Intra- and inter-individual variability in measurements of biomarkers for oxidative damage in vivo: Nutrition and Breast Health Study.

Oxidative stress has been implicated in the pathogenesis of various chronic diseases, such as cancer, cardiovascular disease and inflammatory conditions, as well as in ageing. Although a number of markers are now available, little is known about the reliability of single measurements of such markers in healthy individuals. The study examined the distribution of variance for three oxidative stress markers, 8-oxo-2'-deoxyguanosine (8-oxodG), 5-hydroxymethyl-2'-deoxyuridine (5-OHmdU) and total 8-isoprostane-F2alpha, which were measured every 3-6 months over 1 year in blood and breast nipple aspirate fluid (NAF) for 103 premenopausal women. For both plasma and NAF, the between-subject variances of 8-isoprostane-F2alpha were consistently greater than the within-subject variances. Consequently, their reliability coefficients were close to the level of those for cholesterol. On the other hand, the within-subject variances were much greater than the between-subjects variances for blood 5-OHmdU, resulting in low reliability coefficients, i.e. <0.3. Overall, the reliability coefficients for blood 8-oxodG were between those of 8-isoprostane-F2alpha and 5-OHmdU, but closer to those of 8-isoprostane-F2alpha. The results suggest that the reliability of oxidative stress markers may vary considerably depending on the type of marker. Caution should be exercised in selecting markers as well as in determining the number of study subjects or the number of samples per subject in a study. There also may be ample room to optimize laboratory techniques to quantify markers of oxidative DNA damage.