BACKGROUND AND AIMS This study evaluated the clinical significance of subclinical ascites in patients with hepatitis B virus-related cirrhosis treated with lamivudine (LMV) or entecavir (ETV).

METHODS This multicenter retrospective study involved 8 hospitals. Patients were classified by degree of ascites: (1) no ascites (no ascites on imaging, no diuretics), (2) subclinical ascites (small amount of ascites on imaging, no diuretics), and (3) clinical ascites (moderate to severe ascites or diuretics).

RESULTS Out of 501 patients, 336 (68%), 51 (10%), and 114 (23%) patients were classified as no-ascites, subclinical ascites, and clinical ascites, respectively. In all, 100 (20%) and 401 (80%) were treated with LMV and ETV, respectively. Over 58±24 months of follow-up, 105 patients (21%) developed hepatocellular carcinoma. The cumulative incidence of hepatocellular carcinoma did not differ between LMV-treated and ETV-treated patients (P=0.61); it was higher in the clinical-ascites group than the no-ascites (P=0.054) and subclinical-ascites (P=0.03) groups, but it was comparable between the latter 2 (P=0.225). Forty-five patients (9%) died during follow-up. Survival was significantly shorter in the clinical-ascites group than the other 2 (both P<0.005), but it was comparable between no-ascites and subclinical-ascites groups (P=0.444). Multivariate analysis showed that mortality was significantly associated with prothrombin time [hazard ratio (HR)=2.42; 95% confidence interval (CI), 1.59-3.70], serum albumin (HR=0.54; 95% CI, 0.29-0.99), and presence of clinical ascites (HR=3.58; 95% CI, 1.54-8.30).

CONCLUSIONS Subclinical ascites did not affect prognosis in patients with hepatitis B virus-related cirrhosis receiving antiviral treatment.