Table 4.
Summary of results from prospective studies and meta-analyses of trials and cohort studies on the association of magnesium and hypertension included in the review.
| Authors/Country | Year | N. of Trials or Prospective Cohort Studies | N. of Participants/Cases | Study Characteristics | Magnesium Dose | Duration of Follow-up or Trials | Summary of Results |
|---|---|---|---|---|---|---|---|
| Witteman et al. USA [36] | 1989 | - | 58,218/3275 | Prospective cohort | - | 4 years | For women with high intakes of magnesium vs. low intakes, the RR of hypertension was 0.65 (95% CI, 0.53–0.80). |
| Ascherio et al. USA [38] | 1992 | - | 30,681/1248 | Prospective cohort | - | 4 years | Among male health professionals, dietary magnesium was significantly associated with lower risk of hypertension after adjustment for age, relative weight, alcohol consumption, and energy intake. |
| Ascherio et al. USA [37] | 1996 | - | 41,541/2526 | Prospective cohort | - | 4 years | Among women who did not report hypertension during follow-up, magnesium was significantly inversely associated with self-reported systolic and diastolic BP, after adjusting for age, BMI, alcohol consumption, and energy intake. Dietary magnesium was not significantly associated with risk of hypertension, after adjusting for age, BMI, alcohol, and energy intake. |
| Peacock et al. USA [41] | 1999 | - | 7731/1577 | Prospective cohort | - | 6 years | Significant trend for the association of serum magnesium and incident hypertension in women, after adjustment for age, race, and other risk factors (p trend = 0.01) but not in men (p trend = 0.16). No association between dietary magnesium intake and incident hypertension. |
| Townsend et al. USA [43] | 2005 | - | 10,033/1045 in NHANES III 2311/299 in NHANES IV | Two waves national survey | - | Similar intakes of magnesium and other minerals in hypertensive and non-hypertensive participants in both surveys. The pattern of significantly lower mineral intake (potassium + calcium + magnesium) emerged as unique to persons with isolated systolic hypertension in both waves. | |
| He et al. USA [39] | 2006 | - | 4637/608 MS | Prospective cohort | - | 15 years | Magnesium intake was inversely associated with incidence of metabolic syndrome after adjustment for major lifestyle and dietary variables and baseline status of each component of the metabolic syndrome. The inverse associations were not modified by gender and race. Magnesium intake was also inversely related with individual component of the metabolic syndrome. |
| Song et al. USA [40] | 2006 | - | 28,349/8544 | Prospective cohort | - | 9.8 years | Among women, magnesium intake was inversely associated with the risk of hypertension (p for trend < 0.0001 of magnesium quintiles). This inverse association was attenuated but remained significant after further adjustment for known risk factors (p for trend = 0.03). Similar associations were observed for women who never smoked and reported no history of high cholesterol or diabetes at baseline. |
| Jee et al. Korea, USA [45] | 2002 | 20 (14 in hypertensives) | 1220 | Meta-analysis of interventional studies | 10–40 mmol/d | 3–24 wks | Apparent dose-dependent effect of magnesium on BP, with reductions of 4.3 mm Hg in systolic BP and of 2.3 mm Hg in diastolic BP for each 10 mmol/d increase in magnesium dose. Limiting the analysis to the 14 trials in hypertensives, for each 10 mmol/d of magnesium SBP was reduced by 3.3 mm Hg and DBP by 2.3 mm Hg. |
| Dickinson et al. UK [226] | 2006 | 12 | 545 | Cochrane review- Meta-analysis of RCTs | 10–40 mmol/d | 8–26 wks | On average, people receiving magnesium achieved slightly but significantly lower DBP (mean difference: −2.2 mmHg). Poor quality and heterogeneity of the trials. None of the studies reported any serious side effects. |
| Kass et al. UK [24] | 2012 | 22 | 1173 | Meta-analysis of interventional studies | 120–973 mg/d | 3–24 wks | Small but significant reduction in SBP of 3–4 mm Hg and DBP of 2–3 mm Hg, with greater increased in trials with crossover design and magnesium dose >370 mg/d. |
| Rosanoff et al. USA [227] | 2013 | 7 | 135 | Meta-analysis of interventional studies | 10.5–18.5 mmol/d | 6–17 wks | Significant mean reduction in SBP (mean −18.7 mmHg) and DBP (mean −10.9 mmHg) in hypertensives on continuous anti-hypertensive medication for at least six months, with no more than a two-week washout, and mean starting SPB > 155 mmHg. |
| Zhang et al. USA, China, Canada, Japan [23] | 2016 | 34 | 2028 | Meta-analysis of RCTs | 238–960 mg/d | 3 wks to 6 months | Significant reduction in SBP (mean −2.0 mmHg) and DBP (mean −1.78 mmHg) accompanied by 0.05 mmol/L rise in serum magnesium vs. placebo. Greater BP reduction found in trials with high quality or low dropout rate. |
| Dibaba et al. USA, Israel [22] | 2017 | 11 | 543 | Meta-analysis of RCTs | 365–450 mg/d | 1–6 months | Significant decrease in BP: mean reduction of 4.18 mm Hg in SBP and 2.27 mm Hg in DBP in participants with insulin resistance, prediabetes, or other noncommunicable chronic diseases. |
| Verma et al. India [44] | 2017 | 28 (19 trials included for HTN analyses, 4 in hypertensives) | 1694 | Meta-analysis of RCTs | 300–1006 mg/d | 4–24 wks | Significant reduction in SBP (weighted mean difference = −3.056 mmHg) with greater beneficial effect in diabetic patients with hypomagnesaemia. High heterogeneity of the trials. In meta-regression, elemental magnesium dose was inversely DBP (p < 0.001). |
| Han et al. China, Sweden, USA, Norway [21] | 2017 | 9 | 180,566/20,119 | Meta-analysis of prospective cohort studies | - | 4–15 years | Inverse association between dietary magnesium intake and the risk of hypertension. A 100 mg/d increment in magnesium intake was associated with a 5% reduction in the risk of hypertension. The association of serum magnesium concentration with the risk of hypertension was marginally significant. |
| Wu et al. China, USA [228] | 2017 | 11 (3 on HTN) | Total: 38,808/4437 HTN: 14,876/3149 |
Meta-analysis of prospective cohort studies | - | 6–8 years | Comparing highest vs. lowest category of circulating magnesium concentration, the pooled RR was 0.91 (95% CI 0.80, 1.02) for incident hypertension. Every 0.1 mmol/L increment in circulating magnesium levels was associated with 4% (RR 0.96; 95% CI: 0.94, 0.99) reduction in hypertension incidence. |
| Ikbal et al. Austria [25] | 2019 | 8 (5 of RCTs, 3 of observational studies) | RCTs: 135–1694 | Summary of meta-analyses | 120–1006 mg/d | RCTs: 3–24 wks; observational studies: 4–15 years |
The summary showed SBP reductions in the range of −0.2 and −18.7 mmHg, and DBP reductions between −0.3 and −10.9 mmHg. The meta-analysis [227] showing the largest effect, included a small sample of treated hypertensive patients, which probably responded highly to magnesium. When omitting this meta-analysis, the BP lowering effects of magnesium were attenuated to a low to moderate level. Observational studies showed a lower risk for hypertension with increasing magnesium intake or higher circulating magnesium levels. |
| Veronese et al. Italy, UK, Australia, Spain [31] | 2020 | 16 meta-analyses | RCTs: 2262 participants in 34 RCTs; Observational studies: 180,566/20119 |
Umbrella review of systematic reviews and meta-analyses | 120–1006 mg/d | RCTs: 3–24 wks; observational studies: 4–15 years |
High class evidence for the association of diastolic blood pressure and magnesium in intervention studies with magnesium supplementation vs. placebo and moderate class evidence for systolic blood pressure. Large heterogeneity found for this outcome. The evidence was suggestive for the association of a higher dietary magnesium intake with a lower risk of stroke in observational studies. |
BMI: body mass index; BP: blood pressure; CI: confidence interval; d: day; DBP: diastolic blood pressure; HTN: hypertension; MS: metabolic syndrome; RR: relative risk; SBP: systolic blood pressure; wks: weeks.