P59 Outcomes of patients with cirrhosis admitted to intensive care: a real-world district general hospital experience

<h3></h3> Patients with cirrhosis are at risk of life-threatening complications which can require higher level support in an ICU setting. A subset of these patients will present with acute-on-chronic liver failure (ACLF) syndrome, with associated high reported short-term mortality rates. Large retrospective studies report 90-day mortality rates of at least 35% in patients developing ACLF. Recent years have seen an increased development of prognostic scores to identify patients who would be likely to benefit from ICU admission. Our Trust has 2 busy district general hospitals which both see a significant burden of advanced liver disease. A retrospective study was performed on the outcomes of patients with cirrhosis admitted to ICU. A list of patients across the two sites was captured from 2020 to 2021 using coding records. Demographic and clinical details were collected. Child Pugh scores (CPS), Sequential Organ Failure Assessment (SOFA) scores and CLIF-C ACLF scores were calculated at set time points throughout admission. Data was collected for 61 patients. 84% of patients had alcohol-related liver disease; of these, 94% were drinking to the point of admission. 38/61 (62%) of patients were admitted for management of ACLF, and 13/61 (21%) for management of upper gastrointestinal bleed (UGIB). Across the two units, 39% of all cirrhotic patients admitted to ICU survived the admission (42% survival in the ACLF cohort; 35% survival in the non-ACLF cohort). Overall mortality was found to be 61% at 30-days, 67% at 6-months, and 71% at 1-year. Mortality was found to be higher for patients admitted for UGIB (85%) than non-UGIB causes (54%), which contradicts previously reported data. CLIF-C and SOFA outperformed CPS in correlating with overall survival. No scores significantly correlated with overall length of stay. This real-world study was limited by small patient numbers, missing data points and the reliance on clinical coding. It is likely the Covid-19 pandemic impacted upon the data set, and may have impacted upon overall survival and timely access to ICU admission. Prognostic scoring with CLIF-C and SOFA perform well in the studied population. It is concerning that real-world data appears to show a higher mortality rate than previously reported studies, in particular for patients admitted with UGIB. Further in-depth analysis of cases and further post-pandemic data comparison will yield valuable insight into how appropriate access to resources and care delivery can be improved in DGH settings.