The efficacy of lamivudine for HBeAg-negative chronic hepatitis B (CHB) Chinese patients has not been fully investigated. The role of the Hepatitis B virus (HBV) genotype on the treatment effect of lamivudine is controversial. Thirty-two consecutive patients with HBeAg-negative CHB were enrolled. All patients were treated with lamivudine 100 mg once daily of 7-12 months duration. The mean total period of follow-up since entry for all patients was 24 +/- 3.5 months. HBV genotypes were classified by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and verified by sequencing. Precore (G1896A) and basic core promoter (BCP, A1762T & G1764A) mutations were determined by PCR and direct sequencing. Twenty-one (65.6%) patients were infected by genotype B and, 11 (34.4%) by genotype C. G1896A was predominant in genotype B infected patients (95.2%vs 63.6%, P = 0.037). At the end of treatment, 31 (96.8%) and 14 (43.8%) patients achieved biochemical and virological responses, respectively. The biochemical and virological response rates were 40.6 and 0% at 12 months after treatment. Eighteen (56.3%) patients had biochemical relapse within 12 months after withdrawal of lamivudine. By multivariate analysis, the pretreatment serum level of HBV DNA >/=12 Meq/mL was the only factor associated with early biochemical relapse (Odds ratio = 9.333, 95% CI = 1.497 approximately 58.197, P = 0.017). In conclusion, the virological effect of lamivudine for HBeAg-negative CHB is transient. Most patients had biochemical relapse within 12 months after lamivudine treatment regardless of HBV genotype. A high pretreatment viral load is the determinant for early biochemical relapse.