SCAI-Defined Perioperative MIs Too Numerous to Be Useful in Cardiac Surgery

The prognosis of patients who have a myocardial infarction during or after coronary artery bypass surgery may depend on how those periprocedural events are defined, according to the results of a new study.

In a cohort of consecutive patients who underwent CABG surgery, periprocedural MIs documented with the Fourth Universal Definition of Myocardial Infarction (UDMI) were associated with a significantly higher risk of death at 1 year, but events diagnosed using the Society for Cardiovascular Angiography and Interventions (SCAI) definition proved far less meaningful.

The data, which were published last week in the European Journal of Cardio-Thoracic Surgery, raise questions about how to best diagnose periprocedural MI, particularly for cardiac surgery, in the era of high-sensitivity troponin (hs-cTn) assays. For example, 87.6% of CABG patients had a periprocedural MI based on the SCAI definition, which is far too high to be useful for physicians or researchers.

“In clinical practice, and in research in the future, most centers will only be offering high-sensitivity troponin testing,” senior investigator Nick Curzen, BM, PhD (University Hospital Southampton NHS Foundation Trust, England), told TCTMD. “What this small study does is highlight how important it is to look at the definitions of periprocedural MI, because otherwise interpreting the implication of a raised high-sensitivity troponin will be of little clinical value.”

Based on their analysis, the SCAI definition, which was derived in the era when creatine kinase-myocardial band (CK-MB) was the preferred biomarker, does not appear “to be fit for practice, because so many people would be defined as having an MI that it wouldn’t mean anything,” said Curzen. “In some ways, there’s not that much point in diagnosing a periprocedural MI if it’s not a risk factor for adverse outcomes.”

In some ways, there’s not that much point in diagnosing a periprocedural MI if it’s not a risk factor for adverse outcomes. Nick Curzen

Patrick Myers, MD (Lausanne University Hospital, Switzerland), secretary-general elect of the European Association for Cardio-Thoracic Surgery (EACTS), said that several studies have now shown that the SCAI definition captures a far higher number of periprocedural MI events among surgery patients—somewhere in the order of three to four times more—than definitions that require accompanying evidence of myocardial damage, like the UDMI.  

While the new study won’t help the research community determine the best perioperative MI definition, it highlights the challenges for physicians and researchers trying to make sense of enzymatic spikes that occur during CABG surgery. Like Curzen, Myers said that hs-cTn assays might be capturing enzymatic leaks not reflective of real myocardial damage. 

“It may be an issue that our assays are just so sensitive that we’re getting to the point where we’re detecting many more perioperative MIs than are actually happening,” he said, adding this is especially problematic for the SCAI definition. “We still don’t know what the best definition is for perioperative MI post-cardiac surgery, because we don’t have very good data at the moment,” he said.

Issam Moussa, MD, MBA (Carle Health/University of Illinois Urbana-Champaign), who led the SCAI consensus group that formulated the periprocedural MI definition, said that defining the optimal threshold of cardiac biomarker elevation after coronary revascularization has been an area of controversy for a long time. He stressed that while conventional troponin can be used to diagnose MI if CK-MB is not available, that doesn’t mean hs-cTn can be swapped in.

For that reason, hs-cTn should not be used to alter decision-making after coronary revascularization until there are studies to understand its prognostic significance, said Moussa. “I can tell you that as high-sensitivity troponin was being approved in the US, there were major concerns that people would be using [the assays] post-PCI and postbypass without knowing what they mean.”

Gregg Stone, MD (Icahn School of Medicine at Mount Sinai, New York), another member of the SCAI consensus group, agreed that hs-cTn is too sensitive for the assessment of periprocedural MI. When using CK-MB or standard troponin assays, or a combination of the two, the rates of periprocedural MI using the SCAI definition are less than 10% and correlate strongly with 5-year mortality. 

“The problem from this study was the biomarker not the concept,” he told TCTMD.   

Moussa pointed out that while high-sensitivity troponin has been approved in Europe for over a decade, it’s never been validated or approved to detect post-CABG and post-PCI MI. “If physicians say, ‘Well, our lab will never use conventional troponin again, or never use CK-MB again,’ then I would say we need a well-designed national or multinational study to validate [hs-cTn’s] role after not only bypass but also PCI.”

That study would require at least 3 to 5 years of follow-up and would preferably require an MRI substudy to document myocardial damage, said Moussa. In 2021, the European Society of Cardiology and European Association of Percutaneous Cardiovascular Interventions drafted a consensus statement for diagnosing type 4a MIs—a PCI-related MI within 48 hours of the procedure—but there is yet no consensus for diagnosing infarctions associated with CABG surgery (type 5).

Not Fit for Clinical Practice

The Fourth Universal Definition defines periprocedural MI as a cTn rise > 10 times the manufacturer-defined upper limit of normal (ULN), in addition to ECG, imaging, or angiographic evidence of infarction, within 48 hours of CABG surgery. The SCAI definition defines periprocedural MIs as a cTn increase > 70 times ULN (or > 35 times ULN with new ECG changes within 48 hours of CABG). Neither definition was developed in the era of high-sensitivity troponin assays.

The present study is a post hoc analysis of a larger observational study looking at the value of measuring hs-cTn in cardiac surgery patients admitted to the cardiothoracic critical care unit (CCCU). Normally these patients wouldn’t have their hs-cTn measured in that setting, but the project involved tracking peak hs-cTn over a few days in the CCCU and this allowed researchers to assess the incidence and prognostic relevance of the two different MI definitions, said Curzen.   

For 245 consecutive patients who underwent CABG, 20.4% had a periprocedural MI as per the Fourth Universal Definition. At 1 year, mortality was 14.0% for those who had a periprocedural UDMI compared with 3.1% for those who did not (P = 0.002). Overall, a periprocedural UDMI was associated with a more than fourfold higher risk of all-cause mortality at 1 year in multivariable regression analysis (HR 4.16; 95% CI 1.28-13.49).

In contrast, 87.6% of bypass patients had a periprocedural MI based on the SCAI definition, the majority of whom had a hs-cTn concentration > 70 times ULN. At 1 year, there was no difference in all-cause mortality between those who did and did not have a SCAI-defined periprocedural MI. 

The study also included 243 patients who underwent open non-CABG cardiac surgery. The periprocedural MI definition is meant to only apply to patients undergoing CABG, but investigators report that 11.1% and 85.2% of these non-CABG patients had a UDMI and SCAI-defined periprocedural MI, respectively. Events with both definitions were not associated with 1-year mortality.

As high-sensitivity troponin was being approved in the US, there were major concerns that people would be using [the assays] post-PCI and postbypass without knowing what they mean. Issam Moussa

To TCTMD, Myers said the incidence of periprocedural UDMIs in the CABG patients is a little high and somewhat surprising, hypothesizing that the types of procedures, including the complexity of surgery, cross-clamp times, and use of myocardial protection strategies, might have had an impact on this rate. Roughly 40% of the CABG patients treated in the study were unplanned admissions, both Moussa and Myers pointed out, and these critical patients might be expected to have higher troponin concentrations before the procedure.

“If we’re just measuring their troponin when they enter the ICU, or their peak, we risk diagnosing a perioperative MI when there’s actually nothing there,” Myers said. For this reason, it may be important to track changes from baseline as opposed to peak hs-cTn to capture periprocedural infarctions, something that some clinical trials have incorporated into their MI definitions, he said.  

Moussa also pointed to what he considers to be some inconsistencies in the data: while peak hs-cTn concentration in the admitted surgical patients was independently associated with 1-year mortality, the SCAI definition was not, even though the latter allows patients to be diagnosed with periprocedural MI on the basis on biomarkers alone.

Additionally, he noted that while SCAI-defined MIs were not associated with mortality at 1 year, there were no deaths among patients who didn’t have a SCAI-defined periprocedural infarction. “I’m not saying it’s evidence, but it’s curious that using the SCAI definition, if you did not have an MI, patients didn’t die at 1 year,” said Moussa. “Obviously, this is just hypothesis-generating—it’s not proof of anything because of the small numbers [of patient deaths].” 

Need to Look at Definitions Again

For Curzen, the study raises an important issue for research, noting that the EXCEL trial—a comparison of CABG surgery versus PCI for left main coronary artery disease—was highly contentious because of numerous questions about how periprocedural MIs were captured in the trial.

In EXCEL, the rate of periprocedural MI after PCI and CABG varied substantially—and consequently so did the primary endpoint—depending on how periprocedural infarctions were defined. Notably, the EXCEL researchers used a modified SCAI definition—one they have argued is prognostically relevant—but were roundly criticized for not publishing the complete UDMI data.

If we’re just measuring their troponin when they enter the ICU, or their peak, we risk diagnosing a perioperative MI when there’s actually nothing there. Patrick Myers

“The EXCEL debate is a good example of how this definition can have enormous implications for the interpretation of the outcome of study interventions,” said Curzen.

Recently, ISCHEMIA investigators published data also showing that the results hinged on how MI was defined in the study. When the secondary MI endpoint was defined using cTn instead of CK-MB, the conservative strategy of guideline-directed medical therapy was associated with a significantly lower risk of the study’s primary endpoint, a difference that was driven by more procedural MIs among patients randomized to the invasive strategy. SYNTAX investigators also showed that event rates in that trial were dependent on how MI was defined. Some have even wondered if it might be time to abandon periprocedural MI as an endpoint because it’s so easy to misconstrue.

To TCTMD, Curzen said the study also suggests that in this era of hs-cTn assays, it might be “time to go back to the drawing board” in order to determine the best periprocedural MI definition for cardiac surgery.

“I do wonder if this requires a group of experts to look at it and maybe come to a consensus,” he said. “It’s so important that interventionalists and surgeons agree prior to large-scale studies, because otherwise the potential for disagreement is very high, as we’ve seen. That’s not a criticism of the EXCEL investigators—it’s more that I think the problems of having a robust definition of periprocedural MI are compounded by the high-sensitivity troponin assays.”

Myers agreed, adding that clinical trials, like any future consensus statements, would benefit from multiple viewpoints. 

“It’s always easy to criticize a trial after the fact, but I would advocate for having as many stakeholders as possible—interventional cardiologists, noninterventional cardiologists, surgeons—actively involved to propose or choose definitions for these comparative trials in ischemic heart disease,” he said. “We need a clear consensus before starting off so we can then reach a more defensible and better answer.”