Abstract
Objective Traditional pairwise meta-analyses indicated that nuts consumption can improve blood pressure. We iamed to determine the dose-dependent effect of nuts on systolic (SBP) and diastolic blood pressure (DBP) in adults.
Methods A systematic search was undertaken in PubMed, Scopus, and ISI Web of Science till March 2021. Randomized controlled trials (RCT) evaluating the effects of nuts on SBP and DBP in adults were included. We estimated change in blood pressure for each 20 g/d increment in nut consumption in each trial and then, calculated mean difference (MD) and 95%CI using a random-effects model. We estimated dose-dependent effect using a dose-response meta-analysis of differences in means. The certainty of evidence was rated using the GRADE instrument, with the minimal clinically important difference being considered 2 mmHg.
Results A total of 31 RCTs with 2784 participants were included. Each 20 g/d increase in nut consumption reduced SBP (MD: -0.50 mmHg, 95%CI: -0.79, -0.21; I2 = 12%, n = 31; GRADE = moderate certainty) and DBP (MD: -0.23 mmHg, 95%CI: -0.38, -0.08; I2 = 0%, n = 31; GRADE = moderate certainty). The effect of nuts on SBP was more evident in patients with type 2 diabetes (MD: -1.31, 95%CI: -2.55, -0.05; I2 = 31%, n = 6). The results were robust in the subgroup of trials with low risk of bias. Levels of SBP decreased proportionally with the increase in nuts consumption up to 40 g/d (MD40g/d: -1.60, 95%CI: -2.63, -0.58), and then appeared to plateau with a slight upward curve. A linear dose-dependent reduction was seen for DBP, with the greatest reduction at 80 g/d (MD80g/d: -0.80, 95%CI: -1.55, -0.04).
Conclusions The available evidence provides a good indication that nut consumption can result in a small improvement in blood pressure in adults. Well-designed trials are needed to confirm the findings in long term follow-up.
Introduction
High blood pressure is the leading risk factor for cardiovascular disease (CVD) and death across the globe. In 2019, elevated blood pressure attributed for about 10.8 million death in the world, accounting for 19% of all deaths (1). Despite effective antihypertensive medications (2), the global burden of high blood pressure is still high (3).
The effect of eating habits on levels of blood pressure have long been investigated. There is evidence that unhealthy eating habits such as high sodium and low potassium intake, alcohol drinking, and poor diet quality may contribute to developing hypertension (4). Adiposity is another diet-related condition that is highly associated with the risk of developing hypertension (5). Epidemiologic research has also suggested that high nut consumption was associated with a lower risk of developing hypertension (6).
Nuts are rich in mono- and polyunsaturated fatty acids, minerals such as magnesium and potassium, fibers, and antioxidants and thereby can confer protection against hypertension (7). Several meta-analyses of randomized controlled trials (RCT) have therefore been undertaken of the effects of nuts on blood pressure levels and have indicated that increasing the consumption of nuts can significantly decrease blood pressure levels (8-11).
However, existing meta-analyses only performed pairwise comparisons between intervention and control groups and did not consider differences between doses of nuts consumption in intervention groups across trials, as well as between intervention and control groups in each trial. The identification of dose-response associations is an important part of the analyses in nutrition research that can help determine whether the increasing the level of dose is effective, and to select the optimal dose for implementing the most effective interventions (12, 13).
Dose-response meta-analysis of differences in means is a new statistical approach that helps to address these considerations when evaluating the effects of a specific intervention on continuous outcomes (14). Dose-response meta-analysis of differences in means can present valuable information on the effects of different doses of an intervention on a continuous outcome that cannot be obtained by traditional pairwise meta-analyses and can determine the shape of dose-dependent effect. We, therefore, aimed to perform a systematic review and dose-response meta-analysis of RCTs to evaluate the effects of different doses of nuts consumption on systolic (SBP) and diastolic blood pressure (DBP) in adults.
Methods
The present dose-response meta-analysis has been reported following the Preferred Reporting Items for Systematic Reviews and Meta-analyses: the PRISMA statement (15). The protocol of the systematic review was registered in PROSPERO (CRD42021247346).
Systematic search
To find potential eligible RCTs for inclusion in this review, we searched PubMed, Scopus, and ISI Web of Science up to March 20201. We supplemented the database search by manually reviewing the reference lists of all existing related reviews. The search in the databases and reference lists was restricted to articles published in English. We combined keywords related to intervention, outcome, and study design to find potential eligible RCTs. The complete search strategy is described in Supplementary Table 1. Teams of two reviewers independently screened titles and abstracts according to the pre-defined inclusion and exclusion criteria to identify potential eligible trials.
Eligibility criteria
We applied PICOS (population, intervention, comparator, outcome, and study design) framework to define our inclusion and exclusion criteria. Published human intervention studies were considered eligible for inclusion in the present meta-analysis if they had the following criteria: 1) RCTs, either with parallel or cross-over design, conducted in adults aged 18 years or older, regardless of health status; 2) evaluated the effect of nuts (mixed nuts or one of the nuts subtypes) on SBP and/or DBP; 3) compared the effect of different doses (g/d) of a specific nut on blood pressure across more than one study arm (e.g, 60 g/d vs 30 g/d pistachio) or compared the effect of the specific amount of nuts (g/d) against a nut-free diet; 4) considered the change in SBP and/or DBP as the primary or one of the secondary outcomes; 5) provided mean and standard deviation (SD) of change in DBP or SBP across study arms or reported sufficient information to estimate those values; and 6) reported the number of participants in each study arm.
Trials with non-randomized design, quasi-experimental studies, trials conducted in adolescents (under 18 years of age), pregnant and lactating women, those that applied an active control group with possible effects on blood pressure (e.g, nuts vs fruit or nuts vs omega-3) (16), trials that did not specify the amount of nuts consumption in the control group (e.g, control group was usual or healthy diet without reporting the amount of nuts consumption), as well as trials that compared the effects of two different types of nuts (e.g, pistachio vs almond) were excluded. We excluded trials that compared the effects of two different types of nuts because of their active control group.
Data extraction
Two reviewers (HS and SM) independently and in duplicate screened the full texts of eligible trials and extracted the following data: author and year, population location, study design and duration, characteristics of the population (mean age +/- SD, baseline BMI, health status), total sample size, intervention characteristics (type and dose of nuts consumption), comparison group, calorie restriction, physical activity, behavioral support, outcome measures and main results for the outcomes included. Disagreements between the two reviewers were resolved by discussion.
Risk of bias (quality) assessment
Two reviewers (AJ and SSB) independently and in duplicate performed risk of bias assessments using the Cochrane risk of bias tool (17). An overall quality score was given to the trials based on bias domains: good (≤1/5 items were unknown and none were high), fair (≤2/5 items were unclear or at least one high), and high risk of bias (≥2/5 items were high). Disagreements regarding the risk of bias assessment were resolved by discussion.
Statistical analysis
We considered weighted mean difference (MD) and 95% confidence interval (CI) of change in SBP and DBP as the effect size for reporting the results of the present systematic review. First, we calculated changes from baseline blood pressure in each study arm. If the mean values and SDs of changes were not available, we calculated these values by using data from measures before and after the intervention, according to the guidelines of the Cochrane Handbook (18). When standard errors instead of SDs were presented, the former was converted to SDs (19). If studies reported medians and interquartile ranges, we used the median to impute the missing mean and calculated SDs by dividing interquartile ranges by 1.35 (19). If none of these options was available, we imputed the missing SDs using pooled SDs obtained from other trials included in our meta-analysis (20).
Second, we used the method introduced by Crippa and Orsini (14) to calculate MD and its corresponding SD of change in SBP and DBP for each 20 g/d increment in nuts consumption in the intervention group relative to the control group in each trial. This method requires dose (g/d) of nuts consumption in each study arm, the mean and its corresponding SD of change in SBP and DBP in each study arm, and the number of participants in each arm. Trial-specific results were pooled using a random-effects model (21).
We then performed a series of pre-defined subgroup analyses based on health status, baseline weight (normal weight or overweight/obese), presence of calorie restriction, physical activity, or behavioral support in the intervention program, as well as the type of funding (industrial vs university), follow-up duration, and risk of bias assessment. Influence analysis was carried out to test the potential impact of each trial on the pooled effect size. The potential for publication bias was tested using Egger’s test (22), Begg’s test (23), and by inspection of funnel plots. We assessed heterogeneity quantitatively using the I2 statistic and performed a χ2 test for homogeneity (Pheterogeneity> 0.10) (24).
Finally, we performed a dose-response meta-analysis to clarify the shape of the effect of different doses of nuts on SBP and DBP (14). Statistical analyses were conducted using STATA software version 16.1. A two-tailed P value of less than 0.05 was considered significant.
Results
The database and reference lists search identified 4502 articles. We excluded 291 duplicates and further 4103 articles based on screening of the title and abstract. Eventually, 108 full texts were fully reviewed for eligibility. Of those, 31 randomised trials with 2784 participants were eligible for inclusion in dose-response meta-analysis (27-57). Detailed screening and data extraction processes are depicted in Supplementary Figure 1, and list of studies excluded based on reviewing the full texts are presented in Supplementary Table 2.
Characteristics of randomised trials
Supplementary Table 3 presents the general characteristics of the trials included in the meta-analysis. Eligible trials were published between 2007 and 2020. Twelve trials were conducted in individuals at high risk of CVD (with at least one cardiometabolic risk factor) (36, 37, 40, 42, 45-47, 49-51, 53, 55), six trials each in patients with type 2 diabetes (30, 38, 41, 43, 48, 56) and healthy individuals (27, 32-34, 50, 52), four in patients with the metabolic syndrome (29, 44, 54, 57), and three in those with a history of CVD (28, 31, 39). Fourteen trials were conducted exclusively in individuals with overweight/obesity (30-33, 35, 36, 40, 42, 45, 47, 49, 50, 52, 53), and the remainders in mixed populations (27-29, 34, 37-39, 41, 43, 44, 46, 48, 51, 54-57).
The dose of nut consumption ranged between 15 to 85 g/d. Three trials compared the effect of two doses of a specific type of nut against a nut free diet (52, 54, 55), and the remainder compared the effects of a nut-supplemented diet against a nut free diet (27-51, 53, 56, 57). One study compared the effect of two different types of nuts against a nut free diet and thus, was included as two study arms in the analysis (44).
Of the trials, 22 lasted ≤12 weeks (28-34, 38, 40-45, 47, 48, 51-55, 57), seven trials lasted >12 to ≤24 weeks (35, 37, 39, 46, 49, 50, 56), and two trials >24 weeks (27, 36). Six trials applied behavioral support (33, 35, 36, 42, 49, 50), five trials implemented a calorie-restricted diet (33, 36, 49, 50, 56), and three implemented exercise (36, 49, 50) alongside nuts supplementation. Of the trials, 12 were rated to have low risk of bias (27, 28, 32, 34, 40, 41, 45, 46, 49, 50, 53, 56), nine trials had moderate risk of bias (some concerns) (29-31, 35, 39, 42-44, 47), and 10 were rated to have high risk of bias (33, 36-38, 48, 51, 52, 54, 55, 57) (Supplementary Table 4).
The effect of nuts on systolic blood pressure
All 31 trials reported data for the effect of nuts on SBP. Each 20 g/d increment in nuts consumption reduced SBP by -0.50 mmHg (95%CI: -0.79, -0.29; I2 = 12%, Supplementary Figure 2). The effect size remained significant after step-wise exclusion of each study from the main analysis (MD range: -0.43 to -0.57).
Table 1 presents the results of the subgroup analyses. The results remained significant in healthy individuals, in patients with type 2 diabetes and those with overweight/obesity, trials with a follow-up ≤12 weeks, those with low risk of bias, and trials that received industrial funding, though tests for subgroup deference were not significant. There was an indication of a larger effect size in trials that implemented calorie restriction (MD: -1.25; 95%CI: -2.16, -0.24). There was no indication for publication bias with Egger’s test (P = 0.24), Begg’s test (P = 0.39), or with the funnel plot (Supplementary Figure 3).
Subgroup analysis of the effect of nuts consumption (20 g/d increase) on systolic blood pressure.
Dose-dependent effects of nuts on levels of SBP are indicated in Figure 1, upper panel and Table 2. Levels of SBP decreased proportionally with the increase in nuts consumption up to 40 g/d (MD40g/d: -1.60, 95%CI: -2.63, -0.58), and then appeared to plateau with a slight upward curve (Pnonlinearity = 0.08, Pdose-response= 0.002).
The effects of different doses of nuts on blood pressure form the nonlinear dose-response meta-analysis (mean difference and 95% confidence interval).
The effect of nuts on diastolic blood pressure
Each 20 g/d increment in nuts consumption significantly reduced DBP (MD: -0.23; 95%CI: - 0.38, -0.08; I2 = 0%, Supplementary Figure 4). Sensitivity analysis did not show evidence of a heavy impact of each individual trial on pooled effect size (MD range: -0.21 to -0.25). Table 3 indicates the subgroup analyses. The effect size was significant in patients at high risk of CVD and those with overweight/obesity, short-term follow-up trials, and in trials with low risk of bias. There was no indication for publication bias (Egger’s test P = 0.72, Begg’s test P = 0.71; Supplementary Figure 5).
Subgroup analysis of the effect of nuts consumption (20 g/d increase) on diastolic blood pressure.
Dose-dependent effect of nuts on DBP are indicated in Figure 1, lower panel, and Table 2. Levels of DBP decreased linearly and slightly (Pnonlinearity = 0.89, Pdose-response = 0.02) up to nuts consumption of 80 g/d (MD80g/d: -0.80, 95%CI: -1.55, -0.04)
Grading the evidence
The certainty of evidence was rated using the GRADE approach. For both outcomes, the certainty of evidence was rated moderate for downgrades for serious risk of bias and imprecision, and an upgrade for dose-response gradient (Supplementary Table 5).
Discussion
The present meta-analysis is the first study to investigate the potential dose-dependent effect of nut intake on blood pressure using randomized trials. The analyses indicated that each 20 g/d increment in nuts intake can significantly decrease SBP and DBP by 0.50 and 0.23 mmHg, respectively. The nonlinear dose-response meta-analyses suggested a beneficial dose-dependent effect. Although the certainty of evidence was rated moderate for both outcomes, the effect of nuts on blood pressure levels were small and lower than our MCID (2 mmHg), suggesting the benefit was trivial.
Our findings regarding the effect of nuts intake on blood pressure are in accordance with the results of the previous pairwise meta-analyses. A meta-analysis of 21 randomized trials (11) indicated that nuts consumption can improve SBP in patients with type 2 diabetes by 1.29 mmHg that was comparable to 1.60 mmHg in the present meta-analysis. The study also indicated that pistachios and mixed nuts can improve DBP by 0.80 and 1.19 mmHg, respectively. The greatest DBP reduction in our meta-analysis was about 0.80 mmHg at a dose of 80 g/d. Other pairwise meta-analyses of RCTs have also indicated that dietary interventions to increase the consumption of pistachios (8), cheshaw (9), and almond (10) can improve levels of blood pressure.
However, previous meta-analyses only performed pairwise comparison between intervention and control groups. As mentioned above, there are large variations in the amount of nuts consumed in interventions groups across trials (ranged from 15 to 85 g/d). In addition, the difference between dose of nuts consumption between intervention and control groups within a trial has not been considered in the traditional pairwise meta-analyses. Optimal dose of nuts consumption for reducing blood pressure levels has also not been determined. Dose-response meta-analyses are useful tools for determining the efficacy of increasing dose and selecting optimal dose for intervention (12, 13).
Our analyses indicated that the greatest reduction for SBP was seen at nuts consumption of 40 g/d. An inverse linear association was seen for DBP up to 80 g/d.The favorable effect of nuts on blood pressure levels can be attributed to their high content of mono- and polyunsaturated fatty acids, magnesium, potassium, antioxidants, and dietary fibers (7). Nuts also have anti-inflammatory properties and can improve insulin resistance (58), the two underlying mechanisms for developing hypertension (59, 60). Nuts are rich in arginine and thereby can induce endogenous production of nitric oxide (61), a potent vasodilator (62). Nuts have also been ranked to be the best food group in reducing levels of high density lipoprotein cholesterol (63) and thus, can protect blood vessels against endothelial dysfunction (64).
Despite these favorable effect, our results indicated that the favorable effect of nuts in reducing SBP and DBP was small and was lower than our MCID for blood pressure (2 mmHg). The magnitude of the effect was small in comparison with antihypertensive medications (65), and was modest in comparison with other lifestyle modifications such as aerobic exercise (2.58 to 3.84 mmHg) (66).
In the analysis of SBP, subgroup analyses indicated a stronger effect in patients with type 2 diabetes, as well as in trials that implemented calorie restriction. However, tests for subgroup difference were not significant. In addition, the analyses of DBP did not indicate such stronger effect sizes. There was also no effect modification by physical activity and behavioral support in the analyses of SBP and DBP.
Although nuts have small antihypertensive properties, they are one of the key components of the healthy eating patterns such as the Mediterranean diet and dietary approaches to stop hypertension dietary pattern. The American Heart Association recommends eating four servings (42 grams) of unsalted, unoiled nuts per week for the primary prevention of CVD (67). Epidemiologic research also indicated that each 28 g/d increase in nuts consumption was associated with 7% to 39% lower risk of cardiometabolic disease and all-cause and cause-specific mortality (68). Nuts have a healthy fatty acid profile and are rich in vegetable proteins, fibers, mineral, phytosterols, and carotenoids and thus have anti-inflammatory, antioxidant, and lipid lowering properties (69).
The findings of the present meta-analysis should be interpreted considering the following limitations. Our main limitation is that 26 potentail RCTs could not be included in the present meta-analysis because the dose of nut consumption was not reported in the control groups. This issue can be considered by future trials. Second, of 31 trials included in this review, 22 trials lasted ≤ 12 weeks. More trials are needed to determine whether the favorable effect of nuts on blood pressure levels can exert more than a slight clinical impact in long term. Third, of the trials, only 12 trials have been rated to have low risk of bias. However, subgroup analyses in studies with low risk of bias confirmed the main findings. Fifth, the compliance of participants to study interventions was rarely reported. Thus, we were unable to assess attrition bias in the included trials and its potential impact on the overall results. Sixth, due to lack of trials conducted exclusively in patients with hypertension, we were unable to assess the effects in hypertensive individuals.
Conclusions
The present dose-response meta-analysis of randomized trials gave a good indication that nut consumption can result in small improvement in blood pressure levels in adults. The results suggested a dose-dependent effect, with the greatest reduction being observed at nuts consumption of 40 g/d for SBP and 80 g/d for DBP. Well-designed trials are needed to confirm the findings in long term follow-up.
Data Availability
The data used in the present study will be provided upon request.
References
