Estimating dietary sodium intake using spot urine samples

Data from laboratory animal experiments, observational epidemiological studies, and randomized controlled trials provide strong evidence for a causal relationship between dietary sodium intake and blood pressure [1]. However, the association between sodium intake and clinical cardiovascular disease (CVD) remains conflicted: some studies have found a positive linear association between dietary sodium intake and risk of CVD [2–4]; others have reported a J or U-shaped association between dietary sodium and CVD [5–7]. Methodologic limitations, including inconsistencies in dietary sodium measurement methods, confounding factors, and reverse causality, are likely to blame for these conflicting findings [8]. If conducted properly, 24-h urinary sodium excretion is considered the most accurate method of estimating dietary sodium intake at the individual level [8]. It is estimated that approximately 90% of consumed sodium is absorbed and excreted in the urine, and the correlation between measured dietary sodium intake and urinary sodium excretion is more than 0.75 [8,9]. To estimate an individual's usual dietary intake, multiple 24-h urinary measurements of sodium are required because of large day-to-day variation in dietary intake [10,11]. However, collection of 24-h urine specimens is burdensome and challenging in free-living populations. Thus, overnight urine collection and spot urine collection have been used as alternative methods to estimate dietary sodium intake. In general, overnight (usually 8 h) urinary sodium excretion is well correlated with 24-h urinary sodium excretion [12,13]. However, because of larger observed intra and interindividual variations in overnight sodium excretion, more numbers of overnight sample collections are required to achieve the same precision as 24-h sample collections in estimating the association between an individual's true mean overnight urine sodium excretion and a study outcome [13]. In addition, hypertensive patients may have a different diurnal pattern of sodium excretion than normotensive study participants: hypertensive patients exhibited a greater rate of sodium excretion during sleep than during daytime hours, a reversal of the usual pattern inwhich excretion of sodium is greater during the day than during the night in individuals without hypertension [13]. Furthermore, overnight vs. daytime sodium excretion rates have been shown to vary by race/ethnicity, sex, and antihypertensive medication use [8]. Therefore, overnight urine collection could overestimate or underestimate 24-h sodium excretion. Owing to the relative convenience of spot urine specimen collection, it has been widely used to estimate dietary sodium intake in large-scale epidemiological studies [5–7]. However, because spot specimens are collected at only one time point, random measurement errors because of intra and interday variations in urine tonicity, sodium excretion, and timing of meals, exercise, and medication use cannot be avoided [14]. Typically, spot urinary sodium concentrations have been converted into 24-h sodium excretion using available estimating equations developed among various populations [15–18]. Systematic differences (biases) between measured 24-h urinary sodium excretion and calculated 24-h urinary sodium excretion using these formulas have been reported [17,18]. Moreover, these biases might relate to dietary sodium intake [8,17,18]. In this issue of the Journal of Hypertension, Polonia and colleagues [19] reported findings from a study which assessed the validity of Tanaka, Kawasaki, INTERSALT, and NHANES formulas [15–18] for estimating 24-h urinary sodium and potassium excretion based on spot urine specimens in 2399 individuals aged 18–96 years from the Portuguese general population. They found that estimated 24-h sodium and potassium excretion were only moderately correlated with measured 24-h sodium and potassium excretion (correlation coefficients ranged from 0.23 to 0.36). In addition, the Kawasaki formula significantly overestimated 24-h sodium excretion by 1289 (SD 1654), 1227 (1889), and 1366 (1790) mg among normotensives, treated hypertensive patients, and untreated hypertensive patients, respectively, whereas the other formulas also produced some degree of bias by under or overestimating 24-h sodium excretion. Furthermore, these biases varied by levels of dietary sodium intake – overestimating among individuals with low sodium intake and underestimating among individuals with high sodium intake. They concluded that all of the formulas provided biased estimates of sodium and potassium excretion and may be unreliable for assessing individual and population levels of daily sodium and potassium excretion. In the previous studies, 24-h sodium excretion equations were established by the regression of sodium/creatinine excretion ratios from spot urine samples on sodium excretion from 24-h urine specimens within the same day [15–18]. In this study, spot urine specimens and 24-h urine specimens were collected 1–7 days apart. Therefore, both intra and interday variations in sodium excretion affected the correlation and bias estimates. However, these estimates might be more meaningful because we are interested in assessing the independent correlation between spot urine sodium excretion and 24-h urine sodium excretion. Sodium excretion from spot urine samples and 24-hoururine samples within the same day, especially sodium/creatinine ratios within the same day, would result in an artificially high correlation, which has been cited as evidence to support the use of spot urine sodium as an index of dietary intake in epidemiological studies [5–7]. The study by Polonia and colleagues [19] provides strong evidence that the Kawasaki equation, which was created in a high-sodium intake population, seriously overestimates sodium intake in western populations. We must keep a close eye on bias beyond correlation. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.