Dietary intake of fish rich in n-3 highly unsaturated fatty acids (HUFAs), such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), has been proposed to decrease cancer risk. In contrast to results from laboratory studies, however, protective effects for breast cancer have proved equivocal in epidemiological studies. In the present case-control study, we examined associations between breast cancer risk and fatty acid compositions in erythrocyte membranes as biomarkers for those intakes. Dietary information and blood samples were collected from 103 incident breast cancer cases and 309 non-cancer controls (matched by age and season) and erythrocyte fatty acids were measured using accelerated solvent extraction and gas-liquid chromatography. Dietary intake of n-3 HUFAs demonstrated a negative association with risk (the highest to the lowest tertile, odds ratio (OR), 0.51; 95% confidence interval (CI), 0.27-0.98; p trend < 0.05), but there was no association with those of saturated fatty acids (SFAs) and meat. Moreover, risk was inversely associated with erythrocyte compositions of EPA (OR, 0.27; 95% CI, 0.14-0.53; p trend < 0.0001), DHA (OR, 0.06; 95% CI, 0.02-0.16; p trend < 0.0001) and n-3 HUFAs (OR, 0.11; 95% CI, 0.05-0.24; p trend < 0.0001), and positively with that of SFAs (OR, 12.29; 95% CI,; p trend < 0.0001) and the ratio of SFAs/n-3 HUFAs (OR, 14.65; 95% CI,; p trend < 0.0001). In conclusion, we showed that erythrocyte compositions of specific fatty acids derived from fish intake, as biomarkers, are associated with lower risk of breast cancer, but further studies are needed to investigate mechanisms linked to the etiology. ' 2007 Wiley-Liss, Inc.