ABSTRACT FINAL ID: 1746
TITLE: Post-Discharge Mortality in a National Cohort of Veterans Affairs Patients with Cirrhosis
SPONSORSHIP - THIS STUDY WAS SPONSORED BY: (IF THIS ABSTRACT WAS NOT SPONSORED PLEASE INDICATE):
This study was supported by Veterans Health Administration Health Services Research & Development (HSR&D) Investigator Initiated Research (IIR 13-052).
Background: Cirrhosis and its complications lead to more 150,000 hospitalizations annually and the number of cirrhotic patients is increasing. Mortality models have been developed that have used large administrative databases, which miss key clinical data such as labs, or use richer datasets but have much smaller sample sizes. Our aim was to look at a broad set of clinical factors that affect mortality in patients with cirrhosis that are hospitalized for any cause in a large national cohort of Veterans Affairs patients.
Methods: We analyzed a retrospective cohort of patients hospitalized in the Department of Veterans Affairs (VA) between 2005 and 2013. We included patients with an established International Classification of Diseases (ICD) 9 code for cirrhosis and/or a cirrhosis complication and selected their first admission for analysis. We built a Cox proportional hazards model to estimate the effect of 161 variables at discharge on post-discharge mortality, censored for liver transplant or end of study period. The variables included demographics (3), most recent laboratory values (18), outpatient (26) and inpatient medications (39), most recent vital signs (7), comorbidities (57), risk scores (6), healthcare utilization (4), and length of stay. We used k-nearest neighbor imputation for missing values.
Results: 76,815 hospitalizations for cirrhotic patients who survived to discharge were identified. Alcoholic cirrhosis comprised 27%, viral hepatitis 22%, viral hepatitis and alcohol 24%, and non-alcoholic fatty liver disease 8%. The 30-day, 90-day, 1-year, and 5-year mortality in our cohort was 7%, 15%, 29%, and 52% respectively. The median follow up was 780 days (IQR: 254-1608). Ninety one factors were statistically significant, of which 14 increased the hazard ratio (HR) by greater than 20%: age (HR 1.21, 95% C.I. [1.20-1.23]), hypertension (1.22, [1.12-1.33]), cancer (1.30, [1.17-1.44]), ascites (1.24, [1.21-1.28]), Hepatorenal syndrome (1.31, [1.22-1.42]), hepatocellular carcinoma (1.62, [1.55-1.68]), metastatic cancer (2.00, [1.14-3.50]), inpatient (IP) midodrine (1.31, [1.18-1.45]), IP anti-tubercular medications (1.31, [1.11-1.54]), IP glucocorticoids (1.24, [120-1.29]), home med antiemetics (1.30, [1.21-1.40]), “Unknown Race” (1.24, [1.13-1.36]), male gender (1.38, [1.28-1.50]), and discharge MELD (1.30, [1.27-1.32]). A history of TIPS was one of the few variables to dramatically lower mortality (0.70, [0.59-0.83])
Conclusions: This study, of a national VA cohort, identified ninety one variables that significantly influence mortality of patients hospitalized for any cause with cirrhosis including novel factors not reported in the literature.