Effect of digitalis treatment on survival after acute myocardial infarction

To determine whether treatment with digitalis is associated with decreased survival after acute myocardial infarction (AMI), data from 504 patients who were enrolled in a postinfarction natural history study were analyzed. At the time of discharge, 229 patients (45%) were taking digitalis. After 3 years of follow-up, the cumulative survival rate for patients discharged on a regimen of digitalis was 66%, compared with 87% for those not treated (p < 0.001). Univariate analysis showed that statistically significant differences existed between the 2 groups with respect to age, previous AMI, left ventricular failure in the coronary care unit, atrial fibrillation in the coronary care unit, peak creatine kinase levels, enlarged heart and pulmonary vascular congestion on the discharge chest x-ray, ventricular arrhythmias and treatment with diuretic, antiarrhythmic and β-blocking drugs. Survival analysis using Cox's regression model showed that the association between digitalis and decreased survival was of borderline significance after adjustment for atrial fibrillation and left ventricular failure. Serum digoxin concentration was measured in 83% of the patients who took digitalis. Survival was inversely and significantly related to serum digoxin, i.e., the higher the serum digoxin concentration, the lower the long-term survival rate. After adjusting for atrial fibrillation and left ventricular failure, serum digoxin was not significantly related to survival. Taken together with the results of 3 other large, nonrandomized studies of digitalis treatment after AMI, this study suggests that digitalis treatment may have adverse effects on survival during follow-up. Until this question is definitively answered by controlled, randomized clinical studies, clinicians should ask themselves, in each case, whether treatment is really needed for left ventricular dysfunction after myocardial infarction and, if so, which treatment has the best risk-benefit ratio.