Current issues of ACP Journal Club are published in Annals of Internal Medicine


Intravenous magnesium for acute myocardial infarction: a meta-analysis

ACP J Club. 1992 May-June;116:68. doi:10.7326/ACPJC-1992-116-3-068

Source Citation

Teo KK, Yusuf S, Collins R, Held PH, Peto R. Effects of intravenous magnesium in suspected acute myocardial infarction: overview of randomised trials. BMJ. 1991 Dec 14;303:1499-503.



To determine the effect of intravenous magnesium on mortality in suspected acute myocardial infarction.

Data sources

All published and unpublished randomized controlled trials of intravenous magnesium in acute myocardial infarction were sought through MEDLINE searches, from references in relevant articles, and through personal inquiries.

Study selection

1 single-blind, and 6 double-blind, randomized, placebo-controlled trials were selected.

Data extraction

Missing data were obtained from the authors whenever possible; mortality data were obtained for 99% of the 1301 patients. Typical (common) odds ratios (COR) for mortality, serious ventricular arrhythmia, and cardiac failure were calculated. Treatment was started soon after admission to the hospital. All studies excluded patients with advanced atrioventricular block, and some studies excluded patients with elevated serum creatinine concentrations, hypotension, severe cardiac failure, cardiogenic shock, or age > 70 years. Magnesium sulfate was given in 5 studies and magnesium chloride in 2 studies, at a total dose of 30 mmol to 90 mmol, over 24 to 48 hours, by infusion or by 20-minute bolus injections.

Main results

In 6 of the 7 studies, short-term (varying from 24 hours to 1 month) mortality was lower in the magnesium-treated group. Overall, 25 patients (3.8%) assigned to magnesium died compared with 53 patients (8.2%) assigned to placebo (COR, 0.45; 95% CI, 0.28 to 0.71; P < 0.001), indicating a 55% reduction in the odds of death. 1 study with 1-year follow-up found little further reduction in mortality after 1 month, and the only other study with 1-year follow-up found no difference for the magnesium group at either time point. Ventricular arrhythmias (variously defined but generally detected during the hospitalization period) were less common among the magnesium group than among the control group (70 patients [11%] vs. 109 patients [17%]; COR, 0.51; CI, 0.36 to 0.73; P < 0.001). Cardiac failure did not appear to be influenced by magnesium administration.


Early treatment with adjunctive intravenous magnesium improved survival in the first few weeks among patients with suspected acute myocardial infarction. Serious cardiac arrhythmias were also decreased.

Source of funding: Not stated.

Address for article reprint: Dr. S. Yusuf, Clinical Trials Branch, Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Bethesda, MD 20892.


Patients in the early hours of an acute myocardial infarction are at risk of sudden death caused by ventricular arrhythmias. Disturbances of serum electrolytes, including hypomagnesemia, can potentiate arrhythmias during myocardial ischemia. One may, therefore, reasonably hypothesize that supplemental intravenous magnesium may reduce arrhythmias and cardiac death during an acute myocardial infarction.

The 7 randomized studies included in this overview are relatively small (median, 115 patients; range, 48 to 400 patients). Meta-analysis can improve statistical power by aggregating small studies; therefore this method is appropriate to apply here. The combined data are reasonably consistent and are clearly statistically significant, with a reduction of 55% in the odds of death and of 49% in the odds of serious arrhythmias.

These findings are interesting and provocative, but for several reasons I do not consider them conclusive. First, the pathophysiologic rationale for magnesium in acute myocardial infarction is plausible but not well established. Second, intravenous magnesium has not been the subject of intense clinical investigation (in contrast to thrombolysis or β-blockade), and "negative" studies may possibly not have been published. The tendency to publish positive findings (publication bias) can lead to optimistic judgments of therapeutic efficacy (1). Third, the risk reduction by intravenous magnesium of 55% is incredibly high, comparable to the risk reduction by thrombolytic therapy given in the first hour, and I find it difficult to believe magnesium is that effective. Thus, I would consider this meta-analysis to be generating a hypothesis, not establishing it.

Fortunately, large multicenter investigations (e.g., LIMIT-2, ISIS-4) are currently under way to test the hypothesis that intravenous magnesium is beneficial in acute myocardial infarction. Until these studies have been reported, intravenous magnesium should be regarded as promising but unproven therapy.

Mark A. Hlatky, MD
Stanford University Stanford, California