Fish to meat intake ratio and cooking oils are associated with hepatitis C virus carriers with persistently normal alanine aminotransferase levels.

Aim:   Dietary habits are involved in the development of chronic inflammation; however, the impact of dietary profiles of hepatitis C virus carriers with persistently normal alanine transaminase levels (HCV-PNALT) remains unclear. The decision-tree algorithm is a data-mining statistical technique, which uncovers meaningful profiles of factors from a data collection. We aimed to investigate dietary profiles associated with HCV-PNALT using a decision-tree algorithm.
Methods:   Twenty-seven HCV-PNALT and 41 patients with chronic hepatitis C were enrolled in this study. Dietary habit was assessed using a validated semiquantitative food frequency questionnaire. A decision-tree algorithm was created by dietary variables, and was evaluated by area under the receiver operating characteristic curve analysis (AUROC).
Results:   In multivariate analysis, fish to meat ratio, dairy product and cooking oils were identified as independent variables associated with HCV-PNALT. The decision-tree algorithm was created with two variables: a fish to meat ratio and cooking oils/ideal bodyweight. When subjects showed a fish to meat ratio of 1.24 or more, 68.8% of the subjects were HCV-PNALT. On the other hand, 11.5% of the subjects were HCV-PNALT when subjects showed a fish to meat ratio of less than 1.24 and cooking oil/ideal bodyweight of less than 0.23 g/kg. The difference in the proportion of HCV-PNALT between these groups are significant (odds ratio 16.87, 95% CI 3.40-83.67, P = 0.0005). Fivefold cross-validation of the decision-tree algorithm showed an AUROC of 0.6947 (95% CI 0.5656-0.8238, P = 0.0067).
Conclusion:   The decision-tree algorithm disclosed that fish to meat ratio and cooking oil/ideal bodyweight were associated with HCV-PNALT.
(© 2012 The Japan Society of Hepatology.)