key: cord-280551-9hoxy5ok
authors: Kim, Donghee; Adeniji, Nia; Latt, Nyann; Kumar, Sonal; Bloom, Patricia P.; Aby, Elizabeth S.; Perumalswami, Ponni; Roytman, Marina; Li, Michael; Vogel, Alexander S.; Catana, Andreea M.; Wegermann, Kara; Carr, Rotonya M.; Aloman, Costica; Chen, Vincent; Rabiee, Atoosa; Sadowski, Brett; Nguyen, Veronica; Dunn, Winston; Chavin, Kenneth; Zhou, Kali; Lizaola-Mayo, Blanca; Moghe, Akshata; Debes, José; Lee, Tzu-Hao; Branch, Andrea; Viveiros, Kathleen; Chan, Walter; Chascsa, David; Kwo, Paul; Dhanasekaran, Renumathy
title: Predictors of Outcomes of COVID-19 in Patients with Chronic Liver Disease: US Multi-center Study
date: 2020-09-17
journal: Clin Gastroenterol Hepatol
DOI: 10.1016/j.cgh.2020.09.027
sha: 
doc_id: 280551
cord_uid: 9hoxy5ok

Background Chronic liver disease (CLD) represents a major global health burden. We undertook this study to identify the factors associated with adverse outcomes in patients with CLD who acquire the novel coronavirus-2019 (COVID-19). Methods We conducted a multi-center, observational cohort study across 21 institutions in the United States (US) of adult patients with CLD and laboratory-confirmed diagnosis of COVID-19 between March 1, 2020 and May 30, 2020. We performed survival analysis to identify independent predictors of all-cause mortality and COVID-19 related mortality, and multivariate logistic regression to determine the risk of severe COVID-19 in patients with CLD. Results Of the 978 patients in our cohort, 867 patients (mean age 56.9±14.5 years, 55% male) met inclusion criteria. The overall all-cause mortality was 14.0% (n = 121), and 61.7% (n = 535) had severe COVID-19. Patients presenting with diarrhea or nausea/vomiting were more likely to have severe COVID-19. The liver-specific factors associated with independent risk of higher overall mortality were alcohol-related liver disease (ALD) (hazard ratio [HR] 2.42, 95% confidence interval [CI] 1.29-4.55), decompensated cirrhosis (HR 2.91 [1.70-5.00]) and hepatocellular carcinoma (HCC) (HR 3.31 [1.53-7.16]). Other factors were increasing age, diabetes, hypertension, chronic obstructive pulmonary disease and current smoker. Hispanic ethnicity (odds ratio [OR] 2.33 [1.47-3.70]) and decompensated cirrhosis (OR 2.50 [1.20-5.21]) were independently associated with risk for severe COVID-19. Conclusions The risk factors which predict higher overall mortality among patients with CLD and COVID-19 are ALD, decompensated cirrhosis and HCC. Hispanic ethnicity and decompensated cirrhosis are associated with severe COVID-19. Our results will enable risk stratification and personalization of the management of patients with CLD and COVID-19.

Chronic liver disease (CLD) is a major international public health concern and its prevalence has been increasing over the past two decades 1,2 . Around 1.5 billion people have CLD worldwide, and it causes more than 2 million deaths per year 3, 4 . With the rapid spread of the global pandemic of coronavirus disease-2019 (COVID- 19) , there has been significant concern that patients with CLD represent a vulnerable population at higher risk for complications.

Initial concerns were based on the observation that severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus that causes COVID-19, is genetically related to SARS-CoV and Middle East respiratory syndrome coronavirus (MERS-CoV), both of which impair liver function 5, 6 . These concerns appear to have been substantiated with early studies reporting elevations in liver enzymes in up to 50% of patients with COVID-19, with higher prevalence in those with worse prognosis 7, 8 . Preliminary studies from the United States (US) and Europe also suggest that patients with CLD who acquire COVID-19 have high rates of hospitalization and mortality [9] [10] [11] . While these reports raise the alarm, it is not known if all patients with CLD are affected equally or if there are specific subgroups at higher risk for COVID-19 related mortality and morbidity.

Identifying predictors of mortality will allow for risk stratification of patients with CLD affected by COVID-19 and help improve healthcare delivery. To comprehensively characterize the clinical outcomes of COVID-19 in patients with CLD, we undertook a multicenter, observational study of patients with CLD who were diagnosed with COVID-19 in 21 centers across the US.

This is a multicenter observational cohort study. The consortium of investigators to study COVID-19 in chronic liver disease (COLD) study was formed on April 14, 2020, and accrual of data started immediately (registered Clinicaltrials.gov NCT04439084). A total of 21 centers from the US participated in the study (Table S1 ). The institutional review board (IRB) of each participating center reviewed and approved the study protocol. Inclusion criteria constituted: age over 18 years, laboratory-confirmed diagnosis of COVID-19 and presence of pre-existing CLD (according to predefined ICD-10 codes listed in Table S2 and confirmed by manual chart review). Patients who had undergone liver transplantation were excluded. Patients with COVID-19 diagnosis based on clinical suspicion were excluded. All participating institutions independently identified patients meeting inclusion criteria by searching their electronic medical records and collected data as per the previously established data accrual plan. The study retrospectively identified cases diagnosed between March 1 and April 14th, and subsequent cases diagnosed with COVID-19 between April 15th and May 30, 2020 were identified prospectively. All data was collected until death or date of last follow up. Death was attributed to COVID-19 if it was clinically related to COVID-19 illness and there were no other unrelated causes of death 12 .

We collected de-identified data using 170 structured and text variables in 10 different categories. Complete details on the data collection tool are available in Table S3 . Diagnosis of To determine the independent risk factors for the outcome, we performed univariate Cox proportional hazards analysis. Variables were selected for inclusion in the models based on clinical plausibility, statistical significance in the univariate model, and availability in more than 90% of the patients. Multivariate analysis was performed using Cox proportional hazards analysis for outcomes regarding all-cause mortality and deaths due to COVID-19. To investigate the independent determining factors for mortality among patients with and without cirrhosis, analyses were performed using backward stepwise logistic regression (probability to enter = 0.05 and probability to remove = 0.1) due to insufficient outcome events. All analyses were performed using STATA 15.1 (StataCorp, College Station, TX, USA). Two-sided P values were used and considered statistically significant if P ≤ 0.05. All authors had access to the study data and reviewed and approved the final manuscript.

J o u r n a l P r e -p r o o f

We collected data from 21 institutions across thirteen states representing all five regions of the United States (Table S1 ). Data were collected from a total of 978 patients with CLD, of which 867 patients met the inclusion criteria ( Figure S1 ). The largest proportion of the cases were from the Northeast (41.8%) and Southeast (28.4%) regions of the US. The overall all-cause mortality in the cohort was 14.0% (n = 121), and 61.7% (n = 535) patients experienced the composite endpoint of severe COVID-19. Table 1 shows the demographic and clinical characteristics of the patients in the overall cohort and also their proportional distribution based on clinical outcomes. The mean age at the time of COVID-19 diagnosis was 56.9 ± 14.5 years, and 271 patients (31.3%) were ≥ 65 years ( Figure S2 

The majority of patients were tested for COVID-19 because they presented with symptoms (772 Figure S8 ). Patients presenting with GI symptoms of diarrhea (OR 1.89, 95% CI: 1.30-2.74) or nausea/vomiting (OR 1.84, 95% CI: 1.27-2.68) were more likely to have severe COVID-19 than patients without GI symptoms (Table 2) . Also, patients presenting with respiratory symptoms like shortness of breath, sore throat, runny nose or confusion were at higher risk both for mortality and severe COVID-19.

Among patients with CLD and COVID-19, 60.4% (n = 524) were hospitalized, 49.9% (n = 433) required supplemental oxygen, 23.0% (n = 199) were admitted to the ICU, 15.7% (n = 136) J o u r n a l P r e -p r o o f received vasopressors, and 17.8% (n = 154) required mechanical ventilation. The majority of the deaths were due to COVID-19 (86.7%, n = 105). Sixteen patients had non-COVID-19 related mortality, and the cause of death was available in 37.5% (n = 6) of these patients. Two of them died of cardiac failure, two from acute liver failure due to acute alcoholic hepatitis, one from bleeding complications due to coagulopathy and one from septic shock in the setting of acute cholecystitis. New or worsening hepatic decompensation during COVID-19 was noted in 67 

To identify the predictors of all-cause mortality and COVID-19 related mortality, we performed univariate and multivariate survival analysis ( (Table 3) , the results were largely identical.

Further, we did not find significant interactions between ALD and decompensated CLD or HCC for overall survival on multivariate analysis (test of interaction P > 0.2) (Table S4 ).

Next, we performed a subgroup survival analysis in patients with cirrhosis and COVID-19 ( Table   4 ). The liver-specific factors associated with higher all-cause mortality in patients with cirrhosis were prior hepatic decompensation (HR 3.89, 95% CI: 2.18-6.95), HCC (HR 3.66, 95% CI: 1.67-8.01). In the subgroup of patients with non-cirrhotic CLD, ALD was associated with higher allcause mortality (HR 4.72, 95% CI: 2.05-10.85) and higher COVID-19 related mortality (HR 7.39, 95% CI: 2.96-18.46) ( Table S5) .

Serial liver related labs were available in a majority of the hospitalized patients, but not in the majority of those managed as outpatient. We performed a subgroup analysis in hospitalized J o u r n a l P r e -p r o o f patients in whom serial lab values were available for analysis. Peak values of AST, bilirubin, alkaline phosphatase and MELD score were observed to predict mortality (Table S6) .

Overall, 535 patients with CLD met criteria for the composite endpoint of severe COVID-19. As shown in Table J o u r n a l P r e -p r o o f DISCUSSION According to the Center for Disease Control (CDC), patients with CLD might be at increased risk for severe illness with COVID-19 16 . CLD represents a clinical spectrum ranging from mild asymptomatic disease to severe decompensated cirrhosis. It is not clear which subgroups of patients with CLD are more vulnerable to adverse outcomes with COVID-19. In this multi-center study, we investigated predictors of mortality and COVID-19 disease severity in patients with CLD and SARS-CoV-2 infection. Among the 867 patients with CLD from 21 centers across the US, we observed an all-cause mortality of 14.0%; 60.4% were hospitalized and 23% were admitted to the ICU. New or worsening hepatic decompensation during COVID-19 was noted in 7.7% of patients. We identified the liver-specific factors ALD, hepatic decompensation and HCC as predictors of adverse outcomes from COVID-19, apart from established factors like older age, hypertension, diabetes and COPD. Additionally, we found that patients of Hispanic ethnicity had a higher risk for severe COVID-19. Thus, our large multicenter study identifies specific subgroups of patients with CLD who have higher mortality with COVID-19.

Given that COVID-19 is a novel pandemic, our knowledge of its impact on patients with CLD is still evolving. Singh et al recently identified 250 patients with COVID-19 who had an underlying CLD using a de-identified research network database, and reported a hospitalization rate of 52% and mortality 12%, similar rates to our study 9 . Preliminary data from an international registry of 152 patients with CLD however reported a higher overall mortality rate of 31% and a hospitalization rate of 95% for patients with cirrhosis 11 . The higher mortality rates in this clinician-reported registry study may have been due to selection bias. Around 90% of the J o u r n a l P r e -p r o o f patients with CLD and COVID-19 in our cohort had mild liver disease with either non-cirrhotic stage disease or compensated cirrhosis at baseline, and they had relatively favorable outcomes.

Patients with decompensated cirrhosis were disproportionately adversely affected by COVID- 19 , with an all-cause mortality rate of 31.4% in this subgroup. These findings are in line with the higher morbidity and mortality in patients with decompensated cirrhosis and influenza pneumonia 17, 18 . We posit the less favorable outcomes noted in patients with decompensated cirrhosis may be due to cirrhosis-associated immune dysfunction and fragile physiological buffers, likely increasing susceptibility to severe COVID-19 19 . Our findings highlight the challenges in taking extra precautions to minimize the risk of exposure to SARS-CoV-2 in the vulnerable patients with decompensated cirrhosis, while continuing to optimally manage their decompensating events.

In our study, ALD was independently associated with a higher risk of poor survival and COVID-19 related mortality. This is a novel association and one that has significant implications for patients with CLD. Patients with ALD are known to be at higher risk for infections due to the underlying dysregulation of the immune system 20 . ALD is associated with a sterile inflammatory state induced by damage-associated molecular patterns (DAMPs), which leads to the systemic production of pro-inflammatory cytokines by various immune cells 21, 22 . We hypothesize that the superimposed cytokine storm triggered by SARS-CoV-2 could exacerbate the heightened inflammatory state in patients with ALD thus leading to worse outcomes 23 .

Moreover, there has been significant concern about increased alcohol use during the COVID-19 pandemic, highlighting the importance of this association 24,25 . In our study, up to a third of J o u r n a l P r e -p r o o f patients with CLD, and an alarming 50% of patients with ALD reported daily alcohol consumption, which was disconcertingly associated with poor outcomes in patients with cirrhosis and COVID-19. These findings emphasize the need to implement an aggressive remote care plan for patients with ALD to manage their alcohol use disorder while simultaneously minimizing the risk of exposure to COVID-19. Future studies will be needed to analyze specific subgroups within the spectrum of alcohol liver disease (ALD) who are at higher risk for adverse outcomes with COVID-19.

Another subgroup that was found in our study to be at significantly high risk for mortality was that of patients with HCC. The all-cause mortality rate in this subgroup was 52.4% (n = 11), almost 7-fold higher than in patients without HCC, however the number of patients is small.

Patients with cancer, in general, have worse clinical outcomes after COVID-19 14, 26 . Patients with HCC may be uniquely susceptible to COVID-19 related complications due to a constellation of active malignancy, presence of cirrhosis, as well as the presence of an active underlying liver disease that led to that cirrhosis, all resulting in compromised immune function, which may be further complicated by HCC-directed treatment.

Our cohort includes a racially and ethnically diverse population that is 31% non-Hispanic white, 31% non-Hispanic black and 25% Hispanic. We found that patients of Hispanic ethnicity had a higher risk of developing severe COVID-19 compared to non-Hispanic Whites, even after adjusting for age, comorbidities and hepatic decompensation. These findings are in line with recent reports showing higher age-adjusted rates of hospitalization in Hispanic patients 27,28 .

The strengths of our study include large sample size, broad geographical distribution of sites across the US as well as the granularity of the collected data. We have included patients treated both as outpatients or inpatients, and also patients with non-cirrhotic or cirrhotic stage CLD, thus making our findings generalizable. Limitations of our study include the retrospectiveprospective timeline which was used mainly because of the rapidly evolving nature of the pandemic. Another limitation of our study is the restriction of SARS-CoV-2 testing during the earlier phase of the pandemic, likely leading to decreased representation of mild COVID-19.

Also, we could have enrollment bias since not all patients with CLD have a documented ICD9/10 code in their electronic health records. Also, despite our best efforts, it is possible that not all patients with CLD and COVID19 were identified from the participating centers. Lastly, the majority of the contributing centers are tertiary medical health systems, potentially introducing referral bias. However, our cohort represents an ethnically diverse population with varying stages of liver disease. Larger and longer-term studies will be needed to confirm these findings.

To date, this is the largest study on COVID-19 among patients with CLD in the United States.

Our cohort of 867 patients with CLD had substantial rates of all-cause mortality (14.0%), hospitalization (60.4%) and ICU admission (23%). We identify decompensated cirrhosis, ALD, and HCC to be determinants of mortality in patients with CLD, and additionally show that Hispanic ethnicity is independently associated with severe COVID-19. These findings can be used to prospectively design protective measures for these vulnerable populations, such as continuing the emphasis on telemedicine, prioritizing them for future vaccinations, as well as actively including these patients in prospective COVID-19 surveillance studies and drug trials. NAFLD, nonalcoholic fatty liver disease; ALD, alcoholic liver disease; COPD, chronic obstructive pulmonary disease.

Data are expressed as the mean ± standard deviation or number (proportion). To identify candidate risk factors of mortality, we performed a stepwise backward logistic regression analysis (probability to enter = 0.05 and probability to remove = 0.1) using all variables in the univariate model.

J o u r n a l P r e -p r o o f The clinical outcomes of patients with chronic liver disease (CLD) and the novel coronavirus disease-2019 (COVID-19) are not well defined. Also, it is not clear which patients with CLD are most vulnerable to adverse outcomes from COVID-19.

In this large study of 867 patients from 21 centers across the US with CLD with COVID-19 we determine that patients with alcohol related liver disease (ALD), decompensated cirrhosis and hepatocellular carcinoma have a high risk for allcause mortality from COVID-19.

Lack of adequate COVID-19 testing during the early phase of the pandemic could have led to decreased representation of patients with CLD and mild COVID-19 in our cohort.

Our findings will enable risk stratification and personalized management of patients with CLD who acquire COVID-19.

Moreover, the association between ALD and poor outcomes with COVID-19 has broad public health implications given recent concerns about increased alcohol consumption during the pandemic.

J o u r n a l P r e -p r o o f Figure S1 : Patient Study Cohort 2. Figure S2 : Age at the time of diagnosis of COVID-19 in patients with CLD stratified by Overall Mortality 3. Figure S3 : Patient Demographics Stratified by Clinical Outcomes 4. Figure S4 : Comorbidities in Patients with CLD and COVID-19. Figure S5 : Etiology of CLD among Patients with COVID-19

6. Figure S6 : Etiology of CLD and severity of COVID-19

7. Figure S7 : Risk factors for COVID-19 and Indications for Testing in Patients with CLD 8. Figure S8 : Presenting Symptoms of COVID-19 among Patients with CLD 9. Figure S9 Liver tests during COVID-19 10 . Figure S10 : Treatment for COVID-19 among Patients with CLD

1 . Table S1 : List of participating institutions and investigators from each site 2. Table S2 . List of ICD-10 Codes for Chronic Liver Disease and COVID-19

3. The multivariate model for all-cause mortality was adjusted for age, gender, race/ethnicity, etiology of chronic liver disease, decompensated cirrhosis, HCC, diabetes, hypertension, cardiovascular disease, COPD, smoking status, alcohol consumption, and the interaction term (alcoholic liver disease*decompensated cirrhosis or alcoholic liver disease*HCC).

The multivariate model for mortality due to COVID-19 was adjusted for age, gender, race/ethnicity, etiology of chronic liver disease, decompensated cirrhosis, HCC, diabetes, hypertension, cardiovascular disease, COPD, smoking status, and the interaction term (alcoholic liver disease*decompensated cirrhosis or alcoholic liver disease*HCC). J o u r n a l P r e -p r o o f

Trends in the Burden of Chronic Liver Disease Among Hospitalized US Adults

The global, regional, and national burden of cirrhosis by cause in 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study

Burden of liver diseases in the world

Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study

SARS-associated viral hepatitis caused by a novel coronavirus: report of three cases

Histopathology of Middle East respiratory syndrome coronovirus (MERS-CoV) infection -clinicopathological and ultrastructural study

Liver injury in COVID-19: management and challenges

Clinical Features of COVID-19-Related Liver Damage

Clinical Characteristics and Outcomes of COVID-19 Among Patients with Pre-Existing Liver Disease in United States: A Multi-Center Research Network Study

High rates of 30-day mortality in patients with cirrhosis and COVID-19

High Mortality Rates for SARS-CoV-2 Infection in Patients with Pre-existing Chronic Liver Disease and Cirrhosis: Preliminary Results from an International Registry

Guidance for certifying deaths due to coronavirus disease 2019 (COVID-19)

Determinants of alcohol use and abuse: Impact of quantity and frequency patterns on liver disease

Clinical impact of COVID-19 on patients with cancer (CCC19): a cohort study

Clinical Characteristics and Outcomes of Coronavirus Disease 2019 Among Patients With Preexisting Liver Disease in the United States: A Multicenter Research Network Study

Coronavirus Disease 2019-People with Certain Medical Conditions

Hepatic decompensation in patients with cirrhosis during infection with influenza A

Influenza virus infection as precipitating event of acute-on-chronic liver failure

Cirrhosis-associated immune dysfunction: distinctive features and clinical relevance

Infection and Alcoholic Liver Disease

Laboratory Characteristics among Hospitalized Patients with Chronic Liver Disease and COVID-19 (n=524)

All-cause Mortality status P value Alive Died ALT Before COVID-19 (n=374)

Abbreviation: COVID-19, coronavirus disease 2019; ALT, alanine aminotransferase; AST, aspartate aminotransferase

Data are expressed as the median (interquartile range). Mann-Whitney U test were performed for comparison between groups