Relative and absolute reliability of measures of linoleic acid-derived oxylipins in human plasma

Modern analytical techniques allow for the measurement of oxylipins derived from linoleic acid in biological samples. Most validatory work has concerned extraction techniques, repeated analysis of aliquots from the same biological sample, and the influence of external factors such as diet and heparin treatment upon their levels, whereas less is known about the relative and absolute reliability of measurements undertaken on different days. A cohort of nineteen healthy males were used, where samples were taken at the same time of day on two occasions, at least 7 days apart. Relative reliability was assessed using Lin's concordance correlation coefficients (CCC) and intraclass correlation coefficients (ICC). Absolute reliability was assessed by Bland-Altman analyses. Nine linoleic acid oxylipins were investigated. ICC and CCC values ranged from acceptable (0.56 [13-HODE]) to poor (near zero [9(10)- and 12(13)-EpOME]). Bland-Altman limits of agreement were in general quite wide, ranging from ±0.5 (12,13-DiHOME) to ±2 (9(10)-EpOME; log10 scale). It is concluded that relative reliability of linoleic acid-derived oxylipins varies between lipids with compounds such as the HODEs showing better relative reliability than compounds such as the EpOMEs. These differences should be kept in mind when designing and interpreting experiments correlating plasma levels of these lipids with factors such as age, body mass index, rating scales etc.