Lack of interaction between hepatitis C virus and alcohol in the pathogenesis of cirrhosis. A statistical study.

Background/aims: In several studies markers of hepatitis C virus infection have been shown to be present in alcoholic patients with cirrhosis. Our work was designed to test the likely hypothesis that this association is due to an interaction between hepatitis C virus and alcohol in the pathogenesis of cirrhosis.

Methods: We compared alcohol consumption and repartition of anti-HCV antibodies detected by an immunoblot recombinant assay in 101 male patients with cirrhosis and in 120 male controls. Interactions between anti-hepatitis C virus, alcohol and cirrhosis were calculated using log linear hierarchical models for frequency data. The basis of the method is that an interaction between hepatitis C virus and alcohol implies that a model built on the hypothesis of a role of hepatitis C virus and alcohol in the disease should be improved by a coefficient associated with multiplicative effects of hepatitis C virus and alcohol.

Results: In patients with cirrhosis the mean alcohol consumption (148 +/- 100 g per day) and the incidence of positivity for anti-HCV antibodies (45%) were significantly higher than in controls. The results were consistent with a theoretical model built with the hypothesis of an independent role of both alcohol and hepatitis C virus. The goodness of fit between this model and the actual distribution of alcohol consumption and hepatitis C virus markers was not improved by introduction of an interaction between hepatitis C virus and alcohol.

Conclusions: In alcoholic subjects with hepatitis C virus infection, the probability to have cirrhosis seemed to be explained by additive effects of alcohol and hepatitis C virus. From a purely statistical point of view, no interaction between hepatitis C virus and alcohol consumption on a multiplicative scale could be demonstrated.