Percutaneous coronary intervention (PCI) has progressively become the primary means of coronary revascularization and is considered a safe procedure with low rates of major procedural complications, and which can be performed even in ambulatory conditions. Cardiac myonecrosis characterized by an elevation of cardiac biomarkers, particularly with the implementation of high-sensitivity cardiac troponin (hs-cTn) remains the most frequent complication of PCI [1]. Periprocedural myocardial injury and infarction are essentially distinguished by the magnitude of cardiac biomarker elevation, and how to best define such complications has been the subject of intense debate as their impact on the future occurrence of major cardiovascular events (MACE) and long-term mortality may significantly vary according to the definition used. Periprocedural myocardial infarction (MI) is also called “MI associated with percutaneous coronary intervention” or “type 4a MI”, while “type 4b MI” corresponds to MI caused by stent thrombosis and “type 4c MI” to infarctions related to in-stent restenosis [2].

Several academic groups have provided expert consensus-based definitions of periprocedural myocardial injury and infarction, including different biomarkers such as creatine kinase myocardial band (CK-MB) or cardiac troponin (cTn), varying thresholds, and the requirement or not for cardiac imaging evaluation, leading to significant variations in terms of sensitivity and specificity. According to the selected definition, recent studies have reported highly variable incidence rates of periprocedural myocardial injury or infarction [2, 3, 4, 5, 6], which were associated [1, 3] or not [7, 8] with MACE and long-term mortality.

Establishing a consensual definition is of importance, as periprocedural myocardial infarction is commonly used as a component of the primary composite endpoint of major clinical trials comparing coronary revascularization methods (i.e., PCI versus coronary artery bypass grafting) and the choice of definition has been demonstrated to influence both the outcomes and the clinical meaning of observed statistical differences, causing some controversy in the medical community [9, 10, 11, 12].

In this review, we detail and compare the most frequently used definitions, and provide insights on risk factors of periprocedural myocardial injury and infarction as well as documented preventive strategies.

Different mechanisms such as acute side branch occlusion (SBO), distal embolization, abrupt vessel closure (mainly secondary to acute thrombosis or dissection), vasospasm or slow-flow/no-reflow phenomenon can be involved in the occurrence of PCI complications, with multifactorial pathophysiological pathways [13, 14].

Acute side branch occlusion is the most commonly reported cause of periprocedural myocardial infarction (MI) in PCI [15] and may result from a plaque shift or an embolization from the main vessel into the side branch, an acute thrombosis or a dissection in the side branch, or a vasospasm involving a side branch [16, 17]. It is characterized by late gadolinium enhancement (LGE) adjacent to the stent using cardiovascular magnetic resonance (CMR). Its impact on outcome depends on the importance of the occluded side branch. The risk of SBO is increased in case of ostial stenosis of the side branch before stenting, a side branch origin within the primary arterial lesion, a small diameter of the side branch, a high balloon-to-artery ratio, the target segment, the stent type and complex procedures such as chronic total occlusion or atherectomy [18, 19, 20, 21, 22, 23, 24, 25].

The second most frequent cause of type 4a MI is a distal embolization of thrombus or atheromatous material, resulting in a slow flow or no reflow phenomenon, imaged by CMR as new LGE distal to the stent. Some strategies are associated with a lower risk of such phenomenon, including primary stenting, avoidance of high-pressure balloon or stent inflation or thrombectomy [26]. Intravascular imaging may also be useful to identify lesion at risk of embolization, by showing a thin-cap fibroatheromas by optical coherence tomography (OCT) or a significant plaque burden by intravascular ultrasound (IVUS) [27].

Abrupt vessel closure can also occur, usually caused by acute thrombosis, dissection, vasospasm, balloon-induced ischemia or even air embolism.

A slow flow or no reflow phenomenon can also be involved, related to distal embolization, loss of capillary autoregulation or microvascular spasm, all resulting in endothelial dysfunction [28]. Well-known cardiovascular risk factors, such as diabetes mellitus, hypertension, active smoking, dyslipidemia, kidney failure or inflammatory processes may also be associated with no reflow [29]. Use of longer stents or a high SYNTAX score II are associated with a higher risk of slow flow or no reflow phenomenon as well [29, 30].

Apart from these causes, a post-PCI elevation of biomarkers without an identifiable cause may still occur in nearly 20% of cases [16].

High-sensitivity cardiac troponin is the most sensitive and specific cardiac biomarker for the diagnosis of periprocedural myocardial injury and type 4a MI [31]. Baseline (pre-PCI) and post-PCI cTn values should be routinely measured at 3–6 h post-PCI to detect such complications [2, 14]. In case of periprocedural flow-limiting complications or following rising post-PCI cTn values, further blood samplings at 12–24 h post-procedure may be considered to document the peak cTn values or confirm the diagnosis of type 4a MI [14]. Of note, baseline values of cTn are needed to correctly analyze any post-PCI elevations as chronic elevations may be present in about 30% of patients because of comorbidities and risk factors [32].

An electrocardiogram (ECG) after PCI is required to detect new ischemic changes: new ST-elevation at the J-point, new horizontal or downsloping ST-depression in two contiguous leads, new pathological Q waves or ST-elevation $\ge$1 mm concordant with the QRS in patients with left bundle branch block (LBBB) [2].

Transthoracic echocardiography (TTE) is also useful by showing new loss of viable myocardium or new regional wall motion abnormality, with the use of tissue Doppler imaging, speckle tracking or contrast agents to further improve sensibility, if needed. A TTE should be performed in patients with diagnosis of type 4a MI or with post-PCI cTn elevation of $\ge$5x 99th percentile upper reference limit (URL) within 48 h post-procedure [2, 14].

Presence of LGE by CMR is currently the gold-standard for the diagnosis and quantification of myocardial injury, as it can detect a mass of new irreversible myocardial injury from 0.8 to 5 g [33, 34, 35]. The amount of myocardial injury diagnosed by CMR directly correlates with post-PCI elevation of biomarkers [34] and the increased risk of MACE [33]. Two different patterns of LGE are described: new LGE adjacent to the stent, related to a side-branch occlusion, or distal to the stent, related to a distal embolization [34]. However, its use is highly restricted due to its limited availability in daily clinical practice.

Coronary angiography can show periprocedural flow-limiting complications such as coronary dissection, occlusion of a major epicardial artery, side branch occlusion or thrombus, disruption of collateral flow, or distal embolization. Intravascular imaging should be considered to better identify mechanical factors that may be responsible for coronary dissection or stent thrombosis and to clarify the pathophysiology of complications [36]. There is not always a correlation between the complications observed during the coronary angiogram and post-PCI cardiac biomarkers variations, as large complications may be associated with a non-significant biomarker elevation, while slight elevations in cardiac biomarkers may be noticed in the absence of any obvious angiographic complication. The coronary angiogram has to be carefully reviewed for subtle complications in patients with post-PCI cTn elevation of $\ge$5x 99th percentile URL within 48h post-procedure [14].

A diagnostic algorithm for periprocedural myocardial injury or infarction has been proposed in a consensus document of the European Association of Percutaneous Cardiovascular Interventions (EAPCI) [14] (Fig. 1, Ref. [37]).

Fig. 1.

Diagnostic algorithm for periprocedural myocardial injury or infarction. Adapted from Bulluck et al. [37]. cTn, cardiac troponin; ECG, electrocardiogram; MACE, major adverse cardiovascular events; PCI, percutaneous coronary intervention; URL, upper reference limit.

Over the past few years, scientific societies have published consensus statements attempting to standardize the definition of periprocedural myocardial injury and infarction (Table 1, Ref. [2, 13, 14, 38]). The main differences between these definitions are related to the type of biomarker considered, the diagnostic thresholds of theses biomarkers, and the place of additional features of ischemia (Table 2, Ref. [2, 13, 14, 38]).

Concerning the chosen biomarker, the Society of Cardiovascular Angiography and Interventions (SCAI) uses CK-MB, while the Universal Definition of Myocardial Infarction (UDMI), the Academic Research Consortium-2 (ARC-2) and the EAPCI prefer cardiac troponin. It has been demonstrated that hs-cTn is the more sensitive and specific to rule out the diagnosis of myocardial infarction but also to detect periprocedural myocardial injury and infarction, compared to CK-MB which might not be as readily available in some centers as the former [31, 39].

The diagnostic thresholds of biomarkers for myocardial injury vary from any post-PCI elevations with the UDMI and EAPCI definition to cTn $>$70x URL with the ARC-2 definition, while the SCAI definition does not specify it. Of note, the EAPCI introduces an interesting distinction between minor periprocedural myocardial injury, defined as any biomarker elevation and major periprocedural myocardial injury defined as post-PCI cTn value $>$5x 99th percentile. A recent pooled-analysis of patient-level data by Silvain et al. [37] reported that myocardial injury defined as any post-PCI elevation in high-sensitivity cTn $>$99th percentile URL occurred in almost 80% of patients and was not associated with all-cause mortality at 1 year. In this analysis, the optimal cut-off for post-PCI cTn elevation to predict 1-year mortality was found to be $>$5x 99th percentile URL, corresponding to the threshold used by the EAPCI and the 4th UDMI, which was present in 18.2% of the patients. Furthermore, the association between the occurrence of type 4a myocardial infarction and all-cause death at one year was robust (adjusted Odds ratio [aOR] 3.21, 95% confidence interval [CI] 1.42–7.27) confirming the additional prognostic value of new onset of ischemic features. Prevalence of periprocedural myocardial injury according to the ARC-2 or SCAI criteria is considerably lower, affecting $\le$2% of the patients [37, 40].

Finally, the UDMI, ARC-2 and EAPCI routinely require associated clinical, echocardiographic, or angiographic findings to establish the diagnosis of periprocedural MI, whereas SCAI only requires these findings among patients with elevated baseline values that are further rising following PCI.

Periprocedural myocardial infarction is one of the components of the primary composite endpoint of pivotal clinical trials. However, the use of different definitions across studies (Table 3, Ref. [41, 42, 43, 44, 45, 46]) have been demonstrated to influence the outcomes and the clinical impact of the observed statistical differences [9, 10, 11, 12]. This is critical in clinical trials comparing PCI with medical therapy or coronary artery bypass grafting (CABG) surgery for the treatment of multivessel or left main coronary artery disease [11, 47]. In the Xience Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial, both the SCAI and UDMI definition of periprocedural MI were collected during the trial and the stringent SCAI definition was finally chosen once released. As detailed above, this definition is the only one driven by cardiac biomarker elevation alone (without requiring additional features of ischemia) except for patients with elevated baseline values that are rising and uses the same threshold for both PCI and CABG. The conclusion of the trial after 3 years and 5 years of follow-up was that PCI was not inferior to CABG according to the composite primary outcomes of death, MI (both spontaneous and periprocedural) and stroke in patients with left main coronary artery disease of low or intermediate anatomical complexity [41, 48]. In a post-hoc analysis, the authors demonstrated that using the 4th UDMI, the rate of periprocedural MI after PCI increased from 3.6% to 4.0%, while it was reduced from 6.1% to 2.2% with CABG [9, 11]. Similar observations were made for the Taxus Drug-Eluting Stent Versus Coronary Artery Bypass Surgery for the Treatment of Narrowed Arteries (SYNTAX) trial, where the rates of periprocedural MI according to the protocol definition (CK-MB peak $>$10% or CK-MB $\ge$5x ULN associated with ECG criteria of ischemia), the 4th UDMI or the SCAI definition varied from 2.7% to 3.0% and to 5.7% respectively in the PCI arm, and from 2.4% to 2.1% and to 16.5% respectively in the CABG arm [10]. Interestingly enough, the occurrence of periprocedural myocardial injury following CABG was not significantly associated with 10-year mortality after adjustment on confounders, as opposed to PCI. Post-CABG periprocedural MI defined by combining biomarkers elevations and ECG or imaging abnormalities significantly predicted all-cause mortality at 1 and 10 years, while definitions based on isolated enzyme elevation did not have significant correlation with survival. Of note, the wide discrepancies in the incidence rates of periprocedural myocardial infarction following CABG have not been limited to randomized controlled trials and have also been reported based on large real-world registries [49]. Significant and isolated cardiac enzyme elevation may be observed following CABG corresponding to global cardiac injury possibly subsequent to cardioplegia and without a specific epicardial coronary artery de novo lesion or loss of graft patency [47]. The clinical implications of such events seems less certain following CABG than after PCI [10, 49].

Interestingly in the International Study of Comparative Health Effectiveness With Medical and Invasive Approaches (ISCHEMIA) trial, two definitions of periprocedural MI were analyzed, the first one based on CK-MB in association with ECG and angiographic findings, the second one using cTn in association with symptoms, ECG, imaging or angiographic evidence of ischemia [42]. In this study, the rate of periprocedural MI at 6 months in the invasive group was 2.6% using the former definition and increased to 7.7% using the latter, emphasizing the impact of the definition used on the outcomes and the interpretation of the clinical trials. It should be noted that patients undergoing coronary revascularization presented with higher rates of periprocedural MI at 6 months but with lower risk of spontaneous MI at 4 years compared to patient treated conservatively. Importantly, the occurrence of periprocedural MI, using both definitions was not associated with all-cause or cardiovascular mortality, compared to spontaneous MI.

A number of patients features, lesions characteristics and periprocedural factors have been reported to be independently associated with the onset of periprocedural myocardial injury, and type 4a MI following PCI [1, 14, 16, 37] (Table 4, Ref. [14]).

Patient-related factors independently associated with myocardial injury and type 4a MI after PCI are age [37], renal failure [50], preprocedural cardiac biomarker elevation [50], and congestive heart failure [51]. These comorbidities are major signs of overall frailty which could explain the higher risk of myocardial injury.

Lesion characteristics that have been demonstrated to be independently associated with myocardial injury and type 4a MI after PCI are multivessel lesions [52], left main disease [53] and bifurcation lesions [50]. Indeed, as the corresponding coronary arteries frequently cover large portion of myocardium, the risk of myocardial injury increases accordingly, while intervention on bifurcation lesions may result in alteration of the flow in the side branch.

Finally, procedure-related factors independently associated with the occurrence of myocardial injury and type 4a MI after PCI are the use of longer and/or larger stents [37], a greater number of implanted stents (aOR 1.5, 95% CI 1.1–2.3 for a number of implanted stents $\ge$3) [1], multivessel PCI [51], complex procedures such as retrograde approach for chronic total occlusion [54] or use of rotational atherectomy [51]. Such procedures usually require multiple balloon inflations across pre- and post-dilatation of the lesion and stent implantation. Each inflation carries the risk of distal embolus, coronary artery dissection or side branch occlusion.

The identification of these factors before the procedure could help individualizing high risk patients, thus allowing early implementation of preventive strategies and/or close monitoring following the PCI.

As previously mentioned, the specific association between periprocedural myocardial infarction and adverse outcomes may significantly vary according to the considered definitions. However, periprocedural myocardial injury (as opposed to MI) has also been associated with various adverse outcomes such as readmission for acute coronary syndrome or heart failure (OR 3.3, 95% CI 1.1–8.8) [55], unplanned revascularization (aHR 1.40, 95% CI 1.04–2.06) or target vessel revascularization (aHR 1.90, 95% CI 1.06–3.38) [56], MACE at 30 days (aHR 3.8, 95% CI 1.9–6.9) and one year (aHR 1.7, 95% CI 1.1–2.6), as well as cardiac death at one year (aHR 7.66, 95% CI 3.64–16.11) [40] or at 3 years (aHR 4.93, 95% CI 1.92–12.69) [57]. This morbid association may be explained by the direct consequence of loss of myocardium but also by the fact that periprocedural myocardial injury and infarction more frequently occur in frail patients with more extensive atherosclerotic disease.

Elevation of cardiac biomarkers following PCI is quite common in daily clinical practice particularly with the implementation of hs-cTn. Various definitions have been proposed in the last decades using different cardiac biomarkers and thresholds and resulting in a wide range of prevalence and associations with adverse outcomes. As dedicated randomized controlled trials have failed to demonstrate the benefit of the use of potent P2Y12 inhibitors to prevent the risk of periprocedural MI, the focus should be made on identifying patient-level and lesion-level risk factors beforehand to better anticipate and rebalance the indication, implement preventive strategies, and ensure adequate post-procedure monitoring.

ARC-2, Academic Research Consortium-2; CI, confidence interval; CK-MB, creatine kinase myocardial band; CMR, cardiovascular magnetic resonance; cTn, cardiac troponin; EAPCI, European Association of Percutaneous Cardiovascular Interventions; ECG, electrocardiogram; hs-cTn, high-sensitivity cardiac troponin; LGE, late gadolinium enhancement; MACE, major cardiovascular events; MI, myocardial infarction; PCI, percutaneous coronary intervention; SCAI, Society of Cardiovascular Angiography and Interventions; UDMI, Universal Definition of Myocardial Infarction; ULN, upper limit of normal; URL, upper reference limit.

Writing - original draft preparation—AF; writing - review and editing—PG, MZ, JS, J-PC. All authors have read and agreed to the published version of the manuscript.

Not applicable.

Not applicable.

This research received no external funding.

The authors declare no conflict of interest.