Implications of Alternative Definitions of Peri-Procedural Myocardial Infarction After Coronary Revascularization

John Gregson, Gregg W. Stone, Ori Ben-Yehuda, Björn Redfors, David E. Kandzari, Marie Claude Morice, Martin B. Leon, Ioanna Kosmidou, Nicholas J. Lembo, W. Morris Brown, Dimitri Karmpaliotis, Adrian P. Banning, Jose Pomar, Manel Sabaté, Charles A. Simonton, Ovidiu Dressler, Arie Pieter Kappetein, Joseph F. Sabik, Patrick W. Serruys, Stuart J. Pocock

Research output: Contribution to journalArticlepeer-review

79 Scopus citations


Background: Varying definitions of procedural myocardial infarction (PMI) are in widespread use. Objectives: This study sought to determine the rates and clinical relevance of PMI using different definitions in patients with left main coronary artery disease randomized to percutaneous coronary intervention (PCI) versus coronary artery bypass grafting (CABG) surgery in the EXCEL (Evaluation of XIENCE versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization) trial. Methods: The pre-specified protocol definition of PMI (PMIProt) required a large elevation of creatine kinase-MB (CK-MB), with identical threshold for both procedures. The Third Universal Definition of MI (types 4a and 5) (PMIUD) required lesser biomarker elevations but with supporting evidence of myocardial ischemia, different after PCI and CABG. For the PMIUD, troponins were used preferentially (available in 49.5% of patients), CK-MB otherwise. The multivariable relationship between each PMI type and 5-year mortality was determined. Results: PMIProt occurred in 34 of 935 (3.6%) patients after PCI and 56 of 923 (6.1%) patients after CABG (difference −2.4%; 95% confidence interval [CI]: −4.4% to −0.5%; p = 0.015). The corresponding rates of PMIUD were 37 (4.0%) and 20 (2.2%), respectively (difference 1.8%; 95% CI: 0.2% to 3.4%; p = 0.025). Both PMIProt and PMIUD were associated with 5-year cardiovascular mortality (adjusted hazard ratio [HR]: 2.18 [95% CI: 1.13 to 4.23] and 2.87 [95% CI: 1.44 to 5.73], respectively). PMIProt was associated with a consistent hazard of cardiovascular mortality after both PCI and CABG (pinteraction = 0.86). Conversely, PMIUD was strongly associated with cardiovascular mortality after CABG (adjusted HR: 11.94; 95% CI: 4.84 to 29.47) but not after PCI (adjusted HR: 1.14; 95% CI: 0.35 to 3.67) (pinteraction = 0.004). Results were similar for all-cause mortality and with varying PMIUD biomarker definitions. Only large biomarker elevations (CK-MB ≥10× upper reference limit and troponin ≥70× upper reference limit) were associated with mortality. Conclusions: The rates of PMI after PCI and CABG vary greatly with different definitions. In the EXCEL trial, the pre-specified PMIProt was associated with similar hazard after PCI and CABG, whereas PMIUD was strongly associated with mortality after CABG but not after PCI.

Original languageEnglish
Pages (from-to)1609-1621
Number of pages13
JournalJournal of the American College of Cardiology
Issue number14
StatePublished - 6 Oct 2020


  • coronary artery bypass grafting surgery
  • coronary artery disease
  • left main disease
  • myocardial infarction
  • percutaneous coronary intervention
  • prognosis
  • revascularization


Dive into the research topics of 'Implications of Alternative Definitions of Peri-Procedural Myocardial Infarction After Coronary Revascularization'. Together they form a unique fingerprint.

Cite this