It has been suggested that the dietary intake of omega-3 polyunsaturated fatty acids could be inversely related to the risk of dementia and cognitive decline. This analysis examined the association between plasma concentration of omega-3 polyunsaturated fatty acids and prevalence and incidence of cognitive impairment and dementia. Data are reported on subjects 65 years or older who had a complete clinical evaluation at the first two waves (1991-1992 and 1996-1997) of the Canadian Study of Health and Aging. Main outcome measures were cognitive impairment and dementia by mean relative plasma concentrations of fatty acids in the phospholipid fraction at baseline. Results were adjusted for age, sex, education, smoking, alcohol intake, body mass index, history of cardiovascular disease, and apolipoprotein E e4 genotype. In the cross-sectional analysis, no significant difference in omega-3 polyunsaturated fatty acid concentrations was observed between controls and both prevalent cases of cognitive impairment and dementia. In the prospective analysis, a higher eicosapentaenoic acid (p < 0.01) concentration was found in cognitively impaired cases compared to controls while higher docosahexaenoic acid (p < 0.07), omega-3 (p < 0.04) and total polyunsaturated fatty acid (p < 0.03) concentrations were found in dementia cases. These findings do not support the hypothesis that omega-3 polyunsaturated fatty acids play a protective role in cognitive function and dementia.