Diagnosis of myocardial infarction at autopsy: AECVP reappraisal in the light of the current clinical classification

In clinical guidelines, a distinction is made between myocardial injury, which encompasses any form of acute myocardial damage or destruction, and MI, resulting from myocardial ischemia only [9]. Evidently, MI is a form of myocardial injury, and both entities share the presence of raised serum levels of cardiac troponin (cTn) in a patient. In order to discriminate clinically a MI from other types of myocardial injury, additional criteria such as angina symptoms and characteristic ECG changes are needed. Most important other forms of myocardial injury are listed in Table 3 and should be considered by pathologists at autopsy in the differential diagnosis of MI. For example, differential diagnosis of myocarditis is not always presented by clinical settings and this differentiation may arise from pathological standpoint [11].

The most frequent cause of acute myocardial ischemia is atherothrombotic occlusion of a coronary artery [3, 12]. The presence of a mural or totally occlusive thrombotic mass can be observed at autopsy in approximately 50–70% of sudden coronary deaths and is a reliable marker of myocardial ischemia, even in absence of microscopically visible necrosis [12, 13]. This implies that in cases of sudden death, acute coronary occlusion can explain arrhythmic death [3].

The underlying pathology of mural or occlusive coronary thrombosis is variable and can be due to plaque ruptures, erosions or, less frequently, protruding calcified nodules. Also, intraplaque haemorrhages contribute to the acute flow reduction in these instances [14‐16]. It is important to note that there can be a considerable time interval between the onset of plaque disruption and the evolving critical stenosis or occlusion by thrombus. And also the onset of necrosis in the heart and the clinical manifestation of symptoms of MI are not always around the same time [17, 18]. It is not rare to find even organized thrombi or significant myocardial necrosis at autopsy in a patient with acute onset of ischemic symptoms [18, 19].

At autopsy, the main coronary arteries and large branches such as diagonal and obtuse marginal are examined by transversely sectioning at 3-mm intervals to identify thrombus and critical stenosis. For reliable interpretation, heavily calcified arteries are decalcified prior to cross sectioning. The most severely affected sections can be sampled for histology [3].

The first assessment of luminal stenosis made by the eye is moderately consistent and accurate measure at low (< 30–50%) and high (> 70–75%) grade stenosis. Accuracy can be improved by a visual aid, similar to that in Fig. 1 [20]. Ideally, gross assessments should be confirmed by histology, taking the internal elastic lamina as the original lumen size. This approach has a good inter-observer reproducibility in high-degree lesions [21]. However, lumen shape may affect the interpretation of stenosis, with pathologists overestimating the stenosis caused by slit-like lumens, and underestimating concentric and eccentric stenosis [22]. Geometric remodelling of the artery contributes importantly to the rate of lumen stenosis. Arteries are dynamic organs, in which compensatory expansive enlargement in association with unstable plaques and constrictive shrinkage in stable collagen-rich plaques is known as positive and negative remodelling respectively [23]. In case of a fixed stenotic stable plaque of > 75% stenosis, concomitant functional alterations such as exercise or spasm can evoke irreversible myocardial injury. Therefore, a stenosis of 75% or more is considered as critical [24, 25]. Stenosis of 90% or more leaves myocardium ischemic even at rest and can be seen as a ‘pinpoint’ lumen less than 1 mm in diameter. Many of these stenoses have stable tissue composition of fibrous tissue and calcifications but without thrombus [25]. In these instances, microscopy of malperfused territories of the myocardium can be a helpful adjunct in diagnosis. Clinically, these infarctions are diagnosed as type 2 MI. However, in some cases, mural thrombus can be identified histologically, which implies a reclassification to type 1 MI. The evidence described so far relates to atherosclerotic CAD. Other conditions that may cause coronary stenosis are listed in the differential diagnosis of type 2 MI (Table 2).

Coronary artery spasm (CAS) is defined as an intense constriction of the vascular wall, which causes total or subtotal occlusion of the coronary arteries. Clinically, spasm is one of the MINOCAs, grouped under type 2 MI. The presence of an area of regional infarction in the myocardium could point towards spasm in the artery when no other explanation for the infarction is provided. It can affect either the epicardial arteries, as initially proposed by Prinzmetal et al. ([26] or the microcirculation, or both. The most frequent pathologic substrate of CAS is atherosclerotic CAD, but it has also rarely been reported in normal vessels [27]. Imaging tools such as computerized tomographic angiography (CTA), intravascular ultrasound (IVUS) and optical coherence tomography (OCT) show that the coronary artery segments where spasm is inducible are typically characterized by diffuse intimal and medial thickening with low lipid or calcium content, negative remodelling and small luminal area [28, 29]. These features are in line with the hypothesis that vascular smooth muscle cell hyper-reactivity is a pathophysiological substrate for spasm [30]. Spasm and plaques are certainly not mutually exclusive features; spasms have been documented distal to stable plaques, in plaque-free segments bordering large eccentric plaques, and may serve as a ‘rupture trigger’ by inducing a plaque rupture followed by thrombosis in ‘high-risk’ plaques. Clinically, the combination of severe CAD and spasm is associated with adverse prognosis [31]. At autopsy, there are no distinctive gross or histologic hallmarks of CAS; thus, the possible morphologic substrates associated with an increased risk of spasm must be searched for, as well as microscopic evidence of ischemia. Drugs (cocaine, amphetamine and derivatives, androgenic anabolic steroids, chemotherapy), physical and mental stress and release of vasoconstrictor agents by activated platelets (mural thrombus) are considered precipitating factors of spasm [32‐35]. Very rarely, CAS can occur in the setting of allergic/hypersensitivity reactions, which is known as Kounis Syndrome and histologically characterized by presence of eosinophils [36].

Small vessel diseases (SVD) can occur without obvious structural changes of the heart. Microvascular dysfunction (MVD) is a functional impairment of pressure and flow in intramyocardial vessels < 500 μm in diameter, and may cause MI. It is particularly seen in hypertrophic and dilated hearts, preferentially in the subendocardial areas of the myocardium. SVD is also a common feature in diabetic hearts and hearts of patients with longstanding hypertension, but structural changes such as some vascular wall thickening, if present, are difficult to interpret in practice and subject to interpretation bias. More rarely, structural diseases of intramyocardial vessels may be identifiable which include vasculitis, amyloidosis, small vessel type fibromuscular dysplasia, of which the latter can be isolated or in association with hypertrophic cardiomyopathy (HCM) or Fabry’s disease [37‐39]. It is important to note that in all these instances, myocardial areas with reduced flow due to concomitant epicardial (large vessel) coronary stenosis are most vulnerable. In transplanted hearts, SVD can be caused by widespread small vessel stenosis due to cardiac allograft vasculopathy [40]. Microvascular coronary embolization with thromboembolic materials distal to a thrombosed epicardial plaque can be found at autopsy in patients who died of MI [16, 39, 41] and occur nowadays even more frequently in acute MI patients who are treated with primary percutaneous coronary intervention (PCI) for acutely thrombosed plaques. In clinical studies, such emboli are an uncommon cause of acute MI, although they may result in microinfarctions occurring scattered through the myocardium. In later stages, they may leave small areas of replacement fibrosis. Moreover, microvascular embolization can occur in patients with atrial fibrillation, prosthetic heart valves, infective endocarditis or cardiac myxoma.

Progressive stenosis or acute thrombotic occlusion of a coronary artery can be treated by means of PCI in order to prolong life, to relieve symptoms, or minimize myocardial necrosis. At present, nearly all PCI procedures involve implantation of a metal stent or more recently, bioresorbable scaffolds (BRS). Stent-related complications, which are thrombosis and fibrocellular restenosis, have reduced significantly over the past years, mainly due to widespread application of drug-eluting stents (DES) [42]. Acute stent thrombosis is very rare and may occur due to stent malposition, dissection, long or angulated stented segments, or sometimes due to hypersensitivity reactions. Late stent thrombosis, occurring even more than a year after placement, also remains a rare but life-threatening complication in approximately 2% of patients and is mostly due to impaired neointimal covering of the coated stent and withdrawal of anti-platelet therapy [42]. Moreover, fibrocellular restenosis remains a problem in the still large group of patients, worldwide treated with non-coated bare metal stents (BMS). A more recently described long-term complication is the occurrence of in-stent neoatherosclerosis. Stent-related pathologies are coronary causes of the clinical type 4 MI. Autologous vein grafts or mammary arteries, which are used for CABG procedures, should also be examined carefully. A patent graft at autopsy is normally an empty collapsed vessel. Acute graft thrombosis can be due to technical failure at anastomosis sites, low left ventricular output failure or poor run-off in the distal vascular bed. Vein grafts may develop diffuse concentric intimal lesions within a few years, eventually complicated by thrombotic occlusion. Mammary arteries are particularly resistant to the development of intimal hyperplasia or neoatherosclerosis. Both in cases of PCI or CABG-related MI (clinical types MI 4 and 5), application of post-mortem coronary angiography should be considered. This method enables to localize stents, to visualize the patency of stents and grafts, and to evaluate the run off into the distal arterial bed and presence of collateral vascularization. Histological sampling of the jeopardized myocardium is crucial to investigate the presence of any form of myocardial injury as will be discussed later.

The onset of ischemic necrosis is not immediate. In animal studies, using controlled occlusion of coronary arteries in healthy hearts, irreversible myocardial ischemia can be detected after 20 min [43]. In humans at autopsy, in which pre-existent (ischemic) diseased hearts are not uncommon, timing is less certain. Ischemic necrosis of myocardial cells can be detected reliably with the current diagnostic methods after circa 2–4 h, but the onset depends on many variables such as collateral circulation, ischemic preconditioning and microvascular pathology. After the onset of occlusion, myocardial necrosis evolves in a wave front-like pattern over several hours from the endocardium to the epicardium [44, 45]. This implies that the loss of viable myocardium in the ‘area at risk’ (perfusion area of the occluded artery) gradually increases, finally ending up in a transmural MI. Therefore, in clinical cardiology, it is of pivotal importance to open the occluded artery by means of PCI or thrombolysis as early as possible, and reduce the size of the MI or even prevent it (‘time is muscle’). During the evolving wave of myocardial necrosis, ventricular arrhythmias may occur in the subendocardium, even at the earliest stage of still reversibly damaged myocardium, but also several hours after occlusion when the wave of necrosis has evolved substantially. This is also important for pathologists to realize: early onset arrhythmia and SCD will reveal only early signs of myocardial injury or no changes at all, whereas later on in the wave of developing myocardial necrosis, a full-blown infarction can be visualized at histology. It should also be noted that later stages of myocardial necrosis and even old fibrotic scars can serve as a substrate for life-threatening arrhythmias [46].

According to the myocardial region involved, MI is classified as either regional when it involves the perfusion area of one epicardial artery, or circumferential when it encompasses the largest part of the circumference of the ventricular wall. Regional MI can be either transmural, usually associated with ST segment elevations on ecg (STEMI), or only subendocardial (non-STEMI). Early reperfusion interrupts the wave front of necrosis, which limits irreversible damage to the subendocardial region only (Fig. 2). The topography of segmental MI corresponds grossly to the perfusion territory of the three large epicardial arteries or, more rarely, one of their branches such as the first diagonal branch or the obtuse marginal artery. Occlusion of the left main stem usually results in immediate death. However, in the presence of extensive collateral circulation, which occurs frequently in chronically ischemic hearts, the association between site/extent of necrosis and the occluded branch is often less obvious (‘paradoxical infarction’). Application of post-mortem angiography to the excised heart or during whole body post-mortem CT-angiography (PMCTA) can provide important information on the presence and extent of collaterals between the vascular beds of major epicardial arteries (shown by a retrograde filling pattern), or neovascularization around chronic total occlusions (‘bridging collaterals’) (Fig. 3). Circumferential MI is mostly due to an overall fall in coronary perfusion pressure, often in the presence of severe multivessel CAD and involves in many cases only the subendocardial region.

Aberrant patterns of ischemic damage such as disseminated or predominantly epicardial locations have been reported in patients who died after resuscitation (see later) or in a setting of septic shock. In the latter, ischemia likely results from inflammation-related microvascular spasm, damage or thrombotic occlusions [47].

Atrial infarctions occur in combination with ventricular infarctions and have variable reported incidence among MI patients ranging from 0.7 to 42%. Isolated atrial infarctions are scarce. The leading cause of atrial MI is coronary atherosclerosis. Pathologic significance is obviously lower than in ventricular infarctions, but for a pathologist, two potential complications of atrial MI should be underlined. The first is mural thrombus formation followed by thromboembolization, mostly pulmonary emboli (> 80% of atrial infarctions are located in right atrium), and the second, the rare cases of atrial rupture that can result in cardiac tamponade [48].

Nitro blue tetrazolium (NBT) staining of a fresh myocardial slice demonstrates early ischemic necrosis and is reported to be positive from 3 h after onset of ischemia. NBT stains only in the presence of intracellular lactate dehydrogenase (LDH). Leakage of enzymes from irreversibly injured myocardium appears as an unstained area in the otherwise deep purple stained vital myocardium [50, 51] [5] (Fig. 5). Unstained areas are not fully specific for MI, but can be caused also by other forms of injury although usually not in a distinct regional pattern. In addition, NBT staining is reported to be somewhat vulnerable to artefacts with false positivity in case of technical failure (inappropriate temperature or incubation time, formalin contamination), or situations like sepsis, CPR or long post-mortem intervals.

Several immunohistochemical markers have been investigated for the diagnosis of early MI, mostly markers for proteins that accumulate (such as fibronectin and C5b-9) or leak out of (such as troponins, myoglobin, S100A1) cardiomyocytes following ischemia. Other protein markers have been shown to undergo early changes in their phosphorylation state, as is demonstrated for connexin 43 [54, 55]. Markers/mediators of early inflammation (CD15, IL-6, TNF-α, IL-15, IL-8, CD18 and tryptase) have also been proposed [56]. Some of them are promising in terms of early expression and specificity (dephosphorylated connexin 43), but they have been mainly investigated in experimental models [55]. When tested in human post-mortem samples, these markers keep their early expression profile, but lose their specificity [4]. Therefore, further investigations are needed before their eventual introduction in routine. Moreover, several pathologic or iatrogenic states such as CPR (including injection of catecholamines), autolysis, pre-existing ischemic events and medical treatments [4, 55] may influence the staining pattern and can interfere with a diagnosis of ischemic injury. Currently, a diagnosis of early myocardial injury in absence of changes in H&E stains could be based on a combination of fibronectin and C5b-9 immune stains (Fig. 6). C5b-9 staining is more sensitive and specific than fibronectin staining, but fibronectin positivity starts earlier than C5b-9 [55]. In most cardiac transplant centres, C4d immunostaining is used routinely to detect myocardial injury on paraffin sections but the exact timing is unknown (Fig. 6). A larger panel of markers, simultaneously detected in the same tissue section (multiplexing), and quantification of their expression could improve the diagnostics of early myocardial ischemia, as has recently been shown by employing mass spectrometry immunohistochemistry [57].