Because of within-individual variation, surveys to estimate an individual's usual food intake must be conducted over many days, in general. Here, using non-invasive biomarkers, we examined the number of measurements required to screen for the usual intake of fruit and vegetables, in addition to sodium, potassium, and the sodium-to-potassium (Na/K) ratio. Participants were 202 subjects aged 40-74 years from five areas of Japan who completed weighed food records (WFR) and five 24-hour urinary collections (24-h UCs) between 2012 and 2013. The number of 24-h UCs required to screen for intake that deviated from guidelines estimated by the WFR and their accuracies were assessed by the area under the curve (AUC) in a receiver-operating characteristics (ROC) analysis. The single urinary excretion of sodium, potassium, and the Na/K ratio showed moderate performance (AUC value: >0.7) in discriminating deviations from their criteria by respective intake based on the WFR. Urinary potassium excretion also showed moderate performance (AUC value: >0.7) in estimating the intake of vegetables but could not be used to estimate fruit intake even after five collections. The non-invasive measurement of biomarkers in a single 24-h UC showed moderate performance in screening the usual intake of vegetables, as measured based on the 12-day WFR, as well as of sodium, potassium, and the Na/K ratio.
Keywords: biomarker; fruit and vegetable intake; receiver-operating characteristic analysis; screening; within- and between- individual variation.