The Science of Salt: A Regularly Updated Systematic Review of Salt and Health Outcomes (August to November 2015)

Abstract

The purpose of this review was to systematically identify, summarize, and critically appraise studies on dietary salt relating to health outcomes that were published from August to November 2015. The search strategy was adapted from a previous systematic review on dietary salt and health. Overall, 15 studies were included in the review: one study assessed cardiovascular events, five studies assessed blood pressure or hypertension incidence, six studies assessed surrogate outcomes for cardiovascular or kidney diseases, and three studies assessed other outcomes (age‐related cataracts, rheumatoid arthritis, and bone mineral density, respectively). Four studies were selected for detailed appraisal and commentary.

Excess salt (sodium) consumption is associated with adverse health effects, particularly increased blood pressure (BP).1, 2 Worldwide, complications of hypertension result in 9.4 million deaths per year.3 Based on systematic review evidence for dietary sodium reduction, the World Health Organization (WHO) recommends that adults consume <2 g/d of sodium (equivalent to 5 g/d of salt), with lower amounts for children.2 As part of an initiative to prevent and control major noncommunicable diseases, the WHO has set a global target of reducing dietary sodium intake by 30% by 2025.4

To keep scientific, clinical, and policy stakeholders up‐to‐date with the growing body of literature regarding dietary salt, regularly updated critical appraisals of studies relating to health outcomes are published in the Journal every few months, alternating with reviews of studies relating to salt reduction implementation programs. The objective of the current review is to critically appraise the literature published from August to November 2015 regarding the effects of salt on health outcomes.

Methodology

A detailed description of the methodological approach used to the identify and evaluate the literature in this review has been previously published.5 Articles were identified on a weekly basis through a MEDLINE search strategy, which was adapted from a previous systematic review used to develop the WHO guideline on dietary sodium intake.1, 2 Studies examining the effects of salt on health outcomes included in this review were published from August 4, 2015, to November 17, 2015.

All included studies were assessed for risk of bias by two independent reviewers. Randomized controlled trials (RCTs) were assessed using the Cochrane risk of bias tool.6 Observational, nonrandomized studies were assessed using a modified Cochrane risk of bias tool.7 For meta‐analyses, the AMSTAR tool was applied.8 Detailed appraisals and commentary were performed for a subset of included studies that were judged by two independent reviewers as impactful based on quality of study design, novelty of findings, and potential for generating public discourse or scientific controversy, or for informing public health policy.

Results

The search identified 1571 citations, of which 145 possibly relevant studies met the criteria for full review on the basis of their findings related to salt and health outcomes (Figure). A total of 15 dietary salt studies were included: 12 observational studies,9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 two RCTs,21, 22 and one meta‐analysis.23 The primary outcomes of the included studies were varied: one study assessed cardiovascular events9; five studies assessed BP or hypertension incidence10, 11, 12, 13, 21; six studies assessed surrogate outcomes for cardiovascular14, 22 or kidney diseases15, 16, 17, 23; and three studies assessed other outcomes, including age‐related cataracts,18 rheumatoid arthritis,19 and bone mineral density.20

Figure 1.

Study flow diagram for studies identified from August to November 2015.

Descriptions of the 15 included studies are listed in the Table. Detailed risk of bias assessments for all included studies are included in Appendix S1. Four studies were selected for detailed commentary below: a cross‐sectional study assessing the association between dietary sodium and obesity,14 an RCT comparing the effect of a light salt substitute vs regular salt on BP,21 an RCT comparing the effect of sodium supplementation and potassium supplementation on endothelial function,22 and a meta‐analysis of the effect of dietary sodium on albuminuria.23 Of the four selected studies, both RCTs21, 22 met the methodological criteria applied in the WHO dietary salt systematic review1 (ie, RCTs of ≥4 weeks duration, with no concurrent interventions, that achieved ≥40 mmol/d sodium reduction between intervention and control, with sodium intake measured by 24‐hour urinary sodium excretion; and prospective cohort studies of ≥1 year duration that measured sodium intake in any way), while the meta‐analysis23 and cross‐sectional study14 did not meet the WHO inclusion criteria but were selected based on quality of study design and novelty. Two other prospective cohort studies9, 13 that were not selected for detailed appraisal also met the WHO salt systematic review inclusion criteria.

Table 1.

Description of Included Studies

Study (Country)	Study Design	Participants	Study Duration	Dietary Sodium “Dose” (Actual Mean Intake Per Day)a		Method of Sodium Intake Measurement	Outcomes	Results
Äijälä et al (Finland)9	Prospective cohort study	N=690 adults treated for hypertension and age‐ and sex‐matched controls (OPERA study); 48% male; mean age 51.4 y	~19 y	First quartile: 877.8–1609.0 mg Na/1000 kcal/d Second quartile: 1609.1–1820.6 mg Na/1000 kcal/d Third quartile: 1821.7–2055.8 mg Na/1000 kcal/d Fourth quartile 2058.3–3812.1 mg Na/1000 kcal/d		7‐d food diary at baseline	Cardiovascular events (coronary artery disease and stroke)	Baseline sodium intake was associated with risk of CVD events (adjusted HR 1.664; 95% CI, 1.048–2.640)
Bae et al (Korea)18	Cross‐sectional case‐control	N=12,693 adults with no chronic disease and who underwent ophthalmic examination (KNHANES 2008–2011); 46% male; mean age 44.7 y	Cross‐sectional	Cataract group, UNa/Cr: 16.2 mmol/mmol Noncataract group, UNa/Cr: 11.7 mmol/mmol		UNa/Cr, measured by spot urine	Age‐related cataracts (evaluated with standardized ophthalmic examination)	High sodium intake (UNa/Cr >16.4 mmol/mmol) was associated with higher risk of cataracts (OR, 1.29; 95% CI, 1.16–1.44).
Chen et al (China)10	Cross‐sectional	N=2024 adults from Shandong province (SMASH Project); 52% male; mean age 41.3 y	Cross‐sectional	Men: 5561 mg/d Women: 5113 mg/d		24‐h urine sodium excretion	BP, measured by electronic device	Sodium intake was positively associated with BP (increase in SBP by 1.15 mm Hg and DBP by 0.67 mm Hg per 1 g increment in urinary sodium excretion)
D'Elia et al (Italy)23	Meta‐analysis of RCTs	N=516 adults from 23 cohorts in 11 RCTs	Range: 1–6 wk	Low‐sodium range: 87–6049 mg/d High‐sodium range: 3726–7429 mg/d Mean sodium reduction: 2116 mg/d (92 mmol/d)		24‐h urine sodium excretion	Albuminuria (urine albumin excretion rate or albumin‐creatinine ratio)	Sodium restriction reduced urinary albumin excretion (MD, −32.1%; 95% CI −44.3% to −18.8%); greater effect in patients taking RAAS‐blocking therapy
de Almeida Barros et al (Brazil)21	RCT; single‐blind parallel group	N=56 adults with hypertension (BP ≥140/90 mm Hg during last visit); 34% male; mean age 55.5 y	Intervention and follow‐up: 4 wk	Light salt substitute group: 2923 mg/d Control group (regular salt): 4200 mg/d		24‐h urine sodium excretion	BP (office and home), measured with semiautomatic digital device	Light salt group had lower office BP at the end of treatment compared with the control group (127.11/75.95 mm Hg vs 137.19/82.75 mm Hg). No differences in home BP between groups
Gijsbers et al (Netherlands)22	RCT; double‐blind crossover	N=36 adults with SBP 130–159 mm Hg; 67% male; mean age 65.8 y	13 wk (1‐wk run‐in, followed by three 4‐wk intervention phases)	Sodium supplementation: 4667 mg/d Potassium supplementation: 2220 mg/d Placebo: 2417 mg/d		24‐h urine sodium excretion	Endothelial biomarkers; Inflammatory biomarkers; FMD; 24‐hour ambulatory BP; microvascular vasomotion	Sodium supplementation group had higher BP compared with placebo (+7.5 mm Hg; 95% CI, 4.4–10.5 for SBP and +2.7 mm Hg; 95% CI, 1.1–4.2 for DBP) at the end of the intervention. No effect on endothelial function/FMD/microvascular vasomotion markers except endothelin‐1.
Hallvass et al. (Brazil)15	Cross‐sectional	N=60 adults with nondialysis CKD of any cause (SALTED study baseline); 53% male; mean age 65 y	Cross‐sectional baseline data	Total sample: 4100 mg/d		24‐h urine sodium excretion; 3‐d food record	Fluid status (ECV, TBW), measured by bioimpedance (body composition monitor)	Positive association between sodium excretion and ECW (R 2=0.180), ICW (R 2=0.171), and TBW (R2=0.205). No associations when using sodium intake by food record
Khaledifar et al (Iran)17	Cross‐sectional	N=100 adults with mild hypertension, referred to hypertension research center; 47% male; mean age 46.4 y	Cross‐sectional	Low‐salt intake group: ≤3036 mg/d Intermediate‐salt intake group: 3059–4278 mg/d High salt intake group: ≥4301 mg/d		24‐h urine sodium excretion	Creatinine clearance, estimated by Cockcroft‐Gault formula	Positive association between sodium intake and creatinine clearance (beta 0.07, P=.016).
Kim S‐W et al (Korea)20	Cross‐sectional	N=2779 postmenopausal women (KNHANES 2008–2011); mean age 62.7 y	Cross‐sectional	First quartile, UNa/Cr: 11.5 mmol/mmol Second quartile, UNa/Cr: 21.8 mmol/mmol Third quartile, UNa/Cr: 31.9 mmol/mmol Fourth quartile, UNa/Cr: 53.4 mmol/mmol		UNa/Cr, measured by spot urine	BMD of lumbar spine, total femur, femoral neck, measured by dual‐energy X‐ray absorptiometry. Osteoporosis and osteopenia defined using WHO criteria	UNa/Cr inversely correlated with lumbar spine BMD. Highest UNa/Cr quartile had increased risk of lumbar spine osteoporosis (OR, 1.654; 95% CI, 1.306–2.093) compared with lowest quartile
Kim MH et al (Korea)11	Cross‐sectional	N=8853 men without self‐reported hypertension at time of survey (KNHANES 2007–2013); mean age 42.1 y	Cross‐sectional	Na intake <4000 mg/d		Single 24‐h dietary recall	Hypertension prevalence, defined as SBP ≥140 mm Hg or DBP ≥90 mm Hg	Moderate exercise in men consuming high sodium intake (>4000 mg/d) had lower risk of developing hypertension, compared with nonexercisers (OR, 0.63; 95% CI, 0.47–0.85)
				No exercise: 2865.3 mg/d Moderate exercise: 2833.9 mg/d Vigorous exercise: 2902.9 mg/dNa intake >4000 mg/dNo exercise: 7270.1 mg/d Moderate exercise: 7322.1 mg/d Vigorous exercise: 7290.4 mg/d
Ma et al (UK)14	Cross‐sectional	N=458 children; 52% male; mean age 10.3 y, and N=785 adults; 47.3% male; mean age 49.1 y (UK NDNS)	Cross‐sectional	Lower tertile: 1720 mg/d (adults); 1240 mg/d (children) Middle tertile: 2880 mg/d (adults); 2040 mg/d (children) Upper tertile: 4600 mg/d (adults); 3400 mg/d (children)		24‐h urine sodium excretion, verified by PABA assay; 4‐day food diary	Weight status (obesity, overweight), waist circumference, body fat mass, and lean body mass	Higher salt intake was associated with greater risk of overweight or obesity (adjusted OR, 1.26; 95% CI, 1.16–1.37 per 1 g/d greater salt intake in adults; OR, 1.28; 95% CI, 1.12–1.45) in children
Nerbass et al (UK)16	Cross‐sectional	N=1733 adults with CKD, from primary care practices (RRID Study); 40% male; mean age 72.9 y	Cross‐sectional	Total sample: 2542 mg/d		Early‐morning spot urine from 3 consecutive days. Estimating equation developed in RRID cohort	MAP and BP measured with oscillometric device, albuminuria, carotid to femoral pulse wave velocity, skin autofluorescence	Multivariable linear regression showed a positive association between high sodium intake (>100 mmol/d) and MAP (beta 1.57; 95% CI, 0.41–2.72) and albuminuria (beta 1.35; 95% CI, 1.02–1.79)
Rodrigues et al (Brazil)12	Cross‐sectional	N=272 adults from Vitória; 47% male; mean age 44 y	Cross‐sectional	Total sample: 4060 mg/d		24‐h urine sodium excretion	BP measured with automated oscillometric device	Linear regression showed SBP increased by 0.95 mm Hg and DBP increased by 0.44 mm Hg per 1 g/d salt intake
Salgado et al (Spain)19	Cross‐sectional case‐control	N=18,555 adult alumni of University of Navarre (SUN cohort); 40% male; mean age 38 y	Cross‐sectional	First quartile: 1774 mg/db Second quartile: 2846 mg/db Third quartile: 3713 mg/db Fourth quartile: 4962 mg/db		Food frequency questionnaire (including average pinches of added salt)	Self‐reported rheumatoid arthritis (from baseline questionnaire)	Highest sodium intake quartile associated with self‐reported rheumatoid arthritis (adjusted OR, 1.5; 95% CI, 1.1–2.1) compared with lowest quartile
Takase et al (Japan)13	Prospective cohort	N=4523 normotensive adults with annual health checkups; 64.2% male; mean age 54.1 y	Median follow‐up: 3.1 y	Without hypertension: 4100 mg/d With incident hypertension: 4500 mg/d		Overnight urine (once annually) Estimating equation: Kamata formula	BP measured with mercury sphygmomanometer; new‐onset of hypertension, defined as SBP ≥140 mm Hg or DBP ≥90 mm Hg or use of antihypertensive medications	High salt intake at baseline (≥9 g/d in men; ≥7.5 g/d in women) was associated with greater risk of hypertension (HR, 1.25; 95% CI, 1.04–1.50), compared with lower salt intake. Yearly increase in sodium intake was associated with hypertension incidence

^aUnless otherwise stated, units are in mg per day of sodium intake. To convert to g per day of salt (sodium chloride), divide by 400. To convert to mmol per day of sodium, divide by 23. ^bMedian.

Abbreviations: BMI, body mass index; BMD; bone mineral density; CI, confidence interval; CKD, chronic kidney disease; CrCl, creatinine clearance; CVD, cardiovascular disease; ECW, extracellular water; FMD, flow‐mediated dilation; HR, hazards ratio; ICW, intracellular water; KNHANES, Korea National Health and Nutrition Examination Survey; MAP, mean arterial pressure; Na, sodium; OPERA, Oulu Project Elucidating Risk of Atherosclerosis; OR, odds ratio; PABA, para‐aminobenzoic acid; RAAS, renin‐angiotensin‐aldosterone system; RCT, randomized controlled trial; RRID, Renal Risk in Derby; SBP, systolic blood pressure; SMASH, Shandong and Ministry of Health Action on Salt reduction and Hypertension; SUN, Seguimiento Universidad de Navarra; TBW, total body water; UK NDNS, UK National Diet and Nutrition Survey; UNa/Cr, urine sodium‐to‐creatinine ratio; WHO, World Health Organization.

• Is there an association between salt intake and obesity?

  Ma Y, He F, MacGregor GA. High salt intake: independent risk factor for obesity? Hypertension. 2015;66:843–849.

  Design: Cross‐sectional observational study.

  Setting: UK National Diet and Nutrition Survey rolling program (NDNS RP) 2008/2009 and 2011/2012.

  Participants: 458 children (mean age 10 years) and 785 adults (mean age 49 years).

  Exposure: Sodium intake, measured by a single 24‐hour urine sodium sample and by 4‐day food diary. Dietary misreporting was assessed by: (1) comparing dietary data with Goldberg cutoffs for energy intake/basal metabolic rate, and (2) energy expenditure measured by doubly labeled water in a subset of 67 children and 117 adults.

  Outcomes: Weight status (obesity was defined as a body mass index [BMI] ≥95th percentile in children and BMI ≥30 in adults; overweight was defined as BMI ≥85th percentile in children and BMI ≥25 in adults); waist circumference; body composition (body fat mass and lean mass).

  Risk of bias:

  • Sampling: Low risk.

  • Representativeness: High risk.

  • Reliability/validity of exposure: Low risk.

  • Reliability/validity of outcome: Low risk.

  • Blinding of outcome assessment: Low risk.

  • Risk of selective outcome reporting: Low risk.

  • Confounding: Low risk.

• Sources of funding: UK Food Standards Agency; Department of Health in England; China Scholarship Council.

  Summary of results. This cross‐sectional observational study assessed the relationship between dietary salt intake and the risk of obesity. Adjusting for various demographic variables, including total energy intake, physical activity, and dietary misreporting, the primary analysis demonstrated that increased salt intake was associated with an increased risk of being overweight or obese in adults (odds ratio [OR], 1.26; 95% confidence interval [CI], 1.16–1.37 per 1 g/d greater salt intake (per 400 mg/d sodium intake) and in children (OR, 1.28; 95% CI, 1.12–1.45). Various sensitivity analyses showed similar results (eg, replacing total energy intake with sugar‐sweetened beverages, using a propensity score of higher salt intake to reduce bias from residual confounding, excluding individuals with potentially misreported dietary intakes, using salt intake‐to‐energy intake ratio as the independent variable instead of salt intake, and replacing 24‐hour urine sodium with salt intake calculated from food diaries).

  Comment. Several previous cross‐sectional and prospective cohort studies have demonstrated positive associations between dietary sodium intake and obesity/adiposity measures,24, 25, 26, 27, 28, 29 but adjustment for total caloric intake, a confounding variable, is difficult to quantify accurately from dietary records. Ma and colleagues addressed this issue by assessing dietary misreporting and also by measuring energy expenditure using the gold‐standard method, doubly labeled water, in a subset of study participants. However, since dietary salt intake and total energy intake are significantly correlated (ie, collinear), the precision of the multivariable logistic regression OR estimate is reduced when both variables are included in the model. A strength of this study was that multiple sensitivity analyses to address total energy intake as a confounder demonstrated consistent, robust findings. However, the study does not account for overall diet quality. The major limitation of this study, as the authors noted, is that the cross‐sectional design does not address causality. Reverse causality is possible (ie, adiposity may predispose to higher salt intake independent of total energy intake). There was a 2‐ to 4‐month lag between baseline height and weight measurements and 24‐hour urine measurements, but assuming fairly stable dietary intake, this is unlikely to have significantly impacted the findings.

• Does dietary salt modulate effect of renin‐angiotensin‐aldosterone system inhibition on albuminuria?

  D'Elia L, Rossi G, Schiano di Cola M, et al. Meta‐analysis of the effect of dietary sodium restriction with or without concomitant renin‐angiotensin‐aldosterone system–inhibiting treatment on albuminuria. Clin J Am Soc Nephrol. 2015;10:1542–1552.

  Design: Meta‐analysis of RCTs.

  Methods:

  • Data sources: MEDLINE (from 1966 to October 2014), EMBASE (from 1980 to October 2014), Cochrane Library (to October 2014).

  • Study selection and assessment: RCTs comparing urine albumin excretion rate (UAE) or albumin to creatinine ratio (ACR) between two different sodium intake regimens in one or more adult cohorts. Eleven RCTs (six crossover, five‐parallel group) met inclusion criteria (n=516 adults from 4 countries, mean age range 47–73 years). Twenty cohorts from these studies were analyzed, 11 of which were taking renin‐angiotensin‐aldosterone system (RAAS)–blocking therapy. Five studies were double‐blind and compared slow sodium vs placebo tablets, while six (55%) studies assessed restricted sodium vs regular sodium diets. Studies were pooled using random effects models. Study quality was assessed with the Cochrane risk of bias tool. Heterogeneity was assessed by subgroup analysis and meta‐regression.

  • Method of sodium intake measurement: All studies used 24‐hour urinary sodium excretion.

  • Outcomes: Albuminuria measured by UAE or ACR.

  • Subgroup analyses: RAAS‐blocking therapy, length of intervention, country, presence of kidney damage, hypertension, diabetes, study blinding, risk of bias, funding source, and dietary protocol.

  Risk of bias:

  • A priori design: Yes.

  • Duplicate study selection/data extraction: Yes.

  • Comprehensive literature search: Yes.

  • Status of publication used as an inclusion criterion: No.

  • List of studies included: Yes.

  • Characteristics of included studies provided: Yes.

  • Quality of studies assessed and documented: Yes.

  • Quality of included studies used appropriately in formulating conclusions: Yes.

  • Methods to combine finding appropriate: Yes.

  • Publication bias assessed: Yes.

  • Conflict of interest: No financial support was received.

• Summary of results. In this meta‐analysis of RCTs, D'Elia and colleagues synthesized evidence regarding the effect of sodium intake on albuminuria. Because the albuminuria outcome was reported heterogeneously among included studies (UAE or ACR), results were expressed as between‐regimen percentage change in albuminuria. The meta‐analysis demonstrated a significant reduction in urinary albumin excretion (mean difference [MD], −32.1%; 95% CI, −44.3% to −18.8%). There were significant interactions between sodium restriction and RAAS‐blocking medications, length of intervention, and preexisting kidney damage (defined as presence of proteinuria, microalbuminuria, or lower estimated glomerular filtration rate). The effect size was greater among patients taking RAAS‐blocking therapy (MD, −41.9%; 95% CI, −56.4% to −27.4%) compared with those not taking RAAS‐blocking therapy (MD, −17.2%; 95% CI, −26.1% to −2.1%).

  Comment. The meta‐analysis by D'Elia and colleagues has a number of strengths. The methodology used was rigorous, with a comprehensive search and duplicate study selection and data extraction. A broad range of participants were included (patients with diabetes, hypertension, renal disease, normoalbuminuria and microalbuminuria, and established proteinuria, as well as healthy participants) and therefore the results are fairly generalizable. Another strength was that all studies used 24‐hour urine sodium measurements to assess sodium intake. Although albuminuria data were expressed as relative change between low‐ and high‐sodium regimens, rather than absolute change, the overall effect size is clinically significant. Meta‐regression and several subgroup analyses were performed to investigate heterogeneity. The favorable effect of sodium intake reduction on albuminuria was greater in patients treated with RAAS blockers and in those with preexisting kidney damage. The positive association between decreases in BP and decreases in albuminuria in the meta‐regression suggests that the effect of lowering salt intake on albuminuria may be mediated in part by BP.

  There are a few limitations of this meta‐analysis. As noted by the authors, most of the studies had a small sample size, ranging from 17 to 140 participants. There were wide ranges in mean sodium intake in both the low‐sodium (range 1035–6049 mg/d sodium, equivalent to 2.6–15.1 g/d salt) and high‐sodium (range 3726–7429 mg/d sodium, equivalent to 9.3–18.6 g/d salt) groups between studies, although all studies were able to achieve ≥40 mmol/d sodium reduction between intervention and control (mean reduction 92 mmol/d, equivalent to 2116 mg/d sodium and 5.3 g/d salt). As a result of between‐study overlap between high‐ and low‐sodium intakes, a true “dose‐response” effect could not be quantified. Study durations were relatively short (1–6 weeks), and thus this study does not address long‐term effects of sodium restriction on albuminuria. Four studies had a duration of only 1 week, with subgroup analysis demonstrating no significant effect of sodium reduction on albuminuria in these short studies. The four studies with a high risk of bias demonstrated a smaller pooled effect size compared with the studies with a low risk of bias. Because individual patient data from crossover RCTs were not available/analyzed, the effect sizes from these RCTs may be underestimated. The meta‐analysis did not report whether there were washout periods for crossover RCTs.

• Does light salt substitution improve blood pressure in hypertensive patients?

  de Almeida Barros CL, Lima Sousa AL, Mendoça Chinem B, et al. Impact of light salt substitution for regular salt on blood pressure of hypertensive patients. Arq Bras Cardiol. 2015;104:128–135.

  Design: RCT (parallel group).

  Setting: Hypertension clinic at the General Hospital of the Federal University of Goiás, Goiâna, Brazil.

  Study duration: 4 weeks.

  Participants: 35 participants aged 20 to 65 years (mean age 55.5 years), with uncontrolled hypertension (BP ≥140/90 mm Hg), taking stable doses of antihypertensive drugs for ≥30 days.

  Intervention: Light salt substitute (composition, per gram: 130 mg sodium, 346 mg potassium, 44 µg iodine) vs regular salt (composition, per gram: 390 mg sodium, 25 µg iodine). Salt was provided to participants, with the amount based on the recommended 3 g/d salt per person and the number of people usually sharing meals, plus 10% of salt per patient added to account for routine changes on weekends. All participants were instructed to consume only the provided salt during the study and to reduce sodium‐rich food intake.

  Achieved sodium intake: 2924 mg/d sodium (7.3 g/d salt) in the light salt group and 4200 mg/d sodium (10.5 g/d salt) in the regular salt group. Sodium intake was measured by 24‐hour urine sodium excretion.

  Outcomes: BP measured at the office (casual) and at home using semiautomatic digital device.

  Risk of bias:

  • Random sequence generation: High risk.

  • Allocation concealment: High risk.

  • Blinding of participants and personnel: High risk.

  • Blinding of outcome assessors: High risk.

  • Incomplete outcome reporting: Low risk.

  • Selective reporting: Low risk.

  • Other sources of bias: High risk.

• Source of funding: Partial funding by Coordenação de Aperfeiçoamento de Pessoal de Nivel Superior.

  Summary of results. The light salt group had lower office BP compared with the regular salt group at the end of the 4‐week intervention (difference in systolic BP −12.5 mm Hg, P=.034; difference in diastolic BP −7.6 mm Hg, P=.046). There was no difference between groups in home BP measurements.

  Comment. In this RCT, group allocation was by alternation and therefore was subject to a high risk of bias. Although patients were technically blinded, the light salt had a distinct taste and “low acceptance” by 89.5% of intervention group individuals. Thus, blinding of patients was not adequately maintained during the study, and study adherence was compromised. The study did not control overall sodium intake; provision of salt to participants addressed discretionary salt use, but not other sources of sodium intake. The 24‐hour urine sodium results demonstrate that although a reduction in sodium intake was achieved in the intervention group, actual sodium consumption in both groups was much higher than expected, and their target sodium intake of 3 g/d salt is likely much lower than the participants’ usual sodium intakes. The light salt substitute contained added potassium, and therefore the BP‐lowering effect of reduced sodium cannot be isolated from the effect of potassium supplementation. However, there was no significant change in potassium excretion in the intervention group during the study and no difference in potassium excretion between groups at the end of the study. As a comparison, the effect size reported in a meta‐analysis of 34 RCTs of at least 4 weeks’ duration assessing modest salt reduction was −4.18 mm Hg (95% CI, −5.18 to −3.18) for systolic BP and −2.06 mm Hg (95% CI, −2.67 to −1.45) for diastolic BP.30

• What is the effect of sodium intake on endothelial function?

  Gijsbers L, Dower JI, Schalkwijk CG, et al. Effects of sodium and potassium supplementation on endothelial function: a fully controlled dietary intervention study. Br J Nutr. 2015;114:1419–1426.

  Design: RCT (crossover).

  Setting: Research center of Wageningen University, Wageningen, The Netherlands.

  Study duration: Three intervention periods of 4 weeks each, with no washout.

  Participants: 36 adults aged 40 to 80 years (mean age 65.8 years) who were nonsmokers and had fasting supine systolic BP 130 to 159 mm Hg. Exclusion criteria: use of cardiovascular medications, BMI >40 kg/m², diabetes, cardiovascular, gastrointestinal, liver or renal diseases, heavy alcohol use, pregnancy, or lactating.

  Intervention: Sodium chloride capsules (containing 3000 mg/d sodium) vs potassium chloride capsules (containing 2800 mg/d potassium) vs placebo capsules (low‐sodium, low‐potassium group). All patients were provided with a diet containing 2400 mg/d sodium and 2300 mg/d potassium for a 2500 kcal intake.

  Achieved sodium intake: Sodium supplementation group: 4667 mg/d (11.7 g/d salt); potassium supplementation group: 2220 mg/d (5.6 g/d salt); placebo group: 2417 mg/d (6.0 g/d salt). Sodium intake was measured by a single 24‐hour urine sodium excretion.

  Outcomes: Endothelial biomarkers (eg, E‐selectin, soluble thrombomodulin, soluble vascular cell adhesion molecule 1), inflammatory biomarkers (eg, C‐reactive protein, interleukin 6), 24‐hour ambulatory BP, flow‐mediated dilation (FMD), and microvascular vasomotion.

  Risk of bias:

  • Random sequence generation: Low risk.

  • Allocation concealment: Unclear.

  • Blinding of participants and personnel: Low risk.

  • Blinding of outcome assessors: Unclear.

  • Incomplete outcome reporting: High risk.

  • Selective reporting: Low risk.

  • Other sources of bias: High risk.

• Source of funding: TI Food and Nutrition, a public‐private partnership on precompetitive research in food and nutrition.

  Summary of results. After 4 weeks of intervention, this crossover RCT demonstrated greater 24‐hour ambulatory BP in the sodium supplementation group compared with placebo (MD +7.5 mm Hg; 95% CI, 4.4–10.5 for systolic BP; and MD +2.7 mm Hg; 95% CI, 1.1–4.2 for diastolic BP). There was a reduction in BP in the potassium supplementation group compared with the placebo group (MD for systolic BP, −3.9 mm Hg; 95% CI, −6.9 to −0.9; MD for diastolic BP, −1.6 mm Hg; 95% CI, −3.2 to −0.1). Twenty‐six endothelial function and inflammatory parameters were tested. The study suggests that there may be a reduction in endothelin 1 with sodium restriction and beneficial effects of potassium supplementation on FMD and interleukin 8.

  Comment. A strength of the study was the double‐blind, placebo‐controlled study design with provision of a controlled, low‐sodium diet, thereby facilitating achievement of sodium intake targets in each phase. Individuals using cardiovascular medications and those with chronic diseases were excluded from the study, thus limiting its generalizability. In addition, no single primary outcome was identified. Most endothelial function parameters demonstrated no difference between groups, but those parameters that were identified as statistically significant are subject to type 1 error because no adjustments were made for multiple testing. The authors note that their study was adequately powered to show a clinically relevant effect on FMD; however, only 22 of 36 included patients were analyzed for the FMD outcome, as a result of exclusion of patients with recordings of inadequate quality.

Discussion

This review identified 15 studies relating dietary sodium to health outcomes. Fourteen of these studies demonstrated positive associations between dietary salt and adverse health outcomes or surrogate markers, and one study was neutral (showing an association between salt intake and creatinine clearance).

Two of the four studies selected for detailed appraisal were RCTs that found beneficial effects of reducing sodium intake on BP,21, 22 lending more support to this well‐established evidence base. In the RCT by Gijsbers and colleagues, several endothelial and inflammatory parameters were evaluated to explore mechanisms by which dietary sodium may affect BP, but there was no significant effect on most endothelial function parameters.22 Both RCTs had a high risk of bias in one or more key domains.

The meta‐analysis by D'Elia and colleagues synthesized RCT evidence regarding the effect of dietary sodium restriction on albuminuria.23 Although the included studies were relatively short‐term, a clinically significant reduction in albuminuria with sodium intake restriction was demonstrated, with greater reductions in patients using RAAS‐blocking therapy and in patients with kidney damage. Longer‐term RCTs are needed to assess the dose‐response relationship between sodium intake and albuminuria, as well as the effect of dietary sodium reduction on progression of renal disease and other “hard” outcomes in the chronic kidney disease population.

Ma and colleagues assessed the association between dietary sodium and weight status (overweight and obesity) in their cross‐sectional analysis of adults and children in the United Kingdom.14 This study, which applied methods to ascertain accurate measures of total energy intake, found a positive association between sodium intake and obesity. However, most evidence on this topic is limited to observational studies; clinical trial data are currently lacking. Because salt intake may influence weight through changes in extracellular volume or through direct effects on body fat metabolism, preferred outcomes in future prospective RCTs would include body composition measured by dual x‐ray absorptiometry or bioimpedance. In addition to total energy intake, diet quality is also an important variable to assess in studies evaluating the association between dietary salt and obesity.

Conclusions

The 15 included studies from August to November 2015 highlight the diversity of recently published dietary salt research from 10 countries. These studies assessed a wide range of health outcomes, including cardiovascular and renal outcomes such as BP, cardiovascular events, endothelial biomarkers, albuminuria, and renal function, as well as other outcomes such as cataracts, bone mineral density, and rheumatoid arthritis. While the studies were consistent with adverse health or surrogate effects from high dietary salt, several of the studies were hypothesis‐generating and had limitations (eg, cross‐sectional design) precluding assessment of cause and effect. Evidence from RCTs and a meta‐analysis in this review support a beneficial effect of sodium reduction on BP and albuminuria.

Disclosure

NC is a member of World Action on Salt and Health (a dietary salt reduction organization) but has no financial interests to declare. JA, AAL, KT, and JAS have no conflicts of interest to declare. MMYW is a research consultant with Arbor Research Collaborative for Health.

Supporting information

Appendix S1. Risk of bias tables for included studies

J Clin Hypertens (Greenwich). 2016;18:1054–1062. DOI: 10.1111/jch.12874. © 2016 Wiley Periodicals, Inc.