Background: Patients with hepatitis C virus (HCV) are at increased risk for hepatocellular carcinoma. Although serum alpha-fetoprotein (AFP) is often used to detect hepatocellular carcinoma in HCV-infected individuals, its utility is unclear.
Purpose: To define the test characteristics of AFP for the diagnosis of hepatocellular carcinoma in patients with HCV.
Data sources: MEDLINE search from 1966 to December 2002 for English- and non-English-language articles examining the test characteristics of AFP for identifying hepatocellular carcinoma.
Study selection: Articles were included if they reported the sensitivity and specificity of AFP for detecting hepatocellular carcinoma in patients with HCV. Articles were excluded if the cause of hepatitis was ambiguous or if 50% or more of the study patients did not have HCV.
Data extraction: Relevant articles were evaluated for quality of evidence; test characteristics were abstracted and calculated.
Data synthesis: Five studies met all inclusion criteria and were analyzed. The overall quality of evidence was limited; only one study universally applied an acceptable gold standard test, and three of five studies used a case-control design that potentially overestimates diagnostic accuracy. By using the most commonly reported cutoff value of a positive test result for hepatocellular carcinoma (AFP level > 20 microg/L), the ranges of test characteristics were as follows: sensitivity, 41% to 65%; specificity, 80% to 94%; positive likelihood ratios, 3.1 to 6.8; and negative likelihood ratios, 0.4 to 0.6.
Conclusions: The paucity of high-quality data calls for more rigorous study of AFP and other diagnostic tests for detecting hepatocellular carcinoma in HCV-infected patients with an accepted gold standard applied to the entire cohort. Even if the "best-case" estimates of AFP sensitivity and specificity are accurate, AFP has limited utility for detecting hepatocellular carcinoma.