The Validity of Predictive Equations to Estimate 24-Hour Sodium Excretion

Abstract

We examined the population distribution of urinary sodium concentrations and the validity of existing equations predicting 24-hour sodium excretion from a single spot urine sample among older adults with and without hypertension. In 2013, 24-hour urine collections were obtained from 554 participants in the Multi-Ethnic Study of Atherosclerosis and the Coronary Artery Risk Development in Young Adults study, who were aged 45-79 years and of whom 56% were female, 58% were African American, and 54% had hypertension, in Chicago, Illinois. One-third provided a second 24-hour collection. Four timed (overnight, morning, afternoon, and evening) spot urine specimens and the 24-hour collection were analyzed for sodium and creatinine concentrations. Mean 24-hour sodium excretion was 3,926 (standard deviation (SD), 1,623) mg for white men, 2,480 (SD, 1,079) mg for white women, 3,454 (SD, 1,651) mg for African-American men, and 3,397 (SD, 1,641) mg for African-American women, and did not differ significantly by hypertensive status. Mean bias (difference) in predicting 24-hour sodium excretion from the timed spot urine specimens ranged from -182 (95% confidence interval: -285, -79) to 1,090 (95% confidence interval: 966, 1,213) mg/day overall. Although the Tanaka equation using the evening specimen produced the least bias overall, no single equation worked well across subgroups of sex and race/ethnicity. A single spot urine sample is not a valid indicator of individual sodium intake. New equations are needed to accurately estimate 24-hour sodium excretion for older adults.