COVID-19-Associated Liver Injury Is a Key Determinant of Patient's Survival in a Low Endemicity Area from Central Italy

Introduction

SARS-CoV-2 was recently identified as a novel highly contagious and pathogenic coronavirus causing severe pneumonia. Since December 2019, the viral infection rapidly spread over the world, with more than 12 million of infected cases and over 560,000 related deaths reported to date [1]. The coronavirus disease 2019 (COVID-19) primarily involves the lungs, causing typical respiratory symptoms and signs. However, gastrointestinal symptoms and liver injury (LI) also occur during the course of the disease. Diarrhea has been reported as the most common digestive symptom, more often occurring within the first week of illness [2], and many COVID-19 patients presented abnormal liver enzymes, a feature possibly related to the disease severity [1]. The mechanisms of gastrointestinal and liver involvement by SARS-CoV-2 likely imply a primary role of angiotensin- converting enzyme 2 (ACE2) cell entry receptors, which are used by the virus to infect different epithelial cells. ACE2 receptors are in fact widely expressed in human small intestinal enterocytes [4], as in bile duct epithelial cells, but not in liver cells [5, 6]. This suggests that ACE2 receptors are not directly involved in the development of SARS-CoV- 2-induced hepatic damage, which rather may conceivably be the result of a systemic cytokine-mediated inflammatory response.

Since March 2019, several hospital-based studies have reported the clinical characteristics of patients with COVID-19, showing different patterns of liver abnormalities, with predominant hepatocellular, cholestatic or mixed damage [7, 8, 9, 10, 11, 12]. Additionally, post-mortem pathological analyses performed in patients died from COVID-19 showed the presence of hepatocyte degeneration associated with lobular focal necrosis and neutrophil infiltration, as well as lymphocytes and monocytes infiltration in portal tracts and congestion of hepatic sinusoids, with evidence of micro- thrombosis [13, 14]. Yet, histological features of liver failure or bile duct injury have not been observed in deceased cases. In a recent study from China [15], 41% of patients with COVID-19 showed abnormal liver tests at hospital admission and 5% had liver injury, based on liver enzyme abnormalities. Patients with abnormal liver tests were older, had a higher body mass index (BMI) and were more often males and had a greater risk of developing severe COVID-19 compared to those with normal liver function. A recent Italian meta- analysis [16], including 11 observational studies with more than 2000 patients with COVID-19, reported a relatively low prevalence (3%) of chronic liver disease at baseline in patients. However, patients with severe forms of COVID-19 developed more frequently abnormalities of liver enzymes and changes in the coagulative and fibrinolytic profiles. While the clinical characteristics of LI occurring in patients with COVID-19 have been extensively reported, little attention has been given so far to the prognostic significance of COVID-19- associated LI, which therefore remains largely unexplored. It is also unclear whether LI presents with the same features in geographic areas with different rates of disease endemicity.

In this study of a cohort of SARS-CoV-2-infected patients consecutively admitted to a COVID hospital located in an area of central Italy with a cumulative incidence of 138,51 cases per 100.000 citizens [17], we aimed to investigate to what extent the onset, type and severity of virus-associated LI predicted the clinical outcomes of the disease.

Patients and Methods

This is a single-center, observational study evaluating the prevalence and the possible clinical impact of LI in patients with COVID-19 admitted to the Clinical Infectious Disease Unit of the Tor Vergata University Hospital in Rome, Italy. Patients’ clinical features and laboratory tests were prospectively collected in a dedicated dataset. In particular, serology for hepatitis B and C viruses were evaluated in all patients at entry. All therapeutic regimens were reported and analysed. Liver injury was defined as the presence of aspartate amino-transferase (AST), alanine amino- transferase (ALT), gamma-glutamyl transferase (GGT), or alkaline phosphatase greater than 1.5 times the upper limit of normal values (ULN) (34 U/l, 55 U/l, 50 U/l and 150 U/l, respectively), or by the presence of total bilirubin greater than 2.0 mg/dl. LI was assessed both at hospital admission as well as if developed during the hospital stay and was classified as cytolytic, cholestatic or mixed type depending on the predominantly observed liver abnormality. Severe LI was defined by the presence of any serum liver test greater than 3 times above ULN.

The aim of the study was to assess whether LI predicts the clinical outcomes, considering death as the primary endpoint.

The study was conducted in compliance with the 1975 Declaration of Helsinki and was approved by our local Ethics Committee. (se disponibile mettere il numero dell’approvazione)

Statistical analysis

Chi-square with Mantel-Haenszel correction or Fisher- exact test were performed to compare categorical variables. A Kolmogorov-Smirnov test for normality was preliminarily executed to define the distribution of each continuous variable; a t-Student test was then performed for normally distributed variables and a Mann-Whitney U non-parametric test for non-normally distributed variables. A backward stepwise multivariable regression model was used to assess the association between all recorded patients’ features and death. Survival was estimated using Kaplan-Meier curves and analysed by the log-rank test. A Cox regression analysis was performed to define the independent predictors of death recorded at time of hospitalization. Statistical significance was defined by a p value <0.05 in a “two-tailed” test, with a 95% confidence interval. Statistical analyses were performed using the IBM Statistical Program for Social Sciences (SPSS®) version 20.0 for Windows® (IBM Corp, Armonk, NY, USA).

Discussion

Development of LI has been already described in earlier reports on COVID-19 patients. Our study provides further evidences that LI is a very frequent event in this setting, as it was found in 57% of SARS-CoV-2-infected patients in our cohort. Comparison of the prevalence rates of LI with the current literature is, however, difficult due to the variable definitions of LI adopted in different studies [15, 18, 19]. In the present study we aimed to address for the first time whether emergence of LI is related to the outcome of COVID-19, specifically looking at the time of onset of LI, its predominant type and its severity. We found, at multivariable analysis, that LI was an independent predictor of poor outcome, as patients who developed LI had a significantly lower survival compared to those who maintained normal liver tests during the hospital stay. Interestingly, 23.8% of the whole cohort of patients already presented with LI at hospital admission and this subgroup of patients had the highest probability of death compared to all other patients. Thus, LI is an early event during the course of the disease and its prompt detection upon entering the hospital, may help identifying patients with a poorer prognosis.

The origin of LI in patients with COVID-19 is currently unknown, yet it is most likely to be multifactorial. LI might in fact be the results of direct viral infection of liver epithelial cells, or follow a strong systemic inflammatory response, or be due to pneumonia-associated hypoxia. In addition, LI may be induced by sepsis-related metabolic acidosis and/or antiviral drug toxicity. Many of these putative mechanisms may indeed coexist and together conceivably trigger an ischemic/hypoxic pathway leading to liver damage, as the likely predominant cause. This interpretation is supported by our finding that an increase of AST and of LDH, observed either at or during hospitalization, but not an increase of ALT, was significantly associated with a greater risk of death, as already observed by Lei Fang et al.20 The possibility of a predominant ischemic/hypoxic nature of LI in COVID-19 is also consistent with the known ability of the virus to cause diffuse micro-thrombosis, associated with high D-Dimer values, as also found in our study in patients with worse outcomes. On the other hand, the possibility of a direct virus- related cytopathic damage appears less likely, also considering that none of the patients with LI died of acute liver failure or presented a clinically significant liver disease. Accordingly, liver biopsies were not performed to further investigate the nature of LI, as there was no clinical justification to perform invasive examinations. However, in our study LI appeared to present in most cases with a cytolytic or mixed biochemical pattern, and only rarely with a predominantly cholestatic type. This finding differs from what reported in a cohort of Asian patients, where a predominantly cholestatic type of LI was described15, suggesting that SARS-CoV-2-related liver pathogenicity might reflect different host-virus interactions, or even patient-related genetic variability.

Our study has limitations. Our cohort is relatively small, yet has the advantage to origin from a single academic center and to include a series of consecutive patients, with a negligible number of missing data. Our facility is a university hospital located in central Italy, where the incidence of COVID-19 has been considerably lower compared to other Italian regions, such as Lombardy or other areas from other countries in which COVID-19 incidence was significantly higher.

Conclusion

LI is as an inherent finding in COVID-19, despite its development is unlikely to be clinically relevant, nor associated with liver failure. Yet, its early emergence represents an unfavorable prognostic marker of the disease and prompt detection is therefore advisable to better identify patients at greater risk of death.

Conflicts of Interest: All Authors have no conflict of interest Funding: No Authors received any financial support for the manuscript