Histopathological impact of SARS-CoV-2 on the liver: Cellular damage and long-term complications

Abstract

Coronavirus disease 2019 (COVID-19), caused by the highly pathogenic severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), primarily impacts the respiratory tract and can lead to severe outcomes such as acute respiratory distress syndrome, multiple organ failure, and death. Despite extensive studies on the pathogenicity of SARS-CoV-2, its impact on the hepatobiliary system remains unclear. While liver injury is commonly indicated by reduced albumin and elevated bilirubin and transaminase levels, the exact source of this damage is not fully understood. Proposed mechanisms for injury include direct cytotoxicity, collateral damage from inflammation, drug-induced liver injury, and ischemia/hypoxia. However, evidence often relies on blood tests with liver enzyme abnormalities. In this comprehensive review, we focused solely on the different histopathological manifestations of liver injury in COVID-19 patients, drawing from liver biopsies, complete autopsies, and in vitro liver analyses. We present evidence of the direct impact of SARS-CoV-2 on the liver, substantiated by in vitro observations of viral entry mechanisms and the actual presence of viral particles in liver samples resulting in a variety of cellular changes, including mitochondrial swelling, endoplasmic reticulum dilatation, and hepatocyte apoptosis. Additionally, we describe the diverse liver pathology observed during COVID-19 infection, encompassing necrosis, steatosis, cholestasis, and lobular inflammation. We also discuss the emergence of long-term complications, notably COVID-19-related secondary sclerosing cholangitis. Recognizing the histopathological liver changes occurring during COVID-19 infection is pivotal for improving patient recovery and guiding decision-making.

Key Words: Liver, SARS-CoV-2, COVID-19, Angiotensin-converting enzyme 2, Histopathology, Liver biopsies, Liver autopsy, In vitro

Core Tip: Severe acute respiratory syndrome coronavirus 2 infection is linked to significant liver injury, emerging from the facilitated entry of the virus into liver cells, including cholangiocytes and endothelial cells, due to increased receptor expression. This invasion triggers critical cellular alterations such as mitochondrial swelling, endoplasmic reticulum dilation, and hepatocyte apoptosis. Confirmed by biopsy or autopsy, the presence of viral particles in liver tissues correlates with extensive histological damage, characterized by necrosis, steatosis, cholestasis, and inflammation. Such findings highlight the acute hepatic impact of coronavirus disease 2019 (COVID-19) and signal the risk of severe long-term complications, such as COVID-19-associated sclerosing cholangitis, emphasizing the profound and enduring effect of the virus on liver health.

INTRODUCTION

The coronavirus disease 2019 (COVID-19) pandemic, driven by the pathogenic severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus, has reshaped the global landscape, causing a catastrophic effect worldwide resulting in over 7 million deaths according to the World Health Organization (https://covid19.who.int/). The primary organ damaged in COVID-19 infection is the lung, but it is now known that SARS-CoV-2 can affect several sites such as the brain, kidney, heart, gastrointestinal tract, and liver. Liver injury is a regular finding in COVID-19 patients and has been reported to be mild, but there are cases of more severe abnormalities[1,2]. Patients usually present with an altered pattern in liver function tests [mild to moderate rise in alanine aminotransferase (ALT) or aspartate aminotransferase (AST) levels], hypoalbuminemia and hyperbilirubinemia[1-3]. Unfortunately, these patients are hospitalized and undergoing several interventions, including mechanical ventilation and drug administration, thus, the mechanisms behind these liver abnormalities remain unclear as in several cases it is not possible to determine if the observed damage is due to direct SARS-CoV-2 action or is attributable to indirect damage due to systemic disturbances[4].

There are several review articles speculating about possible mechanisms behind the observed abnormalities, including discussions about the hepatotoxic effects of antiviral drugs and steroids, direct cytopathic effect of SARS-CoV-2 infection, systemic immune response, cytokine storm disorder, or a combination of all of them[5-8]. However, due to the complexity of the disease and its systemic involvement, it is difficult to draw a definitive conclusion.

It is known that SARS-CoV-2 infiltrates host cells by connecting its spike glycoprotein (S protein) to angiotensin-converting enzyme 2 (ACE2) receptors. Transmembrane protease, serine 2 (TMPRSS2) plays a crucial role by facilitating the activation of the spike protein, enabling viral entry while circumventing antiviral proteins[9,10]. The high expression of ACE2 in cholangiocytes and other liver cell types, makes the liver a relevant target for SARS-CoV-2 infection[11,12]. Histologically, liver injury after SARS-CoV-2 infection is shown by diverse manifestations including inflammation, necrosis, fibrosis, and steatosis, which are further classified in severity in the literature discussed.

This review comprehensively examines the physiopathology of COVID-19 liver injury, placing a strong emphasis on histopathological evidence. We discuss, based on in vitro and in vivo evidence, the direct SARS-CoV-2 liver infection and the damage caused by the virus at the cellular and the tissue level. We offer a detailed overview of receptor expression in the different types of cells comprising the liver and its implication in infection and injury, establishing this organ as an important target of SARS-CoV-2 virus infection. In conclusion, our study establishes a robust foundation for future research on understanding the liver's response to COVID-19 infection and the resultant long-term complications.

MECHANISMS OF VIRAL ENTRY AND RECEPTOR EXPRESSION IN LIVER TISSUE

The SARS-CoV-2 virus enters host cells by binding its S protein to the ACE2 receptor[9]. Subsequently, the S protein undergoes cleavage by furin and cathepsin L[13]. With the assistance of the TMPRSS2, the S protein is activated to facilitate viral entry[10] (Figure 1).

Figure 1 Mechanisms of viral entry and receptor expression in liver tissue. The severe acute respiratory syndrome coronavirus 2 virus enters the host cell using its spike protein (S protein) which binds to the angiotensin I converting enzyme 2 receptor. The S protein gets cleaved by furin and cathepsin L, then with the help of transmembrane serine protease 2 the S protein gets activated to facilitate viral entry. These cells are transduced by the virus, which enters the cell through endocytosis, where lysosomal Phosphatidylinositol 3-phosphate 5-kinase aids in endosome formation and permits fusion of viral and host membranes. ACE2: Angiotensin-converting enzyme 2; CTSL: Cathepsin L; TMPRSS2: Transmembrane serine protease 2; SARS-CoV-2: Severe acute respiratory syndrome coronavirus 2.

The virus primarily enters cells through endocytosis, with a higher level of transduction in cells expressing ACE2. In this process, phosphatidylinositol 3-phosphate 5-kinase plays a crucial role in endosome formation[14]. The susceptibility of various organs to SARS-CoV-2 entry has been found to correlate with the expression levels of ACE2[11]. However, it has been demonstrated that TMPRSS2 provides SARS-CoV-2 with a replication advantage by enabling viral entry independent of the endosome pathway, thereby evading antiviral proteins such as interferon-induced transmembrane protein. This was corroborated using modified spike variants to infect various human cell lines, including those from intestinal and respiratory epithelium. Interestingly, even among cells expressing ACE2 to varying degrees, those lacking TMPRSS2 exhibited reduced viral entry[15].

The SARS-CoV-2 virus has a highly organized system for infecting host cells, where each component plays a critical role in its virulence. This includes not only the receptors responsible for the initial attachment to host cells, but also various proteins involved in cleaving the S protein, such as furin and cathepsin L. Notably, cathepsin L levels have been observed to increase in the circulation of COVID-19 patients, and this increase has shown a positive correlation with the progression and severity of the disease[13]. In vitro evidence further supports the crucial role of cathepsin L in enhancing viral entry[13].

The ACE2 receptor is widely distributed throughout various organs and tissues in the body, including the lungs, heart, kidneys, intestines, and the endothelial lining of blood vessels. Endothelial cells, which line the blood vessels in all organs, exhibit a high expression of ACE2, making them particularly susceptible to infection and damage by the virus[16]. This susceptibility helps explain the widespread organ damage often observed in COVID-19 patients[17]. It occurs due to endothelial dysfunction, leading to extensive microvascular impairment and disruption of vascular homeostasis[18]. This shift towards vasoconstriction results in ischemia, inflammation, a procoagulant state, and edema, aligning with our current understanding of COVID-19's pathogenesis[16]. Indeed, postmortem samples have revealed diffuse endothelial inflammation and the presence of viral inclusions within endothelial cells[16].

Human pluripotent stem cells have been utilized to create organoids representing different lineages for the assessment of ACE2 expression, SARS-CoV-2 tropism, and the response to infection across the entire organism[11].

Using this model, it was observed that endoderm-derived lineages, including pancreatic cells, exhibited ACE2 expression in both alpha and beta cells but not in delta cells. Conversely, in liver cells, ACE2 was detected in the majority of albumin positive (ALB+) hepatocytes. ACE2 expression was also observed in lineages originating from the mesoderm, such as CD31+ endothelial cells, cardiomyocytes, CD206+ macrophages, and microglia. However, ACE2 expression was found to be low in cells derived from the ectoderm, such as cortical neurons[11].

Additional evidence was gathered by utilizing liver bile duct-derived progenitor cells to create human liver ductular organoids, which confirmed the presence of ACE2 and TMPRSS2 in cholangiocytes[19].

This confirmation was further supported by two separate studies that utilized healthy liver tissue and ribonucleic acid sequencing (RNA-seq) analysis. These studies revealed the highest expression of ACE2 in cholangiocytes, exceeding even hepatocyte expression[20,21]. In fact, the ACE2 expression in cholangiocytes was found to be comparable to that observed in alveolar type 2 cells in the lungs[20]. Moreover, several other receptors, previously identified as crucial components, have been shown to be expressed across different cell types in the liver. For instance, TMPRSS2 exhibits widespread expression in cholangiocytes, hepatocytes, periportal liver sinusoidal endothelial cells, erythroid cells, and, to a lesser extent, non-inflammatory macrophages[21].

Similarly, the cleaving enzyme furin is broadly expressed in all cell types throughout the liver, with hepatocytes and cholangiocytes exhibiting the strongest expression, along with endothelial cells[21].

In patients with cirrhosis, we have previously demonstrated a notable increase in hepatic ACE2 and TMPRSS2 expression, alongside increased proinflammatory markers such as interleukin (IL)-6, IL-8, and monocyte chemoattractant protein 1 in the liver. Notably, we observed higher mRNA-level expression of both ACE2 and TMPRSS2 in patients with more advanced disease states, including decompensated cirrhosis and acute on chronic liver failure[22].

Further investigations have demonstrated that, at the protein level, liver expression of both receptors is significantly elevated in patients with non-alcoholic fatty liver disease[23]. Additionally, these patients exhibit higher circulating ACE2 levels compared to those with chronic hepatitis[23]. It is known that patients with metabolic syndrome are very susceptible to developing severe manifestations of COVID-19 infection. Indeed, obese patients with non-alcoholic steatohepatitis have increased hepatic expression of ACE2 and TMPRSS2[24]. Collectively, these findings help partially explain the severe outcomes observed in patients with chronic underlying diseases[1].

IN VITRO EVIDENCE OF SARS-COV-2 INFECTION IN LIVER CELLS

In vitro evidence strongly supports the ability of SARS-CoV-2 to infect various types of liver cells. Understanding the interactions between the virus and these liver cells is essential for uncovering the molecular mechanisms responsible for the hepatic effects of SARS-CoV-2 infection.

Cholangiocytes, the lining epithelium of the bile ducts and the gallbladder, exhibit varying levels of ACE2 expression. Among them, cholangiocytes from the gallbladder display the highest ACE2 expression within the biliary tree, presumably rendering them the most susceptible to viral infection[11,12].

Cholangiocytes play a crucial role in transporting bile acids secreted by hepatocytes into the bile ducts, vital to optimal liver function[25]. Interestingly, it has been observed that the chenodeoxycholic acid (CDCA), a bile acid, can modulate ACE2 expression through the farnesoid X receptor (FXR) signaling pathway[12]. FXR is a direct regulator of ACE2 transcription in a variety of tissues affected by COVID-19, such as the gastrointestinal system and respiratory systems[12].

Cholangiocytes from the gallbladder exposed to CDCA and subsequently infected with SARS-CoV-2 exhibit a notably high level of viral infection. Susceptibility to this infection can be reduced when FXR signaling is suppressed using compounds such as ursodeoxycholic acid (UDCA) or Z-guggulsterone. UDCA was proven to reduce ACE2 and viral infection ex vivo in experiments using human lungs and livers perfused ex situ after exposure at physiologically elevated concentrations of UDCA, though the exact mechanism remains unknown[12].

Ductal hepatic organoid cells expressing ACE2 and TMPRSS2, when inoculated with SARS-CoV-2, exhibited rapid expression of the virus nucleocapsid N and the formation of syncytia after infection[19]. Moreover, they displayed a significant increase in viral load after 24 hours. SARS-CoV-2 also suppressed the expression of certain proteins and genes critical for maintaining barrier integrity and bile acid transport in these cells[19]. Additionally, SARS-CoV-2 expression induced the upregulation of genes associated with cell death and apoptosis. These significant alterations indicate that cholangiocytes are damaged upon infection, leading to the impairment of the liver's bile acid transport mechanisms[19]. Consequently, direct cholangiocyte injury resulting from SARS-CoV-2 infection contributes to liver damage in COVID-19 patients[19]. Indeed, gamma glutamyl transferase (GGT), a biomarker for cholangiocyte injury, has been found to be elevated in hospitalized COVID-19 patients and associated with severe manifestations[26].

In vitro studies have also demonstrated that SARS-CoV-2 can damage hepatocytes. When hepatocyte organoids, derived from human pluripotent stem cells, were exposed to SARS-CoV-2, a significant percentage of viral RNA expression was observed, as confirmed through immunostaining[11]. This infection resulted in the upregulation of chemokines, including C-X-C motif chemokine ligand 1 (CXCL1), CXCL3, and CXCL5, among others. Simultaneously, it led to the downregulation of essential metabolic markers in hepatocytes, such as cytochrome P450 7A1 (CYP7A1), CYP2A6, CYP1A2, and CYP2D6. These changes suggest a metabolic shift towards an immune-like cell state during active SARS-CoV-2 infection[11].

IN VIVO EVIDENCE OF SARS-COV2 INFECTION IN THE LIVER

One of the earliest studies to provide evidence of SARS-CoV-2 hepatic infection was conducted by Wang et al[27] They examined liver samples from two deceased COVID-19 patients and demonstrated the presence of viral particles in the cytoplasm of hepatocytes. These particles closely resembled SARS-CoV-2 virions, as they exhibited an envelope with corona-like spikes[27].

Additional cytologic features, such as mitochondrial swelling, endoplasmic reticulum dilatation, shedding of microvilli, and the presence of apoptotic hepatocytes, indicated a cytopathic lesion caused by the SARS-CoV-2 virus[27]. Nevertheless, the absence of viral identification through quantitative polymerase chain reaction (PCR) in the liver biopsies and the limited sample size posed significant limitations in confirming direct viral liver injury by the SARS-CoV-2 virus[27]. Despite that, several other reports have confirmed the presence of viral liver infection[28,29], and isolated case reports provide clear confirmation of direct liver damage caused by the virus, as exemplified by the case reported by Orandi et al[30]. The patient developed acute liver failure (ALF) and was urgently listed for liver transplantation with status 1A, meaning a few days to live without a transplant, due to the severe liver damage following COVID-19 infection. No additional risk factors for ALF were identified, and extensive testing to rule out other potential causes was performed[30]. ALF prohibited the use of remdesivir, so casirivimab/imdevimab was administered. The patient showed improvement in coagulopathy and mental status. A liver biopsy was taken ten days after clinical improvement and showed an acute hepatitis pattern of injury, with severe necrosis and cholestasis, residual hepatocytes undergoing ballooning degeneration, and prominent nucleoli. They were able to detect replicating SARS-CoV-2 RNA in hepatocytes using in situ hybridization, a finding confirmed by immunostaining for the SARS-CoV-2 nucleocapsid protein[30].

There have been several other reports of COVID-19-infected patients who, while not directly demonstrating the presence of SARS-CoV-2 in the liver, have contributed significantly to our understanding of the morphological and pathological changes observed following infection. Among these changes, hepatocellular regenerative features such as mitotic figures and antigen Ki67-positive nuclei have been observed[31]. Additionally, there is evidence of vacuolar degeneration and edematous mitochondria in hepatocytes, as well as an enlargement of the endoplasmic reticulum[32]. Sweed et al[33] reported the presence of trichrome-positive intrahepatic cytoplasmic globules and ballooning degeneration, which could serve as potential histopathological clues indicative of COVID-19-induced hepatitis.

Cholestatic hepatitis secondary to SARS-CoV-2 has also been reported in patients without pre-existing liver diseases, presenting with markedly elevated total bilirubin levels and exhibiting severe histological findings[34,35]. The main pathological features were observed in the cholangiocytes and included cytoplasmic vacuolization, degenerative changes, and mitosis[34]. Additionally, evidence of viral RNA, as well as spike and nucleocapsid proteins of SARS-CoV-2, has been found in endothelial cells, Kupffer cells, and portal macrophages[29]. Interestingly, in COVID-19 patients where SARS-CoV-2 has been detected by PCR in liver tissue, liver enzymes, including ALT and AST, were significantly higher compared to those with COVID-19 but without SARS-CoV-2 detection in the liver. However, histological evidence of acute hepatitis did not demonstrate higher liver enzymes in comparison with those that did not show lobular necroinflammation[28]. While these reports do not conclusively demonstrate that the liver damage observed in COVID-19 patients is directly caused by a cytopathic injury, they clearly show that the virus is capable of infecting various types of parenchymal and non-parenchymal liver cells, subsequently leading to liver injury.

HISTOPATHOLOGICAL LIVER CHANGES OBSERVED AFTER SARS-COV2 INFECTION

Inflammation, necrosis and fibrosis in liver tissue

Liver injury has been reported to be mild in most patients following COVID-19 infection[36]. However, in patients with moderate to severe liver injury, characterized by a notable elevation of hepatic enzymes, liver involvement plays an important role in the disease course by affecting the production of ALB, acute phase reactants, and coagulation factors[30,37]. Therefore, hepatic dysfunction may contribute to the development of multisystemic manifestations of SARS-CoV-2, such as acute respiratory distress syndrome (ARDS), coagulopathy, and multiorgan failure[4].

COVID-19 is characterized by an overactivation of the immune system[38], and the liver plays a crucial role as it houses the largest reservoir of macrophage in the body[39,40]. The liver’s parenchymal and non-parenchymal cells are well-equipped to sense and initiate immune responses, facilitated by its extensive blood supply[40]. This makes the liver highly effective in recognizing pathogens and mounting immune responses. Due to this immune response, multiple pathological changes occur following COVID-19 infection, including inflammation, necrosis, fibrosis, and combinations thereof (Figure 2).

Figure 2 Direct coronavirus disease 2019 liver injury is observed through various histopathological changes. After analyzing several cohorts of liver biopsies, the most common findings associated with severe acute respiratory syndrome coronavirus 2 hepatic infection were reported and described. I: Centrilobular areas of confluent necrosis. II: The inflammatory infiltrate surrounding necrosis showed lymphoplasmacytic infiltrate with histiocytes, few eosinophils, and neutrophils. III: Periportal fibrosis and presence of both micro and macrovesicular steatosis. IV: Ballooning degeneration shown by swelling and rounding up of hepatocytes. V: Canalicular cholestasis with discrete ductular reaction, predominantly in zone 3. VI: Chronic venous congestion, endotheliitis, portal vein thrombosis and platelet-fibrin thrombi in hepatic sinusoids.

After COVID-19 infection, the inflammatory process in the liver manifests in various forms. In some reports, it appears to be an acute liver injury, which can range from mild to moderate lobular necroinflammation[28]. There have also been isolated cases with significant liver injury, leading to ALF[30,37]. However, it is more common to observe cases characterized by portal or lobular inflammation, which accounts for approximately 50% of all reported cases[28,32,41-44] (Table 1).

Table 1 Liver injury in patients with coronavirus disease 2019 infection, n (%).

Ref.	Condition	Present
Lagana et al[28]	Lobular necroinflammation	20/40 (50)
Sonzogni et al[45]	Inflammation and fibrosis	24/48 (50)
Chu et al[32]	Inflammation and necrosis	24/24 (100)
Schmit et al[43]	Centrilobular necrosis	11/13 (85)
	Lobular inflammation	8/13 (62)
Duarte-Neto et al[51]	Centrilobular/midzonal necrosis	49/75 (65)
Sweed et al[33]	Inflammation and necrosis	1/1 (100)
Vishwajeet et al[44]	Lobular inflammation	13/20 (65)
	Kupffer cell hypertrophy	17/20 (85)
Yurdaisik et al[48]	Extensive necrosis	2/7 (29)
	Patchy necrosis	4/7 (57)
Canillas et al[52]	Fibrosis	8/14 (57)
Ramos-Rincon et al[47]	Necrosis	5/39 (13)
Santana et al[49]	Necrosis	26/27 (96)
	Lobular inflammation	2/27 (7)
Pesti et al[29]	Fibrosis	78/150 (52)
	Necrosis	103/145 (71)
	Chronic inflammation	63/146 (43)

Inflammation cell types varied between neutrophils, monocytes, and lymphocytes.

Some of the lobular inflammation is characterized by CD4 Lymphocyte infiltrates, with severity ranging from mild (48%) to moderate (2%)[45]. In a case of fatal ALF associated with SARS-CoV2 infection, the predominant inflammatory infiltrate in the portal area consisted mainly of CD8 cytotoxic T cells. T cells accounted for 60% of the observed infiltrate in the liver of this patient[37].

In patients with COVID-19 and liver involvement, it is common to observe lobular inflammation characterized by neutrophilic infiltration[41,44]. Additionally, increased Kupffer cells are often present, and occasionally eosinophils have been reported[28,31,44]. In a series of 26 liver samples from COVID-19-related autopsies, three of them showed common histopathological features related to the inflammatory process and immune cell activation that led to hepatocellular regenerative changes, with the presence of mitotic figures and Ki67-positive cells[31].

Hepatobiliary damage can be observed in patients with severe presentations of COVID-19[46]. More severe histopathological liver findings in patients with COVID-19 include evidence of liver necrosis, which ranges from moderate to severe[32-43]. Necrosis predominantly affects centrilobular areas, and occasionally midzonal areas with confluent necrosis have been reported[30,32,43,47]. Necrosis is a common finding, reported in as many as 86% of cases[43,48]. Within hepatic parenchyma, multiple foci of necroinflammatory activity have been observed[33], sometimes with extensive hepatocyte necrosis or patchy necrosis[33,48-50]. In some cases, necrosis was a common alteration and thought to be due to shock[49,51]. Inflammation and necrosis are common histopathological liver findings in patients who died from severe COVID-19 complications (Table 1). However, most of this information arises from post-mortem liver biopsies conducted in critically ill patients, in whom cardio-respiratory dysfunction characteristic of severe COVID-19 cases could also account for some of the ischemia-related liver alterations.

The inflammatory infiltrate surrounding the necrotic areas consists of mononuclear cell infiltration, including lymphocytes, plasmacytic infiltrate with histiocytes; eosinophils and neutrophils have been also observed (Figure 2)[47,48].

Liver fibrosis has also been documented in patients with COVID-19 infection. It is a common phenomenon that affects different areas of the liver, but predominantly, portal or periportal fibrosis has been reported[29,31,33,44,45,52]. It is not clear whether fibrosis is the result of the severe inflammatory process of COVID-19 that affects the liver, at least during the acute phase of infection. However, most of the cases documenting fibrosis report it as mild in severity, suggesting it is not a consequence of COVID-19, but rather linked with preexisting medical conditions and comorbidities frequently found in clinically significant COVID-19 cases (Figure 2)[29,53].

What has been even more controversial, is the debate surrounding whether a pre-existing chronic liver disease is exacerbated by the acute phase of COVID-19 or not[1,3,54].While some reports provide evidence of aggravation, others assert the absence of severe acute alterations, such as extended necrosis, hemorrhage, and inflammation[29]. Overall liver injury is prominent in SARS-CoV-2 infection. Clinically, it is important to closely monitor liver chemistries during treatment and recovery, particularly when the patient has additional risk factors for hepatic dysfunction, such as the use of antibiotics and vasopressors, or the presence of ischemic or hypoxic injury following circulatory or respiratory failure.

Steatosis

Liver steatosis has been shown to be prevalent in patients with infection by the SARS-CoV-2 virus, ranging between 54% to 75% according to some case series[28,32,45]. Reports from various authors have characterized steatosis as ranging from mild (< 10%) to severe (> 10%), with distinctions made between microvesicular and macrovesicular steatosis (Table 2)[32,45].

Table 2 Liver steatosis in patients with coronavirus disease 2019 infection, n (%).

Ref.	Histological findings	Present
Lagana et al[28]	Macrovesicular steatosis	30/40 (75)
Sonzogni et al[45]	Steatosis	26/48 (54)
Chu et al[32]	Microvesicular steatosis	20/24 (83)
	Macrovesicular steatosis	5/24 (21)
Vishwajeet et al[44]	Macrovesicular steatosis	18/20 (90)
Duarte-Neto et al[51]	Steatohepatitis	3/75 (4)
	Steatosis	42/75 (56)
Schmit et al[43]	Steatosis	9/13 (69)
Sweed et al[33]	Macrovesicular steatosis	1/1 (100)
Santana et al[49]	Steatosis	17/27 (63)
Yurdaisik et al[48]	Macrovesicular steatosis	4/7 (57)
Canillas et al[52]	Steatohepatitis	10/14 (71)
	Steatosis	1/14 (7)
Ramos-Rincon et al[47]	Steatosis	12/39 (31)
Pesti et al[29]	Steatosis	93/147 (63)

Some authors refer to the severity of steatosis as mild < 10% or severe > 10%, and others as macrovesicular or microvesicular steatosis.

Macrovesicular steatosis often presents as fat droplets with panlobular distribution affecting liver zones 1, 2, and 3[28]. Steatosis tends to develop early during COVID-19 infection. For instance, mild steatosis was detected in a 15-year-old girl who underwent a liver biopsy on hospital day 15 after displaying clinical signs of liver disease[30]. However, while liver steatosis is common, steatohepatitis occurs at a lower rate[44]. For example, Duarte-Neto et al[51] reported 3 cases of steatohepatitis among 75 COVID-19 patients, while 42 patients exhibited only steatosis.

While the exact mechanisms behind hepatic steatosis remain elusive, a contentious debate persists regarding its causality. On one hand, compelling evidence suggests that elevated levels of serum IL-6, IL-10, and tumor necrosis factor-α along with subsequent inflammatory signaling, might indeed contribute to its development[55]. On the other hand, there is a prevailing belief that steatosis in COVID-19 affected livers may not be intrinsically tied to the severity of SARS-CoV-2 infection; instead, it could be construed as a secondary consequence associated with comorbidities or treatments administered during or prior to the infection[29]. What is clear is that liver steatosis is a common pathological feature in COVID-19 patients being found in 30%-90% of patients with a confirmed SARS-CoV-2 infection, as reported by multiple studies. The assessment of liver steatosis in these studies was made through histopathological findings from autopsies following COVID-19-related deaths. The studied populations were from different countries around the world and included predominantly middle-aged and elderly individuals, some of whom had at least one risk factor for developing liver steatosis. Despite liver steatosis being a common finding, the results of these studies indicate a prevalence that surpasses the reported rates in the general population worldwide (Table 2)[28,29,32,43,44,51,52,56].

Typical risk factors for hepatic steatosis are chronic alcohol use, higher body mass index, male gender, older age, longer waist circumference, and higher levels of cholesterol[57]. While unaware of most patients' status on hepatic steatosis previous to their COVID-19 infection, it can be observed in young patients without risk factors and could be a predictive factor for a faster progression of liver disease.

Autoimmune hepatitis

SARS-CoV-2 infection can trigger autoimmune responses in genetically predisposed individuals. It is widely acknowledged that COVID-19 is associated with increased serum pro-inflammatory cytokines. This cytokine storm is directly linked to disease severity and is a major factor contributing to COVID-19-related deaths. Due to the molecular resemblance between human proteins and viral components, this immune system hyperstimulation can potentially lead to the production of autoantibodies and the development of autoimmune liver diseases[58,59].

Multiple autoantibodies including antinuclear antibodies (ANA), anticardiolipin antibodies, and anti-β2-glycoprotein I antibodies are frequently present in patients with SARS-CoV-2 infection, particularly in those with COVID-19-associated pneumonia, potentially indicating a role of immune dysregulation in disease severity. Patients who tested positive for autoantibodies had a worse prognosis, with higher mortality and respiratory rates compared to those without autoantibodies[60]. ANA reactivity in COVID-19 patients has been reported in several cohorts, with frequencies ranging between 25% and 50%, and has consistently been associated with increased disease severity[60,61].

There is also evidence of a possible association between the development of autoimmune hepatitis (AIH), following SARS-CoV-2 infection. While the mechanisms are not well understood, molecular mimicry is hypothesized to play a role[62]. It has been previously reported that viral infections and certain drugs may serve as potential stimuli for the development of AIH, suggesting that these stimuli might share epitopes resembling self-antigens that disrupt self-tolerance. Additionally, it has been demonstrated that peripheral CD4+ and CD8+ T cells show reduction and hyperactivation in patients with severe SARS-CoV-2 infection[63]. In addition, a defective subpopulation of CD4+ CD25+ regulatory T cells is a well-described mechanism involved in the impaired regulation of self-antigens observed in AIH, which can be an additional factor that predisposes patients to the development of this disease following COVID-19[62]. However, it remains uncertain whether SARS-CoV-2 infection induces an impaired immune response that results in de novo AIH, or if this response unmasks a possible pre-existing latent autoimmune disease[59].

Osborn et al[64], described a possible association between the development of a type 2 AIH, diagnosed by an elevated anti-liver-kidney-microsomal antibody titer of 1:1280, following severe hepatic damage proceeded by a SARS-CoV-2 infection. In this case, a 3-year-old patient experienced ALF, characterized by extensive hepatic necrosis, lobar collapse, and substantial inflammatory infiltrate following a mild COVID-19 infection. Isolated severe liver dysfunction related to SARS-CoV-2 is rare and has been measly reported, however this pediatric case had an excellent response to high-dose steroids followed by maintenance immunosuppressive therapy with azathioprine. One year after, a follow-up biopsy revealed complete liver recovery with only mild residual periportal and portal inflammation and ductular reaction[64]. In this case, there was a slight genetic predisposition conferred by a family history of Hashimoto’s thyroiditis and type 1 diabetes mellitus in first-degree relatives. While it is not possible to confirm whether SARS-CoV-2 infection caused AIH in this case, the temporal association between the infection and the development of fulminant hepatic failure accentuates the importance of evaluating underlying causes of liver injury in patients with isolated severe hepatic dysfunction following SARS-CoV-2 infection[64].

Although AIH involves a complex interplay of genetic, immunologic, and environmental factors, SARS-CoV-2 infection and vaccination have been associated with the development of several autoimmune diseases in adults, such as AIH, myocarditis, immunological thrombocytopenic purpura, immune-mediated nephropathy, and type 1 diabetes[65-67].

With the global distribution of COVID-19 vaccines, documented case reports of COVID-19 vaccine-associated AIH-like syndromes have been published, with an estimated risk as low as 1 in 14 million. In contrast, the estimated incidence of idiopathic AIH ranges from 0.67 to 2 cases per 100000 people per year[68]. A systematic review involving 39 cases of COVID-19 vaccine-associated AIH-like syndromes reported a marked female predominance (76.9%), while most cases occurred in patients over 50 years old (61.5%). Interestingly, in this study, 64.7% of the patients had a history of autoimmune disease, liver disease, or were taking AIH-inducing drugs[69]. Most of the patients in this study population developed symptoms after receiving the first dose of the vaccine, with a median time to symptom onset of two weeks. All patients in this cohort underwent liver biopsy, which revealed interface hepatitis, centrilobular necrosis, and lymphocyte or plasma cell infiltration as the main findings. The overall prognosis was favorable after initiating steroid-based treatment, and only four deaths were observed, two of which were related to complications of liver disease[69].

The precise mechanism underlying the development of AIH following COVID-19 vaccination remains incompletely understood[69]; however, it is hypothesized that molecular mimicry, where similarities between vaccine peptides and human self-peptides result in the production of homologous self-antigens, leads to autoimmune-mediated tissue damage. Recent research has demonstrated that 21 out of 50 tissue antigens had moderate to strong reactions with the SARS-CoV-2 antibodies, suggesting that cross-reaction between SARS-CoV-2 proteins and a various tissue antigens could be responsible for the development of different vaccine-related autoimmune diseases[70]. Other theories propose that the formation of immune complexes may affect the balance between effector and regulatory T-lymphocytes, while certain molecular adjuvants added to most vaccines might trigger cell activation, leading to an exaggerated immune response[71].

Various types of vaccines can induce autoimmunity through different mechanisms. The incorporation of lipid nanoparticles and adenovirus vectors in authorized vaccines could potentially induce an amplified immune response. Moreover, mRNA vaccines can bind to pattern recognition receptors triggering multiple pro-inflammatory cascades, and T-cell and B-cell immune responses. Additionally, viral vector vaccines may activate immune responses by the involvement of toll-like receptor 9, leading to type I interferon secretion[67,72,73].

Clinicians should remain vigilant regarding the potential onset of AIH following SARS-CoV-2 infection or COVID-19 vaccination. Clinical manifestations such as jaundice, choluria, pruritus, or prolonged fatigue and anorexia, coupled with abnormal liver function test results, should prompt suspicion of this complication, particularly if there is temporal alignment with exposure to risk factors[69].

BILIARY TRACT INJURY AND SCLEROSING CHOLANGITIS AS A LONG-TERM SEQUELA OF COVID-19

Hepatocyte injury and a pattern of hepatocellular damage are observed in most specimens. Nevertheless, the elevated expression of ACE2 in cholangiocytes raises the possibility that they could potentially act as a reservoir within the liver or serve as an entry point for the virus[19,74-76]. Several reports have found histological abnormalities in the intrahepatic biliary tract. In some cases, only mild lobular cholestasis with intact interlobular bile ducts, or canalicular cholestasis has been observed[30,32,37,44,47]. Some other cases have noted cholestasis accompanied by discrete bile duct proliferation[43], or a more critical bile duct injury, including profound cholestasis, ductular reaction, and bile infarcts[77]. Severe cholestatic hepatitis with intense zone 3 hepato-canalicular cholestasis and prominent bile duct damage has also been observed[34]. Cytokeratin 7 (CK7), a marker of bile ducts, has been commonly used to detect intralobular canalicular cholestasis[29].

Sclerosing cholangitis

Recently, cases of secondary sclerosing cholangitis (SSC) have been observed in individuals who developed cholestatic liver disease as a result of cholangiocyte injury following acute COVID-19 infection. While the reports are not so common, there are some of them describing progressive cholestatic disease following severe COVID-19, which upon confirmation with biopsy or imaging techniques, have been determined to be COVID-19-SSC[78-82].

Post-COVID-19-SSC in critically ill patients rarely develops in patients treated in a critical care unit[78] from recent studies have found a significant increase in the incidence of SSC in critically ill COVID-19 patients, more than 46 times higher compared with data from before the pandemic[79].

In a study involving 1082 hospitalized patients with COVID-19 pneumonia who required invasive mechanical ventilation, it was determined that approximately one in every 43 of these invasively ventilated COVID-19 patients developed SSC, with an incidence rate of 2.3 per 100 patients (95%CI: 1.5-3.4)[79].

Some mechanisms related to hyperinflammation, cytokine release, and ischemia of the biliary ducts have been proposed as potential causes of COVID-19-SSC. However, a direct mechanism of damage mediated by SARS-CoV-2 towards cholangiocytes has not been identified.

Most COVID-19-SSC cases documented in the literature are from cohorts of critically ill patients who were diagnosed using magnetic resonance cholangiopancreatography, endoscopic retrograde cholangiography, or histopathological findings assessed by liver biopsy or autopsy.

Among these cohorts, one includes 334 patients admitted to the intensive care unit (ICU), where six patients began experiencing alterations in the intrahepatic biliary tract. Notably, these patients had no prior history of biliary tract issues, and these changes manifested during their hospitalization. The patients spent an average of 35.5 days in the ICU and all of them developed ARDS, necessitating invasive mechanical ventilation. After discharge, they were monitored for a median duration of 282 days (ranging from 89 to 452 days). Interestingly, these patients did not return to normal levels in their liver laboratory parameters, and three of them even developed portal hypertension, one of them experiencing decompensation. These complications strongly suggest the presence of COVID-19-related SSC in these patients[81].

Some unique histopathological findings were reported in a small series of three cases, which involved intrahepatic microangiopathy with superimposed injury to cholangiocytes. It was believed that these findings could represent a confluence of SSC and direct hepatic injury from SARS-CoV-2[83]. The patients were three young adults. Each patient had a prolonged hospitalization because of acute hypoxemic respiratory failure requiring mechanical ventilation while also having additional COVID-19 complications. On admission, liver chemistries ranged from normal to mildly elevated, however, on repeat parameters each patient had severe but brief aminotransferase elevations which were attributed to ischemic hepatitis by clinicians. Further on their hospitalization they developed marked cholestasis with associated jaundice that persisted even after cardiopulmonary and renal recovery. Laboratory studies, serologies and liver imaging ruled out other liver pathologies. Nonetheless, imaging of patient 1 showed intrahepatic beading with multiple short segmental strictures and intervening dilatation, patient 2 intra and extrahepatic dilatation, and patient 3 had no signs of bile duct dilatation. Percutaneous liver biopsies exhibited at least moderate portal and periportal fibrosis and extensive degenerative cholangiocyte injury, with prominent cholangiocyte vacuolization, regenerative changes, apoptosis, and necrosis of the cholangiocyte epithelial layer of terminal bile ducts and marginal ductules. Only patient 3 demonstrated metaplastic expression of CK7 in periportal hepatocytes[83].

A separate study also reported three cases presenting with severe COVID-19 and shortly after developed persistent cholestasis and chronic liver disease. They all required ICU admission, mechanical ventilation, vasopressor support, and broad-spectrum antibiotics due to secondary infections. All 3 patients had a history of chronic kidney disease type 2 diabetes mellitus and systemic arterial hypertension. In case 1, SARS-CoV-2 infection was confirmed on admission by reverse transcription PCR. During a 33-day stay in the ICU the patient experienced gastrointestinal bleeding, acute renal failure, and metabolic acidosis, necessitating hemodialysis. Liver tests showed cholestatic patterns, with persistently elevated alkaline phosphatase (ALP) levels, despite normal imaging results initially. After discharge, he developed jaundice, pruritus, and hypercholesterolemia. Imaging revealed bile duct dilatation and sludge. Endoscopic procedures were performed, showing bile duct obstruction and stenosis. Despite treatment, liver chemistry remained altered. Various liver conditions were considered but ruled out via specific tests, and hepatic biopsy was performed, which showed signs of cholestasis, inflammation, and fibrosis. In case 2, the patient experienced severe respiratory failure requiring mechanical ventilation and prone positioning. Complications included bacteremia and ventilator-associated pneumonia, leading to multiorgan dysfunction and necessitating hemodialysis, red blood cell transfusions, and vasopressor support. Liver tests showed a cholestatic pattern, with persistently elevated ALP and GGT levels despite initial negative imaging findings. After discharge, the patient developed jaundice, with further liver tests indicating cholestasis. Despite negative viral hepatitis and autoimmune disease panels, imaging revealed stenosis consistent with SSC. Endoscopic retrograde cholangiopancreatography (ERCP) identified bile duct filling defects, but liver function did not improve. In case 3, the patient initially presented with normal liver chemistry but elevated inflammatory markers. Within 72 hours of admission, she required ICU admission and mechanical ventilation due to severe hypoxemia. During her 20-day ICU stay, she underwent hemodialysis, received high positive end-expiratory pressure, norepinephrine, and antibiotics for ventilator-associated pneumonia. Liver chemistry progressively showed cholestasis, with significantly elevated bilirubin, GGT, and ALP levels. Imaging revealed intra and extrahepatic biliary dilation, and other potential causes were ruled out. Magnetic resonance cholangiography demonstrated irregular bile duct morphology without obstruction, indicating a severe cholestatic liver injury[84]. All three cases were subsequently treated with UDCA with no clinical improvement on liver chemistries.

Unfortunately, there is no effective clinical treatment for COVID-19-SSC, and liver transplantation remains the only option for selected patients[78,82]. The 1-year transplant-free survival rate of COVID-19-associated SSC has been shown to be 40%[79]. Clinicians have chosen to treat post-COVID cholangiopathy with UDCA and obeticholic acid, which halt hepatic damage by reducing the accumulation of bile acids that are not excreted through the digestive system. Both drugs have even been combined to enhance treatment potency and efficacy following unsuccessful monotherapy with UDCA. In severe cases with an unfavorable prognosis, even after drug treatment fails, clinicians have resorted to invasive procedures such as ERCP and sphincterotomy. The aim is to prevent biliary tract obstruction caused by sedimentation formed due to alterations produced by the virus in the biliary ducts[85].

All this evidence suggests that cholestatic liver disease and SSC may be long-term sequelae of COVID-19 acute illness as a longstanding manifestation of critical illness.

VASCULAR THROMBOSIS AND ENDOTHELIAL INFLAMMATION

Some of the proposed pathophysiological mechanisms of liver injury associated with COVID-19 include endothelial damage which leads to vascular changes due to coagulopathy and endotheliitis[29,86]. Coagulopathy has been described to be exacerbated by the severe inflammatory response syndrome associated with SARS-CoV2 infection[86,87]. There have been several reports providing histological evidence of vascular alterations in hepatic parenchyma. These include abnormal vessels in the portal or periportal areas[45,46]. Herniated portal veins in periportal tissue ranging from focal, to diffuse have been also reported[45]. Zone 3 sinusoidal ectasia with significant red cell congestion and portal tract vein luminal dilation have been common findings in some reports, observed in up to 84% and 72% of the cases respectively[31]. Other findings include sinusoidal diffuse platelet-fibrin microthrombi (PMT), sinusoidal erythrocyte aggregation (SEA), portal vein thrombosis, centro-acinar ischemic-type hepatic necrosis[31,51,56]. PMT was found to be associated with increased liver injury but not SEA[56]. In patients that undergo shock, liver histological evidence in some reports shows hepatic centrilobular congestion in 50%-100% of the patients, and centrilobular/midzonal necrosis causing hepatocytic loss[49,51]. CD31, CD34 and claudin-5 have been used as markers to assess the integrity or damage of endothelial cells in some of these cases (Table 3)[29].

Table 3 Vascular thrombosis and endothelial inflammation in patients with coronavirus disease 2019 infection, n (%).

Ref.	Histological findings	Present
Pesti et al[29]	Endothelial cell damage	119/119 (100)
	Sinus dilatation	119/119 (100)
	Fibrin	119/119 (100)
Duarte-Neto et al[51]	Centrilobular congestion	75/75 (100)
	Sinusoidal thrombosis	5/75 (7)
Chu et al[32]	Portal inflammation	23/24 (96)
Schmit et al[43]	Portal inflammation	12/13 (92)
Santana et al[49]	Centrilobular congestion	23/27 (85)
	Portal inflammation	14/27 (52)
Fassan et al[31]	Sinusoidal ectasia	21/25 (84)
	Sinusoidal microthrombi	5/25 (20)
	Portal vein thrombosis	3/25 (12)
Sonzogni et al[45]	Portal inflammation	32/48 (66)
Kondo et al[56]	Sinusoidal thrombosis	23/43 (53)
	SEA	19/43 (44)
	PMT	14/43 (32)
Lagana et al[28]	Portal inflammation	20/40 (50)
Vishwajeet et al[44]	Portal inflammation	10/20 (50)

PMT: Platelet microthrombi; SEA: Sinusoidal erythrocyte aggregation.

Different series of postmortem liver biopsies from patients affected by severe COVID-19 have reported liver sinusoidal microthrombosis and partial portal thrombosis in at least 50% of the patients. Data for these series are mostly obtained from autopsies of patients whose cause of death was completely related to severe COVID-19. Therefore, the population studied in these series presents the classical risk factors for COVID-19 clinical progression and mortality (older age, comorbid conditions). Although the mechanism of liver vascular thrombosis is likely multifactorial, the reported findings are rare to observe and could hardly be triggered by other comorbidities of the patients[56]. Endotheliopathy is a key component of the prothrombotic imbalance in patients with COVID-19, which is related to a worse outcome in this disease and has been reported to persist following COVID-19. Some markers of endothelial cell activation, such as soluble thrombomodulin levels and the von Willebrand Factor antigen, have been shown to correlate with mortality[88].

While these findings align with the possibility of vascular-related injury in the hepatic parenchyma[45], and considering that SARS-CoV-2 has been detected in blood clots and endothelial cells, it remains uncertain whether these phenomena stem from direct endothelial damage caused by the virus or result from the immunological and inflammatory response triggered by the virus.

CONCLUSION

In conclusion, the comprehensive exploration of liver involvement in COVID-19 presented in this article underscores the virus´s capacity to impact hepatic tissue directly and indirectly[27]. The study of viral entry mechanisms and in vitro observations allow for an improved understanding of direct liver injury and a superior analysis of tissular involvement in pathogenesis[9]. The evidence surrounding hepatic parenchyma histopathological changes such as necrosis, steatosis, cholestasis, and lobular inflammation, endothelial dysfunction-induced endotheliitis and procoagulant factors support the claim that SARS-CoV-2 infection produces liver injury through a variety of mechanisms even when other factors such as COVID-19 and vasoactive medication, invasive ventilation, and infection-induced cytokine storm, also produce significant liver injury[29,86]. This study provides evidence that will support future insight when producing new therapeutic targets to reduce morbidity and mortality due to systemic complications following COVID-19 infection.