A study of inter- and intrasubject variability in seven-day weighed dietary intakes with particular emphasis on trace elements.

In this study of inter- and intrasubject variance of trace element (TE) dietary intake data, 14 female university students recorded weighed, self-selected food intakes for seven consecutive days, starting on a randomly selected day of the week. Daily intakes of energy, dietary fiber, macronutrients, 7 vitamins, and Fe, Zn, Cu, Mn, and Se were calculated using food composition tables and literature values. Analysis of variance showed a significant weekend effect for energy and TE intakes that was reduced when data were expressed on the basis of nutrient densities. No day of the working week or training effect was noted for energy or any nutrient. The ratios of the inter- to intrasubject coefficient of variation for all the nutrients were surprisingly constant. However, the overall coefficients of variation-Fe, 36.5%; Zn, 45.7%; Cu, 43.7%; Mn, 58.2%, and Se, 52.7%-were markedly lower than for vitamin A (109.0%), Vitamin D (117.1%), and linoleic acid (82.2%), but slightly higher than those for energy (27.2%) and carbohydrate (30.1%). The widespread distribution of TE in the diets results in comparable reliability for dietary studies of TE and macronutrients. The 95% confidence intervals for mean individual TE intakes ranged from 25% (Zn) to 40% (Se). These relatively large uncertainties make it essential that any assessment of the significance of correlations between dietary TE intakes and biochemical indices takes into account the intra-individual variation associated with both variables.