ABSTRACT
Background
Excess sodium intake can increase blood pressure, and high blood pressure is a major risk factor for cardiovascular disease. Accurate population sodium intake estimates are essential for monitoring progress toward reduction, but data are limited on the amount of sodium consumed from discretionary salt.
Objectives
The aim of this study was to compare measured sodium intake from salt added at the table with that estimated according to the Healthy People 2020 (HP 2020) methodology.
Methods
Data were analyzed from the 2014 Salt Sources Study, a cross-sectional convenience sample of 450 white, black, Asian, and Hispanic adults living in Alabama, Minnesota, and California. Sodium intake from foods and beverages was assessed for each participant through the use of 24-h dietary recalls. Estimated sodium intake from salt used at the table was assessed from self-reported frequency and estimated amounts from a previous study (HP 2020 methodology). Measured intake was assessed through the use of duplicate salt samples collected on recall days.
Results
Among all study participants, estimated and measured mean sodium intakes from salt added at the table were similar, with a nonsignificant difference of 8.9 mg/d (95% CI: −36.6, 54.4 mg/d). Among participants who were non-Hispanic Asian, Hispanic, had a bachelor's degree or higher education, lived in California or Minnesota, did not report hypertension, or had normal BMI, estimated mean sodium intake was 77–153 mg/d greater than measured intake (P < 0.05). The estimated mean sodium intake was 186–300 mg/d lower than measured intake among participants who were non-Hispanic black, had a high school degree or less, or reported hypertension (P < 0.05).
Conclusions
The HP 2020 methodology for estimating sodium consumed from salt added at the table may be appropriate for the general US adult population; however, it underestimates intake in certain population subgroups, particularly non-Hispanic black, those with a high school degree or less, or those with self-reported hypertension. This study was registered at clinicaltrials.gov as NCT02474693.
Keywords: discretionary salt, sodium, table salt, Healthy People 2020, NHANES
Introduction
High blood pressure is a major risk factor for cardiovascular disease (1). In recognition of the direct and positive relation between blood pressure and sodium intake, and excess intake, Healthy People 2020 (HP 2020) includes an objective to “reduce consumption of sodium in the population aged 2 years and older” (2). The majority of US intake is estimated to come from sodium added to food outside the home by commercial manufacturers and food preparers, but discretionary salt added by the consumer can be an important contributor (3). In a recent study, ∼5% of sodium intake came from salt added at the table and ∼6% from in-home food preparation; among certain population subgroups, e.g., black participants, 10% of intake, or ∼400 mg sodium/d, came from salt added at the table. Reducing mean population sodium intake by 400 mg/d has been projected to save billions of dollars in health care costs each year and prevent 20,000–40,000 new cases of coronary heart disease (4). Accurate monitoring of changes in the total sodium intake, and the sources of this, over time is essential for understanding the impact of reduction strategies.
Estimating discretionary sodium intake from salt used at the table or during home food preparation is a challenge. Typically, estimates of mean US sodium intake exclude sodium from salt added at the table, and since 2009, the USDA discontinued adjustment for sodium added during home food preparation in NHANES because of the challenges of estimating changes in how foods are prepared (5). To monitor progress toward the HP 2020 objective, the HP 2020 method combines data from the NHANES 24-h dietary recall with an estimate of sodium intake from salt added at the table. This estimate is based on self-reported frequency of discretionary salt use from responses to additional questions in NHANES and an estimate of the amount of sodium typically used when salt is added taken from a study conducted >20 y ago (6). Understanding how the reported frequency of discretionary salt use compares with measured amounts can help improve monitoring of total sodium intake and better support sodium-reduction strategies.
Data from the 2014 Salt Sources Study provide an opportunity to evaluate the validity of questions used to monitor discretionary salt use in NHANES (3). The objective of this analysis was to compare the estimated sodium intake (both total and from discretionary salt used at the table) through the use of the HP 2020 method (as described above), with self-reported intake, i.e., from dietary recalls and duplicate salt sample collections. Secondary objectives were 2-fold: 1) to determine whether the mean sodium intake per day from discretionary salt added at the table and during home food preparation varied across participant subgroups and reported frequency of use; and 2) to compare a measured frequency of salt added at the table and during home food preparation (from duplicate salt sample collection) with self-reported frequency according to the questions in NHANES. Understanding how questions used to monitor sodium consumed from discretionary salt use compare with measured values can assist in public health efforts to better monitor total sodium intake and inform sodium-reduction strategies.
Methods
Overview
Data were analyzed from the 2014 Salt Sources Study (3), a cross-sectional study of adults living in 3 geographic locations in the United States. Data on the amount and proportion of dietary sodium obtained from a variety of sources including discretionary salt use from salt added at the table or during home food preparation were previously published, as well as information on study design and data collection methods (3).
Participants
In 2014, 450 adults were recruited as a convenience sample from 3 distinct geographic locations: Minneapolis-St Paul, MN; Birmingham, AL; and Palo Alto, CA. Details regarding recruitment are available elsewhere (3). Participants with missing responses to the NHANES discretionary salt use questions (described below) (n = 22) and those who reported consuming light salt or salt substitutes (n = 5) were excluded from analysis. The final sample for this analysis included 423 participants from the full study sample of n = 450 (94%) (Supplemental Figure 1).
Study design and data collection
Over a period of 11 d each participant visited the clinic at least once and participated in 4 24-h dietary recalls over the telephone (3). At the baseline clinic visit, interviewers administered a detailed questionnaire, scheduled participants for 4 24-h dietary recalls, and provided them with kits and detailed instructions for collecting duplicate samples of discretionary salt. Each kit included 1 bag/d for salt added at the table and 6 bags/d for salt added during home preparation, with a separate bag used for each food item prepared with salt. Participants were instructed to share the instructions and collection bags for home food preparation with anyone else in the household who may have prepared food at home on the day of collection. Salt added at the table included salt added to foods at and away from home. Questions included demographic and health characteristics, self-reported weight and height, and the frequency of adding salt to food at the table and in home food preparation (NHANES discretionary salt use questions, see Supplemental Methods). Participants then completed 4 telephone 24-h dietary recalls (on 3 weekdays and 1 weekend day) and used the provided kits and detailed instructions to collect duplicate portions of salt added at the table and during home food preparation on days corresponding with the 24-h recall period. Dietary recalls were collected by trained staff with the use of the 2012 version of the Nutrition Data Systems for Research nutrient calculation software. Duplicate salt samples were mailed to study centers, and the contents of each bag were weighed and recorded.
Measured sodium intake from discretionary salt
Measured sodium intake was based on the returned salt collection kits. For each recall day the amount of sodium from discretionary salt was estimated from the weight of the salt collected in the duplicate salt sample bags, separately by type of use, i.e., salt added at the table or during home food preparation.
For sodium from salt added during home food preparation, the weight of the duplicate salt samples collected was adjusted to reflect the proportion of the food prepared that was consumed (servings consumed/total servings made) by the participant. Also, absorption fractions for foods were used to adjust the sodium intake from salt added during preparation of foods such as pasta in which only a portion of salt added was absorbed in the food and the remainder was discarded with the water used to cook the food (3).
If no duplicate salt samples were submitted for a recall day for the participant and type of use, i.e., at the table or during home food preparation, 0 mg sodium intake from discretionary salt was assumed for the corresponding use for that day. For recall days with ≥1 duplicate samples of salt returned, sodium estimates for each sample were summed to derive the total sodium intake per day from discretionary salt used at the table or during home food preparation.
For each participant, the measured sodium intake per day from discretionary salt for each type of use was expressed as both: 1) the mean sodium intake from discretionary salt across all 4 recall days; and 2) the geometric mean of sodium intake consumed from discretionary salt only across days each type of salt was used.
Measured frequency of discretionary salt use
For each participant, the measured frequency of adding discretionary salt by type of use was estimated as the sum of the number of days a participant returned ≥1 bag with the discretionary salt (>0 mg sodium) divided by 4, the number of dietary recall days. Participants who added salt on none of their 4 recall days had a frequency of 0; 1 day, 0.25; 2 days, 0.50; 3 days, 0.75, and 4 days, 1.00.
Estimated sodium intake from salt added at the table
Estimated mean sodium intake from salt added at the table and total sodium intake were calculated according to HP 2020 methodology (see Supplemental Methods). The HP 2020 method estimates total sodium intake by summing sources of sodium intake (i.e., foods and beverages, drinking water, supplements, and antacids) and an estimated amount of sodium from salt added at the table (2). For ordinary salt added at the table, the mean amount of sodium is estimated at 580 mg/d, based on the mean sodium intake from salt added at the table among people who routinely added salt (581 mg/d) in the 1991 study (6). This is multiplied by a value assigned to the self-reported frequency of adding salt from the NHANES discretionary salt use questions (see Supplemental Methods). Those who report “never” adding salt at the table receive a frequency value of 0, for “rarely”, the sodium value is multiplied by 0.25, for “occasionally,” it is multiplied by 0.50, and for “very often,” it is multiplied by 1. Therefore, HP 2020 estimates of sodium intake from discretionary salt used at the table can be 0, 145, 290, or 580 mg/d (2). To replicate the HP 2020 method for estimating sodium intake from salt added at the table, 580 mg was also used for the current analysis, as well as the same frequencies based on self-report. To estimate total sodium intake in this study according to this HP 2020 methodology, the measured sodium intake from salt used at the table based on the duplicate salt sample collections was subtracted from the measured total sodium intake of participants, and the estimated value of sodium from salt used at the table based on HP 2020 methodology was added. For salt added during home preparation, self-reported frequency of salt use was also based on responses to the corresponding NHANES question (see Supplemental Methods). Participants could respond that they added salt during home preparation “never,” “rarely,” “occasionally,” or “very often.” For this analysis, because very few respondents reported “never” adding salt, “never” and “rarely” were combined into 1 group, referred to as the “never/rarely” group.
Ethics
All participants provided informed consent and the Institutional Review Board at each participating institution and the CDC reviewed and approved the study protocol and procedures.
Statistical analysis
Covariates of interest included sex, age, race/ethnicity, highest level of education, study site, self-reported hypertension, and BMI based on self-reported height and weight. To maintain comparability with the previous publication (3), arithmetic means were used to estimate the mean sodium intake from salt added at the table and total sodium intake across all study participants and by subgroup. Spearman correlation coefficients (ρ) were used to determine the associations between measured sodium intakes and estimated sodium intakes according to the HP 2020 methodology for total intake and for sodium consumed from salt added at the table. Paired t tests were used to determine the statistical significance of differences in mean sodium intake between the 2 assessment methods overall and by subgroups based on age, sex, race/ethnic group, geographic location, education level, self-reported hypertension, and BMI. Linear regression models were performed to examine the association between covariates and the mean differences between estimated and measured sodium intake for each participant.
The distributions of the sodium intake consumed from discretionary salt on intake days (when discretionary salt intake was >0 mg) were significantly skewed. Therefore, data were log transformed to normalize the distributions and geometric means presented, overall and by demographic and health characteristics. An ANOVA test was used to assess whether differences in geometric means varied across subgroups, after adjustment for age, sex, race/ethnicity, education, and BMI in linear regression models. In supplemental analyses, the geometric means of sodium intake from discretionary salt use by demographic and health characteristics were stratified by study site.
The mean measured frequency of discretionary salt use, i.e., proportion of days salt was added, was reported overall, by demographic and health characteristics, and by self-reported frequency based on response to the NHANES discretionary salt use questions. Linear regression models were used to test for differences across self-reported frequency of salt use, adjusting for age, sex, race/ethnicity, education, and BMI.
For all analyses, estimates with a relative SE (SE/estimate) ≥30% were marked as statistically unreliable (7). A P value of <0.05 was considered statistically significant. No corrections were made for multiple tests. A correlation of <0.4 was considered low, 0.4–0.7 moderate, and >0.7 strong (8). All analyses were completed with SAS software version 9.3 (SAS Institute, Inc.).
Results
The mean measured sodium intake from salt added at the table among the participants across all 4 days of dietary recall was 224 mg/d and the mean total sodium intake from all sources (including measured table salt) was 3544 mg/d. Overall, among all study participants combined, the estimated mean sodium intake from salt added at the table calculated according to HP 2020 methodology was not statistically significantly different from measured sodium, with a difference of 8.9 mg/d (95% CI: −36.6, 54.4 mg/d); however, significant differences in estimated compared with measured intake were noted among population subgroups (Table 1).
TABLE 1.
Mean sodium intake and Spearman correlations by assessment method for sodium consumed from salt added at the table and total sodium intake among US adults, overall and by demographic and health characteristics1
| Sodium from salt added at the table, mg/d | Total sodium intake, mg/d | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Measured2 | Estimated3 | ρ4 | Measured5 | Estimated6 | ρ4 | Difference7 | P 8 | |
| Overall | 423 | 224 ± 24.99 | 233 ± 7.7 | 0.40*** | 3540 ± 71.7 | 3550 ± 64.5 | 0.96*** | 8.9 (−36.6, 54.4)10 | 0.70 |
| Sex | |||||||||
| Male | 208 | 238 ± 41.3 | 224 ± 10.7 | 0.35*** | 3940 ± 110 | 3920 ± 97.6 | 0.96*** | −14.1 (−90.8, 62.6,) | 0.72 |
| Female | 215 | 210 ± 28.4 | 241 ± 11.1 | 0.45*** | 3160 ± 84.9 | 3200 ± 77.4 | 0.95*** | 31.2 (−19.4, 81.8) | 0.23 |
| Age, y | |||||||||
| <45 | 212 | 178 ± 30.5 | 232 ± 10.7 | 0.42*** | 3580 ± 101 | 3630 ± 94.0 | 0.96*** | 54.2 (−2.3, 110.7) | 0.06 |
| ≥45 | 211 | 270 ± 39.2 | 234 ± 11.1 | 0.39*** | 3510 ± 102 | 3480 ± 88.1 | 0.96*** | −35.6 (−107.9, 34.8) | 0.31 |
| Race or ethnicity | |||||||||
| Non-Hispanic white | 141 | 229 ± 36.6 | 238 ± 13.6 | 0.36*** | 3630 ± 128 | 3640 ± 120 | 0.95*** | 8.6 (−61.9, 79.2) | 0.81 |
| Non-Hispanic black | 91 | 597 ± 87.1 | 304 ± 19.0 | 0.47*** | 4170 ± 171 | 3870 ± 135 | 0.91*** | −292 (−452, −132) | 0.0005 |
| Non-Hispanic Asian | 97 | 21.5 ± 6.6* | 173 ± 12.5 | 0.26** | 3420 ± 141 | 3570 ± 140 | 0.99*** | 152 (129, 175) | <0.0001 |
| Hispanic | 94 | 64.2 ± 17.2 | 218 ± 14.2 | 0.16 | 2951 ± 107 | 3100 ± 105 | 0.99*** | 153 (112, 195) | <0.0001 |
| Education | |||||||||
| High school graduate or less | 58 | 561 ± 113 | 260 ± 20.1 | 0.33** | 3940 ± 235 | 3630 ± 179 | 0.93*** | −301 (−514, −87.5) | 0.007 |
| Some college or technical school | 92 | 293 ± 64.4 | 263 ± 17.6 | 0.46*** | 3930 ± 180 | 3900 ± 159 | 0.97*** | −29.8 (−143, 83.6) | 0.60 |
| Bachelor's degree or greater | 271 | 130 ± 18.2 | 217 ± 9.4 | 0.38*** | 334 ± 76.5 | 3420 ± 74.7 | 0.96*** | 87.4 (52.1, 123) | <0.0001 |
| Study site | |||||||||
| Alabama | 132 | 566 ± 65.9 | 274 ± 15.0 | 0.46*** | 4210 ± 143 | 3920 ± 119 | 0.92*** | −293 (−413, −172) | <0.0001 |
| California | 145 | 29.8 ± 6.1 | 179 ± 9.6 | 0.30*** | 3270 ± 107 | 3420 ± 106 | 0.99*** | 149 (130, 168) | <0.0001 |
| Minnesota | 146 | 107 ± 23.6 | 249 ± 13.9 | 0.30**** | 3220 ± 107 | 3360 ± 105 | 0.98*** | 142 (95.1, 190) | <0.0001 |
| Self-reported hypertension | |||||||||
| Yes | 97 | 421 ± 72.9 | 235 ± 15.7 | 0.57*** | 3810 ± 170 | 3620 ± 143 | 0.95*** | −187 (−317, −56.0) | 0.006 |
| No | 322 | 157 ± 22.9 | 233 ± 9.0 | 0.38*** | 3430 ± 76.6 | 3510 ± 71.6 | 0.97*** | 76.8 (34.3, 119) | 0.0004 |
| BMI,11 kg/m2 | |||||||||
| <25.0 | 149 | 89.0 ± 27.2* | 220 ± 12.8 | 0.38*** | 3260 ± 112 | 3400 ± 109 | 0.97*** | 131 (78.0, 184) | <0.0001 |
| 25.0–29.9 | 143 | 257 ± 47.4 | 218 ± 12.5 | 0.38*** | 3570 ± 122 | 3530 ± 107 | 0.96*** | −38.5 (−127, 50.5) | 0.39 |
| ≥30 | 131 | 342 ± 50.8 | 263 ± 14.6 | 0.46*** | 3840 ± 136 | 3760 ± 119 | 0.95*** | −78.0 (−168, 12.0) | 0.09 |
Data are from the Salt Sources Study, 2014. *Estimates are considered statistically unreliable [relative SE (SE/estimate) ≥0.30)]. **P < 0.05. ***P < 0.001. HP 2020, Healthy People 2020.
Measured = mean sodium intake from table salt over 4 days of recall, measured from duplicate salt collection.
Estimated = HP 2020 estimate.
Spearman correlation for measured compared with estimated.
Measured = mean total sodium intake from all sources as measured from dietary recalls and duplicate salt collections.
Estimated = total sodium intake from all sources, i.e., measured sodium from table salt + HP2020 estimate of sodium from table salt.
Difference = estimated intake (HP 2020) – measured intake (dietary recalls and duplicate salt collection).
t test P value of whether difference is statistically significantly different for each group.
Mean ± SE (all such values).
Mean (95% CI) of the differences between estimated and measured sodium intake.
BMI = weight (kg)/height (m)2; <25 = normal weight; 25.0–29.9 = overweight; ≥30 = obese.
The estimated mean total sodium intake according to HP 2020 estimates for salt added to food at the table was significantly greater than that measured among non-Hispanic Asians, Hispanics, those with a bachelor's degree or higher, participants from the California and Minnesota study sites, those without self-reported diagnosis of hypertension, and those with normal BMI. Estimated sodium intake was 77–153 mg/d greater than measured intake in these groups. In contrast, the estimated mean total sodium intake was 186–300 mg/d lower than measured intake among non-Hispanic black participants and participants with high school degree or less, living in Alabama, or who reported having hypertension (Table 1). After adjustment for age, race/ethnicity, sex. education, and BMI, factors that remained statistically significantly associated with differences between estimated and measured sodium intake were hypertension status, being a high school graduate or less, being of non-Hispanic black race, and being Hispanic (Table 2).
TABLE 2.
Linear regression coefficients (β) from univariate and multivariate analyses of differences between measured and estimated sodium intake from salt added at the table among US adults, by demographic and health characteristics1
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| Difference as a dependent variable2 | β | 95% CI | P | β | 95% CI | P |
| Sex | ||||||
| Male | Ref. | Ref. | ||||
| Female | −45.3 | −136, 45.7 | 0.33 | −18.1 | −103, 67.1 | 0.68 |
| Age,3 y | 4.3 | 1.4, 7.2 | 0.004 | 1.6 | −1.3, 4.4 | 0.28 |
| Race or ethnicity | ||||||
| Non-Hispanic white | Ref. | Ref. | ||||
| Non-Hispanic black | 301 | 183, 419 | <0.0001 | 223 | 101, 344 | 0.0004 |
| Non-Hispanic Asian | −143 | −259, −27.5 | 0.02 | −104 | −219, 11.9 | 0.08 |
| Hispanic | −145 | −262, −27.8 | 0.02 | −222 | −340, 104 | 0.0002 |
| Education | ||||||
| High school graduate or less | 388 | 257, 519 | <0.0001 | 336 | 198, 473 | <0.0001 |
| Some college or technical school | 117 | 8.2, 226 | 0.04 | 25.5 | −84.6, 136 | 0.65 |
| Bachelor's degree or greater | Ref. | Ref. | ||||
| Study site | ||||||
| Alabama | Ref. | Ref. | ||||
| California | −442 | −544, 340 | <0.0001 | −301 | −447, −156 | <0.0001 |
| Minnesota | −435 | −537, −333 | <0.0001 | −362 | −481, −243 | <0.0001 |
| Self-reported hypertension | ||||||
| Yes | 263 | 159, 368 | <0.0001 | 156 | 45.3, 267 | 0.006 |
| No | Ref. | Ref. | ||||
| BMI,3 kg/m2 | 11.9 | 5.9, 17.8 | 0.0001 | 5.1 | −0.8, 11.1 | 0.09 |
Data are from the Salt Sources Study, 2014. Univariate analyses were linear regression models performed for each covariate (independent variable) and the difference (dependent variable). The multivariate models included sex, age, race/ethnicity, education, and BMI. HP 2020, Healthy People 2020.
Difference between sodium intake = estimated intake (HP 2020) – measured intake (dietary recalls and duplicate salt collection).
Age and BMI modeled continuously. BMI = weight (kg)/height (m)2.
Overall, the estimated mean sodium intake per day from salt added at the table was moderately and statistically significantly correlated with the measured intake for all participants (ρ = 0.40, P < 0.001). Correlations ranged from ρ = 0.16 among Hispanics (P = 0.12), to ρ = 0.57 among those with self-reported hypertension (P < 0.001). When accounting for all sources of sodium intake, with the only difference being method of measuring sodium from salt added at the table, measured and estimated amounts of total sodium intake were highly correlated across all subgroups (ρ > 0.90, P < 0.001) (Table 1).
On days participants used salt at the table (consumption days), the mean sodium intake from table salt was 400 mg/d (Table 3). Intake varied substantially by participants’ demographic and health characteristics. On consumption days, the mean sodium intake from table salt was >1100 mg/d among non-Hispanic black participants, participants with a high school degree or less, and participants living in Alabama, and >944 mg/d among participants with hypertension. In contrast, the mean sodium intake from table salt on consumption days was <200 mg/d among Hispanic and non-Hispanic Asian participants, participants living in California and Minnesota, and among those with BMI <25 kg/m2 (Table 3). After adjustment for demographic characteristics and BMI, on consumption days, the mean sodium intake from table salt significantly varied by race/ethnicity, education, study site, hypertension status, and BMI, but not by sex, age group, or self-reported frequency of table salt use based on NHANES discretionary salt use questions. Per the methods of the study, selection of participants by race/ethnicity varied by study, making it difficult to disaggregate results for participants in different racial/ethnic groups from where they live. Data suggest that non-Hispanic white participants in Alabama added a greater amount of salt at the table on days they added salt to food when compared with non-Hispanic white participants in Minnesota or California (Supplemental Table 1).
TABLE 3.
Geometric means and range of sodium intake from salt added at the table and during home cooking/preparation, on days the specified salt was used, among US adults, overall and by demographic and health characteristics1
| n | Table salt, mg/d | Maximum – minimum, mg/d | n | Cooking/preparation salt, mg/d | Maximum – minimum, mg/d | |
|---|---|---|---|---|---|---|
| Overall | 186 | 400 ± 48.02 | 7650 | 283 | 388 ± 27.9 | 6790 |
| Sex | ||||||
| Male | 80 | 423 ± 79.0 | 4250 | 131 | 482 ± 49.6 | 6790 |
| Female | 106 | 384 ± 59.8 | 7650 | 152 | 322 ± 31.4 | 3700 |
| P3 | 0.96 | 0.004 | ||||
| Age, y | ||||||
| <45 | 87 | 314 ± 58.7 | 4600 | 138 | 347 ± 35.0 | 4090 |
| ≥45 | 99 | 495 ± 74.7 | 7650 | 145 | 432 ± 43.7 | 6790 |
| P3 | 0.75 | 0.24 | ||||
| Race or ethnicity | ||||||
| Non-Hispanic white | 65 | 399 ± 85.0 | 3440 | 89 | 332 ± 48.0 | 6790 |
| Non-Hispanic black | 59 | 1130 ± 162 | 7650 | 58 | 827 ± 84.0 | 3310 |
| Non-Hispanic Asian | 23 | 105 ± 31.6* | 711 | 69 | 262 ± 36.5 | 4090 |
| Hispanic | 39 | 187 ± 35.9 | 4600 | 67 | 372 ± 47.8 | 2830 |
| P3 | <0.0001 | 0.0009 | ||||
| Education | ||||||
| High school graduate or less | 26 | 1430 ± 379 | 7650 | 41 | 598 ± 94.2 | 2810 |
| Some college or technical school | 48 | 517 ± 105 | 4240 | 59 | 522 ± 79.0 | 6780 |
| Bachelor's degree or greater | 112 | 267 ± 40.6 | 3520 | 181 | 324 ± 29.1 | 4090 |
| P3 | 0.01 | 0.64 | ||||
| Study site | ||||||
| Alabama | 75 | 1690 ± 76.1 | 6950 | 80 | 846 ± 75.6 | 6710 |
| California | 47 | 137 ± 21.6 | 1470 | 95 | 285 ± 33.1 | 2700 |
| Minnesota | 64 | 163 ± 33.1 | 4600 | 108 | 285 ± 34.8 | 4090 |
| P3 | <0.0001 | <0.0001 | ||||
| Self-reported hypertension | ||||||
| Yes | 50 | 944 ± 177 | 7650 | 66 | 471 ± 73.3 | 6790 |
| No | 132 | 276 ± 38.7 | 4600 | 216 | 365 ± 29.5 | 4090 |
| P2 | 0.07 | 0.49 | ||||
| BMI,4 kg/m2 | ||||||
| <25.0 | 58 | 177 ± 33.7 | 4250 | 102 | 258 ± 28.9 | 2090 |
| 25.0–29.9 | 57 | 459 ± 105 | 7650 | 91 | 432 ± 52.5 | 4090 |
| ≥30 | 71 | 699 ± 117 | 3520 | 90 | 554 ± 70.4 | 6790 |
| P2 | 0.02 | 0.04 | ||||
| Self-reported frequency of salt use | ||||||
| Never/rarely5 | 80 | 286 ± 54.9 | 4600 | 45 | 527 ± 81.1 | 3670 |
| Occasionally | 60 | 443 ± 95.1 | 7650 | 82 | 298 ± 47.8 | 6790 |
| Very often | 46 | 630 ± 119 | 4230 | 156 | 408 ± 35.3 | 4080 |
| P 3 | 0.42 | 0.13 | ||||
Data are from the Salt Sources Study, 2014. *Estimates are considered statistically unreliable [relative SE (SE/estimate) ≥0.30)].
Geometric mean ± SE (all such values); means calculated only over days salt was added.
P values adjusted for age, sex, race/ethnicity, education, and BMI.
BMI = weight (kg)/height (m)2; <25 = normal weight; 25.0–29.9 = overweight; ≥30 = obese.
“Never” group included n = 28 participants; combined with “rarely.”
On days participants used salt during home cooking/food preparation (consumption days), the mean sodium intake from cooking/preparation salt was 388 mg/d (Table 3). On consumption days mean sodium intake from cooking/preparation salt was >800 mg/d among participants who were non-Hispanic black or lived in Alabama. In contrast, sodium from salt added during home food preparation was <300 mg/d among non-Hispanic Asians, participants in California and Minnesota, those with BMI <25, or those who reported they used salt in home food preparation occasionally. After adjustment for demographic characteristics, sodium intake from salt added during home food preparation differed by sex, race/ethnicity, study site, and BMI category (Table 3), but not by frequency of self-reported use during home food preparation. When analyses were further stratified by study site (Supplemental Table 2), results were consistent with data on salt added at the table, in that non-Hispanic white participants from Alabama seemed to add more salt during cooking/preparation than non-Hispanic white participants in Minnesota or California.
The mean measured frequency of salt added at the table was estimated as 0.22, i.e., on average, participants added salt to ≥1 foods at the table on ∼22% of the days assessed (Table 4). This measured frequency of use was greater for individuals with greater self-reported frequency of use among participants overall, and across subgroups. In some subgroups, however, estimates were statistically unreliable due to small sample sizes (Table 4). Overall, on average, persons who reported adding salt at the table “never/rarely” added salt at the table on 12% of the days assessed, ranging from 6% to 22% across subgroups; “occasionally” 27% overall, ranging from 13% to 42% across subgroups; and “very often” 52% overall, ranging from 43% to 66% across subgroups (Table 4).
TABLE 4.
Measured frequency of salt added at the table, across self-reported frequency of salt use, among US adults, overall and by demographic and health characteristics1
| “How often do you add ordinary salt to your food at the table?” | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Total | n | Never/rarely2 | n | Occasionally | n | Very often | P 3 | |
| Overall | 423 | 0.22 ± 0.014 | 255 | 0.12 ± 0.01 | 110 | 0.27 ± 0.03 | 58 | 0.52 ± 0.05 | <0.0001 |
| Sex | |||||||||
| Male | 208 | 0.20 ± 0.02 | 128 | 0.11 ± 0.02 | 55 | 0.27 ± 0.04 | 25 | 0.47 ± 0.08 | <0.0001 |
| Female | 215 | 0.24 ± 0.02 | 127 | 0.14 ± 0.02 | 55 | 0.27 ± 0.04 | 33 | 0.56 ± 0.06 | <0.0001 |
| Age, y | |||||||||
| <45 | 212 | 0.19 ± 0.02 | 123 | 0.10 ± 0.02 | 62 | 0.25 ± 0.04 | 27 | 0.49 ± 0.07 | <0.0001 |
| ≥45 | 211 | 0.24 ± 0.02 | 132 | 0.15 ± 0.02 | 48 | 0.30 ± 0.05 | 31 | 0.55 ± 0.07 | <0.0001 |
| Race or ethnicity | |||||||||
| Non-Hispanic white | 141 | 0.23 ± 0.03 | 81 | 0.13 ± 0.02 | 40 | 0.31 ± 0.05 | 20 | 0.46 ± 0.08 | <0.0001 |
| Non-Hispanic black | 91 | 0.37 ± 0.04 | 36 | 0.22 ± 0.05 | 31 | 0.36 ± 0.06 | 24 | 0.60 ± 0.07 | 0.0002 |
| Non-Hispanic Asian | 97 | 0.09 ± 0.02 | 76 | 0.06 ± 0.02* | 16 | 0.13 ± 0.06* | † | † | ‡ |
| Hispanic | 94 | 0.18 ± 0.03 | 62 | 0.14 ± 0.03 | 23 | 0.18 ± 0.05 | † | † | ‡ |
| Education | |||||||||
| High school graduate or less | 58 | 0.27 ± 0.05 | 30 | 0.16 ± 0.05* | 19 | 0.37 ± 0.09 | † | † | ‡ |
| Some college or technical school | 92 | 0.27 ± 0.04 | 47 | 0.14 ± 0.03 | 28 | 0.26 ± 0.05 | 17 | 0.66 ± 0.1 | <0.0001 |
| Bachelor's degree or greater | 271 | 0.19 ± 0.02 | 176 | 0.11 ± 0.02 | 63 | 0.25 ± 0.04 | 32 | 0.47 ± 0.05 | <0.0001 |
| Study site | |||||||||
| Alabama | 132 | 0.30 ± 0.03 | 64 | 0.15 ± 0.03 | 41 | 0.40 ± 0.06 | 27 | 0.53 ± 0.07 | <0.0001 |
| California | 145 | 0.13 ± 0.02 | 108 | 0.09 ± 0.02 | 31 | 0.19 ± 0.04 | † | † | ‡ |
| Minnesota | 146 | 0.22 ± 0.03 | 83 | 0.15 ± 0.03 | 38 | 0.19 ± 0.04 | 25 | 0.52 ± 0.08 | <0.0001 |
| Self-reported hypertension | |||||||||
| Yes | 97 | 0.27 ± 0.03 | 59 | 0.14 ± 0.03 | 25 | 0.42 ± 0.07 | 13 | 0.60 ± 0.10 | 0.0001 |
| No | 322 | 0.20 ± 0.02 | 192 | 0.11 ± 0.01 | 85 | 0.23 ± 0.03 | 45 | 0.50 ± 0.05 | <0.0001 |
| BMI,5 kg/m2 | |||||||||
| <25.0 | 149 | 0.16 ± 0.02 | 90 | 0.09 ± 0.02 | 42 | 0.21 ± 0.04 | 17 | 0.43 ± 0.08 | <0.0001 |
| 25.0–29.9 | 143 | 0.22 ± 0.03 | 93 | 0.12 ± 0.02 | 34 | 0.35 ± 0.06 | 16 | 0.52 ± 0.10 | <0.0001 |
| ≥30 | 131 | 0.28 ± 0.03 | 72 | 0.17 ± 0.03 | 34 | 0.27 ± 0.06 | 25 | 0.59 ± 0.07 | <0.0001 |
Data were from the Salt Sources Study, 2014. Measured frequency of salt added was assessed for each participant as the sum of the number of days a participant returned ≥1 bag with salt added at the table, divided by 4, the number of days of collection. Self-reported frequency of salt use was based on NHANES categoric frequency-of-use questions that were asked at the start of the 11-d window of data collection. *Estimates are considered statistically unreliable [relative SE (SE/estimate) ≥0.30)]. †Estimates suppressed for small cell sample sizes (n < 10). ‡P values not calculated for subgroups that include unreliable estimates or small cell sample sizes.
“Never” group included n = 28 participants; combined with “rarely.”
P values adjusted for age, sex, race/ethnicity, education, and BMI.
Mean ± SE (all such values).
BMI = weight (kg)/height (m)2; <25 = normal weight; 25.0–29.9 = overweight; ≥30 = obese .
The mean measured frequency of adding salt during home food preparation was 0.36, i.e., on average, participants added salt during preparation slightly more than one-third of the time (Table 5). Overall, measured frequency of salt use was greater for individuals who reported they added salt more often. However, this trend was not significant for some participant subgroups, including non-Hispanic black participants, participants with a high school degree or less, and participants with self-reported hypertension.
TABLE 5.
Measured frequency of salt added during home cooking/preparation, across self-reported frequency of salt use, among US adults, overall and by demographic and health characteristics1
| “How often is ordinary or seasoned salt added in cooking or preparing foods in your household” | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Total | n | Never/rarely2 | n | Occasionally | n | Very often | P 3 | |
| Overall | 423 | 0.36 ± 0.024 | 97 | 0.20 ± 0.03 | 132 | 0.28 ± 0.02 | 194 | 0.49 ± 0.03 | <0.0001 |
| Sex | |||||||||
| Male | 208 | 0.34 ± 0.02 | 39 | 0.21 ± 0.04 | 69 | 0.28 ± 0.04 | 87 | 0.49 ± 0.04 | <0.0001 |
| Female | 215 | 0.37 ± 0.02 | 58 | 0.19 ± 0.03 | 63 | 0.27 ± 0.03 | 113 | 0.49 ± 0.03 | <0.0001 |
| Age, y | |||||||||
| <45 | 212 | 0.31 ± 0.02 | 47 | 0.18 ± 0.03 | 71 | 0.23 ± 0.03 | 94 | 0.44 ± 0.03 | <0.0001 |
| ≥45 | 211 | 0.40 ± 0.02 | 50 | 0.23 ± 0.04 | 61 | 0.34 ± 0.04 | 100 | 0.53 ± 0.04 | <0.0001 |
| Race or ethnicity | |||||||||
| Non-Hispanic white | 141 | 0.30 ± 0.03 | 46 | 0.19 ± 0.04 | 51 | 0.26 ± 0.03 | 44 | 0.45 ± 0.05 | <0.0001 |
| Non-Hispanic black | 91 | 0.34 ± 0.04 | 28 | 0.25 ± 0.05 | 22 | 0.32 ± 0.07 | 41 | 0.41 ± 0.06 | 0.52 |
| Non-Hispanic Asian | 97 | 0.39 ± 0.04 | 15 | 0.12 ± 0.06* | 30 | 0.25 ± 0.04 | 52 | 0.55 ± 0.05 | ‡ |
| Hispanic | 94 | 0.42 ± 0.04 | † | † | 29 | 0.30 ± 0.06 | 57 | 0.51 ± 0.05 | ‡ |
| Education | |||||||||
| High school graduate or less | 58 | 0.40 ± 0.05 | 13 | 0.31 ± 0.09 | 15 | 0.33 ± 0.08 | 30 | 0.48 ± 0.07 | 0.41 |
| Some college or technical school | 92 | 0.32 ± 0.03 | 20 | 0.16 ± 0.06* | 23 | 0.30 ± 0.06 | 49 | 0.40 ± 0.05 | ‡ |
| Bachelor's degree or greater | 271 | 0.36 ± 0.02 | 64 | 0.19 ± 0.03 | 93 | 0.26 ± 0.03 | 114 | 0.53 ± 0.03 | <0.0001 |
| Study site | |||||||||
| Alabama | 132 | 0.29 ± 0.03 | 56 | 0.21 ± 0.03 | 42 | 0.30 ± 0.05 | 34 | 0.41 ± 0.06 | 0.02 |
| California | 145 | 0.35 ± 0.03 | 22 | 0.13 ± 0.05* | 52 | 0.22 ± 0.04 | 71 | 0.51 ± 0.04 | ‡ |
| Minnesota | 146 | 0.42 ± 0.03 | 19 | 0.25 ± 0.07 | 38 | 0.33 ± 0.05 | 89 | 0.50 ± 0.04 | 0.01 |
| Self-reported hypertension | |||||||||
| Yes | 97 | 0.35 ± 0.03 | 19 | 0.21 ± 0.05 | 33 | 0.32 ± 0.05 | 45 | 0.43 ± 0.05 | 0.16 |
| No | 322 | 0.36 ± 0.02 | 76 | 0.20 ± 0.03 | 98 | 0.26 ± 0.03 | 148 | 0.51 ± 0.03 | <0.0001 |
| BMI,5 kg/m2 | |||||||||
| <25.0 | 149 | 0.38 ± 0.03 | 30 | 0.21 ± 0.05 | 49 | 0.22 ± 0.03 | 70 | 0.56 ± 0.04 | <0.0001 |
| 25.0–29.9 | 143 | 0.31 ± 0.03 | 40 | 0.15 ± 0.03 | 47 | 0.33 ± 0.05 | 56 | 0.41 ± 0.04 | 0.0004 |
| ≥30 | 131 | 0.38 ± 0.03 | 27 | 0.27 ± 0.05 | 36 | 0.28 ± 0.04 | 68 | 0.48 ± 0.05 | 0.01 |
Data were from the Salt Sources Study, 2014. Measured frequency of salt added was assessed for each participant as the sum of the number of days a participant returned ≥1 bag with salt added at the table, divided by 4, the number of days of collection. Self-reported frequency of salt use was based on NHANES categoric frequency-of-use questions that were asked at the start of the 11-d window of data collection. *Estimates are considered statistically unreliable [relative SE (SE/estimate) ≥0.30)]. †Estimates suppressed for small cell sample sizes (n < 10). ‡P values not calculated for subgroups that include unreliable estimates or small cell sample sizes.
“Never” group included n = 28 participants; combined with “rarely.”
P values adjusted for age, sex, race/ethnicity, education, and BMI.
Mean ± SE (all such values).
BMI = weight (kg)/height (m)2; <25 = normal weight; 25.0–29.9 = overweight; ≥30 = obese.
Dicussion
In this diverse sample of adults living in 3 US geographic locations, the HP 2020 methodology for estimating sodium from salt added at the table appeared to accurately estimate mean total sodium intake among participants in this study overall and across selected demographic and health characteristics. In addition, our results suggest total intake based on the use of the HP 2020 methodology and measured intake are highly correlated. However, differences in the mean amounts of discretionary sodium consumed by assessment method suggest that the HP 2020 estimates may not reflect intake as well among certain population subgroups, such as those with hypertension or at a higher risk for hypertension. These differences appear related to substantial variabilities in the amounts of sodium consumed from salt added at the table, e.g., by race/ethnicity or geographic location. Thus, differences in total sodium intake by demographic or other population subgroups based on HP 2020 estimates may need to be interpreted with caution. The lack of association between self-reported and measured frequency of salt used during home food preparation in some subgroups supports the earlier decision by the USDA to discontinue the adjustment of the sodium content of foods based on NHANES questions (6).
Limited data exist on the validity of measures of self-reported frequency of discretionary salt use, and many studies that have examined salt use in relation to self-reported frequency are >20–30 y old. Data from these previous studies suggest some agreement between self-reported and measured table salt use, similar to the current study (9–12). However, none of these studies were conducted in a diverse sample of US adults, and differences in the questions used to determine frequency of salt use, years conducted, and study populations make results difficult to compare.
The participants in the Salt Sources Study differ in sociodemographic characteristics from the US population and from the original 1991 study used for HP 2020, which may explain some of the differences in results (6). In the original 1991 study, >70% participants were non-Hispanic white and all were from the northeastern US, whereas participants in the Salt Sources Study were purposely recruited from locations across 3 diverse US geographic regions and 4 racial/ethnic groups (6). Participants in the 1991 study were reported to be generally healthy, whereas in the Salt Sources Study, approximately a quarter of the sample self-reported hypertension (6). The HP 2020 methodology based on the 1991 study may not reflect consumption across key demographic and health characteristics in the US population.
Although surveys on attitudes and behaviors regarding sodium reduction indicate that non-Hispanic blacks are more likely to report taking action to reduce sodium intake and to use spices/herbs other than salt to flavor foods (13, 14), this sample of non-Hispanic black participants consumed more sodium from salt added at the table on days when salt was added than any other racial/ethnic group. However, in the current study the racial/ethnic differences may be related to regional differences in salt use, as the majority of the non-Hispanic black participants in the current study were recruited from the Alabama site. Both non-Hispanic white and non-Hispanic black participants from Alabama had similarly high sodium intakes from table salt (>1600 mg/d). Salt is a common component of the traditional diet in the southeastern United States, a diet that is associated with hypertension and cardiovascular disease (15, 16). Both non-Hispanic black participants and participants from Alabama, regardless of race or ethnicity, also added more sodium during cooking than other racial/ethnic groups and participants in other study sites. Previous findings on differences in sodium intake by race and region are mixed, with 1 study indicating that the southern region of the United States consumed the most sodium, but showed no difference in reported “salt added at the table” compared with other US regions (17). Other studies have reported that black individuals and individuals living in the “stroke belt” had a lower intake of sodium overall, although this may be due to use of an FFQ to estimate sodium intake (18, 19). The most recent publication from the Salt Sources Study found that sodium from salt added at the table accounted for >9% of total sodium intake among Alabama participants, compared with 2–4% in the other sites (3). Further investigation regarding these regional differences is warranted.
The current study results also suggest that on days that they add salt at the table, persons with less education or with hypertension may add more salt independent of differences in age, sex, race/ethnicity, and BMI. These differences were smaller when stratified by geographic location, but sample sizes were small. Socioeconomic status (SES) and education have been shown to be associated with dietary behaviors and diet quality, including use of discretionary salt (20, 21). Individuals with hypertension often report watching salt intake and adding salt to food less frequently than individuals without hypertension, but these reported behaviors may not always translate into lower sodium consumption, particularly if a greater amount of sodium is added at each occasion (22–24). Although adults with hypertension may consume slightly less sodium from foods and beverages than nonhypertensive adults, these estimates can exclude salt added at the table (25). Additional studies are required to replicate and extend the current study results for adults with hypertensive compared with without hypertension across additional geographic areas.
For those with less education or with self-reported hypertension (or both), self-reported frequency of salt added during cooking did not correspond as well with measured sodium intake on days salt was added, or with the measured frequency of adding salt during cooking. This may be explained by underreporting or difficulty in reporting, i.e., if the people who are reporting the amount of salt added during cooking or preparation are not the same as those who are doing the actual cooking or preparation, respondents may not know if salt is added during cooking or how often (6). Frequency of home cooking may also play a role; for example, some data indicate that lower educational attainment or SES may actually be associated with less frequent home cooking, whereas higher SES may lead to more time spent cooking at home (26–29), in which case interpretation of the frequency responses may differ.
Compared with responses in NHANES, the current analysis found similar results for self-reported very often and occasional table salt use; however, relatively few respondents in the current study reported never adding salt at the table (<10%) compared with 30% of adults in NHANES 2009–2012 (30). According to Mattes and Donnelly (6), 30% of all respondents to their recruitment advertisements indicated that they never used table or cooking salt. On the other hand, in the current study 56% (n = 237) of participants in the current study sample did not return a duplicate salt portion for salt added at the table. This may be an overestimate of the prevalence of participants not adding salt at the table as only 4 days were assessed across an 11-d period. Although mean sodium intake from table salt in this sample, on consumption days, was slightly lower than the estimate used from Mattes and Donnelly (400 compared with 580 mg/d), this may be related to differences in characteristics of participants or changes in salt use over time (6).
Some limitations of this analysis should be noted. First, participants in the Salt Sources Study were a convenience sample of adults from only 3 geographic locations that might not represent US discretionary salt use, particularly as some racial/ethnic groups were selected primarily from 1 geographic location. Most non-Hispanic black participants were from Birmingham, AL, whereas most non-Hispanic Asian participants were from Palo Alto, CA. Although the prevalence of hypertension and overweight/obesity in the study sample was similar to the US adult population, the sample was slightly more educated than the general population (31–33) and, by design, had higher proportions of non-Hispanic black, Hispanic, and Asian participants (33). Second, determining sodium intake from discretionary salt through the use of the duplicate portion method required reliance on participants’ data collection instead of a biomarker, and collection occurred on 4 days over a short period of time, which may not capture usual long-term salt use. Third, because so few participants reported consuming “light salt” or salt substitutes, we could not examine this group's intake separately. Fourth, participants’ 24-h dietary recalls were scheduled in advance, which may have influenced their diet on those days, and thereby their sodium intake.
Monitoring total dietary sodium intake in the United States is an important public health strategy, especially given efforts to reduce sodium content in the food supply. Previous analyses and HP 2020 have relied on self-reported responses from the NHANES discretionary salt use questions to estimate sodium intake from salt added at the table. However, these questions are subject to participant interpretation and have not been validated. In this study, the HP 2020 estimate of total sodium intake was significantly correlated with measured total sodium intake (when salt added at the table was measured from duplicate salt portion collection). However, differences by sociodemographic and geographic characteristics suggest that alternative estimates may be needed to examine sodium intake from salt added at the table or differences in total sodium intake by demographic and health characteristics. Further research is needed to replicate and extend these findings to determine if differences in salt use (amounts and frequency) are related to geographic location, types of foods consumed, or other factors. Estimating sodium intake through the use of urine biomarkers along with validating measures of discretionary salt use in specific subgroups could also provide more accurate information on total sodium intake. Accurate assessment of sodium consumed from discretionary sources can help in formulating and monitoring sodium reduction strategies particularly among groups for whom discretionary salt is a substantive contributor to intake.
Supplementary Material
Acknowledgments
The authors’ responsibilities were as follows—LJH, CDG, JMS, LMS, MEC, and ZSQ: designed the research; LJH, CDG, JMS, and LMS: conducted the research; ZSQ, LZ, and CM: analyzed the data or performed statistical analysis; ZSQ and MEC: wrote the paper; ZSQ: had primary responsibility for final content; and all authors: read and approved the final manuscript
Notes
Funding was provided through a research contract (Broad Agency Announcement) with the CDC (contract no. 200-2012-M-52155). The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC. Additional funding was also provided (to ZSQ) by the National Heart, Lung, and Blood Institute (grant no. T32HL130025).
Author disclosures: ZSQ, LZ, LJH, CDG, JMS, LMS, CG, AM, and MEC, no conflicts of interest.
Supplemental Figure 1, Supplemental Methods and Supplemental Tables 1 and 2 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.
References
- 1. Eckel RH, Jakicic JM, Ard JD, de Jesus JM, Miller NH, Hubbard VS, Lee I-M, Lichtenstein AH, Loria CM, Millen BE et al.. 2013 AHA/ACC Guideline on lifestyle management to reduce cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129:S76–99. [DOI] [PubMed] [Google Scholar]
- 2. US Department of Health and Human Services, Office of Disease Prevention and Health Promotion. Healthy People 2020 topics and objectives[Internet]. Updated November 23, 2016; [cited 2016 Nov 25] Available from: https://www.healthypeople.gov/2020/topics-objectives. [Google Scholar]
- 3. Harnack LJ, Cogswell ME, Shikany JM, Gardner CD, Gillespie C, Loria CM, Zhou X, Yuan K, Steffen LM. Sources of sodium in US adults from 3 geographic regions. Circulation. 2017;135:1775–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bibbins-Domingo K, Chertow GM, Coxson PG, Moran A, Lightwood JM, Pletcher MJ, Goldman L. Projected effect of dietary salt reductions on future cardiovascular disease. N Engl J Med. 2010;362(7):590–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Sebastian RS, Wilkinson Enns C, Steinfeldt LC, Goldman JD, Moshfegh AJ. Monitoring sodium intake of the US population: impact and implications of a change in what we eat in America, National Health and Nutrition Examination Survey dietary data processing. J Acad Nutr Diet. 2013;113(7):942–9. [DOI] [PubMed] [Google Scholar]
- 6. Mattes RD, Donnelly D. Relative contributions of dietary sodium sources. J Am Coll Nutr. 1991;10(4):383–93. [DOI] [PubMed] [Google Scholar]
- 7. Johnson CL, Paulose-Ram R, Ogden CL, Carroll MD, Kruszon-Moran D, Dohrmann SM, Curtin LR. National Health and Nutrition Examination Survey: analytic guidelines, 1999–2010. National Center for Health Statistics. Vital Health Stat. 2013;2(161):1–24. [PubMed] [Google Scholar]
- 8. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed.Mahwah (NJ): Lawrence Erlbaum; 1988. [Google Scholar]
- 9. Kumanyika SK, Jones DY. Patterns of week-to-week table salt use by men and women consuming constant diets. Hum Nutr Appl Nutr. 1983;37(5):348–56. [PubMed] [Google Scholar]
- 10. Pietinen P, Tanskanen A, Tuomilehto J. Assessment of sodium intake by a short dietary questionnaire. Scand J Soc Med. 1982;10(3):105–12. [DOI] [PubMed] [Google Scholar]
- 11. Mittelmark MB, Sternberg B. Assessment of salt use at the table: comparison of observed and reported behavior. Am J Public Health. 1985;75(10):1215–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Nowson C, Lim K, Grimes C, O'Halloran S, Land MA, Webster J. Dietary salt intake and discretionary salt use in two general population samples in Australia: 2011 and 2014. Nutrients. 2015;7(12):10501–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Patel D, Cogswell ME, John K, Creel S, Ayala C. Knowledge, attitudes, and behaviors related to sodium intake and reduction among adult consumers in the United States. Am J Health Promot. 2017;31(1):68–75. [DOI] [PubMed] [Google Scholar]
- 14. Jackson SL, Coleman King SM, Park S, Fang J, Odom EC, Cogswell ME. Health professional advice and adult action to reduce sodium intake. Am J Prev Med. 2016;50(1):30–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Shikany J, Safford M, Newby PK, Durant R, Brown T, Judd S. Southern dietary pattern is associated with hazard of acute coronary heart disease in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study. Circulation. 2015;132(9):804–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Smith S, Quandt S, Arcury T, Wetmore L, Bell R, Vitolins M. Aging and eating in the rural, southern United States: beliefs about salt and its effect on health. Soc Sci Med. 2006;62(1):189–98. [DOI] [PubMed] [Google Scholar]
- 17. Hajjar I, Kotchen T. Regional variations of blood pressure in the United States are associated with regional variations in dietary intakes: the NHANES-III data. J Nutr. 2003;133:211–4. [DOI] [PubMed] [Google Scholar]
- 18. Newby PK, Noel SE, Grant R, Judd S, Shikany JM, Ard J. Race and region are associated with nutrient intakes among black and white men in the United States. J Nutr. 2011;141:296–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Newby PK, Noel SE, Grant R, Judd S, Shikany JM, Ard J. Race and region have independent and synergistic effects on dietary intake in black and white women. Nutr J. 2012;11:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sarmugam R, Worsley A, Wang W. An examination of the mediating role of salt knowledge and beliefs on the relationship between socio-demographic factors and discretionary salt use: a cross-sectional study. Int J Behav Nutr Phys Act. 2013;10:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kant AK, Graubard BI. Secular trends in the association of socio-economic position with self-reported dietary attributes and biomarkers in the US population: National Health and Nutrition Examination Survey (NHANES) 1971–1975 to NHANES 1999–2002. Public Health Nutr. 2007;10(2):158–67. [DOI] [PubMed] [Google Scholar]
- 22. Fang J, Cogswell ME, Park S, Jackson SL, Odom EC. Sodium intake among U.S. adults—26 states, the District of Columbia, and Puerto Rico, 2013. MMWR Morb Mortal Wkly Rep. 2015;64(25):695–8. [PMC free article] [PubMed] [Google Scholar]
- 23. Ayala C, Gillespie C, Cogswell M, Keenan NL, Merritt R. Sodium consumption among hypertensive adults advised to reduce their intake: national health and nutrition examination survey, 1999–2004. J Clin Hypertens (Greenwich). 2012;14(7):447–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Slade AN, Kim H.. Dietary responses to a hypertension diagnosis: evidence from the National Health and Nutrition Examination Survey (NHANES) 2007–2010. Behav Med. 2014;40(1):1–13. [DOI] [PubMed] [Google Scholar]
- 25. Jackson SL, King SM, Zhao L, Cogswell ME. Prevalence of excess sodium intake in the United States—NHANES, 2009–2012. MMWR Morb Mortal Wkly Rep. 2016;64(52):1393–7. [DOI] [PubMed] [Google Scholar]
- 26. Virudachalam S, Long JA, Harhay MO, Polsky DE, Feudtner C. Prevalence and patterns of cooking dinner at home in the USA: National Health and Nutrition Examination Survey (NHANES) 2007–2008. Public Health Nutr. 2014;17(5):1022–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. McGowan L, Pot GK, Stephen AM, Lavelle F, Spence M, Raats M, Hollywood L, McDowell D, McCloat A, Mooney E et al.. The influence of socio-demographic, psychological and knowledge-related variables alongside perceived cooking and food skills abilities in the prediction of diet quality in adults: a nationally representative cross-sectional study. Int J Behav Nutr Phys Act. 2016;13(1):111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Thornton LE, Crawford DA, Ball K. Who is eating where? Findings from the SocioEconomic Status and Activity in Women (SESAW) study. Public Health Nutr. 2011;14(3):523–31. [DOI] [PubMed] [Google Scholar]
- 29. Mancino L, Newman C. Who has time to cook? How family resources influence food preparation. [Internet]. Economic Research Report no. 40. Washington (DC). Economic Research Service, US Department of Agriculture, 2007. [Accessed 2017 July 18]. Available from: https://www.ers.usda.gov/webdocs/publications/45797/11722_err40_1_.pdf?v=41056. [Google Scholar]
- 30. Quader ZS, Patel S, Gillespie C, Cogswell ME, Gunn JP, Perrine CG, Mattes RD, Moshfegh A. Trends and determinants of discretionary salt use: National Health and Nutrition Examination Survey 2003–2012. Public Health Nutr. 2016;19(12):2195–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Yoon SS, Fryar CD, Carroll MD. Hypertension prevalence and control among adults: United States, 2011–4.. NCHS Data Brief, no 220. Hyattsville (MD): National Center for Health Statistics; 2015. [PubMed] [Google Scholar]
- 32. Ogden C, Carroll M, Kit B, Glegal K. Prevalence of childhood and adult obesity in the United States, 2011–2012. J Am Med Assoc. 2014;311:806–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. US Census Bureau. 2010 census data. [Internet]. Available from: https://www.census.gov/2010census/data/. [Accessed 2017 July 18]. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.