Zhang · Clinical nutrition (Edinburgh, Scotland) 2018 · prospective cohort study · n=14353

Serum magnesium concentrations and all-cause, cardiovascular, and cancer mortality among U.S. adults: Results from the NHANES I Epidemiologic Follow-up Study.

Cited 36 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective cohort study with long-term mortality follow-up

PubMed 28890274 · doi:10.1016/j.clnu.2017.08.021 · record verified 2026-08-30

What was done

Analyzed prospective cohort data from 14,353 US adults aged 25–74 years from the NHANES I Epidemiologic Follow-up Study (NHEFS) who had baseline serum magnesium (Mg) measured between 1971 and 1975. Mortality was tracked through December 31, 2011 (median follow-up of 28.6 years). Weighted multivariate-adjusted Cox proportional hazards models were used to estimate hazard ratios (HRs) for all-cause, cardiovascular disease (CVD), cancer, and stroke mortality across seven serum Mg categories (<0.70, 0.70–0.74, 0.75–0.79, 0.80–0.89 [reference], 0.90–0.94, 0.95–0.99, and ≥1.00 mmol/L).

What was found

During follow-up, 9,012 total deaths occurred (3,959 CVD, 1,923 cancer, and 708 stroke deaths). Compared with the reference group (0.80–0.89 mmol/L), participants with serum Mg <0.70 mmol/L had a significantly increased risk of all-cause mortality (HR 1.34, 95% CI: 1.02–1.77) and stroke mortality (HR 2.55, 95% CI: 1.18–5.48). Non-significant increases in risk for the <0.70 mmol/L category were observed for cancer mortality (HR 1.39, 95% CI: 0.83–2.32) and CVD mortality (HR 1.28, 95% CI: 0.81–2.02). HRs for all-cause mortality in the remaining categories were: 0.70–0.74 mmol/L (HR 0.94, 95% CI: 0.75–1.18), 0.75–0.79 mmol/L (HR 1.08, 95% CI: 0.97–1.19), 0.90–0.94 mmol/L (HR 1.05, 95% CI: 0.95–1.16), 0.95–0.99 mmol/L (HR 0.96, 95% CI: 0.79–1.15), and ≥1.00 mmol/L (HR 0.98, 95% CI: 0.76–1.26).

Why it matters

This study provides long-term, nationally representative evidence that clinical hypomagnesemia (<0.70 mmol/L) is associated with substantially increased risks of all-cause and stroke mortality in the general population.

Limits

Serum magnesium was measured only once at baseline (1971–1975), potentially introducing misclassification over almost three decades of follow-up. The observational design cannot prove causality or rule out residual confounding. Wide confidence intervals for cause-specific outcomes indicate limited statistical power in the lowest magnesium category.

Cited by