BACKGROUND Animal and human data suggest that magnesium may play an important role in ischaemic heart disease. Few prospective epidemiological studies have related serum magnesium concentrations to mortality from ischaemic heart disease (IHD) or all-causes. METHODS Data from the National Health and Nutrition Examination Survey Epidemiologic Followup Study were used to examine the association between serum magnesium concentration, measured between 1971-1975, and mortality from IHD or all-causes in a national sample of 25-74-year-old participants followed for about 19 years. RESULTS The analytical samples for IHD and all-cause-mortality included 12 340 and 12 952 participants, respectively (1005 IHD deaths, 2637 IHD deaths or hospitalizations, 4282 total deaths). Hazard ratios for IHD mortality from proportional hazards analysis comparing the second (1.59-<1.68 mEq/l), third (1.68-<1.77 mEq/l), and fourth (> or =1.77 mEq/1) quartiles of serum magnesium concentration with the lowest quartile were 0.79 (95% CI: 0.58-1.08), 0.66 (95% CI: 0.47-0.93), 0.69 (95% CI: 0.52-0.90), respectively. For all-cause mortality, hazards ratios were 0.82 (95% CI: 0.72-0.93), 0.84 (95% CI: 0.73-0.96), 0.85 (95% CI: 0.75-0.95). No significant interactions between serum magnesium concentration and age, sex, race, and education were observed. CONCLUSIONS Serum magnesium concentrations were inversely associated with mortality from IHD and all-cause mortality.