In the context of a case-control study undertaken in Greece, we examined the role of six flavonoid classes in the etiology of hepatocellular carcinoma (HCC), by viral status, and of cholangiocarcinoma (CAC). Data and blood samples were collected between 1995 and 1998. Information about dietary intakes and covariates, including chronic infection with hepatitis B (HBV) and C (HCV) virus, were available for 250 HBV and/or HCV positive HCC cases, 83 HBV and HCV negative HCC cases, six CAC cases, and 360 hospital controls. In logistic regression models including gender, age, education, tobacco smoking, and total energy intake, there were no distinct patterns with respect to either HCC virus positive and HCC virus negative in relation to total flavonoids or any class of flavonoids, with the exception of flavones. Flavone intake, mostly derived from spinach and peppers, was inversely associated with both virus positive (P-trend, 0.049) and virus negative (P-trend, 0.084) HCC. There was also a suggestion of an inverse association of CAC with flavan-3-ols, anthocyanidins, and total flavonoids which, however, has to be taken with due caution on account of the small number of cases of this rare tumor. We conclude that flavones may be inversely associated with HCC risk, irrespective of its dominant etiology (viral or non viral).