P21 Predictors of relapse to alcohol after hospital admission for patients with alcohol related cirrhosis

<h3>Background and Aims</h3> Hospital admission provides an opportunity to achieve abstinence for patients with alcohol related cirrhosis. Unfortunately relapse to alcohol is common. We examined the risk factors for relapse in this setting. <h3>Method</h3> This was a retrospective single centre study. We included patients with a diagnosis of alcohol related cirrhosis and/or alcohol-associated hepatitis with a hospital admission between Sept 2014 and March 2023. Where there were multiple admissions the first admission on or after the diagnosis was used. We collected a granular dataset which included clinical and psychosocial factors felt to be relevant in predicting abstinence. The outcome of interest was return to alcohol use at one year. Continuous variables were compared using Mann-Whitney U and categorical variables using chi squared test. Variables that were statistically significant (p ≤ 0.05) on univariate analysis were entered into a backward stepwise elimination variable selection procedure (multivariate logistic regression). <h3>Results</h3> We identified 384 patients with alcohol use data at one year. Median age was 57 (IQR 50–66) and 35% were female. The rate of ongoing alcohol use at one year was 45%. Patients who were drinking alcohol at one year were less likely to have been admitted with decompensated liver disease (43% vs 64%, p&lt;0.001) and had a lower MELD score at admission (median 10 vs 14 p = 0.001). They also had higher alcohol use at admission (median 100 vs 50 units/week, p&lt;0.001). These patients were more likely to be single (58% vs 36% p&lt;0.001), live alone (43% vs 27% p&lt;0.001), be unemployed (50% v 32% p = 0.002) and have a co-morbid psychiatric diagnosis (26% vs 15% p = 0.011). Those who had relapsed were also more likely to have missed at least two clinic appointments in the year after admission (32% vs 6% p&lt;0.001). After backwards stepwise elimination the following categorical variables remained statistically significant on multiple logistic regression for predicting relapse at one year: admission with decompensated liver disease (OR 0.27 p&lt;0.001) and poor clinic attendance (OR 4.33 p&lt;0.001). In addition, for every 10 units additional alcohol consumption/week at admission, likelihood of relapse increased by 4% (OR 1.04 p = 0.005). For each one unit increase in MELD, likelihood of relapse decreased by 6% (OR 0.94 p 0.006). <h3>Conclusion</h3> Heavier alcohol use, less severe liver disease and poor clinic attendance predict return alcohol use in patients with alcohol related cirrhosis.