Infections in Patients With Cirrhosis Have a Changing Profile Over Time and Are Linked to a Higher Risk of Death Compared to Patients Without Cirrhosis

Introduction: Intensive care outcomes in patients with cirrhosis are relatively poor. The comparison between outcomes, especially related to infections, remains unclear in those with and without cirrhosis. With the emergence of resistant and fungal organisms, the changes in infection profiles over time are important to analyze. The aim of this study is to determine the impact of cirrhosis and infections on inpatient death over time in a qSOFA-matched cohort of patients with and without cirrhosis. Method(s): Inpatients admitted to ICUs throughout 2015-2021 were analyzed. Patients with cirrhosis were matched 1:1 by age, gender, and admission qSOFA to patients without;COVID-positive patients were excluded. Admission demographics, labs, the reasons for ICU transfer, infections, and inpatient death or hospice referral were obtained for each patient. Comparisons were made between patients with and without cirrhosis and those who died/referred to hospice versus not. Logistic regression for death/hospice was performed. In patients with cirrhosis, the culture results were compared over the years. Result(s): 1669 patients;833 cirrhosis and 836 non-cirrhosis patients were included. Patients with cirrhosis had higher rates of infection, positive culture, abdominal infection, and bacteremia. They also had higher gram-positive and fungal infections with a higher rate of VRE. They showed a greater organ failure load, death, and hospice referral compared to patients without cirrhosis. Logistic regression showed that cirrhosis (OR 4.0, p< 0.0001), admission qSOFA (1.60, p< 0.0001), WBC (1.02, p=0.003), reasons for ICU (altered mental status 1.69, hypotension 1.79, renal support 2.77, respiratory failure 1.79, CVA 1.96, all p< 0.0001) with Infection (1.77, p< 0.0001, >1 microbe isolated 1.86, p=0.05) were risk factors for death/hospice. The infection trend in the cirrhosis group showed a significant decrease in positive cultures and gram-negative infections and an increase in fungal and gram-positive infections over time. Conclusion(s): Despite matching for demographics and qSOFA, patients with cirrhosis had higher risks of death and organ failures. They were more likely to develop gram-positive and fungal infections with multiple organisms and VRE. Time trends in cirrhosis showed lower rates of positive cultures and gram-negative infections and an increase in fungal and gram-positive infections over time, which should encourage re-evaluation of diagnostic and prophylactic strategies in cirrhosis-related infections. (Figure Presented).

Low Serum Albumin at Admission Is Associated with Mortality among Inpatients with Cirrhosis: A Multi-National Experience

Background: MELD and Child-Pugh scores have traditionally been used as prognostic indicators in patients with cirrhosis. Albumin infusions in outpatients have been associated with improved outcomes, but not in transplant waitlisted patients or inpatients. This aim of this study was to assess whether low serum albumin (sAlb) on admission alone is a poor prognostic indicator among cirrhotic inpatients from a new multi-national cohort. Method(s): The CLEARED study is a global study that enrolled consecutive non-electively admitted inpatients without organ transplant or COVID-19 from 6 continents. Admission demographics, medical history, laboratory data, inpatient course, death/hospice transfer and mortality at 30 days post-discharge were recorded. Patients were divided into 3 groups: sAlb <28gm/L(A), sAlb >=28 but <35gm/L (B), and sAlb>=35gm/L (C) were compared. Multi-variable logistic regression was performed using inpatient mortality and overall 30-day mortality as outcomes. Result(s): 2429 patients were enrolled from 21 countries worldwide. The distribution was A:49%, B:39%, C:12%. Gp A patients were significantly younger (54yrs vs. 57yrs vs 58yrs p<0.0001) but with similar gender distribution, and higher MELD-Na score of 25 vs. 20 vs. 17 (p<0.0001). Gp A patients were more likely to have alcohol as etiology of cirrhosis (49% vs. 45% vs 38%, p=0.004), and were more likely to have either infection (27% vs. 18% vs. 13%, p<0.0001), HE (39% vs. 33% vs. 23%, p=0.005) or fluid related issues as a reason for admission (p<0.0001). More patients in Gp A received albumin infusion during their hospital stay (120gm vs. 100gm vs. 100gm p=0.0004), mostly for the indications of AKI (47% vs. 49% vs. 47%, p=0.79) and performance of large volume paracentesis (44% vs. 42% vs. 41%, p=0.80), followed by bacterial peritonitis indication (22% vs. 17% vs. 11%, p=0.01). Group A patients had longer hospital stays (9 days vs. 8 days vs. 7 days (p<0.001), but similar ICU transfer (23% vs. 22% vs. 20%, p=0.55). group A patients were more likely to die while inpatients (19% vs. 11% vs. 5%, p<0.0001), or by 30 days post-discharge (29% vs. 20% vs. 9%, p<0.0001). Table shows the admission variables associated with a poor outcome. Conclusion(s): Hypoalbuminemia is extremely common among admitted cirrhotic patients, with sAlb of <28gm/L occurring in almost half. Together with MELD-Na score and infection at admission, a low sAlb is associated with a poor outcome in these patients. Future studies will need to validate these findings and to assess whether albumin infusions will improve the outcome of these patients. (Figure Presented).

Differences in Mortality Rates in Inpatients with Cirrhosis Depends on the Low, Middle or High Income Country Location: Multi-National Cleared Consortium Experience

Background: Although cirrhosis is a major cause of mortality worldwide, there could be disparities in outcomes. This needs a global consortium to study disparities in inpatient cirrhosis care Aim: Define the impact of location in prediction of outcomes in inpts with cirrhosis. Method(s): CLEARED prospectively enrolled non-electively admitted cirrhosis pts without COVID from all continents. To ensure equity, we allowed only 50 pts/site. Admission details, cirrhosis history, inpatient & 30-day course were recorded. World bank classification of low/low middle income (LMI), upper middle (UMI) & High income (HI) were used. Cirrhosis details, inpatient & 30-day outcomes were compared between groups. Multi-variable regression was performed using inpatient & 30-day mortality as outcomes. Result(s): 2758 pts from 21 countries from all continents, including Africa & Australia, were included.727 were L/LMI, 1050 UMI & 981 pts were from HICs. More men & younger pts were in LMI. Cirrhosis details: More pts in HI gp had 6M hospitalizations & infections, HE & ascites while prior variceal bleeding was higher in LMI . Prior HCC & transplant listings were lower in LMI but similar in UMI/HI. Alcohol & NASH was highest in HI. Viral hepatitis & cryptogenic were highest in UMI.Admissions: Admission MELD was highest in LMI. LMI pts were admitted more for GI Bleed, HE, & DILI, while anasarca & HBV flares were higher in UMI. Higher SBP (36% vs 24% vs 21% p<0.0001) & lowest skin/soft-tissue infections were in LMI (5% vs 5% vs 10% p=0.008);rest were similar. Nosocomial infections, driven by UTI were highest in LMI & HI pts (15% vs 14% vs 11% UMI, p=0.03). Admission diuretics, PPIs, Lactulose & statins were highest & antivirals lower in HI. SBP prophylaxis & rifaximin were highest in LMI pts. Outcome(s): More LMI pts needed ICU & had more organ failures (Fig B). Discharge MELD was highest in LMI. In-hospital mortality was highest & transplant lowest in LMI. This extended to 30-day mortality & transplant in LMI patients vs HI pts.Regression: In-hospital mortality was linked with age, infections, MELD & being in a LMI/UMI vs HIC while being on a transplant list, diabetes, & SBP prophylaxis were protective (Fig C). 30-day mortality predicted by age, ascites, HCC, discharge MELD, organ failures, LMI/UMI vs HIC but rifaximin was protective(Fig D). In-hospital transplant was higher with high MELD, admission rifaximin & listed pts &lower in LMI (OR 0.26) & UMI (OR 0.22) & age. 30-day transplant was higher in those with hyponatremia, ascites & HRS, on the list & on rifaximin and lower in LMI (OR 0.24) & UMI (OR 0.59) vs HI. Conclusion(s): In a global study of inpatients with cirrhosis, there were major differences in outcomes. Not being in a high-income country significantly increased the risk of inpatient and 30-day mortality independent of demographics, medications, in-hospital course, and cirrhosis severity likely due to disparities in access to transplant, which should be accounted for in global models. (Figure Presented).

CLEARED Global consortium results highlight regional variation and need for equity in inpatient outcomes in hospitalized patients with Chronic Liver Disease

Background and aims: A global study with equitable participation for cirrhosis and chronic liver disease (CLD) outcomes is needed. We initiated the Chronic Liver disease Evolution And Registry for Events and Decompensation (CLEARED) study to provide this global perspective. Aim to evaluate determinants of inpatient mortality and organ dysfunction in a multi-center worldwide study. Method: We prospectively enrolled pts with CLD/Cirrhosis >18 years without organ transplant or COVID-19 who were admitted non-electively. To maintain equity in outcome analysis, a maximum of 50 pts/site were allowed. Data for admission variables, hospital course, and inpatient outcomes (ICU, death, organ dysfunction [ODF]) were recorded. This was analyzed for death and ODs using significant variables on admission and including World Bank classification of low/middle-income countries (LMIC). A model for in-hospital mortality for all variables during the hospital course, including ODs) was analyzed. Results: 1383 pts (55 ± 13 yrs, 64% men, 39% White, 30% Asian, 10% Hispanic, 9% Black, 12% other) were enrolled from 49 centers (Fig A). 39% were from high-income while the rest were from LMICs. Admission MELDNa 23 (6–40) with history in past 6 months of hospitalizations 51%, infections 25%, HE 32%, AKI 23%, prior LVP 15%, hydrothorax 8% and HCC 4%. Leading etiologies were Alcohol 46% then NASH 23%, HCV 11% and HBV 13%. Most were on lactulose 52%, diuretics 53%, PPI 49% and statins 11%, SBP prophylaxis 16%, beta-blockers 35% and rifaximin 31%. 90% were admitted for liver-related reasons;GI bleed 30%, HE 34%, AKI 33%, electrolyte issues 30%, anasarca 24% and 25% admission infections. In-hospital course: Median LOS was 7 (1–140) days with 25% needing ICU. 15% died in hospital, 3% were transplanted, 46% developed AKI,15% grade 3–4 HE, 14% shock, 13% nosocomial infections and 13% needed ventilation. Logistic Regression: Fig B shows that liver-related/unrelated factors on admission which predicted in-hospital mortality and development of organ dysfunction with MELDNa and Infections being common among all models. Nosocomial infections and organ dysfunctions predicted mortality when all variables were considered. High-income countries had better mortality outcomes likely due to transplant and ICU availability. AUCs were >0.75 (Figure Presented) Conclusion: In this worldwide equitable experience, admission cirrhosis severity and infections are associated with inpatient outcomes, which are greater in low-income settings. Liver-related and unrelated factors and regional variations are important in defining critical care goals and outcome models in inpatients with cirrhosis.

PATIENTS WITH CIRRHOSIS EXPERIENCE HIGHER ICU MORTALITY COMPARED TO NON-CIRRHOTIC PATIENTS WHICH CONTINUED WITH OR WITHOUT COVID-19 DESPITE CONTROLLING FOR ORGAN FAILURE SEVERITY

Patients with cirrhosis have a relatively poor prognosis in intensive care (ICU) that could be affected by the9 pandemic. However, the impact of cirrhosis care compared to noncirrhotic patients is unclear pre and post-pandemic. Aim: Define impact of cirrhosis on mortality in ICU patients before & after COVID-19. Methods: ICU pts from a large tertiary hospital who were admitted for >24 hours were divided into pre-COVID (2019) and postpandemic (2020) eras. We excluded patients where cirrhosis diagnosis was unclear. Within the 2020 cohort, we further divided pts into COVID-positive or negative based on PCR. Pts with cirrhosis were matched 1:1 to non-cirrhotic pts with respect to age, ICU admission qSOFA & ICU length of stay in both cohorts. Reasons for ICU admission, infections, organ failures and discharge information were collected. We first compared only COVID negative cirrhosis vs other pts in the pre and post cohorts & then further compared these within the COVID positive pts. Logistic regression with death/hospice as the outcome was used with cirrhosis status, qSOFA, reason for ICU admission and organ failures as independent variables in the three matched cohorts (pre-COVID, post-COVID positive & post-COVID negative). Finally, to evaluate the relative impact of cirrhosis vs COVID-19, we combined the 2020 cohort and determined death/hospice determinants. Results: We included 200 age/LOS/qSOFA-matched pts with/without cirrhosis in pre-COVID cohort. Post-COVID similarly, 200 pts were included in the COVID negative group. 64 COVID+ pts (with/ without cirrhosis) were also included. More non-cirrhotic pts were admitted for procedural observation & stroke while altered mental status (AMS) were similar. Remaining organ failures were higher in cirrhosis in pre- and post-COVID settings (Table 1). In COVIDpositive pts, cirrhosis had lower infections, respiratory failure and intubation but trended towards higher death. Cirrhosis comparison pre vs post-COVID: Post-COVID cirrhosis pts had a higher MELDNa score (15.4±7.9 vs 22.3±10.2, p=0.004)and qSOFA (2.4 vs 1.7, p<0.001) compared to pre-COVID. Logistic regression for death/hospice (Table 2): Pre COVID was significant for cirrhosis, qSOFA , altered mental status & Pressors. Post-COVID in COVID-negative pts it was again significant for cirrhosis, Infection, renal failure & qSOFA. For only COVID positive patients, only renal failure was significant. In the entire 2020 cohort, COVID-19 positive status was not significant in death/hospice prediction, but cirrhosis remained significant. Conclusions: Cirrhosis remains a major cause of mortality in patients admitted to intensive care that continues regardless of COVID-19 pandemic-induced changes in the health system. Cirrhosis is predictive of death independent of COVID-19 despite controlling for demographics and organ failure severity. (Table Presented)

TRACKING CIRRHOSIS CARE DURING COVID-19 PANDEMIC

Background and Aims: The COVID-19 pandemic has resulted in profound disruptions in the delivery of health care. These disruptions have disproportionately impacted patients with complex diseases, such as cirrhosis. This study aimed to evaluate the impact of the COVID-19 pandemic on the implementation of high-value care among veterans with cirrhosis. Method: All US Veterans with cirrhosis who were enrolled in VA health care in the fiscal year 2019 were identified using ICD-10-CM codes. This cohort was then followed through the end of the fiscal year 2021. Quarterly outcomes measured included every 6- months HCC screening (HCC-6), endoscopic variceal surveillance or treatment with nonselective beta-blockers (EVST), and all-cause hospitalizations. Joinpoint trend analysis was used to identify the time point at which health care measures declined during the pandemic. Multivariate logistic regression was used to compare pre-and post-COVID care, using January 2020 (onset of COVID pandemic) as the cutpoint. Results: There were 71,552 veterans with cirrhosis in VA care in 2019. This cohort was 96% male, with a median age of 66 (IQR: 61-70), and predominantly alcohol- or hepatitis C-related cirrhosis (Table 1). Over the following 3 years, 22,043 were censored due to death, withdrawal from the VA, or liver transplantation. HCC-6, which was stable at an average of 38% before COVID-19, had a sharp decline in the early phase of COVID-19 to a nadir of 31% (p<0,01) and is now gradually recovering (Figure 1A). Although the rates of EVST were decreasing before COVID- 19 onset, they dropped precipitously starting in 2020-Q1 and have not yet recovered (Figure 1B), with the rate of 49% dropping to 43% (p<0.01). There was no clear point at which hospitalization changed in the joinpoint analysis, though this measure has decreased slowly during COVID-19 from a 7.6% quarterly admission rate to 5.8% during the pandemic (Figure 1C). Similarly, on multivariate analysis, the odds of HCC-6 (0.68[0.63-0.65]), EVST (0.58[0.56-0.61]) and hospitalization (0.85[0.83-0.87]) were all lower during the COVID- 19 period. Conclusion: Despite the ongoing COVID-19 pandemic, HCC surveillance is recovering in VA. However, EVST rates remain low. Further research is needed to understand the decrease in hospitalization for patients with cirrhosis and identify measures to improve cirrhosis care. (Table Presented) (Figure Presented)

Implementation strategies associated with improved cirrhosis care during COVID-19 in the veterans health administration

Background: The COVID-19 pandemic has presented tremendous hurdles to the continuation of preventative services worldwide, including surveillance for HCC for patients with cirrhosis. The Veterans Health Administration (VA) Hepatic Innovation Team Learning Collaborative (HIT) aims to improve care for Veterans with cirrhosis nationally. This evaluation aimed to assess 1) the approaches that sites used to providing cirrhosis care (implementation strategies) during the pandemic and 2) the implementation strategies associated with improved HCC surveillance rates during the pandemic. Methods: VA hepatology clinicians were surveyed about the use of 73 implementation strategies used to improve cirrhosis care in fiscal year (FY) 2020 and whether strategy use was affected by COVID-19. Descriptive and bivariate statistics defined frequencies and associations with HCC surveillance in the year. Results: Survey responses were received from 72 (55%) VA sites caring for over 42,000 Veterans with cirrhosis. Over the course of the pandemic in FY20, the HCC surveillance rate nationally declined from 51% to a nadir of 39%, with an overall rate of 40% at the end of FY20. Sites reported using a median of 10 (IQR 4-20) of 73 implementation strategies to improve cirrhosis care in FY20. Implementation strategy use shifted during the pandemic such that, relative to pre-COVID, sites engaged in more academic collaborations, local technical assistance, and networking to problem solve and less outreach to patients, clinical performance data, site visits, and external facilitation. Implementation strategy selection was associated with HCC surveillance;sites using more implementation strategies had significantly higher HCC surveillance rates (p=.046). Six of 73 individual implementation strategies were significantly associated with increased HCC surveillance: 1) use clinical reminders or note templates, 2) create new clinical teams, 3) use data experts, 4) tailor care to meet local needs, 5) learn from experts in cirrhosis care, and 6) use interdisciplinary workgroups to provide input into cirrhosis policies and practices. Conclusion: While VA experienced an expected national decline in HCC surveillance during FY20, sites used a variety of shifting strategies to adapt to the changing environment. A subset of implementation strategies, including interdisciplinary team formation and using medical record tools and data, was associated with higher surveillance rates during the pandemic.

Functional capacity in cirrhosis is predicted by additive impact of covert hepatic encephalopathy and frailty in a multi-center outpatient cohort

Background: Daily functional capacity is a major determinant of outcomes in patients with chronic medical conditions. Given that it can be affected by disease-specific factors as well as physical and cognitive impairment, it may be of increasing relevance in patients with cirrhosis given the changing demographics and increasing co-existing conditions including HE and physical frailty. An integrated multi-site approach combining cirrhosis-related factors, comorbidities, cognitive function, and frailty metrics is needed. Aim: To determine the integrated effect of frailty and CHE on functional capacity in outpatients with cirrhosis. Methods: NACSELD-3 (North American Consortium for the Study of End-Stage Liver Disease) is a new cohort of outpatients with cirrhosis recruited from 11 centers across North America. We enrolled pts able to consent, without HIV/illicit drug use or current alcohol misuse. Demographics, cirrhosis severity/history, comorbidities, medications were recorded. DASI (Duke Activity Status Index, Low=worse), studies that assess functional capacity, Liver frailty index instruments (LFI, high=worse) & EncephalApp Stroop (High time=worse) were administered. Norms were used to classify pts as having CHE on EncephalApp & frailty on LFI. Pts divided into having none, either or both CHE & frailty. Regression analyses were performed for DASI using all clinical variables collected. Results: Demographics: 220 patients (61.7±10.6 yrs, 74% men, 76% White, 6% Latinx) were enrolled;EncephalApp & LFI were complete in 182 pts (redgreen color blindness, logistic issues or COVID restrictions) Cirrhosis details: Major etiologies were 37% alcohol, 26% NAFLD, 17% HCV and 10% HCV+alcohol. Mean MELD was 13.9±8.5, 36% had prior HE, and 19% had difficult to control ascites. Mean Charlson comorbidity index was 5.1±2.2 Cognition, frailty and DASI: EncephalApp total was 184.2±55.6 seconds and 148 (64%) pts had CHE. Mean LFI score was 3.89±0.63 and 37 (16%) were deemed frail. 49 (27%) had neither CHE nor frailty, 104 (58%) had either CHE or Frailty , and 26 (15%) had both (Fig A). EncephalApp & LFI were positively correlated (r=0.36, p<0.001) and both were correlated with DASI (EncephalApp r=-0.33, LFI r=-0.27, both p<0.001). DASI was lower with both CHE & Frailty (Fig B). Regression: Variables associated with lower DASI (poor capacity) were higher MELD score (T-value -2.1, p=0.03), higher CCI (T-value -3.6, p<0.0001) and being frail+CHE versus either or none (T-value -2.6, p=0.01). No interaction between LFI and EncephalApp was seen. Conclusion: In this multi-center experience combined frailty and covert hepatic encephalopathy and cirrhosis-unrelated comorbidities significantly add to MELD score in predicting functional capacity in outpatients with cirrhosis.