Lamivudine (LAM) has been shown to prevent de novo hepatitis B virus (HBV) infection in recipients of hepatitis B core antibody (HBcAb)-positive liver transplants (LT) but primarily in small studies with limited follow-up. We conducted a retrospective cohort study of HBcAb+ graft recipients at our institution from October 1999 to August 2008. One hundred nineteen recipients without prior HBV were identified (median age, 54 years; 70% male), of which 62 received LAM. The median follow-up was 2.6 years overall and 5.3 years in the LAM group. Among LAM recipients, 44% were HBV naïve (HBsAb-/HBcAb-) at LT, of which 6% developed HBsAb+ and 3% developed HBcAb+ after LT. Eight percent developed de novo HBV: two recipients became hepatitis B surface antigen positive at 70 and 23 months and three experienced breakthrough with HBV DNA more than 2000 IU at 1 to 9 months after LT. Sixty percent (3 of 5) were HBV naïve. Four (6%) other recipients also had transiently detectable HBV less than 2000 IU, which did not require any changes to their prophylaxis regimen. When compared with recipients who received other nucleos(t)ide analogues, there was no difference in de novo rates: LAM 8% (5 of 62), adefovir 15% (5 of 33), tenofovir 0% (0 of 3), entecavir 0% (0 of 1), and 5% (1 of 20) for those not given prophylaxis (P=0.59). LAM monoprophylaxis was effective in preventing de novo HBV in the vast majority of recipients over long-term follow-up. Adefovir had a higher rate of de novo infections numerically, whereas tenofovir and entecavir had no cases and may be more effective, but this was limited by a small sample size.