The prognostic impact of periprocedural myocardial infarction and injury

This editorial refers to ‘Periprocedural myocardial infarction and injury in elective coronary stenting’†, by M. Zeitouni et al., on page 1100. The appropriate metrics with which to diagnose periprocedural myocardial infarction (MI) have been controversial for years. The original appreciation of elevations in biomarkers, at that time CK-MB (creatine kinase-MB), was described in the 1980s.1 , 2 It was thought that elevations would not occur in the absence of complications with percutaneous coronary intervention (PCI). Subsequently, patients presenting with acute coronary syndrome were studied who had putatively normal CK-MB values at baseline and increases were found after PCI.3 This then led to the presumption that myocardial injury occurring during PCI was of prognostic importance since patients with elevated biomarker values after PCI seemed to be at increased risk.4 Later, it was realized that many of those increases in reality came from the original insult that was not detected with an insensitive biomarker such as CK-MB but was detected when cardiac troponin (cTn) was measured.5 It also became apparent that the prognostic significance of post-PCI cTn elevations was ablated when the baseline cTn value was included in the analysis.5 This construct was easily missed if an inappropriate cut-off of the baseline cTn values was used, which often happened in the early days of cTn and, for that reason, the data for the cTn evaluations were disparate in part because of the varying baseline values that were employed.6 This issue was eagerly discussed when preparing the consensus document on the Redefinition of Myocardial Infarction in 2000.7 At that time it was believed that increases of cardiac biomarker values during PCI were indicative of myocardial injury, most probably due to myocardial ischaemia and so, combined with a rising and/or falling pattern, should lead to the diagnosis of an acute MI.7 PCI operating cardiologists were infuriated because, viewed from their perspective, even small increases in post-PCI cTn values were called a periprocedural MI, irrespective of the success or failure of the procedure, merely due to the cTn elevation. This controversy gave rise to an advocacy for a higher biomarker threshold [three times the 99th percentile upper reference limit (URL)] when developing the second Universal Definition of MI (UDMI).8 However, the complaint still was that cardiologists wanted a stricter clinical approach to determine whether a periprocedural MI has occurred at all and whether the clinical outcomes could be ascribed to the post-procedural values or those that were measured before the procedure. This decision is often difficult in patients with an acute MI undergoing PCI, since it is hard to determine how much of the myocardial injury is related to the acute MI and how much is induced by the PCI itself. This point at issue may also arise in patients with chronic cTn elevations who undergo PCI because these elevations are a reflection of the magnitude and complexity of the underlying coronary artery disease and co-morbidity and thus reduce the prognostic significance of post-PCI values.9 This disputed point led some to suggest that different thresholds should be utilized to define myocardial injury and to strengthen the criteria for defining PCI-related MI. Thus, the task force of the third UDMI recommended an increase of the biomarker (preferably cTn) threshold to five times the 99th percentile URL in patients with normal baseline cTn values or a rise of cTn values >20% if the baseline values were elevated. In addition, it required either symptoms suggestive of myocardial ischaemia, new ECG changes indicative of ischaemia, coronary angiographic-documented complications, or imaging demonstration of new loss of viable myocardium.10 Another group involved in coronary angiography and interventions suggested as an alternative a substantial stand-alone increase of the cut-off point, preferably CK-MB, to diagnose PCI-related MI (the SCAI criteria) in order to increase the specificity and the prognostic significance of this entity.11 However, these definitions do not substantiate whether the prognosis linked to the biomarker pattern after PCI is related to the procedure itself or to a preceding event. This uncertainty has not been solved in the past with the older biomarkers, and now there is the question: are we getting closer to an explanation when the new high sensitivity (hs)-cTn assays are in use for diagnosing myocardial injury and infarction? Zeitouni et al. have tackled this problem in their manuscript in this issue of the journal.12 What they did was to exclude patients with pre-procedural elevation of hs-cTnT values in order to avoid any relationship to possible post-PCI elevations. In doing so, interesting observations were generated. They studied 3416 PCI procedures, 48.7% of which were performed in an elective setting with a normal hs-cTnT value at baseline. The post-PCI hs-cTnT measurements were available in 1390 of these individuals, who were derived from an all-comers population that included a fairly sick group involving patients with PCI for chronic total occlusion (6.8%) and allograft vasculopathy in heart transplantation patients (4.2%). The authors then probed the diagnosis of periprocedural MI with the third UDMI criteria and subsequently evaluated the prognostic significance in those patients with de novo post-PCI hs-cTnT elevations. Next, they compared those results with the prognostic impact of elevated post-PCI hs-cTnT values using the second UDMI definition,8 which advocated for a lower multiple of the 99th percentile URL, and the SCAI criteria that demand a much higher cut-off point for post-procedural biomarker values.11 When employing the third UDMI criteria within 48 h after elective PCI, 7.0% of the patients had periprocedural MI and 21.6% myocardial injury.12 The latter was diagnosed in situations where either signs or symptoms of ischaemia were absent despite elevation of hs-cTnT values or where the increase did not reach five times the 99th percentile URL. The rate of periprocedural MI using the criteria of the second UDMI was substantially higher (43.8%) and much lower by the SCAI criteria (1.5%). Importantly, there was a five-fold increase in cardiovascular events driven by ischaemia in patients who had periprocedural MI or myocardial injury at 30 days diagnosed with the third UDMI criteria. Taken separately, periprocedural MI and myocardial injury both had a significant impact on cardiovascular events at 30 days, with high hazard ratios of 6.6 for periprocedural MI and 4.4 for myocardial injury. At 1 year, the risk of ischaemic events after discharge was also increased significantly in those with periprocedural MI and myocardial injury. The outcomes with periprocedural MI reached statistical significance, whereas those associated with myocardial injury did not.12 Interestingly, patients with a post-PCI increase of hs-cTnT values below the cut-off point of 70 ng/L (third UDMI criteria) and without evidence of acute ischaemia had a similar rate of events as patients who had no hs-cTnT elevations at all. The SCAI criteria provided a greater percentage increase in events at 30 days and 1 year, but this was at the expense of an important drop in sensitivity.12 It appeared that the optimal threshold using hs-cTnT to predict cardiovascular adverse events at 30 days and 1 year was very close to the five-fold increase suggested by the third UDMI.10 , 12 By eliminating patients with elevation of pre-procedural hs-cTnT values, the present data are the first to document an adverse influence of periprocedural MI and myocardial injury as determined by hs-cTnT on ischaemic events up to 1 year after PCI.12 That may have implications for the selection and duration of the post-PCI treatment in elective PCI patients. How should we consider these new results and why do they differ so markedly from the findings of Ndrepepa et al., who studied the prognostic impact of hs-cTnT elevation after elective PCI in 5626 patients with or without raised baseline values of hs-cTnT?13 These authors did not find any prognostic importance, assessed by all-cause mortality at 3 years, of post-procedural hs-cTnT elevations in 2721 patients compared with 742 patients without pre-procedural elevations.13 Sure enough, it is tricky to compare demographics, and there were some modest differences between the two studies.12 , 13 However, the long-term 3-year overall mortality rate in the study of Ndrepepa et al. for patients with a normal baseline hs-cTnT value was only 3.8%, which seems lower than the 2.3% cardiovascular short-term mortality rate reported by Zeitouni et al. at 1 year.12 , 13 Was revascularization less extensive? Was medical therapy less aggressively applied? Several things do strike us as critical in analysing these data. The results presented by Zeitouni et al. suggest that there may be important short-term prognostic significance in periprocedural hs-cTnT elevations in elective PCI patients who are not confounded by an increase of hs-cTnT values at baseline.12 However, these short-term clinical outcomes of elevated periprocedural hs-cTnT values seem to vanish in the long term when considering the 3-year mortality that includes causes of death other than those strictly related to a coronary procedure.13 Unfortunately, these data for hs-cTnT may not be easily extrapolated to other hs-cTn assays or to less sensitive cTn assays. Nevertheless, if confirmed, the hs-cTn measurements have opened up an important new gateway not only where periprocedural elevations can be detected in an orderly manner but also where there may be a prognostic significance that so many have desperately been seeking. Conflict of interest: K.T. has no conflicts of interest. A.S.J. consults outside the submitted work for most of the major diagnostic companies.