Comparison of three random spot urine-based calculations with 24-hour urine sodium measurement for mean salt intake estimation: a sampling study in adult residents of Anhui province

  Objective  To compare disparities between three random spot urine-based calculations and 24-hour urine sodium measurement in mean salt intake estimation among adult residents in Anhui province for evaluating the reliability of the three spot urine-based mean salt intake estimation methods.

  Methods  Random spot and 24-hour urine samples were collected from permanent adult residents (18 – 69 years old) participating in a hypertension and sodium intake baseline survey during 2018 in Anhui province. The eligible samples included spot urine samples from 3 329 participants and 24-hour urine samples from 1 500 participants, of which 1 493 had both spot and 24-hour urine samples. Mean salt intake of the participants was estimated according to determination of spot urine samples using the three (INTERSALT/Kawasaki/Tanaka) equations and the calculated results were compared with the measurement from 24-hour urine samples. Differences between the mean salt intake estimations based on spot urine samples and 24-hour urine samples were examined with paired t-test, intraclass correlation coefficient (ICC) analysis, and Bland-Altman plot test.

  Results  The mean salt intake based on 24-hour urine measurement was 9.1 g/day (95% confidence interval CI: 8.9 – 9.4); the spot urine-based mean salt intakes (in g/day) calculated using INTERSALT, Kawasaki, and Tanaka equation were 8.6 (95%CI: 8.5 – 8.7), 12.7 (95%CI: 12.6 – 12.9), and 9.5 (95%CI: 9.4 – 9.6), respectively. The paired comparison showed that the estimation derived from Tanaka equation was the closest to the 24-hour urine-based measurement, with a difference of 0.2 g/d (95% CI: 0.0 – 0.4). Among the three equations, the ICC was the highest (0.34) for Tanaka, followed by that of INTERSALT (0.33) and Kawasaki (0.23). The three spot urine-based estimations of mean salt intake were in good agreement with that of 24-hour urine-based measurement; while the estimations derived from INTERSALT and Tanaka equation could be higher under the condition of lower level salt intake but lower under the condition of higher level salt intake.

  Conclusion  The three random spot urine-based estimations of salt intake were consistent with 24-hour urine sodium measurement for adult residents in Anhui province and the estimation derived from Tanaka equation was more accurate.