OBJECTIVE We designed a randomized trial to examine whether increase or preservation of serum albumin levels was attained with administration of branched-chain amino acid (BCAA) granules for compensated cirrhosis, compared with decompensated cirrhosis. METHODS Sixty-five patients with HCV-related cirrhosis with serum albumin level less than 4.0 g/dl were enrolled in this study. Half of the patients were randomly assigned to receive 14.22 g/day of BCAA granules orally, and half were assigned to a control group. Patients were evaluated at entry and at 1-year intervals for at least 2 years. The parameters were divided into 3 categories. Class 1 was decompensated cirrhosis with serum albumin level less than 3.5 mg/dl. Class 2 was compensated cirrhosis with serum albumin level over 3.6 mg/dl and molar ratio of BCAA to tyrosine (BTR) less than 4. Class 3 was compensated cirrhosis with serum albumin level over 3.6 mg/dl and BTR over 4. RESULTS In class 1 and class 2, the BCAA group exhibited significantly higher rates of maintaining serum albumin level than the control group for 2 years. In contrast, there was no significant difference between the BCAA group and control group in rate of maintaining serum albumin levels in class 3. CONCLUSIONS Those results suggested that if cirrhotic patients were in the compensated stage at the entry but with lower BTR, as for decompensated cirrhosis, oral BCAA supplementation might be effective in maintaining serum albumin level for 2 years.