Considerations for an optimal definition of procedural myocardial infarction

This commentary refers to ‘Periprocedural myocardial injury in the EXCEL trial’, by J.L. Bjørnstad and B. Bendz, 2020;41:1703. We thank Drs Bjørnstad and Bendz for their interest in our manuscript from the EXCEL trial1 2 in which we demonstrated a strong relationship between large procedural myocardial infarction (MI) after both percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG) and subsequent 3-year cardiovascular and all-cause mortality in patients with left main coronary disease (LMCAD). The distribution of creatine kinase-MB (CK-MB) levels after each procedure is shown in Figure 1 of the manuscript. As shown in Table 4 in the manuscript, the pre-specified protocol definition of procedural MI was independently predictive of 3-year all-cause mortality [hazard ratio (HR) 2.28, 95% confidence interval (CI) 1.22–1.49; P = 0.01], with similar hazards after PCI (HR 2.18) and CABG (HR 2.25), although given the number of events these relationships did not reach statistical significance. However, as expected, procedural MI was very strongly related to cardiovascular mortality, but not to non-cardiovascular mortality. In this regard, after multivariable analysis the hazard of procedural MI after CABG was independently predictive of 3-year cardiovascular death (HR 2.74, 95% CI 1.02–7.38; P = 0.047), whereas the hazard of procedural MI after PCI was not (HR 1.82, 95% CI 0.42–7.81; P = 0.42). However, given the modest number of events we caution declaring this difference definitive. Similarly, there were too few events to parse out the predictors of early vs. late mortality after each procedure with high confidence. In EXCEL, post-revascularization myonecrosis was assessed after both PCI and CABG by serial assessment of CK-MB. Troponin assessments were optional and were collected in the minority of patients, complicating determination of procedural MI using alternative definitions. However, as shown in Table 5 of the manuscript, CK-MB elevations 5–10× URL with additional electrocardiographic, angiographic, or imaging evidence of MI (which would include almost all infarctions defined by the 3rd Universal Definition not meeting the EXCEL stand-alone biomarker definition of large procedural MI) were not significantly associated with death (HR 1.44, 95% CI 0.20–10.60; P = 0.72).3 Notwithstanding the limitations of biomarker availability, further analyses reporting the rates and 5-year predictive implications of the protocol and alternative definitions of procedural MI from EXCEL will be reported later this year. In EXCEL, the higher procedural MI risk after CABG compared with PCI was counterbalanced by a higher non-procedural MI risk during follow-up after PCI compared with CABG, such that the 5-year rates of all MI were similar between the two procedures, underlying the similar 5-year risk of long-term cardiovascular mortality observed after both procedures.4 The difference in mortality in EXCEL was driven by greater non-cardiovascular mortality due mainly to cancer and sepsis arising between 1 and 5 years after PCI compared with CABG, a finding that has not previously been observed in any prior trial and lacks mechanistic plausibility. Indeed, in an individual patient data pooled analysis of 9 randomized trials (including EXCEL) in which 4478 patients with LMCAD were randomized to PCI vs. CABG, there was nearly identical mortality at 5 years (HR 1.07, 95% CI 0.87–1.33; P = 0.52).5 Finally, among 705 patients with LMCAD randomized in the SYNTAX trial, all-cause mortality at 10 years occurred in 26.1% PCI patients vs. 26.7% CABG patients (HR 0.90, 95% CI 0.68–1.20; P = 0.47).6 Thus, there would seem to be no meaningful difference in long-term survival after PCI and CABG for most patients with LMCAD in the contemporary era. Conflict of interest: Dr Stone discloses having received speaker honoraria from Cook, Terumo; having served as a consultant to Valfix, TherOx, Vascular Dynamics, Robocath, HeartFlow, Gore, Ablative Solutions, Miracor, Neovasc, V-Wave, Abiomed, Ancora, MAIA Pharmaceuticals, Vectorious, Reva, Matrizyme; having equity/options from Ancora, Qool Therapeutics, Cagent, Applied Therapeutics, Biostar family of funds, SpectraWave, Orchestra Biomed, Aria, Cardiac Success, MedFocus family of funds, Valfix; Dr Ben-Yehuda reports that his employer has received funding by Abbott Vascular; Dr Sabik has served as a consultant for Medtronic and has received grants from Edwards and Abbott Vascular; Dr Kappetein is a full-time employee of Medtronic; Dr Serruys has served as a consultant for Abbott Laboratories, AstraZeneca, Biotronik, Cardialysis, GLG Research, Medtronic, Sino Medical Sciences Technology, Société Europa Digital Publishing, Svelte Medical Systems, Stentys France, Phillips, St. Jude Medical, Qualimed and Xeltis.