Hospital mortality over time in patients with specific complications of cirrhosis

Abstract

Hospital mortality secondary to cirrhosis is high.

Aim: To evaluate hospital mortality in patients admitted for specific complications of cirrhosis over time.

Material and methods: Registry-data from Administrative Inpatient Dataset of acute care hospitals were collected at discharge from 2003 to 2010. Inclusion criteria were as follows: hospital admissions where one of the diagnoses was cirrhosis and the reason for admission was a specific complication of cirrhosis (ascites, encephalopathy, hepatorenal syndrome and haemorrhage from varices, bacterial spontaneous peritonitis). Analysis of variance was used for comparisons of quantitative variables and Chi-square for qualitative variables. Logistic regression was performed to identify the risk factors associated with hospital mortality; the Hosmer and Lemeshow test was applied to evaluate calibration and the ROC curve for discrimination respectively.

Results: A total of 12,671 hospital admissions were analysed; 67.7% were men. Mean hospitalization stay was 10.9 (SD 9.2) days and the most frequent causes were encephalopathy (44.2%) and ascites (30.9%). Global hospital mortality was 11.6%. Logistic regression showed that once all factors had been adjusted, hepatorenal syndrome conveyed the highest risk for death (49.2%; OR = 8.1(95%CI:6.6-9.9). Risk of death was also increased by associated comorbidities and older age. Hospital mortality in the period 2006-2010 was 27% inferior to the period 2003-2005. The area under the ROC curve (AUROC) was 0.77 (95%CI 0.76-0.78).

Conclusions: Hospital mortality as a result of specific complications of cirrhosis is high, but has been declining in recent years.