The effect of digoxin on survival in patients with heart failure: an answer
at last?
J. R. Hampton, Division of Cardiovascular Medicine, Universital Hospital Nottingham,
Queen´s Medical Centre, Nottingham
William Withering described the value of an extract of the foxglove for 'the cure of
dropsy' in 1785, and for about 150 years digitalis was the only effective remedy for
treating heart failure. With the introduction of the mercurial and then the thiazide
diuretics, digitalis became considered to be less important, perhaps especially in the
United Kingdom, and there were doubts whether it had any useful effect other than the
control of the ventricular rate in patients with atrial fibrillation. A long series of
studies - mainly involving the withdrawal of digoxin from treated patients - left doubt
whether digoxin improved exercise time in patients with heart failure, and eventually
evidence accumulated suggesting that it probably does not have any useful effect . There
has been fairly convincing evidence from clinical trials (the Randomized Assessment of
Digoxin on Inhibitors of the Angiotensin Converting Enzyme, RADIANCE , and the Prospective
Randomized Study of Ventricular Failure and the Efficacy of Digoxin, PROVED ) that
worsening heart failure and hospitalization occur less often in patients treated with
digoxin, but the main concern about digoxin has been that it might actually increase
fatality in patients with heart failure.
There has for many years been speculation that treatment with digoxin after myocardial
infarction is associated with an increased risk of death. It has always been difficult to
determine whether this was simply a reflection of the fact that patients with heart
failure after myocardial infarction are at higher risk, but at least one study has
identified digoxin therapy as having an independent deleterious effect. Furthermore,
treatment with a range of unrelated compounds known to increase cardiac performance and to
improve heart failure symptoms has been shown to be associated with an increase in
fatality: the list of such treatments includes xamoterol , milrinone , enoximone ,
flosequinan and ibopamine . A trial specifically designed to study the effect of digoxin
on mortality in patients with heart failure but who were in sinus rhythm was therefore of
extreme importance, and the results of what is usually referred as the DIG Trial have now
been published .
In the main part of this trial 6800 patients with clinically-diagnosed heart failure and
who had a left ventricular ejection fraction (LVEF) of less than 45% were randomly
allocated on a double blind basis to treatment with digoxin or placebo. In a smaller
parallel study, patients with heart failure but who had a higher ejection fraction were
similarly randomised. The primary outcome was mortality, and in the main part of the trial
there were during an average follow-up of 37 months 1181 deaths (34.8%) among the patients
with digoxin and 1194 deaths (35.1%) among those allocated to placebo treatment. The
effect of digoxin was therefore neutral so far as the main trial endpoint was concerned.
However, the trial outcomes that were defined as 'secondary' are of equal interest. These
outcomes included death or hospitalization from worsening heart failure, and both of these
were reduced by digoxin treatment.
In the digoxin group 394 deaths were attributed to worsening heart failure, compared to
449 in the placebo group (risk ratio 0.88, 95% confidence interval 0.77 - 1.01, p = 0.06).
There were, however, a few more deaths from all cardiovascular causes in the digoxin group
(1016 deaths, compared with 1004 in the placebo group) and the two treatment groups had
nearly identical outcomes in total mortality because there were more deaths due to cardiac
causes other than heart failure in the digoxin group (508 compared with 444 in the placebo
group). These non-heart failure deaths, which were not described in detail, included those
due to sudden death, arrhythmias, myocardial infarction, and cardiac surgery.
Significantly fewer patients in the digoxin group (910) than in the placebo group (1180)
were admitted to hospital because of heart failure (risk ratio 0.72, 95% confidence
interval 0.66 - 0.75, p less than 0.001).
The results of the DIG trial can therefore be interpreted in different ways, and the
effect of the trials on clinical practice has to depend on this interpretation. Strictly
speaking, all the trial showed (from its primary endpoint) was that digoxin had no effect
on mortality in heart failure. However, it is not unreasonable to go further and accept
that digoxin reduced deaths and hospital admissions due to heart failure itself; but if we
accept this we must also accept that digoxin has an unwanted effect in increasing cardiac
deaths involving mechanisms other than heart failure. Since hospitalization due to heart
failure implies an increased severity of symptoms, most patients with heart failure would
probably prefer to be treated with digoxin, knowing that their symptoms should be improved
even though their mode of death (but not their likelihood of death) might well be changed.
The DIG trial does not help us to decide whether or not digoxin should remain the
treatment of choice in patients with atrial fibrillation, for such patients were excluded
from the study. There are alternative ways of controlling the ventricular rate in atrial
fibrillation (including drugs such as verapamil, diltiazem, amiodarone etc.) but how these
might compare with digoxin in their effect on survival is completely unknown.
