- 2026-04-06
- Nutrients 18(7)
Study Design
- Type
- Meta-Analysis
- Sample size
- n = 1,100
- Population
- type 2 diabetic patients with retinopathy (1100) and without retinopathy (1132)
- Methods
- Systematic review and meta-analysis; comprehensive search of PubMed, ScienceDirect, Google Scholar, and Scopus from inception to January 2026; pooled standardized mean difference estimated using random-effects model
Background/objectives
Magnesium is an intracellular cation that plays important roles in metabolism and insulin signaling. The evidence of association between magnesium levels and diabetic retinopathy is limited by small study effects. Therefore, this systematic review and meta-analysis aim to update the current evidence.Methods
A comprehensive search of PubMed, ScienceDirect, Google Scholar, and Scopus was conducted from database inception to January 2026 to identify studies examining magnesium levels and diabetic retinopathy. The pooled standardized mean difference in magnesium levels between type 2 diabetic patients with retinopathy and those without retinopathy was estimated using the "meta" package in R software.Results
We included seventeen studies which assessed magnesium levels in 1100 patients with diabetic retinopathy and 1132 diabetic controls without retinopathy. The random-effects model indicated significantly lower magnesium levels in patients with diabetic retinopathy compared to diabetic controls [SMD = -1.19, 95% CI (-1.68; -0.70); p < 0.0001; I2 = 95%]. Sensitivity analysis retained all studies, and no evidence of publication bias was detected. Subgroup analyses demonstrated consistent findings across geographic regions (Asian versus non-Asian), study designs (case-control versus cross-sectional), and magnesium assay methods except enzymatic method. Meta-regression analysis revealed that year of publication (coefficient = 0.061; p = 0.009) and non-Asian studies (coefficient = 2.376; p = 0.001) were positively associated with the pooled effect size, while the NOS quality score was inversely associated (coefficient = -0.709; p = 0.035). The magnesium levels were significantly lower in patients with proliferative diabetic retinopathy compared with those with non-proliferative diabetic retinopathy using a fixed effect model [SMD = -1.41, (95% CI: -1.83; -1.00); p < 0.01; I2 = 32%; Cochran's Q statistic (Q = 1.46, p < 0.23)]. The certainty of the generated evidence is rated as low certainty.Conclusions
This systematic review and meta-analysis conclude that magnesium levels are significantly lower in patients with diabetic retinopathy than in diabetic controls without retinopathy. A potential association between hypomagnesemia and the development of diabetic retinopathy in individuals with type 2 diabetes is suggested; therefore, the clinician may check and adjust magnesium levels in patients with type 2 diabetes mellitus.
Research Insights
| Supplement | Dose | Health Outcome | Effect Type | Effect Size | Source |
|---|