Effect of grape products on blood pressure: a systematic review and meta-analysis of randomized controlled trials

ABSTRACT Previous studies have suggested that grape and its products may possess blood pressure (BP)-lowering properties. Due to inconsistencies in results, we aimed to systematically examine the effect of grape products on BP by conducting a meta-analysis of randomized controlled trials (RCTs). PubMed, Scopus, Web of Science (ISI), and Cochrane Library databases were comprehensively searched until March 2020. Human clinical trials which reported the effect of grape products supplementation on systolic BP (SBP) and diastolic BP (DBP) were included. Data were pooled using a random-effects model and expressed as a weighted mean difference (WMD) with a 95% confidence interval (CI). Twenty-eight studies comprising a total of 1344 subjects were included in our meta-analysis. The overall outcome of the meta-analysis indicates that grape products consumption can significantly reduce SBP (WMD: −3.40 mmHg, 95% CI: −6.55, −0.24, p = .03, I2 = 93.4%) and DBP (WMD: −1.69 mmHg, 95% CI: −3.12, −0.27, p = .01, I2 = 80.4%). This meta-analysis found a moderate and statistically significant reduction for either SBP or DBP with grape products compared with controls. Additional high-quality studies are needed to further evaluate the causal conclusions.


Introduction
Hypertension (HTN), a medical condition when systolic/diastolic BP reaches more than 140/ 90 mmHg, is one of the primary risk factors for several health problems such as cardiovascular diseases (CVDs), renal failure, and sudden death, affecting approximately one billion individuals worldwide. [1] Using medications and lifestyle changes including dietary supplements, low sodium intake, and exercise are common treatments for HTN. [2] However, there has been a surge of interest to find new agents with BP-modifying properties to be used as adjuncts to low dose antihypertensive drugs in patients who cannot tolerate higher doses.
Recent studies have reported that oxidative stress, the state of overproduction of reactive oxygen species (ROS), can play a crucial role in developing HTN. [3][4][5] Increasing number of evidences suggest that improvement in systemic antioxidant activity has beneficial effect on reversing deleterious changes in arteries' endothelium and BP. [6,7] Plant polyphenolic compounds are of powerful selection process, disagreements between researchers were resolved by face-to-face discussion to achieve consensus.

Data extraction
We recorded study characteristics as follows: first author's last name, publication year; design details, including whether parallel or crossover; study duration; number of participants; daily dose of intervention. Participant characteristics including health status, mean age, mean body mass index (BMI) and baseline SBP and DBP were also recorded. When aforesaid characteristics were not reported in available publications, we contacted the corresponding author to acquire the necessary data. Two of the authors (O.A. and E.Gh) independently performed the data extraction, and disagreements resolved by discussion.

Risk for bias assessment
Two reviewers (O.A. and E.Gh.) independently assessed the quality of each study according to the Cochrane risk of bias. [22] This scale involves of 7 criteria to assess the risk of the bias which are as follows: random sequence, generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other biases. Bias is assessed as a judgment (high, low, or unclear) for individual elements, which are interpreted as high risk, low risk and unknown risk, respectively.

Statistical analyses
Data were analyzed using Stata version 12.0 software (StataCorp, College Station, Texas, USA). Blood pressure was measured in mmHg. The effect size of each study was calculated from mean and standard deviation (SD) of the outcomes before and after the intervention and presented as weighted mean difference (WMD) with 95% confidence intervals (CI). If only SD for the baseline and final values was provided, SD for the net changes was assigned based on the Follmann method [23] using a correlation coefficient of 0.5. Where standard error (SE) was only reported, SD was estimated as follows: SD = SE × sqrt (n), where n is the number of participants in each group. Due to the fact that selected RCTs were carried out in different settings, the random-effects model was employed to calculate the overall effect from effect sizes. Heterogeneity was examined using the I-squared (I 2 ) index. [24] An I 2 value >50% was considered to indicate substantial heterogeneity between trials. [25] To explore the source of heterogeneity, as well as the possible influences of study designs and participant characteristics on combined effect sizes, we further conducted pre-specified subgroup analyses stratified by trial duration, baseline BMI of subjects, type of intervention, health status, and baseline SBP. Sensitivity analysis was also performed to explore the extent to which inferences might depend on a particular trial using the leaveone-out method. Publication bias was assessed by visual inspection of funnel plots and formally complemented by Begg's test, where P < .10 was considered evidence of small study effects. All tests were two-sided. P values ˂0.05 were considered statistically significant, except where otherwise specified.

Study selection
From 1198 provided articles in initial search, 323 duplicated studies excluded. After screening of title and abstract 843 unrelated studies discarded because of primary evaluation of inclusion criteria: unrelated title or abstract (n = 740), animal studies (n = 63) and review studies (n = 40). The remaining 32 studies were screened based on full text for eligibility, 4 studies were excluded due to lack of BP reporting. [26,27] Ultimately, 28 studies with 1344 participants were included. The PRISMA flow diagram of search process is illustrated in Figure 1.

Study characteristics
The characteristics of studies included are outlined in Table 1. Included studies were published between 2004 and 2017. The follow-up period ranged from 2 [28] to 52 [29] weeks. The sample size of   Of the 190 subjects that underwent actual screening, eightyfour subjects were eligible for the study. Of these, ten subjects were randomly excluded and four subjects were assigned to be spare subjects to allow for some dropout before the start of the intervention.

Effect of grape products on systolic blood pressure
The effect of grape products supplementation on SBP was investigated in 28 trials with 33 arms (772 cases and 729 control subjects). Overall, meta-analysis indicated that SBP decreased significantly following grape products supplementation (WMD: −3.40 mmHg, 95% CI: −6.55, −0.24, p = .03). Due to a significant heterogeneity between studies (I 2 = 93.4%, p < .001) (Figure 2), subgroup analyses were performed based on baseline SBP, duration of intervention, intervention type, health status, and baseline BMI. Betweenstudy heterogeneity was decreased or disappeared after subgroup analysis by baseline SBP, intervention type, health status, and baseline BMI. However, after classifying the studies, the results remained significant only in the following subsets:  Table 3).

Effect of grape products on diastolic blood pressure
The effect of the grape products supplementation on DBP was examined in 33 arms from 28 studies (772 cases and 729 control subjects). Overall, current meta-analysis revealed significant effects of grape products on DBP (WMD: −1.69 mmHg, 95% CI: −3.12, −0.27, p = .01). There was significant heterogeneity among studies (I 2 = 80.4%, p < .001) (Figure 3). Subgroup analysis based on intervention type, health status, and baseline BMI was decreased or disappeared between-study heterogeneity.   Table 3).

Sensitivity analysis
To discover the influence of each single study on the combined effect size, we removed each trial from the analysis, step by step. We found that after removing the Sivaprakasapillai   Robinson (WMD: −3.10 mmHg, 95% CI: −6.75, 0.53) studies for SBP and Belcaro (A) (WMD: −1.54 mmHg, 95% CI: −3.14, 0.06) study for DBP the statistical results were changed to insignificant.

Publication bias
Visual inspection of funnel plots delivered no evidence for publication bias in studies involved in the current meta-analysis ( Figure 4). The results of the Begg's test were SBP (P = . 19) and DBP (P = .54).

Discussion
As far as we know, this is the first quantitative review evaluating the effects of grape products such as grape extract, grape juice, grape seed extract, and raisins on BP parameters including SBP and DBP. In the current systematic review and meta-analysis of 28 clinical trials were published between 2004 and 2017, we found that grape products supplementation in a period of 2 to 52 weeks intervention, significantly reduced SBP and DBP levels. In the subgroup analysis, we investigated a significant reduction in both SBP and DBP levels when baseline SBP ≥130 mmHg and DBP ≥85 mmHg, when grape seed extract was administered, in healthy subjects and participants with baseline BMI >30 kg/m 2 .
In line with our results, a systematic review of 39 studies that assessed the effects of grape polyphenols on metabolic syndrome components, exhibited that seven studies found a significant decrease in BP indices, but only one was of high quality. [54] Another meta-analysis of 10 clinical trials published in 2015 found a significant reduction in SBP by 1.48 mmHg after grape polyphenol intake. Contrarily, there were no significant changes in DBP levels in the grape polyphenols group as compared to controls. [55] It should be noted, mentioned meta-analysis had a small sample size leading to unsteady approximates of therapeutic effects and also restricted the capacity of randomization to lessen the possible effect of confounding factors. A further review of 16 RCTs and 810 study subjects which evaluated the impact of grape seed extract supplementation on the BP parameters, found a significant reduction in SBP and DBP levels. [20] However, in mentioned review, because of retrieving all included articles from the English-language literature, it remained a possibility of selection bias. Furthermore, there were relatively small sample sizes in stratified analysis. Some published clinical trials demonstrated that various grape products supplementation improved BP indices [,42,48-50] whereas the other trials did not. [29,36,45] In agreement with our findings, a study indicated that phenolic compounds from purple grape juice improved BP parameters, upon receiving supplementation (10 mg/kg/day) for 28 days. This study found a significant decrease in SBP by 5.3 mmHg, while there were no significant changes in DBP levels and mean BP. [51] As can be noted, because participants in mentioned study were normotensive and exercise practitioners, the ability of exercise training in reducing BP led to the absence of a hypotensive effect on diastolic component. [56] Altogether, different biological activity of grape products contributed to dissimilarities in the results explained here. Differences in the context of grape varieties, geographical and botanical grape's origin and, production process affect the biological activity of grape extracts. [57][58][59] Grape is a phenol-rich fruit. Proanthocyanidins, anthocyanins, flavonols, flavanols, resveratrol, and phenolic acids are Phenolic compounds in grape. [19] Several polyphenolic compound of grape and its antioxidants prevent cell damage due to free radicals. [60] Stimulation and promotion of the release of NO, resulting in the vasorelaxation, might be the main cause of hypotensive effect of grape polyphenols. [55] In addition, polyphenols endothelial function could be increased by nitric oxide bioactivity and eventually lowering BP. Resveratrol could enhance the expression and activity of eNOS, possibly through the activation of PI3K/Akt pathway. [6,61] Low-molecular-weight procyanidinrich grape seed extract (LM-GSPE) administration to rats increased 6-keto-prostaglandin F1α (PGF1α) plasma levels, relaxed SHR aorta rings, and finally exhibited antihypertensive effect. [62] PGF1α, a stable metabolite of prostacyclin, is an important vasodilator endothelial factor. [63] Angiotensin-converting enzyme (ACE), a zinc metalloenzyme, converts angiotensin-I into SBP DBP Figure 4. Funnel plots for SBP, and DBP. angiotensin-II, which is a vasoconstrictor. [64] Phenolic compounds deactivate metal ions by its chelation ability. [65] Because ACE is a metalloenzyme, phenolic compounds bond with its zinc ion and therefore decrease its activity. [2] Altogether, the hypotensive effect of grape products may be related to the level of prostacyclin and reduction of ACE activity, which are influenced by phenolic compounds. However, additional confirmation needed.
Despite the interesting findings of the current meta-analysis, many potential limitations should be addressed. First, grape products supplementation was used in different dosages and various types. Second, variable and wide duration of intervention led to the bias in our meta-analysis. Third, subjects involved in included studies had different physiological status and various age groups. And fourth, lifestyle modifications during the intervention of grape products were not reported in a large number of included articles. Dissimilar lifestyle modifications and diets may affect the impact of grape polyphenols on BP. In addition, the present study was not registered in the International Prospective Register of Systematic Reviews (PROSPERO), which may be a limitation as well. However, this review and meta-analysis was designed and performed according to the Cochrane guidelines. Although, our study has important strengths. All studies included in our review were high quality, well-designed, randomized, double-blinded trials. Furthermore, using Begg's test, there was no publication bias for SBP and DBP. In addition, majority of considered studies permitted a complete subgroup analysis contributing to determine plausible sources of heterogeneity.

Conclusion
Conclusively, our results exhibited that grape products administration improved BP parameters including SBP and DBP levels, and this effect was more obvious in healthy subjects, as well as in subjects with baseline BMI >30 kg/m 2 . Larger, better designed trials, that specifically include hypertensive subjects, are required to verify our results in the future.