Blood Pressure Reduction Effects of Salt Substitutes in Different Application Settings A Meta-Analysis of Randomized Controlled Trials

Figures

• 

  Figure 1. Flowchart of Literature Search and Selection The selection process for randomized controlled trials (RCTs) evaluating the effects of salt substitutes on blood pressure. A total of 296 records were identified. After removing duplicates, screening titles, and abstracts, and assessing full texts, 21 RCTs were included in the final analysis: 18 were conducted in home-cooking settings and 3 were in collective-cooking settings.

• 

  Figure 2. Effects of Salt Substitutes on Systolic and Diastolic Blood Pressure The squares represent the effect of salt substitutes on blood pressure in each RCT. The diamonds represent the blood pressure–lowering effect of salt substitutes within each application setting and overall, respectively. The horizontal lines represent the prediction interval. Random-effects meta-analysis with DerSimonian-Laird estimation was used to pool effect sizes. DBP = diastolic blood pressure; IV = inverse variance; MD = mean difference; SE = standard error; SBP = systolic blood pressure; other abbreviation as in Figure 1.

• 

  Figure 3. Results From Sensitivity Analyses Except for the Leave-One-Out Analysis The square represents the effect of salt substitutes on blood pressure in each application setting. MD and 95% CIs were pooled by meta-analysis using random-effects model. P values for difference were derived from Cochran’s Q test. Chi square statistics and corresponding P values for heterogeneity for each sensitivity analysis are provided in Supplemental Table 3. Abbreviations as in Figures 1 and 2.

• 

  Figure 4. Results From the Leave-One-Out Sensitivity Analysis The squares and diamonds represent effects of salt substitutes on blood pressure in home-cooking and collective-cooking settings, respectively, based on leave-one-out analysis. MD and 95% CIs were pooled by meta-analysis using random-effects model. P for difference were estimated by Cochran’s Q test. Chi square statistics, I2, and corresponding P values for heterogeneity are provided in Supplemental Table 4. Abbreviations as in Figures 1 and 2.

• 

  Central Illustration. Blood Pressure–Lowering Effect of Salt Substitutes in Different Application Settings Evidence from 21 RCTs (29,570 participants). Of these, 18 were conducted in home-cooking settings and 3 in collective-cooking settings. Pooled analyses show larger reductions in SBP with salt substitutes in collective-cooking than in home-cooking settings. The reduction in DBP was borderline significant. DBP = diastolic blood pressure; RCT = randomized controlled trial; SBP = systolic blood pressure.