Two hundred ninety patients (203 men, 87 women), age 7 to 74 years (mean: 39.1 years), with chronic hepatitis B virus infection, were prospectively followed for a period of 1 to 4 years to determine the value of α-fetoprotein monitoring in the early detection of hepatocellular carcinoma. At presentation, 66% of the patients were asymptomatic, 19% had chronic hepatitis and 15% had established cirrhosis.

Forty-four (15%) patients had elevated α-fetoprotein levels on one or more occasions during the study period. Twenty patients with normal α-fetoprotein levels at presentation developed elevated α-fetoprotein levels during the course of follow-up, whereas 24 patients had elevated α-fetoprotein levels at presentation. Six (14%) of these 44 patients (five men and one woman), age 23 to 66 years, had persistent or progressive increase in α-fetoprotein levels and were confirmed to have hepatocellular carcinoma. In four patients, the α-fetoprotein levels were below 500 ng per ml at the time of tumor localization. Only three patients had resectable tumors. All six patients would have been missed if α-fetoprotein screening was restricted to men above the age of 40 with cirrhosis and anti-HBe. Of the remaining 38 patients, elevation in α-fetoprotein levels in 18 patients was associated with exacerbations of the underlying liver disease and/or significant changes in level of hepatitis B virus replication, but in 20 patients, no apparent cause could be identified. The elevation in AFP levels exceeded 200 ng per ml in 26% and persisted beyond 6 months in 15% of these patients.

We observed that elevation in α-fetoprotein levels occurred quite frequently in patients with chronic hepatitis B virus infection. The differentiation between benign and malignant causes of α-fetoprotein elevation was at times difficult. Thus, α-fetoprotein monitoring alone is not a satisfactory tool in the early diagnosis of hepatocellular carcinoma.