The Spectrum of Digestive Tract Histopathologic Findings in the Setting of Severe Acute Respiratory Syndrome Coronavirus-2 Infection: What Pathologists Need to Know

Abstract

Although the novel severe acute respiratory syndrome coronavirus-2 is known primarily to affect the respiratory system, current evidence supports its capability to infect and induce gastrointestinal tract injury. Data describing the histopathologic alterations of the digestive system in patients infected with severe acute respiratory syndrome coronavirus-2 are becoming more detailed, as the number of studies is increasing and the quality of our insight into the infection and the histopathologic findings is improving. This review highlights the range of pathologic findings that could be observed in gastrointestinal specimens from patients infected with coronavirus disease 2019 and the potential underlying pathogenetic mechanisms of this disease.

In late December 2019, the novel severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), the leading cause of coronavirus disease 2019 (COVID-19), emerged in Wuhan, China, and caused the current pandemic responsible for a significant number of infected patients and deaths.¹ The major transmission route is through aerosols and droplets, which are exhaled by an infected individual. These small droplets then can be inhaled and infect the upper/lower respiratory tract or can also be swallowed and gain access into the gastrointestinal (GI) tract.² Although the clinical presentation of COVID-19 is diverse and ranges from asymptomatic infection to severe disease with multiorgan dysfunction leading to death, respiratory tract-related symptoms such as fever, cough, fatigue, and dyspnea are the most common disease-related symptoms. In addition to the respiratory system, COVID-19 shows varying degrees of digestive tract involvement. About 29% of patients complained of GI symptoms, mostly in association with respiratory symptoms; however, 4% of patients had GI symptoms alone.³ The most prevalent GI symptom, although nonspecific, is anorexia, which occurred in about 18% of patients, followed by diarrhea in 15%, nausea or vomiting in 10%, and abdominal pain in 6% of infected individuals.⁴ There is increasing evidence that the digestive system is a potential target for SARS-CoV-2 infection. Indeed, SARS-CoV-2 RNA has been detected in stool samples in 53% of studied patients and 24% of them continued to show positive results for nearly 5 weeks, even after viral clearance from the respiratory tract.^5,6 Knowledge about the histopathological findings and alterations of GI organs caused by SARS-CoV-2 infection is essential to understand the pathogenesis of this novel disease. As many elective endoscopic procedures were suspended during the peak of COVID-19, pathologic evaluation of the digestive system during the pandemic has been largely limited to autopsy studies and resection specimens removed for life-threatening indications. Unfortunately, autopsy studies have several limitations related to postmortem autolytic changes and poor preservation, which prevented a thorough evaluation of the digestive system as well as assessment of viral presence through various ancillary studies.

The purpose of this article is to give a comprehensive review that pathologists need to know on the spectrum of histopathologic findings of the digestive tract seen in the context of SARS-CoV-2 infection, including those that evaluated the viral expression at the level of protein and/or RNA in GI, hepatobiliary, and pancreatic tissues, as well as studies that conducted the ultrastructural analysis. Moreover, the potential pathophysiologic mechanisms that underlie COVID-19-related GI and hepatobiliary injury are discussed in detail. This would aid practicing pathologists when signing out cases showing unusual microscopic features in the context of COVID-19, some of which may have an impact on the patient’s management. Thus, a thorough literature review was carried out through PubMed, MEDLINE, Science Direct, and Google Scholar databases for all relevant English-language full texts/abstracts using a search query constructed with the following Medical Subject Headings (MeSH) terms: (“COVID-19” or “SARS-CoV-2”) and (“gastrointestinal tract” or “enterocytes” or “stomach” or “duodenum” or “ileum” or “small intestine” or “colon” or “liver” or “gallbladder” or “pancreas”).

PATHOGENESIS OF GASTROINTESTINAL INFECTION BY SEVERE ACUTE RESPIRATORY SYNDROME CORONAVIRUS-2

For a better understanding of the histopathological alterations of COVID-19, the disease pathogenesis should be clearly elucidated. SARS-CoV-2 causes a systemic disease that may be complicated by multiorgan damage, morbidity, and mortality. The pathogenesis of GI injury in COVID-19 is probably multifactorial. The presumed mechanisms underlying this injury include both direct SARS-CoV-2 cytopathic effect and indirect systemic effects from the excessive release of inflammatory cytokines and the hypercoagulable state impairing microcirculation.² The direct viral cytopathic effect is established by the interaction of the SARS-CoV-2 structural spike glycoprotein with the angiotensin-converting enzyme-2 (ACE2) receptor, the major functional receptor of SARS-CoV-2.⁷ This interaction facilitates viral entry into target cells; however, successful viral entry depends also on the cleavage of the viral spike protein by the action of cellular serine proteases, such as furin and transmembrane protease serine 2. This cleavage is essential for the attachment of the virus to both the ACE2 receptor and the cellular membrane.⁷ The distribution of viral ACE2 receptors in different cell types of diverse tissues can reveal the viral tropism, its potential transmission routes, and its pathogenesis.⁸ The ACE2 mRNA and protein are widely distributed in human tissues, with the digestive tract showing the highest expression levels. This included expression in esophageal keratinocytes, gastric foveolar epithelium, absorptive enterocytes of the ileum, colonocytes of the rectum, cholangiocytes, pancreatic ductal epithelial cells, endocrine cells, and microvasculature of both the endocrine and exocrine compartments.^8–11 Single-cell RNA sequencing analyses in healthy livers have shown gene expression levels for ACE2 to be highest in cholangiocytes (comparable to alveolar type-2 cells), followed by sinusoidal endothelial cells and hepatocytes.⁸ High expression levels have been reported in intestinal enterocytes along with the endothelium and vascular smooth muscle cells of the colon. Indeed, data from the Human Protein Atlas portal indicated that the ACE2 mRNA expression level was highest in the brush border of intestinal enterocytes of all tissues.^7,12,13 Interestingly, ACE2 expression in the small intestine and colon is ~40 times and ~3 times, respectively, higher than that of the lungs, whereas transmembrane protease serine 2 expression in the small intestine and colon is ~2 times and ~20 times, respectively, higher than in the lungs.⁶ Moreover, the protein expression level of ACE2, as evaluated by immunohistochemical analysis, was relatively higher in the small intestine, including the duodenum, jejunum, and ileum, than that in other tissues. The cell-type–specific expression analysis of ACE2 revealed distinct expression in the small intestinal epithelial cells, including absorptive and crypt enterocytes. However, ACE2 expression was not observed in an intestinal goblet, Paneth, and neuroendocrine cells.^7,12 Despite this observation, Livanos et al¹³ demonstrated that the SARS-CoV-2-infected intestinal cells were predominantly goblet cells. ACE2 receptor has an essential role in amino acid homeostasis, innate immunity, and maintaining intestinal microbiota. This may explain the relatively high incidence of diarrhea observed in COVID-19 patients.⁷ There may be a competitive action of SARS-CoV-2 and ACE2 receptors, which may inhibit intestinal absorption of dietary tryptophan and result in alteration of the gut microbiota and susceptibility to colitis.⁷ In contrast, the distribution of ACE2 receptors may point toward the indirect mechanisms of digestive tract injury. As discussed, the ACE2 receptor is expressed in endothelial cells of various tissues; this may elicit endothelial cell infection (endotheliitis) by SARS-CoV-2, which results in microvascular injury leading to vasoconstriction and dissemination of microthrombi causing ischemic changes that are observed in various infected tissues.¹⁴ All these facts point toward the high susceptibility of the digestive tract as an important potential target for SARS-CoV-2 infection that would possibly result in distinct pathologic alterations of the infected tissues. Interestingly, the extent of organ-specific pathologic alterations would be related to the severity of COVID-19, which in turn depends on underlying comorbidities and risk factors for severe infection. The most common patient comorbidities include hypertension, diabetes, obesity, cardiovascular disease, chronic kidney disease, chronic liver disease/cirrhosis, cancer, and transplantation/immunosuppression.¹⁵

HISTOPATHOLOGICAL FINDINGS IN DIGESTIVE TRACT ORGANS

Esophagus

COVID-19-related histopathological alterations of the esophagus have only been described in a handful of cases and showed mainly nonspecific findings that are related to underlying chronic diseases.¹⁶ Endoscopic findings included features of “acute esophageal necrosis”, in the form of diffuse circumferential black discoloration of the esophagus with sharp demarcation at the Z-line.^17–19 The microscopic appearance of the lesion reveals mucosal ulceration with extensive necrosis extending to the muscularis propria.¹⁹ Infiltration of the esophageal mucosa by lymphocytes and histiocytes was noted in a few cases, some of which demonstrated SARS-CoV-2 protein and RNA in the esophageal epithelium.^5,20

Stomach

There are isolated reports of COVID-19-associated gastric mucosal alterations in the literature. The endoscopic appearances of gastric mucosa include the presence of diffuse, multifocal, punctate mucosal hemorrhage; dark-red discoloration of the mucosa; multiple easily-bleeding irregular-shaped ulcers; and rarely ischemic changes.^14,21–24 Microscopic findings ranged from nonspecific alterations, including lamina propria (LP) edema, dilated and congested capillaries, and an inflammatory infiltrate comprised of lymphocytes, plasma cells, and macrophages, to ischemic mucosal changes with scattered capillary microthrombi (Fig. 1A).^5,14 SARS-CoV-2 nucleocapsid protein and RNA were detected in the cytoplasm of gastric glandular epithelial cells (Fig. 1B).⁵ In a study conducted through minimally invasive autopsy, incidental findings of degeneration, necrosis, and shedding of the gastric mucosal epithelium were observed.²⁵

FIGURE 1: A, Gastric mucosal changes in the setting of coronavirus disease 2019 (COVID-19) showing predominantly lamina propria (LP) neutrophilic infiltrate and vascular congestion (hematoxylin-eosin, original magnification×200). B, Scattered dot-like positivity for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) spike protein in the gastric epithelium [immunohistochemistry (IHC), original magnification×200].

Small Intestine

The most frequently reported pathologic finding in small bowel resections performed in the setting of COVID-19 was ischemic enteritis. In a case series, the ileum was the most common site of involvement by ischemic injury, followed by the jejunum, then the colon.²⁶ On gross examination, the most striking feature was the presence of a distinct patchy to circumferential, tan to bright yellow serosal discoloration that was observed along the antimesenteric border in a subset of cases.^26–28 The ischemic bowel had an extremely thin and friable variably edematous wall with significant luminal dilation, serosal hemorrhage and congestion, palpable pneumatosis, occasional perforation, rare saponification, and patent mesenteric vessels.^26–30 Mucosal alterations included variable discoloration, granularity, gray-brown exudates, and ulcerations.^27,30,31 Microscopically, acute ischemic alterations were observed in most cases, represented as variable patterns of diffuse mucosal necrosis ranging from coagulative to liquefactive to hemorrhagic to purulent form of necrosis; the latter was frequently associated with pseudomembrane formation. Occasional cases demonstrated sharply demarcated patchy/multifocal mucosal necrosis. Ischemic necrosis involved the muscularis propria, either partially or transmurally, in a subset of cases. The non-necrotic mucosa showed atrophic changes, erosions, and ulcerations, with edematous villi, LP hemorrhage, and variable degrees of LP acute inflammatory infiltrate with occasional microabscess formation as well as some lymphoplasmacytic infiltrate.^{26,27,29–33}

The most striking feature of COVID-19-induced ischemic enteritis is the presence of vascular injury. Vascular alterations ranged from severe congestion to intravascular fibrin strands to more obvious fibrin microthrombi involving mucosal and submucosal small to medium-sized vessels. The vascular thrombi were associated occasionally with fibrinoid degeneration of the vessel wall and perivascular neutrophilic infiltrate reminiscent of leukocytoclastic vasculitis.^{26,27,29–31,33,34} Features of microvascular damage manifesting as endothelial cell shrinkage, leukocyte adhesion, endotheliitis, and partial destruction of the vessel wall were described in a subset of cases.³² Indeed, Varga et al³⁵ were the first to describe the presence of endotheliitis and apoptotic bodies in the submucosal vessels of an ischemic enteritis case. Infrequently, thrombosis and endotheliitis of the mesenteric vessels were seen.³³ These findings suggest that COVID-19-associated ischemic enteritis could be attributed to direct SARS-CoV-2 infection of the intestinal vascular endothelium as demonstrated by the presence of viral protein and/or RNA in a subset of cases.^16,29,31 Few ischemic enteritis cases showed features of both acute and chronic ischemia, the latter manifesting as submucosal edema and varying degrees of muscularis propria fibrosis.²⁷ Additional microscopic features described include pneumatosis cystoides intestinalis and acute serositis.^27,31,33 Cytologic changes suggestive of viral inclusions in the glandular epithelial cells were described in an isolated case of ischemic enteritis; however, this was not confirmed by ancillary testing for SARS-CoV-2.³⁴

Apart from the alterations described, the histopathological findings of COVID-19–associated ischemic enteritis resembled those of the usual ischemic bowel due to other etiologies.²⁷ Small intestinal findings outside the context of ischemic enteritis ranged from endoscopically and histologically normal-appearing bowel wall with an aberrant epithelial expression of SARS-CoV-2 to active enteritis with abnormal villous architecture, intraepithelial lymphocytosis (mostly CD4⁻, CD8⁻, T lymphocytes), increased number of goblet cells, crypt hyperplasia, and mild lymphoplasmacytic (CD4⁺, CD8⁺, T lymphocytes) infiltrate in the LP associated with scant neutrophils and striking depletion of dendritic cells.^13,36 Moreover, a weird finding observed in some critically ill patients and post–COVID-19 individuals is the presence of small intestinal erosions and ulcerations, accompanied by lymphocyte-rich LP inflammation. The submucosal blood vessels showed dilatation and thickening of the walls, associated with fibrinoid necrosis, with protrusion through the overlying mucosa (Dieulafoy lesions).^37,38 Transmission electron microscope (TEM) study on these cases revealed a significantly altered microvillous morphology, which in the ulcerated areas appeared shorter than in the nonulcerated epithelium with disorganization and relevant cytopathic effect.³⁸

Duodenal biopsies from COVID-19 patients showed variable villous blunting, intraepithelial lymphocytosis, and expression of viral protein and RNA in the epithelial cells, in an otherwise endoscopically unremarkable duodenal mucosa.² In contrast, severe duodenitis reportedly affects about 8% of critically ill COVID-19 patients, manifesting microscopically as small duodenal crypts with intracytoplasmic and intranuclear presumably viral inclusions that express SARS-CoV-2 spike protein by immunohistochemistry.³⁹ COVID-19-associated severe hemorrhagic ulcerative duodenitis has been rarely described, featuring neutrophilic and lymphoplasmacytic infiltrate in the LP, epithelial regeneration with cytoplasmic depletion on microscopic evaluation, and negative polymerase chain reaction testing. These alterations were suggested to be related to the systemic COVID-19 infection.⁴⁰

Distinct histopathological findings that would suggest a direct viral cytopathic effect were described in 2 case studies. The first described a COVID-19 patient who developed severe watery diarrhea. The endoscopic evaluation showed congestion and loss of mucosal folds with a decreased vascular pattern in the duodenum and jejunum. Biopsies from the small intestine showed uninflamed mucosa with mostly preserved villous architecture and diffuse crypt hyperplasia. There were striking surface and crypt epithelial alterations in the form of regenerative changes manifested by marked cytoplasmic depletion, nuclear enlargement, increased mitotic figures in the crypts, decreased goblet cells, and displaced endocrine cells beyond the crypt border. However, SARS-CoV-2 was not detected in these biopsies.⁴¹ The second described a COVID-19 patient who had a Mycobacterium tuberculosis coinfection and demonstrated peculiar histopathologic findings in the terminal ileum, including patchy short, wide, and blunted villi, with erosions and overlying fibrinopurulent exudate (Figs. 2A, B). There was a focal detachment of enterocytes with a striking cuboidal to rounded “hobnail” appearance, accompanied by nuclear alterations including obvious enlargement, coarse chromatin pattern, prominent nucleoli, thickening of the nuclear membrane, and occasional binucleation. These abnormal enterocytes showed striking, predominantly subnuclear intracytoplasmic vacuoles, imparting a signet ring-like appearance. There was an obvious loss of microvilli as well as a marked reduction in goblet cells (Fig. 2C). The LP showed marked edema, congestion, and some fibrin microthrombi, and the crypts were small and elongated, showing cytoplasmic depletion, and a regenerative appearance. SARS-CoV-2 spike protein was detected in the cytoplasm of the epithelial cells supporting the direct viral cytopathic effect in this particular case (Fig. 2D).⁴²

FIGURE 2: A, Normal terminal ileum mucosa for comparison with SARS-CoV-2 affected mucosa (hematoxylin-eosin, original magnification×100). B, COVID-19-affected terminal ileum mucosa featuring marked shortening and blunting of the villi, surface erosions with overlying fibrinopurulent exudate, LP edema and congestion, and regenerative-appearing crypts (hematoxylin-eosin, original magnification×100). C, Higher magnification of the terminal ileum mucosa showing striking “hobnail” appearance with the detachment of enterocytes. Scattered signet ring-like cells and loss of goblet cells as depicted here (hematoxylin-eosin, original magnification×200). D, SARS-CoV-2-infected enterocytes showing cytoplasmic immunoexpression for the viral spike protein (IHC, original magnification×200).

It is worth mentioning that in rare instances, COVID-19 might induce acute graft-versus-host disease-like changes in the intestine, characterized by prominent villous atrophy, mucosal sloughing, extensive crypt loss, increased crypts cell apoptosis, and crypt abscesses with lymphoplasmacytic infiltration in the LP.⁴³ Moreover, COVID-19 might induce a pauci-inflammatory thrombogenic vasculopathy mimicking lupus enteritis. This is characterized by platelet/fibrin-rich thrombi and marked endothelial cell injury in the microvasculature of the duodenal mucosa and submucosa associated with ischemic mucosal necrosis and extravasated fibrin in the LP. SARS-CoV-2 envelop and spike proteins were expressed in the injured mucosal and submucosal vessels in these cases.⁴⁴ SARS-CoV-2 proteins were also expressed in a subset of studied cases, some of which were seen in uninflamed bowel mucosa, with exclusive localization in the epithelial cells as a cytoplasmic dot-like staining pattern. Although this was detected mostly in ileum biopsies with diffuse distribution, it was also observed in a patchy distribution in duodenal biopsies. This is explained by the increased number of goblet cells in the ileum as compared with the duodenum.^13,36 Viral-like particles suggestive of SARS-CoV-2 were observed through the ultrastructural analysis of small intestinal specimens.¹³

Large Intestine

As in the small intestine, the most common COVID-19-associated injury of the large bowel is ischemic colitis, showing a predilection for the right colon.³¹ Gross findings include dilation of the colon with some thickening of the wall, hemorrhagic-dusky necrotic appearance of the serosal surface, petechiae, fibrinous exudates, occasional perforations, and air bubbles.^{31,36,45–47} Mucosal alterations included edema, patchy petechiae and erythema, occasional bleeding, focal loss of mucosal folds, erosions and areas of ulceration, areas of necrosis with dark-brown discoloration, and fibrinous exudates.^{14,31,32,48,49} In about 18.5% of studied cases of ischemic colitis, a colonoscopic examination was essentially normal.¹⁴ However, microscopic appearance showed features typical of acute or chronic ischemic colitis due to non–COVID-19 etiology, such as mucosal erosions and ulcerations, fibrinopurulent exudate, LP edema, hemorrhage, fibrin deposition, vascular congestion, withered atrophic crypts with cryptitis, transmural acute and chronic inflammation with serositis, mucosal to transmural ischemia, fibrin microthrombi in the mucosal and submucosal vessels with rare features of vasculitis, features suggestive of pneumatosis cystoides intestinalis, mesenteric vascular thrombi, and fibrosis of the wall in cases of chronic ischemic colitis (Fig. 3).^{14,27,29,31,32,36,45–48,50–52}

FIGURE 3: COVID-19-associated acute ischemic colitis featuring mucosal erosions, ulcerations, LP hemorrhage and fibrin deposition, and submucosal early vascular thrombus formation (arrow) (hematoxylin-eosin, original magnification×40).

SARS-CoV-2 has been detected in some cases of ischemic colitis as a cytoplasmic dot-like staining pattern in the surface and crypt epithelial cells as well as in the LP cells of both inflamed and noninflamed colonic mucosa.^{14,36,46,47,50} Features suggestive of vascular injury of the colon due to SARS-CoV-2 have been described, including prominent multifocal small and medium-sized vasculitis of the submucosal vessels with bizarre-appearing endothelial cells, luminal obliteration, early thrombi formation, and bleeding into vessel walls. These observations have been confirmed through the detection of SARS-CoV-2 in the cytoplasm of the injured endothelial cells.^49,53 On rare instances, SARS-CoV-2 might induce graft-versus-host disease-like changes of the colon analogous to those described in the small intestine, with prominent mucosal sloughing and ulceration, lymphoplasmacytic and eosinophilic infiltration, and crypt cell apoptosis.^43,54 It is worth mentioning that patients with preexisting inflammatory bowel disease would show worsening of the pathologic findings.³⁶

Uncommonly, the infected colonic mucosa showed no significant injury with preserved architecture and only scant lymphoplasmacytic and histiocytic infiltrate in the LP with the absence of a notable viral cytopathic effect. Viral protein and RNA were detected in these cases, with localization to the epithelial cells and LP inflammatory cells as well as viral-like particles seen in the cytoplasm of the epithelial cells by TEM.^5,55 In contrast, viral-like particles as well as viral RNA were demonstrated in some autopsy cases that showed essentially normal colonic mucosa.²²

Peculiar findings of the colon were described in a study of 2 patients. One of them was a colonic resection specimen that showed a localized area of mucosal erythema surrounded by recent hemorrhage on gross examination. The microscopic evaluation did not show features of ischemic colitis or increased inflammation; instead, it revealed distinctive epithelial changes throughout the mucosa. Tufts of overlapping cells with mildly enlarged hyperchromatic nuclei were present in the surface epithelium. Epithelial cells in the superficial crypts contained large cytoplasmic blebs projecting into their lumina. SARS-CoV-2 spike protein and RNA were detected in the cytologically abnormal mucosa, and viral particles were observed in cytoplasmic vacuoles and at the luminal surface by ultrastructural analysis.⁴¹ The other case was a colonic biopsy sample exhibiting prominent crypt injury without significant inflammation. The injured crypts featured a range of changes, including some crypts exhibiting dilatation with intralumenal mucin admixed with necrotic cellular debris, and other crypts showing focal epithelial cytoplasmic mucin depletion with eosinophilic cytoplasm and disorganized nuclei and necrotic luminal debris. However, SARS-CoV-2 was not detected in this case.⁴¹ Attempts to visualize SARS-CoV-2-like particles under TEM have been performed; however, care should be taken when interpreting ultrastructural findings as some of the visualized particles would represent cross-sections of the rough endoplasmic reticulum of the studied cells.⁵⁶

Liver

COVID-19-associated liver injury is principally defined as any type of liver damage occurring during the disease course and/or management of COVID-19 patients, with or without preexisting liver disease.⁵⁷ It mostly occurred in patients with the moderate-to-severe illness. Up to 76% of COVID-19 patients had liver enzyme abnormalities, with a predominantly hepatocellular pattern of injury characterized by elevated serum aminotransferase levels.^58–60 Pathogenesis of liver injury in the setting of COVID-19 is probably multifactorial, driven by several potential local and systemic pathogenetic mechanisms. Significant hepatic pathology was primarily related to systemic factors including hemodynamic and vascular alterations that resulted from COVID-19-associated coagulopathy and endotheliitis, hypoxic injury and septic shock, COVID-19-associated systemic inflammatory response syndrome, multiorgan failure, and the presence of comorbidities such as hypertension, diabetes, obesity, malignancies, and transplant history.^{15,57,58,61–65} Local factors cannot be ignored as well, as high levels of ACE2 receptors have been detected in the hepatobiliary system. These included intracellular active SARS-CoV-2 replication leading to direct viral cytopathic injury, drug-induced liver injury from potentially hepatotoxic drugs as one of the most important factors, and exacerbation of preexisting chronic liver diseases.^57,58,60,63 Indeed, ACE2 expression is upregulated in patients with underlying chronic liver disease, which may predispose them to worse outcomes, such as the development of acute-on-chronic hepatic failure.^60,63

To date, several studies have evaluated hepatic pathology in the context of SARS-CoV-2 infection, most of which were autopsy-based examinations. The liver showed nonspecific gross findings, such as hepatomegaly, congestion, dark red/ischemic appearance of the outer surface, macroscopically apparent infarction, varying degrees of fatty change, and chronic changes related to preexisting liver disease; however, no other significant gross pathology was seen.^{16,22,59,66,67} Microscopic findings were generally nonspecific, predominantly reflected preexisting chronic liver diseases, and principally showed mixed patterns of hepatic injury.²² The most frequently encountered finding of COVID-19-associated hepatic injury was steatosis. Macrovesicular steatosis was the predominant type, observed in 50% to 88% of studied cases, followed by mixed micro and macrovesicular steatosis, and less frequently microvesicular steatosis was observed in nearly 3% of cases.^{59,62,66–73} Mostly minimal/mild to rarely moderate degrees of macrovesicular steatosis, involving about 5% to 60% of hepatocytes, with a predominantly panlobular or centrilobular distribution, were observed.^{15,58,59,63–65,70,74–79} It has been postulated that the presence of steatosis in the liver of COVID-19 patients is most likely associated with underlying diabetes mellitus, obesity, and/or hyperlipidemia.^15,73 Active steatohepatitis with ballooning degeneration and Mallory-Denk bodies were quite rare.⁵⁹ Variable degrees of sinusoidal dilatation and congestion, predominantly in the centrilobular zone, is another common finding observed in the liver, although nonspecific, and most probably attributed to shock.^{15,16,22,58,59,63,65,68–71,74,75,80,81} The lobular architecture was mostly preserved. Lobular inflammation was an uncommon finding, manifesting as a mild to moderate focal lymphocytic infiltrate with a variable admixture of neutrophils, plasma cells, and histiocytes, and localization in zone 1 and 3 and adjacent to areas of necrosis.^{58,59,62,64,65,68,69,77,78,80} Uncommonly, the lobular injury could manifest as severe acute hepatitis with massive lobular necrosis and present as acute liver failure.^59,66,82 Furthermore, features of centrilobular parenchymal atrophy secondary to marked sinusoidal congestion along with some regenerative changes have been rarely described.⁶⁵ Hepatocyte alterations included ballooning degeneration, syncytial giant cell metaplasia, bi and multinucleation, glycogenated nuclei, occasional acidophil bodies/apoptotic hepatocytes, rare mitosis, and focal hepatocyte drop-out.^{15,25,59,62,64–66,74,75,77,78,80} Portal tracts were essentially normal in most cases or demonstrated only mild lymphoplasmacytic or rarely neutrophilic and eosinophilic inflammatory cell infiltrates, with no significant interface activity, and associated with mild to rarely moderate fibrosis and mild ductular reaction.^{25,58,59,62–65,70,73,74,77} Histiocytic proliferations exhibiting vacuolated cytoplasm and expressing SARS-CoV-2 protein were seen in portal tracts around bile ducts and in connective tissue areas in a subset of cases, as well as scattered ceroid-laden macrophages. Megakaryocytes were described occasionally in liver sinusoids.⁶²

Features suggestive of significant alterations of intrahepatic vasculature secondary to systemic derangement caused by SARS-CoV-2 have been described. These vascular changes cannot be overlooked, as ACE2 is highly expressed in hepatic sinusoidal and portal endothelial cells.^8,63 These vascular alterations encompassed 2 features: (1) vascular wall alterations and (2) thrombotic events. Alterations of blood vessel walls affected primarily the portal veins and sinusoids, spared the central vein branches, and were represented by phlebosclerosis (reminiscent of veno-occlusive disease), variable degrees of the portal vein and capillary endotheliitis, increased numbers of portal vein branches associated with marked luminal dilatation and wall fibrosis, and focal herniation of the lumen beyond portal tract borders into periportal hepatic parenchyma that became completely circumscribed by hepatocytes. The smooth muscle layer of portal veins was fragmented, partially lost, and occasionally infiltrated by lymphocytes, along with overlying endothelium. Immunohistochemistry for smooth muscle actin antibody highlighted numerous perivascular activated pericytes within portal vein walls and in adventitial tissue.^{59,65,66,68,69} Portal arteriolar muscular hyperplasia, hyalinosis of the vessel wall, and fibrinoid necrosis with endothelial apoptosis were present in a subset of cases.⁵⁹ Partial or complete occlusion of intrahepatic vasculature (arterial, sinusoidal, central vein, and portal vein) by platelet-rich/fibrin microthrombi have been described in about 40% of cases.^{59,61,62,65,66,68–70,72,74,79,80,83,84}

An interesting finding observed in 2 cases is the presence of pale eosinophilic ovoid sinusoidal (thrombotic) bodies that expressed CD61, supporting that these structures were rich in platelets.⁵⁹ Less frequent findings observed included cholestasis, necrosis, and granulomata. Cholestasis was either canalicular, lobular (mild and focal), or ductular in distribution, the latter occurring in the setting of sepsis.^{15,16,59,70–72,74,76,78} Bile plugs have been rarely described, whether in canaliculi at zone 3 or in small bile ducts (focally and mostly associated with sepsis).^{25,58,63,70,75,76} Necrosis was another less commonly observed finding, ranging from individual cell necrosis, to patchy or multifocal, to diffuse/massive or confluent areas of necrosis. It showed predominantly a centrilobular localization, less commonly periportal, and even rarely panlobular. Focal centrilobular hepatocyte necrosis has been attributed to hypoperfusion injury, whereas confluent or panlobular necrosis indicates ischemic injury, seen particularly in cases with underlying septic shock secondary to COVID-19 or underlying chronic liver disease.^{15,22,25,58,63–66,68,70,72,74,75,78–81}

Cholangiocytes showed rare alterations including mild nuclear pleomorphism and eosinophilic cytoplasm with densely pyknotic nuclei indicative of apoptosis; the latter finding indicates severe bile duct injury analogous to what is usually seen in acute cellular rejection after liver transplantation.^15,66,70 Kupffer cells seemed activated and hyperplastic, exhibiting abundant cytoplasm that contained some necrotic debris, and showed hemophagocytosis.^{66,69,70,72,74,77} Indeed, nodular proliferation or aggregates of Kupffer cells have been observed in about 70% of studied cases.^65,70 Granulomata were an uncommon finding, appearing either as lobular fibrin ring-like collections of histiocytes (probably related to drug injury) or as a scattered portal or periportal non-necrotizing granulomata described in about 1% of cases.^59,68 Peculiar large basophilic structures with scant or no cytoplasm were observed in the sinusoids and were thought to represent remnants of degenerated cells.⁷³

It is worth mentioning the pathologic features of “post–COVID-19 cholangiopathy”, a novel entity that occurs rarely during recovery from severe COVID-19 infection in absence of preexisting chronic liver disease and may lead to progressive liver injury with the potential need for liver transplantation. This is manifested clinically as marked liver enzyme elevations associated with jaundice and radiologic evidence of bile duct injury.^85,86 At the microscopic level, liver biopsy exhibits features that resemble “secondary sclerosing cholangitis of the critically ill patient” with prominent cholangiocytes injury accompanied by intrahepatic microangiopathy. Marked cholangiopathy reflects degenerative ischemic bile duct changes featuring striking cholangiocytes cytoplasmic vacuolization and swelling, apoptosis, necrosis of cholangiocytes, bile duct paucity, cytokeratin 7 metaplasia of periportal hepatocytes, and ductular reaction. Mild to moderate mixed portal inflammation and a moderate degree of the portal and periportal fibrosis, which is a consistent feature, with occasional onion skinning of bile ducts have been described. Microvascular alteration is another important feature of this entity, characterized by hepatic artery endothelial swelling, portal vein endophlebitis, fibrotic obliteration of the terminal hepatic vein, and sinusoidal obstruction syndrome. Hepatocellular and canalicular cholestasis, and in some cases zone 3 hepatocellular necrosis, can be seen.^85–88

SARS-CoV-2 protein showed mostly weak and patchy granular cytoplasmic immunostaining in hepatocytes and cholangiocytes as demonstrated in various studies.^22,62,74,84 Viral RNA has been identified in hepatocytes as well as in endothelial cells and vascular thrombi in half of the cases, and this was not significantly associated with transaminase elevations.^22,59,64 Moreover, SARS-CoV-2 RNA has been isolated from bile samples; although the significance of this remains unclear, this could be indirect evidence of SARS-CoV-2 cholangiocellular infection.^84,89 Direct evidence of hepatocellular injury was demonstrated through ultrastructural analysis, in form of accumulation of lipid droplets in the cytoplasm, decreased glycogen granules, and markedly swollen mitochondria with obscure cristae. Furthermore, viral-like particles were seen in the cytoplasm of injured hepatocytes, cholangiocytes, and sinusoidal endothelial lining cells. Portal histiocytes contained lipid droplets, lysosomes, dilated rough endoplasmic reticulum, and microvesicular bodies containing viral-like particles morphologically consistent with SARS-CoV-2.^62,66,76,77

Gallbladder

Pathologic alterations of the gallbladder were uncommon and mostly observed in cholecystectomy specimens. The most commonly reported pathologic diagnosis was “acute acalculous gangrenous cholecystitis”. Microscopic evaluation of these cholecystectomy specimens revealed extensive mucosal ulceration, mucosal to transmural necrosis, hemorrhage, widespread fibroblastic proliferation, transmural inflammatory cell infiltrate, vasculitis with thrombosis of medium-sized vessels, hemorrhagic infarction of the gallbladder wall, and hypertrophy of nerve fibers.^46,90–94 Spontaneous perforation of the gallbladder in the setting of COVID-19 has been described in a case of acute cholecystitis; that showed ischemic mucosal necrosis upon microscopic evaluation of the cholecystectomy specimen.⁹⁵ SARS-CoV-2 nucleocapsid protein and RNA were demonstrated in the gallbladder wall in 2 studies, one of which was for 2 COVID-19 patients who presented with features of acute acalculous cholecystitis and had a microscopically uninflamed gallbladder.^46,96

Pancreas

The pancreas is one of the organs that has not been commonly investigated in COVID-19 postmortem studies, possibly due to the vulnerability of this organ to autolytic/degenerative changes by the action of local enzymes. The included studies focused primarily on the ultrastructural analysis of pancreatic tissue as well as the detection of viral protein and/or genomic material. However, studies have concluded that both the endocrine and exocrine aspects of the pancreas as well as endothelial cells of pancreatic microvasculature are susceptible to SARS-CoV-2 infection.⁹⁷ The first autopsy study performed showed no obvious microscopic alterations in the exocrine compartment and only focal degeneration of islet cells.²⁵ Subsequent studies described varying degrees of pancreatitis, manifesting as lymphocytic infiltration, that was observed in about 25% of postmortem examinations. It ranged from focal acute pancreatitis in a grossly unremarkable pancreas, observed in 5% of cases, to chronic pancreatitis affecting 14% of cases, to a rare occurrence of hemorrhagic pancreatitis with secondary mucormycosis infection, to clinically diagnosed cases of acute necrotizing pancreatitis.^{67,68,98–100} Variable degrees of interstitial fibrosis have been described in the pancreas, attributed to activation of pancreatic stellate cells possibly secondary to SARS-CoV-2 infection and the accompanying inflammatory response.^97,101 Additional microscopic features observed in a nondiabetic individual included alterations primarily affecting the exocrine compartment such as severe replacement of the acinar cell component by adipose tissue (lipomatosis), centrilobular fibrosis with residual acinar cells surrounding ductules, scattered islets within fibrotic areas, acinar atrophy, and moderate arteriosclerosis.^9,10,101 In contrast, the pancreas of diabetic patients exhibited moderate fatty replacement and limited centrilobular fibrosis, rare dystrophic calcification of adipocytes, numerous islets, and rare microvascular thrombi.⁹ These histopathological findings were compatible with the normal range of expected changes within the exocrine compartment from aged persons and those with type-2 diabetes mellitus.⁹

Expression of SARS-CoV-2 protein and RNA ranged from focal to wide distribution throughout the pancreatic compartments.^{9–11,101–104} This would support the fact that SARS-CoV-2 infection causes a systemic illness and would not spare any organ.¹⁰ Ultrastructural evaluation of postmortem pancreatic specimens demonstrated significant cell injury in form of marked edema and disintegration of acinar epithelial cells, cell membrane damage, swelling and enlargement of intracytoplasmic organelles, vacuolization of the endoplasmic reticulum—Golgi apparatus complex of islet cells containing viral-like particles within these vacuoles, vacuolated irregularly-shaped nuclei with clear nuclear membrane, and large nucleoli. There was abundant rough endoplasmic reticulum with few secretory granules, and the cell gap was significantly widened. Intact viral-like particles were identified also in insulin-secretory granule-containing cells.^{10,11,103,104}

Ultrastructural alterations in beta-cell structure and morphology were analogous to those commonly observed in patients with type-2 diabetes, as shown by reduced numbers of mature insulin granules associated with features of beta-cell damage. Beta-cells from patients with COVID-19 who had been newly hyperglycemic were found to contain numerous circular granules and several granules with crystalline morphology, a typical feature of degeneration and hyperstimulation. Moreover, a conspicuous number of vacuoles, which may be suggestive of the presence of viral particles, were observed in beta-cells as well as in alpha cells, delta cells, endothelial cells, pericytes, and histiocytes.¹⁰⁴ Studies have concluded that pancreatic endocrine cells are susceptible to infection with SARS-CoV-2 at a higher degree than the exocrine compartment, and this infection may induce a local inflammatory response and necrotic cell death in islets, thus leading to variable degrees of metabolic dysregulation.¹⁰

CONCLUSIONS

The histopathologic findings of the digestive tract in the setting of COVID-19 are generally nonspecific and largely attributed to the indirect systemic viral effect. The most significant pathologic findings are SARS-CoV-2-induced (micro) vascular injury and ischemic changes. In addition to this, isolated reports describing peculiar pathologic findings of the intestinal epithelium and hepatic cholangiocytes that would be regarded as a direct viral cytopathic effect, have seemed. From a practical point of view, these findings could be faced by any practicing pathologist, who should be aware of them for differential diagnostic considerations with proper integration of the pathologic findings and the clinical data. Finally, these pathologic observations would contribute to our understanding of the novel COVID-19 manifestations and pathogenesis.