Premium
Fish to meat intake ratio and cooking oils are associated with hepatitis C virus carriers with persistently normal alanine aminotransferase levels
Author(s) -
Otsuka Momoka,
Uchida Yuki,
Kawaguchi Takumi,
Taniguchi Eitaro,
Kawaguchi Atsushi,
Kitani Shingo,
Itou Minoru,
Oriishi Tetsuharu,
Kakuma Tatsuyuki,
Tanaka Suiko,
Yagi Minoru,
Sata Michio
Publication year - 2012
Publication title -
hepatology research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.123
H-Index - 75
eISSN - 1872-034X
pISSN - 1386-6346
DOI - 10.1111/j.1872-034x.2012.01014.x
Subject(s) - alanine transaminase , medicine , odds ratio , alanine aminotransferase , aspartate transaminase , hepatitis c virus , transaminase , receiver operating characteristic , gastroenterology , food science , virus , biology , immunology , alkaline phosphatase , biochemistry , enzyme
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.