Related Experiment Videos
Beta blockade and the secondary prevention of myocardial infarction
Abstract:
In 1965 Snow reported a clinical trial in which treatment with propranolol significantly reduced mortality following myocardial infarction. Unfortunately the design of this trial was inadequate by modern standards, and the results must be discounted. None of the studies published since then have provided convincing evidence that beta blockers are useful following myocardial infarction. Survival after a myocardial infarction depends principally upon the amount of heart muscle that has been destroyed, and it is probably unreasonable to expect any treatment to reduce mortality by more than 20 or 30%. In patients who survive an infarction for more than 48 hours, the expected fatality rate in the next year is approximately 15%. To detect a reduction of this mortality to 10%, 2300 patients would be needed in the trial. The analysis of several of the published trials has been made difficult by a lack of data, the failure on the part of authors and journal editors to agree on a common method of presenting results, and disagreement as to whether results should be analysed on an "all patients--intention to treat" basis or by considering only "clinical effectiveness" among patients who remained on treatment. Examples of these problems will be given, together with a suggested scheme for data presentation. The 95% confidence intervals of all the randomised and double blind studies of beta blockers after myocardial infarction show that by "intention to treat" analysis, none demonstrate a statistically significant reduction in mortality among treated patients. Only the practolol trial was of a reasonable size, but even here total mortality was not significantly reduced; the result is in any case of only theoretical interest. The trials showing a reduction of mortality in association with alprenolol treatment are not convincing, and the stratified trial design of one of these may have given misleading results because of the relatively small number of patients involved.