Digoxin?Doubts to dampen enthusiasm

Abstract

Dear Sir: Controversy concerning the use of digoxin in patients with heart failure in sinus rhythm has existed for almost the whole of the present century [1-4]. The debate is not finished. Several recent withdrawal studies [5-7] have revived enthusiasm for the use of digoxin, and the results of the large mortality study currently being undertaken by the National Institute of Health in America are awaited. The article by F.I. Marcus [8] in this journal puts forward an optimistic view. He believes that several clinical trials have shown unequivocally that digoxin decreases symptoms of cardiac failure and declares that the major unresolved question concerning digoxin is safety. I beg to differ. The first issue is to identify what are the key clinical questions concerning the use of digoxin. No physician doubts the value of digoxin in controlling the fast heart rate of patients with atrial fibrillation with or without heart failure. The question is not whether digoxin in patients with heart failure and in sinus rhythm is positively inotropic either acutely or chronically. Nor is the question whether digoxin in selected groups of patients under particular circumstances can be shown to be beneficial for heart failure. Such an argument could be used to justify the continual use of leaches. The key question for the clinician is whether, after a patient has received optimal t reatment with ACE inhibitors and a combination of diuretics, the addition of digoxin confers further advantage. The reason that this question is key is that the current initial t reatment for overt heart failure is the careful and thoughtful use of [9] an ACE inhibitor combined with a diuretic. The results from the study by Lee et al. in 1982 showed that patients with fluid overload and a third heart sound benefitted from the use of digoxin. That is to be expected, but exactly the same result might have been obtained by the propitious use of an ACE inhibitor and a diuretic. Indeed the clinical and hemodynamic benefit [10] in patients is probably related to change in total body weight [11]. There may be subgroups of patients in whom digoxin is particularly useful, but those subgroups need to be identified in carefully controlled clinical trials. Physicians may well have their own suspicions of who those particular groups might be. There have been a very large number of studies on the use of digoxin in the last century [3,4,12,13]. Most have been withdrawal or observational studies. Some included patients in atrial fibrillation. Very few of these studies would fulfill current criteria for a proper clinical trial. The historical data have been reviewed. One meta-analysis [4] and one overview have been published recently [3]. The meta-analysis found seven trials in the world literature that fulfill the criteria for a modern study but only six of these were included in the overview. The one additional study [5] was rejected from the overview because it was a withdrawal study and did not fulfill the appropriate criteria. The meta-analysis came to the conclusion that there was evidence that the number of patients in the control group who were withdrawn from studies was statistically different from those withdrawn in the group on digoxin. The overview was unable to show any benefit to patients in terms of symptoms or exercise capacity. The numbers are far too small to show any benefit in terms of mortality. These data do not lie easily with the claim by Dr. Marcus that digoxin unequivocally decreases symptoms of heart failure. Furthermore, withdrawal is not a rigorous or robust endpoint for a trial, nor does it answer the question as to whether the patients would be more appropriately treated with diuretics and ACE inhibitors. In one recent trial, withdrawal gave an advantage to patients because those patients were not then exposed to the risk of continuing with the drug [14]. Recently three withdrawal studies have been reported. One has been published as a full paper [5] and the other two as abstracts [6,7]. The problem with this approach is that it does not answer the clinical