key: cord-0904295-ytm32tr2
authors: Haslbauer, Jasmin D; Tzankov, Alexandar; Mertz, Kirsten D; Schwab, Nathalie; Nienhold, Ronny; Twerenbold, Raphael; Leibundgut, Gregor; Stalder, Anna K; Matter, Matthias; Glatz, Katharina
title: Characterisation of cardiac pathology in 23 autopsies of lethal COVID‐19
date: 2021-04-09
journal: J Pathol Clin Res
DOI: 10.1002/cjp2.212
sha: b3059c8566b1fe34f0e69630f3faf86d9cb7eb4d
doc_id: 904295
cord_uid: ytm32tr2

While coronavirus disease 2019 (COVID‐19) primarily affects the respiratory tract, pathophysiological changes of the cardiovascular system remain to be elucidated. We performed a retrospective cardiopathological analysis of the heart and vasculature from 23 autopsies of COVID‐19 patients, comparing the findings with control tissue. Myocardium from autopsies of COVID‐19 patients was categorised into severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) positive (n = 14) or negative (n = 9) based on the presence of viral RNA as determined by reverse transcriptase polymerase chain reaction (RT‐PCR). Control tissue was selected from autopsies without COVID‐19 (n = 10) with similar clinical sequelae. Histological characteristics were scored by ordinal and/or categorical grading. Five RT‐PCR‐positive cases underwent in situ hybridisation (ISH) for SARS‐CoV‐2. Patients with lethal COVID‐19 infection were mostly male (78%) and had a high incidence of hypertension (91%), coronary artery disease (61%), and diabetes mellitus (48%). Patients with positive myocardial RT‐PCR died earlier after hospital admission (5 versus 12 days, p < 0.001) than patients with negative RT‐PCR. An increased severity of fibrin deposition, capillary dilatation, and microhaemorrhage was observed in RT‐PCR‐positive myocardium than in negatives and controls, with a positive correlation amongst these factors All cases with increased cardioinflammatory infiltrate, without myocyte necrosis (n = 4) or with myocarditis (n = 1), were RT‐PCR negative. ISH revealed positivity of viral RNA in interstitial cells. Myocardial capillary dilatation, fibrin deposition, and microhaemorrhage may be the histomorphological correlate of COVID‐19‐associated coagulopathy. Increased cardioinflammation including one case of myocarditis was only detected in RT‐PCR‐negative hearts with significantly longer hospitalisation time. This may imply a secondary immunological response warranting further characterisation.

Coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has rapidly evolved into one of the most significant public health crises of the 21st century. While chiefly a respiratory disease, a wide spectrum of SARS-CoV-2-associated cardiovascular sequelae have been described in the literature. These include elevated cardiac troponins [1] , associated with worse outcome and an increased mortality rate [2, 3] , heart failure [4] , arrhythmias, bundle branch block, tachycardiabradycardia syndrome, and/or non-specific repolarisation changes in electrocardiograms (ECG) [5] and occasional ST-segment elevation myocardial infarction [6, 7] . Most notably, a higher mean age and incidence of cardiovascular risk factors and comorbidities such as hypertension, coronary heart disease, heart failure, obesity, diabetes mellitus, and cardiac amyloidosis have been observed in patients with severe disease [8, 9] . SARS-CoV-2 utilises its spike (S) protein to bind to the membrane-bound form of angiotensin-converting enzyme 2 (ACE2) receptor [10] . The heart has been proposed as a potential direct target of viral entry, as ACE2 has been found to be expressed in cardiomyocytes, fibroblasts, endothelium, and pericytes [11, 12] . Electron microscopy (EM) has detected SARS-CoV-2 both in the interstitial compartment of the myocardium [13, 14] and occasionally in the cardiomyocytes themselves [15] ; these findings were mirrored by in situ hybridisation (ISH) [12, 16] .

Recently performed in vitro studies demonstrated that SARS-CoV-2 was able to directly infect pluripotent stem cell-derived cardiomyocytes [17] . These data are strongly suggestive of a potential cardiotropic effect of SARS-CoV-2 which may contribute to the cardiovascular manifestations discussed above, although its mechanisms may be more complex than initially imagined and warrant further study.

Secondary dysregulation of the immune response in the form of a cytokine storm has been observed in COVID-19 [18] . Rapid viral replication leads to a disproportionate mobilisation of neutrophils, monocytes, and endothelial dysfunction leading to acute lung injury and multi-organ failure, including cardiac dysfunction [19] . An increased risk for thromboembolic events, likely due to microvascular dysfunction [20, 21] , has been observed in both clinical and autopsy series with severe or lethal outcome despite receipt of anticoagulation [9, 22, 23] . In addition, postmortem investigations have proposed the development of global endothelialitis in lethal disease [24] .

COVID-19-associated myocarditis has been delineated in several case studies. Clinical diagnosis was based on serology, ECG, echocardiography, and cardiac magnetic resonance [13, [25] [26] [27] . However, histological findings varied, ranging from lymphocytic myocarditis without the presence of SARS-CoV-2 [28] to mild interstitial and endocardial inflammation without fulfilling diagnostic criteria [13] . A comparison between patients with increased versus insignificant myocardial viral load revealed no significant difference in inflammatory infiltrate [16] . Several potential pathophysiological mechanisms behind these findings have since been suggested, including direct viral injury, endothelialitis, interleukin (IL)-6-induced cytokine storm, and auto-antibodies [27] , which calls for further studies characterising the cardiac inflammatory effects in COVID-19.

To further analyse the full spectrum of SARS-CoV-2-associated cardiopathy, we performed a systematic cardiopathological characterisation of lethal COVID-19 in 23 autopsies.

Twenty-three autopsies of patients with COVID-19 were performed at the pathology departments of the University Hospital of Basel (n = 12) and the Cantonal Hospital Baselland, Liestal (n = 11), from March to June 2020. All decedents were diagnosed with COVID-19 as per antemortem nasopharyngeal swab. Full body autopsy was performed in 21 cases (91%); a partial autopsy of the upper respiratory tract, lungs, and heart (n = 2, 9%) was conducted in cases with excessive overweight or according to patient or relatives' wishes. In all 23 cases, the cause of death at autopsy was COVID-19-associated respiratory failure (by either acute respiratory distress syndrome/diffuse alveolar damage [ARDS/DAD], pulmonary embolism [PE] , and/or bronchopneumonia). Myocardial reverse transcriptase polymerase chain reaction (RT-PCR) C T values (cycle threshold) were utilised to classify cases into RT-PCR-positive (n = 15) and -negative (n = 8) subgroups. Autopsies without COVID-19 (n = 10) but with similar sequelae (ARDS/DAD n = 7; PE n = 2; bronchopneumonia n = 1) were selected as controls. The overall mean time to fixation was 29.14 h (5.3-84.5). Table 1 lists clinical features, comorbidities, pharmacology, laboratory results, and ECG findings of COVID-19 patients and controls. Angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, renin inhibitors, and aldosterone inhibitors were defined as renin angiotensin aldosterone system (RAAS) inhibitors. Anticoagulants included heparin and its derivatives, new oral anticoagulants, coumarin derivatives, and/or platelet aggregation inhibitors. This study was approved by the Ethics Committee of Northwestern and Central Switzerland (ID 2020-00629).

All hearts and greater vessels were dissected in the direction of blood flow. The extent of coronary atherosclerosis, stenosis, aortic atherosclerosis, and atrial dilatation was recorded ordinally and/or nominally. Cardiac hypertrophy was determined by gender-dependent percentile curves according to body mass index (BMI) and heart weight (see Supplementary material for gross findings and scoring methodology).

Representative tissue sections (0.4 × 1 × 2 cm) of the left anterior and posterior wall, the interventricular septum, and papillary muscles of the left ventricle and interventricular septum were extracted at autopsy and preserved in 4% phosphate-buffered formalin before being processed. In seven cases, a sample of the right 327 Cardiac pathology of lethal COVID-19 ventricle was extracted. Standard histological staining protocols (haematoxylin and eosin [H&E] and chromotrope aniline blue [CAB]) were applied. In cases with suspected cardiac amyloidosis, Congo red staining was performed. In addition, immunohistochemical staining of fibrin, cluster of differentiation (CD) 3, CD68, ACE2, κ and λ immunoglobulin light chains, antisera to amyloid A, and transthyretin (ATTR) for amyloid subtyping was performed in line with previously described protocols [29, 30] . Immunohistochemical detection of ACE2 (Abcam ab15348, polyclonal, rabbit, 1:10,000) (ABCAM, Cambridge, UK) and staining for viral nucleocapsid (N)-protein (Rockland #200-401-A50, polyclonal, rabbit, 1:2,000) (Rockland Immunochemicals Inc, Limerick, PA, USA) was performed in all COVID-19 cases. Histomorphological characteristics, such as hypertrophy, ischaemic changes (contraction band necroses/ Increased numbers of intravascular granulocytes and mononuclear cells and intramyocardial lymphohistiocytic inflammatory infiltrates outside of areas of fibrosis and unassociated with tissue repair were assessed on H&E. Intramyocardial inflammatory cells were additionally quantified by immunohistochemistry (IHC; CD3-, CD20-, and CD68-positive cells/mm 2 ). Seven or fewer (≤7) single CD3-positive T-lymphocytes/mm 2 and 3 or fewer (≤3) single CD68-positive histiocytes outside areas of fibrosis were considered physiological. Increased cardioinflammatory infiltrate without necrosis was defined by the presence of clusters (≥3 CD3-positive T-lymphocytes in non-fibrotic myocardium) and foci (≥14 leucocytic cells/mm 2 including ≤4 monocytes/mm 2 and ≥7 CD3-positive T-lymphocytes/mm 2 ). Myocarditis was defined as increased cardioinflammatory infiltrate with associated cardiomyocyte necrosis.

The histomorphological distribution of ACE2 expression in stromal and endothelial cells was assessed qualitatively. ACE2-positive epicardial fat cells served as a positive internal control.

For a detailed description of RT-PCR, ISH, and EM protocols, see Supplementary material.

Statistical analyses were performed using IBM® SPSS® Statistics, Version 25.0 (2017) (Armonk, NY, USA). Significance testing was performed between three groups (RT-PCR positive, RT-PCR negative, and controls) and two groups (COVID-19 patients and controls). In parametric data sets, a one-way analysis of variance (ANOVA) and an independent sample t-test were performed. For non-parametric or ordinal variables, Kruskal-Wallis one-way ANOVA and Mann Whitney U-test were used. Appropriate post hoc testing was subsequently performed on all statistically significant variables with more than two groups (Tukey's honestly significant difference for parametric and Dunn's test for non-parametric variables; see Supplementary material for all values). A twotailed Fisher's exact test was performed for categorical variables. Correlation analyses of myocardial characteristics was performed using Spearman's rank-order correlation. Pearson's correlation was utilised to correlate viral copy numbers between the lung and myocardium (see Supplementary material). P values were considered significant at the 0.05 level.

Clinical characteristics are displayed in Table 1 . The median BMI was 27 and 78% were male. Amongst COVID-19 patients, myocardial RT-PCR-negative cases experienced longer hospitalisation times than positive cases (12 versus 5 days; p < 0.005). COVID-19 patients had a significantly higher mean age (76 versus 64 years; p = 0.01) than controls. Hypertension (91 versus 70%; p < 0.001) and a history or current diagnosis of coronary artery disease (61 versus 20%; p = 0.04) and diabetes mellitus (48 versus 10%; p = 0.04) were more common amongst COVID-19 patients compared to controls. An intake of RAAS inhibitors was more likely amongst myocardial RT-PCR-positive patients than negatives or controls (71 versus 22 versus 30%; p = 0.04). Intake of anticoagulants and/or platelet aggregation inhibitors before and during hospital admission was more common amongst COVID-19 patients than controls (before: 65 versus 20%, p = 0.04; during: 96 versus 50%, Figure 1 . Temporal evolution of COVID-19-associated cardiopathy: early microvasculopathy followed by secondary cardioinflammatory response. Myocardium from autopsies of COVID-19 patients (n = 23) was classified based on SARS-CoV-2 presence determined by RT-PCR. Cases with detectable viral loads (RT-PCR positive) displayed higher severity of capillary dilatation, fibrin deposition, and microhaemorrhage, a histomorphological correlate of microvascular dysfunction. In RT-PCR-negative cases (n = 9), a high incidence of predominantly subtle cardioinflammatory pathologies was observed (including one case with lymphohistiocytic myocarditis), along with a longer hospitalisation time.

Cardiac pathology of lethal COVID-19 p = 0.004). Most recently available laboratory findings before death revealed significantly higher C-reactive protein (176.4 versus 56 mg/dl; p = 0.03) in COVID-19 patients versus controls. Leucocyte differentiation showed severe lymphopenia (median: 0.61 × 10 −9 /l) and mild neutrophilia (median: 6.8 × 10 −9 /l) in COVID-19 patients. High-sensitivity troponin T (hs-troponin T) was measured in eight (34%) COVID-19 patients; its median value in RT-PCR-positive and -negative cases was 77 and 98 ng/l, respectively. 

A graphical overview of histological findings can be found in Figure 1 . Gross findings are summarised in Table 2 . Heart weight ranged from 430 to 550 g amongst COVID-19 patients (median: 468 g). There was a significantly higher incidence of predominantly eccentric hypertrophy amongst RT-PCR-positive hearts (93%) compared to RT-PCR negatives (67%) or controls (60%; p = 0.04). Thirteen percent of COVID-19 cases presented with coronary stenosis. Histological analysis of fibrotic changes revealed no difference between categories (interstitial, replacement, and infarct scar) when comparing COVID-19 cases to controls. Small foci of non-territorial myocyte necroses were observed in 26% of COVID-19 patients, and one case (4%) presented with acute transmural type 1 myocardial infarction of the anterior and lateral wall of the left ventricle (diameter: 7 cm) with perforation and fatal pericardial tamponade ( Figure 2 ). Analysis of coronary vessels revealed an acute thrombotic occlusion of the left anterior descending artery as the cause of infarction. Histological analysis of the right ventricle in 7 of 23 cases revealed unremarkable findings devoid of right heart dilatation and/or stress; none of the samples showed contraction band necroses.

RT-PCR of COVID-19 patients revealed positive C T values in 14 (61%) out of 23 cases. Myocardial SARS-CoV-2 genomes/10 6 RNase-P copies correlated significantly with values measured in lung tissue of the same cohort (Pearson correlation between myocardial and pulmonary SARS-CoV-2 genomes/10 6 RNase-P copies: r = 0.622; p = 0.001). Pulmonary viral load was significantly higher in the lungs than in the myocardium (p = 0.0057, Student's t-test) and there were no cases of isolated myocardial positivity (see Supplementary material). IHC for N-protein and ISH revealed scarce positivity in interstitial cells without detection in cardiomyocytes (Figure 3 ). EM performed in two RT-PCR-positive cases did not detect viral particles in cardiomyocytes or interstitial cells.

An overview of microvascular histology findings is displayed in Figure 4 . While its incidence did not differ between groups, the distribution of capillary dilatation varied significantly between RT-PCR positives, negatives, and controls (p = 0.04)it was particularly severe in RT-PCR-positive myocardium (64% marked) compared to negatives (22% marked). The overall incidence of marked capillary dilatation was 48% in COVID-19 patients versus 20% in controls. An overall higher incidence of both focal (35%) and extensive (22%) microhaemorrhages was observed in COVID-19 patients versus controls, which also yielded a statistically significant difference in distribution (RT-PCR positive versus negative versus controls: p = 0.03; COVID-19 versus controls: p = 0.04). All COVID-19 cases manifested capillary fibrin deposition; a significant difference in its distribution was observed between all groups (p = 0.04) and between COVID-19 patients and controls (p = 0.04); 64% of RT-PCR-positive myocardium presented with extensive deposition versus 22% of RT-PCR negatives. There were significant positive correlations between capillary dilatation, microhaemorrhage, and capillary fibrin deposition (capillary dilatation-microhaemorrhage: ρ = 0.61, p < 0.01; capillary dilatation-capillary fibrin: ρ = 0.36, p = 0.04; capillary fibrin-microhaemorrhage: ρ = 0.35, p = 0.04).

In lieu with fibrin deposition, erythrocyte deformation and anisocytosis in capillaries were observed in 61% of COVID-19 patients. The myocardial capillaries of two COVID-19 patients examined by EM contained densely packed, fragmented, and deformed erythrocytes coated with a thin layer of fibrin and extensive filamentous fibrin deposits along capillary walls ( Figure 4C ). In areas of haemorrhage, capillaries were replete with erythrocytes presenting with focal wall ruptures. However, no statistically significant incidence of erythrocyte anisocytosis was observed compared to controls.

Six patients (26%) with COVID-19 presented with cardiac amyloidosis, which in all cases was immunohistochemically subtyped as senile ATTR amyloidosis.

Increased numbers of interstitial CD68-positive macrophages were observed in COVID-19 patients compared to controls (39 versus 20%; 67% of RT-PCR-negative cases). A total of four COVID-19 cases (17%) presented with an increased lymphohistiocytic infiltrate without the presence of myocyte necrosis. In one case (4%), both active lymphohistiocytic myocarditis without fibrosis characterised by several active, necrotising inflammatory foci and an acute transmural infarction resulting in lethal perforation of the myocardial wall with associated ischaemic inflammation in a different locus were observed ( Figure 2 ). All cases with increased cardioinflammatory infiltrate were RT-PCR negative. Only one case in the control group exhibited histological signs of increased cardioinflammation without necrosis (10%).

Myocardial ACE2 was detected in stromal cells with either granular cytoplasmic and/or membranous positivity. No notable differences in staining patterns between COVID-19 patients and controls were observed.

In this study, we assessed cardiopathological characteristics in 23 lethal cases of COVID-19. The 

JD Haslbauer et al incidence of hypertension, coronary artery disease, and diabetes mellitus was high amongst COVID-19 patients, in line with the high incidence of cardiovascular comorbidities in patients with lethal COVID-19 that has repeatedly been shown in the literature [4, 8] . Our findings mirror a meta-analysis of comorbidity prevalence of seven clinical studies, which reported pooled odds ratios of 2.36 for hypertension and 3.42 for cardiovascular disease between mild and severe disease outcome [31] , thus further underlining the critical role of cardiovascular comorbidities in the pathogenesis of severe COVID-19. This is further supported by a high incidence of eccentric hypertrophy amongst COVID-19 patients (78%; Table 2 ), which is considered as long-term adaptation to hypertension and may thus be a predisposing factor for severe disease. A complex relationship between cardiovascular risk factors and immune response [32, 33] , as well as a potential role of immunosenescence [34] are all plausible pathophysiological explanations behind these observations. Identification of these patient subgroups should therefore play an instrumental role in risk stratification for early therapeutic intervention.

Sixty percent of cases in this cohort (14/23) displayed myocardial RT-PCR positivity by SARS-CoV-2 PCR. It however remains to be elucidated whether this finding is a reflection of parenchymal viral load or merely due to packaged virions in contaminated blood. A significant positive correlation with the viral load in lungs (see Supplementary material) seems to support the latter, although ancillary techniques (IHC against the N-protein, ISH, and EM) did not detect viral presence in cardiomyocytes in this study. A broad array of studies have detected viral presence in both interstitial compartments and occasionally in cardiomyocytes [12, 13, 15, 16] , although data are presently sporadic and do not exclude a potential cardiotropic nature of SARS-CoV-2.

Histomorphological analysis of the myocardium in COVID-19 patients revealed significantly higher levels of capillary fibrin deposition, capillary dilatation, and parenchymal microhaemorrhages compared with 10 autopsies without SARS-CoV-2 infection but similar sequelae. These morphological alterations of the microvasculature correlated positively amongst each other, suggesting potential causality and also suggesting a myocardial manifestation of the hypercoagulable state in severe COVID-19 in the form of microangiopathic disease. We postulate that capillary fibrin deposition causes microcirculatory stasis, resulting in focal microhaemorrhages due to increased intravascular pressure and permeability of the endothelial barrier [35] . Erythrocyte anisocytosis in capillary lumina, detected on both light microscopy and EM, further supports these findings as a possible histomorphological correlate of disseminated intravascular coagulation and complement activation. Our initial post-mortem series analysing the same cohort parallels these observations, reporting an increased incidence of pulmonary microthrombi indicative of vascular dysfunction [9] . Another autopsy case series observed a systemic incidence of microthrombi in the microvasculature of several organ systems including the heart, thus confirming the systemic nature of COVID-19-associated microvasculopathy [36] . Our results are similarly in line with clinical observations which report an increased incidence of thromboembolic events in severe COVID-19 [22, 37] . Remarkably, no necroses associated with the microhaemorrhages were histologically detectable. It therefore remains questionable whether these microangiopathic changes have a significant impact on myocardial function. In our cohort, senile cardiac ATTR amyloidosis amongst COVID-19 patients (26%) was more common than its incidence amongst autopsies performed in Basel between 2018 and 2019 [9] . These findings may be relevant, as cardiac amyloidosis has previously been reported to impact microvascular dysfunction [38] , purportedly predisposing to the aforementioned observations in COVID-19 vasculopathy.

As denoted in previous observational studies, acute coronary syndrome can be triggered by COVID-19 and even be its primary symptom [7] . Possible pathophysiological explanations of angina pectoris in COVID-19 include plaque rupture, occlusive thrombosis, or coronary spasm [39] . In our study, ECG of COVID-19 patients did not reveal significant or specific changes related to ischaemic events, although non-specific conduction and/or repolarisation abnormalities were reported (Table 1) . Furthermore, only one case of acute territorial transmural infarction in conjunction with pericardial tamponade was observed amongst COVID-19 patients (Figure 2 ). In this case, we believe that the infarction, caused by thrombotic occlusion of the left anterior descending artery, was solely responsible for the wall rupture despite concurrently diagnosed active myocarditis. Moreover, infrequent small foci of acute and subacute non-territorial ischaemic necroses in 17% of COVID-19 patients did not statistically differ from controls, mirroring previous findings [14] . Small foci of myocardial necrosis can occur due to a broad spectrum of aetiologies and are a frequent incidental post-mortem finding irrespective of the cause of death. Taken together, these findings suggest that ischaemic events are an uncommon observation in this cohort, although more extensive post-mortem series are needed for further characterisation.

Four out of five cases with increased cardioinflammatory infiltrate presented without cardiomyocyte necrosis (n = 4) while one case presented as active, lymphohistiocytic myocarditis. These findings are in line with a literature review assessing the autopsy incidence of SARS-CoV-2-associated myocarditis, confirming that clinically relevant cardioinflammatory changes are generally a rare finding in COVID-19 [40, 41] . Of note, all these cases were deemed RT-PCR negative. This finding is highly relevant when considering increased hospitalisation time amongst RT-PCR-negative patients, as it seems to mirror the behaviour of other types of viral myocarditis such as Coxsackie, citing infection as a precursor to delayed-onset cardiac inflammation in a complex, sequential cytokine response model [42] . Whether the myocardial effects observed in our study occur as a consequence of a systemic cytokine storm, are induced by auto-antibodies, or are the result of direct virally mediated injury remain to be elucidated. Lindner et al performed a cytokine gene expression panel on the myocardium of COVID-19 patients, noting that cases with increased viral load present with higher expression of tumour necrosis factor-α, interferon-γ, CCL5, IL-6, IL-8, and IL-18 [16] ; however, the unique myocardial cytokine signature in response to COVID-19 is still unclear and the subject of ongoing studies. An increased influx of CD68-positive macrophages in 39% of COVID-19 patients, more prominently in RT-PCRnegative cases (75%), observed in our study is consistent with results from previous post-mortem reports [13, 43] . This is supported by findings of ISH and N-protein IHC conducted in this study which revealed positivity in interstitial cells of likely monocytic origin ( Figure 3B ).

The role of ACE2 as a mediator of COVID-19-associated cardiopathy remains unclear. In our study, immunohistochemical analysis of ACE2 predominantly revealed positivity in stromal cells, in line with previous

JD Haslbauer et al gene expression analyses of myocardial ACE2 [10, 12] . However, staining patterns did not markedly differ in control tissue. This contradicts a previously conducted investigation in SARS patients which observed weaker staining in patients versus controls, also denoting a significant decrease in ACE2 expression amongst SARS patients [44] . Notwithstanding these observations, the evaluation of immunohistochemical staining intensity and viability of RNA in post-mortem tissue varies with the extent of autolysis. Future studies are required to systematically study the basic biology of RAAS metabolites to shed light on the potential cardiotropic nature of SARS-CoV-2 and the role pharmacological inhibition of RAAS inhibitors may have on COVID-19 susceptibility.

One of the key observations of this study was significantly longer hospitalisation times (12 versus 5 days; p = 0.005) amongst myocardial RT-PCR-negative cases. This may either be because of a more serious disease progression in cases with higher myocardial viral load, leading to more rapid death, or due to gradual viral elimination as the disease progresses. The latter was suggested in a differential gene expression analysis conducted on the lungs of the same autopsy cohort, which identified two distinct immunopathological profiles based on interferon-stimulated gene (ISG) expression: ISG high cases were identified as having an upregulation of ISGs, high viral load, and minimal pulmonary damage, while ISG low cases featured downregulated ISGs, lower viral loads, and more extensive pulmonary damage with signs of remodelling [45] . Significantly longer hospitalisation times were observed in ISG low cases. As all three genomic targets of myocardial RT-PCR viral load show significant positive correlation to pulmonary values (see Supplementary material), we postulate that there is some degree of concordance between the immunopathological profiles and the subgroups denoted in our study. A biphasic dynamic of COVID-19 progression may similarly be prevalent in the myocardium; microvascular changes, such as capillary dilatation, microhaemorrhage, and capillary fibrin deposition, show decrease in severity when comparing RT-PCR positive with negative cases (marked capillary dilatation: 64 to 22%; incidence of microhaemorrhage: 64 to 45%; extensive capillary fibrin: 64 to 22%), implying a dominance of microvascular pathology in early COVID-19 ( Figure 1) . Conversely, the extent of CD68-positive infiltrate (67 versus 21%) and the overall incidence of myocarditis are increased (22 versus 0%) in RT-PCR-negative cases compared to positives. Interestingly, ISG low cases were also associated with increased presence of pulmonary CD68 + macrophages, insinuating a migration from the lung to the myocardium in later stages of the disease. When comparing these trends with differential blood counts, an overall improvement of leucopenia (0.5 to 1.0 × 10 −9 /l) and decrease of neutrophilia (7.4 to 6.1 × 10 −9 /l) towards normal values further support our hypothesis that RT-PCR-positive and -negative subgroups represent a temporal evolution profile of immune viral response in COVID-19.

This study has several limitations. Its retrospective design resulted in inconsistent clinical data acquisition, such as missing hs-troponin T and N-terminal pro-brain natriuretic peptide values, in more than half of the data set. Moreover, the selection of controls as cases with COVID-19 suggestive sequelae was a challenge and may display several confounding factors; for instance, controls were significantly younger and presented with a high incidence of intensive care unit admission which may impact cardiac histomorphology. Moreover, the utilisation of ordinal scales in most histological analyses may comport limitations of statistical power in this study.

In addition, there are currently no systematic data on assay sensitivity of SARS-CoV-2 RT-PCR in formalinfixed, paraffin-embedded (FFPE) tissue, although consistency between ante-mortem swabs has been established in all our cases [9] and recent studies have been able to detect viral load in FFPE even in subclinical patients [46, 47] ; nonetheless, assay standardisation studies are required for both RT-PCR as well as other ancillary techniques for the detection of SARS-CoV-2 in FFPE tissue.

Lastly, due to their mostly subtle manifestation, it is questionable whether the cardioinflammatory changes identified in this study would have been detected in an endomyocardial biopsy; in fact, small inflammatory foci without necrosis have previously been suggested be a non-specific finding at autopsy [48] .

In conclusion, this cardiopathological study of myocardium in COVID-19 patients confirms a systemic manifestation of the SARS-CoV-2 hypercoagulable state extending to the myocardium, as well as an occurrence of subtle cardioinflammatory conditions in a subset of patients with longer hospitalisation duration and without detectable myocardial viral load. assisting with dissection; PD Dr Thomas Menter, Dr Katharina Marston, Dr Fermin Person, Dr Daniel Turek, and Dr Simon Häfliger for conducting autopsies; and Susi Grieshaber, Petra Hirschmann, Valeria Perrina, Jan Schneeberger, and Martin Herzig for their expertise in IHC, ISH, RT-PCR, and EM. This study was funded by the Botnar Research Centre for Child Health.

Cardiac troponin for assessment of myocardial injury in COVID-19: JACC Review Topic of the Week

Cardiovascular implications of fatal outcomes of patients with coronavirus disease

Association of cardiac injury with mortality in hospitalized patients with COVID-19 in Wuhan, China

Clinical characteristics of 113 deceased patients with coronavirus disease 2019: retrospective study

Electrocardiographic features of patients with COVID-19 pneumonia

ST-segment elevation in patients with Covid-19 -a case series

ST-elevation myocardial infarction in patients with COVID-19: clinical and angiographic outcomes

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Post-mortem examination of COVID19 patients reveals diffuse alveolar damage with severe capillary congestion and variegated findings of lungs and other organs suggesting vascular dysfunction

Cell type-specific expression of the putative SARS-CoV-2 receptor ACE2 in human hearts

The ACE2 expression in human heart indicates new potential mechanism of heart injury among patients infected with SARS-CoV-2

ACE2 (angiotensinconverting enzyme 2) and TMPRSS2 (transmembrane serine protease 2) expression and localization of SARS-CoV-2 infection in the human heart

Myocardial localization of coronavirus in COVID-19 cardiogenic shock

Unexpected features of cardiac pathology in COVID-19 infection

Evidence of SARS-CoV-2 transcriptional activity in cardiomyocytes of COVID-19 patients without clinical signs of cardiac involvement

Association of cardiac infection with SARS-CoV-2 in confirmed COVID-19 autopsy cases

SARS-CoV-2 infects and induces cytotoxic effects in human cardiomyocytes

The COVID-19 cytokine storm; what we know so far

Innate immunity and the failing heart: the cytokine hypothesis revisited

Pulmonary vascular endothelialitis, thrombosis, and angiogenesis in Covid-19

COVID-19 sepsis and microcirculation dysfunction

Pulmonary embolism in patients with COVID-19: awareness of an increased prevalence

Autopsy findings and venous thromboembolism in patients with COVID-19: a prospective cohort study

Endothelial cell infection and endotheliitis in COVID-19

Cardiac involvement in a patient with coronavirus disease 2019 (COVID-19)

Fulminant myocarditis due to COVID-19

SARS-CoV-2-related myocarditis-like syndromes Shakespeare's question: what's in a name?

Acute myocarditis presenting as a reverse Tako-Tsubo syndrome in a patient with SARS-CoV-2 respiratory infection

A more accurate approach to amyloid detection and subtyping: combining in situ Congo red staining and immunohistochemistry

Histologic features of hematopoietic stem cell transplant-associated thrombotic microangiopathy are best percepted in deep skin biopsies and renal biopsies, while showing a significant overlap with changes related to severe acute graft-versus-host disease in gastrointestinal biopsies

Prevalence of comorbidities and its effects in patients infected with SARS-CoV-2: a systematic review and meta-analysis

Role of adaptive and innate immunity in type 2 diabetes mellitus

Immune mechanisms in arterial hypertension

Remodeling of the immune response with aging: immunosenescence and its potential impact on COVID-19 immune response

Pericyte alteration sheds light on micro-vasculopathy in COVID-19 infection

Megakaryocytes and platelet-fibrin thrombi characterize multi-organ thrombosis at autopsy in COVID-19: a case series

Systematic assessment of venous thromboembolism in COVID-19 patients receiving thromboprophylaxis: incidence and role of D-dimer as predictive factors

Coronary microvascular dysfunction is related to abnormalities in myocardial structure and function in cardiac amyloidosis

Acute coronary syndromes and Covid-19: exploring the uncertainties

Myocarditis is rare in COVID-19 autopsies: cardiovascular findings across 277 postmortem examinations

Pathological evidence for SARS-CoV-2 as a cause of myocarditis

Critical cytokine pathways to cardiac inflammation

Pathological features of COVID-19-associated myocardial injury: a multicentre cardiovascular pathology study

SARS-coronavirus modulation of myocardial ACE2 expression and inflammation in patients with SARS

Two distinct immunopathological profiles in autopsy lungs of COVID-19

SARS-CoV-2 detection in formalin-fixed paraffin-embedded tissue specimens from surgical resection of tongue squamous cell carcinoma

Detection of SARS-CoV-2 in formalin-fixed paraffin-embedded tissue sections using commercially available reagents

Guidelines for autopsy investigation of sudden cardiac death: 2017 update from the Association for European Cardiovascular Pathology

Criteria and scoring of pathological parameters C. SARS-CoV-2 RT-PCR methodology and results D

Electron microscopy (EM) methodology

The authors thank all patients included in this study and their relatives; Ralph Schoch and Thomas Rost for