Rehabilitating digoxinThe opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.

The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology. This editorial refers to ‘Digoxin and reduction in mortality and hospitalization in heart failure: a comprehensive post hoc analysis of the DIG trial’† by A. Ahmed et al., on page 178 Digitalis glycosides were among the first cardiovascular therapeutic agents used in medicine, but despite descriptions over 200 years ago of their utility in the treatment of dropsy (heart failure), there was a lingering controversy concerning any clinical benefit. To resolve this uncertainty, a large simple randomized trial comparing digoxin to placebo was designed1 and executed2 in early 1990s. This National Heart, Lung, and Blood Institute and the Department of Veteran's Affairs sponsored trial was performed by the Digitalis Investigators’ Group (DIG).2 The DIG trial involved 302 clinical centres in the United States and Canada, randomized 7788 heart failure patients with normal sinus rhythm, and followed them for an average of 37 months. The main study involved 6800 patients with an ejection fraction <45%, whereas an ancillary arm included 988 patients with normal ejection fraction. The primary outcome was total mortality, and a limited number of secondary analyses including cardiovascular death, death from congestive heart failure, and hospitalizations were also examined. In the main trial, digoxin when compared with placebo had no affect on total mortality [risk ratio (RR) 0.99, 95% confidence interval (CI) 0.91–1.07]. This internecine result comprised a 1.6% absolute reduction with digoxin in mortality due to worsening heart failure (RR 0.88, 95% CI 0.77–1.01) but accompanied by a 1.9% absolute increase in presumed arrhythmic deaths (RR 1.14, 95% CI 1.01–1.30). There were 6% fewer hospitalizations overall among the digoxin group driven largely by the 7.9% absolute reduction in hospitalizations for worsening heart failure (RR 0.72, 95% CI 0.66–0.79). Results from the ancillary trial were generally consistent with those of the main trial. The DIG authors concluded that digoxin did not reduce overall mortality but did reduce the rate of hospitalization both overall and for worsening heart failure. Ahmed et al.3 report on a comprehensive post hoc re-analysis of the DIG trial that accounts for serum digoxin concentrations (SDCs). These authors studied 5549 of the original 7788 DIG patients, comprising the random sample of digoxin patients who had a measured 1 month SDC (1687) as well as the 3861 placebo patients alive at 1 month. As in the original study, the main outcomes reported are all-cause mortality and hospitalizations due to congestive heart failure. The authors have shown that compared with placebo, digoxin patients with a 1 month SDC of 0.5–0.9 ng/mL not only had a 38% relative (9.8% absolute) reduction (RR 0.62, 95% CI 0.54–0.72) in the need for CHF hospitalization, consistent with the original DIG publication findings, but also had a statistically significant 23% reduction in all-cause mortality (RR 0.77, 95% CI 0.67–0.89). This represents a 3.6% absolute reduction in mortality, which is of definite clinical importance, and contrasts with the original conclusions. Ahmed et al.3 showed that those with SDC >1.0 ng/mL had an increased crude mortality rate when compared with placebo but this difference disappeared (RR 1.06, 95% CI 0.93–1.20) after statistical adjustments for unbalanced confounders. Patients with SDC >1.0 ng/mL also had lower hospitalization rates. Importantly, after adjusting for SDC, Ahmed et al.3 did not notice any interaction between digoxin and gender. This suggests that an earlier publication4 showing an increased mortality for women in the digoxin group was perhaps a spurious result due to residual confounding from SDC. Does the present comprehensive post hoc re-analysis of the DIG trial finally remove any lingering doubts regarding a survival benefit with digoxin? Can the authors' conclusions that digoxin at low SDC reduces both mortality and morbidity be accepted? An earlier publication5 of this same database by different authors also observed increasing mortality in men with increasing SDC levels. This earlier article5 followed a similar protocol to the present study, but in addition to limiting the population to men considered three slightly differing groups (0.5–0.8, 0.9–1.1, and ≥1.2 ng/mL). Patients with an SDC of 0.5–0.8 mg/mL had the lowest mortality, a 6.3% (95% CI 2.1–10.5) absolute reduction compared with placebo. The middle group had no difference in mortality when compared with placebo, whereas those with an SDC ≥1.2 ng/mL had 11.8% (95% CI 5.7–18.0) increase in mortality. This earlier study concluded that higher SDCs in men were associated with increased mortality and suggested that the effectiveness of digoxin therapy would be optimized by use of the lower dose range. The present study reinforces and extends this earlier work5,6 by examining the effect of SDC on both men and women and by considering pertinent clinical interactions. Nevertheless, while reassuring to have confirmation by independent investigators, care must be taken in drawing definitive conclusions from these post hoc analyses. For example, we do not know how the authors elected to form the digoxin groups. Were the groups constructed a priori or only after seeing the data? Such post hoc analysis may give spurious associations as elegantly and convincingly demonstrated in the ISIS-27 clinical trial where no association was observed between aspirin and mortality following acute myocardial infarction for the 1400 patients with a Libra or Gemini Zodiac sign when compared with astrong benefit for the other 15 600 patients (P<0.000001). Moreover, as the hazard ratio for digoxin over time is unknown and because these analyses are based on 30 day survivors, the true relationship between mortality and digoxin even at low SDC remains unknown. Most importantly, patients have not been randomized to a specific digoxin concentration and physicians in the original study had the possibility of adjusting digoxin dose as clinically indicated. A substantial bias would be introduced if patients who are the most ill systematically received the highest digoxin doses. Support for this possibility comes from Table 1 of Ahmed's et al.3 where we notice that those with the highest SDC were generally older, had a longer duration of heart failure, and more Class III and Class IV symptoms. The rather large differences between the crude and adjusted values underline this unbalancing of covariates. Although the authors have taken care to adjust for these unbalances in their analysis, this lack of randomization certainly raises the possibility that other unmeasured, but equally important, prognostic variables might result in residual confounding and an over-estimation of any survival benefit for low dose SDC. Although a biased estimation of the survival benefit with low dose SDC cannot be excluded, we remain reassured by the original trial2 showing at worst a neutral impact on mortality and by the consistency in the reduction of hospitalizations. Despite the inability of this data to conclusively prove improved survival with low SDC, the temporal association, dose–response relationship, biological plausibility, and overall consistency with our general knowledge of digoxin toxicity lend credibility to this possibility. The authors are also to be congratulated for re-emphasizing the large morbidity benefit associated with digoxin, certainly at least as large as that seen for ACE-inhibitors in heart failure, and raising the tantalizing hypothesis of a survival benefit with low dose digoxin exposure. Perhaps, the spectre of a survival benefit will encourage a global re-examination of the role of digoxin in heart failure. Attaining high prescription rates for evidence-based therapies remain problematic for all heart failure treatment, but especially for digoxin which is currently prescribed at only 50% the rates of ACE-inhibitors.8,9 Interestingly, as the overall DIG study demonstrated unequivocally the overall safety of digoxin accompanied by substantial morbidity benefits, there has been a widespreadtrend in the last decade to remove digoxin from the heart failure therapeutic armamentarium.8 The exact reasons for this behaviour while unknown are probably multi-factorial, possibly relating to concerns about digoxin toxicity and the increasing polypharmacy of digoxin, diuretics, ACE-inhibitors and/or angiotensin receptor blockers, and most recently beta-blockers for heart failure. Could physicians’ prescribing patterns also have been influenced by the lack of extensive marketing for this low cost generic drug? Expert panels with high profile consultants in conjunction with intensive pharmaceutical detailing have been shown to successfully increase prescribing patterns. Conversely, might their absence negatively influence prescribing patterns? As an example, consider the routine prescription of ACE-inhibitors to vascular patients without left ventricular dysfunction. The EUROPA trial10 randomized over 12 000 patients to perindopril or placebo and demonstrated no mortality benefit and the need to treat 70 (absolute risk benefit 1.4%) patients for 4 years to avoid ahospitalization for a non-fatal myocardial infarction. Despite the PEACE11 trial's confirmation of no benefit in similar vascular patients, ACE-inhibitors have become a standard treatment in vascular patients without left ventricular dysfunction. The under prescribing of digoxin is problematic on several grounds. First, this ignores, or at least minimizes, the very substantial benefit in morbidity seen with digoxin (number needed to treat for 3 years to avoid one costly heart failure hospitalization=10–12). Moreover, the large randomized trials showing the benefits of ACE-inhibitors and beta-blockers have generally been performed with a background therapy of digoxin. Although these medications may most likely exert their benefits independently of any interaction with digoxin, this is not known with certainty. The precise economic consequences of digoxin therapy will require detailed local analyses. However, given its low cost, a fraction of that of ACE-inhibitors, and potential large health benefits, a back of the envelope calculation suggests a favourable cost-effectivenessratio for both developed and developing countries.12 The paper of Ahmed et al. will hopefully contribute to the rehabilitation of digoxin in the treatment of heart failure. Conflict of interest: none declared. doi:10.1093/eurheartj/ehi687

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.