BACKGROUND: Up to 30% of the cirrhotic patients may clinically be classified as having diabetes. The aims of this retrospective study were to evaluate the prevalence of diabetes, its association with the severity of cirrhosis, and subsequent implication on mortality in a broad population of cirrhotic patients in Taiwan.

METHODS: From January 2004 to October 2008, 4,127 cirrhotic patients were reviewed. The latest demographic data were gathered after the latest laboratory measurement in the outpatient or inpatient department. In all, 2,945 patients were included on the basis of the etiology of cirrhosis: hepatitis B, hepatitis C, hepatitis B, and hepatitis C co-infection, or alcoholics.

RESULTS: The observed cases of diabetes were significantly higher in cirrhotic patients than general population. Neither the Child-Pugh score (odds ratio = 0.99, p = 0.527) nor the model for end-stage liver disease score (MELD) (odds ratio = 1.00, p = 0.352) was associated with diabetes. The presence of diabetes was dissociated with renal function by means of serum creatinine (β = -0.01, p = 0.604) or estimated glomerular filtration rate (eGFR) (β = -0.02, p = 0.233), or in-hospital mortality (odds ratio = 1.08, p = 0.362).

CONCLUSION: The prevalence of diabetes was higher in cirrhotic patients in Taiwan. The prevalence of diabetes was dissociated with Child-Pugh or MELD score. The diagnosis of diabetes did not affect the renal function or in-hospital mortality in cirrhotic patients. Since the clinical outcomes of diabetes were different from cirrhotic patients to general population, it is unclear whether the treatment of diabetes and improved glycemic control will benefit cirrhotic patients.