Translating plasma and whole blood fatty acid compositional data into the sum of eicosapentaenoic and docosahexaenoic acid in erythrocytes.

Specific blood levels of eicosapentaenoic plus docosahexaenoic acid (EPA+DHA, wt% of total) in erythrocytes or "the omega-3 index" have been recommended for cardio-protection, but fatty acids are often measured in different blood fractions. The ability to estimate the % of EPA+DHA in erythrocytes from the fatty acid composition of other blood fractions would enable clinical assessments of omega-3 status when erythrocyte fractions are not available and increase the ability to compare blood levels of omega-3 fatty acids across clinical studies. The fatty acid composition of baseline plasma, erythrocytes and whole blood samples from participants (n=1104) in a prospective, multicenter study examining acute coronary syndrome were determined. The ability to predict the % of EPA+DHA in erythrocytes from other blood fractions were examined using bivariate and multiple linear regression modelling. Concordance analysis was also used to compare the actual erythrocytes EPA+DHA values to values estimated from other blood fractions. EPA+DHA in erythrocytes was significantly (p<0.001) correlated EPA+DHA in plasma (r(2)=0.54) and whole blood (r(2)=0.79). Using multiple linear regression to predict EPA+DHA in erythrocytes resulted in stronger coefficients of determination in both plasma (R(2)=0.70) and whole blood (R(2)=0.84). Concordance analyses indicated agreement between actual and estimated EPA+DHA in erythrocytes, although estimating from plasma fatty acids appears to require translation by categorization rather than by translation as continuous data. This study shows that the fatty acid composition of different blood fractions can be used to estimate erythrocyte EPA+DHA in a population with acute coronary syndrome.