Several studies have examined the association between serum magnesium and hyperglycemia, but findings have been inconsistent. Some prospective cohort studies report inverse associations, while others show no significant relationship, and cross-sectional studies have yielded mixed effect sizes. This systematic review and dose-response meta-analysis examines the association between serum magnesium and hyperglycemia. A comprehensive and systematic search was undertaken utilizing PubMed, Scopus, ISI Web of Science, and Google Scholar, extending up to January 2026. Epidemiologic studies presenting relative risks or odds ratios with 95% confidence intervals regarding serum magnesium and hyperglycemia, type 2 diabetes mellitus (T2DM), and prediabetes were included. The GRADE approach was utilized to evaluate the body of evidence. A total of 24 studies (7 cohorts, 16 cross-sectional, 2 cross-sectional analyses of cohort baseline data, and 1 case-control) were included. In cohort studies including 46,190 participants and 4,330 cases, negative associations were observed in the highest versus lowest category of serum magnesium with hyperglycemia (RR = 0.67; 95%CI: 0.51, 0.78) and T2DM (RR = 0.67; 95%CI: 0.56, 0.82); this relation was not significant in the case of prediabetes (RR = 0.69; 95%CI: 0.42, 1.02). In non-cohort studies with 40,834 participants and 8,161 cases, reductions in odds were observed for hyperglycemia (OR = 0.43, 95%CI: 0.32, 0.57), T2DM (OR = 0.34, 95%CI: 0.24, 0.46), and prediabetes (OR = 0.53, 95%CI: 0.36, 0.76). In addition, each 0.20 mmol/L increase in serum magnesium level was associated with lower risk of T2DM (RR = 0.92; 95%CI: 0.87, 0.97) in cohort studies and lower odds of hyperglycemia (OR = 0.79; 95%CI: 0.72, 0.86) and T2DM (OR = 0.77; 95%CI: 0.69, 0.87) in cross-sectional studies. In non-linear dose-response analysis, increasing serum magnesium concentration from deficient to sufficient values (> 0.85 mmol/L) was associated with a decreasing trend in hyperglycemia and T2DM risk. There were significant inverse dose-response associations between serum magnesium and hyperglycemia and T2DM. The association with prediabetes was weak and uncertain.
Rashki et al. (Mon,) studied this question.