Journal of Epidemiology
Online ISSN : 1349-9092
Print ISSN : 0917-5040
ISSN-L : 0917-5040
Original Article
Intake Frequency of Fish and Serum Levels of Long-chain n-3 Fatty Acids: A Cross-sectional Study within the Japan Collaborative Cohort Study
Kenji WakaiYoshinori ItoMasayo KojimaShinkan TokudomeKotaro OzasaYutaka InabaKiyoko YagyuAkiko Tamakoshi
Author information

2005 Volume 15 Issue 6 Pages 211-218


BACKGROUND: Several investigations have adopted self-reported intake frequency of fish as a surrogate for intake of long-chain n-3 fatty acids, for which protective effects against cancer have been suggested. Whether reported fish consumption reflects the fatty acid intake, however, has to be elucidated.
METHODS: We examined the association between intake frequency of fish and serum long-chain n-3 fatty acids (weight percentage of total fatty acids) among 1,257 control subjects (631 men and 626 women), aged 40-79 years, in case-control studies nested in the Japan Collaborative Cohort Study. All the subjects were not fasting when blood was drawn. Serum fatty acids were determined by gas chromatography.
RESULTS: In men, intake frequency of fresh fish and dried or salted fish was significantly but weakly correlated with serum levels of eicosapentaenoic (EPA), docosapentaenoic (n-3) (DPA), and docosahexaenoic (DHA) acids; the age-adjusted Spearman correlation coefficients ranged from 0.11 to 0.18. In women, fresh fish consumption was somewhat associated with serum EPA (Spearman correlation coefficient = 0.12) as was dried or salted fish consumption with serum DPA (0.11). A rising trend in geometric means of serum EPA, DPA, and DHA was found with an increasing intake frequency of fresh or dried/salted fish in both sexes. The geometric means adjusted for age and participating institution in the highest intake category were higher than those in the lowest by 7% to 40%.
CONCLUSIONS: A population with high self-reported frequency of fish intake, as a group, may have higher bioavailability of long-chain n-3 fatty acids than one with low frequency.
J Epidemiol 2005; 15: 211-218.

Information related to the author
© 2005 by Japan Epidemiological Association
Previous article Next article