revascularisation + MT vs. MT alone.4 Our analysis of 19 806 patients (17 454 for cardiac mortality) objectively included the entire pertinent evidence according to prespecified criteria.1 Trial chronology did not impact the primary findings.1 Statistical heterogeneity was low (I2 = 21%) for both cardiac mortality and spontaneous MI.1 Heterogeneity for cardiac death was explained by variable follow-up.1 Robust trial sequential analysis mandated a sample of ≥15 234 patients with adequate follow-up to cross the monitoring boundaries for efficacy, confirming that the results for cardiac mortality were conclusive.1 We thank Dr White for providing insightful comments. Cardiac mortality, adequate power, and objective inclusion of the entire evidence are key to accurately define the long-term effect of revascularisation vs. medical therapy alone in stable coronary syndromes. Conflict of interest: E.P.N. reports research grants from Abbott and Amgen and lecture fees/honoraria from Amgen, AstraZeneca, Bayer, Pfizer, and Sanofi-Regeneron, outside the submitted work. F.A. reports lecture/consultancy fees from Amgen, Bayer, BMS/Pfizer, and Daiichi Sankyo, outside the submitted work.