Skip main navigation

Urinary Sodium Excretion, Blood Pressure, and Risk of Future Cardiovascular Disease and Mortality in Subjects Without Prior Cardiovascular Disease

Originally published 2019;73:1202–1209


Hypertension is a risk factor for cardiovascular disease. Increased urinary sodium excretion, representing dietary sodium intake, is associated with hypertension. Low sodium intake has been associated with increased mortality in observational studies. Further studies should assess whether confounding relationships explain associations between sodium intake and outcomes. We studied UK Biobank participants (n=457 484; mean age, 56.3 years; 44.7% men) with urinary electrolytes and blood pressure data. Estimated daily urinary sodium excretion was calculated using Kawasaki formulae. We analyzed associations between sodium excretion and blood pressure in subjects without cardiovascular disease, treated hypertension, or diabetes mellitus at baseline (n=322 624). We tested relationships between sodium excretion, incidence of fatal and nonfatal cardiovascular disease, heart failure, and mortality. Subjects in higher quintiles of sodium excretion were younger, with more men and higher body mass index. There was a linear relationship between increasing urinary sodium excretion and blood pressure. During median follow-up of 6.99 years, there were 11 932 deaths (1125 cardiovascular deaths) with 10 717 nonfatal cardiovascular events. There was no relationship between quintile of sodium excretion and outcomes. These relationships were unchanged after adjustment for comorbidity or excluding subjects with events during the first 2 years follow-up. No differing risk of incident heart failure (1174 events) existed across sodium excretion quintiles. Urinary sodium excretion correlates with elevated blood pressure in subjects at low cardiovascular risk. No pattern of increased cardiovascular disease, heart failure, or mortality risk was demonstrated with either high or low sodium intake.


Hypertension is a modifiable risk factor for development of cardiovascular disease (CVD)—the leading cause of death in high income countries.1,2 Large cohort studies showed associations between high sodium consumption and raised blood pressure.3–5 Clinical trials have demonstrated that reduction in sodium intake leads to lowering of blood pressure,6–8 and international guidelines advocate a dietary sodium intake of 1.5 to 2.4 g per day.9 However, observational cohort studies (in ≈3600 to 102 000 participants) consistently demonstrate J-shaped relationships between sodium intake and CVD risk whereby low sodium intake is associated with higher cardiovascular risk.10–12 In the Prospective Urban Rural Epidemiology study (n=101 945), both low (<3 g/day) and high (>6 g/day) sodium intakes were associated with higher risk of CVD. Some observational data controversially suggest that salt intake relates positively to life expectancy.13 Thus, conventional dietary guidance on reduction of sodium intake conflicts with observational population data regarding CVD risk, even after correction for baseline CVD, blood pressure, and cholesterol.10

One explanation often proposed is of “reverse causality” where subjects with comorbid conditions (eg, hypertension and heart failure) have low sodium intake either as recommended therapy or are malnourished because of poor dietary (and sodium) intake. This explanation becomes less robust in large cohort studies using appropriate statistical analysis to account for comorbidities and baseline individual cardiovascular risk profile.3,10,14 We aimed to test whether reverse causality accounted for J-shaped relationships between sodium intake and CVD or mortality.

Materials and Methods

UK Biobank data are available on application to the UK Biobank for data access (, and authors have returned derived data to UK Biobank as per their requirements. Between April 2007 and December 2010, over 500 000 participants were recruited by UK Biobank and had baseline measurements recorded at 1 of 22 assessment centers in Scotland, England, or Wales, as described elsewhere.15 UK Biobank received ethical approval from the North West Multi-Center Research Ethics Committee (REC reference: 11/NW/03820). All participants gave written informed consent before enrollment, which was conducted in accord with the principles of the Declaration of Helsinki.

Ethnicity was coded as white, South Asian, black, or other, with white used as the referent group. Body mass index (BMI) was calculated as weight (kg)/height2 (m). Smoking status was self-reported and categorized as never, former, or current smoker, and area-based sociodemographic status was derived from postcode of residence, using Townsend score,16 where higher scores relate to greater deprivation. Diagnosis by a doctor of diabetes mellitus (DM), stroke, myocardial infarction, transient ischemic attack, or brain hemorrhage was self-reported at baseline. Family history of CVD (stroke or heart disease in participant’s mother, father, or sibling) was self-reported.

A midstream urine sample collected was refrigerated between 2 and 8°C until potentiometric measurement for sodium and potassium content using a Beckman Coulter, Ltd (United Kingdom), AU5400 analyzer. Sodium and potassium concentrations (in mmol/L) were capped at sex-specific mean ±4 SDs. Kawasaki formulae17 were used to convert spot sodium and potassium measurements into estimated 24-hour excretion (eNa24h and eK24h, in g/day). Two resting blood pressure measurements were taken, according to a standard operating procedure ( at least 1 minute apart, by trained staff, using either an automated sphygmomanometer with an appropriate cuff size relative to the participant’s arm or manually if an automated reading was not possible. Mean systolic and diastolic blood pressures were calculated and used to derive an estimated mean arterial blood pressure (MABP); MABP=(systolic blood pressure+2×diastolic blood pressure)/3, and pulse pressure=(systolic blood pressure−diastolic blood pressure). Lean body mass as fat-free mass was derived from whole-body bioimpedance measures using the Tanita BC418MA body composition analyzer. For full data collection protocols, refer to UK Biobank online (

The process for generation of the cohort for analysis is shown in Figure S1 in the online-only Data Supplement. Charlson comorbidity index was calculated for all subjects,18 and a secondary analysis was conducted after excluding those participants with Charlson score >0. Use of diuretics and blood pressure medications were detected by searching medicines reported by participants in the 63rd revision of the British National Formulary.

CVD events (myocardial infarction, stroke, heart failure admission International Classification of Diseases, Tenth Revision, codes I20-23, I24.1, I25.2, I60, I61, I62, I63, or I64) were identified by linkage with routine hospital data, and date and cause of death (where appropriate) were obtained from death certificates held by the National Health Service Information Centre for participants from England and Wales and the National Health Service Central Register Scotland for participants from Scotland. All-cause mortality (ACM) was defined as death from all causes. Start of follow-up was date of assessment, and end of follow-up for mortality end points was recorded as the end of data collection for the assessment center attended (November 30, 2015, for centers in Scotland; January 31, 2016, for centers in England/Wales), or date of death, whichever came first. For nonfatal outcomes, end of follow-up was date of death, first date of hospitalization for nonfatal CVD, or end of follow-up.

Statistical Analyses

We excluded participants hospitalized within 30 days of assessment or those missing blood pressure measurements from all analyses. Continuous variables (age, BMI [or lean body mass], Townsend deprivation score, eNa24h) were summarized as mean±SD or median±interquartile range (Q1–Q3) if the distribution skewed. Categorical variables (sex, ethnicity, family history of CVD, smoking status) were summarized as numbers and percentages. MABP and eNa24h were divided into sex-specific quintiles to lessen impact of outliers. Independent variables were examined for differences across quintiles of eNa24h and MABP. Differences between means or groups were tested using ANOVA, Kruskal-Wallis, or Pearson χ2 tests for normal, skewed continuous or categorical variables, respectively.

Blood Pressure Models

Associations of MABP with sex-specific quintiles of eNa24h were examined using sex-specific unadjusted linear regression models. Each model was then adjusted for the following; age, lean body mass, ethnicity, family history of CVD, smoking status, eK24h, and deprivation score. Linearity of associations was further explored using restricted cubic splines (data not shown).

CVD Models

Associations of fatal or nonfatal CVD or ACM with eNa24h were examined using eNa24h in sex-specific quintiles in Cox proportional hazard models. In these main analyses, we excluded those with treated hypertension, baseline CVD, or baseline DM (Figure S1). All models were adjusted for age, Townsend deprivation index, eK24h, MABP (used as continuous variable), smoking status, ethnicity, lean body mass, family history of CVD, rheumatoid arthritis, atrial fibrillation, and a binary variable for Charlson comorbidity score >0. We checked for effect mediation by MABP through models that did not adjust for this variable. The proportional hazard assumption was checked by visual inspection of Schöenfeld residuals.

To minimize effect of reverse causality on predictors of early CVD events, we performed additional landmark sensitivity analyses, analyzing events occurring after 2 years follow-up. Further sensitivity analyses were conducted excluding participants with Charlson comorbidity index of >0.

Incident heart failure was treated as specific event of interest, on the basis that excess sodium ingestion may lead to sodium and water retention triggering heart failure. Cox proportional hazard models were used to test for associations with eNa24h quintiles, using the methods as for other CVD events.

All analyses were performed using STATA 14 (StataCorp LP, College Station). A P of <0.05 was considered statistically significant.


Baseline Demographics, Blood Pressure, and Urinary Sodium Excretion in All Subjects and Those Without Comorbid Disease

Baseline characteristics for all subjects categorized by quintile of MABP are shown in Table 1. Subjects in the highest MABP quintiles were more likely to be older, have greater BMI and lean body mass, a higher prevalence of family history of CVD, less likely to have a positive smoking history, reside in an area of greater socioeconomic deprivation, and have greater eNa24h excretion. When categorized by quintiles of eNa24h excretion (Table 2), subjects in the highest eNa24h quintiles were younger, had higher BMI and lean body mass, lower prevalence of family history of CVD, lower smoking prevalence, more likely to reside in areas of lower socioeconomic deprivation, and have higher MABP. Excluding those with treated hypertension, baseline CVD, or DM did not substantially attenuate these results (Tables S1 and S2). In regression models assessing relationships between eNa24h and MABP performed in subjects without CVD, DM, or treated hypertension, there were linear relationships between increasing quintiles of eNa24h and increased MABP in both women and men (Figure 1). These associations were consistent before and after adjustment for baseline characteristics and were similar when restricted to subjects with Charlson score of zero to minimize influence of residual confounding.

Table 1. Distribution of Variables by Quintiles of MABP in Women (Upper) and Men (Lower) Among All Included UK Biobank Participants

FactorMABP, mm Hg; QuintilesP Value
50.7–88.7; n=51 40089–95.7; n=51 21896–102; n=49 240102.3–110; n=51 163110.3–174.7; n=49 993
48.3–93.7; n=41 58694–100.3; n=41 446100.7–106.3; n=40 772106.7–113.7; n=40 642114–177.7; n=40 024
Age, y52.92 (8.06)55.08 (8.08)56.50 (7.82)57.73 (7.53)58.75 (7.19)<0.001
54.45 (8.54)55.52 (8.40)56.41 (8.19)57.32 (7.89)58.15 (7.53)<0.001
Ethnic backgroundWhite46 785(91.4)47 076 (92.3)45 242 (92.2)47 071 (92.4)45 829 (92.0)<0.001
37 102(89.8)37 715 (91.6)37 307 (92.0)37 466 (92.7)37 032 (93.0)<0.001
South Asian735 (1.4)677 (1.3)677 (1.4)653 (1.3)679 (1.4)
957 (2.3)807 (2.0)758 (1.9)673 (1.7)545 (1.4)
Black660 (1.3)742 (1.5)803 (1.6)932 (1.8)1155 (2.3)
724 (1.8)610 (1.5)571 (1.4)585 (1.4)636 (1.6)
Other3011 (5.9)2529 (5.0)2326 (4.7)2274 (4.5)2136 (4.3)
2553 (6.2)2057 (5.0)1922 (4.7)1692 (4.2)1610 (4.0)
BMI, kg/m224.70 (3.95)26.27 (4.59)27.31 (5.07)28.08 (5.30)28.76 (5.66)<0.001
26.14 (3.82)27.35 (3.97)28.00 (4.16)28.45 (4.21)28.80 (4.29)<0.001
Lean body mass, kg43.53 (4.33)44.17 (4.71)44.62 (5.05)44.90 (5.20)45.19 (5.50)<0.001
62.30 (7.42)63.52 (7.61)63.98 (7.77)64.21 (7.87)64.17 (7.99)<0.001
Family history of CVD26 162(50.9)28 586 (55.8)29 067 (59.0)31 471 (61.5)31 378 (62.8)<0.001
20 451(49.2)21 184 (51.1)21 356 (52.4)22 401 (55.1)22 367 (55.9)<0.001
Smoking statusNever30 075 (58.6)30 555 (59.7)29 646 (60.3)30 863 (60.4)30 904 (61.9)<0.001
21 726 (52.3)21 298 (51.4)20 391 (50.1)19 975 (49.2)19 246 (48.1)<0.001
Former15 501 (30.2)15 726 (30.7)15 237 (31.0)16 303 (31.9)15 572 (31.2)
13 564 (32.7)14 581 (35.2)15 460 (38.0)16 092 (39.6)16 330 (40.8)
Current5754 (11.2)4887 (9.6)4300 (8.7)3943 (7.7)3475 (7.0)
6243 (15.0)5521 (13.3)4873 (12.0)4537 (11.2)4418 (11.0)
Townsend deprivation index−1.25 (3.08)−1.37 (3.02)−1.37 (3.02)−1.39 (3.01)−1.43 (2.99)<0.001
−1.07 (3.26)−1.27 (3.15)−1.34 (3.10)−1.41 (3.07)−1.38 (3.07)<0.001
Estimated Na24h, g3.57 (2.92–4.26)3.68 (3.01–4.39)3.75 (3.06–4.47)3.82 (3.11–4.56)3.92 (3.21–4.67)<0.001
4.16 (3.43–4.93)4.32 (3.60–5.08)4.41 (3.65–5.18)4.46 (3.69–5.23)4.53 (3.75–5.31)<0.001
Rheumatoid arthritis563 (1.1%)656 (1.3%)698 (1.4%)781 (1.5%)831 (1.7%)<0.001
263 (0.6%)271 (0.7%)281 (0.7%)291 (0.7%)335 (0.8%)0.005
Atrial fibrillation130 (0.3%)108 (0.2%)120 (0.2%)116 (0.2%)115 (0.2%)0.68
302 (0.7%)214 (0.5%)224 (0.5%)174 (0.4%)230 (0.6%)<0.001
Baseline CVD3736 (7.3%)7459 (14.6%)11 021 (22.4%)16 639 (32.5%)22 991 (46.0%)<0.001
6484 (15.6%)9208 (22.2%)12 066 (29.6%)15 759 (38.8%)19 520 (48.8%)<0.001
Antihypertensive medications2592 (5.0%)5420 (10.6%)7984 (16.2%)11 600 (22.7%)13 824 (27.7%)<0.001
4936 (11.9%)7010 (16.9%)8884 (21.8%)10 955 (27.0%)11 644 (29.1%)<0.001
DM1200 (2.3%)1755 (3.4%)2010 (4.1%)2207 (4.3%)1894 (3.8%)<0.001
2534 (6.1%)2777 (6.7%)2854 (7.0%)2649 (6.5%)2128 (5.3%)<0.001

Numbers in parentheses are percentages for ordinal variables, SDs for continuous variables, or Q1 and Q3 for skewed variables. Tests of differences between groups are χ2 and 1-way ANOVA. BMI indicates body mass index; DM, diabetes mellitus; and MABP, mean arterial blood pressure.

Table 2. Distribution of Variables by Quintiles of Na24h in Women (Upper) and Men (Lower) Among All Included UK Biobank Participants

FactorEstimated Urinary Sodium Excretion, g/d; QuintilesP Value
0.71–2.89, n=47 830; 0.74–3.43, n=38 1932.89–3.48, n=47 829; 3.43–4.08, n=38 1933.48–4.01, n=47 829; 4.08–4.66, n=38 1934.01–4.66, n=47 829; 4.66–5.35, n=38 1934.66–7.00, n=47 829; 5.35–7.00, n=38 192
Age, y57.37 (7.88)56.45 (7.93)56.07 (7.95)55.84 (7.98)55.31 (8.07)<0.001
58.10 (7.86)57.11 (8.05)56.43 (8.18)55.77 (8.22)54.71 (8.32)<0.001
Ethnic backgroundWhite44 421 (93.2%)44 256 (92.9%)44 116 (92.5%)43 912 (92.2%)43 179 (90.7%)<0.001
34 916 (91.9%)35 163 (92.5%)35 039 (92.2%)35 040 (92.2%)34 602 (91.1%)<0.001
South Asian410 (0.9%)464 (1.0%)494 (1.0%)669 (1.4%)1057 (2.2%)
644 (1.7%)563 (1.5%)606 (1.6%)696 (1.8%)787 (2.1%)
Black700 (1.5%)765 (1.6%)806 (1.7%)787 (1.7%)804 (1.7%)
586 (1.5%)556 (1.5%)587 (1.5%)542 (1.4%)577 (1.5%)
Other2110 (4.4%)2178 (4.6%)2262 (4.7%)2280 (4.8%)2558 (5.4%)
1830 (4.8%)1720 (4.5%)1762 (4.6%)1713 (4.5%)2005 (5.3%)
BMI, kg/m226.48 (4.82)26.40 (4.73)26.67 (4.85)27.09 (5.07)28.05 (5.65)<0.001
26.99 (3.91)27.18 (3.82)27.44 (3.91)27.91 (4.08)28.69 (4.54)<0.001
Lean body mass by impedance, kg43.97 (4.74)43.97 (4.71)44.27 (4.82)44.57 (4.96)45.33 (5.37)<0.001
62.15 (7.49)62.68 (7.33)63.29 (7.46)64.06 (7.63)65.39 (8.06)<0.001
Family history of CVD28 404 (59.4%)27 682 (57.9%)27 647 (57.8%)27 446 (57.4%)27 525 (57.5%)<0.001
20 815 (54.5%)20 395 (53.4%)20 069 (52.5%)19 988 (52.3%)19 521 (51.1%)<0.001
Smoking statusNever28 754 (60.2%)29 186 (61.1%)29 098 (60.9%)28 840 (60.4%)28 094 (58.8%)<0.001
19 292 (50.6%)19 743 (51.7%)19 632 (51.4%)19 105 (50.1%)18 573 (48.7%)<0.001
Former14 917 (31.2%)14 539 (30.4%)14 692 (30.7%)14 788 (30.9%)15 151 (31.7%)
14 051 (36.8%)14 025 (36.7%)14 018 (36.7%)14 332 (37.6%)14 528 (38.1%)
Current4111 (8.6%)4060 (8.5%)4000 (8.4%)4159 (8.7%)4531 (9.5%)
4818 (12.6%)4402 (11.5%)4508 (11.8%)4713 (12.4%)5047 (13.2%)
Townsend deprivation index−1.48 (2.98)−1.50 (2.96)−1.45 (2.97)−1.37 (3.01)−1.16 (3.11)<0.001
−1.32 (3.17)−1.49 (3.06)−1.44 (3.05)−1.35 (3.07)−1.10 (3.17)<0.001
MABP, mm Hg97.00 (89.00–105.67)97.67 (89.67–106.67)98.67 (90.33–107.33)99.67 (91.33–108.33)101.33 (93.00–110.00)<0.001
101.67 (93.67–110.00)102.33 (94.67–110.67)103.00 (95.33–111.00)104.00 (96.33–112.00)105.00 (97.67–113.00)<0.001
Rheumatoid arthritis803 (1.7%)579 (1.2%)636 (1.3%)542 (1.1%)727 (1.5%)<0.001
335 (0.9%)258 (0.7%)234 (0.6%)241 (0.6%)252 (0.7%)<0.001
Atrial fibrillation138 (0.3%)105 (0.2%)97 (0.2%)107 (0.2%)96 (0.2%)0.029
321 (0.8%)211 (0.6%)210 (0.5%)164 (0.4%)161 (0.4%)<0.001
Baseline CVD11 807 (24.7%)10 112 (21.1%)10 308 (21.6%)11 176 (23.4%)14 124 (29.5%)<0.001
12 497 (32.7%)10 499 (27.5%)10 471 (27.4%)11 194 (29.3%)12 867 (33.7%)<0.001
Antihypertensive medications7741 (16.2%)6251 (13.1%)6611 (13.8%)7476 (15.6%)10 425 (21.8%)<0.001
8670 (22.7%)6864 (18.0%)6932 (18.1%)7571 (19.8%)9355 (24.5%)<0.001
DM1624 (3.4%)1410 (2.9%)1440 (3.0%)1627 (3.4%)2145 (4.5%)<0.001
2423 (6.3%)2075 (5.4%)2080 (5.4%)2208 (5.8%)2792 (7.3%)<0.001

BMI indicates body mass index; CVD, cardiovascular disease; DM, diabetes mellitus; and MABP, mean arterial blood pressure.

Figure 1.

Figure 1. Linear regression analyses describing the association between mean arterial blood pressure (MABP) and sex-specific quintiles of eNa24h. Unadjusted (A and C) and fully adjusted models are shown (B and D), using all participants (A and B; n=322 624) and excluding participants with a Charlson score of >0 (C and D; n=160 920). All plots exclude participants with baseline cardiovascular disease, diabetes mellitus, or treated hypertension. Na24h Quintiles represent quintile of eNa24h as per Table 2.

Relationship Between Urinary Sodium Excretion, Future CVD Events, Heart Failure Admissions, and ACM in Subjects Without Baseline CVD

In 322 624 subjects without baseline CVD, DM, and treated hypertension during median follow-up time of 6.99 years (interquartile range, 6.29–7.64 years), there were 6742 deaths, 3016 of which were in women (44.7%). There were 740 fatal CVD events in men and 364 in women, and 6972 nonfatal CVD events in men and 3739 in women.

There was no association between quintiles of eNa24h and risk of CVD events or ACM in any model (Figure 2). These relationships were unaffected by adjustment for self-reported chronic kidney disease or statin use (Figure S2). These relationships did not change either after limiting the adjustment model to not include MABP as a potential effect mediator, or following landmark analysis excluding subjects who died within 2 years of follow-up, or after substituting lean body mass with conventional BMI (Figure S3). Data on the incidence of subtypes of CVD (myocardial infarction, ischemic stroke, and hemorrhagic stroke) are shown in Table S3 and Figure S4.

Figure 2.

Figure 2. Adjusted hazard ratios of sex-specific quintiles of eNa24h on the risk of all-cause mortality and fatal or nonfatal cardiovascular disease events. All-cause mortality is shown in upper panels (A and B) and nonfatal cardiovascular disease events in lower panels (C and D). Analyses are for all participants (A and C; n=322 624) and excluding participants with a Charlson score of >0 (B and D; n=160 920). All plots exclude participants with baseline cardiovascular disease, diabetes mellitus, or treated hypertension. Na24h Quintiles represent quintile of eNa24h as per Table 2.

There were 283 incident heart failure hospitalizations during follow-up in the 322 624 participants without baseline CVD, DM, or treated hypertension. There were no statistically significant differences in risk of incident heart failure across eNa24h quintiles in either men or women (Figure 3).

Figure 3.

Figure 3. Adjusted hazard ratios of sex-specific quintiles of eNa24h on incident heart failure in 322 624 UK Biobank participants, excluding those with baseline cardiovascular disease, diabetes mellitus, or treated hypertension. Na24h Quintiles represent quintile of eNa24h as per Table 2.


In a large prospective cohort study, we demonstrate a consistent relationship between estimated urinary sodium excretion as a marker of sodium intake and elevation in blood pressure. These relationships were present both in the whole cohort and when restricted to subjects free of baseline comorbidity, specified by restricting analysis to those without baseline CVD, DM, or treated hypertension and with Charlson comorbidity score of zero. These findings are consistent with other cohort studies, interventional clinical trials, and meta-analyses.3–8

The mechanisms by which dietary sodium contributes to raised blood pressure have been reviewed extensively elsewhere.19 Sodium excess has been associated with excess activation of the sympathetic nervous system, increased vascular tone, and endothelial dysfunction, as well as end organ damage with left ventricular hypertrophy, glomerulosclerosis, and arteriolosclerosis. In industrialized countries, diets rich in sodium and low in potassium are the norm. Processed meals are rich in sodium. Societies following a nonindustrialized diet low in sodium have low prevalence (<1%) of hypertension.20–22 Meta-analyses demonstrate that reduction in dietary sodium intake is associated with average reduction in systolic blood pressure of 3 to 5.4 mm Hg in hypertensive subjects.7,8 Our data add weight to the hypothesis that reduction of dietary sodium intake may lead to lower blood pressure. Unfortunately, we do not have long-term blood pressure changes to assess the influence of sodium intake on development of hypertension.

The relationship between sodium excretion and future nonfatal CVD and ACM is more complex and essentially showed no strong relationship between higher or lower sodium intake with risk of mortality or CVD. These null relationships did not appear to be altered by adjustment for comorbid disease using the Charlson comorbidity index. The landmark analysis excluding subjects who died within 2 years of follow-up showed similar lack of a relationship between sodium excretion and risk of CVD death. This refutes the perception that reverse causality underpins relationships whereby those with the lowest sodium intake have greater mortality risk because of underlying (perhaps undiagnosed) disease.

Our results demonstrating a lack of straightforward linear relationships between high sodium intake and increased risk of mortality of CVD are broadly in keeping with other studies. Although much public policy and conventional advice focuses on benefits of reduction of sodium intake on future cardiovascular health, large cohort studies, including subjects with comorbidities associated with increased CVD risk, such as chronic kidney disease and DM, have demonstrated there is a nadir of sodium consumption below which increased risk of future CVD or ACM increases.12,23–27 Alternatively, lower sodium intake may risk via activation of the renin-angiotensin-aldosterone system to maintain sodium and water homeostasis exposing the cardiovascular system to the deleterious effects of aldosterone.28 As heart failure represents a syndrome of renal salt avidity driving fluid retention, in the setting of cardiac dysfunction,29 we specifically analyzed the influence of sodium intake on incidence of heart failure hospitalization. However, we could find no evidence that sodium excretion as a marker for sodium intake was associated with incident heart failure, albeit in a small number of events. Alternatively, if renal sodium avidity was expected to be a driver for future heart failure, one might expect lower sodium excretion to precede future heart failure events, but this was not the case. Unlike Prospective Urban Rural Epidemiology, we do not demonstrate any association with low sodium excretion on CVD events or ACM supporting the notion that public health policy reducing sodium intake in healthy individuals is not harmful and will be associated with reduced blood pressure. It may take longer follow-up than this study to demonstrate the benefits of sodium reduction on CVD events in low-risk individuals.

Only a large randomized trial of differing sodium intake targets will address what the optimal sodium intake should be in an industrialized nation. In the face of public health advice, competing pressures from the food industry, and trends toward a diet based on convenience foods, such a trial would be challenging to implement and would require regular monitoring to ensure participants complied with the prescribed sodium intake. Experts have designated this such a pressing issue that undertaking such a trial in institutionalized environments such as prisons may be justified.30

There are limitations to this study. The population recruited to UK Biobank were volunteers and, therefore, may not be representative of older or more comorbid populations. As observational data, we describe associations and cannot infer causal relationships. We used the Kawasaki formula to estimate daily sodium excretion. Estimating sodium excretion from a single urine sample may be inaccurate. We chose to use the Kawasaki formula to allow comparability between studies because it is felt to be the least biased of a number of estimation algorithms for estimating 24-hour sodium excretion.31 The relationship between sodium and fatal and nonfatal CVD was not in keeping with perceived wisdom that lower sodium intake leads to lower risk of CVD. Sodium intake was linearly associated with blood pressure, reducing the probability that exposure misclassification biased results to the null. The absence of serum biochemistry is limitation although out with electrolyte disorders, serum sodium is homeostatically maintained in a physiological range in the face of a wide dietary intake. There was no assessment of salt sensitivity in this study, which would provide more detailed insight into which groups are at greater risk from increased dietary salt intake.32 Finally, recent sodium balance studies demonstrate that relationships between dietary sodium intake and urinary excretion are not as simple as previously thought with cyclical handling of sodium over several days and sequestering of sodium in organs such as skin and muscle.33–36 Although this opens new paradigms of thought on sodium homeostasis, for large cohort epidemiological research, using estimates of sodium intake and excretion from spot urine and derivation from sodium:creatinine ratio is widely accepted.10,17


In summary, we confirm the relationship between increasing sodium intake and blood pressure is linear, but there is a fairly limited relationship between salt intake and risk of fatal/nonfatal CVD and ACM in subjects at low cardiovascular risk. Based on these observational data from a large cohort, we demonstrate no strong association of high or low sodium excretion, as a marker of dietary salt intake, with increased risk of cardiovascular events or death. These data may further inform public health debate around recommended dietary daily sodium intake.


This research was conducted using the UK Biobank resource. We thank the participants of the UK Biobank. The work was performed under UK Biobank project number 9310.


This paper was sent to Takayoshi Ohkubo, Guest Editor, for review by expert referees, editorial decision, and final disposition.

The online-only Data Supplement is available with this article at

Correspondence to Patrick B. Mark, BHF Glasgow Cardiovascular Research Centre, 126 University Pl, Glasgow G12 8TA, United Kingdom. Email


  • 1. Benjamin EJ, Blaha MJ, Chiuve SE, et al; American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics-2017 update: a report from the American Heart Association.Circulation. 2017; 135:e146–e603. doi: 10.1161/CIR.0000000000000485LinkGoogle Scholar
  • 2. Olsen MH, Angell SY, Asma S, et al. A call to action and a lifecourse strategy to address the global burden of raised blood pressure on current and future generations: the Lancet Commission on hypertension.Lancet. 2016; 388:2665–2712. doi: 10.1016/S0140-6736(16)31134-5CrossrefMedlineGoogle Scholar
  • 3. Mente A, O’Donnell MJ, Rangarajan S, et al; PURE Investigators. Association of urinary sodium and potassium excretion with blood pressure.N Engl J Med. 2014; 371:601–611. doi: 10.1056/NEJMoa1311989CrossrefMedlineGoogle Scholar
  • 4. Khaw KT, Barrett-Connor E. The association between blood pressure, age, and dietary sodium and potassium: a population study.Circulation. 1988; 77:53–61.LinkGoogle Scholar
  • 5. Jackson SL, Cogswell ME, Zhao L, Terry AL, Wang CY, Wright J, Coleman King SM, Bowman B, Chen TC, Merritt R, Loria CM. Association between urinary sodium and potassium excretion and blood pressure among adults in the United States: National Health and Nutrition Examination Survey, 2014.Circulation. 2018; 137:237–246. doi: 10.1161/CIRCULATIONAHA.117.029193LinkGoogle Scholar
  • 6. Sacks FM, Svetkey LP, Vollmer WM, Appel LJ, Bray GA, Harsha D, Obarzanek E, Conlin PR, Miller ER, Simons-Morton DG, Karanja N, Lin PH; DASH-Sodium Collaborative Research Group. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group.N Engl J Med. 2001; 344:3–10. doi: 10.1056/NEJM200101043440101CrossrefMedlineGoogle Scholar
  • 7. He FJ, Li J, Macgregor GA. Effect of longer term modest salt reduction on blood pressure: cochrane systematic review and meta-analysis of randomised trials.BMJ. 2013; 346:f1325. doi: 10.1136/bmj.f1325CrossrefMedlineGoogle Scholar
  • 8. Aburto NJ, Ziolkovska A, Hooper L, Elliott P, Cappuccio FP, Meerpohl JJ. Effect of lower sodium intake on health: systematic review and meta-analyses.BMJ. 2013; 346:f1326. doi: 10.1136/bmj.f1326CrossrefMedlineGoogle Scholar
  • 9. Eckel RH, Jakicic JM, Ard JD, et al; American College of Cardiology/American Heart Association Task Force on Practice Guidelines. 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(25 suppl 2):S76–S99. doi: 10.1161/01.cir.0000437740.48606.d1LinkGoogle Scholar
  • 10. O’Donnell M, Mente A, Rangarajan S, et al; PURE Investigators. Urinary sodium and potassium excretion, mortality, and cardiovascular events.N Engl J Med. 2014; 371:612–623. doi: 10.1056/NEJMoa1311889CrossrefMedlineGoogle Scholar
  • 11. O’Donnell MJ, Yusuf S, Mente A, Gao P, Mann JF, Teo K, McQueen M, Sleight P, Sharma AM, Dans A, Probstfield J, Schmieder RE. Urinary sodium and potassium excretion and risk of cardiovascular events.JAMA. 2011; 306:2229–2238. doi: 10.1001/jama.2011.1729CrossrefMedlineGoogle Scholar
  • 12. Stolarz-Skrzypek K, Kuznetsova T, Thijs L, Tikhonoff V, Seidlerová J, Richart T, Jin Y, Olszanecka A, Malyutina S, Casiglia E, Filipovský J, Kawecka-Jaszcz K, Nikitin Y, Staessen JA; European Project on Genes in Hypertension (EPOGH) Investigators. Fatal and nonfatal outcomes, incidence of hypertension, and blood pressure changes in relation to urinary sodium excretion.JAMA. 2011; 305:1777–1785. doi: 10.1001/jama.2011.574CrossrefMedlineGoogle Scholar
  • 13. Messerli FH, Hofstetter L, Bangalore S. Salt and heart disease: a second round of “bad science”?Lancet. 2018; 392:456–458. doi: 10.1016/S0140-6736(18)31724-0CrossrefMedlineGoogle Scholar
  • 14. Mente A, O’Donnell M, Rangarajan S, et al. Urinary sodium excretion, blood pressure, cardiovascular disease, and mortality: a community-level prospective epidemiological cohort study.Lancet. 2018; 392:496–506. doi: 10.1016/S0140-6736(18)31376-XCrossrefMedlineGoogle Scholar
  • 15. Sudlow C, Gallacher J, Allen N, et al. UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age.PLoS Med. 2015; 12:e1001779. doi: 10.1371/journal.pmed.1001779CrossrefMedlineGoogle Scholar
  • 16. Townsend P PM, Beattie A. Health and Deprivation. Inequality and the North. Bristol, UK: Croom Helm; 1988.Google Scholar
  • 17. Kawasaki T, Itoh K, Uezono K, Sasaki H. A simple method for estimating 24 h urinary sodium and potassium excretion from second morning voiding urine specimen in adults.Clin Exp Pharmacol Physiol. 1993; 20:7–14.CrossrefMedlineGoogle Scholar
  • 18. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis. 1987; 40:373–383.CrossrefMedlineGoogle Scholar
  • 19. Adrogué HJ, Madias NE. Sodium and potassium in the pathogenesis of hypertension.N Engl J Med. 2007; 356:1966–1978. doi: 10.1056/NEJMra064486CrossrefMedlineGoogle Scholar
  • 20. He J, Klag MJ, Whelton PK, Chen JY, Mo JP, Qian MC, Mo PS, He GQ. Migration, blood pressure pattern, and hypertension: the Yi Migrant Study.Am J Epidemiol. 1991; 134:1085–1101.CrossrefMedlineGoogle Scholar
  • 21. Mancilha-Carvalho JJ, de Oliveira R, Esposito RJ. Blood pressure and electrolyte excretion in the Yanomamo Indians, an isolated population.J Hum Hypertens. 1989; 3:309–314.MedlineGoogle Scholar
  • 22. Page LB, Damon A, Moellering RC. Antecedents of cardiovascular disease in six Solomon Islands societies.Circulation. 1974; 49:1132–1146.LinkGoogle Scholar
  • 23. Alderman MH, Madhavan S, Cohen H, Sealey JE, Laragh JH. Low urinary sodium is associated with greater risk of myocardial infarction among treated hypertensive men.Hypertension. 1995; 25:1144–1152.LinkGoogle Scholar
  • 24. Mills KT, Chen J, Yang W, et al; Chronic Renal Insufficiency Cohort (CRIC) Study Investigators. Sodium excretion and the risk of cardiovascular disease in patients with chronic kidney disease.JAMA. 2016; 315:2200–2210. doi: 10.1001/jama.2016.4447CrossrefMedlineGoogle Scholar
  • 25. Thomas MC, Moran J, Forsblom C, Harjutsalo V, Thorn L, Ahola A, Wadén J, Tolonen N, Saraheimo M, Gordin D, Groop PH; FinnDiane Study Group. The association between dietary sodium intake, ESRD, and all-cause mortality in patients with type 1 diabetes.Diabetes Care. 2011; 34:861–866. doi: 10.2337/dc10-1722CrossrefMedlineGoogle Scholar
  • 26. Lelli D, Antonelli-Incalzi R, Bandinelli S, Ferrucci L, Pedone C. Association between sodium excretion and cardiovascular disease and mortality in the elderly: a cohort study.J Am Med Dir Assoc. 2018; 19:229–234. doi: 10.1016/j.jamda.2017.09.004CrossrefMedlineGoogle Scholar
  • 27. Ekinci EI, Clarke S, Thomas MC, Moran JL, Cheong K, MacIsaac RJ, Jerums G. Dietary salt intake and mortality in patients with type 2 diabetes.Diabetes Care. 2011; 34:703–709. doi: 10.2337/dc10-1723CrossrefMedlineGoogle Scholar
  • 28. Graudal NA, Hubeck-Graudal T, Jurgens G. Effects of low sodium diet versus high sodium diet on blood pressure, renin, aldosterone, catecholamines, cholesterol, and triglyceride.Cochrane Database Syst Rev. 2017; 4:CD004022. doi: 10.1002/14651858.CD004022.pub4MedlineGoogle Scholar
  • 29. Mullens W, Verbrugge FH, Nijst P, Tang WHW. Renal sodium avidity in heart failure: from pathophysiology to treatment strategies.Eur Heart J. 2017; 38:1872–1882. doi: 10.1093/eurheartj/ehx035CrossrefMedlineGoogle Scholar
  • 30. Jones DW, Luft FC, Whelton PK, Alderman MH, Hall JE, Peterson ED, Califf RM, McCarron DA. Can we end the salt wars with a randomized clinical trial in a controlled environment?Hypertension. 2018; 72:10–11.LinkGoogle Scholar
  • 31. Mente A, O’Donnell MJ, Dagenais G, et al. Validation and comparison of three formulae to estimate sodium and potassium excretion from a single morning fasting urine compared to 24-h measures in 11 countries.J Hypertens. 2014; 32:1005–1014; discussion 1015CrossrefMedlineGoogle Scholar
  • 32. Chen J, Gu D, Huang J, Rao DC, Jaquish CE, Hixson JE, Chen CS, Chen J, Lu F, Hu D, Rice T, Kelly TN, Hamm LL, Whelton PK, He J; GenSalt Collaborative Research Group. Metabolic syndrome and salt sensitivity of blood pressure in non-diabetic people in China: a dietary intervention study.Lancet. 2009; 373:829–835. doi: 10.1016/S0140-6736(09)60144-6CrossrefMedlineGoogle Scholar
  • 33. Titze J, Dahlmann A, Lerchl K, Kopp C, Rakova N, Schröder A, Luft FC. Spooky sodium balance.Kidney Int. 2014; 85:759–767. doi: 10.1038/ki.2013.367CrossrefMedlineGoogle Scholar
  • 34. Kopp C, Linz P, Dahlmann A, Hammon M, Jantsch J, Müller DN, Schmieder RE, Cavallaro A, Eckardt KU, Uder M, Luft FC, Titze J. 23Na magnetic resonance imaging-determined tissue sodium in healthy subjects and hypertensive patients.Hypertension. 2013; 61:635–640. doi: 10.1161/HYPERTENSIONAHA.111.00566LinkGoogle Scholar
  • 35. Schneider MP, Raff U, Kopp C, et al. Skin sodium concentration correlates with left ventricular hypertrophy in CKD.J Am Soc Nephrol. 2017; 28:1867–1876. doi: 10.1681/ASN.2016060662CrossrefMedlineGoogle Scholar
  • 36. Rakova N, Kitada K, Lerchl K, et al. Increased salt consumption induces body water conservation and decreases fluid intake.J Clin Invest. 2017; 127:1932–1943. doi: 10.1172/JCI88530CrossrefMedlineGoogle Scholar

Novelty and Significance

What Is New?

  • We demonstrate that in a large cohort of subjects at low risk of cardiovascular disease, urinary sodium excretion as a marker for salt intake is linearly associated with blood pressure but with no relationship between sodium excretion and cardiovascular events or mortality.

What Is Relevant?

  • Current public health advice to reduce sodium intake is likely to lead to lower blood pressure and is unlikely to be harmful in low-risk individuals. Based on these findings, it is less clear that this has benefits on cardiovascular events during follow-up of ≈7 years.


Lower urinary sodium excretion correlated with reduced blood pressure in subjects at low cardiovascular risk, but this was not associated with reduced cardiovascular events.


eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.