Liver Injury in patients with COVID-19: Clinical Profiles, CT findings and their Correlation with the Severity


 BACKGROUND & AIMS: Liver injury is found in some of patients with COVID-19. Liver injury of patients based on severity grading and abdominal radiological signs have not been reported until now. The aim of our study is to determine clinical profiles of the patients based on severity grading, describe abdominal radiological signs, and investigate the correlations of the severity with clinical profiles and radiological signs.METHODS: This retrospective cohort study included 115 patients with COVID-19 from Jan 2020 to Feb 2020. Medical records of the patients were collected and CT images were reviewed.RESULTS: Common clinical manifestations of the patients with COVID-19 were fever (68.70%), cough (56.52%), fatigue (31.30%); some of them had gastrointestinal symptoms (diarrhea,12.17%; nausea or vomiting 7.83%; inappetence, 7.83%). Abnormal liver function was observed in some of the patients with COVID-19. Significant differences in the levels of AST, albumin,CRP were observed among different groups classified by the severity. Common findings of upper abdominal CT scan were liver hypodensity (26.09%) and pericholecystic fat stranding (21.27%); liver hypodensity were more frequently found in critical cases (58.82%). The semi-quantitative CT score of pulmonary lesions, CT-quantified liver/spleen attenuation ratio correlated with severity grading in patients with COVID-19.CONCLUSIONS: Some of patients with COVID-19 displayed liver damage revealed by liver functional tests and upper abdominal CT imaging, and some of liver functional tests and CT signs correlate with severity grading; thus, it will allow an earlier identification of high-risk patients for early effective intervention.


Introduction
2019 novel coronavirus disease(COVID-19) is a zoonotic illness caused by a new viral pathogen named as the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), previously known as 2019 novel coronavirus(2019-nCoV) 1 .The rapid spread of COVID-19 has created a global health crisis. Lung has been considered as a major organ attacked by SARS-CoV-2. Besides the respiratory system, COVID -19 leads to various degrees of damage to other organs such as the liver, heart, kidney and gastrointestinal tract [2][3][4][5] . The majority of patients with COVID-19 only suffer mild symptoms; while some patients with COVID-19 progressed rapidly with acute respiratory distress syndrome(ARDS) and eventually followed by multiple organ failure (MOF) 1,6 .
Previous studies showed that some of patients infected by SARS-CoV-2 had gastrointestinal symptoms such as diarrhea, abdominal pain, inappetence, nausea and vomiting [7][8][9] . More importantly, liver injury occurred in patients with COVID [8][9][10] .The abnormality of liver function tests were observed in some of COVID-19 patients at admission 11 . The liver biopsy specimens of patients with COVID-19 showed that hepatocyte degeneration accompanied by focal necrosis, moderate microvascular steatosis, infiltration of leukocytes into lobular and portal area, and sinusoidal congestion 12 . Thus, these indicated liver involvement in patients infected by SARS-CoV-2. While previous studies collected the data from the whole cohort of patients without the exclusion of underlying liver diseases.
Chest CT, a routine imaging tool for pneumonia diagnosis, plays a central role in the diagnosis and classification of COVID-19 pneumonia [13][14][15] . Moreover, radiological signs of chest CT varied with the course of diseases, and the changes of radiological signs might be associate with the outcomes of diseases 15,16 .Liver injury in patients infected by SARS-CoV-2 might result in radiological change in liver. However, radiological findings of abdominal CT scan have not been explored so far.
To get a better understanding of liver injury in patients with COVID-19, 115 COVID-19 patients enrolled were classified into three groups according to the severity, clinical profiles and abdominal CT signs were described then the correlations of the severity with clinical profiles and CT signs were investigated in our study.

Study Population
Our retrospective study was performed at Cancer Center of Union Hospital and main campus of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology from Jan 24, 2020 to Feb 17,2020.178 consecutive patients with COVID 19 pneumonia were enrolled. The diagnosis of the patients with COVID 19 are: 17 fever and/or respiratory symptoms;(2) positive nucleic acid of the virus (RT-PCR analysis of throat swab specimens). A severity-grading system for patients with COVID 19 were issued by National Health Commission of the People's Republic of China and National Administration of Traditional Chinese Medicine of the People's Republic of China. The severity of the patients with COVID 19 is classified into four grades: mild, ordinary, severe and critical 13 . Mild cases: symptomatic infection include fever, cough, sore throat, fatigue, headache or myalgia, but there were no pneumonia signs on chest imaging. Ordinary cases: symptomatic infection include fever, cough, sore throat, fatigue, headache or myalgia, and pneumonia signs were showed on chest imaging, meanwhile, these patients do not have any of the severe or critical symptoms and complications described below. Severe cases, disease progresses to meet any of the following conditions:①significantly increased respiration rate: RR ≥ 30/min; ②SpO2 ≤ 93% from a finger oximeter;③PaO2/FiO2≤ 300mmHg; ④patients with >50% lesions progression within 24 to 48 h in pulmonary imaging. Critical cases, with any of the following conditions: ①respiratory failure which requires mechanical ventilation; ②shock; ③accompanied by other organ failure that needs ICU monitoring and treatment. Inclusion criteria were as follows: ①confirmed cases with COVID 19; ②pneumonia manifestation on chest CT scan; ③CT scan and laboratory examinations were performed within 2 days after admission. Exclusion criteria included: ①the patients who received antiviral treatment for more than 3 days before admission, such as interferon, chloroquine phosphate, arbidol, lopinavir/ritonavir, which may result in the alleviation of COVID 19 and side effects including GI symptoms and drug-induced liver injury, etc.(n=30).② CT images of upper abdomen were not obtained (n=19); ③the patients with hematological malignancy who was receiving chemotherapy(n=6);④the patients with severe underlying pulmonary diseases(n=4); ⑤the patients with chronic liver diseases and liver cancer(n=4) (Fig.1).This study was conducted in accordance with the Declaration of Helsinki. Informed consent was waived because of the retrospective nature of the study and the usage of anonymous clinical data.

Data collection
The pertinent data of the patients was collected from medical records and tabulated in a database. The pertinent data of the patients included demographic data; presenting symptoms and signs; medical history; laboratory tests; imaging findings; therapeutic process.

Imaging technique and imaging analysis
All CT examinations were performed with one of the following three scanners: 64-detector spiral CT scanner (Somatom Definition AS, Siemens, Germany), gemstone spectral CT (Discovery CT750 HD, GE, USA), 16-detector spiral CT scanner (Activion 16, Toshiba, Japan).Images were reconstructed into lung window and soft tissue mediastinal window with a slice thickness of 1.5mm, 2 mm or 5 mm and an interval of 1.5mm, 2 mm or 5 mm, respectively. The radiological evaluation included pulmonary involvement and abnormal signs of upper abdomen. All CT images were independently reviewed and the findings recorded by two experienced radiologists who were unaware of clinical data. For the semi-quantitative and qualitative imaging evaluation, discrepant evaluation was jointly re-evaluated to reach a final consensus [18][19][20] . Whereas intraclass correlation coefficient was introduced to assess interobserver consistency for the quantitative imaging evaluation. The severity of pneumonia was evaluated by a semi-quantitative scoring system, i.e., each of the 5 lung lobes was visually scored from 0 to 5 as: 0, no involvement; 1, <5% involvement; 2, 5%-25% involvement; 3, 26%-49% involvement; 4, 50%-75% involvement; 5, >75% involvement. The total CT score was the sum of the individual lobar scores and ranged from 0 (no involvement) to 25 (maximum involvement) 21,22 . Radiologic signs reviewed in upper abdominal CT scan including: hepatomegaly, the density of liver, periportal edema, pericholecystic fat stranding, portal lymphadenopathy, biliary ductal dilatation, ascites. Liver hypodensity was defined when the density of liver is homogeneous or heterogeneous lower than or nearly equal to that of spleen. CT-quantified hepatic density was measured by liver-to-spleen (L/S) attenuation ratio. Hepatic and splenic CT attenuation values were measured using regions of interest (ROI) >100 mm 2 in area 23 on the PACS (CARESTRAM 3.1.S12.1, Canada). Two ROIs were in the right liver lobe, one ROI in the left liver lobe and three ROIs in the spleen. ROI should include a greater area of the liver and spleen whenever possible, while exclude regions of hepatic vessels and bile ducts. Then L/S ratio was calculated by taking mean HU measurement of liver lobe ROIs and dividing it by the mean splenic HU measurement of its ROIs.

Statistical analysis
Continuous variables were expressed as mean and standard deviation for normal distribution data and median (25th-75th percentiles) if the distribution was skewed. Categorical variables were presented as count (percentage). The interobserver agreement for measurement of CT-quantified hepatic density was determined using the intraclass correlation efficient (ICC). The level of agreement was generally recognized as follows: poor, ICC<0.20; fair, 0.2<ICC≤0.40; moderate, 0.4<ICC≤0.60; good, 0.6<ICC≤0.80; and very good, 0.8<ICC≤1.0.The differences of categorical variables among groups were analyzed by chi-square test (the Fisher exact test was used when the data was limited) and continuous variables by Kruskall-Wallis H test. The correlation of categorical variables was analyzed by Spearman correlation test and continuous variables by Pearson correlation test. All statistical tests were two-sided. P value less than 0.05 was considered to indicate statistical significance. Statistical analyses were performed using SPSS version 22.0 (SPSS Inc., Chicago, Illinois, USA).

Clinical presentation
115 patients with COVID-19 pneumonia were consecutively enrolled in our study. The patients were classified into mild group, ordinary group, severe group and critical group according to the severity 13 . Mild cases without pneumonia manifestation did not receive inpatient hospital care in Wuhan, China, so mild cases were not included in our study. Thus, the enrolled patients included ordinary cases (50 patients), severe cases (48 patients) and critical cases(17 patients) (Fig.1). The median age of the enrolled patients were 66 years (IQR: 60-70), there were 58 males and 57 females, with a male-to-female ratio of 1.02:1 (Supplementary Table.1). Significant differencein patient age existed among different groups,which revealed that older patients were at higher risk for severe illness.Then, clinical manifestations of the patients were analyzed, and the results showed fever in 68.70% (79/115) of the patients, cough in 56.52%(65/115), fatigue in 31.30%(36/115), and shortness of breath in 13.91% (16/115) of the patients (Supplementary Table.1) . Thirdly, gastrointestinal symptoms were determined in the patients, and we found that 12.17% (14/115) of the cases had diarrhea,7.83% (9/115) of the cases with nausea or vomiting,7.83% (9/115) of the cases with inappetence (Supplementary Table.1). In addition, no significant differences in clinical manifestionswere evident among different groups.

Laboratory tests (peripheral blood routine examination and clinical biochemistry)
Laboratory results of the patients containing peripheral blood routine examination and clinical biochemistry were summarized in Table 1 at the time of baseline evaluation. Firstly, blood routine examination of the patients was determined. Generally, the values of erythrocyte, leukocyte and platelet were within normal limit in most of the patients (Table 1); Then, the parameters of liver functional tests were analyzed. Most of the patients had normal serum levels of AST, ALT, ALP,γGT and total bilirubin. Significant difference in AST level was observed among different groups. Liver synthetic function was also assessed by the detection of albumin and PT, and we found that 60.00% (69/115) of the cases had hypoalbuminemia,3.88% of the cases had increased prothrombin time; Thirdly, biomarkers of renal function indicated by serum urea and creatinine showed that most of the patients had normal renal function (Table 1); Finally, abnormal CRP level was detected in 64.42% of the patients. Elevated AST level, hypoalbuminemia and increased CRP level were frequently found in critical cases.

Imaging findings of chest and upper abdominal CT in COVID-19 patients
Pneumonia in COVID-19 patients was commonly manifested as ground-glass opacity, smooth or irregular interlobular septal thickening, consolidation with air bronchogram, crazy-paving pattern, and pleura effusion ( Fig.2 and 3). Most of the patients suffered pneumonia in multiple lobes. The severity of pneumonia was evaluated by a semi-quantitative scoring system and the median scores of lung involvement were 11 (IQR: 8-15) ( Table.2). More importantly, significant difference in semi-quantitative scores of lung involvement was observed among different groups (Table.2).
Then, imaging features of abdominal CT were determined. Common CT findings in upper abdominal CT scan were liver hypodensity and pericholecystic fat stranding (Table 2). Hepatomegaly, splenomegaly, periportal edema, edema of the gallbladder walls, periportal lymphadenopathy, biliary dilatation and ascites were not observed. Homogeneous or heterogeneous hepatic hypodensity was the most common CT finding ( Fig. 2 and 3). Liver hypodensity was observed in 26.09% (30/115) of the patients and statistical analysis revealed significant differencesamong different groups (Table.2).Moreover, the liver-to-spleen(L/S) attenuation ratio was measured and the interobserver agreement for assessment of L/S attenuation ratios was very good (ICC=0.98). A decrease in the L/S attenuation ratio (<1) was detected in 16.52% of the whole cohort, up to 47.06% of critical cases. Meanwhile, significant difference in the L/S attenuation ratio was observed among different groups (Table.2). Finally, pericholecystic fat stranding was evaluated. Unfortunately, 21 cases were excluded due to the patients with cholecystolithiasis or cholecystectomy. Pericholecystic fat stranding was also observed in21.27% (21/94) of the patients (Fig.3), but no significant differences were evident among different groups.

Correlation of severity grading with radiologic features and liver function tests
To evaluate the value of radiological features and liver function tests, the correlations of severity grading with CT features and liver function tests were determined. Firstly, the correlations of semi-quantitative scoring of pulmonary lesions with a panel of liver function tests and severity grading were investigated in the patients with COVID-19, and the correlation tests demonstrated that the score of pulmonary lesions associated with liver function tests, severity grading and CRP (Table.3). Secondly, the correlations of the liver-to-spleen attenuation ratio with a panel of liver function tests and severity grading, and the results revealed the L/S attenuation ratio correlated with severity grading, the score of pulmonary lesions and CRP. Thirdly, the correlation tests showed the level of CRP correlated with some of liver functional tests (AST, T-Bil, ALB, PT) and severity grading (Table.3).

COVID-19 has turned into a global pandemic disease. Previous studies demonstrated coronavirus COVID-19
(SARS-CoV-2)resulted in liver injury in the whole cohort 8-10 , we described clinical profiles and abdominal radiological features of the patients with COVID-19 according to severity grading, and then determined the correlation of radiological features with clinical profiles. Abnormal liver function was observed in some of the patients infected with COVID-19, radiological signs of the liver in non-contrast CT scan include liver hypodensity and pericholecystic fat stranding. The correlation tests demonstrated that disease severity correlated with AST, albumin, CRP, the score of pulmonary lesions, the L/S attenuation ratio. Thus, it will allow an earlier identification of high-risk patients for immediate effective intervention in future.
Gastrointestinal symptoms of the patients with COVID-19 were analyzed, and common GI symptoms included diarrhea, nausea, vomiting, and inappetence. The proportion of the patients with GI symptoms in our study was similar to that of previous research 7 .GI symptoms in the patients with COVID-19 may result from the infection of COVID-19 in intestinal tissues 7 . In addition, interaction between brain and gut might cause uncomfortable of gastrointestinal tract. A panel of liver functional tests were determined, the result showed that some of the patients had abnormal liver function, and clinically significant liver injury is uncommon even in critical cases. The mechanism underlying liver damage has not been exploited until now. Liver damage in patients with COVID-19might be caused by the viral infection of liver cells, adverse drug reactions, systemic inflammation, underlying liver diseases, hypoxic hepatitis, etc. [3][4][5] . Xu L et al reported the drugs used for COVID-19 probably induce liver damage, such as antibiotics, antiviral agents, etc. 5 . Thus, to exclude drug-induced liver injury, liver functional tests were collected within 2 days after admission. In addition, liver dysfunction occurred in the patients with chronic liver diseases, so the patients with chronic liver diseases were excluded in our study. Thus, we speculate liver injury in patients with COVID-19 might resulted from viral infection of liver cells, inflammatory responses, and hypoxic hepatitis in our study. Preliminary studies showed that angiotensin-converting enzyme 2 (ACE2) receptor expression is enriched in cholangiocytes 24 and SARS-CoV-2 enter the target cell through ACE2 receptor 24 , which resulted in abnormality of serum γGT, a diagnostic biomarker for cholangiocyte injury. Interestingly, hypoalbuminemia and increased AST level were frequently found in critical cases, and the correlation analysis revealed the correlation of albumin and AST level with severity grading. Thus, biomarkers of liver function containing AST and albumin should be detected in patients with COVID-19.
Common findings of upper abdominal CT scan in patients with COVID-19 were mainly manifested as homogeneously or heterogeneously liver hypodensity and pericholecystic fat stranding. Further study revealed the L/S attenuation ratio and pulmonary lesions correlated positively with the severity of the disease. These indicated that the infection of SARS-CoV-2 may result in multiple organ dysfunction syndrome (MODS) containing liver injury. Pathological examination of liver showed moderate microvascular steatosis and mild lobular and portal activity in a patient with COVID-19 12 . Liver hypodensity was a common signs of liver steatosis or acute hepatitis. While other imaging signs of acute hepatitis, such as hepatomegaly, periportal edema, edema of the gallbladder wall, and periportal lymphadenopathy were not observed in patients with COVID-19. Thus, liver hypodensity and a decrease in the L/S attenuation ratio might be attributed to liver steatosis. Hepatic steatosis is observed in some of the patients with diabetes or chronic liver diseases, thus the patients with chronic liver diseases were excluded and no difference was observed in different groups classified by the severity (Supplementary Table 1). However, the hypothesis should be confirmed by a large sample of pathological examinations in future and the mechanism underlying liver steatosis should be explored in the patients infected with SARS-CoV-2. In addition, pericholecystic fat stranding, a common sign of abdominal CT scan, might be associated with inflammation. Interestingly, enlargement of hilar lymph nodes and spleen were not found.
The correlations of clinical profiles and radiological signs with severity grading were investigated, the results showed severity grading correlated with AST, albumin, CRP, semi-quantitative scoring of pulmonary lesions, the liver-to-spleen (L/S) attenuation ratio. These indicated elevated AST level, hypoalbuminemia, a high CRP, a low L/S attenuation ratio, a high score of pulmonary lesions might be indicators of severe or critical cases.
Obviously, our study had several limitations. Firstly, it was a single-center retrospective study. Selection bias occurred in retrospective cohort studies, the patients with mild system and normal lung manifestation were excluded in our study. In addition, most of the enrolled patients were middle-aged and aged people. Secondly, the patients did not receive contrast-enhanced CT scan. Thirdly, the limited number of cases in this cohort might resulted in the bias of the results. In conclusion: the patients displayed abdominal and pulmonary CT imaging abnormalities and some of the patients have abnormal liver function. Furthermore, some of CT signs and CRP are positively correlated with severity grading; thus, it will allow an earlier identification of high-risk patients. Further studies should be performed to confirm our conclusion and develop suitable therapeutic strategies, and investigate the prognosis in future. Note: L/S, liver-to spleen; * Pearson correlation test, § spearman correlation test.   CT findings in an 80-year-old male patient with severe COVID-19. (A) and (B) Transverse thin-section chest CT scan showed ground-glass opacities in both sides, partial consolidation and small amount of pleural effusion in right side, the semi-quantitative scoring of pulmonary lesions was 10: (C) Transverse thin-section CT scan revealed homogeneous liver hypodensity, the L/S attenuation ratio is 0.99.

Figure 3
Flowchart of the patients' selection, inclusion, and exclusion

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download. Supplementarytable.docx