Respiratory failure appears to be the ultimate mechanism of death in most patients with severe coronavirus disease 2019 (COVID-19) infection. Studies of postmortem COVID-19 lungs largely report diffuse alveolar damage and capillary fibrin thrombi, but we have also observed other patterns.
To report demographic and radiographic features along with macroscopic, microscopic, and microbiologic postmortem lung findings in patients with COVID-19 infections.
Patients with confirmed COVID-19 infection and postmortem examination (March 2020–May 2020) were included. Clinical findings were abstracted from medical records. Lungs were microscopically reviewed independently by 4 thoracic pathologists. Imaging studies were reviewed by a thoracic radiologist.
Eight patients (7 men, 87.5%; median age, 79 years; range, 69–96 years) died within a median of 17 days (range, 6–100 days) from onset of symptoms. The median lung weight was 1220 g (range, 960–1760 g); consolidations were found in 5 patients (62.5%) and gross thromboemboli were noted in 1 patient (12.5%). Histologically, all patients had acute bronchopneumonia; 6 patients (75%) also had diffuse alveolar damage. Two patients (25%) had aspiration pneumonia in addition. Thromboemboli, usually scattered and rare, were identified in 5 patients (62.5%) in small vessels and in 2 of these patients also in pulmonary arteries. Four patients (50%) had perivascular chronic inflammation. Postmortem bacterial lung cultures were positive in 4 patients (50%). Imaging studies (available in 4 patients) were typical (n = 2, 50%), indeterminate (n = 1, 25%), or negative (n = 1, 25%) for COVID-19 infection.
Our study shows that patients infected with COVID-19 not only have diffuse alveolar damage but also commonly have acute bronchopneumonia and aspiration pneumonia. These findings are important for management of these patients.
Acute lung injury appears to be the most serious complication of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and the subsequent clinical disease, coronavirus disease 2019 (COVID-19), resulting in respiratory failure and death in a subset of patients. Epidemiologic studies show that 3.4% of COVID-19 patients develop acute respiratory distress syndrome (ARDS).1 However, ARDS was found in 72% to 93% of COVID-19 patients who died.2,3 In addition, many of these patients have thromboembolic complications. Although it has been shown that only 0.1% of all COVID-19 patients develop disseminated intravascular coagulopathy,1 this presentation was again much more common (8%) in COVID-19 patients who died.2 Moreover, COVID-19 patients with ARDS had thrombotic complications (11.7%) at more than double the rate of patients with ARDS of other causes (4.8%), commonly in the form of pulmonary thromboembolisms.4
Histopathologic studies have confirmed the increased incidence of diffuse alveolar damage (DAD), the histomorphologic correlate of ARDS, and thromboembolic complications in COVID-19 patients. A few case series and reports of postmortem studies have identified DAD in 67% to 100% of COVID-19 patients.5–12 In addition, thromboembolic findings were commonly observed, with one study5 reporting capillary thrombi in all cases, arterial thrombi in 57%, and venous thrombi in 71%. Another study reported fibrin thrombi in small pulmonary arteries in 87% of cases.6 Only occasional cases have been described to also show acute bronchopneumonia on postmortem examination.9,11,13 Furthermore, the long-term consequences for patients who survive COVID-19 infection of the lungs are still unknown.14
Herein, we report the postmortem macroscopic, microscopic, and microbiological pulmonary findings in 8 patients who tested positive for COVID-19 antemortem. In addition, we reviewed the imaging studies that were available at the time of COVID testing and thereafter.
MATERIALS AND METHODS
All autopsies performed at our institution (March 2020–May 2020) on patients who tested positive for SARS-CoV-2 by antemortem nasopharyngeal or oropharyngeal swab were included, with detailed examination of the lungs. All patients underwent another nasopharyngeal or oropharyngeal swab, and, if indicated, a serologic test for immunoglobulin (Ig) G SARS-CoV-2 antibodies at time of autopsy. Viral swabs at autopsy were analyzed for SARS-CoV-2 based on real-time polymerase chain reaction technique as previously described.15 In a single case, postmortem serum was also analyzed for SARS-CoV-2 IgG antibodies by enzyme-linked immunosorbent assay as previously described.16 Clinical information was abstracted from medical records. Immediate and underlying causes of death were recorded from the death certificate and autopsy reports.
Legal next of kin consented for all autopsies, providing permissions that allowed for research and education. Because of the postmortem nature of the cases, the study was exempted by the Mayo Clinic Rochester (Rochester, Minnesota) Institutional Review Board.
At autopsy, the lungs were removed from the heart and the trachea and were individually weighed. Based on gross findings and medical history, lung cultures were performed at the discretion of the autopsy physician in a subset of patients. For lung cultures, an approximately 1-cm3 piece of tissue was placed in a sterile container and submitted for bacterial and fungal cultures. Cultures were taken from areas of gross consolidation if present. Subsequently, both lungs were individually perfused with formalin. After about 30 minutes to 1 hour, the lungs were serially sectioned, parasagitally, at 1-cm thickness. Multiple (at least 5) sections were taken for microscopy from both lungs, including areas of consolidation and grossly fibrotic areas.
Any available imaging studies of the chest from between COVID-19 testing and death were reviewed by a thoracic radiologist (T.F.J.) blinded to clinical information and histomorphologic findings. Patterns on computed tomography imaging were categorized as typical, indeterminate, or negative related to COVID-19 based on the recent expert consensus statement by the Radiological Society of North America.17 Chest radiographs were evaluated using a similar approach.
Gross findings of lung specimens were abstracted from the autopsy report. All lung specimens were morphologically reviewed by 4 thoracic pathologists (A.C.R., M.C.B., M.C.A., and J.M.B.) independently and blinded to radiologic findings. Ancillary testing including Grocott methenamine silver, Verhoeff–van Gieson, sulfated Alcian blue, and Congo red stains was performed in selected cases based on morphologic findings. Consensus morphologic features were recorded.
To summarize numerical data, a median was calculated and the data range was provided.
Demographics and Clinical Findings of Patients With COVID-19 Infection
Eight patients were included in the study. These patients tested positive for COVID-19 by nasopharyngeal or oropharyngeal swab, died within less than 1 to 10 weeks after diagnosis, and had a complete autopsy. Demographics of the study population are summarized in Table 1. The majority of patients were men (7 of 8; 87.5%). All patients were older than 69 years, with a median age of 79 years. The majority of patients (5 of 8; 62.5%) were former smokers; the remainder were never smokers. The time between onset of symptoms and death was less than 4 weeks in most patients (6 of 8; 75%). In 1 patient (12.5%), symptoms began 100 days before death. The exact date of onset of symptoms was unknown in another patient. Three patients (37.5%) were intubated prior to death for a median of 10 days (range, 7–12 days). Four of 5 patients (80%) received thrombolytics within a week prior to death, including heparin (n = 2, 50%), aspirin (n = 1, 25%), and aspirin followed by heparin (n = 1, 25%); that information was unknown in 3 patients.
All patients died from complications of COVID-19 infection (Table 1).
Six of 8 patients (75%) tested positive for COVID-19 by nasopharyngeal or oropharyngeal swab at autopsy. For 1 patient (12.5%), the postmortem nasopharyngeal swab was negative for COVID-19 but the SARS-CoV-2 IgG serum enzyme-linked immunosorbent assay test was positive. In another patient (12.5%), the postmortem oropharyngeal and bronchial swabs were negative.
Most of the patients had other significant diseases, most commonly cardiovascular disease (5 of 8; 62.5%) and/or advanced dementia (4 of 8; 50%) (Table 1). One patient was obese and 1 patient had diabetes mellitus type II. One patient had a history of an esophageal stricture that had been dilatated multiple times. At autopsy a slight stricture of the esophagus was noted in that patient. Two patients had a known underlying lung disease in the form of fibrosing interstitial lung disease that was most consistent with combined pulmonary fibrosis and emphysema at autopsy and chronic obstructive pulmonary disease, respectively. Another patient was clinically thought to have a fibrosing interstitial lung disease possibly due to prior chemotherapy for diffuse large B-cell lymphoma; however, extensive sampling of the postmortem lungs revealed only very focal fibrosis that could not be further classified, in part because of the exuberant acute lung injury. A fourth patient was found to have pulmonary arterial changes at autopsy that were suggestive of pulmonary arterial hypertension.
Gross Findings in Patients With COVID-19 Infection
The gross findings of all lungs are summarized in Table 2. With a median combined right and left lung weight of 1220 g (range, 960–1760 g), all lungs were heavier than the expected weight of 650 to 850 g.18 Consolidation was found in the majority of lungs (5 of 8; 62.5%), although, in general, it was patchy and did not involve the entire lung (Figure 1, A). Thromboemboli were noted only grossly in a single case in distal pulmonary arteries (Figure 1, B and C). In 2 cases, there appeared to be an underlying fibrosing process. In half of the patients, pleural effusions were noted, which were usually serous.
Histomorphologic Findings in Patients With COVID-19 Infections
A median of 9.5 (range, 5–15) sections were taken and microscopically reviewed. The histomorphologic findings are summarized in Table 2 and in more detail in Supplemental Table 1 (see supplemental digital content at https://meridian.allenpress.com/aplm in the January 2021 table of contents). All patients had an acute bronchopneumonia either superimposed on DAD (6 of 8 cases; 75%) or as sole finding (2 cases; 25%) (Figure 2, A and B). Diffuse alveolar damage was identified as acute (ie, hyaline membranes present; n = 3; Figure 3, A through C), organizing (ie, interstitial fibroblast proliferation; n = 1), or acute and organizing (ie, hyaline membranes and interstitial fibroblast proliferation; n = 2; Figure 3, D and E). In 2 patients (25%), at least some of the acute lung injury was attributable to aspiration pneumonia (Figure 3, F). Both of these patients suffered from advanced dementia; one also had recurrent esophageal strictures. In one case, organizing pneumonia was focally associated with early fibrosis.
Thromboemboli were seen in 5 patients (62.5%) in small vessels, with pulmonary arterial involvement in 2 of these patients (Figure 4, A and B). However, in most patients, these fibrin thromboemboli were scattered and rare rather than a diffuse finding (Figure 4, C). Only 1 patient (12.5%) had thromboemboli identified diffusely in pulmonary arteries of all calibers, and also in pulmonary veins. In this patient, gross examination also revealed pulmonary arterial thromboemboli. Scattered thromboemboli were identified in small vessels in the submucosa of the trachea (Figure 4, D) and bronchi of another patient; however, the pulmonary interstitium showed only scattered fibrin thromboemboli in small vessels.
In 4 cases (50%) there was perivascular chronic inflammation (Figure 5, A and B). Chronic inflammation was also observed around large airways (3 patients; 37.5%), small airways (2 patients; 25%), or around both large and small airways (2 patients; 25%) (Figure 5, C and D).
Postmortem lung cultures were performed in 6 cases and are detailed in Table 2. In 3 of 6 cases (50%), the culture results indicated definite bacterial infection (4+), mostly due to Staphylococcus aureus, whereas cultures in the 3 other cases (50%) resulted in inconclusive findings because of possible contamination.
In addition, the lung parenchyma showed alveolar septal amyloid in a patient who also was found to have cardiac amyloidosis of ATTR (transthyretin) type. In one patient an underlying interstitial fibrosing lung disease was morphologically most consistent with combined pulmonary fibrosis and emphysema. Additional morphologic findings are detailed in Supplemental Table 1.
Radiologic Findings in Patients With COVID-19 Infection
Imaging studies performed between the day of positive COVID-19 testing and death were available in only half (4 of 8) of the patients and mostly included chest radiographs. All imaging findings are summarized in Table 3. Two cases had imaging findings that were regarded as typical for COVID-19 infection. The computed tomography case showed extensive mixed ground-glass opacities and consolidation involving both lungs essentially throughout the left lung and peripherally in the right upper lobe. The chest radiograph case showed patchy peripheral and basilar-predominant airspace opacities with subsequent development of consolidation within areas of airspace opacities (Figure 6, A and B). In another case the pattern was indeterminate. In that case, although there were airspace opacities in mid and lower lungs, these were superimposed on interstitial fibrosis. This case did show acute bronchopneumonia and DAD on microscopic examination. A fourth case did not show any abnormalities on chest radiograph. Gross examination of the lungs of that case also did not show any consolidation. Microscopy revealed focal acute bronchopneumonia. This patient had underlying atherosclerotic and hypertensive cardiovascular disease, which might have contributed to his death.
Our study of postmortem lungs from patients who died after testing positive for SARS-CoV-2 by nasopharyngeal or oropharyngeal swab found that all lungs had morphologic findings of an acute lung injury. However, the extent and spectrum of morphologic patterns of acute lung injury were quite variable. All cases showed acute bronchopneumonia and most cases also showed a pattern of DAD. In 2 cases, at least some of the acute lung injury was attributable to aspiration pneumonia. That was further highlighted by the results of lung cultures that confirmed bacterial infections in at least half of these cases, most of which grew S aureus. Not surprisingly, acute bronchopneumonia was considered the immediate cause of death in 87.5% of these cases. Interestingly, in contrast to other reports, thromboembolism was not a uniform finding in our study population, and, if present, was usually not a diffuse or marked feature. In cases with available imaging studies, radiologic and histopathologic findings appeared to correlate in most cases. In 1 case, a chest radiograph 7 days before death did not reveal any parenchymal abnormalities, whereas the lungs did show focal acute bronchopneumonia at autopsy. This discrepancy might be due to the interval of 7 days between the imaging study and the death of the patient, in addition to the acute bronchopneumonia being a focal finding.
Although our results confirmed earlier reports of DAD being a common finding in postmortem lungs of COVID-19–infected patients,5–7 the high percentage of acute bronchopneumonia that we observed has not been previously described. Only occasional cases of acute bronchopneumonia at autopsy have been reported, including 25% to 33% in small autopsy series9,11 and a single case report.13 However, the common finding of acute bronchopneumonia is not particularly surprising, as our patients were elderly, with a median age of 79 years. Moreover, many of our patients had advanced dementia and lived in an assisted-care facility, features that predisposed them to a higher likelihood of acute bronchopneumonia and aspiration pneumonia. Furthermore, one of our patients had a slight esophageal stricture and another patient was obese, features that might also have contributed to aspiration. Indeed, the acute lung injury in 2 of our patients was, at least in part, attributed to aspiration pneumonia.
There are myriad potential explanations for the common finding of acute bronchopneumonia. One hypothesis for the high percentage of acute bronchopneumonia in these patients might be that COVID-19 infection of the lungs increases susceptibility to acute bacterial superinfection, possibly by further weakening the immune system and altering the defense mechanisms of the respiratory tract. Alternatively, it could be that an occult acute bacterial bronchopneumonia and/or aspiration pneumonia preceded SARS-CoV-2 infection in at least some of our patients, which then led to an exacerbation of a preexisting acute lung injury. It was surprising to us that other studies5,6 did not find higher numbers of acute bronchopneumonia even though their populations comprised patients of similar ages, with mean and median ages of 69 and 78 years (range, 32–96 years). Furthermore, although in all studies5,6,8 patients were found to have various comorbidities, including hypertension (58%–100%), cardiovascular disorders (35%), diabetes mellitus (29%–43%), immunosuppression (14%), and mild chronic obstructive pulmonary disorders (10%), none of the studies reported dementia in any patients. One reported case13 of acute bronchopneumonia at autopsy in the context of COVID-19 infection was a 42-year-old man with myotonic dystrophy who also had aspiration pneumonia.
A third possibility might center on increased time between first symptoms of COVID-19 infection and death. However, in our study, the median time between onset of symptoms and death was 17 days, ranging from 6 to 100 days, which was only slightly longer than in the study by Carsana et al6 (mean time, 16 days; range, 5–31 days) and longer than in the study by Fox et al8 (median, 10.5 days; range, 1–32 days).
A fourth hypothesis could be that the COVID-19 virus is eliciting acute bronchopneumonia. Although acute bronchopneumonia is usually caused by bacterial infections, it might be that this particular virus elicits an acute bronchopneumonia pattern, especially in cases that are negative by culture. Similarly, a combination of DAD and acute bronchopneumonia (with and without positive cultures) has been described in postmortem lungs from patients who died because of H1N1 influenza.19 However, negative cultures in cases of acute bronchopneumonia might also be due to sampling, as cultures were taken from the periphery of the lungs before they were perfused with formalin and sectioned.
In the study by Ackermann et al,5 all lung specimens from the COVID-19 group had DAD with necrosis of alveolar lining cells, type II pneumocyte hyperplasia, and linear intra-alveolar fibrin deposition (ie, hyaline membranes). Similarly, in the study of Carsana et al,6 87% of lungs had hyaline membranes. Although the majority of lungs examined in our study (6 of 8; 75%) also showed DAD, in 1 case DAD was exclusively of the organizing phase without hyaline membranes, and 2 other cases did not show definite findings of DAD. In 3 cases in our study there was a combination of acute and organizing DAD.
Evidence suggests that SARS-CoV-2 infections are associated with frequent activation of the coagulation system, and COVID-19 disease has been associated with high rates of venous thromboembolism and particularly acute pulmonary embolism.20–22 Moreover, autopsy studies have proposed that the inflammatory process in the microcirculation of the lung may cause in situ immunothrombi, providing an alternative explanation to the conventional thromboembolic mechanism of pulmonary embolism. One study5 suggested that lungs from COVID-19 patients show severe endothelial injury associated with intracellular virus and disrupted cell membranes, and that alveolar capillary microthrombi are 9 times as prevalent in COVID-19 patients as in patients with influenza. In addition, it has been shown that lungs from COVID-19 patients have a larger amount of new vessel growth, predominantly through a mechanism of intussusceptive angiogenesis, when compared with patients with influenza.5 Another study20 suggested that the phenotype of pulmonary embolism in COVID-19–infected patients might differ from the phenotype in non–COVID-19 patients. Indeed, imaging studies revealed that pulmonary embolism in COVID-19 patients predominantly affected segmental (70%) and subsegmental pulmonary arteries (13%), in contrast to pulmonary embolism in non–COVID-19 patients, in whom embolism to the main/lobar pulmonary arteries (38%) and segmental pulmonary arteries (41%) more commonly occured.20 In addition, D-dimer was reported to be higher in COVID-19 patients (median, 7551 ng/mL) when compared with non–COVID-19 patients (median, 2637 ng/mL).20 Although the reason for these findings is not entirely clear, the authors20 suggested that COVID-19–associated pulmonary embolism more likely represents a combination of thromboembolic disease and in situ immunothrombosis. Supporting this literature, our patient who was found to have macroscopic pulmonary embolism demonstrated thromboemboli in distal pulmonary arteries.
Previous reports5,6 showed that most, if not all, autopsy lungs of COVID-19 patients (87%–100%) had fibrin thrombi within alveolar capillaries. Pulmonary embolism and/or microthrombi were identified in at least 58% of patients in another autopsy study9 of COVID-19–infected patients, although it was stated that microthrombi were regularly found in capillaries of the lungs and 33.3% of the cases had pulmonary embolism in the setting of deep vein thrombosis. Moreover, in the study by Ackermann et al,5 57% of cases had pulmonary arteries with a diameter of 1 to 2 mm that harbored thrombi without complete luminal obstruction. Our experience was somewhat different. We encountered only a single case with rather diffuse pulmonary arterial thromboemboli in variously sized pulmonary arteries. Another case showed a single fibrinous pulmonary arterial thromboembolus. Although we did identify fibrin thromboemboli in small vessels, this finding was apparent in only 62.5% of the cases and was in general only a focal finding, not diffuse.
Arguably, fibrin thromboemboli are challenging to identify even under the microscope. However, in our study, 4 pulmonary pathologists independently evaluated all slides with special emphasis on these previously described findings. Therefore, failure to recognize fibrin thromboemboli might be unlikely to explain the lower number of cases with capillary thromboemboli in our study. Furthermore, it is very difficult to determine the significance of small vascular thromboemboli in the setting of DAD. It is well known that thromboemboli in the small pulmonary vasculature frequently occur in the setting of DAD because of any number of etiologies, presumably because of in situ thrombosis secondary to endothelial damage and possible hypercoagulable microenvironment in the setting of acute lung injury.23 Therefore, it is difficult to attribute any special COVID-19–related significance to small vessel thromboemboli in the lung when DAD pattern is present.
In addition, we noted a recurrent and often rather prominent chronic inflammation around airways, including bronchi and bronchioles. Indeed, chronic inflammation around airways was identified in 87.5% of cases. Chronic inflammation around airways has been described in other postmortem histopathologic reports of lungs of COVID-19 patients; however, its prevalence is difficult to estimate based on the available studies. For instance, Barton et al13 reported mild chronic inflammation within bronchi and bronchioles in 1 of 2 autopsy cases. Carsana et al6 mentioned mild transmural lymphocytic and monocytic infiltrates of main bronchi and bronchioles; however, it is not clear how common this finding was. Fox et al8 described an inflammatory cell infiltrate composed of a mixture of CD4-positive and CD8-positive lymphocytes around larger bronchioles in 2 of 10 patients with early alveolar damage. In the study by Wichmann et al,9 chronic bronchitis was described as a histologic finding in 2 of 12 patients; however, the authors also stated that lymphocytic infiltration of the bronchi was often visible as a preexisting condition. Other autopsy reports10,24 of COVID-19–infected patients do not comment on chronic inflammation of airways. The etiology of the chronic airways-centered inflammation is not clear. Although the transmission of SARS-CoV-2 is not fully understood, it is thought that the virus is predominantly transmitted by droplet inhalation or direct contact and that the virus can be inhaled and exhaled from the lungs.25,26 Therefore, the chronic inflammation around airways might be elicited by the virus itself or might be the result of an immunologic reaction to other changes elicited by the virus or the aerosol droplets that contain the virus.
Similar to the descriptions by Ackermann et al5 and Konopka et al,11 we found patchy chronic inflammation around small vessels in 50% of our cases, which were identified as T-cell inflammation by Ackermann et al.5 Perivascular inflammation has been identified in previous studies in association with other viral infections such as cytomegalovirus or fungal infection such as Pneumocystis jirovecii,27 although the exact mechanism for that finding is not entirely clear.
Limitations of our study included the small number of cases and the lack of virus-confirmatory studies such as ultrastructural, immunohistochemical, or molecular studies. These studies are still under development in our institution and therefore were not available to us at the time of this study. Furthermore, detailed imaging studies were lacking from many of our patients. For instance, except for a single computed tomography scan that was done as chest computed tomography with pulmonary embolism protocol, all other available imaging was chest radiographs.
In conclusion, our study shows that patients infected with COVID-19 might die because of respiratory failure due to conditions other than DAD. These conditions include acute bronchopneumonia and aspiration pneumonia, and possibly pulmonary embolism. These findings are important for treatment and management of these patients.
The authors have no relevant financial interest in the products or companies described in this article.
Supplemental digital content is available for this article at https://meridian.allenpress.com/aplm in the January 2021 table of contents.