The tumor suppressor gene p16INK4A is mainly inactivated by an epigenetic change involving promoter hypermethylation in hepatocarcinogenesis. The possible clinical impact of p16INK4A methylation and the potential risk factors for this epigenetic alteration have not been thoroughly investigated. We studied the methylation status and mRNA and protein expression of p16INK4A in 50 hepatocellular carcinomas and corresponding nonneoplastic liver lesions using methylation-specific PCR, reverse transcription-PCR, and immunohistochemical techniques. p16INK4A hypermethylation was observed in 58% (29 of 50) of the hepatocellular carcinomas and 16% (6 of 38) of the corresponding chronic hepatitis and cirrhosis tissue samples. p16INK4A methylation was significantly associated with mRNA and protein expression (P <0.001 and P=0.003, respectively). All of the p16INK4A-methylated tumors were positive for hepatitis B virus or hepatitis C virus markers, but none of the virus-negative tumors exhibited p16INK4A methylation (P=0.006). The frequency of p16INK4A hypermethylation tended to be higher in hepatitis C virus-related tumors (23 of 32, 72%) than in hepatitis B virus-related tumors (6 of 13, 46%; P=0.1). Aberrant methylation of p16INK4A was also related significantly to increasing age, female gender, and normal levels of serum PIVKA-II (P=0.02, 0.04, and 0.04, respectively). No statistically significant difference in survival was observed between patients with p16INK4A hypermethylation and those without. Our observations suggest that p16INK4A hypermethylation may contribute to hepatocarcinogenesis from an early stage and that multiple risk factors, such as viral infections, age, and gender, may be associated with p16INK4A hypermethylation in hepatocarcinogenesis.