Nine RCTs (n=370) were included: 5 parallel-group and 4 crossover.
The Jadad quality scores ranged from 2 to 5 out of 5; the median score was 4.
Glycaemic control.
There was no statistically significant difference between magnesium and control in HbA1c (WMD -0.31, 95% confidence interval, CI: -0.81, 0.19, p=0.22; 6 studies with no significant heterogeneity). Magnesium supplements were associated with a significant reduction in fasting glucose compared with control (WMD -0.56 mmol/L, 95% CI: -1.10, -0.01, p=0.03; 8 studies showing significant heterogeneity, p=0.02). The results were similar after the exclusion of one outlier study and significant heterogeneity was no longer found. The funnel plot, based on glycaemic outcomes, showed no evidence of publication bias.
Secondary outcomes.
Magnesium supplements were associated with a significant increase in high-density lipoprotein cholesterol compared with control (WMD 0.08 mmol/L, 95% CI: 0.03, 0.14; no significant heterogeneity). There were no significant differences between magnesium and control for other lipid measures, blood-pressure, BMI or plasma calcium. Magnesium supplements were also associated with a significant increase in plasma/serum magnesium levels compared with control (WMD 0.06 mmol/L, 95% CI: 0.04, 0.08; no significant heterogeneity was detected). The meta-regression analysis showed no significant linear dose-response relationship between magnesium dose and plasma magnesium levels.
Adverse effects (6 studies).
No severe adverse effects were reported. The most common adverse events were gastrointestinal symptoms.