The International Consortium for Quality Research on Dietary Sodium/Salt (TRUE) position statement on the use of 24‐hour, spot, and short duration (<24 hours) timed urine collections to assess dietary sodium intake

The International Consortium for Quality Research on Dietary Sodium/Salt (TRUE) is a coalition of intentional and national health and scientific organizations formed because of concerns low‐quality research methods were creating controversy regarding dietary salt reduction. One of the main sources of controversy is believed related to errors in estimating sodium intake with urine studies. The recommendations and positions in this manuscript were generated following a series of systematic reviews and analyses by experts in hypertension, nutrition, statistics, and dietary sodium. To assess the population's current 24‐hour dietary sodium ingestion, single complete 24‐hour urine samples, collected over a series of days from a representative population sample, were recommended. To accurately estimate usual dietary sodium at the individual level, at least 3 non‐consecutive complete 24‐hour urine collections obtained over a series of days that reflect the usual short‐term variations in dietary pattern were recommended. Multiple 24‐hour urine collections over several years were recommended to estimate an individual's usual long‐term sodium intake. The role of single spot or short duration timed urine collections in assessing population average sodium intake requires more research. Single or multiple spot or short duration timed urine collections are not recommended for assessing an individual's sodium intake especially in relationship to health outcomes. The recommendations should be applied by scientific review committees, granting agencies, editors and journal reviewers, investigators, policymakers, and those developing and creating dietary sodium recommendations. Low‐quality research on dietary sodium/salt should not be funded, conducted, or published.


| INTRODUC TI ON
High intake of dietary sodium is considered one of the leading global health risks, and reducing dietary sodium is estimated to be one of the most cost-effective strategies to improve population health. 1,2 As a result, reducing dietary sodium by 30% is one of nine World Health Assembly-World Health Organization endorsed targets to reduce the global burden of non-communicable disease by 25% by 2025. 3 Based on comprehensive systematic reviews of the evidence, multiple national and international health and scientific governmental, and non-governmental organizations have recommended reducing dietary sodium at the population level. 4 However, there are research studies that associate lowering dietary sodium with harm. 5,6 Concern has been expressed by several national and international health and scientific organizations that the association of reduced dietary sodium with harm results, in part, from the use of inappropriate low-quality research methods. 7  The recommendations were reviewed and approved by consensus of the TRUE expert committee. The manuscript was then reviewed for approval and supported by the organizations listed in Table 1.
One of the major methodologic challenges in research on dietary sodium is accurate assessment of dietary sodium. 10,16 Outside of tightly controlled trials, in most circumstances, individuals' diets vary widely from meal to meal, day to day, and workday to weekend day, and have many other temporal sources of variation related to factors such as seasonal availability of foods, holidays, cultural practices, and climatic change (flood, drought, heat, and cold) causing altered food availability. 16 Small, rigorous studies have directly and carefully measured the sodium content of foods and beverages, and the amount consumed. 13 Such intensive methods are not feasible in studying the health impact of dietary sodium in large, long-term studies conducted in non-institutionalized populations due to the difficulty in measurement of portion sizes and discretionary salt use, errors in self-report, and inaccuracies in food composition databases with respect to sodium concentration. 10,[17][18][19] Hence, other means of assessing sodium intake have evolved. 10,11,16 population sample, were recommended. To accurately estimate usual dietary sodium at the individual level, at least 3 non-consecutive complete 24-hour urine collections obtained over a series of days that reflect the usual short-term variations in dietary pattern were recommended. Multiple 24-hour urine collections over several years were recommended to estimate an individual's usual long-term sodium intake. The role of single spot or short duration timed urine collections in assessing population average sodium intake requires more research. Single or multiple spot or short duration timed urine collections are not recommended for assessing an individual's sodium intake especially in relationship to health outcomes. The recommendations should be applied by scientific review committees, granting agencies, editors and journal reviewers, investigators, policymakers, and those developing and creating dietary sodium recommendations. Low-quality research on dietary sodium/salt should not be funded, conducted, or published. Study methodologies to assess dietary sodium may differ depending on the purpose of the research. Some studies are primarily designed to assess the average sodium intake of a population (eg, to assess the overall impact of a population-wide dietary intervention) while others are primarily designed to assess individual sodium consumption (eg, to relate individuals' sodium intake to health outcomes). Estimation of average sodium consumption at the population level is minimally impacted by random error because random high and low individual estimates offset each other. In contrast, random error adds variability when estimating a population distribution and can result in inaccurate estimates at high or low levels of intake (eg, percentiles, prevalence of inadequate, or excess intake). When examining the relationship between individual-level sodium intake and health outcomes, random error will influence the association.
Systematic error will affect estimates of sodium consumption for both population averages and individual-level assessments, independent of the sample size. Systematic error can be constant or vary depending on the level of sodium intake or other factors (eg, varying degrees of non-adherence with urine collection). This implies that when assessing population average intake, the prime concern is to minimize systematic error (taking into account that random error can affect percentile and prevalence estimates of high or low intake levels), while research assessing sodium intake in relation to health outcomes of individuals must minimize both random error and systematic error. 9,20 A further challenge in assessing dietary sodium is related to the time period being assessed (current, short-term [~1 year] and/ or long-term intake [years]). Because of temporal variations in sodium intake, sodium excretion from a single 24-hour urine collection does not reflect usual short-or long-term intake in individuals.
Between-day variation in sodium intake in an individual can be as high as inter-individual differences in intake. 16 Sodium excretion from a single 24-hour urine collection may not reflect short-and long-term population average intake either, due to seasonal variability in food intake, age-related changes in food intake (eg, in aging populations), as well as population interventions to decrease sodium intake.

| E XCRE TI ON OF D IE TARY SOD I UM IN URINE
Under homeostatic circumstances of constant sodium intake in healthy people, approximately 93% of ingested sodium is excreted in the urine. 13 Similarly, studies in non-institutionalized populations on their usual diets also find that about 90% of sodium is excreted in urine. 13,16 In one small study with limited documentation, ingested sodium was exponentially excreted within 6 hours. 21 Similarly, following acute intravenous administration of 106 mg/kg sodium, most sodium was excreted within 5-10 hours and all within 40 hours. 22 It takes longer to fully excrete ingested sodium when there is an overall change in the usual amount of sodium ingested (ie, a change in homeostasis). 23,24 When there is a significant change in dietary sodium, a new homeostasis requires 2-7 days to be achieved. 13,21,[23][24][25][26] Even at a constant sodium intake, the individual daily variation in sodium excretion is large and some studies find a weekly cycle in sodium excretion associated with cycles in aldosterone excretion. [27][28][29]

| 24 -HOUR URINE COLLEC TI ON S TO A SS E SS AVER AG E P OPUL ATI ON SOD I UM I NTA K E
Twenty-four-hour urine collections capture approximately 93% of the current average population sodium intake. 13 Hence, 24-hour urinary sodium can be used to provide a close estimate of current 24-hour dietary sodium in population studies. Assessing average population sodium intake is minimally affected by the random component of day-to-day variation in sodium excretion in individuals as the random over-and underestimates of individual sodium intake are balanced out in calculating the population average. The caveats are that the 24-hour urine collections need to be complete, collected on days that are representative of the usual population pattern of sodium intake (eg, a mixture of weekend and weekdays) and the participants need to be representative of the population in question. If the intent is to assess usual short-term sodium intake in a population, the study design also needs to take into consideration seasonal or cultural variations (if any) in dietary patterns through inclusion of measurements across the time period (eg, a year) and cultural groups of interest. Seasonal variation may affect the extent to which "discretionary" or added salt contributes to total salt consumption in some diets and there may be agricultural or climatic influences on diet that need to be accounted for in the study design. Shortterm sodium intake may be used to relate to short-term changes in outcomes such as change in blood pressure. Long-term (>1 year) estimation of dietary intake also needs to consider changes in diet over time (eg, population interventions to reduce dietary sodium, reduced food intake with age in aging populations). Estimating the long-term sodium intake in a population is usually most relevant for studies assessing the impact of population policies or strategies to reduce dietary sodium, for example, that require stepwise changes in the sodium content of the food supply, or the relationship between dietary sodium and chronic disease outcomes (eg, stroke).

| 2-HOUR URINE COLLEC TI ON S TO A SS E SS HE ALTHY IND IVIDUAL' S SOD I UM I NTA K E
Sodium ingestion in individuals varies from day to day, and 24-hour urinary sodium excretion also varies in individuals at a constant sodium intake. Hence, multiple days of 24-hour urine collections are needed to assess an individual's usual sodium intake. 16,30,31 Fortyfive to fifty percent of respondents switched tertile of sodium intake when a single 24-hour urine collection was used to estimate long-term sodium intake vs sequential 24-hour urine collections in a study from Amsterdam. 32 The required number of 24-hour urine collections to obtain a stable estimate of usual sodium intake is likely to differ with different dietary patterns, populations, and settings, and has been estimated to be at least 3 non-consecutive days. 27,29,31-39 Weaver et al 29 found that ten 24-hour urine collections were required to have a 75% reliability to estimate an individual's sodium intake when on a constant sodium diet and this number is likely higher when there is substantial day-to-day variation in sodium intake. The strength of the association between dietary sodium and health outcomes is highly influenced by the number of urine collections. 32,38,40 The issues relating to current, short-term, and long-term estimates of sodium intake for populations also apply to individuals. Current sodium intake estimates require multiple 24-hour urine collections that account for usual daily changes in dietary patterns (eg, weekday vs weekend day). Short-term sodium intake estimates for 1 year need to account for annual cyclic changes in diet by the timing of 24-hour urine collections, and longterm sodium intake needs to have 24-hour urine collections taken throughout the study timeframe.

| CHALLENG E S IN COLLEC TING 24 -HOUR URINE COLLEC TIONS
There are challenges in collecting complete 24-hour urine collections that can reduce their utility in assessing dietary sodium. 10 Study conduct and quality control must be rigorous to ensure complete urine collection. 27 The systematic undercollection of 24-hour urine collections commonly seen in less rigorously conducted studies will underestimate both population and individual sodium intake.
Overcollection, which is less common, will do the opposite. There is also considerable respondent burden in collecting 24-hour urines, such that a sizable proportion of potential respondents may decline to enter studies that involve 24-hour urine collection. 10,16

| US E OF S P OT URINE AND S HORT DUR ATI ON TIMED URINE COLLEC TI ON S TO A SS E SS AVER AG E P OPUL ATI ON SOD I UM I NTA K E
Estimates of a population's average sodium intake with spot and short duration timed urine collections are not likely to be influenced by random error (each measurement is likely to be randomly above or below the average) but are very likely to be influenced by systematic errors. Several formulae used to estimate 24-hour urine sodium from spot collections have relatively small systematic errors in estimating average population sodium intake; however, some formulae used in different settings result in more substantive systematic error (>400 mg sodium). 14,15 Currently, all commonly used formulae systematically overestimate sodium intake at lower 24-hour urine sodium and underestimate intake at higher 24-hour urine sodium. 14,15,40,43,44 Thus, changes in dietary sodium intake at the population level (both increases and decreases) will be sys-

| US E OF S P OT AND S HORT DUR ATI ON TIMED URINE COLLEC TI ON S TO A SS E SS USUAL CURRENT HE ALTHY IND IVIDUAL' S SODIUM INTAKE
In a systematic review, correlations between 24-hour sodium esti- in studies of sodium association with health outcomes. 14,43,44 Using a spot urine collection to assess an individual's sodium intake will be influenced by both random and systematic error, and hence, large inaccuracies occur. Indeed, the errors in estimating an individual's 24hour urine sodium with this technique can exceed 8000 mg, which is greater than an adult's mean daily intake in most populations. 52,53 Further, studies examining spot and short duration timed urine collections have used a single 24-hour urine collection as an indicator of usual current sodium intake for individuals. As previously discussed, multiple 24-hour urine collections are required to reflect usual current intake in individuals. A few studies have investigated the potential for multiple spot urine collections to estimate usual sodium intake as assessed by multiple 24-hour urine collections. 47,54 As is the case for single spot urine collections, multiple spot urine collections underestimate 24-hour urine sodium at lower 24-hour urine sodium and overestimate at higher 24-hour urine sodium and differences between the methods can be as much 7000 mg sodium. 47 Current data do not support using single or multiple spot or short-term timed urine collections to assess individual sodium intake. 47,54,55

| US E OF S P OT AND S HORT DUR ATI ON TIMED URINE COLLEC TI ON S TO A SS E SS SOD I UM INTAKE AND ITS REL ATI ON S HIP TO D IS E A S E
Because sodium intake varies widely between meals, days, and seasons, and because most ingested sodium is exponentially and rapidly excreted within hours when eating a usual diet, there is little scientific rationale to expect the sodium concentration or quantity from a single spot or short duration timed urine collection to reflect current or long-term sodium intake. Sodium concentration in spot and short duration timed urine collections will largely reflect the sodium content of food and beverages consumed within hours of the urine collection. 21 The quantity and concentration of To partially account for variation in hydration, some investigators have examined the urine sodium in relationship to creatinine.
Creatinine is secreted by the renal tubules and less impacted by state of hydration than sodium, which is avidly reabsorbed in the renal tubules when there is dehydration. 58 Changes in the fractional excretion of sodium relative to creatinine are used, clinically, to assess dehydration as a cause of renal dysfunction. 58 A person's hydration status is a confounder in assessing sodium consumption using the ratio of sodium to creatinine in urine. The relationship between sodium and creatinine excretion is also changed by diuretics, such as caffeine, several kidney diseases, and illnesses. 58 Several formulae have been developed to estimate 24-hour urine sodium from a spot urine sodium collection among adults. Most of these formulae utilize age and sex as variables that are predictive of average sodium intake. On average, sodium intake is lower in females than males, lower in older adults than younger adults, and lower in children than adults. Many formulae also incorporate urine creatinine concentration, potentially to correct for changes in sodium concentration related to urinary dilution/concentration or possibly because creatinine is closely related to muscle mass (and indirectly to physical activity), and hence may relate to food intake. 59,60 Some formulae also incorporate weight, height (or body mass index), and urine potassium (potentially related to sodium-potassium exchange in the renal tubules or to the types of food consumed). 61 Age and sex are strong predictors of death and cardiovascular events. Body mass index has a complex relationship with health outcomes as both high and low values are major health risks. 62 Further, creatinine (as a reflection of impaired renal function and muscle mass), 63,64 and potassium (either directly or as a marker of diet quality) are also predictive of major health outcomes. 65 The associations between sodium intake estimated by for- The lack of scientific rationale to support the hypothesis, serious methodological issues in validation studies, incorporation of major confounding risk factors in formulae to estimate 24-hour urine sodium, and systematic differences in error with different levels of dietary sodium have led several to recommend that spot and short duration timed urine collections not be used. 14,45 Finally, these shortcomings create distortions in the associations between estimated salt consumption and health outcomes. 37 The current data do not support using single or multiple spot or short-term timed urine collections to assess sodium intakes in association with health outcomes.  3. Rigorous attention to quality control, including careful training of research leads, field workers, and study participants, should be in place to ensure a high participation rate and a high rate of complete 24-hour urine collections.

| POSITION ON THE USE OF SPOT OR SHORT DURATION TIMED URINE COLLECTIONS (<24 HOURS) TO ASSESS DIETARY SODIUM INTAKE
1. The role of single spot or short duration timed urine collections in assessing population average sodium intake requires more research for a definitive position and should be used cautiously for this purpose (Table 2). Where a single spot or short duration timed urine sample is used to assess average population sodium intake, a simultaneous calibration study with complete 24-hour urine samples should ideally be conducted

Setting Recommendation
Current average population intake 2. Single or multiple spot or short duration timed urine collections are not recommended for assessing an individual's sodium intake especially in relationship to health outcomes.

| D ISCUSS I ON
The TRUE Consortium recommends 24-hour urine collections be retained to assess population and individual sodium intake, with a cautious and currently unclear role for spot and short-term timed urine collections to assess population average sodium intake. For individual sodium intake, three and up to ten 24-hour urine collections are needed to obtain a reliable estimate. The important caveats to 24-hour urine collection include that rigorous methods are used to ensure complete urine collection and to assess completeness of the 24-hour urine collections, that the participants are representative of the population being studied, and that the timing of the urine collections meets the needs of the research question (current, annual, or long-term sodium consumption). The Consortium recommends not to use a spot and short-term timed urine collection to assess individual sodium intake.
The TRUE consortium's systematic reviews indicate that lowquality research has been commonly used in assessing dietary sodium. Previous systematic reviews on the use of 24-hour urine collections, food frequency questionnaires, dietary records, and 24-hour diet recall to assess sodium intake found serious methodological issues to be common. 11,13 Few high-quality validation studies were identified in our searches. A priority is to develop minimum methodologic standards for the conduct of validation studies. A regularly updated systematic review of the literature assessing the association of sodium intake to clinical outcomes found the majority of studies could not meet even minimum methodologic criteria. [69][70][71][72][73][74] The initial TRUE consortium position on blood pressure measurement in research studies was also developed because few research studies used the recommended methods to assess blood pressure. 13 Institutions funding research, journals, and scientists need to carefully assess the validity of the methods used in research relating to dietary sodium to ensure reliable guidance to public health programs.
Increasing evidence relates low-quality research methodology on dietary sodium to findings of "U-shaped," "J-shaped," or "inverse linear" associations between sodium intake and health outcomes. Meta-analyses that use criteria to exclude cohort studies with major methodological weaknesses find positive associations between increasing dietary sodium and cardiovascular events, especially stroke. 75,76 In contrast, meta-analyses that do not exclude studies with major methodological weaknesses find J-or U-shaped associations with dietary sodium and cardiovascular outcomes. 77,78 The Using the best current evidence, high dietary sodium intake has been stated to be a leading risk for death and disability globally, with reducing dietary sodium being one of the most cost-effective mechanisms to improve population health. 2,80 However, some research finds reducing dietary sodium to be associated with harm. 77 The TRUE consortium and others have expressed concern that low-quality research methods, including inaccurate assessment of dietary sodium and not accounting for confounding health risks (ie, use of formulae and spot or short-term timed urine collections), have caused some of the controversy around reducing dietary sodium. 11,13,15,46,66,81,82 Systematic review of the use of dietary records, food recall, and food frequency questionnaires has led the TRUE Consortium to recommend against using those methods for assessing sodium intake in individuals. Studies on dietary sodium need to be done rigorously and reproducibly with appropriate methods to further scientific knowledge and support public health action.
In contrast, low-quality research can generate false controversy, and misleading results thus confusing policymakers and the public with a strong potential to harm the ongoing public health efforts to reduce cardiovascular disease burden globally. The recommendations are intended to guide scientific review committees, granting agencies, editors and journal reviewers, investigators, policymakers, and those developing and creating dietary sodium recommendations.
Low-quality research on important public health topics should not be funded, conducted, or published.

Dr. Whelton was supported by a Centers for Biomedical Research
Excellence grant from the National Institute of General Medical Sciences (P20GM109036).

CO N FLI C T O F I NTE R E S T
NRCC was a paid consultant to the Novartis Foundation (2016)(2017) to support their program to improve hypertension control in low-