Effect of Magnesium Supplements on Improving Glucose Control, Blood Pressure and Lipid Profile in Patients With Type 2 Diabetes Mellitus: A systematic review and meta-analysis

Summary

This study evaluated the effects of magnesium (Mg) supplementation on glycaemic control in type 2 diabetes mellitus through a meta-analysis of 23 randomised controlled trials involving 1,345 participants. Mg supplementation significantly increased serum Mg levels (weighted mean difference [WMD] = 0.69, 95% CI: 0.32 to 1.06) and reduced fasting blood glucose levels (WMD = –0.58, 95% CI: –0.87 to –0.28). However, the impact on glycated haemoglobin was minimal (WMD = –0.16, 95% CI: –0.32 to 0.00). Subgroup analysis showed a greater reduction in glycated haemoglobin among participants aged ≥65 years and those receiving longer durations of supplementation. Additionally, Mg supplementation was linked to lower diastolic blood pressure and potential improvements in lipid profile. This review highlights the importance of considering patient characteristics, dosage and intervention duration in future research on Mg supplementation for glycaemic control in diabetes.

1. Introduction

Noncommunicable diseases (NCDs), including hypertension and diabetes, are the primary contributors to global mortality and morbidity.¹ The World Health Organization reports that NCDs account for approximately 41 million deaths annually, representing 71% of the total global deaths.² As of 2014, the global prevalence of diabetes was estimated at 8.5%, with projections indicating a significant rise by 2030, particularly in developing nations.³ Various approaches have been proposed to address diabetes-related complications, such as dietary modifications, increased physical activity and dietary supplement use.⁴ One hypothesis suggests that dietary supplements, notably magnesium (Mg), may enhance diabetes management.

Mg is an essential mineral that plays a crucial role in over 300 biochemical reactions necessary for maintaining homeostasis in the human body.⁵ Its functions range from nucleic acid production to influencing adenosine triphosphate-fuelled reactions and modulating activities linked to intracellular calcium concentration fluxes, such as insulin release.⁶

The literature indicates that low circulating total serum Mg is linked to hyperglycaemia, insulin resistance and an elevated risk of metabolic syndrome, type 2 diabetes mellitus (T2DM), chronic kidney disease and cardiovascular disease.^7,8,9,10,11 Studies have shown varying prevalence rates of hypomagnesaemia among individuals with T2DM, ranging from 13.5% to 47.7%, compared to 2.5–15% among healthy controls without diabetes.¹² However, these findings are influenced by factors such as race, gender and overall health status.^13,14,15 Hypomagnesaemia in patients with diabetes has been associated with insulin resistance, poor glycaemic control, hypertension and retinopathy.¹⁶ The intricate relationship between Mg and insulin suggests that Mg supplementation in T2DM patients and hypomagnesaemia may enhance insulin sensitivity.¹⁷

A meta-analysis of 24 trials published in 2023 showed a statistically significant reduction in fasting blood sugar (FBS), glycated haemoglobin (HbA1c), systolic blood pressure (SBP) and diastolic blood pressure (DBP) associated with Mg supplementation.¹⁸ Another meta-analysis of 18 randomised controlled trials (RCTs) conducted in 2022 indicated that oral Mg supplements resulted in significant benefits in FBS and HbA1c.¹⁹ Despite the abundance of research, the efficacy of Mg supplements for glycaemic control in diabetic patients remains inconclusive. Some studies indicate that Mg supplementation improves glycaemic control and helps prevent chronic complications,^20,21 while others do not support such outcomes.²² Given the conflicting findings in the literature and the increasing number of RCTs published in recent years, there is a pressing need to systematically review and meta-analyse pooled data from controlled trials involving adult human subjects to assess the impact of Mg supplements on glycaemic control, specifically FBS and HbA1c, in T2DM patients. Thus, this review aimed to conduct a meta-analysis of these RCTs to evaluate the effectiveness of Mg supplementation in controlling blood sugar levels. Specifically, the primary outcome of this review was to assess the effects of Mg supplementation on the mean reductions in FBS and HbA1c levels. The secondary outcome was to investigate the effects of Mg supplementation on C-peptide levels, blood pressure (BP) control, obesity, lipid profiles and overall cardiovascular risk.

2. Methods

2.1. Protocol and registration

This systematic review and meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (2020) guidelines.²³

2.2. Eligibility criteria

All studies were eligible for inclusion if they met the following criteria: RCTs that administered Mg supplements either orally or intravenously to T2DM patients.

2.3. Search strategy

Two researchers (NM and SB) conducted a thorough and independent search of electronic databases, including PubMed, Scopus, CENTRAL and ClinicalTrials.gov, covering the period from inception to November 2023. Supplementary studies were identified by screening the references of pertinent articles. The search was not limited to articles published in the English language, and no temporal restrictions were applied to the articles [Supplementary Table 1].

2.4. Study selection

The screening process, which involved evaluating titles and abstracts followed by a review of full-text articles, was conducted independently by two researchers (NM and SB). Any discrepancies between the researchers were resolved through consultation with a third researcher (either JM or AA). The inclusion criteria encompassed fully published articles in peer-reviewed journals, and RCTs were considered without language restrictions. The Covidence platform for systematic reviews (https://www.covidence.org) was utilised to organise and manage the studies.

2.5. Quality assessment

Two researchers (NM and SB) performed the quality assessment independently, while JM resolved any discrepancies. The modified Cochrane risk of bias scale was used to assess the quality of the included clinical trials.²⁴ The modified Cochrane risk of bias scale contains 5 questions that evaluate the following criteria: sequence generation, allocation concealment, blinding of participants and personnel and completeness of assessment and reporting of outcomes. A 0–39% score indicates a high risk of bias, a 40%–69% reflects a moderate risk of bias and a 70%–100% signifies a low risk of bias.

2.6. Data extraction

All pertinent details, including the first author, publication year and relevant baseline characteristics (age, gender, body mass index [BMI], HbA1c, fasting blood glucose [FBG], fructosamine, insulin, C-peptide, homeostatic model assessment for insulin resistance [HOMA-IR], SBP, DBP, total cholesterol [TC], total triglycerides [TGs], low-density lipoprotein [LDL] cholesterol, high-density lipoprotein [HDL] cholesterol, glucagon, serum Mg level and erythrocyte Mg), were systematically extracted by the researchers NM and SB. Additionally, the study design and outcome measures (BMI, HbA1c, FBG, fructosamine, insulin, C-peptide, HOMA-IR, SBP, DBP, TC, TG, LDL, HDL, glucagon, serum Mg and erythrocyte Mg) were meticulously documented by the same researchers (NM and SB).

2.7. Quantitative data synthesis

All descriptive values of the outcomes related to BP, blood glucose measurements and lipid profiles, initially presented by the articles as medians and interquartile ranges (IQRs), were converted to means and standard deviations (SDs) using the formula: mean equals median and SD = (Q75 - Q25/1.35).²⁵ The random-effects model was used to identify the weighted mean difference (WMD) for continuous variables, accounting for some heterogeneity between studies, and was illustrated using forest plots for continuous outcomes between the intervention and control groups. The I² statistic was used to identify between-study heterogeneity, with a significance threshold set at P < 0.1 (0–25%, 25.1–75% and 75.1–100% representing low, moderate and high degrees of heterogeneity, respectively).²⁶ Furthermore, for outcomes exhibiting significant heterogeneity, the regression-based Egger test for small-study effects was performed to identify publication bias, with a significance threshold of P < 0.05. Additionally, a funnel plot was used for illustration. To explain the causes of heterogeneity, the random-effects meta-regression model was applied to compare subgroup differences and identify independent factors. Statistical analysis was performed using STATA, Version 17.0 (StataCorp, College Station, Texas, USA).

3. Results

3.1. Study selection

This study identified 7,487 articles from research databases and clinical trial registries. After meticulously screening and subsequently removing 1,951 duplicate references and excluding 5,471 studies that did not meet the predefined inclusion criteria, 65 potential studies remained. However, 10 of these studies were unretrievable, leading to the assessment of eligibility for 55 studies. Following a comprehensive full-text review, 34 studies were excluded for the following reasons: 1 study had incorrect outcomes, 10 studies used incorrect investigations, 8 studies had flawed study designs, 14 studies involved the wrong patient population and 1 study did not report any results. Consequently, 21 studies were included, and an additional 2 studies were identified through reference screening, resulting in a total of 23 studies incorporated into the final evaluation [Supplementary Figure 1].

3.2. Study characteristics

A total of 23 studies met the inclusion criteria, encompassing 1,345 participants, with 741 individuals in the intervention arm.^{22,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48} Among the participants, 36.4% were men, and the mean age was 58.4 ± 7.6 years. The average duration of diabetes among participants was 28.9 ± 28.0 months, with a mean HbA1c level of 8.4 ± 1.6. The diabetes treatments varied: 6 studies employed a diabetic diet, 7 studies implemented insulin therapy and 18 studies utilised oral anti-hyperglycaemic agents. The Mg supplements used in the studies were formulated with different salt components, including organic, inorganic or a combination of both.⁴⁹ Of the 23 included studies, 5 employed organic Mg, 16 utilised inorganic Mg and 1 used a mixed combination. The calculated mean daily elemental Mg dose was 369.2 ± 265.7 mg [Table 1].

Table 1.

Characteristics of included randomised controlled trials on magnesium supplements in patients with type 2 diabetes mellitus.

								Diabetic treatment			Magnesium supplements
Author and year of publicaton	Country	n	Mean age in years ± SD	Mean duration of DM in years ± SD	FBG in mean ± SD/ median (IQR)	HbA1c in % in mean ± SD/ median (IQR)	BMI in kg/m2 in mean ± SD/ median (IQR)	On diabetic diet	On OHA	On insulin	Magnesium formulation	Type of formulation	Route	Total salt dose per day	Elemental dose per day	Duration of treatment in months
Paolisso et al.40 (1989)	Italy	8	67.6 ± 4.7	8.5 ± 3.3	8.1 (7.1–10.4)	8.3.(7.3–9.3)	30.5 ± 0.8	Yes	Yes	No	Diet with magnesium	-	Oral	-	324 mg	1
Paolisso et al.47 (1989)	Italy	8	72.2 ± 2.0	11.5 ± 1.1	9.0 ± 0.4	-	-	Yes	Yes	No	Diet with magnesium	-	Oral	2 g	170 mg	1
Corica et al.28 (1994)	Italy	43	63 ± 5.0	10.7 ± 3.0	8.8 ± 2.2	7.0 ± 0.2	24.4 ± 0.4	Yes	Yes	No	Magnesium pidolate	Organic	Oral	4.5 g	393 mg	1
Gullestad et al.35 (1994)	Norway	54	64 ± 8.0	9.8 ± 8.6	9.6 ± 3.2	7.8 ± 1.5	25.4 ± 3.7	-	-	-	Magnesium lactate citrate	Organic	Oral	15 mmol	365 mg	4
Paolisso et al.39 (1994)	Italy and Belgium	9	73 ± 2.5	7.9 ± 1.2	8 ± 0.1	-	25.8 ± 0.3	-	-	-	Magnesium pidolate MAGZ	Organic	Oral	4.5 g	381 mg	1
Purvis et al.42 (1994)	USA	28	53.8 ± 12.8	19 ± 8.0	-	-	32.2 ± 7.1	Yes	Yes	No	Sustained-release magnesium chloride (Slo-Mag)	Inorganic	Oral	384 mg	98 mg	1.5
Eibl et al.31 (1995)	Austria	38	63 ± 8.0	7.6 ± 6.9	-	7.2 ± 0.7	27.5 ± 3.2	-	-	-	Magnesium citrate	Organic	Oral	30 mmol	729 mg	3
Eriksson and Kohvakka33 (1995)	Finland	27	61 ± 2.0	-	-	9.3 ± 0.3	28.9 ± 0.8	-	-	-	Magnesium	-	Oral	-	600 mg	3
de Lordes Lima et al.29 (1998)	Brazil	128	55.4 ± 10.2	7.2 ± 4.9	10.3 ± 3.3	10.2 ± 2.8	25.3 ± 8.0	Yes	Yes	No	Magnesium oxide	Inorganic	Oral	20.7 mmol	505 mg	1
			51.2 ± 11.0	7.1 ± 5.5	12.6 ± 4.2	9.0 ± 2.4	25.5 ± 6.5	Yes	Yes	No	Magnesium oxide	Inorganic	Oral	41.9 mmol	1,016 mg	1
de Valk et al.30 (1998)	Netherlands	50	63.0 ± 8.2	16.1 ± 8.1	11.8 ± 3.6	8.65 ± 1.45	28.7 (0.7–30.9)	No	No	Yes	Magnesium-aspartate-HCl	Inorganic	Oral	15 mmol	365 mg	3
Rodríguez-Morán and Guerrero-Romero34 (2003)	Mexico	65	59.7 ± 8.3	8.8 ± 4.9	12.8 ± 5.6	11.5 ± 4.1	27.6 ± 9.1	No	Yes	No	Magnesium chloride solution	Inorganic	Oral	2.5 g	630 mg	4
Singh et al.48 (2003)	India	120	-	-	8.0 ± 0.1	-	-	-	-	-	Magnesium chloride	Inorganic	Oral	300 mg	76 mg	4
Guerrero-Romero and Rodríguez-Morán22 (2009)	Mexico	79	58.9 ± 8.5	10.4 ± 6.3	13.6 ± 3.7	13.4 ± 3.9	29.9 ± 5.2	No	Yes	No	Magnesium chloride	Inorganic	Oral	2.5 g/50 mL	450 mg	4
Navarrete-Cortes et al.41 (2014)	Mexico	56	52.84 ± 8.42	<15	8.88 (6.38–12.52)	7.7 (6.4–10.4)	30.55 ± 5.72	No	Yes	No	Magnesium lactate	Organic	Oral	1.5 g	360 mg	3
Solati et al.46 (2014)	Iran	54	46.76 ± 9.0	4.11 ± 4.23	10.2 ± 0.9	8.33 ± 1.47	26.19 ± 2.86	No	Yes	No	Magnesium sulphate	Inorganic	Oral	.	300 mg	3
ELDerawi et al.30 (2018)	Palestine	40	51.15 ± 7.0	<1	8.8 ± 2.4	7.90 ± 0.95	29.02 ± 5.07	No	Yes	No	Jamieson magnesium (oxide, gluconate and lactate)	Mixed of organic and inorganic	Oral	.	250 mg	3
Razzaghi et al.44 (2018)	Iran	70	60.1 ± 11.1	-	12.6 ± 5.0	8.3 ± 1.9	28.2 ± 5.2	No	Yes	Yes	Magnesium oxide	Inorganic	Oral	250 mg	150 mg	3
Rashvand et al.43 (2019)	Iran	96	49.89 ± 7.83	6.47 ± 3.43	8.4 ± 3.9	7.72 ± 1.87	29.69 ± 3.24	No	Yes	No	Magnesium oxide	Inorganic	Oral	500 mg	302 mg	2
Soliman and Nofal36 (2019)	Cairo	122	55.70 ± 12.62	-	-	5.72 ± 1.10	-	-	Yes	Yes	Magnesium sulphate	Inorganic	IV	5g/50 mL	1,000 mg once	3
Talari et al.37 (2019)	Iran	54	58.8 ± 10.1	-	5.9 ± 1.9	7.1 ± 0.9	27.2 ± 5.6	-	-	-	Magnesium oxide	Inorganic	Oral	250 mg	150 mg	6
Sadeghian et al.45 (2020)	Iran	80	41.2 ± 8.8	13.2 ± 8.6	9.1 ± 2.6	7.6 ± 1.4	31.2 ± 5.5	No	Yes	Yes	Magnesium oxide	Inorganic	Oral	-	250 mg	3
Albaker et al.27 (2022)	Saudi Arabia	102	57.5 ± 7.04	-	8.1 (7.1–10.4)	8.3 (7.3–9.3)	31.39 (29.4–34.8)	No	Yes	Yes	Magnesium chloride added to desalinated water	Inorganic	Oral	20 mg/L	2.4 mg/L	3
			55.9 ± 8.9	-	7.9 (6.6–10.0)	8.2 (7.5–8.9)	31.3 (28.5–35.0)	No	Yes	Yes	Magnesium chloride added to desalinated water	Inorganic	Oral	40 mg/L	6 mg/L	3
Drenthen et al.38 (2024)	Netherlands	14	67 ± 6.0	19 ± 8.0	-	7.4 ± 0.9	31.3 ± 4.9	No	Yes	Yes	Magnesium gluconate	Inorganic	Oral	-	360 mg	1.5

SD = standard deviation; DM = diabetes mellitus; FBG = fasting blood glucose; IQR = interquartile range; HbA1c = glycated haemoglobin; BMI = body mass index; OHA = oral hypoglycaemic agents.

3.3. Quality assessment

Quality analysis was conducted using the modified Cochrane risk-of-bias tool for clinical trials.²⁴ The majority of the included studies showcased a low risk of bias in 17 out of 23 included trials (60%), while 6 out of 23 trials (40%) were categorised as having a moderate risk of bias. Notably, none of the studies fell into the high-risk category. The overall risk-of-bias assessment is visually represented using the Robvis tool [Supplementary Figure 2].⁵⁰

3.4.1. Serum Mg concentrations

A total of 19 RCTs were pooled in the meta-analysis of serum Mg concentrations. The analysis indicated that Mg supplementation led to a significant increase in serum Mg concentrations (WMD = 0.69, 95% CI: 0.32 to 1.06; P < 0.01). However, there was considerable between-study heterogeneity (I² = 87.87%; P < 0.01), along with a significant publication bias (P < 0.001), as demonstrated in the funnel plot [Fig. 1]. The subgroup analysis showed that Mg supplementation in individuals aged ≥65 years was associated with increased serum Mg levels (WMD = 1.882, 95% CI: 0.979 to 2.784; P < 0.01) [Supplementary Table 2].

Fig. 1.

Forest plot for the impact of magnesium supplements on serum magnesium concentration (n = 19).

3.4.2. FBG

A total of 20 RCTs were pooled in the FBG meta-analysis. Overall, Mg supplementation was associated with a significant reduction in FBG levels compared to the control group in T2DM patients (WMD = –0.58, 95% CI: –0.87, –0.28; P < 0.01). However, there was considerable between-study heterogeneity (I² = 84.25%; P < 0.01), with no evidence of small-study effects indicating publication bias (P = 0.33) [Fig. 2A]. In terms of the cumulative effect of Mg supplementation on reducing FBG (stratified by elemental dosage, duration of treatment and formulation variation, respectively), the effect of dose on reducing FBG starts at an elemental Mg dose of 150 mg/day, while the effect of treatment duration on reducing FBG becomes significant after 1 month. Additionally, the formulation appears to have a greater impact when using inorganic Mg, peaking with a mixed formulation [Fig. 2B–D]. However, the subgroup analysis revealed a significant reduction only in FBG levels in patients aged ≥65 years (WMD = –2.367, 95% CI: –3.073 to –1.661; P < 0.01) compared to younger patients [Supplementary Table 2].

Fig. 2.

A: Forest plot showing the impact of magnesium supplements on fasting blood glucose (FBG) (n = 20). B: Forest plot showing the cumulative impact of magnesium elemental dose in mg/day on FBG. C: Forest plot showing the cumulative impact of magnesium treatment duration in months on FBG. D: Forest plot showing the cumulative impact of magnesium formulation (organic, inorganic and mixed) on FBG.

3.4.3. HbA1c in %

A total of 16 RCTs were pooled in the meta-analysis of HbA1c levels. Overall, Mg supplementation was not associated with a significant reduction in HbA1c levels compared to the control group in T2DM patients (WMD = –0.16; 95% CI: –0.32 to 0.00; P = 0.05). Additionally, there was considerable between-study heterogeneity (I² = 37.03%; P = 0.06), along with a significant publication bias (P < 0.01). The cumulative effects of Mg supplementation stratified by elemental dose, duration of treatment and formulation variation, respectively, showed insignificant results [Fig. 3B–D]. The subgroup analysis showed that a diabetes duration of ≤8 years is associated with a significant reduction in HbA1c levels compared to patients with a longer duration of diabetes (WMD = –0.049, 95% CI: –0.295 to –0.198; P = 0.043) [Supplementary Table 2]. Furthermore, regression analysis, adjusted for Mg elemental dose (mg/day), identified that a longer duration of Mg treatment is more likely to result in a clinically and statistically significant reduction in HbA1c levels (adjusted coefficient = –0.174; P < 0.01).

Fig. 3.

Forest plot showing the impact of magnesium supplements on glycated haemoglobin (HbA1c) in % (n = 16) in patients with type 2 diabetes mellitus. A: Forest plot showing the impact of magnesium supplements on HbA1C in % (n = 20). B: Forest plot showing the cumulative impact of magnesium elemental dose in mg/day on HbA1C in %. C: Forest plot showing the cumulative impact of magnesium treatment duration in months on HbA1C in %. D: Forest plot showing the cumulative impact of magnesium formulation (organic, inorganic and mixed) on HbA1C in %.

3.4.4. Fructosamine

Only 2 RCTs were pooled in the meta-analysis of fructosamine, which demonstrated homogeneity (I² = 33.92%; P = 0.22) [Supplementary Figure 3A]. However, the analysis revealed no significant effect of Mg supplementation on serum fructosamine levels when compared to the control group (WMD = –0.11, 95% CI: –0.56 to 0.33; P = 0.61).

3.4.5. C-peptide in ng/mL

Again, only 2 RCTs were pooled in the meta-analysis of C-peptide, which exhibited a significant between-study heterogeneity (I² = 97.92%; P < 0.01) [Supplementary Figure 3B]. However, the analysis revealed no significant effect of Mg supplementation on serum C-peptide levels when compared to the control group (WMD = –1.79, 95% CI: –5.73 to 2.15; P = 0.37).

3.4.6. Insulin in umol/L

A total of 11 RCTs were pooled in the meta-analysis of insulin levels, which revealed that Mg supplementation did not significantly alter serum insulin levels when compared to the control group (WMD = –0.13, 95% CI: –0.35 to 0.08; P = 0.22) [Supplementary Figure 3C]. However, there was a considerable between-study heterogeneity (I² = 42.46%; P = 0.05), along with no evidence of small-study effects indicating publication bias (P = 0.76).

3.4.7. HOMA-IR

A total of 9 RCTs were pooled in the HOMA-IR meta-analysis and revealed that Mg supplements did not lead to a reduction following Mg supplements compared to the control group (WMD = –0.28, 95% CI: –0.57 to 0.01; P = 0.06) [Supplementary Figure 3D]. However, there was a considerable between-study heterogeneity (I² = 66.50%; P < 0.01), along with no evidence of small-study effects indicating publication bias (P = 0.07).

3.4.8. BMI in kg/m²

The meta-analysis of BMI included 9 RCTs and demonstrating a strong homogenous study (I² = 0.00%; P = 0.65) [Supplementary Figure 4A]. However, the analysis indicated that Mg supplementation did not significantly affect patients' BMI when compared to the control group (WMD = 0.03, 95% CI: –0.13 to 0.20; P = 0.68).

3.4.9. SBP in mmHg

A total of 8 RCTs were pooled in the meta-analysis of SBP, which demonstrated a non-significant reduction in SBP following Mg supplementation when compared to the control group (WMD = –0.50, 95% CI: –1.06 to 0.06; P = 0.08) [Supplementary Figure 4B]. However, there was a significant between-study heterogeneity (I² = 87.69%; P < 0.01) and no evidence of small-study effects indicating publication bias (P = 0.31).

3.4.10. DBP in mmHg

The meta-analysis of DBP included 8 RCTs and showed a statistically significant reduction in DBP following Mg supplementation when compared to the control group (WMD = –0.36, 95% CI: –0.66 to –0.06; P = 0.02) [Supplementary Figure 4C]. However, there was a significant between-study heterogeneity (I² = 59.14%; P = 0.02) and no evidence of small-study effects indicating publication bias (P = 0.99).

3.4.11. LDL in mmol/L

A total of 11 RCTs were pooled in the meta-analysis of LDL levels, which demonstrated no clinical and statistical effect of Mg supplementation on LDL levels (WMD = –0.13, 95% CI: –0.50 to 0.76; P = 0.68) [Supplementary Figure 4D]. However, there was a strong significant between-study heterogeneity (I² = 93.71%; P < 0.01) and a highly significant small-study effect indicating publication bias (P = 0.01).

3.4.12. HDL in mmol/L

The meta-analysis of HDL levels included 13 RCTs which demonstrated no clinical and statistical effect of Mg supplementation on HDL levels (WMD = 0.60, 95% CI: –0.16 to 1.35; P = 0.12) [Supplementary Figure 4E]. However, there was a strong significant between-study heterogeneity (I² = 96.27%; P < 0.01) and a highly significant small-study effect indicating publication bias (P < 0.01). Additionally, regression analysis, adjusted for Mg elemental dose (mg/day), identified that a longer duration of Mg treatment is more likely to result in a clinically and statistically significant increase in HDL levels (adjusted coefficient = 0.841; P < 0.01).

3.4.13. TG in mmol/L

A total of 13 RCTs were pooled in the meta-analysis of TG levels, demonstrating homogeneity among the studies (I² = 0.00%; P = 1.00) and no evidence of publication bias (P = 1.00) [Supplementary Figure 4F]. The analysis indicated no clinical or statistical effect of Mg supplementation on TG levels (WMD = 0.00, 95% CI: –0.14 to 0.14; P = 1.00). However, the subgroup analysis showed that ≥250 mg elemental doses of Mg supplementation were associated with a significant reduction in TG levels compared to patients receiving <250 mg/day (WMD = 0.157, 95% CI: –0.011 to 0.325; P = 0.023) [Supplementary Table 2].

4. Discussion

This systematic review and meta-analysis present a detailed analysis of the potential therapeutic effects of Mg supplements on glycaemic control and other associated metabolic derangements in T2DM patients. Additionally, it includes recently conducted, high-quality RCTs that have not been previously evaluated in other systematic reviews/meta-analyses. Moreover, this study is among the first to compare both intravenous and oral Mg supplementation. This approach emphasises robust methodological quality and offers a more reliable analysis of Mg's effects on glycaemic control, including thorough subgroup analyses—a aspect that has been overlooked in earlier systematic reviews/meta-analyses.

The results of meta-analyses indicate that Mg supplements are associated with higher serum Mg concentrations. This finding was also observed in several clinical studies, including a meta-analysis of 16 RCTs, which demonstrated that Mg supplements significantly increased serum Mg concentrations (mean difference = 0.15 mg/dL, 95% CI: 0.06 to 0.23; P = 0.001) as well as urinary Mg levels (WMD = 1.99 mg/dL, 95% CI: 0.36 to 3.62; P = 0.017) among T2DM patients compared to the control groups.¹⁸ The study findings suggest a potential correlation between Mg supplementation and its impact on serum Mg levels in elderly patients (WMD = 1.882, 95% CI: 0.979 to 2.784; P < 0.01), which can be explained by 2 hypotheses. First, the reduced renal excretion associated with ageing may contribute to elevated serum Mg levels, thereby influencing insulin sensitivity and glucose metabolism. Second, the decrease in gut motility with age may lead to prolonged Mg retention in the digestive tract, potentially enhancing absorption and subsequently affecting physiological processes related to glucose metabolism. Consequently, higher Mg levels may positively impact glycaemic control.^51,52 It is essential to gain a comprehensive understanding of Mg homeostasis in the elderly through further research.

The study found that Mg supplements led to an overall reduction in FBG levels. It was observed that higher elemental doses of Mg supplements corresponded to greater reductions in FBG levels. Additionally, the study noted that a longer duration of Mg supplementation was associated with improved glycaemic control, as measured by FBG levels.^53,54 It is believed that intracellular Mg plays a crucial role in regulating insulin action and insulin-mediated glucose uptake. Moreover, low intracellular Mg concentrations can impair tyrosine kinase activity, worsening insulin resistance in diabetic patients.⁵⁵ Hence, increasing Mg concentrations may reverse the effects of hypomagnesaemia on insulin activity and resistance, potentially leading to a reduction in FBG levels.

This meta-analysis suggests that Mg supplementation in T2DM patients resulted in a non-significant reduction in HbA1c levels. Previous studies have reported conflicting results regarding the effects of Mg supplements on HbA1c. A meta-analysis by Xu et al. revealed that Mg administration had a significant effect on HbA1c levels in Asians with T2DM, while another meta-analysis by Chua et al. showed that routine Mg supplementation had no effect on HbA1c.^{18,19,55,56,57} The study observed a substantial improvement in HbA1c levels with extended Mg supplementation, specifically when patients were supplemented for 4 months or longer. This notable improvement was evident in 2 out of the 4 RCTs that investigated the impact of Mg supplements lasting 4 months or more.^34,37 The observed effect was less apparent in the other 2 RCTs; 1 had elevated baseline HbA1c levels of 13.4 ± 3.9,²² and the other utilised inorganic Mg citrate.³⁵ However, the reduction in HbA1c levels seems to be more significant when Mg supplements are administered over 6 months. Therefore, it is recommended that future studies incorporate such a prolonged supplementation period to further validate this observation.³⁷

On the other hand, the reduction in FBG levels was statistically significant with a shorter duration of Mg supplementation. However, the duration of Mg supplementation plays a crucial role in influencing HbA1c levels. Additionally, the differing mechanisms and timeframes of FBG and HbA1c measurements contribute to this discrepancy, as FBG levels respond more rapidly to changes in glycaemic control compared to HbA1c, which reflects a longer-term average. This can be attributed to the understanding that HbA1c reflects glycaemic control over the preceding three months, and a shorter duration of Mg supplementation may not be sufficient to demonstrate improvements in HbA1c levels.⁵⁸ Subgroup analysis showed that HbA1c levels significantly reduced in T2DM subjects for less than 8 years. However, despite the statistical significance of this result, the clinical importance is negligible, as the difference in reduction is very small. Previous studies have shown that a longer duration of diabetes adversely affects glycaemic control. A retrospective observational study assessing factors correlated with poor glycaemic control in T2DM revealed that worse glycaemic control was observed in diabetic patients for 5–10 years duration (OR = 1.74) and in patients with a history of diabetes lasting more than 10 years compared to those with less than five years of diabetes (OR = 2.55).⁵⁹

Literature indicates that organic Mg supplements exhibit superior bioavailability. However, the solubility of Mg salts is more critical for tissue accumulation than the dosage itself, whereas absorption is primarily dose-dependent.⁴⁹ Despite the low solubility of Mg oxide, an inorganic salt, it provides a high concentration of ionised Mg, which contrasts with the findings observed for other Mg forms.^60,61 This systematic review found that inorganic Mg supplements led to a reduction in FBG and HbA1c levels. This can be explained by the fact that 8 out of the 16 trials examining inorganic Mg supplements use Mg oxide. Moreover, the trials that tested inorganic Mg used higher doses of elemental Mg and had a longer duration of supplementation.^29,37,43,45

This analysis demonstrated that Mg supplements are associated with a significant reduction in DBP and a non-significant reduction in SBP. To date, there is no consensus about using Mg supplements for BP reduction, as several studies conducted in this area have yielded conflicting results.⁶² However, Mg supplements are believed to be more effective in reducing BP when combined with potassium and calcium supplements than when administered alone.^22,63 One proposed mechanism by which Mg lowers BP is its function as a calcium channel blocker. Mg promotes endothelial-dependent vasodilation by competing with sodium for binding sites on vascular smooth muscle cells and increasing prostaglandin-E levels.^22,63

In this analysis, a slight improvement in HDL levels was observed with Mg supplements in the three enrolled RCTs. However, no association was found between Mg supplements and changes in body BMI, LDL levels or TG levels. This finding aligns with previous research from another systematic review focused on lipid profiles; however, the reduction effect was not apparent across all lipid types.⁶⁴ It seems that the effect of Mg supplements on lipid metabolism varies among different ethnic groups. Mg serves as a crucial rate-limiting factor in the biosynthesis of cholesterol, as it regulates the enzymatic activity of HMG CoA reductase, lecithin cholesterol acyltransferase and desaturase.⁶⁵ If the cellular activity of these enzymes differs significantly between visceral and subcutaneous adipocytes, the effects of Mg may vary among populations.⁶⁶

The systematic review excluded open-label studies that compared Mg against no supplements to enhance reliability and validity while minimising performance and detection biases. By focusing on more rigorous designs, such as double-blind, placebo-controlled trials that studied Mg against placebo supplements, the review ensures stronger and more reliable conclusions. The study involved a detailed evaluation of relevant demographic, biochemical and pharmaceutical characteristics of the included RCTs. The meta-analysis suggests that using inorganic Mg supplements for a longer duration may play a potential role in improving glycaemic control and addressing other associated metabolic derangements in patients with diabetes mellitus. However, it is not without limitations; the inclusion of RCTs with high levels of heterogeneity due to the diverse characteristics of participants and intervention protocols may affect the generalisability and interpretation of the findings.

5. Conclusion

Mg supplements have demonstrated positive effects on serum Mg concentrations and FBG levels in T2DM patients, particularly among individuals aged 65 years or older and those with a shorter duration of diabetes (≤8 years). Although HbA1c reduction was not statistically significant, subgroup analyses suggest potential benefits associated with prolonged Mg use, especially in patients with shorter disease duration and those receiving inorganic Mg supplements. Notably, reductions in FBG levels were observed with an elemental Mg dose starting at 150 mg/day, becoming noticeable after 1 month of treatment. Prolonged Mg treatment may gradually lead to a clinically meaningful reduction in HbA1c levels. Additionally, Mg supplementation was associated with a reduction in DBP and a slight improvement in HDL levels. However, the heterogeneity among the included RCTs and the varying quality of evidence represent the limitations of this analysis. The review highlights the importance of tailoring Mg supplementation for glycaemic control based on patient characteristics, Mg dosage, Mg salt type and intervention duration in future research.

Authors' Contribution

Nasiba Al Maqrashi: Investigation, Investigation, Methodology, Resources, Validation, Writing - Original Draft. Salim Al Busaidi: Investigation, Investigation, Methodology, Project administration, Resources, Validation, Writing - Original Draft. Sara Al-Rasbi: Investigation, Validation, Writing - Original Draft. Abdullah M. Al Alawi: Conceptualization, Methodology, Project administration, Resources, Supervision, Visualization, Writing - Review & Editing. Juhaina Salim Al-Maqbali: Conceptualization, Formal analysis, Methodology, Resources, Software, Supervision, Writing - Review & Editing.

Ethics Statement

This systematic review and meta-analysis were registered in PROSPERO (CRD42023454167).

Data Availability

Data are available upon reasonable request from the corresponding author.