Table 3.

Prediction of Plasma Phospholipid Total Very-Long-Chain Omega-3 Fatty Acid Content From Estimated Dietary Intake of EPA + DHA in Multivariate Models^a, According to MMSE Score Cutoff and Clinical Consensus Diagnosis, NAME Study, Boston, Massachusetts, 2002–2008

	No. of Subjects	Model Statistic
		β Estimate	P Value	Model R2
All subjects	273	0.157	<0.0001	0.273
Cognitive functionb
High	186	0.146	<0.0001	0.303
Low	87	0.177	<0.0001	0.282
Clinical diagnosisc
Normal	129	0.171	<0.0001	0.287
Mild cognitive impairment	82	0.137	0.0006	0.273
Dementia	62	0.160	<0.0001	0.403

Abbreviations: DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; MMSE, Mini-Mental State Examination; NAME, Nutrition, Aging, and Memory in Elders.

^aAll results were adjusted for age, sex, race, total energy intake, and home-care agency. The dependent variable was plasma phospholipid EPA + DPA + DHA (mol%) concentration; the independent predictor was estimated dietary EPA + DHA intake (mg/day). Data on both plasma and dietary fatty acid variables were log(e)-transformed.

^bHigh and low cognitive function were defined by MMSE scores of >24 points and ≤24 points, respectively.

^cClinical diagnosis of normal functioning, mild cognitive impairment, or dementia as determined by consensus diagnostic criteria.