Estimating mean population salt intake using spot urine samples in Nepal: a cross-sectional study

Introduction:Little is known about the usefulness of spot urine testing compared with 24-h urine samples to estimate salt intake in low-income settings. This is given 24-h urinary collection can be costly, burdensome, and impractical in population surveys. The primary objective of the study was to compare urinary sodium levels (as an estimate of salt intake) of Nepalese population between 24-h urine and spot urine using previously established spot urine-based equations. Additionally, this study explored the 24-h prediction of creatinine and potassium excretion from spot urine samples using available prediction equations.Methods:The sample population was derived from the community-based survey conducted in Nepal in 2018. Mean salt intake was estimated from spot urine samples comparing previously published equations, and this was then contrasted with mean salt intake estimations from 24-h urine samples, using paired t test, Pearson correlation coefficient, intraclass correlation coefficient, and Bland-Altman plots.Results:A total of 451 participants provided both complete 24-h and morning spot urine samples. Unweighted mean (& PLUSMN;SD) salt intake based on 24-h urine collection was 13.28 & PLUSMN; 4.72 g/day. The corresponding estimates were 15.44 & PLUSMN; 5.92 g/day for the Kawasaki, 11.06 & PLUSMN; 3.17 g/day for the Tanaka, 15.22 & PLUSMN; 16.72 g/day for the Mage, 10.66 & PLUSMN; 3.35 g/day for the Toft, 8.57 & PLUSMN; 1.72 g/day for the INTERSALT with potassium, 8.51 & PLUSMN; 1.73 g/day for the INTERSALT without potassium, 7.88 & PLUSMN; 1.94 g/day for the Whitton, 18.13 & PLUSMN; 19.92 g/day for the Uechi simple-mean and 12.07 & PLUSMN; 1.77 g/day using the Uechi regression. As compared with 24-h urine estimates, all equations showed significant mean differences (biases); the Uechi regression had the least difference with 9% underestimation (-1.21 g/day, P < 0.001).Proportional biases were evident for all equations depending on the level of salt intake in the Bland-Altman plots.Conclusion:None of the included spot urine-based equations accurately corresponded to 24-h salt intake in the present study. These equations may be useful for longitudinal monitoring of population salt intake in Nepal, our study highlights that there are limitations on using existing equations for estimating mean salt intake in Nepali population. Further studies are warranted for accuracy and validation.