Interpreting Myocardial Infarction Analyses in ISCHEMIA

Separating Facts From Fallacy

Raffaele De Caterina; David L. Brown


Eur Heart J. 2021;42(31):2986-2989. 

In This Article

Problems in Downplaying Periprocedural Myocardial Infarction

An advancing narrative in the aftermath of ISCHEMIA is that, compared to 'spontaneous' MI, periprocedural MI events are inconsequential and prognostically unimportant. ISCHEMIA, for unknown reasons, used a primary definition of periprocedural MI that required, for PCI, an increase of creatine kinase (CK)-MB >5 times the upper reference limit (URL) (preferred) or an increase in cardiac troponin (cTn) >35 times the URL with either new electrocardiographic changes or angiographic evidence of reduced flow or coronary dissection. In the absence of ancillary findings, CK-MB and cTn had to be >10 times and >70 times the URL, respectively. For CABG-related MI, definitions were even more stringent.[1] Thus, the frequent patient with 'only' an increase in troponin T (URL: 14 ng/mL)—say, from 6 to 910 ng/mL (= URL × 65)—would not be classified as having suffered an MI.

ISCHEMIA also considered alternative 'secondary' definitions of MI, consistent with the Third Universal Definition of Myocardial Infarction (UDMI):[11] in patients with normal baseline cTn (as in most patients with CCS), elevations of cTn >5 and >10 times the URL occurring within 48 h of PCI or CABG, respectively—plus ancillary findings—would be defined as a Type 4a or 5 MI, respectively—not just 'myocardial injury'. These 'secondary' definitions, recently reiterated in the 4th UDMI[12] based on expert consensus in the absence of definitive trial-based data, have now been corroborated by a recent analysis from the SYNTAX trial[13] and a meta-analysis of all PCI trials reporting on cardiac biomarkers in patients with CCS,[14] both concluding that such post-procedural troponin elevations are indeed associated with mortality. When the UDMI-consistent secondary ISCHEMIA definition is adopted, Type 4a MIs in the invasive arm in ISCHEMIA would more than triple, from 26 to 98.[1] Using the secondary MI definition, there would be overall 106 Type 1 and Type 2 spontaneous MIs in the invasive arm and 186 in the conservative arm, but this would be counterbalanced by far many more periprocedural (Types 4a and 5) and late procedure-related (Types 4b and 4c) MIs in the invasive arm (224 vs. 44): in aggregate, 330 MIs in the invasive strategy vs. 230 MIs in the conservative strategy[1] (Graphical abstract). The occurrence of some procedural MIs in the conservative arm was due to some cross-over from the conservative to the invasive arm during the course of the trial. In any case, in reality, according to current UDMI definitions, more MIs with adverse downstream effects on mortality occurred in the invasive than in the conservative arm.