Evidence Summaries

A systematic review ¹ including 5 studies with a total of 1,734 subjects was abstracted in DARE. Of the 5 included studies, one did not use an appropriate gold standard for case patients, and two did not for control patients. Only one study reported blinding investigators to index test results, and two that consecutive patients were recruited. Four studies were prone to partial verification bias, while in one study verification bias was not applicable since US was used as the gold standard, which was considered unacceptable. Cut-off at 20 microg/L AFP (5 studies). The sensitivity ranged from 41 to 65%, the specificity from 80 to 94%, the positive LR from 3.1 to 6.8, and the negative LR from 0.4 to 0.6. Cut-off at 200 microg/L AFP (4 studies). The sensitivity ranged from 20 to 45%, the specificity from 99 to 100%, the positive LR from 29 to infinity, and the negative LR from 0.6 to 0.8.

Comment: The quality of evidence is downgraded by poor quality of the included studies.

References

1. Gupta S, Bent S, Kohlwes J. Test characteristics of alpha-fetoprotein for detecting hepatocellular carcinoma in patients with hepatitis C. A systematic review and critical analysis. Ann Intern Med 2003 Jul 1;139(1):46-50.  [PMID:12834318]