Variability in 24-hour urine sodium excretion in children.

How many 24-hour urine sodium measurements are adequate for characterizing a child's salt intake? Can overnight urine specimens accurately replace 24-hour collections for salt assessment? A sample of 73 6th-8th grade children was taken from two parochial schools in Chicago to investigate systematically these questions. Seven consecutive 24-hour-urine specimens were collected from each child. The estimated ratio of intra- to inter-individual variances was 1.94 for 24-hour-urine sodium. Based on this value, eight 24-hour specimens are necessary to limit to 10% the diminution of the estimated correlation coefficient between 24-hour-urine sodium and blood pressure. Six measurements are required to reduce to 0.01 the probability of misclassifying a child in tertile 1 versus tertile 3. The overnight specimens show a moderate consistency with the 24-hour collections in detecting children with high or low salt intake. For example 92% and 85% of children in the fifth quintile and the third tertile respectively of the true mean overnight sodium have their true mean 24-hour Na in the upper half of the distribution. These results suggest that in a large scale epidemiologic study, overnight specimens may be reasonable alternatives when 24-hour-urine sodium is practically very difficult to collect.