Introduction

Coronavirus disease 2019 (COVID-19) is associated with elevated liver biochemistries in approximately half of hospitalized patients [1,2,3,4]. COVID-19 has been most frequently associated with an aspartate aminotransferase (AST)-predominant hepatocellular injury pattern, though other patterns exist [4,5,6]. Multiple potential etiologies for elevated liver biochemistries in COVID-19 have been suggested, including direct SARS-CoV-2 liver injury, drug-induced liver injury (DILI), non-liver origin, and ischemia, as well as etiologies not specific to COVID-19 such as cholestasis of sepsis and chronic liver disease [5, 7, 8]. In this context, hepatologists are consulted for diagnostic and therapeutic guidance [9]. Our aim was to assess degree of hepatologist agreement on the etiology of abnormal liver biochemistries as well as diagnostic recommendations in COVID-19.

Methods

We reviewed the characteristics of twenty inpatient hepatology consultations performed at Massachusetts General Hospital and Brigham and Women’s Hospital for abnormal liver biochemistries in adult patients with a positive SARS-CoV-2 RT-PCR. Three senior hepatologists independently reviewed the data available at the time of consultation. They provided a rank-order list of up to three possible etiologies and any additional testing they would request. Reviewers provided post-discharge recommendations after receiving the liver biochemistries at discharge. A Fleiss kappa statistic was used to calculate agreement between multiple reviewers, using SAS version 9.4 (Cary, NC). This study was deemed exempt by our institutional review board.

Results

From March 20 to April 21, 2020, our inpatient hepatology teams performed 20 consultations on patients with COVID-19 and abnormal liver biochemistries; eight had chronic liver disease, and twelve required ICU admission (Table 1 for patient characteristics).

Table 1 Patient characteristics
Full size table

Seventeen (85%) patients had at least one abnormal liver biochemistry on admission. Mean (range) AST at consultation was 343 U/L (41-3300), ALT 288 U/L (25-1591), alkaline phosphatase 223 U/L (51-976), and total bilirubin 4.0 mg/dL (0.3-19.3). At the time of hepatology consultation, 10 (50%) patients had a hepatocellular pattern of liver injury, 7 (35%) had a cholestatic pattern, and 3 (15%) had a mixed pattern, determined by the ALT to alkaline phosphatase ratio [10].

The suspected diagnosis by the original consultant and three senior hepatologists is shown in Fig. 1. The original consultant and three hepatology reviewers had Kappa agreement of 0.10 (p = 0.03) for the most likely primary etiology of elevated liver biochemistries. The two most common diagnoses were SARS-CoV-2-related liver injury and DILI. Between all four hepatologists, there was 0.17 Kappa agreement (p = 0.03) that SARS-CoV-2-related liver injury was on the differential and 0.51 Kappa agreement (p < 0.0001) that DILI was on the differential.

Fig. 1
figure 1

Etiology of elevated liver biochemistries in COVID-19. a Number of cases with each primary diagnosis by reviewer. b Global frequency of each diagnosis in the top 3 differential. Includes diagnoses made by original consultant and 3 senior hepatologists

Full size image

Of the 14 patients with at least one reviewer suspecting DILI, 9 (64%) had hepatocellular, 4 (29%) had cholestatic, and 1 (7%) had mixed liver injury patterns. The medications administered to those patients included: hydroxychloroquine in 13 (93%), cephalosporin 13 (93%), acetaminophen 13 (93%), statin 12 (86%), azithromycin 11 (79%), placebo-controlled remdesivir trial 5 (36%), lopinavir–ritonavir 1 (7%), and tocilizumab 1 (7%).

Prior to consultation, medical teams performed the following evaluation: hepatitis B virus (HBV) serologies (85%), HCV antibody (80%), HCV PCR (15%), hepatitis A virus (HAV) serologies (15%), ultrasound (15%), ultrasound with Doppler (30%), and cross-sectional abdominal imaging (5%).

The hepatology consultant recommended: liver biochemistry monitoring (100%), HBV PCR (10%), HAV serologies (10%), ultrasound (5%), ultrasound with Doppler (15%), autoimmune markers (15%), and other viral studies (20%). They did not recommend liver biopsy or cross-sectional imaging in any initial consultation. They recommended against ultrasound in 40% for lack of necessity and to minimize health care worker exposure. Hepatologist reviewers made similar diagnostic recommendations.

Seven patients were discharged, six had improved liver biochemistries, and none died at the end of the study period on April 22, 2020. For the 7 discharged patients, hepatologist reviewers recommended hepatology follow-up in 3 (43%) and liver biochemistries in 6 (86%).

Discussion

Determining the cause of elevated liver biochemistries in patients with COVID-19 is a challenge. The patients described here were a heterogeneous group by liver disease history and degree and pattern of liver biochemistry elevation. The differential diagnosis includes SARS-CoV-2-related liver injury, DILI, complications of the infection including myositis (particularly when AST > ALT), ischemia, and more. The challenge is highlighted by the fact that, without liver biopsy, there was little agreement between hepatologists regarding the most likely etiology of elevated liver biochemistries.

The hepatologists attributed approximately half of cases to phenomena specific to the COVID-19 pandemic: SARS-CoV-2-related liver injury or DILI. The remaining cases were attributed to non-COVID-specific etiologies. The greatest consensus was around DILI diagnosis. Patients with suspected DILI were on multiple potentially hepatotoxic drugs, which underscores the complexity of these consults, as well as the challenge of assessing hepatic safety of COVID-19-specific therapies.

These consultations were performed in March and April 2020, early in the US pandemic, at large academic institutions, in a city with relatively high incidence of disease, and mainly for patients with severe liver biochemistry elevations. We suspect that issues related to hepatology consultation will differ by phase of the pandemic, patient population, location, and as more literature about SARS-CoV-2 and its treatments become available. Our data are also limited by lack of a gold standard comparison (i.e., liver biopsy) or extended follow-up.

Despite lack of consensus about etiology of liver biochemistries in inpatient hepatology consultations, there was agreement that the diagnostic strategy for the majority of cases should be serial liver biochemistry monitoring, as opposed to multiple diagnostic tests, especially those that expose health care workers to SARS-CoV-2. Nevertheless, liver biopsy should be considered in select cases of severe liver injury with multiple potential causes. These comments are consistent with expert guidance, which suggests limiting transportation of patients with COVID-19 for liver-related testing unless it is likely to change management [9].