Background and objective Cirrhosis is the number one cause of non-cancer deaths among gastrointestinal diseases and is responsible for significant morbidity and healthcare utilisation. The objectives were to measure the 30-day readmissions rate following index hospitalisation, to determine the predictors of readmission, and to estimate the cost of 30-day readmission in patients with decompensated cirrhosis.

Methods We performed a retrospective cohort study of patients with decompensated cirrhosis using 2014 Nationwide Readmission Database from January to November. Decompensated cirrhosis was identified based on the presence of at least one of the following: ascites, hepatic encephalopathy, variceal bleeding, spontaneous bacterial peritonitis and hepatorenal syndrome. We excluded patients less than 18 years of age, pregnant patients, patients with missing length of stay data, and those who died during the index admission.

Results Among 57 305 unique patients with decompensated cirrhosis, the 30-day readmission rate was 23.2%. The top three predictors of 30-day readmission were leaving against medical advice (AMA), ascites and acute kidney injury, which increased the risk of readmission by 47%, 22% and 20%, respectively. Index admission for variceal bleeding was associated with a lower 30-day readmission rate by 18%. The estimated total cost associated with 30-day readmission in our study population was US$234.4 million.

Conclusion In a nationwide population study, decompensated cirrhosis is associated with a 30-day readmission rate of 23%. Leaving AMA, ascites and acute kidney injury are positively associated with readmission. Targeted interventions and quality improvement efforts should be directed toward these potential risk factors to reduce readmissions.