Admission serum magnesium level does not predict the hospital outcome of patients with acute myocardial infarction.

OBJECTIVE: To establish whether hypomagnesemia at admission predicts excessive morbidity, particularly cardiac arrhythmias, and mortality in patients with acute myocardial infarction.

METHODS: We compared hypomagnesemic and normomagnesemic patients with acute myocardial infarction in 517 patients admitted to the coronary care unit. The serum magnesium concentration, along with a large array of other parameters, was measured on admission to the emergency department. Other baseline attributes and variables related to the patients' hospital course were used to compare the 2 groups.

RESULTS: The 132 patients (25.9%) with low serum magnesium concentrations at admission (mean +/- SD, 0.61 +/- 0.06 mmol/L [1.48 +/- 0.15 mg/dL]) were comparable to the patients with normal serum magnesium concentrations (0.81 +/- 0.11 mmol/L [1.96 +/- 0.26 mg/dL]) except for a higher rate of prehospital use of diuretic agents (32.6% vs 22.5%, P = .02) and earlier presentation after onset of symptoms (mean +/- SD, 3.2 +/- 4.1 vs 4.8 +/- 6.6 hours, P = .003). There was no correlation between serum magnesium and potassium concentrations in the emergency department (r = 0.14). No difference was detected between the hypomagnesemic and normomagnesemic cohorts in rates of total mortality (18.9% vs 18.5%, P = .91), cardiac mortality (15.2% vs 15.3%, P = .99), atrial fibrillation (13.6% vs 13.8%, P = .97), ventricular tachycardia (18.2% vs 15.3%, P = .44), or ventricular fibrillation (15.2% vs 13.5%, P = .63). Management of the 2 cohorts was not different, except for higher rates of use of magnesium (17.4% vs 1.3%, P < .001) and potassium (59.8% vs 42.1%, P < .001) supplements and antiarrhythmic drugs (62.9% vs 48.7%, P = .005) in the hypomagnesemic patients. An endogenous rise in serum magnesium level was documented in a subgroup of 161 patients who had a repeated measurement (0.74 +/- 0.12 mmol/L [1.79 +/- 0.29 mg/dL] in the emergency department vs 0.77 +/- 0.09 mmol/L [1.88 +/- 0.23 mg/dL] in the coronary care unit, P < .001).

CONCLUSIONS: We conclude that hypomagnesemia is seen in approximately one fourth of patients with myocardial infarction, is not linked to hypokalemia, has some relationship to preadmission use of diuretic agents, is associated with early presentation to the hospital, and is not a predictor of increased morbidity or mortality.