Analysis of long-chain n-3 and n-6 fatty acid (FA) concentrations is used to evaluate their potential health effects in epidemiological studies, and, recently, also to counsel patients with a suboptimal intake of n-3 FA. Data on the method's ability to track and detect differences within and between individuals in appropriate populations are, however, lacking. The present study provides such data for twenty-nine plasma phospholipid (PL) FA concentrations and indices measured in 214 newly diagnosed type 2 diabetic patients at baseline and after 3 years. 20 : 3n-6 and the 20 : 4n-6:20 : 3n-6 ratio showed the highest tracking coefficients (Spearman's r 0·68), while DHA, EPA and PLN3-index (EPA+DHA) coefficients were 0·60, 0·47 and 0·55, respectively. Fish consumption measured simultaneously with EPA, DHA, sum n-3 and PLN3 index showed Spearman's correlation coefficients of 0·47, 0·44, 0·48 and 0·49, respectively, decreasing to 0·20, 0·19, 0·22 and 0·21 when measured 3 years apart. The within-subject CV of EPA, DHA and PLN3 index were 39·9, 14·3 and 18·0 %, respectively. The corresponding between-subject CV were 33·6, 16·5 and 18·7 %, while the reference change values were 112, 41 and 52 %. In conclusion, PL n-3 FA concentrations showed a significant long-term tracking and were positively correlated with marine food intake. Analytical precision, biological variability, reference change value and the index of individuality of EPA, DHA and PLN3 index are similar to commonly used clinical biomarkers, supporting their validity as dietary markers in clinical and epidemiological work.