Meta-analysis or Systematic Review

M - Click here for explanation of classification scheme.

POSITIVE: See Quality Criteria Checklist below.

To determine the effects of moderate, longer-term salt reduction on blood pressure, plasma lipids and hormones. This effect was further examined by blood pressure status, sex and ethnic group.

• Trials included randomization to either a control (usual salt intake) or intervention (moderately reduced salt intake) group
• Reduction in 24-hour urinary sodium within the range of 40mmol to 120mmol (equal to 2.3g to 7.0g per day salt)
• Duration of salt reduction four or more weeks
• Age more than 18 years
• Individuals could have hypertension.

• Concomitant interventions including non-pharmacological interventions, anti-hypertensive or other medications in both groups
• Children
• Pregnant women
• Individuals with diabetes or heart failure.

Recruitment

• Search strategy included the following electronic databases:
  • Medline (1950 to November 2012)
  • Embase (1980 to November 2012)
  • Hypertension Group Specialised Register
  • Cochrane Central Register of Controlled Trials.
• Reference lists were reviewed for original and review articles
• No language restrictions.

Design

• Data was extracted using a standard form by two of the authors. Differences were resolved by discussion with the third author.
• Study quality was assessed with the following criteria:
  • Group allocation was judged adequate or inadequate:
    • Adequate: Central randomization or other method where participants and investigators could not foresee assignment
    • Inadequate: Open list of random numbers or other method where participants and investigators could determine assignment.
  • Blinding:
    • Double blind
    • Observer blind to blood pressure
    • Open study.
  • Completeness of follow-up:
    • Intention to treat vs. non-intention to treat
    • Number of participants lost to follow-up after randomization.

Statistical Analysis

• Treatment effect for outcomes was determined:
  • Crossover trials: Difference between the end of the intervention and control periods
  • Parallel trials: Difference between the two treatment groups from baseline to end of follow-up.
• Variance of the treatment effect for outcomes was calculated for each trial using:
  • Standard deviations or standard errors
  • Confidence intervals and exact T-test or P values
  • Imputed by inverting a boundary P value (e.g., P<0.05 became P=0.05) or assuming a correlation coefficient of 0.5 between the initial and final measurement.
• Mean effect sizes were estimated using pooled data and the inverse variance method in random effects meta analysis
• Heterogeneity  was assessed with the I² test (I² more than 50% was considered significant); the source of heterogeneity was examined with meta regression analysis
• Relationship between the change in 24-hour urinary sodium and the change in blood pressure was evaluated with meta regression analysis
• Funnel plot asymmetry assessed publication bias and Egger's regression test estimated funnel plot asymmetry
• Sub-group analyses:
  • Groups were further divided by blood pressure status, sex and ethnicity
  • If more than 85% of the participants in a trial were white, then the ethnic group was considered white
  • If information on ethnic group was not provided, the study was excluded from this sub-group analysis
  • Not performed for lipid and hormone data due to the small number of trials reporting these outcomes.

Timing of Measurements

• Extracted data included:
  • Characteristics of the study
  • Design:
    • Crossover
    • Parallel.
  • Type of blinding:
    • Open
    • Single blind
    • Double blind.
  • Randomization method
  • Method of blinding:
    • Placebo
    • Random zero or automated sphygmomanometers
    • Observer blind to blood pressure
    • Open.
  • Study duration:
  • Results before and after intervention
  • Changes in systolic and diastolic blood pressure
  • Twenty-four hour urinary sodium excretion
  • Lipids:
    • Total cholesterol
    • LDL cholesterol
    • HDL cholesterol
    • Triglycerides.
  • Hormones:
    • Plasma renin activity
    • Aldosterone
    • Noradrenaline
    • Adrenaline.
  • Statistics that could be used to estimate the variance of the outcome measures.

Dependent Variables

• Primary: Blood pressure:
  • Trials with normotensive individuals
  • Trials with hypertensive individuals
  • All trials.
• Secondary:
  • Lipids
  • Hormones
  • Study quality
  • Publication bias.

Independent Variables

Moderate salt reduction vs. usual salt intake.
 

Initial N

• All: N=34 trials (30 trials, in which four trials included both normotensive and hypertensive individuals, so each counted as four separate trials), for a total of 3,230 participants:
  • Hypertensive: N=22 studies (990 participants)
  • Normotensive: N=12 studies (2,240 participants).
• Sex:
  • Men:
    • Hypertensive: N=9 or 10 studies, depending on outcome (227 to 239 participants)
    • Normotensive: N=6 studies (1,391 participants).
  • Women
    • Hypertensive: N=9 or 10 studies depending on outcome (181 to 193 participants)
    • Normotensive: N=6 studies (691 participants).
• Hormone trials:
  • Plasma renin: N=14 studies (455 participants)
  • Aldosterone: N=9 studies (340 participants)
  • Noradrenaline: N=6 studies (129 participants)
  • Adrenaline: N=4 studies (84 participants).
• Lipid trials:
  • Total cholesterol: N=8 studies (365 participants)
  • LDL cholesterol: N=5 studies (262 participants)
  • HDL cholesterol: N=6 studies (278 participants)
  • Triglycerides: N=6 studies (309 participants).

Attrition (Final N)

N=34.

Age

Median 50 years (range 22 years to 73 years):

• Hypertensive individuals: Median age 50 years (range 24 years to 73 years)
• Normotensive individuals: Median age 50 years (22 years to 67 years).

Ethnicity

• White:
  • Hypertensive: N=16 or 17 studies depending on the outcome (599 participants)
  • Normotensive: N=12 studies (1,901 participants).
• Black:
  • Hypertensive: N=5 studies (171 participants)
  • Normotensive: N=3 studies (412 participants).
• Asian:
  • Hypertensive: N=1 studies (29 participants)
  • Normotensive: N=0 studies.

Other Relevant Demographics

• Median blood pressure: 148/93mm Hg
  • Hypertensive: 148/93mm Hg
  • Normotensive: 127/77mm Hg.
• Median 24-hour urinary sodium: 160mmol (range 125mmol to 200mmol) equivalent to salt intake of 9.4g per day (range 7.3g to 11.7g per day)
  • Hypertensive: 162mmol (range 125mmol to 191mmol) equivalent to salt intake of 9.5g per day (range 7.3g to 11.7g per day)
  • Normotensive: 153mmol (range 128mmol to 200mmol) equivalent to salt intake of 8.9g per day (range 7.5g to 11.7g per day).
• Median study duration: Four weeks (range four weeks to three years)
  • Hypertensive: Five weeks (range four weeks to one year)
  • Normotensive: Four weeks (range four weeks to three years).

Location

• Queen Mary University, London, UK
• Affiliated Hospital of LuZhou Medical College, Luzhou, Sichuan, China.

 

 

Key Findings
 

Variables	All Trials  Mean (range)	Hypertensive Trials Mean (range)	Normotensive Trials Mean (range)
Change in 24-hour urinary sodium, mmol	-75 (-40 to -118)	-4.4 (-3.1 to -6.8)	-4.4 (2.3 to -6.9)
Reduction in salt intake, g per day	- 4.4 (-2.3 to -6.9)	-4.4 (-3.1 to -6.8)	-4.4 (2.3 to -6.9)
Change in systolic blood pressure, mm Hg	-4.18 95% CI: -5.18 to -3.18, P<0.001 I2=75%	-5.39  95% CI: -6.62 to -4.15 P<0.001 I2=61%	-2.42 95% CI: -3.56 to -1.29 P<0.001 I2=66%
Change in diastolic blood pressure, mm Hg	-2.06  95% CI: -2.67 to -1.45 P<0.001 I2=68%	-2.82 95% CI: -3.54 to -2.11 P<0.001 I2=52%	-1.00 95% CI: -1.85 to -0.15 P=0.02 I2=66%

 

Source of Heterogeneity

• All trials:
  • Change in 24-hour urinary sodium, age, blood pressure status and ethnic group (regression analysis) were significantly associated with the reduction in systolic blood pressure accounting for 68% of the variance between studies
  • After adjusting for age, blood pressure status and ethnic group, a 100mmol reduction in 24-hour urinary sodium (equivalent to 6g per day of salt) resulted in a reduction of 5.8mm Hg (95% CI: 2.5 to 9.2; P=0.001) in systolic blood pressure
  • Age, ethnic group, blood pressure status and 24-hour urinary sodium accounted for 41% of the variation in diastolic blood pressure between studies; however, only ethnic group was significantly associated with changes in diastolic blood pressure (P=0.021).
• Trials with hypertensive individuals:
  • Change in 24-hour urinary sodium and ethnic group were significantly associated with reduced systolic blood pressure. Including age with these two variables accounted for 46% of the variance between studies.
  • After adjustment for age and ethnic group, a 100mmol reduction (6g per day of salt) in 24-hour urinary sodium resulted in a 10.8mm Hg (3.5 to 18.2, P<0.01) decrease in systolic blood pressure
  • None of the variables (24-hour urinary sodium, ethnic group, age) were significantly related to the decrease in diastolic blood pressure. These variables explained 11% of the variation between studies.
• Trials with normotensive individuals:
  • Three variables (age, ethnic group, change in 24-hour urinary sodium) accounted for 51% of the variation seen in systolic blood pressure between studies; however, only change in 24-hour urinary sodium was significant associated with a decrease in the systolic blood pressure
  • Systolic blood pressure was reduced 4.3mg Hg (95% CI: 0.1 to 8.5, P<0.05) with a 100mmol reduction in 24-hour urinary sodium (6g per day of salt)
  • Age, ethnic group and change in 24-hour urinary sodium accounted for 43% of the variation in diastolic blood pressure between studies; only ethnic group was significantly associated with a decrease in diastolic blood pressure (P=0.04).

	Hypertensive Men	Normotensive Men	Hypertensive Women	Normotensive Women
Change in 24-hour urinary sodium, mmol	-86.07 95% CI: -100.17 to -71.97 P<0.001	-67.26 95%CI: -81.90 to  -52.62 P<0.001	-69.56 95% CI:         -77.56 to -61.55 P<0.001	-62.98 95% CI:         -88.59 to -37.37 P<0.001
Reduction in salt intake, g per day	-5.04 95% CI: -5.86 to -4.21	-3.93 95% CI: -4.79 to -3.08	-4.07 95% CI: -4.54 to -3.60	3.68 95% CI: -5.18 to -2.19
Change in systolic blood pressure, mm Hg	-6.4 95% CI: -8.00 to -4.80 P<0.001	-3.39 95% CI: 5.63 to -1.16 P=0.003	-3.39 95% CI: -5.63 to -1.16 P<0.001	-4.26 95%CI: -88.59 to -37.37 P<0.001
Change in diastolic blood pressure, mm Hg	-3.96 95% CI: -5.47 to -2.46 P<0.001	-1.78  95% CI: -3.01 to -0.55 P=0.005	-2.18 95% CI: -2.95 to -1.41 P<0.001	-2.18 95% CI: -2.95 to -1.41 P<0.001

	Hypertensive Blacks	Normotensive Blacks	Hypertensive Whites	Normotensive Whites
Change in 24-hour urinary sodium, mmol	-66.87 95% CI: -82.79 to -50.95 P<0.001	-40.31 95% CI: -97.16 to 16.55 P=0.16	-77.46 95% CI: -85.22 to -69.66 P<0.001	-76.45 95% CI: -89.52 to -63.38 P<0.001
Reduction in salt intake, g per day	-3.91  95% CI: -4.84 to -2.98	-2.36  95% CI: -5.68 to 0.97	4.53 95% CI: -4.99 to -4.08	-4.47 95% CI: -5.24 to -3.71
Change in systolic blood pressure, mm Hg	-7.83 95% CI: -10.96 to -4.7 P<0.001	-4.02 95% CI: -7.44 to -0.61 P=0.02	-5.12 95% CI: -6.27 to -3.96 P<0.001	-2.11 95% CI: -3.03 to -1.19 P<0.001
Change in diastolic blood pressure, mm Hg	-4.08 95% CI: -5.90 to -2.26 P<0.001	-1.98 95% CI: -4.45 to 0.49 P=0.12	-2.66 95% CI: -3.37 to -1.95 P<0.001	-0.88 95% CI: -1.68 to -0.08 P=0.03

 

Effect of Salt Reduction on Hormones
 

	Change	95% CI	P value
Plasma renin activity, ng per ml per hour	0.26	0.17 to 0.36	<0.001
Aldosterone, pmol per ml	73.20	44.92 to 101.48	<0.001
Noradrenaline, pmol per ml	187	39 to 336	0.01
Adrenaline, pmol per L	37	-1 to 74	0.06

 

Effect of Salt Reduction on Lipids

No significant changes in plasma lipid levels were observed relative to reduced 24-hour urinary sodium.

Study Quality

• Randomization:
  • N=26 (with adequate concealment)
  • N=8 (not reported).
• Intention to treat:
  • Intention to treat analysis: Seven trials
  • A total of 6.7% participants were lost to follow up following randomization.
• Blinding:
  • Double blind: N=22
  • Observer blind: N=11
  • Open: N=1.

Publication Bias

• Systolic blood pressure:
  • Funnel plot was asymmetrical around effect size, indicating bias (P=0.025)
  • With removal of two larger and long-term studies, asymmetry test was no longer significant (P=0.247).
• Diastolic blood pressure: Funnel plot was symmetrical around effect size (P=0.416).

Other Findings

• Study design:
  • Crossover: N=23
  • Parallel: N=12.
• Blinding:
  • Double blind: N=22
  • Observer blind: N=11
  • Not reported: N=1.
• Study duration:
  • Hypertensive trials:
    • Four weeks: N=10
    • Five weeks: N=2
    • Six weeks: N=3
    • Eight weeks: N=5
    • Twelve weeks: N=1
    • Six months: N=1
    • Twelve months: N=1.
  • Normotensive trials:
    • Four weeks: N=9
    • Six weeks: N=3
    • Eighteen months: N=1
    • Thirty-six months: N=1.

• A moderate reduction in salt intake (4.4g per day) will have a significant effect on blood pressure in hypertensive and normotensive individuals as well as men, women, white and black participants
• Results support a dose response relationship:A larger reduction in salt intake will result in a greater decrease in blood pressure
• Target population salt intake should be 3g per day.

University/Hospital:

Queen Mary University, London; Affiliated Hospital, LuZhou Medical College, Luzhou Sichuan, China

• No search terms were described
• The median study duration was only four weeks, which was short-term.