Abstract
Objective:
The use of digoxin in patients having atrial fibrillation (AF) with or without heart failure (HF) is not without controversy. The aim of this study was to examine the impact of digoxin therapy on mortality stratified by HF.
Methods:
Gulf Survey of Atrial Fibrillation Events was a prospective, multinational, observational registry of consecutive patients with AF recruited from the emergency department of 23 hospitals in 6 countries in the Middle East. Patients were recruited between October 2009 and June 2010 and followed up for 1 year after enrollment. Analyses were performed using univariate and multivariate statistical techniques.
Results:
The study included a total of 1962 patients with AF, with an overall mean age of 56 ± 16 years, and 52% (n = 1026) were males. At hospital discharge, digoxin was prescribed in 36% (n = 709) of the patients, whereas HF was present in 27% (n = 528) of the cohort. A total of 225 (12.1%) patients died during the 12-month follow-up period after discharge (5.3% [n = 104] were lost to follow-up). Patients with HF were consistently associated with higher mortality at 1 month (5.1% vs 2.1%; P < .001), 6 months (17.2% vs 5.0%; P < 0.001), and 12 months (24.3% vs 7.6%; P < .001) when compared to those without HF. When stratified by HF, digoxin therapy was associated with significantly higher mortality in those without HF at 6 months (8.7% vs 3.7%; adjusted odds ratio (aOR), 5.07; P < .001) and 12 months (12.3% vs 6.0%; aOR, 4.22; P < .001) but not in those with HF (6 months: 18.6% vs 14.7%; aOR, 1.62; P = .177 and 12 months: 25.4% vs 22.4%; aOR, 1.37; P = .317).
Conclusions:
In patients with AF and HF, digoxin did not offer any survival advantages. However, in those without HF, digoxin therapy was, in fact, associated with significantly higher long-term mortality.
Introduction
With the introduction of more evidence-based and effective medications as well as interventional procedures for atrial fibrillation (AF) and heart failure (HF) management, the use of digoxin has declined over the last 2 decades. 1 –4 In the American College of Cardiology/the American Heart Association/ the Heart Rhythm Society (ACC/AHA/HRS) guidelines for rate control in AF, digoxin, although mentioned in the text, is not given any class recommendation, but in HF, in the absence of preexcitation, intravenous digoxin is recommended to control heart rate acutely (class IB) and chronically, and digoxin is effective to control resting heart rate in patients with HF having reduced ejection fraction (EF; class IC). 1 In addition, a combination of digoxin and a β-blocker (or a nondihydropyridine calcium channel antagonist for patients with HF having preserved EF) is reasonable to control resting and exercise heart rate in patients with AF (class IIA). 1 The European Society of Cardiology (ESC) AF guidelines recommend digoxin as third choice after β-blockers and calcium antagonists for rate control in patients with AF either alone or in combination (class IB). 2 In HF, the ESC guidelines recommend intravenous digoxin (class IB), and for chronic rate control, they recommend digoxin in patients with HF, left ventricular (LV) dysfunction, and sedentary lifestyle (class IIA). 2
However, it has been observed that in practice, digoxin is either overused or used inappropriately in patients with AF, especially in the elderly population, consequently leading to digoxin toxicity. 5 Of late, there are conflicting reports about digoxin use in patients having AF with or without HF with regard to morbidity and mortality. A number of studies have reported an increased mortality and morbidity associated with digoxin use in AF, 6 –11 whereas other studies have showed no increase in cardiac events with digoxin use in AF. 12 –15 To our knowledge, there are currently no studies that have evaluated the impact of digoxin on mortality in patients having AF with or without HF in the Middle East. In view of increasing safety concerns about digoxin use in recent reports, more evidence is warranted on its use in AF before any policy change. The aim of the current study was to evaluate the impact of digoxin therapy on mortality stratified by HF in 6 Middle Eastern countries.
Methods
Gulf Survey of Atrial Fibrillation Events (Gulf SAFE) was a prospective, multinational, observational study of consecutive patients with AF recruited and admitted to hospital from emergency department (ED) of 23 hospitals in 6 countries in the Middle East. Patients were recruited between October 2009 and June 2010 and followed up for 1 year after enrollment. Full details of the methods and definitions of data variables have already been published elsewhere. 16,17 Briefly, patients were eligible if they were ≥18 years, found to have AF lasting more than 30 seconds on a 12-lead electrocardiogram or rhythm strip while in ED, and provided informed consent. Patients were enrolled regardless of the reason for their ED visit. All management decisions were left to the discretion of the treating physician. The study complies with the Declaration of Helsinki, and the protocol was approved by the ethics committees of each participating institution in each of the 6 countries.
Data were collected on a standardized case report form (CRF) and entered online. Data collected included patient demographics; medical/surgical history; AF history including type of AF; ED investigations and management; whether by rate control or rhythm control; outcome of ED visit; and medications before, during, and at discharge. Medications included but not limited to diuretic, angiotensin-converting enzyme inhibitor (ACEI), angiotensin receptor blocker, β-blocker, statin, aspirin, clopidogrel, and warfarin. Data collected were in accordance with the ACC/AHA key data elements and definitions for AF. 18 Variables required for Congestive heart failure; Hypertension; Age >75 years; Diabetes mellitus; and Stroke/transient ischemic attack (CHADS2) score calculation (previous stroke/transient ischemic attack [TIA], age, hypertension, diabetes mellitus, and HF) were also included in the CRF. Heart failure was defined according to the Framingham criteria as mentioned in the ACC/AHA data variables. 18 Systolic LV dysfunction was defined as EF < 40% by transthoracic echocardiography. 18 The EF data were collected from patient charts if echocardiogram was performed within the past 6 months or done during the index admission. Data on mortality were collected during both hospital stay and follow-up (1 month, 6 months, and 1 year). One, 6, and 12 months of follow-up occurred via clinic visit or telephone interview.
Statistical Analysis
Patient characteristics were summarized using frequencies and percentages for categorical variables and means and standard deviations or medians and interquartile ranges for continuous variables, wherever appropriate. Group comparisons for categorical variables were performed using Pearson χ2 test, whereas for continuous variables, Student t test or Wilcoxon–Mann–Whitney test was used, wherever appropriate. The impact of digoxin therapy on all-cause mortality (1 month, 6 months, and 1 year) was evaluated using multivariable logistic regression utilizing stepwise-backward elimination method. Confounding variables included those variables (Table 1) that were correlated with mortality (at P < .1). The covariates were age, gender, body mass index (BMI), LV systolic dysfunction, hypertension, diabetes mellitus, coronary artery disease (CAD), chronic obstructive pulmonary disease (COPD), prior stroke/TIA, peripheral vascular disease (PVD), serum creatinine, medications at discharge (diuretic, β-blocker, statin, aspirin, clopidogrel, warfarin), CHADS risk score, and AF type.
Demographic, Clinical, Medications, and Mortality Characteristics of the Study Cohort Stratified by Digoxin Therapy at Discharge (N = 1962).a
Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; BMI, body mass index (was missing in 6% [n = 109] of the patients); CHADS, Congestive heart failure; Hypertension; Age >75 years; Diabetes mellitus; and Stroke/transient ischemic attack; COPD, chronic obstructive pulmonary disease; LV, left ventricular; SD, standard deviation; TIA, transient ischemic attack.
aSerum creatinine was missing in 499 patients (25%). Analyses were conducted using Student t test, Wilcoxon–Mann–Whitney test, or Pearson χ2 tests, whenever appropriate.
bMortality numbers and percentages do not add to 100% due to losses at follow-up, which reached a total of 3.4% (n = 67; N = 1895) at 1 month, 4.4% (n = 86; N = 1876) at 6 months, and 5.3% (n = 104; N = 1858) at 1 year.
The goodness of fit of the logistic model was examined using the Hosmer and Lemeshow goodness-of-fit statistic. The discriminatory power of the logistic model was assessed by the area under the receiver operating characteristic curve, also known as the C-index. An a priori 2-tailed level of significance was set at the .05 level. Statistical analysis was conducted using STATA version 13.1 (STATA Corporation, College Station, Texas, USA).
Results
From a total of 2043 patients with AF enrolled in the Gulf SAFE registry, 81 (3.96%) patients died either in the ED (n = 2) or during admission (n = 79). The remaining 1962 patients, who were discharged alive from the hospital, represent the sample size for this study. Demographic and clinical characteristics of the cohort are shown in Table 1. The overall mean age of the cohort was 56 ± 16 years, with 52% (n = 1026) being males. At discharge, digoxin was prescribed in 36% (n = 709) of the patients, whereas HF was present in 27% (n = 528) of the cohort. The digoxin cohort was more likely to be associated with HF than the group without digoxin (47% vs 15%; P < .001).
The digoxin group was associated more with females (56% vs 43%; P < .001), LV systolic dysfunction (32% vs 10%; P < .001), CAD (33% vs 26%; P = .001), rheumatic heart disease (32% vs 6%; P < .001), prior stroke/TIA (14% vs 10%; P = .016), COPD (6.5% vs 4.3%; P = .035), PVD (3.8% vs 1.4%; P = .001), lower BMI (27 vs 29 kg/m2; P < .001), higher serum creatinine (86 vs 80 mmol/L; P < .001), higher CHADS scores of ≥ 2 (49% vs 40%; P < .001), and permanent AF type (54% vs 22%; P < .001). However, the group that was not on digoxin at discharge was associated more with “first AF attack” (44% vs 25%; P < .001) and persistent AF (23% vs 9%; P < .001). With regard to medications at discharge, the digoxin group was associated with the use of diuretics (82% vs 41%; P < .001), ACEIs (44% vs 35%; P = .002), and warfarin (75% vs 49%; P < .001), whereas the cohort that was not on digoxin was more likely to be prescribed β-blockers (61% vs 45%; P < .001), statins (50% vs 38%; P < .001), aspirin (56% vs 47%; P = .001), and clopidogrel (12% vs 8%; P = .037).
A total of 5.3% (n = 104) of the patients were lost to follow-up during the 12-month follow-up period after discharge. There were a total of 2.9% (n = 55), 8.3% (n = 155), and 12.1% (n = 225) of the patients who died cumulatively during the 1-month, 6-month, and 12-month follow-up periods, respectively (Table 1). As shown in Table 2, patients with HF were consistently associated with higher mortality at 1 month (5.1% vs 2.1%; P < .001), 6 months (17.2% vs 5.0%; P < .001), and 12 months (24.3% vs 7.6%; P < .001) when compared to those patients having AF without HF. However, adjusting for other factors in the stepwise-backward multivariable logistic model, digoxin therapy was associated with significantly higher mortality in those without HF at 6 months (8.7% vs 3.7%; adjusted odds ratio [aOR], 5.07; P < .001) and 12 months (12.3% vs 6.0%; aOR, 4.22; P < .001) but not in those with HF (6 months: 18.6% vs 14.7%; aOR, 1.62; P = .177 and 12 months: 25.4% vs 22.4%; aOR, 1.37; P = .317).
Impact of Digoxin Prescribing at Hospital Discharge on 1 Month, 6 Months, and 1 Year Mortalities of the Study Cohort Stratified by HF.a
Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; CHF, chronic heart failure; HF, heart failure; HL, Hosmer and Lemeshow P value; ROC, area under the receiver operating curve also known as C-statistic.
aPercents are row percentages. Multivariate analyses were conducted using logistic regression model utilizing stepwise-backward elimination method adjusting for age, gender, body mass index, left ventricular systolic dysfunction, hypertension, diabetes mellitus, coronary artery disease, chronic obstructive pulmonary disease, prior stroke/transient ischemic attack, peripheral vascular disease, serum creatinine, medications at discharge (diuretic, β-blocker, statin, aspirin, clopidogrel, and warfarin), Congestive heart failure; Hypertension; Age >75 years; Diabetes mellitus; Stroke/transient ischemic attack (CHADS) risk score, and atrial fibrillation type. From an original discharged cohort of 1962, losses to follow-up reached a total of 3.4% (n = 67; N = 1895) at 1 month, 4.4% (n = 86; N = 1876) at 6 months, and 5.3% (n = 104; N = 1858) at 1 year.
Discussion
In this prospective, multinational, multicenter study from the Middle East, the predominant finding is that in patients with AF without HF, digoxin therapy was associated with significantly higher long-term mortality compared to those who were not on digoxin therapy. In addition, in patients with AF and HF, digoxin did not offer any survival advantages.
In a large recent study of 122 465 newly diagnosed patients with AF in the United States, with a mean age of 72.1 years, 23.4% received digoxin. 9 In comparison, in the present study, the overall mean age of patients with AF was 56 years with 36% of patients receiving digoxin at discharge, which is relatively high. In a subanalysis of Stroke Prevention using an ORal Thrombin Inhibitor in atrial Fibrillation [SPORTIF] III and V trials, among patients with AF receiving digoxin, a mortality of 4.22%/patient-year compared to 2.66%/patient-year in those who did not received digoxin with an adjusted hazard ratio (HR) of 1.53 was observed. 19 Similar HRs of 1.42 and 1.41 were observed in digoxin-treated patients in the Registry of Information and Knowledge about Swedish Heart Intensive care Admissions (RIKS-HIA) and AF Follow-Up Investigation of Rhythm Management (AFFIRM) studies. 6,7
The Stockholm Cohort of Atrial Fibrillation (SCAF) trial and another smaller study showed that digoxin was mainly given to an elderly frail population with other comorbidities, and after adjusting for these confounders, digoxin use was neutral for long-term mortality and major cardiac events in patients with AF. 12,13 In the post hoc propensity-matched analysis of the AFFIRM trial, Gheorghiade et al reported that there was no significant increase in mortality in patients with AF treated with digoxin treatment irrespective of the presence or absence of HF (HR, 1.06). 14 However, in the AFFIRM study by Whitbeck et al, digoxin was treated as a time-dependent covariate, considering changes in digoxin use over time, but in the AFFIRM study by Gheorghiade et al, digoxin use was assessed at a fixed time point of initial therapy at randomization similar to our registry. 7,14 But, our prospective registry, even though analyzed digoxin use at a fixed time point at discharge, results are contrary to the post hoc findings of the AFFIRM study demonstrating high long-term mortality in patients with AF on digoxin. All the above studies quote comorbidities and elderly age as confounding factors, but in this Middle East registry, with a mean age of 56 years with all confounding factors adjusted, there was still highly significant increase in long-term mortality in patients on digoxin without HF.
In the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) study, digoxin treatment was also associated with a significant increase in all-cause mortality, vascular death, and sudden death in patients with AF. 20 However, with regard to patients with HF, the Digitalis Investigation Group (DIG) study, one of the largest randomized study recruiting 7000 patients, although excluded patients with AF, showed safety of digoxin in patients with HF. There was no effect on mortality, but digoxin reduced the rate of hospitalizations for HF leading to its recommendation by the guidelines in HF. 21 In another study by Pastori et al, 22 similar findings were observed implicating digoxin with mortality using propensity score-matched analysis. The cohort of this study is significantly older (56 vs 73 years), however, the Gulf SAFE group only had 54% of the overall cohort anticoagulated compared to all patients in the study by Pastori et al. 22
The exact reasons behind digoxin-associated increased mortality in patients with AF without HF are not known. Digoxin is also being linked to malignancy, a prothrombotic state, with increased markers of endothelial activation, stroke, increased mortality in end-stage renal failure patients, and digoxin-induced arrhythmias. 19,23,24 Overall, the above studies, 19,23,24 including this current registry’s findings, confirm the increased mortality rates related to digoxin in patients with AF without HF. Thus, in relatively healthy young patients with AF as their only problem with no HF, the risk of death at 1 year is high if discharged on digoxin therapy. In addition, one important difference needs to be considered while comparing Middle East registry with Western registries. The Middle East population (mean age, 56 years) is a decade younger than Western AF populations (mean age, 72 years), 9 and this high mortality shows that digoxin use in patients with only AF is an alarming sign that needs urgent intervention from health authorities in educating physicians with regard to digoxin use in Middle East patients with AF.
A major limitation of this study is its observational design that limits its ability to assess causal relationships. The study was also not able to confirm whether patients continued digoxin during follow-up or changed to other medications. Furthermore, digoxin doses as well as digoxin serum levels were not available, and hence, it is difficult to assess whether digoxin levels could have contributed to mortality in any way. All-cause, rather than cardiac-specific, mortality was recorded, and hence, the impact of digoxin on all-cause mortality might have been clouded with other unmeasured confounding factors. Finally, the multivariate Cox proportional hazards regression employing a time to death model would have been a more appropriate model to use; however, the variable “time to death” was not captured in the Gulf SAFE registry. What was captured was only the death occurred within the 1-year period.
Conclusion
This is the first study from the Middle East region to show that in patients with AF without HF, digoxin therapy was associated with significantly higher long-term mortality compared to those who were not on digoxin therapy. In addition, in patients with AF and HF, digoxin did not offer any survival advantages. These findings should be seriously considered given the younger age of AF population in this part of the world. These findings indicate that physicians should strongly consider other drug options before prescribing digoxin as a rate-controlling drug in patients with AF without HF.
Footnotes
Author Contributions
I. Al-Zakwani contributed to conception and design, contributed to acquisition, analysis, and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. P. Panduranga contributed to conception and design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. M. Zubaid contributed to conception and design, contributed to acquisition, analysis, and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. K. Sulaiman contributed to conception and design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. W. Rashed contributed to conception and design, contributed to acquisition and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. A. Alsheikh-Ali contributed to conception and design, contributed to analysis and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. W. AlMahmeed contributed to conception and design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. A. Shehab contributed to conception and design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. A. Al Qudaimi contributed to design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. A. Asaad contributed to conception and design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. H. Amin contributed to conception and design, contributed to interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy.
Acknowledgments
The authors thank the patients, physicians, nurses, and support staff who participated in the Gulf SAFE for their invaluable cooperation. The study has also been presented as a poster at the 12th Gulf Heart Association and 4th Kuwait Cardiac Society Conference in Kuwait on March 4-7, 2015.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Prof Zubaid and Drs Alsheikh-Ali and AlMahmeed received speaker honoraria from Sanofi and Boehringer Ingelheim. Dr Al-Zakwani received honoraria from Sanofi for data analysis.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Gulf SAFE is a Gulf Heart Association project supported by Sanofi-Aventis, Paris, France. The sponsors had no involvement in the study conception or design; data collection, analysis, or interpretation of data; writing, review, or approval of the manuscript; or the decision to submit the manuscript for publication.