Uncontrolled studies have suggested a beneficial effect of lamivudine in patients with decompensated cirrhosis caused by replicating hepatitis B virus (HBV). We analyzed the outcome of lamivudine treatment in 23 consecutive patients with severely decompensated HBV-cirrhosis defined as a Child-Pugh-Turcotte (CPT) score of ≥10, and compared with a historical untreated control group of 23 patients matched for age, gender, and baseline CPT score. Significant clinical response, defined as a decrease in the CPT score by ≥3 points, was observed in 14 of 23 (60.9%) treated patients versus none of the controls (P < .0001). The median change in CPT scores was −3.0 (range, −6 to +3) in the treated group versus +1.0 in the controls (range, −1 to +2) (P = .016). Orthotopic liver transplantation (OLT) was performed in 34.8% of treated patients (median, 3.5; range, 1-32 months), versus 73.9% of controls (median, 3.0; range, 1-14 months) (P = .04). Excluding transplanted patients, there were no deaths in the treated group versus 6 deaths in the control group (P = .009). Time to death or OLT was significantly longer in treated patients than in controls (P < .001). Two patients developed lamivudine resistance after 9 and 12 months, respectively. Our results suggest that lamivudine significantly improves hepatic function in over half of the patients with decompensated cirrhosis and replicating HBV, and may confer a survival advantage. However, the small sample size and the use of a retrospective control cohort preclude drawing definitive conclusions. Expedited OLT remains the only viable treatment for lamivudine nonresponders.