**Background:** Fragility fractures are a major public health problem, with lifetime risks of 22% in men and 46% in women at age 50. While calcium and vitamin D have been extensively studied, the role of magnesium (Mg) in bone health is less established. Magnesium is a major bone component (67% of body Mg is in bone) and is involved in over 600 enzymatic reactions. Experimental studies suggest Mg deficiency increases osteoclast activity and decreases osteoblast activity. This systematic review and meta-analysis aimed to investigate the association between serum Mg concentrations and incident fracture risk.
**Methods:** The review followed PRISMA guidelines with a pre-planned but unpublished protocol. Two investigators independently searched PubMed/Medline and Scopus from inception to May 24, 2022, using the search strategy (magnes* AND fractur*). Inclusion criteria were: observational studies (case-control or longitudinal), evaluation of serum Mg at baseline, reporting incident osteoporotic fractures, and written in English. Exclusion criteria included non-human studies, dietary Mg assessment, and cross-sectional designs. Data extraction used a standardized Excel spreadsheet. Quality assessment employed the Newcastle-Ottawa Scale (NOS). The primary outcome was incidence of any osteoporotic fracture (total and by site: hip, vertebral, wrist). Risk ratios (RRs) with 95% confidence intervals (CIs) were calculated using random-effect models, with the highest Mg quantile as reference. Heterogeneity was assessed using I² and χ² statistics. Publication bias was evaluated via funnel plots and Egger's test, with trim-and-fill analysis.
**Key Results:** From 1,332 initial records, 17 full texts were retrieved, and 4 studies met inclusion criteria (2 from Japan, 2 from Europe). Three studies were prospective cohorts, one was retrospective. Two studies involved hemodialysis patients, two involved general populations. Total sample size was 119,755 participants, followed for a mean of 79 months (range 24–161). Mean age was 62 years, 33% were women. All four studies scored 9/9 on the Newcastle-Ottawa Scale (low risk of bias). Lower serum Mg concentrations were significantly associated with higher fracture risk (RR = 1.579; 95% CI: 1.216–2.051; p = 0.001; I² = 46.9%). Heterogeneity was not significant (I² = 31.2%, p = 0.201). Egger's test showed no publication bias (0.94 ± 0.43; p = 0.10). After trim-and-fill adjustment, the effect size was slightly reduced but remained significant (RR = 1.25; 95% CI: 1.09–1.43). Individual study findings: Hayhoe et al. (2016) reported significant trends in fracture risk across serum Mg in men for spine (p = 0.02) and total fractures (p = 0.02). Kunutsor et al. (2017) found fully adjusted HR for total fracture of 1.80 (1.10–2.94) and for femoral fracture of 2.13 (1.13–3.99) comparing bottom vs. top quartile. Sakaguchi et al. (2018) reported a 1.23-fold higher hip fracture risk for lower Mg quartile (95% CI: 1.06–1.44; p < 0.01), with each 1-mg/dL increase associated with 14.3% decreased hip fracture risk (95% CI: 3.8–23.8; p < 0.01). Hori et al. (2021) found cumulative fracture incidence of 17.6% vs. 5.2% (lower vs. higher Mg), with multi-adjusted HR of 2.31 (95% CI: 1.03–5.17; p = 0.03).
**Clinical Implications:** This meta-analysis provides strong evidence that lower serum Mg is associated with increased fracture risk across different populations, including both general populations and hemodialysis patients. The findings are supported by multiple biological mechanisms: Mg modulates osteoclast and osteoblast activity; Mg is a cofactor for vitamin D synthesis and metabolism; Mg deficiency impairs PTH secretion and action; and Mg has anti-inflammatory properties. The population attributable fraction for hip fracture due to low serum Mg was 13.7% (95% CI: 3.7–22.7%), higher than for calcium, phosphate, or PTH. Given that approximately 49–60% of the US population has Mg intake below estimated requirements, hypomagnesemia represents a potentially modifiable risk factor. Limitations include the small number of studies (4), geographic restriction to Japan and Europe, inclusion of hemodialysis patients in two studies, and the limitation of serum Mg as a marker (only 1% of total body Mg is in blood, potentially underestimating mild to moderate deficits). Further research in diverse populations and development of practical intracellular Mg measurement methods are needed.