In June of 2002, the National Institutes of Health (NIH) convened a consensus development conference on the management of hepatitis C.1 This conference differed from a previous conference2 on management of hepatitis C by including presentations from a detailed systematic review of evidence on key questions3–6 and presentations by leading experts in the field.
The Consensus Statement makes many important recommendations regarding the management of chronic hepatitis C, but we would like to highlight the evidence on two aspects of management that may not receive as much attention as the issue of treatment choice. These aspects are (1) liver biopsy—should all patients with chronic hepatitis C be referred for biopsy? and (2) hepatocellular carcinoma (HCC)—should patients with chronic hepatitis C be screened for this disease?
Because these are questions that arise when patients are initially diagnosed with hepatitis C, primary care physicians should be familiar with the Consensus Statement and its supporting evidence.
In patients with hepatitis C, liver biopsy provides information about fibrosis and histology that clinicians may use to determine prognosis and treatment. Liver biopsy also helps identify the presence of liver disease related to other causes. Experts at the Consensus Development Conference advocated for the use of liver biopsy primarily because biopsies may help in treatment decisions and allow for monitoring of disease progression, and have a low risk of serious complications (3 percent).7 In particular, patients have a low risk of progression to cirrhosis if they have persistently normal or slightly elevated alanine transaminase levels with minimal or no fibrosis on biopsy. Such patients may think that the potential adverse effects of antiviral treatment may outweigh the potential benefit of treatment, at least until more fibrosis develops. The patients' views may vary, depending on the hepatitis C virus (HCV) genotype, because the efficacy of antiviral treatment is much better in patients with HCV genotype 2 or 3 than in those with genotype 1.
To assess the utility of liver biopsy in predicting treatment response, we evaluated randomized trials of treatment for chronic hepatitis C infection where treatment assignment was not dependent on liver biopsy results and patients were followed for at least 24 weeks after treatment ended.5 We found 12 studies that reported at least some information on the relation of pretreatment histology to virologic or histologic outcomes of treatment. In addition, six studies reported enough data to enable an estimate of whether the difference in virologic response between treatment regimens depends on pretreatment histology.
These studies were relatively consistent in finding that the presence of advanced fibrosis or cirrhosis may predict a modest decrease in the likelihood of a virologic response to any of the interferon-based treatment regimens evaluated in the trials. The Consensus Development Conference Panel weighed this evidence along with the testimony of clinical experts and concluded that “liver biopsy is a useful part of the informed consent process regarding treatment.”1
The Panel emphasized that a biopsy allows patients to make a more informed choice about the initiation or postponement of treatment, although “the appropriate interval for subsequent evaluations is yet to be determined.”1 Because liver biopsy can cause morbidity and is expensive, some have questioned the utility of liver biopsy and have recommended alternatives to assess the extent of liver fibrosis.8,9 Therefore, we sought to identify biochemical or other serologic tests that accurately predict the degree of fibrosis seen on liver biopsy. We found that serum transaminase levels correlate only modestly with the degree of fibrosis.5 Although some evidence suggested that the ratio of aspartate transaminase to alanine transaminase was a relatively specific indicator of significant fibrosis, that ratio had poor sensitivity for detecting fibrosis.
New tests of extracellular matrix proteins, especially when combined into panels of tests, were better than transaminase levels at predicting the presence or absence of significant fibrosis. However, no serologic test, alone or in combination, was able to consistently identify patients with intermediate stages of fibrosis.5 Accordingly, the Panel indicated that noninvasive tests could not replace the information provided by a liver biopsy.
Hepatocellular Carcinoma Screening
The incidence of HCC in patients with hepatitis C ranges between zero and 1.6 percent per year.10 The risk of HCC is higher in patients with hepatitis C who have cirrhosis—estimates range from 1 to 6 percent per year.11–15 Mortality from HCC is substantial, with survival rates as low as 1 percent at two years in untreated patients.16,17 Nevertheless, screening for HCC in patients with hepatitis C infection has been a matter of controversy, because no randomized controlled trials have demonstrated the efficacy of screening in this population.
Despite the lack of trial evidence, many physicians screen patients for HCC with serum alpha-fetoprotein (AFP) levels or hepatic ultrasound examinations, believing that screening will improve survival. Although we were unable to identify any study demonstrating a screening or surveillance protocol that improved long-term survival, one study18 suggested that HCC was detected earlier and was resectable more often in patients who had twice-yearly screening with serum AFP determinations and hepatic ultrasonography than in those who had usual care, but no survival analysis was performed.
In addition, we evaluated studies that looked at performance characteristics of different screening tests, including serum AFP and ultrasonography. The studies demonstrated that the sensitivity and specificity of a serum AFP level in identifying HCC depends on the thresholds used.4 For example, a threshold range from 10 to 19 ng per mL (10 to 19 mg per L) had a sensitivity of 75 to 80 percent and a specificity of 65 to 80 percent, while a threshold value of 400 ng per mL (400 mg per L) had a sensitivity of 4 to 48 percent, with a specificity of about 100 percent. Ultrasound studies reported variable sensitivities, between 11 and 100 percent, with a high degree of specificity.4
Therefore, the Panel concluded that the value of screening for HCC is uncertain because no data demonstrate the clinical impact of this screening on the management of HCC or associated mortality.1 The Panel also concluded that hepatic ultrasonography is more sensitive than serum AFP testing but is more expensive and can lead to invasive and unnecessary follow-up tests. The Panel recommended that routine screening for HCC not be performed in patients with hepatitis C in the absence of cirrhosis because HCC is so rare in this group.
Because there were significant limitations of the evidence for both the liver biopsy and the HCC screening issues, the Panel used both clinical expert opinion and published evidence to arrive at these conclusions. Perhaps the most important recommendations are that further studies should be done, particularly to “more clearly establish the role of liver biopsy in the therapeutic management of patients with hepatitis C” and “to assess screening tests in patients at greatest risk of HCC.”1