The effect of magnesium supplementation on blood pressure in individuals with insulin resistance, prediabetes, or noncommunicable chronic diseases: a meta-analysis of randomized controlled trials

Abstract

Background: To our knowledge, the effect of magnesium supplementation on blood pressure (BP) in individuals with preclinical or noncommunicable diseases has not been previously investigated in a meta-analysis, and the findings from randomized controlled trials (RCTs) have been inconsistent.

Objective: We sought to determine the pooled effect of magnesium supplementation on BP in participants with preclinical or noncommunicable diseases.

Design: We identified RCTs that were published in English before May 2017 that examined the effect of magnesium supplementation on BP in individuals with preclinical or noncommunicable diseases through PubMed, ScienceDirect, Cochrane, clinicaltrials.gov, SpringerLink, and Google Scholar databases as well as the reference lists from identified relevant articles. Random- and fixed-effects models were used to estimate the pooled standardized mean differences (SMDs) with 95% CIs in changes in BP from baseline to the end of the trial in both systolic blood pressure (SBP) and diastolic blood pressure (DBP) between the magnesium-supplementation group and the control group.

Results: Eleven RCTs that included 543 participants with follow-up periods that ranged from 1 to 6 mo (mean: 3.6 mo) were included in this meta-analysis. The dose of elemental magnesium that was used in the trials ranged from 365 to 450 mg/d. All studies reported BP at baseline and the end of the trial. The weighted overall effects indicated that the magnesium-supplementation group had a significantly greater reduction in both SBP (SMD: −0.20; 95% CI: −0.37, −0.03) and DBP (SMD: −0.27; 95% CI: −0.52, −0.03) than did the control group. Magnesium supplementation resulted in a mean reduction of 4.18 mm Hg in SBP and 2.27 mm Hg in DBP.

Conclusion: The pooled results suggest that magnesium supplementation significantly lowers BP in individuals with insulin resistance, prediabetes, or other noncommunicable chronic diseases.

INTRODUCTION

Magnesium has been hypothesized to have a beneficial effect on hypertension by interacting with calcium (1, 2), reducing peripheral vascular resistance (1, 3, 4), increasing nitric oxide release (5) and endothelial prostaglandin I₂ secretion (2, 6), and enhancing the effect of antihypertensive medications (7). A few meta-analyses and systematic reviews have summarized studies that were conducted mainly in the general populations and showed that magnesium supplementation provides a modest reduction in blood pressure (BP) (8–10). Studies have shown that even a modest reduction in BP is clinically relevant in reducing the risk of coronary heart diseases and stroke (11). In addition, a recent meta-analysis of 34 trials that was conducted in mixed participants showed an inverse association between magnesium supplementation and BP (12). Eight of the studies overlapped with the studies included in the present study (13–20). However, the authors of the meta-analysis did not conduct a stratified analysis by underlying metabolic conditions. In general, previous meta-analyses of randomized controlled trials (RCTs) have mixed studies in apparently healthy participants and participants with primary hypertension with studies in participants with preclinical health conditions or participants with chronic noncommunicable diseases. Note that individuals with or without preclinical or chronic disease may respond to magnesium supplementation differently. In the present study, we included only original studies in participants with underlying preclinical conditions or noncommunicable chronic diseases.

Some preclinical health conditions [e.g., insulin resistance (IR) and prediabetes] and noncommunicable chronic diseases (e.g., type 2 diabetes) are established risk factors for hypertension. Magnesium intake is inversely associated with these health conditions or diseases (21, 22). However, findings from RCTs (13–20, 23–25) on magnesium supplementation and BP in individuals with preclinical or chronic diseases have been inconsistent. To determine the weighted effect and provide additional information to the literature, we conducted this meta-analysis on the effect of magnesium supplementation on BP and hypertension in individuals with diagnosed IR, prediabetes, or other noncommunicable chronic diseases.

METHODS

Data sources and study selection

We searched for original trials that were published in English before May 2017 that related magnesium supplementation to BP or hypertension in individuals with IR, prediabetes, or noncommunicable chronic diseases including type 2 diabetes, cardiovascular diseases, renal diseases, or cancer. We searched for the journal articles in online databases including PubMed (www.ncbi.nlm.nih.gov/pubmed), ScienceDirect (www.sciencedirect.com), Cochrane, (www.cochrane.org) clinicaltrials.gov (www.clinicaltrials.gov), SpringerLink (www.link.springer.com), and Google Scholar (https://scholar.google.com). We used Medical Subject Headings terms including “magnesium,” OR “magnesium” AND “supplementation,” OR “mineral supplementation,” OR “dietary supplementation” combined with “secondary hypertension” OR “hypertension” OR “blood pressure” and “participants with insulin resistance” OR “participants with prediabetes” OR “participants with type 2 diabetes” OR “participants with cardiovascular diseases” OR “participants with renal disease” OR “participants with cancer.” We searched additional trials with the use of the reference lists of relevant articles. The study-selection process is presented in Figure 1.

FIGURE 1.

Study selection process. BP, blood pressure. Web addresses of searched online databases: PubMed (www.ncbi.nlm.nih.gov/pubmed), ScienceDirect (www.sciencedirect.com), Cochrane (Cochrane.org), clinicaltrials.gov, SpringerLink (Link.Springer.com/), and Google Scholar (scholar.google.com).

Inclusion and exclusion criteria

An original study was considered for inclusion if the design was an RCT that was conducted in humans. We included a trial in the meta-analysis if the exposure was magnesium supplementation; participants had IR, prediabetes, or any noncommunicable chronic diseases; and the main outcome was hypertension or BP. To be included, the trial should have reported differences in systolic blood pressure (SBP) and diastolic blood pressure (DBP) with SD, Hedges g, or Cohen’s d values or BP data with SDs that were reported at baseline and at the end of trial. A study was excluded if it was an observational study, an animal study, a review or meta-analysis, a trial in the general population, a trial without relevant effect measures, or a nonmagnesium-supplementation trial.

Data extraction

Two authors (DTD and PX) independently extracted the data and resolved discrepancies by a group discussion with a third author (KH). The extracted data included the first author’s name, year of publication, number of participants in each arm, type of magnesium supplementation, type of placebo, duration of intervention, baseline mean SBP and DBP, mean SBP and DBP at the end of the study, the SD in BP, and changes in BP.

Data synthesis

The pooled SD (S_p) was computed with the use of the following equation:

where S₁ and S₂ are the SDs, and n₁ and n₂ are sample sizes of the magnesium-supplementation group and control group, respectively (26). For trials that reported changes in BPs as mean differences and associated 95% CIs, the CIs were converted into SDs of changes as follows:

where the upper CI (UCI) and lower CI (LCI) are upper and lower confidence limits, respectively (26). For trials that did not report information on changes in BP, we estimated the SDs of changes in BP as follows:

where S₀ and S₁ are the SDs of BP at baseline and at the end of the study, respectively, SD_D is the standard deviation of the changes in BP, and r is the assumed correlation between BP at baseline and the end of the study (27). We assumed r = 0.5 in our main analysis, and used r = 0.3 and 0.7 in the sensitivity analyses.

Statistical analysis

All of the included trials are RCTs that assessed the effect of magnesium supplementation on BP or hypertension in individuals with IR, prediabetes, type 2 diabetes, or cardiovascular diseases. We did not find a trial on the effect of magnesium supplementation on BP or hypertension in individuals with renal diseases or cancer.

We conducted a random-effects and fixed-effects meta-analysis to pool standardized mean differences (SMDs) between the magnesium supplementation group and the control group in the trials. The results from both methods are presented in forest plots, and we report the pooled summary from either method on the basis of heterogeneity. The SMD was calculated as follows:

where μ_Mg is the mean of the magnesium-supplementation group (mean change from baseline to the end of the trial), μ_P is the mean of the placebo arm or control group, and S_P is the pooled SD.

We used Cochran’s chi-square test to examine the heterogeneity in the included trials and computed I², which is the proportion of the total variation that was due to the heterogeneity between the trials, to quantify the degree of inconsistency across trials. We tested the null hypothesis of SMD = 0 with the use of the Z test. Egger’s test (28) and visual inspection of funnel plots were used to assess for the presence of publication bias. We conducted sensitivity analyses by eliminating one study at a time and replacing fixed-effects models with random-effects models. All analyses were conducted with the use of STATA statistical software (version 14; STATA Corp LP). P ≤ 0.05 was considered statistically significant.

RESULTS

Characteristics of included trials

Eleven trials (13–20, 23–25) with a total of 543 participants (278 subjects in the intervention groups) were identified. Eight trials (13, 16–20, 23, 25) used magnesium chloride for the magnesium-supplementation group and a similar-appearing placebo for the control group (Table 1). Two trials (15, 16) used magnesium aspartate hydrochloride with a similar-appearing placebo; 1 trial (14) used magnesium pidolate and a control with no placebo, and the other study (24) used Magnosolv-Granulat (a magnesium product by Meda that contains a total of 365 mg Mg) and an unspecified placebo. All trials (13–20, 23–25) used a parallel-group design.

TABLE 1.

Characteristics of trials included in the meta-analysis¹

										SBP, mm Hg		DBP, mm Hg
Study (ref)	Year	Country	Preclinical or chronic condition	Duration	Group	Magnesium-supplementation regimen per day	n	Age at baseline, y	Sex, M/F, n	Baseline	End of the study	Baseline	End of the study
Rodríguez-Morán and Guerrero-Romero (18)	2014	Mexico	Insulin resistance	4 mo	Treatment	30 mL MgCl2 5% solution (equivalent to 383 mg elemental Mg)	24	31.92	NA	111.3 ± 14.5	109.4 ± 12.4	71.5 ± 6.6	68.8 ± 7.4
					Placebo	30 mL placebo	23	39.52	NA	109.4 ± 12.4	116.6 ± 11.5	71.4 ± 9.3	76.8 ± 7.6
Simental-Mendía et al. (19)	2014	Mexico	Prediabetes	12 wk	Treatment	30 mL MgCl2 5% solution (equivalent to 383 mg elemental Mg)	29	39.8 ± 163	16/13	114.8 ± 31.1	117.5 ± 18.6	76.9 ± 12.9	75 ± 14.5
					Placebo	30 mL NaHCO3 0.1% solution	28	41.1 ± 13.1	17/11	115.7 ± 21.4	123.4 ± 22.2	73.2 ± 10.5	76.9 ± 10.8
Simental-Mendía et al. (25)	2012	Mexico	Prediabetes	12 wk	Treatment	30 mL MgCl2 5% solution (equivalent to 383 mg elemental Mg)	11	44.2 ± 10.8	4/7	116.9 ± 7.6	115.5 ± 18	66.7 ± 6.9	65.6 ± 10.4
					Placebo	30 mL NaHCO3 0.1% solution	11	43.2 ± 7.8	4/7	118.4 ± 8.2	114 ± 9.7	71.8 ± 7.5	66.6 ± 8.7
Mooren et al. (16)	2011	Germany	Insulin resistance	6 mo	Treatment	Verum (magnesium aspartate hydrochloride), 365 mg Mg	25	NA	NA	137.7 ± 14.9	131.4 ± 16.4	85.3 ± 9.4	81.6 ± 9.8
					Placebo	Not specified	22	NA	NA	134.8 ± 15	133.1 ± 21.9	82.5 ± 9.6	83.2 ± 12.1
Barbagallo et al. (14)	2010	Italy	Type 2 diabetes	1 mo	Treatment	4.5 g Mg pidolate (equivalent to 368 mg Mg ion)	30	71.0 ± 4.9	18/12	150 ± 7	148 ± 5	82 ± 5	79 ± 5
					Placebo	No placebo was used in the control	30	71.2 ± 4.6	17/13	148 ± 8	147 ± 6.5	83 ± 5	82 ± 5
Guerrero-Romero and Rodríguez-Morán (13)	2009	Mexico	Type 2 diabetes	4 mo	Treatment	2.5 g MgCl2 (equivalent to 450 mg elemental Mg)	40	58.9 ± 8.5	19/21	161.1 ± 26	140.7 ± 11.9	88.4 ± 14.5	79.7 ± 7.1
					Placebo	Inert placebo	39	60.5 ± 9.4	19/20	154.5 ± 21.2	149.8 ± 20.6	84.9 ± 12.4	83.8 ± 9.7
Barragán-Rodríguez et al. (23)	2008	Mexico	Type 2 diabetes	12 wk	Treatment	50 mL MgCl2 5% solution (equivalent to 450 mg elemental Mg)	12	69 ± 5.9	NA	134.1 ± 19.2	135.2 ± 20.5	77.2 ± 3.1	77 ± 3.6
					Placebo	Imipramine 50 mg	9	66.4 ± 6.1	NA	141.0 ± 20.1	143.7 ± 19.9	84.7 ± 6.1	87.6 ± 6.4
Guerrero-Romero et al. (20)	2004	Mexico	Type 2 diabetes	12 wk	Treatment	2.5 g MgCl2 (equivalent to 450 mg elemental Mg)	32	43.0 ± 7.9	NA	110 ± 8.4	108 ± 8.1	73 ± 7.5	72.3 ± 7.4
					Placebo	Not specified	31	42.2 ± 6.8	NA	111 ± 12	110 ± 11	73 ± 9	72.4 ± 8.9
Rodríguez-Morán and Guerrero-Romero (17)	2003	Mexico	Insulin resistance	16 wk	Treatment	2.5 g MgCl2 (equivalent to 450 mg elemental Mg)	32	59.7 ± 8.3	NA	148.2 ± 32.3	140.2 ± 28.1	86.3 ± 17	82.7 ± 16.4
					Placebo	Not specified	31	54.1 ± 9.6	NA	138.1 ± 25.6	135 ± 19.6	80.5 ± 14.6	79.1 ± 13.5
Shechter et al. (24)	2000	United States	Coronary artery disease	6 mo	Treatment	Magnesium product4 containing 365 mg total Mg (equivalent to 200 mg elemental Mg 2 times/d)	25	68 ± 10	21/3	147 ± 21	143 ± 20	69 ± 10	69 ± 10
					Placebo	Not specified	25	66 ± 12	19/6	145 ± 18	140 ± 19	69 ± 14	68 ± 7
de Valk et al. (15)	1998	Netherlands	Type 2 diabetes	3 mo	Treatment	15 mmol Mg aspartate HCl (365 mg Mg)	18	63.0 ± 8.2	12/6	162.6 ± 23.3	158.7 ± 20	84 ± 11.5	82.9 ± 8.3
					Placebo	Not specified	16	62.0 ± 7.3	8/8	157.4 ± 23.6	146.9 ± 21.8	83 ± 14.2	77.1 ± 8.4

¹DBP, diastolic blood pressure; NA, not available; ref, reference; SBP, systolic blood pressure.

²Mean.

³Mean ± SD (all such values).

⁴Magnosolv-Granulat (Meda).

The duration of the trials ranged from 1 to 6 mo (with an average of 3.6 mo) (Table 1). The mean age of the participants ranged from 31.9 to 71.0 y in the intervention groups and from 39.5 to 71.2 y in the control groups. Two trials (13, 18) reported a significant beneficial effect of magnesium supplementation on both SBP and DBP, whereas the rest of the trials (14–17, 19, 20, 23–25) did not achieve a significant reduction in SBP. One trial reported a significant reduction only in DBP (14).

Change in BP from baseline to the end of the trials

Eleven trials provided data that compared mean changes in BP from baseline to the end of the trials between the magnesium-supplementation group and the control group (13–20, 23–25). With the use of a random-effects model, the pooled results suggest that magnesium supplementation significantly reduces BP in the defined population (for SBP, see Figure 2), and for DBP, see Figure 3). These reductions corresponded to a weighted mean decrement in SBP by 2.22 mm Hg and in DBP by 2.54 mm Hg when comparing the supplementation group with the control group. The magnesium supplementation resulted in a mean SBP reduction of 4.18 mm Hg and a mean DBP reduction of 2.27 mm Hg for the BP at the end of trial compared with baseline.

FIGURE 2.

Forest plot of SMDs (95% CIs) of changes in SBP from baseline to the end of trials for the comparison of the Mg-supplementation group with the placebo group. The horizontal line next to each trial indicates the SMD and associated 95% CI. For each study, the size of the rectangular box indicates the relative weight of the trial and the diamond at the center indicates the point estimate. The dashed line indicates where the point estimate in each study lies compared to the pooled estimates, and the line passes through the pooled estimates. The solid line passes through SMD = 0, and 95% CIs that cross it are not significant. The diamonds at the bottom of the figure indicate pooled SMDs (95% CIs). D+L, DerSimonian and Laird for the random-effects method; ID, identifier; I-V, inverse variance for the fixed-effects method; SBP, systolic blood pressure; SMD, standardized mean difference.

FIGURE 3.

Forest plot of SMDs (95% CIs) of changes in DBP from baseline to the end of trials for the comparison of the Mg-supplementation with the placebo group. The horizontal line next to each trial indicates the SMD and associated 95% CI. For each study, the size of the rectangular box indicates the relative weight of the trial and the diamond at the center indicates the point estimate. The dashed line indicates where the point estimate in each study lies compared to the pooled estimates, and the line passes through the pooled estimates. The solid line passes through SMD = 0, and 95% CIs that cross it are not significant. The diamonds at the bottom of the figure indicate pooled SMDs (95% CIs) of all trials. DBP, diastolic blood pressure; D+L, DerSimonian and Laird for the random-effects method; I-V, inverse variance for the fixed-effects method; SMD, standardized mean difference.

Egger’s test suggested that there was no strong evidence of a publication bias for both SBP (P = 0.07) and DBP (P = 0.49). No significant heterogeneity was shown in the trials for SBP [χ²(10) = 10.21, P = 0.42], but for DBP, there was significant heterogeneity in the trials [χ²(10) = 20.03, P = 0.03].

A sensitivity analysis showed that the mean change in SBP was significantly different from zero in all analyses except in the analysis that excluded Guerrero-Romero and Rodríguez-Morán (13). In all the sensitivity analyses except those that excluded Simental-Mendía et al. (19), Mooren et al. (16), Barbagallo et al. (14), Guerrero-Romero and Rodríguez-Morán (13), or Barragán-Rodríguez et al. (23), the change in DBP was significantly different from zero (r was assumed to be 0.5) (Table 2). The result also remained the same when r = 0.7 was used but disappeared when r = 0.3 was used (Table 2).

TABLE 2.

Sensitivity analysis changes in blood pressure¹

Study omitted (ref)	ΔSBP	ΔDBP
Rodríguez-Morán and Guerrero-Romero (18)	−0.18 (−0.36, 0.00)	−0.23 (−0.40, −0.05)*
Simental-Mendía et al. (19)	−0.20 (−0.38, −0.02)*	−0.25 (−0.52, 0.02)
Simental-Mendía et al. (25)	−0.22 (−0.40, −0.05)*	−0.32 (−0.50, −0.15)*
Mooren et al. (16)	−0.20 (−0.38, −0.02)	−0.26 (−0.53, 0.02)
Barbagallo et al. (14)	−0.21 (−0.39, −0.03)*	−0.26 (−0.53, 0.02)
Guerrero-Romero and Rodríguez-Morán (13)	−0.12 (−0.30, 0.07)	−0.23 (−0.49, 0.04)
Barragán-Rodríguez et al. (23)	−0.21 (−0.38, −0.04)*	−0.25 (−0.51, 0.01)
Guerrero-Romero et al. (20)	−0.22 (−0.40, −0.04)*	−0.30 (−0.57, −0.03)*
Rodríguez-Morán and Guerrero-Romero (17)	−0.21 (−0.39, −0.03)*	−0.29 (−0.56, −0.01)*
Shechter et al. (24)	−0.23 (−0.41, −0.05)*	−0.31 (−0.57, −0.05)*
de Valk et al. (15)	−0.24 (−0.41, −0.06)*	−0.33 (−0.57, −0.10)*
None†	−0.20 (−0.37, −0.03)*	−0.27 (−0.52, −0.03)*
None‡	−0.25 (−0.42, −0.08)*	−0.34 (−0.65, −0.03)*
None§	−0.18 (−0.35, 0.01)	−0.25 (−0.42, 0.08)

¹All values are standardized mean differences (95% CIs). In each row, results are from a fixed-effects model without the inclusion of the study listed. *Pooled results were significant when omitting the study listed in the row. ^†r = 0.5, ^‡r = 0.7, ^§r = 0.3. DBP, diastolic blood pressure; ref, reference; SBP, systolic blood pressure; Δ, difference.

DISCUSSION

The results of this meta-analysis indicate that magnesium supplementation significantly reduces both SBP and DBP in individuals with IR, prediabetes, or other noncommunicable chronic diseases. The findings from this study are generally consistent with results from previous meta-analyses of RCTs in the general population or in participants with primary hypertension and provide additional evidences to the literature supporting the beneficial effect of magnesium supplementation on reducing BP (8–10, 12). In the previous meta-analyses, studies in individuals with or without underlying preclinical health conditions and chronic diseases were combined. This use of this method might partially explain the differences in the magnitude of the effect sizes in the present and previous meta-analyses because participants with underlying preclinical metabolic disease or noncommunicable chronic disease may respond to magnesium supplementation differently from participants without these conditions. In addition, the variations in the population, follow-up period, baseline BP, and dosage and form of magnesium supplementation in included studies may have contributed to the differences in the magnitude of the effect sizes.

The average reductions in blood pressure that were due to magnesium supplementation as observed in the present study (SBP : 4.18 mm Hg; DBP: 2.27 mm Hg) might have relevant clinical effects on cardiovascular health. A clinical trial examined the effects of antihypertensive medications on cardiac outcomes and showed that a reduction of 0.8–2 mm Hg SBP was clinically relevant in reducing the incidence of coronary heart disease, heart failure, and stroke. The trial suggested that a reduction of BP by 2–3 mm Hg might account for a difference of stroke rate by 6–12% between antihypertensive medications (11). This suggestion was also supported by other clinical trials (29) and observational studies (30). Thus, the magnitude of BP reduction in the present meta-analysis is of great clinical significance (31).

The mechanism of the beneficial effect of magnesium on hypertension may include the modulation of vascular tone (1, 3, 4) and the prevention of endothelial dysfunction, carotid intima thickness (32–35), atherosclerosis (36), IR, and hyperglycemia (4). The beneficial effect of magnesium on BP is 2-fold. In one way, magnesium supplementation may directly lower BP, whereas in contrast, it may also improve preclinical conditions and chronic diseases that commonly predispose individuals to hypertension. In addition, magnesium was suggested to have synergetic effects with antihypertensive medications (7). Moreover, accumulated evidences have suggested that magnesium intake is inversely associated with preclinical conditions such as IR (17) and prediabetes (21) and noncommunicable chronic diseases such as type 2 diabetes (37, 38), which are well-known risk factors for hypertension. A meta-analysis showed that magnesium supplementation improves glycemic control in type 2 diabetes patients (39). Furthermore, hypertension affects ∼70% of individuals with diabetes, which is twice the percentage of individuals without diabetes (40). The findings from previous studies have also linked hypertension, hyperinsulinemia, and cardiovascular diseases together (41, 42). A study suggested that hypertension and type 2 diabetes have a synergistic negative health impact on the development of other chronic diseases (43). Approximately 60–80% of individuals with diabetes die of cardiovascular complications and nearly 75% of the cardiovascular complications have been attributed to hypertension (44). Studies have also reported that magnesium has beneficial effects on cardiovascular health (22, 33, 45). In particular, one study suggested that the inverse association between magnesium intake and risk of fatal coronary heart disease may be mediated through hypertension (46). A trial documented that magnesium supplementation improved endothelial function, which resulted in a relatively more flow-mediated vasodilation compared with the placebo group and concluded that the potential mechanism through which magnesium intake is beneficial in coronary artery disease might be through an improvement of endothelial function (24). Another trial reported a nonlinear J-shaped relation of serum magnesium concentrations with all-cause mortality and cardiovascular mortality (47). Thus, magnesium supplementation may play role in breaking the cycle between the preclinical conditions and type 2 diabetes and cardiovascular diseases through the lowering of BP.

Our study has some strengths to highlight. To the best of our knowledge, this study is the first meta-analysis on the effect of magnesium supplementation on BP in participants with IR, prediabetes, type 2 diabetes, and cardiovascular diseases. In this study, the statistical method to calculate the SDs of the pooled changes in SBP and DBP was based on the assumed correlation between the baseline and trial-end BPs when the original trials did not report them. However, the robustness was checked using sensitivity analyses with varying correlation coefficients. The sensitivity analyses indicate that, except in the analysis that excluded Barbagallo et al. (14) (diastolic) or Guerrero-Romero and Rodríguez-Morán (13) or when the correlation coefficient between the baseline and the end of study was assumed to be 0.3, the association remained. In addition, the data indicate no evidence of publication bias. In addition, the studies are relatively more uniform than other previous studies because we included only studies in participants with underlying metabolic conditions. The absence of marked heterogeneity in SBP provided good evidence of the uniformity of the included studies.

However, our study also has some limitations that are in part due to inherent limitations in the original trials. The limitations of the study include relatively short trial durations, limited trials in participants with cardiovascular diseases, and a lack of trials in subjects with renal diseases or cancer, thereby possibly confounding the effect of dietary magnesium intake and the heterogeneity between some of the trials, particularly for DBP. The heterogeneity might partially be attributable to differences in dietary magnesium intake, the duration of study, the dose of supplemental magnesium, the type of supplementation used, unadjusted confounders such as antihypertension drugs, antigastritis drugs, and glucose-lowering drugs, differences in participants’ characteristics (e.g., diet and age), and study design.

In conclusion, the results of this study suggest that magnesium supplementation has a beneficial effect on BP or hypertension in participants with IR, prediabetes, or noncommunicable chronic diseases such as diabetes or cardiovascular disease. Because of the results of our study and the findings from previous studies, patients with hypertension and underlying preclinical metabolic conditions including IR and prediabetes and patients with type 2 diabetes or cardiovascular diseases may benefit from magnesium supplementation in the reduction of BP and the amelioration of the underlying health conditions. The findings from this study add complementary evidence to the literature on the effect of magnesium supplementation on BP. Because of the heterogeneity in the included trials on DBP and no study, to our knowledge, in individuals with renal disease or cancer, future large-scale, well-designed, double-blind, randomized, placebo-controlled clinical trials are warranted to provide more solid evidence of the benefit of magnesium supplementation on BP and possibly on disease outcomes in patients with IR, prediabetes, or noncommunicable chronic diseases.