Original Research6 May 2020

Autopsy Findings and Venous Thromboembolism in Patients With COVID-19

A Prospective Cohort Study
    Author, Article, and Disclosure Information



    The new coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has caused more than 210 000 deaths worldwide. However, little is known about the causes of death and the virus's pathologic features.


    To validate and compare clinical findings with data from medical autopsy, virtual autopsy, and virologic tests.


    Prospective cohort study.


    Autopsies performed at a single academic medical center, as mandated by the German federal state of Hamburg for patients dying with a polymerase chain reaction–confirmed diagnosis of COVID-19.


    The first 12 consecutive COVID-19–positive deaths.


    Complete autopsy, including postmortem computed tomography and histopathologic and virologic analysis, was performed. Clinical data and medical course were evaluated.


    Median patient age was 73 years (range, 52 to 87 years), 75% of patients were male, and death occurred in the hospital (n = 10) or outpatient sector (n = 2). Coronary heart disease and asthma or chronic obstructive pulmonary disease were the most common comorbid conditions (50% and 25%, respectively). Autopsy revealed deep venous thrombosis in 7 of 12 patients (58%) in whom venous thromboembolism was not suspected before death; pulmonary embolism was the direct cause of death in 4 patients. Postmortem computed tomography revealed reticular infiltration of the lungs with severe bilateral, dense consolidation, whereas histomorphologically diffuse alveolar damage was seen in 8 patients. In all patients, SARS-CoV-2 RNA was detected in the lung at high concentrations; viremia in 6 of 10 and 5 of 12 patients demonstrated high viral RNA titers in the liver, kidney, or heart.


    Limited sample size.


    The high incidence of thromboembolic events suggests an important role of COVID-19–induced coagulopathy. Further studies are needed to investigate the molecular mechanism and overall clinical incidence of COVID-19–related death, as well as possible therapeutic interventions to reduce it.

    Primary Funding Source:

    University Medical Center Hamburg-Eppendorf.

    Since it was first detected in December 2019, the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spread from the central Chinese province of Hubei to almost every country in the world (1, 2). Most persons with COVID-19 have a mild disease course, but about 20% develop a more severe course with a high mortality rate (3). As of 26 April 2020, more than 2.9 million persons have been diagnosed with COVID-19 and 210 000 of them have died (4). Why the new coronavirus seems to have a much higher mortality rate than the seasonal flu is not completely understood. Some authors have reported potential risk factors for a more severe disease course, including elevated D-dimer levels, a high Sequential Organ Failure Assessment score, and older age (5, 6). Because of the novelty of the pathogen, little is known about the causes of death in affected patients and its specific pathologic features. Despite modern diagnostic tests, autopsy is still of great importance and may be a key to understanding the biological characteristics of SARS-CoV-2 and the pathogenesis of the disease. Ideally, knowledge gained in this way can influence therapeutic strategies and ultimately reduce mortality. To our knowledge, only 3 case reports have been published about COVID-19 patients who have undergone complete autopsy (7, 8). Therefore, in this study we investigated the value of autopsy for determining the cause of death and describe the pathologic characteristics in patients who died of COVID-19.

    Study Design

    In response to the pandemic spread of SARS-CoV-2, the authorities of the German federal state of Hamburg ordered mandatory autopsies in all patients dying with a diagnosis of COVID-19 confirmed by polymerase chain reaction (PCR). The legal basis for this was section 25(4) of the German Infection Protection Act. Because of legal regulations, no COVID-19 death was exempted from this order, even if its clinical cause seemed obvious. The case series demonstrated herein consists of 12 consecutive autopsies, starting with the first known SARS-CoV-2–positive death occurring in Hamburg (the second largest city in Germany, with 1.8 million inhabitants). All autopsies were performed at the Department of Legal Medicine of University Medical Center Hamburg-Eppendorf. The Ethics Committee of the Hamburg Chamber of Physicians was informed about the study (no. WF-051/20). The study was approved by the local clinical institutional review board and complied with the Declaration of Helsinki. In all deceased patients, postmortem computed tomography (PMCT) and a complete autopsy, including histopathologic and virologic evaluation, were performed. Clinical records were checked for preexisting medical conditions and medications, current medical course, and antemortem diagnostic findings.

    PMCT, Autopsy, and Histologic Examination

    Computed tomographic examination was done at the Department of Legal Medicine with a Philips Brilliance 16-slice multidetector scanner in accordance with an established protocol (9). In brief, full-body computed tomography was performed from top to thigh (slice thickness, 1 mm; pitch, 1.5; 120 kV; 230 to 250 mAs), complemented by dedicated scans of the thorax with higher resolution (slice thickness, 0.8 mm; pitch, 1.0; 120 kV; 230 to 250 mAs). We performed external examinations and full-body autopsies on all deceased persons with SARS-CoV-2 positivity (PCR confirmed) as soon as possible after taking proper safety precautions (using personal protective equipment with proper donning and doffing), following guidelines from the German Association of Pathologists, which are closely aligned with relevant international guidelines. The recently published recommendations for the performance of autopsies in cases of suspected COVID-19 were taken into account (10). The interval from death to postmortem imaging and autopsy (postmortem interval) ranged from 1 to 5 days. During autopsy, tissue samples for histology were taken from the following organs: heart, lungs, liver, kidneys, spleen, pancreas, brain, prostate and testes (in males), ovaries (in females), small bowel, saphenous vein, common carotid artery, pharynx, and muscle.

    For virologic testing, we took small samples of heart, lungs, liver, kidney, saphenous vein, and pharynx and sampled the venous blood.

    Tissue samples for histopathologic examination were fixed in buffered 4% formaldehyde and processed via standard procedure to slides stained with hematoxylin–eosin. For the lung samples, we also used the keratin marker AE1/AE3 (Dako) for immunohistochemistry.

    Quantitative SARS-CoV-2 RNA Reverse Transcription PCR From Tissue

    Tissue samples were ground by using ceramic beads (Precellys lysing kit) and extracted by using automated nucleic acid extraction (MagNA Pure 96 [Roche]) according to manufacturer recommendations. For virus quantification in tissues, a previously published assay was adopted with modifications (11). One-step real-time PCR was run on the LightCycler 480 system (Roche) by using a 1-step RNA control kit (Roche) as master mix. The Ct (cycle threshold) value for the target SARS-CoV-2 RNA (fluorescein) and whole-process RNA control (Cy5) was determined by using the second derivative maximum method. For quantification, standard in vitro–transcribed RNA of the E gene of SARS-CoV-2 was used (12). These samples were also analyzed in a study focusing on renal tropism of SARS-CoV-2 (Puelles V, et al. Multi-organ and renal tropism of SARS-CoV-2. In preparation).

    Statistical Analysis

    Data that were normally distributed are presented as means (SDs); data outside the normal distribution are presented as medians (ranges). Categorical variables were summarized as counts and percentages. All data were analyzed with Statistica, version 13 (StatSoft).

    Role of the Funding Source

    The sponsor was not involved in the design or conduct of the study, nor in the analysis of the data or the decision to submit the manuscript.

    Clinical Data

    The median age of the 12 patients included in this study was 73 years (interquartile range, 18.5 years); 25% were women. For all patients, preexisting chronic medical conditions, such as obesity, coronary heart disease, asthma or chronic obstructive pulmonary disease, peripheral artery disease, diabetes mellitus type 2, and neurodegenerative diseases, could be identified (Table 1). Two patients died out of the hospital after unsuccessful cardiopulmonary resuscitation, 5 died after treatment in the intensive care unit, and the remaining 5 had an advanced directive for best supportive care and died in the non–intensive care ward. Laboratory results for clinical chemistry, hematology, and coagulation were not available for the patients who died out of the hospital. In the remaining patients, the most striking features of the initial laboratory test were elevated levels of lactate dehydrogenase (median, 7.83 µkat/L [range, 2.71 to 11.42 µkat/L]), D-dimer (available for 5 patients; median, 495.24 nmol/L [range, 20.38 to >1904.76 nmol/L]), and C-reactive protein (median, 189 mg/L [range, 18 to 348 mg/L]), as well as mild thrombocytopenia in 4 of 10 patients. A procalcitonin test had been performed in 6 patients, and the results were negative in all but 1 patient with pneumonia (case 10). Table 2 provides an overview of the initial laboratory results.

    Table 1. Patient Characteristics and Autopsy Findings
    Table 2. Overview of Laboratory Results Taken at the Time of Hospitalization*
    Postmortem Computed Tomography

    In 2 cases (2 and 4), PMCT was not possible for logistic reasons. In the remaining cases, PMCT demonstrated mixed patterns of reticular infiltrations and severe, dense, consolidating infiltrates in both lungs in the absence of known preexisting pathology (such as emphysema or tumor). A juxtaposition of antemortem and postmortem findings is demonstrated in Figure 1. A complete summary of PMCT findings is presented in Table 1.

    Figure 1. Antemortem versus postmortem computed tomographic (CT) imaging (case 3).

    Top. Contrast medium–enhanced CT scan demonstrates the antemortem findings: bilateral ground glass opacities in the lower lobes of both lungs (yellow asterisks) and a chest tube (yellow arrow), which has been introduced to treat a pneumothorax (yellow arrowheads). Bottom. CT scan without contrast medium enhancement demonstrates the corresponding postmortem findings. For technical reasons, the postmortem image has a lower resolution. To protect the staff from potential infection, bodies were scanned in a double-layer body bag with the arms positioned alongside the body. Although the findings correspond to the antemortem images, ground glass opacities in both lower lobes (yellow asterisks) and a chest tube (yellow arrow) are seen. In addition, a central venous line (red arrowhead) and gastric tube (red arrow) are visible.


    In 4 cases (1, 3, 4, and 12), massive pulmonary embolism was the cause of death, with the thrombi deriving from the deep veins of the lower extremities. In another 3 cases (5, 8, and 11), fresh deep venous thrombosis was present in the absence of pulmonary embolism. In all cases with deep venous thrombosis, both legs were involved (Figure 2). In 6 of the 9 men (two thirds) included in the study, fresh thrombosis was also present in the prostatic venous plexus (Appendix Figure 1).

    Figure 2. Macroscopic autopsy findings.

    A. Patchy aspect of the lung surface (case 1). B. Cutting surface of the lung in case 4. C. Pulmonary embolism (case 3). D. Deep venous thrombosis (case 5).

    Appendix Figure 1. Thrombosis of the prostatic vein (case 1) (arrows).

    In all 12 cases, the cause of death was found within the lungs or the pulmonary vascular system. However, macroscopically differentiating viral pneumonia with subsequent diffuse alveolar damage (a histologic diagnosis) from bacterial pneumonia was not always possible. Typically, the lungs were congested and heavy, with a maximum combined lung weight of 3420 g in case 11. The mean combined lung weight was 1988 g (median, 2088 g). Standard lung weights for men and women are 840 g and 639 g, respectively (13, 14). Only cases 6 and 9 presented with a relatively low lung weight: 550 g and 890 g, respectively (Appendix Table 1). The lung surface often displayed mild pleurisy and a distinct patchy pattern, with pale areas alternating with slightly protruding and firm, deep reddish blue hypercapillarized areas. On the cutting surfaces, this pattern was also visible (Figure 2). The consistency of the lung tissue was firm yet friable. In 8 cases, all parts of the lungs were affected by these changes. Cases 6, 7, and 9—occurring in the 3 women of the case series—presented with changes compatible with focal purulent bronchopneumonia. Macroscopically, no changes were observed outside the lungs and respiratory tract, except for splenomegaly in 3 cases, which suggested a viral infection.

    Appendix Table 1. Weights of Individual Organs, in Grams, for All Cases*

    During autopsy, all cases except for case 6 presented with preexisting heart disease, including high-grade coronary artery sclerosis (7 of 12); myocardial scarring, indicating ischemic heart disease (6 of 12); and congestive cardiomyopathy. Mean heart weight was 503 g (median, 513 g). In addition to this finding, the most common accompanying diseases were pulmonary emphysema (6 of 12) and ischemic enteritis (3 of 12). Often these conditions were known to the treating physician before death (compare columns 4 and 10 of Table 1). The macroscopic autopsy findings are presented organ by organ in Appendix Table 2 and the lung findings in Table 1.

    Appendix Table 2. Macroscopic Autopsy Findings in Organs Other Than the Lung in Patients Dying of COVID-19*

    A clear trend toward obesity was observed among the cases (mean body mass index, 28.7 kg/m2; median, 28.7 kg/m2). However, case 9, involving a patient with known neuroendocrine tumor of the lung, presented with severe cachexia (body mass index, 15.4 kg/m2). The comorbid conditions found are summarized in Table 1.


    Histopathology of the lungs showed diffuse alveolar damage, consistent with early acute respiratory distress syndrome in 8 cases. Predominant findings were hyaline membranes (Figure 3, A and B), activated pneumocytes, microvascular thromboemboli, capillary congestion, and protein-enriched interstitial edema. As described by Wang and colleagues (15), a moderate degree of inflammatory infiltrates concurred with clinically described leukopenia in patients with COVID-19 and predominant infiltration of lymphocytes fit the picture of a viral pathogenesis. In later stages, squamous metaplasia was present (Figure 3, C). Long-term changes, such as destruction of alveolar septae and lymphocytic infiltration of the bronchi, were often visible as preexisting conditions. Four cases (6, 8, 9, and 10) showed no diffuse alveolar damage but extensive granulocytic infiltration of the alveoli and bronchi, resembling bacterial focal bronchopneumonia. Histologically, thromboemboli were detectable in cases 1, 3, 4, and 5 (Figure 3, D). Microthrombi were regularly found within small lung arteries, occasionally within the prostate, but not in other organs.

    Figure 3. Histopathologic findings.

    A. Diffuse alveolar damage with hyaline membranes (case 4) (hematoxylin–eosin [H&E] stain; original magnification, × 50). B. Hyaline membranes (case 4) (cytokeratin AE1/AE3 stain; original magnification, × 50). C. Squamous metaplasia in the lung (case 5) (H&E stain; original magnification, × 100). D. Pulmonary embolism (case 1) (H&E stain; original magnification, × 100).

    In addition to the lung changes described in Table 1, there were isolated histologic findings that might indicate a viral infection. The pharyngeal mucosa was examined in 7 cases. In 6 of them, hyperemia and alternating dense, predominantly lymphocytic infiltrates were found as signs of chronic pharyngitis. In 1 case (case 3), lymphocytic myocarditis was seen in the right ventricle (Appendix Figure 2). The remaining histologic changes were compatible with shock changes in part of the deceased patient (liver, kidneys, intestine) or corresponded to the macroscopically determined virus-independent preexisting pathology (such as ischemic cardiomyopathy).

    Appendix Figure 2. Mononuclear infiltrations consisting of lymphocytes (arrows) in the myocardium of the right ventricle (case 3) (hematoxylin–eosin stain; original magnification, × 100).

    Apart from findings related to SARS-CoV-2 infection, patients showed other histopathologic findings related to their chronic preexisting conditions, including hypertrophy of myocardial fibers or scarring of the myocardium. The peripheral veins, including those occluded by thrombi, showed no abnormalities on hematoxylin–eosin staining.

    PCR Results

    Quantitative reverse transcription PCR detected SARS-CoV-2 RNA in the lungs of all 12 patients (range, 1.2 × 104 to 9 × 109 copies/mL) and in the pharynx of 9 patients. Six patients showed moderate viremia (<4 × 104 copies/mL). In 5 of these patients, viral RNA was also detected in other tissues (heart, liver, or kidney) in concentrations exceeding viremia. Patients without viremia showed no or a low virus load in the other tissues. Only 4 patients had detectable viral RNA in the brain and saphenous vein.


    In this autopsy study of 12 consecutive patients who died of COVID-19, we found a high incidence of deep venous thrombosis (58%). One third of the patients had a pulmonary embolism as the direct cause of death. Furthermore, diffuse alveolar damage was demonstrated by histology in 8 patients (67%).

    To our knowledge, this is the first case series summarizing and comparing clinical data of consecutive COVID-19 cases with findings obtained by a full autopsy, supplemented by PMCT, histology, and virology.

    The high rate of death-causing pulmonary embolism at autopsy correlates well with the unsuccessful resuscitation of 3 of 4 patients, 2 of whom died out of the hospital. Apart from that, no preclinical evidence had been reported of pulmonary embolism or deep venous thrombosis.

    In studies that examined deceased patients with COVID-19 without relying on autopsy, no increased rates of pulmonary embolism were observed clinically. However, it is known that many cases of pulmonary embolism remain clinically overlooked and are often associated with sudden, unexpected death. This may have been aggravated by the method for diagnosing COVID-19 in Germany, which is based on PCR tests rather than computed tomographic imaging because of concerns about infection of medical staff and other patients. A recent report described clinical features of 85 fatal cases of COVID-19 from Wuhan (16). Besides respiratory failure, the cause of death was multiorgan failure in 16% and cardiac arrest in 9%. No autopsies were performed. The gold standard for identifying cause of death is still the autopsy (17). However, in-hospital autopsy rates have declined worldwide over the past decades. Also, because of pathologists' potential risk for SARS-CoV-2 infection, very few autopsies have been performed worldwide (18). To our knowledge, only 3 case reports have been published on patients with COVID-19 who have undergone complete autopsy and a few more in which only lung tissue was examined (7, 8).

    Other researchers have described coagulopathy as a common complication in patients with severe COVID-19 (5, 6, 19). In a recent study of 191 patients with COVID-19, 50% of those who died had coagulopathy, compared with 7% of survivors. D-dimer levels greater than 1000 µg/L were associated with a fatal outcome (6).

    COVID-19 may predispose to venous thromboembolism in several ways. The coagulation system may be activated by many different viruses, including HIV, dengue virus, and Ebola virus (20, 21). In particular, coronavirus infections may be a trigger for venous thromboembolism, and several pathogenetic mechanisms are involved, including endothelial dysfunction, characterized by increased levels of von Willebrand factor; systemic inflammation, by Toll-like receptor activation; and a procoagulatory state, by tissue factor pathway activation (22). In a subgroup of patients with severe COVID-19, high plasma levels of proinflammatory cytokines were observed (23). The direct activation of the coagulation cascade by a cytokine storm is conceivable. With COVID-19, severe hypoxemia develops in some patients (24). Thrombus formation under hypoxic conditions is facilitated both in animal models of thrombosis and in humans. The vascular response to hypoxia is controlled primarily by the hypoxia-inducible transcription factors, whose target genes include several factors that regulate thrombus formation (25). Lastly, indirect causes, such as immune-mediated damage by antiphospholipid antibodies, may partially contribute, as speculated by Zhang and colleagues (26).

    The macroscopic findings in our autopsy series—with rather heavy, consolidated, friable, basically air-free lungs in most of the cases—were impressive and explain the difficulties in sufficiently ventilating some of these patients. The histopathologic changes in most of our cases with diffuse alveolar damage as the main finding resemble those described by Xu and colleagues (7) and Barton and colleagues (8), who reported single cases; Zhang and colleagues (26), who reported on lung biopsy in a patient with SARS-CoV-2 positivity; and Tian and colleagues (27), who described macroscopic and histologic pulmonary findings in 2 patients with lung cancer who received positive results on SARS-CoV-2 testing. However, the full-blown picture of diffuse alveolar damage seems to be more prevalent in younger patients with fewer preexisting diseases and longer survival, whereas older patients with more comorbid conditions tend to die in the early stages of the disease.

    In line with clinical, macroscopic, and histopathologic findings, PCR detected the highest concentration of SARS-CoV-2 RNA in lung and pharyngeal tissue. Of interest, in most patients with disease, high titers of RNA were also detected in postmortem samples. The clinical relevance of this is not yet clear. Clearance of viral RNA from blood 7 days after transfusion of COVID-19 convalescent plasma was associated with substantial clinical improvement, but studies have not shown a correlation between viremia and acute respiratory distress syndrome in patients with severe COVID-19 (28, 29). As in patients with SARS-CoV-1, in whom viral replication could be detected in other organs, including the liver, kidney, spleen, and cerebrum (30), we detected viral RNA at high titers in other organs (liver, kidney, and heart) in 5 patients. These data suggest that SARS-CoV-2 may spread via the bloodstream and infect other organs. To prove this, replication intermediates must be detected.

    The current study had some limitations. First, the sample size was small, possibly leading to overestimation of the rate of pulmonary embolism. However, both the clinical and postmortem observations agree well with the current knowledge about SARS-CoV-2 pathology. This includes the sex and age distribution as well as the preexisting conditions among the patients, but also the histologic findings. Second, although viral titers in swabs (pharynx) taken longitudinally up to 7 days after death remained similar, we lack data on how postmortem processes affect viral titers and dynamics in different tissues and body fluids. Moreover, the quantitative PCR assay used cannot discriminate between genomic and subgenomic RNA. As stated earlier, to prove viral replication, detection of replication intermediates or antigenomic RNA would be necessary.

    In conclusion, we found a high incidence of thromboembolic events in patients with COVID-19. When hemodynamic deterioration occurs in a patient with COVID-19, pulmonary embolism should always be suspected. That patients with COVID-19 who have increased D-dimer levels, a sign of coagulopathy, may benefit from anticoagulant treatment seems plausible (31). As demonstrated in our cohort, this might be important for hospitalized patients and outpatients. In this context, some professional societies have already made recommendations for antithrombotic therapy for patients with COVID-19 (32). Robust evidence, however, remains scant, and further prospective studies are urgently needed to confirm and validate these results.



    Sign In to Submit A Comment
    Alok Srivastava, Jecko Thachil,13 May 2020
    Autopsy findings in COVID-19 - Is it thrombosis or embolism?

    Sir, We read with interest the excellent detailed report on autopsy findings in patients with COVID-19 infection.(1) The authors’ description of the gross and histopathology of changes in the lung emphasizing the thrombosis in microvasculature and haemorrhage in the alveoli is very significant. It would have been useful to include more details about the microthrombi in the arteries particularly whether it was found in all decedents. The presence of extensive microvascular thrombi is highly suggestive of a local thrombotic process.(2)

    Even though 7 out of the 12 patients evaluated in this series had evidence of deep venous thrombosis (DVT), in all probability, this was a late effect in patients who were seriously ill for several days in the hospital. The primary event is very likely to have been pulmonary thrombosis as evidenced by extensive microthrombi in small pulmonary arteries of these decedents. The autopsy features described in this report support the hypothesis that the predominant pathology in these patients with COVID-19 associated hemostasis abnormality (CAHA) is microvascular thrombosis. Even at its early stage in ambulant patients, the breakdown of these micro-clots tend to cause raised d-dimer levels.(3) With disease progression, the marked coagulation activation and extensive microthrombi lead to extremely high d-dimer levels which have been shown to correlate with worse clinical outcomes.(4) It is important to recognize this spectrum of early to late CAHA to plan timely interventions. In this report, d-dimer levels were highly elevated (20-2000 fold) in all five patients for whom this data was available. Therefore it was extensive thrombosis rather than ‘pneumonia’ which was the cause of respiratory failure.

    We would like to emphasize that pulmonary vascular changes in CAHA is distinct from classical ‘thromboembolism’. In COVID-19, the cause of thrombi in pulmonary vasculature is not distal thrombosis embolizing to the lung but rather de novo thrombosis in the microvasculature. The fact that alveolar type 2 cells and the endothelium in the lung share receptors which mediate SARS-CoV-2 infection further supports this hypothesis. Recognizing this difference between early, localized, organ-specific pulmonary thrombosis leading to respiratory failure and systemic thromboembolism in the late stages is critical to planning suitable investigations and management strategies. Needless to say, early use of anticoagulants in patients with high or rising d-dimer levels is paramount and have already been shown to improve survival. (5)


    1.Wichmann D, Sperhake JP, Lutgehetmann M, Steurer S, Edler C, Heinemann A, et al. Autopsy Findings and Venous Thromboembolism in Patients With COVID-19: A Prospective Cohort Study. Ann Intern Med (in press) https://doi.org/ 10.7326/M20-2003. . 2020.

    2.Thachil J, Srivastava A. SARS-2 Coronavirus–Associated Hemostatic Lung Abnormality in COVID-19: Is It Pulmonary Thrombosis or Pulmonary Embolism? Seminars in Thrombosis and Hemostasis (in press) https://doi.org/10.1055/s-0040-1712155. 2020.

    3.Thachil J, Cushman M, Srivastava A. A Proposal for Staging COVID‐19 Coagulopathy. Research and Practice in Thrombosis and Haemostasis (in press) https://doi.org/10.1002/rth2.12372. 2020.

    4.Tang N, Li D, Wang X, Sun Z. Abnormal coagulation parameters are associated with poor prognosis in patients with novel coronavirus pneumonia. J Thromb Haemost. 2020;18(4):844-7.5.Tang N, Bai H, Chen X, Gong J, Li D, Sun Z. Anticoagulant treatment is associated with decreased mortality in severe coronavirus disease 2019 patients with coagulopathy. J Thromb Haemost. 2020;18(5):1094-9.

    Antonio Mirijello, MD; Elvira Grandone, MD; Salvatore De Cosmo, MD13 May 2020
    Comment on Wichmann D et al.: Thrombotic complications of COVID-19


    We read with great interest the article by Wichmann and colleagues [1] examining twelve consecutive
    patients deceased because of Sars-Cov-2 infection. Autoptic studies are pivotal in understanding
    mechanisms of new and unknown diseases, particularly for COVID-19.

    The main finding was the high prevalence of deep vein thrombosis (DVT) in seven patients (58%), being
    pulmonary embolism (PE) the direct cause of death in four patients (33%) [1].

    In line with literature reports [2], all the evaluated patients were affected by chronic comorbidities (e.g.
    cardiovascular, metabolic, respiratory, neurological, oncological).

    Among the seven patients with DVT [1], four underwent mechanical ventilation: all developed venous
    thromboembolism (VTE) and three died of PE. Only two out of seven DVT patients had received prophylaxis
    for VTE with low molecular weight heparin (LMWH). However, this treatment was not effective in
    preventing VTE, as both died because of PE. D-dimer for both were not available. Indeed, D-dimer levels
    had been assessed in only five out of twelve patients; excluding two out of hospital deaths, only five out of
    ten patients (50%) had a D-dimer assay in their clinical records. No autoptic signs of VTE were found in
    those two patients on treatment with direct-acting oral anticoagulants (DOACs).

    These are our considerations: COVID-19 patients are heterogeneous in terms of characteristics and clinical
    management (ICU vs general wards). As underlined [1], COVID-19 is associated with thrombotic
    manifestations and coagulopathy, negatively influencing the disease course [3]. Besides VTE, the
    mechanism of pulmonary vascular thrombosis has been hypothesized as a consequence of interstitial
    pneumonia causing a severe acute inflammation and prothrombotic complement/cytokines-mediated
    endothelial dysfunction [4]. In this context, anticoagulant treatment seems to reduce mortality in severe
    patients with coagulopathy (e.g. high D-dimer) [5].

    Thrombotic risk of acute medical patients is often underestimated given the lack of clinical signs of
    thrombosis (e.g. swollen leg, Homan’s sign). However, acute conditions (i.e. respiratory failure) together
    with comorbidities significantly raise this risk. In the setting of COVID-19, validated prediction scores and D-
    dimer testing are useful to assess thrombotic risk and for risk stratification.

    Whether all COVID-19 patients should receive standard or intermediate-doses LMWH prophylaxis for the
    prevention of thrombotic complications remains an open question. Similarly, the utility of lower limbs
    compression ultrasonography or pulmonary CT angiography to high-risk patients as well as the role of
    DOACs need further evaluation.


    1. Wichmann D, Sperhake JP, Lütgehetmann M, et al. Autopsy Findings and Venous Thromboembolism in
      Patients With COVID-19: A Prospective Cohort Study. Ann Intern Med. 2020 May 6. doi: 10.7326/M20-
    2. Guan WJ, Ni ZY, Hu Y, et al. Clinical Characteristics of Coronavirus Disease 2019 in China. N Engl J Med.
      2020 Apr 30;382(18):1708-1720. doi: 10.1056/NEJMoa2002032.
    3. Violi F, Pastori D, Cangemi R, Pignatelli P, Loffredo L. Hypercoagulation and Antithrombotic Treatment
      in Coronavirus 2019: A New Challenge. Thromb Haemost. 2020 Apr 29. doi: 10.1055/s-0040-1710317.
    4. Marongiu F, Grandone E, Barcellona D. Pulmonary thrombosis in 2019-nCoV pneumonia? J Thromb
      Haemost. 2020 Apr 15. doi: 10.1111/jth.14818.
    5. Tang N, Bai H, Chen X, Gong J, Li D, Sun Z. Anticoagulant treatment is associated with decreased
      mortality in severe coronavirus disease 2019 patients with coagulopathy. J Thromb Haemost. 2020
      May;18(5):1094-1099. doi: 10.1111/jth.14817.


    Authors declare no conflict of interest.

    Dominic Wichmann18 May 2020
    Reply to Srivastava and Thachil

    We appreciate the comments of Srivastava and Thachil in which they give an explanation for the origin of micro vascular thrombemboli (MVT) and discuss if the pulmonary embolisms (PE) observed in 4 of our 12 patients were most likely explained by the severe illness of the patients and the prolonged cause of the disease.We would like to state that to the best of our judgement MVT and PE result from different pathological mechanisms. We observed MVT ubiquitously in all parts of the lung, but in contrast to Srivastava and Thachil we do not consider this as a specific feature of SARC-CoV-2-pneumonia.

    We think it is rather a long known finding in viral pneumonias, resulting from the interaction of the innate immunity and a viral pathogen. Identical findings have been first describer in patient during the 1918 Influenza pandemic and for many other viral pathogens later (1).With respect to the comment on the origin of PE in our patients we admit that many of them were severely ill and were at high risk for PE. But half of the patients who died from PE were only mildly ill and had a cardiac arrest as outpatients. Which emphasizes the need for urgent research in this area.

    1) Taubenberg JK and Morens DM. The Pathology of Influenza Virus Infections. Annu Rev. 2008. Aug. 11. doi: 10.1146/annurev.pathmechdis.3.121806.154316

    Dominic Wichmann18 May 2020
    Reply to Mirijello et al.
    We appreciate the thoughtful comments of Antonio Mirijello and colleagues. Indeed, the underlying conditions leading to thromboembolisms in COVID-19 patients are not completely understood. As demonstrated in our manuscript the potential benefit of D-dimer testing was also not clear to us. As a result of this study we have implemented the test into our regular routine for COVID-19-patients.We also value the comment on the uncertainty regarding the clinical implications of our findings. Which patient groups might benefit from an intensified prophylaxis with LMWH or if there are certain “high-risk” groups which might even benefit from a prolonged treatment with oral anticoagulants (DOAC or VKA) remains subject to ongoing studies.
    Oleg Epelbaum MD FACP19 May 2020
    Venous Thomboembolism in Fatal COVID-19

    TO THE EDITOR:  The autopsy series from Germany reported by Wichmann et al (1) was a welcome addition to the growing literature on the postmortem lung histology of SARS-CoV-2 lung disease, though unfortunately the finding of diffuse alveolar damage (DAD) in terminal cases does little to illuminate the process in its earlier stages.  Worrisome, however, was the authors’ emphasis on the “venous thromboembolism” aspect of their findings, which appears in the title and dominates the concluding paragraph.  If assimilated without context by the clinical community, this extract from the study’s results could further fuel the pervasive but unsubstantiated belief that COVID-19 is a uniquely hypercoagulable state.

    Case 1 in the study is a patient who sustained an out-of-hospital cardiac arrest and was found to have PE as the likely cause.  This is unsurprising, since well before the emergence of SARS-CoV-2, PE has been recognized as the most common non-cardioaortic etiology of unsuccessfully resuscitated community arrests (2).  The other three PE cases in the series were managed in the intensive care unit (ICU); all were obese and mechanically ventilated.  We are not informed whether these patients received appropriate venous thromboembolism (VTE) prophylaxis, which remains a global deficiency from which Germany is not exempt (3).  The antemortem detection rate of incidental PE in general ICU populations receiving mechanical ventilation can approach 20% (with obesity being a risk factor) and exceed that figure in autopsy studies, but this finding has not been linked to inferior outcomes clinically and has rarely been deemed a precipitant of death pathologically (4,5).  Turning to specifics pertinent to the patients in the Wichmann series, death with DAD is nearly universally accompanied by pulmonary vascular thrombosis, including macrovascular, so thrombi should not be construed as a unique feature of SARS-CoV-2 lung disease when DAD is present (6).  Furthermore, SARS-CoV-2 hardly stands out next to other viruses in regard to postmortem VTE; a study of eight autopsies performed on fatal H1N1 influenza cases revealed a higher percentage of PE than did the Wichmann series: 5/8 (63%) versus 4/12 (33%) (7).

    Although severe COVID-19 promotes hemostatic dysregulation, it is not alone among critical illnesses, and the findings of the Wichmann series do not advance the theory that critically ill COVID-19 patients are unusually predisposed to VTE and therefore merit an unprecedented approach. The authors, however, allude to the opposite.  They invoke an international guidance document as corroboration (8).  Majority of its expert author panel, however, voted against routine empirical anticoagulation.


    1. Wichmann D, Sperhake JP, Lütgehetmann M et al. Autopsy Findings and Venous Thromboembolism in Patients With COVID-19: A Prospective Cohort Study. Ann Intern Med 2020. [PMID: 32374815] doi: 10.7326/M20-2003.
    2. Virkkunen I, Paasio L, Ryynänen S et al. Pulseless electrical activity and unsuccessful out-of-hospital resuscitation: what is the cause of death? Resuscitation 2008;77:207-10. [PMID: 18249482] doi: 10.1016/j.resuscitation.2007.12.006.
    3. Kröger K, Moerchel C, Bus C, Serban M. Venous thromboembolism in Germany: results of the GermAn VTE registry (GATE-registry). Int J Clin Pract 2014;68:1467-72 [PMID: 25333964]. doi: 10.1111/ijcp.12504.
    4. Minet C, Lugosi M, Savoye PY et al. Pulmonary embolism in mechanically ventilated patients requiring computed tomography: Prevalence, risk factors, and outcome. Crit Care Med 2012;40:3202-8. [PMID: 23164766]. doi: 10.1097/CCM.0b013e318265e461.
    5. McLeod AG, Geerts W. Venous thromboembolism prophylaxis in critically ill patients. Crit Care Clin 2011;27:765-80. [PMID: 22082513]. doi: 10.1016/j.ccc.2011.07.001.
    6. Tomashefski JF Jr, Davies P, Boggis C, Greene R, Zapol WM, Reid LM. The pulmonary vascular lesions of the adult respiratory distress syndrome. Am J Pathol 1983;112:112-26. [PMID: 6859225].
    7. Harms PW, Schmidt LA, Smith LB et al. Autopsy findings in eight patients with fatal H1N1 influenza. Am J Clin Pathol 2010;134:27-35. [PMID: 20551263]. doi: 10.1309/AJCP35KOZSAVNQZW.
    8. Bikdeli B, Madhavan MV, Jimenez D et al. COVID-19 and Thrombotic or Thromboembolic Disease: Implications for Prevention, Antithrombotic Therapy, and Follow-up. J Am Coll Cardiol 2020;Apr 15:S0735-1097(20)35008-7. [PMID: 32311448]. doi: 10.1016/j.jacc.2020.04.031.
    Dominic Wichmann22 May 2020
    Reply to Epelbaum

    In his comment Dr. Epelbaum raises concerns that focusing on venous thromboembolism (VTE) or pulmonary embolism (PE) in COVID-19 is misguiding the attention of colleagues to an epiphenomenon. We strongly disagree with him about this issue especially because in our opinion the references he gives, do not backup his statements. Regarding the statement that VTE prophylaxis “remains a global deficiency…” the reference cited a study of Kröger et al. (1) . The main topic of this study was to investigate if patients presenting with VTE/PE in Germany had a risk factor which could have been identified previously. The majority of patients had no medical condition and no identifiable risk factor. Consequently, the algorithm decided against a prophylaxis. Which hardly can be extrapolated to COVID-19 patients. The statement that PE is a common (but rarely fatal) finding in CT of critically ill patients is correct (2) and most likely due to improved CT performance in recent years. Contrasting to this, one third of our patients had a fatal PE. Furthermore Dr. Epelbaum states that we corroborate the consensus statement of Bikdeli et al. to demand general anticoagulation treatment for COVID-19 patients (3). This is not correct, we stress the need for further studies on this subject. The consensus statement focuses in large parts on the effects of COVID-19 associated coagulopathies on anticoagulation strategies for cardio-vascular interventions. A small paragraph cited by Dr. Epelbaum deals with empiric anticoagulation therapy in COVID-19 patients: “The majority of panel members consider prophylactic anticoagulation, although a minority consider intermediate-dose or therapeutic dose to be reasonable.”. The paper was written before our and other studies have been published (4). Because the panel decision was based mainly on a single retrospective study from China in which only laboratory abnormalities have been presented and no autopsies have been conducted (5), our study and the one of Baldi et al. add substantial value to the claim of the consensus statement that for anticoagulation treatment the “optimal dosing in patients with severe COVID-19 remains unknown and warrants further prospective investigation.”

    1. Kroger K, Moerchel C, Bus C, Serban M. Venous thromboembolism in Germany: results of the GermAn VTE registry (GATE-registry). Int J Clin Pract. 2014;68(12):1467-72.

    2. Minet C, Lugosi M, Savoye PY, Menez C, Ruckly S, Bonadona A, et al. Pulmonary embolism in mechanically ventilated patients requiring computed tomography: Prevalence, risk factors, and outcome. Crit Care Med. 2012;40(12):3202-8.

    3. Bikdeli B, Madhavan MV, Jimenez D, Chuich T, Dreyfus I, Driggin E, et al. COVID-19 and Thrombotic or Thromboembolic Disease: Implications for Prevention, Antithrombotic Therapy, and Follow-up. J Am Coll Cardiol. 2020.

    4. Baldi E, Sechi GM, Mare C, Canevari F, Brancaglione A, Primi R, et al. Out-of-Hospital Cardiac Arrest during the Covid-19 Outbreak in Italy. N Engl J Med. 2020.

    5. Tang N, Bai H, Chen X, Gong J, Li D, Sun Z. Anticoagulant treatment is associated with decreased mortality in severe coronavirus disease 2019 patients with coagulopathy. J Thromb Haemost. 2020.

    Ferdinando D’Amico1,2 Silvio Danese1, Laurent Peyrin-Biroulet225 May 2020
    Pathogenesis of COVID-19 infection: the forgotten colon

    Dear Editors,

    We read with great interest the article by Wichmann and colleagues recently published in Annals of Internal Medicine (1). Autopsies were performed on 12 patients who died from coronavirus disease 2019 (COVID-19) to further investigate the pathogenesis of this disease. They first confirmed the risks of deep venous thrombosis and ischemic heart disease in individuals infected with COVID-19. Unexpectedly, they also found ischemic enteritis (3/12, 25%) on small bowel biopsies. Polymerase chain reaction (PCR) confirmed the presence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in the lungs of all patients, while in 5 cases viral RNA was found in kidneys, heart, or liver. A growing body of evidence indicates that the colon may be affected by SARS-CoV-2 infection. It is noteworthy that two central proteins involved in the pathogenesis of the new coronavirus infection, serine protease TMPRSS2 and angiotensin-converting enzyme 2 (ACE2) receptor, are highly expressed within the colonic mucosa (2). The former regulates spike protein cleavage allowing its activation, while the latter mediates virus entry into the host cell. Interestingly, the virus has been detected in the feces of positive subjects and gastrointestinal symptoms, among which diarrhea, are experienced in about 10% of patients (3). SARS-CoV-2 has been also identified in endoscopic rectal biopsies of two patients with severe forms of COVID-19, and higher levels of fecal calprotectin, which is known to reflect intestinal inflammation, have been found in patients with diarrhea compared to those without diarrhea (4,5).

    Unfortunately, no colonic biopsies were performed by Wichmann et al. (1). COVID-19 mainly affects the tracheobronchial tree and lung parenchyma, and respiratory symptoms are the most frequently encountered. It is the reason why most articles about COVID-19 focused on the respiratory tract at the beginning of the pandemic. However, COVID-19 is now recognized as a systemic disease and gastrointestinal symptoms should not be underestimated. Greater attention should be paid to the gastrointestinal tract, especially the colon. Studies reporting colonic histology of COVID-19 patients are needed to better understand the pathogenesis of this disease, to define whether ischemic and thromboembolic events may occur in the colon, and to explain why patients with gut involvement may have a more severe disease course.


    1. Wichmann D, Sperhake JP, Lütgehetmann M, et al. Autopsy Findings and Venous Thromboembolism in Patients With COVID-19 [published online ahead of print, 2020 May 6]. Ann Intern Med. 2020;M20-2003. doi:10.7326/M20-2003.

    2. Burgueño JF, Reich A, Hazime H, et al. Expression of SARS-CoV-2 Entry Molecules ACE2 and TMPRSS2 in the Gut of Patients With IBD. Inflamm Bowel Dis. 2020;26(6):797‐808. doi:10.1093/ibd/izaa085.

    3. D'Amico F, Baumgart DC, Danese S, et al. Diarrhea During COVID-19 Infection: Pathogenesis, Epidemiology, Prevention, and Management [published online ahead of print, 2020 Apr 8]. Clin Gastroenterol Hepatol. 2020;S1542-3565(20)30481-X. doi:10.1016/j.cgh.2020.04.001

    4. Lin L, Jiang X, Zhang Z, et al. Gastrointestinal symptoms of 95 cases with SARS-CoV-2 infection. Gut. 2020;69(6):997‐1001. doi:10.1136/gutjnl-2020-321013.

    5. Effenberger M, Grabherr F, Mayr L, et al. Faecal calprotectin indicates intestinal inflammation in COVID-19 [published online ahead of print, 2020 Apr 20]. Gut. 2020;gutjnl-2020-321388. doi:10.1136/gutjnl-2020-321388.


    F D’Amico declares no conflict of interest. S Danese has served as a speaker, consultant, and advisory board member for Schering-Plough, AbbVie, Actelion, Alphawasserman, AstraZeneca, Cellerix, Cosmo Pharmaceuticals, Ferring, Genentech, Grunenthal, Johnson and Johnson, Millenium Takeda, MSD, Nikkiso Europe GmbH, Novo Nordisk, Nycomed, Pfizer, Pharmacosmos, UCB Pharma and Vifor. L Peyrin-Biroulet has served as a speaker, consultant and advisory board member for Merck, Abbvie, Janssen, Genentech, Mitsubishi, Ferring, Norgine, Tillots, Vifor, Hospira/Pfizer, Celltrion, Takeda, Biogaran, Boerhinger-Ingelheim, Lilly, HAC- Pharma, Index Pharmaceuticals, Amgen, Sandoz, For- ward Pharma GmbH, Celgene, Biogen, Lycera, Samsung Bioepis, Theravance.

    Hasan Yazici MD27 May 2020
    Better avoid percentages when reporting about 12 autopsies
    Wichmann and colleagues give a very informative and thought provoking account of these 12 autopsies. However I must point out their very liberal use of percentages (13 times in the whole manuscript), I am afraid, unfortunately lessens the impact of what they aim to convey.
    Dominic Wichmann29 May 2020
    Reply to Yazici
    We were not aware about the restrictions to use of percentages and would like to apologize for its excessive use. Also we do not understand how this lessens the impact of our manuscript we must state that we did not intended to disguise any results. Due to the small base number of cases presented in our manuscript we assumed that the academic readership would be easily able to transfer percentages into real numbers.
    Dominic Wichmann29 May 2020
    Reply to D'Amico et al.

    In their comment Prof. D’Amico and colleagues raise an interesting point: the aspect of multi-organ involvement in SRAS-CoV-2-infections. In fact focusing on pulmonary pathology alone may not show the whole picture of COVID.(1) With the fast growing knowledge about pathology and tissue tropism of SARS-CoV-2 the scientific community may learn interesting things in the near future.

    1. Puelles VG, Lutgehetmann M, Lindenmeyer MT, Sperhake JP, Wong MN, Allweiss L, et al. Multiorgan and Renal Tropism of SARS-CoV-2. N Engl J Med. 2020.

    Dong Ji 1, Mingjie Zhang 2, Gregory Cheng 2,3, Feng Wang 4 George Lau 1,326 August 2020
    Is non-alcoholic fatty liver disease (NAFLD) is a risk factor of venous thrombosis in COVID-19 ?

    We read with great interest the article by Wichmann and colleagues [1] showing a prevalence of deep vein thrombosis (DVT) in 7/12(58%) patients.

    We previously reported a 25% DVT prevalence among 81 COVID-19 patients admitted to ICU and identified high D-dimer levels as a predictive factor for  DVT [2].We also observed that NAFLD was associated with a high risk of progression to severe COVID-19 [3]. NAFLD was characterized by a hypercoagulable state, with elevated plasma levels of von Willebrand factor, and increased levels of circulating plasminogen activator inhibitor type 1 [4]. It is possible that the hypercoagulable state in NAFLD may contribute to thrombosis in COVID-19. We retrospectively studied the prevalence of NAFLD among our DVT subjects [2] and compared the D-dimer levels of NAFLD subjects with non-NAFLD subjects in our previous COVID-19 cohort [3]. NAFLD was identified as hepatic steatosis index [HSI = 8× (ALT/AST) + BMI + 2 if type 2 diabetes and + 2 if female] more than 36 points and/or by abdominal ultrasound examination. 

    NAFLD was present in 76% (16/21) of DVT subjects as compared with 45%(27/60) prevalence in non-DVT subjects p=0.01. Alternatively, DVT was detected in 37.2% (16/43) and 13.2% (5/38) of NAFLD and non-NAFLD subjects respectively (p=0.01). The mean admission and peak D-dimer levels were also significantly higher in NAFLD subjects as compared with non-NAFLD group, 0.72 ± 1.10 ug/ml vs 0.38 ± 0.46 ug/ml, p=0.003 and 1.81 ± 4.1mg/ml vs 0.63 ± 0.41mg/ml, p=0.003 respectively. The association of NAFLD with admission and peak D-dimer levels remain significant in multivariate analysis, p=0.046 and p=0.028.

    Among the seven patients with DVT/PE (Case 1,3,4,5,8 11,12) reported by Wichmann et al [1], four ( Case, 1,4,8,12) had NAFLD risk factors such as obesity or diabetes, case 12 had macroscopic fatty changes at autopsy. It would be interesting to analyze whether there was a high prevalence of hepatic steatosis in the seven subjects with DVT in Wichmann's case series.  A high prevalence will support the hypothesis that NAFLD is a risk factor of thrombosis in COVID-19. As suggested, patients with severe COVID-19 had high plasma levels of pro-inflammatory cytokines [1], the liver is a front-line immune organ and increased production of pro-inflammatory cytokines by adipose cells and Kupffer cells had been reported in NAFLD patients [5] .Therefore,COVID-19 patients with underlying NAFLD may have higher likelihood of activation of the coagulation cascade by pro-inflammatory cytokines and subsequent thrombosis.


    1.       Wichmann D, Sperhake JP, Lütgehetmann M, et al. Autopsy Findings and Venous Thromboembolism in Patients With COVID-19. Ann Intern Med. 18 August 2020; M20-2003. doi:10.7326/M20-2003

    2.       Cui S, Chen S, Li X, Liu S, Wang F. Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia. J Thromb Haemost. 2020;18(6):1421-1424.

    3.       Ji D, Qin E, Xu J, Zhang D, Cheng G, Wang Y, et al. Non-alcoholic Fatty Liver Diseases in Patients With COVID-19: A Retrospective Study. J Hepatol. 2020 Apr 8. S0168-8278(20)302063 doi:10.1016/j.jhep.2020.03.044

    4.       Verrijken A FS, Mertens I, Prawitt J, Caron S, Hubens G, Van Marck E, et al. Prothrombotic factors in histologically proven NAFLD and NASH. Hepatology 2014; 59:121-129.1.   

     5.       Braunersreuther V, Viviani GL, Mach F , Montecucco F. Role of cytokines and chemokines in non-alcoholic fatty liver disease World J Gastroenterol. 2012 Feb 28; 18(8): 727–735


    Dominic Wichmann31 August 2020
    Reply to Dong Ji et al.

    We greatly appreciate the comment of Dong Ji and colleagues. Even though patients with non-alcoholic fatty liver disease (NAFLD) in general, represent a group with a large range of overlapping risk factors for thromboembolism, their well established thrombophilic and hyperinflammatory state represents a good explanation for an increased rate of VTE in the context of COVID-19. The ongoing SARS-CoV-2 epidemic with millions of patients involved will provide the chance to further investigate this in a larger context, including ethnical and social aspects.