Effect of interferon-alpha on progression of cirrhosis to hepatocellular carcinoma: a retrospective cohort study. International Interferon-alpha Hepatocellular Carcinoma Study Group

Abstract

Background: There is debate about whether interferon-alpha treatment lowers the risk of progression to hepatocellular carcinoma in patients with chronic viral hepatitis and cirrhosis and whether any effect is limited to certain subgroups. We investigated these issues by retrospective analysis of data for 913 patients from Italy and Argentina.

Methods: 21 centres reported patients from their records who had chronic viral hepatitis and Child's A cirrhosis, were positive for HBsAg or hepatitis-C-virus antibodies (anti-HCV), and had been screened yearly for at least 3 years by ultrasonography and alpha-1-fetoprotein testing. Prognostic risk factors for hepatocellular carcinoma defined by multivariate Cox regression analysis and individual observation time were used for group matching and conditional logistic regression analysis of the independent interferon-alpha treatment effect.

Findings: After group matching, the number of patients was reduced to 637. Age, male sex, and portal hypertension were significant risk factors for hepatocellular carcinoma (each p < 0.001); hepatic inflammation (p = 0.21) and iron storage (p = 0.18) were also included in the model 66 (19%) of 356 untreated patients and 29 (10%) of 281 treated patients developed hepatocellular carcinoma (relative risk 1.99 [95% CI 1.09-3.64]); the corresponding proportions for anti-HCV-positive patients were 48 (18.5%) of 259 versus 21 (9.1%) of 232 (3.14 [1.46-6.80]), and those for hepatitis-B-virus-infected (HBV) patients were 18 (10%) of 97 and eight (16%) of 49 (0.98 [0.33-2.92]). Among anti-HCV patients without HBV markers, 29 (20%) of 129 untreated and six (5%) of 116 treated patients developed hepatocellular carcinoma (6.28 [1.65-23.8]).

Interpretation: Interferon treatment lowered the rate of progression to hepatocellular carcinoma two fold. The risk reduction was apparently greater for patients with chronic hepatitis C and no evidence of HBV infection. Future studies should stratify HCV-infected patients by HBV status.