Fatal pulmonary thromboembolism in asymptomatic COVID-19

This study primarily reviewed post-mortem examinations (reports) performed by one pathologist at Dublin City Mortuary of 6 asymptomatic COVID-19 deaths. Autopsies were conducted between 7 April 2020 and 27 May 2020. These included the first 5 COVID-19 autopsies conducted in Ireland [9]. Post-mortem reports from the same facility and pathologist for the years 2019 and 2020 were also reviewed to establish the incidence of ABPTE.

Due to the threat of COVID-19 to mortuary staff, limited autopsy, histology and toxicology were performed for these cases. Normal autopsy components not performed included examination of the head (including the brain) and the removal of abdominal and pelvic contents. Contents instead were examined in situ with histology samples taken of the liver, spleen, kidneys and any abnormality identified. Histology samples were also taken from all lung lobes as well as from the left and right heart ventricles. Toxicology specimens were taken from all cases.

Each post-mortem report was reviewed for gross examination, histology and toxicology findings. A review of the current literature on COVID-19 autopsy findings, COVID-19 pathogenesis, thrombosis in COVID-19 and asymptomatic SARS-CoV-2 infection was also conducted using PubMed.

The first 5 cases tested positive for COVID-19 in the community as they were close contacts of symptomatic patients. The sixth case tested positive in the mortuary (testing of all bodies was introduced to Dublin City Mortuary in mid-April 2020). Immunohistochemistry used to confirm COVID-19 infection was not performed as it was not available at the time.

The cause of death in 4 of the 6 cases (1, 2, 3 and 5) was ABPTE (Fig. 1). Case 4 died of ischaemic and hypertensive heart disease and case 6 died of congestive cardiac failure. Of those who died of ABPTE, 2 cases (1 and 5) had suspect pelvic deep venous thrombosis with either mild or severe benign prostatic hyperplasia. The other cases (2 and 3) had left calf swelling.

Significant cardiac findings were seen in cases 3, 4, 5 and 6. These included cardiomegaly (cases 4, 5 and 6), coronary artery disease (cases 4, 5 and 6), ventricular dilation (cases 3 and 6), ventricular hypertrophy (cases 4 and 5), valve degeneration (cases 5 and 6) and aortic atheroma (cases 5 and 6). Case 6 had subendocardial fibrosis, fibrosis and inflammation of the pericardium (pericarditis) as well as conjunctival swelling.

As expected, pulmonary oedema (cases 1, 3, 5 and 6) and congested liver (cases 1, 5, and 6) were also seen.

Targeted histology was conducted for cases 1, 2, 3 and 4. Findings are summarized in Table 2.

In the lungs, emphysema was seen in cases 1, 2 and 4. Mild chronic inflammation was seen in cases 3 and 4. At least one micro-thromboembolism was seen in cases 1 and 3. Case 3 showed anthracosis, type 2 pneumocyte hyperplasia and early acute bronchopneumonia. In addition, occasional hyaline membranes were also identified.

In the heart, hypertensive nuclear changes were seen in cases 1, 2 and 3. Fibrosis was seen in cases 1 and 2. Lymphocytes were seen in the pericardium of cases 2 and 3. Case 2 had focal pericarditis.

Case 1 had additional histology performed which included mild fatty change of the liver as well as mild fibrosis and arteriolosclerosis of the kidney.

A retrospective review of all 140 post-mortems carried out by the same pathologist at Dublin City Mortuary in the year 2019 showed two cases of BPTE, a less severe form of ABPTE. The year 2020 showed that of 156 post-mortems, there were 5 cases of ABPTE (the four cases in this study and a fifth case in a COVID-19-negative patient in December 2020).

A literature review on the pathogenesis of thrombosis in COVID-19 highlighted the significant role of the endothelium in eliciting the hypercoagulable state seen in COVID-19 [10‐15].

Figure 2 demonstrates how SARS-CoV-2 enters a host cell via the transmembrane protein, angiotensin-converting enzyme 2 (ACE2), replicates and exits in order to infect other host cells, killing the host cell in the process [10, 11]. Crucially, ACE2, the means by which SARS-CoV-2 infects cells, is expressed by (among others) type II pneumocytes and endothelial cells that line arteries and veins. During homeostasis, ACE2 functions to lower blood pressure by converting angiotensin II (a vasoconstrictor) to angiotensin 1–7 (a vasodilator). However, ACE2 is internalized during viral entry and in turn downregulated, increasing the amount of circulating (vasoconstricting) angiotensin II.

Figure 3 shows how viral entry induces the host cell to produce interferons as part of the innate immune response to try to shut down viral replication in that cell and neighbouring cells. Interferons recruit macrophages to attack the virus which release proinflammatory cytokines IL-1 and TNF-α. Neighbouring cells are instructed to undergo apoptosis or to destroy RNA, all to reduce viral spread. IL-1 and TNF-α are also released during host cell death and apoptosis. If, however, the virus is successful in replicating virus particles (virions), these virions exit the cell and can enter the bloodstream via alveolar capillaries. Now the virus can infect endothelial cells.

Figure 4 shows the effect that IL-1 and TNF-α production has on thrombosis pathogenesis in COVID-19. IL-1 and TNF-α target uninfected endothelial cells as they contain the proinflammatory transcriptional hub, nuclear factor-κB, which results in more of these proinflammatory cytokines being produced. Angiotensin II described previously also stimulates nuclear factor-κB. IL-1 causes endothelial cells to produce IL-6, which acts on the liver to induce the acute phase response. IL-6 is also produced by macrophages. The liver produces fibrinogen, plasminogen activator inhibitor-1 (PAI-1) and C-reactive protein (CRP) as a result. Fibrinogen is a precursor of fibrin used to form thrombi. PAI-1 inhibits the activation of plasminogen into plasmin, which is responsible for fibrinolysis.

The term cytokine storm has been used abundantly when describing severe COVID-19 infection. In the context of the endothelium, a cytokine storm is when IL-1 produced by endothelial cells induces its own gene expression, causing it to continuously produce, creating a procoagulant state in the bloodstream (Fig. 4). IL-1 also induces the production of TNF-α, and TNF-α in turn induces the production of IL-1. A number of factors can pre-empt a cytokine storm in COVID-19 including impaired viral clearance, a low level of type 1 interferons, excessive neutrophil extracellular traps (NETs) (which are usually anti-viral) and viral apoptosis with the associated release of proinflammatory cytokines (pyroptosis).

The left part of Fig. 5 lists the properties of a normal endothelium during homeostasis. The right part of Fig. 5 shows what happens to the endothelium when infected and activated by IL-1 and TNF-α. During homeostasis, the endothelium is both anti-coagulant and profibrinolytic in nature. However, viral infection of endothelial cells can activate subendothelial tissue factor, which acts as the spark to start the coagulation cascade. Endothelial cells also store von Willebrand factor, which can be released, encouraging platelet aggregation and eventual clot formation. Under the same proinflammatory stresses, endothelial cells can release PAI-1, which inhibits fibrinolysis. The combination of prothrombic acute phase reactants produced by the liver and the procoagulant effects of IL-1 and TNF-α on endothelial cells causes an imbalance between thrombosis and fibrinolysis, resulting in excessive clot formation.

All cases in our autopsy study, on the surface, had at least one risk factor for severe COVID-19 infection. Autopsy and histological examination revealed additional risk factors. Case 1 had mild emphysema as well as arteriosclerosis and fibrosis of the kidneys. Case 2 had hypertensive nuclear changes and fibrosis of the heart as well as pericarditis. Case 3 had chronic inflammation and anthracosis of the lungs as well as dilated cardiomyopathy and hypertensive nuclear changes of the heart. Case 4 had emphysema and chronic inflammation of the lungs as well as ventricular hypertrophy and severe coronary artery disease. Case 5 had ventricular hypertrophy, coronary artery disease and aortic atheroma. Case 6 had significant heart disease. In addition, cases 2 and 3 displayed common non-specific histological findings often seen in COVID-19 and other viral infections [16]. Case 3 displayed type II pneumocyte hyperplasia, nuclear enlargement and occasional hyaline membranes in the lungs. Case 3 also displayed lymphocytic infiltration of the pericardium as did case 2. The combination of known risk factors, autopsy findings and histology findings underlines the fact that all cases in our study were at risk of severe COVID-19 infection despite not displaying symptoms.

The cause of death in 4 of the 6 cases was ABPTE. Several other autopsy studies of symptomatic patients report ABPTE or pulmonary thrombus to be a significant cause of death in COVID-19 patients [2-4, 17, 18]. ICU studies of COVID-19 patients showed the incidence of pulmonary embolism to be twice as frequent compared to a control group [19], while 49% of COVID-19 ICU patients displayed thrombotic complications [20]. In contrast, our retrospective review of 2 years’ worth of post-mortems showed ABPTE to be a rare event in non COVID-19 deaths. These findings suggest that pulmonary embolism is a major contributory factor in COVID-19 deaths.

There have been a handful of papers written about pulmonary thrombus or thromboembolism in asymptomatic COVID-19 infection [21‐25]. These papers mostly describe individuals who were initially asymptomatic but later became symptomatic. However, the cases in our report all remained asymptomatic. One paper describes a case of fatal acute bilateral pulmonary thrombus in an individual who remained asymptomatic [24]. To our knowledge, this is the first report of fatal ABPTE in a case of asymptomatic COVID-19 infection.

In Ireland, of those with COVID-19 who died of acute respiratory distress syndrome, COVID-19 was documented as a causative condition in part 1 of the death certificate. After detailed discussions between the coroner and pathologist, in the 4 cases of ABPTE, a decision was made to record COVID-19 as a contributory condition in part 2 of all 4 death certificates. Where COVID-19 lies in the death certification process requires further research. However, it could be argued, based on our initial observations and the data from the literature we cite, that going forward, in the case of asymptomatic COVID-19 deaths attributed to ABPTE, COVID-19 should be documented as a contributory condition. Whether COVID-19 can be documented as a causative condition of ABPTE requires confirmation from larger statistically significant studies. Perhaps when the COVID-19 pandemic has subsided and mandatory autopsy testing is no longer required, cases of ABPTE should continue to be tested for COVID-19.

In summary, our limited study of 6 asymptomatic COVID-19 deaths, related retrospective review of past post-mortems and literature review on the pathogenesis of thrombosis in COVID-19 have shown that thrombosis and massive thromboembolism may be a significant cause of death.