1. Objectives

To estimate the effects of low sodium versus high sodium intake on systolic and diastolic blood pressure (SBP and DBP), plasma or serum levels of renin, aldosterone, catecholamines, cholesterol, high-density lipoprotein (HDL), low density lipoprotein (LDL) and triglycerides.

2. How studies were identified

The following databases were searched in March 2016:

• Cochrane Hypertension Specialised Register
• CENTRAL (The Cochrane Library 2016, Issue 3)
• MEDLINE Ovid
• MEDLINE Ovid In-Process and Other Non-Indexed Citations
• EMBASE
• ClinicalTrials.gov

Reference lists of relevant articles were also searched

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials, including both parallel and crossover trials

3.2 Study participants

Normotensive and hypertensive individuals of all ages and ethnicities were included

(Studies in which participants had diseases other than elevated blood pressure were excluded)

(Hypertension is normally defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg)

3.3 Interventions

Interventions comparing the effect of high versus low dietary sodium intakes on blood pressure, renin, aldosterone, catecholamines, cholesterol, and triglyceride, with sodium intakes estimated using 24-hour urinary sodium excretion. Studies including other concomitant antihypertensive interventions were included only if the intervention was identical in both high and low sodium diet groups

3.4 Primary outcomes

Effects on:

• SBP
• DBP
• Renin
• Aldosterone
• Adrenaline
• Noradrenaline
• Triglyceride
• Cholesterol
• LDL
• HDL

4. Main results

4.1 Included studies

One hundred and eighty-five studies were included in this review:

• In studies of untreated hypertensive individuals, the mean blood pressure at baseline was 151/93 mmHg, and in studies of treated hypertensive individuals, the mean baseline blood pressure was 144/88 mmHg
• In studies of people with normal blood pressure, the mean baseline blood pressure was 119/71 mmHg
• The intake of sodium during the intervention in the high sodium group ranged from 100 mmol/day to 795 mmol/day, with a mean of 201 mmol/day. The mean intake of sodium in the low sodium group was 66 mmol/day, with the majority of studies (82%) reporting intakes below 100 mmol/day

4.2 Study settings

How the data were analysed
The change in outcome measure from baseline to the end of treatment during a low versus a high sodium diet was defined as the treatment effect. The effect of low versus high sodium diets on blood pressure, cholesterol, triglyceride, aldosterone, and catecholamines was compared using the mean difference (MD), and the standardized mean difference (SMD) was used for renin, as the units of measurement differed between studies. For studies reporting blood pressure as mean arterial pressure (MAP) only, SBP and DBP were estimated using the following formulae: SBP=1.3*MAP + 1.4, and DBP=0.83*MAP - 0.7. Separate meta-analyses were performed for each outcome measure. In blood pressure analyses, participants were stratified by race (Caucasians, Blacks, Asians) and level of blood pressure (normotensive or hypertensive).

Subgroup analyses were performed excluding studies lasting less than seven days and studies with reported sodium intakes above 250 mmol/day.

Results
Blood pressure
Caucasians
Among normotensive Caucasians, the MD for a reduction in sodium intake was a statistically significant decrease in SBP of -1.09 mmHg (95% confidence interval (CI) [-1.63 to -0.56], p=0.0001; 89 studies/8569 individuals); and a non-significant decrease in DBP of -0.03 mmHg (95% CI [-0.37 to 0.43], p=0.89; 90 studies/8833 individuals). Among hypertensive Caucasians, SBP and DBP were significantly lowered (MD -5.51 mmHg in SBP, 95% CI [-6.45 to -4.57], 84 comparisons/5925 participants; and -2.88 mmHg in DBP (95% CI [-3.44 to -2.32], 85 comparisons/6001 individuals). For both normotensive and hypertensive Caucasians, these results did not differ meaningfully when restricted to trials of at least seven days’ duration with sodium intakes ≤250 mmol/day.

Blacks
Among normotensive Blacks, the MD for a reduction in sodium intake was a significant decrease in SBP of -4.02 mmHg (95% CI [-7.37 to -0.68], p=0.02); and in DBP of -2.01 mmHg (95% CI [-4.37 to 0.35], p=0.09; 7 studies/506 individuals).

Among hypertensive Blacks, the MD for a decrease in sodium intake was a statistically significant reduction in SBP of -6.64 mmHg (95% CI [-9.00 to -4.27], p=0.00001); and in DBP of -2.91 mmHg (95% CI [-4.52 to -1.30], p=0.0001; 8 studies/619 individuals).

Asians
Among Asians with normal BP, the MD for a reduction in sodium intake was a non-significant decrease in SBP of -0.72 mmHg (95% CI: [-3.86 to 2.41], p=0.65); and a non-significant decrease in DBP of -1.63 mmHg (95% CI [-3.35 to 0.08], p=0.062; 3 studies/393 individuals).

In Asians with elevated BP, a lowered sodium intake led to a significant decrease in SBP (MD -7.75 mmHg (95% CI [-11.44 to -4.07], p=0.000037); and in DBP (MD -2.68 mmHg (95% CI [-4.21 to -1.15], p=0.00058; 8 studies/501 individuals).

Renin
The SMD for a reduction in sodium intake was a significant increase in renin of 1.22 ng/mL/hour (95% CI [1.07 to 1.37], p<0.00001; 88 trials/5498 individuals). The result did not differ meaningfully in subgroup analysis restricted to trials lasting seven days or more with sodium intakes ≤250 mmol/day.

Aldosterone
In 65 studies including measures of aldosterone, the MD was a significant 97.81 pg/mL (95% CI [82.56 to 113.05], p<0.00001; 4884 individuals). The MD in aldosterone was similar in pooled analysis of 34 trials lasting at least seven days with sodium intakes ≤250 mmol/day (MD 95.59 pg/mL, 95% CI [74.12 to 117.05], 3128 individuals).

Catecholamines
Noradrenaline
The MD for a reduction in sodium intake was a statistically significant increase in noradrenalin of 63.56 pg/mL (95% CI [42.66 to 84.46], p<0.00001; 36 studies/1736 individuals). In 23 comparisons of at least seven days’ duration with sodium intakes ≤250 mmol/day, the difference was reduced, but remained statistically significant (MD 48.66 pg/mL, 95% CI [28.88 to 68.44], p<0.00001).

Adrenaline
In 16 studies including measurements of adrenaline, the MD was a statistically significant increase of 7.55 pg/mL (95% CI [0.85 to 14.26], p=0.027; 662 individuals), and this finding remained significant in pooled analysis of 12 studies of at least seven days’ duration with sodium intakes ≤250 mmol/day.

Cholesterol
A reduction in sodium intake led to a significant rise in total cholesterol (MD 5.64 mg/dL, 95% CI [2.46 to 8.82], p=0.0005; 27 studies/1800 participants), and this finding remained statistically significant in analysis of studies of at least seven days’ duration with sodium intakes ≤250 mmol/day.

LDL and HDL cholesterol
There was no significant effect of a reduction in sodium intake on LDL (MD 3.12 mg/dL, 95% CI [-0.41 to 6.64]) or HDL (MD -0.29 mg/dL, 95% CI [-1.66 to 1.08]; both p≥0.08).

Triglyceride
In 19 studies including measures of triglyceride, there was a significant increase in triglyceride (MD 7.04 mg/dL, 95% CI [3.04 to 11.05], p=0.00057). In comparisons of at least seven days’ duration with sodium intakes ≤250 mmol/day, the increase was also significant (MD 6.92 mg/dL, 95% CI [1.82 to 12.02], p=0.0078; 12 studies).

Additional analyses
Among studies in Caucasian populations, no meaningful differences in SBP were observed in analyses comparing trials at high versus low risk of bias for blinding and blinding of outcome detection. In sensitivity analyses excluding studies contributing to funnel plot asymmetry, effect sizes were reduced and were non-significant.

5. Additional author observations*

The lowering of mean dietary sodium intakes from 201 to 66 mmol/day in the trials included in this review is comparable to the reduction recommended by most health institutions. Reducing sodium intakes decreased SBP among Black and Caucasian normotensive and hypertensive populations and Asian hypertensives by a modest but statistically significant amount. The reduction in SBP among Asian and Black hypertensives was greater than that among Caucasian hypertensives. In Blacks, the effect of sodium reduction on blood pressure was similar among normotensives and hypertensives, while among Asian and Caucasian populations, the effect size was smaller in normotensives compared with hypertensives. These observed differences in effect size between ethnic groups may be due to study-specific confounding factors, such as differences in baseline blood pressure, age, and degree of sodium reduction achieved. Future studies in mixed populations should report data separately by ethnicity to enable further investigation of these findings.

An increase in plasma renin, aldosterone, adrenaline and noradrenaline, total cholesterol and triglyceride was observed with sodium reduction. Due to these potential adverse effects, sodium reduction may not have a net beneficial effect for normotensive Caucasians. The compensatory activation of the renin-angiotensin-aldosterone system, as evidenced by the increase in plasma renin and aldosterone, may explain why a reduction in sodium intake failed to have a large effect on blood pressure.

Due to large variations in study populations with regard to baseline blood pressure, age, sodium intake, and level of sodium reduction, blood pressure and hormone meta-analyses were heterogeneous; however, little heterogeneity was observed in the lipid meta-analyses. Few trials adequately described allocation concealment and many trials were not double-blinded; however, no difference was found between trials at low versus high risk of detection bias. Although long-term randomized controlled trials with mortality and morbidity outcomes would address whether or not the beneficial effect of sodium reduction on blood pressure outweighs the associated increases in renin, aldosterone, adrenaline, noradrenaline, total cholesterol, and triglyceride, such studies may not be practicable.