Correspondence
Digoxin for the Treatment of Heart Failure
N Engl J Med 2003; 348:661-663February 13, 2003
- Article
To the Editor:
The report by Rathore et al. (Oct. 31 issue)1 emphasizes differences according to sex in the outcomes of digoxin therapy in patients with congestive heart failure. Their post hoc analysis showed that outcomes differed between men and women. With respect to the characteristics of the patients, there were striking differences between men and women in the causes of congestive heart failure. Women more often had cardiomegaly, severe heart failure, diabetes, and idiopathic heart failure; men more often had ischemic heart failure and heart failure due to prior myocardial infarction. Digoxin may be a useful treatment for some causes of congestive heart failure but may be harmful for others. Differences in the causes of heart failure would be better candidates than sex to explain the observed differences in outcomes.
1 ReferencesSophie Moulias, M.D.
Fabien Tigoulet, M.D.
Sylvie Meaume, M.D.
Hôpital Charles Foix, 94200 Ivry/Seine, France
sophie.moulias@cfx. ap-hop-paris. fr 1
Rathore SS ,Wang Y ,Krumholz HM . Sex-based differences in the effect of digoxin for the treatment of heart failure. N Engl J Med 2002;347:1403-1411
Full Text | Web of Science | Medline
To the Editor:
Rathore et al. suggest that women with heart failure who are treated with digoxin have a greater risk of death than women who are not treated with digoxin. No difference in mortality was found among the men in the study. We wish to address two issues. First, did the authors adjust for the right predictors? In the study, women were significantly older and were more likely to be in New York Heart Association functional class III or IV or to have diabetes than men. These factors have been shown to predict mortality in heart failure. We would thus expect an increased mortality among women. In fact, the Kaplan–Meier estimates show that their overall mortality was lower than that of men. More than 50 variables have been shown to be predictors of mortality in patients with heart failure.1 Most of them, however, including the use of beta-blockers, were not adjusted for in the current analysis. Therefore, we believe that there may be some residual confounding in the study, although its effect on the results is difficult to assess.
Second, the results would be more convincing if there were greater biologic plausibility, such as a dose effect. Unfortunately, the serum levels of digoxin are not reported. It may be that women also benefit from digoxin but are more susceptible than men to toxic effects, perhaps because of their smaller body size, or that women have a therapeutic window that differs from that of men.
The implications of this study for women are serious. Alternative explanations for the findings must be considered before we decide to withhold this medication from them.
1 ReferencesMarie Hudson, M.D.
Louise Pilote, M.D., M.P.H., Ph.D.
Montreal General Hospital, Montreal, QC H3G 1A4, Canada
mhudso@po-box.mcgill. ca To the Editor:
Rathore et al. found that digoxin therapy is associated with an increased risk of death among women but not among men. The trial had insufficient statistical power to test whether the interaction between sex and digoxin therapy was dependent on sex-based differences in serum digoxin levels. Was the digoxin dose too high for women, as Eichhorn and Gheorghiade suggest in their Perspective?1
We investigated the serum levels of digoxin in patients undergoing therapeutic-drug monitoring at our hospital over the past three years (Table 1Table 1
Sex Differences in Serum Digoxin Levels and Adverse Drug Reactions.). Patients were included if a steady-state level had been reached, if trough serum values were available, and if information on concomitant treatment and serum creatinine was available. Interestingly, the mean serum digoxin level was significantly higher in women than in men. Adjustment for age, dose, and serum creatinine had only a minor effect on the estimates.We next searched the Swedish national register of adverse drug reactions3 to see whether adverse reactions to digoxin are more common among women than among men. The results are shown in Table 1. Evidently, these sex differences cannot be explained by an interaction between digoxin and progestin in combined hormone-replacement therapy, as suggested by Rathore et al., nor can they be explained by differences in digoxin prescriptions.2
We conclude that digoxin therapy in women should be administered with heightened attention to the appropriate dose.
3 ReferencesPär Hallberg, M.D.
Karl Michaëlsson, M.D., Ph.D.
Håkan Melhus, M.D., Ph.D.
Uppsala University Hospital, S-751 85 Uppsala, Sweden
hakan.melhus@medsci. uu. se 1
Eichhorn EJ ,Gheorghiade M . Digoxin -- new perspective on an old drug. N Engl J Med 2002;347:1394-1395
Full Text | Web of Science | Medline2
Drug statistics Sweden. Stockholm, Sweden: Apoteket AB, 2003. (Accessed January 24, 2003, at http://www.apoteket.se/.)
3
Swedish Drug Information System. Uppsala, Sweden: Swedish Medical Products Agency, 2003. (Accessed January 24, 2003, at http://www.mpa.se/.)
Author/Editor Response
Moulias et al. and Hudson and Pilote contend that our finding of a digoxin-associated increased risk of death among women enrolled in the Digitalis Investigation Group trial may reflect the more clinically severe condition of women enrolled in the trial, rather than an interaction between sex and digoxin. Our analysis considered a wide range of potentially confounding factors (any characteristic associated with mortality at P≤0.50 on univariate analysis), suggesting that residual confounding is an unlikely possibility. Furthermore, we found no differences between women randomly assigned to digoxin and women randomly assigned to placebo (P>0.05 for all comparisons in Table 1 of our article), as would be expected in a subgroup analysis in a randomized, controlled trial. The rate of adverse reactions to digoxin presented by Hallberg et al. is markedly low (approximately 1 event per 100,000 daily doses), raising questions about the validity of their data for evaluating the combination of digoxin and hormone-replacement therapy.
Hallberg et al. and Hudson and Pilote, as well as Eichhorn and Gheorghiade, suggest that sex differences in serum digoxin levels explain the increased risk of death among women randomly assigned to digoxin. Because the serum digoxin level was assessed one month after randomization in only 330 women, the Digitalis Investigation Group trial lacked sufficient statistical power to assess the interaction between sex and digoxin among patients in whom serum digoxin levels were measured. The interaction persisted after we accounted for factors associated with the serum digoxin level (e.g., renal function) and also after we assessed reports of digoxin toxicity. In addition, the difference in the average serum digoxin level between men and women was small at 1 month (0.9 ng per milliliter in women vs. 0.8 ng per milliliter in men, P=0.007), and no difference was observed at 12 months. Yet women randomly assigned to receive digoxin continued to have higher mortality rates than women randomly assigned to receive placebo over the full duration of the trial, suggesting that a factor other than the serum digoxin level may have a role. Together, these observations support the hypothesis that digoxin may be associated with an increased risk of death in women.
Ultimately, it is up to women and their physicians to determine, on the basis of currently available evidence, the clinical value of digoxin by deciding whether the sole demonstrated benefit of this drug — a small decrease in the risk of hospitalization1 — is worth a possible digoxin-associated increased risk of death.
1 ReferencesSaif S. Rathore, M.P.H.
Harlan M. Krumholz, M.D., S.M.
Yale University School of Medicine, New Haven, CT 06520-80251
The Digitalis Investigation Group
Full Text | Web of Science | Medline
- Citing Articles (1)
Citing Articles
1
María G. Crespo Leiro, María J. Paniagua Martín. (2006) Insuficiencia cardiaca. ¿Son diferentes las mujeres?. Revista Española de Cardiología 59:7, 725-735
CrossRef
