Abstract
6 min readSince the DIG study,1 which demonstrated that whilst digoxin reduced hospitalization for heart failure (HF), there was no significant overall effect on mortality, the place of digoxin in treating HF has steadily declined over recent years and has been debated.2 Selection of heart rate control therapy in patients with permanent atrial fibrillation (AF), particularly those with coexisting HF, is still an issue.3 In the recent RATE-AF trial that included patients with permanent AF and symptoms of HF, treatment with digoxin or bisoprolol did not result in a significant difference in quality of life at 6 months.4 We analysed whether the clinical outcomes differed among such patients with permanent AF and HF who are treated with either digoxin or beta-blocker in a ‘real-life’ observational study. All patients who were treated for AF in our academic institution were identified using a local database. The information for each patient was extracted from computerized data of hospitalization and consultation of our institution. We examined the clinical course of 8962 consecutive patients with AF seen over a 10-year period (2000–2010).5, 6 Among them, 1787 patients met the inclusion criteria of RATE-AF (permanent AF, age ≥60 years and New York Heart Association functional class ≥2), of whom 512 patients (29%) were treated with beta-blocker alone, 425 (24%) were treated with digoxin alone and 237 (13%) were treated with both a beta-blocker and digoxin. Adverse outcomes were investigated during follow-up and we identified the causes of death. Owing to the non-randomized nature of the study, and considering for significant differences in baseline characteristics, propensity-score matching was used to control for potential confounders of the treatment outcome relationship considering the covariates listed in Table 1. Outcomes in patients treated with beta-blocker alone or digoxin alone were compared after 1:1 propensity-score matching using the one-to-one nearest neighbour method (with a calliper of 0.1 of the standard deviation of the propensity score on the logit scale). Proportional hazard model was used for the outcome analyses and proportional hazard assumption was checked by plotting the log–log Kaplan–Meier curves and scaled Schoenfeld residuals against time plots. After propensity score matching, 270 patients treated with beta-blocker (bisoprolol 70%, carvedilol 17%, nebivolol 11%, metoprolol 2%) were matched with 270 patients treated with digoxin (dose range, 125–250 µg/day; mean dose, 194 µg/day) (Table 1). The mean follow-up period was 2.2 (±2.7) years [median 1.1 (interquartile range 0.1–3.5) years]. In total, 125 all-cause deaths (including 72 cardiovascular deaths) were recorded, corresponding to an annual death rate of 10.4%. Major clinical events (all-cause death, myocardial infarction, ischaemic stroke, or major bleeding) were recorded in 192 patients, corresponding to an annual rate of 19.1%. Risk was not statistically different in the two matched groups (beta-blocker vs. digoxin) for all-cause death [hazard ratio (HR) 0.95, 95% confidence interval (CI) 0.67–1.35], cardiovascular death (HR 1.23, 95% CI 0.77–1.96) or major clinical events (HR 0.98, 95% CI 0.74–1.31) (Figure 1 and online supplementary Figure S1). There was no statistical interaction with year of inclusion (2005–2010 vs. 2000–2004). We found that among patients with permanent AF and symptoms of HF, there were no significant differences in the risk of all-cause death, cardiovascular mortality and major clinical events between those treated with digoxin or beta-blocker. The strengths of our study are that it was based on a cohort of fair size, the population studied was representative of the majority of unselected patients with AF and we were able to take into account the severity of HF. Although not being a randomized study, our analysis included more patients and had a longer follow-up than in the RATE-AF trial,4 resulting in a 10-fold higher number of clinical events. Some secondary/exploratory outcomes differed between the two arms in RATE-AF, including several cardiovascular outcomes and adverse events that were in favour of digoxin. This was not the case in the present study and in this regard, the results of the two studies are not fully convergent. There remain significant knowledge gaps about how digitalis works and how it should be used in the modern treatment of AF. Since the DIG study,1 which demonstrated that whilst digoxin reduced HF hospitalization there was no significant overall effect on mortality, the place of digoxin in treating HF has steadily declined over recent years. In one of the very few trials in this area, the combination of carvedilol and digoxin appears generally superior to either carvedilol or digoxin alone in the management of AF in patients with HF.7 An increased mortality among digoxin-treated patients has been found in some observational studies in the last two decades.8 However, there was no adjustment to severity of HF in most of these studies. Results in other cohorts of AF patients found a neutral association of digoxin with outcomes leading to renewed questions.9 We do not intend to compare our study to a randomized clinical trial. It was based on an observational cohort design and subject to the risk of residual confounders. This has to be considered since digoxin is often used after beta-blockers have failed. The use of propensity matching may not be able to account for such inherent bias in selection of patients. We were not able to evaluate baseline and achieved heart rate in our analysis. Also, serum concentrations of digoxin are a possible determinant of adverse clinical events. Even if we were not able to adjust for serum drug concentration, unavailable in our data sources, measures of renal function (with matched adjustment) and dose of digoxin treatment were collected, which is an added value in comparison with previous studies. Finally, considering the putative mechanisms of either improved or worsened outcome with either beta-blocker or digoxin and the relatively wide CIs in our results, a trial or a propensity analysis with hard endpoint might require a quite larger sample size than that in the present study. While an increase in sudden death may be a concern with digoxin, we acknowledge that we have an insufficient sample size to prove any difference in outcome for sudden death. In conclusion, among patients with permanent AF and symptoms of HF, there was no significant difference in the risk of all-cause death, cardiovascular mortality and major clinical events between those treated with digoxin or beta-blocker when differences in patient characteristics were accounted for. Concerns regarding the use of digoxin, such as the narrow therapeutic window and drug interactions, were not issues resulting in worse clinically relevant cardiovascular outcomes with the approach used in the current study for HF patients with permanent AF. Conflict of interest: G.Y.H.L.: consultant and speaker for BMS/Pfizer, Boehringer Ingelheim and Daiichi-Sankyo. No fees are received personally. L.F.: consultant or speaker fees of small amounts for Bayer, BMS/Pfizer, Boehringer Ingelheim, Medtronic and Novartis outside of this work. All other authors have nothing to disclose. Figure S1. Cumulative incidence of all-cause death (top left panel), cardiovascular death (top right panel), stroke/thromboembolic events (middle left panel), acute coronary syndrome (middle right panel) and major clinical events (all-cause death, acute coronary syndrome, need for revascularization, ischaemic stroke/thromboembolic event or major bleeding, lower panel) in matched patients with permanent atrial fibrillation and heart failure treated with digoxin or beta-blocker during follow-up of 2.5 ± 3.0 years (median 1.2, interquartile range 0.3–4.4 years). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Discussion(0)
No comments yet. Be the first to comment.