Author + information
- Mintu P. Turakhia, MD, MAS∗ ()
- Department of Medicine and Center for Digital Health, Stanford University School of Medicine, Stanford, California; and the Veterans Affairs Palo Alto Health Care System, Palo Alto, California
- ↵∗Address for correspondence:
Dr. Mintu P. Turakhia, Veterans Affairs Palo Alto Health Care System, 3801 Miranda Avenue, 111C, Palo Alto, California 94304.
Digoxin is a curious drug with an unusual history. The use of digitalis genus plants was first described in 1785 by British physician Dr. William Withering, for the treatment of edema (“dropsy”) and other peculiar conditions (1). Following the advent of isolation of digoxin from the foxglove plant in the 1930s, the use of cardiac glycosides steadily soared to be a staple therapy of atrial fibrillation (AF) and heart failure for the past few decades. As of 2015, the World Health Organization still includes it on its Model List of Essential Medications to treat heart failure and arrhythmias (2). Digoxin is used as an abortifacient and, more recently, has been extolled for potential anticancer effects.
However, digoxin has an unusual and even macabre lore in medicine. Called “Dead Man’s Thimbles” in Ireland, the narrow therapeutic window of digitalis has captivated the profession. Indeed, Dr. Withering’s uncontrolled and unblinded case series of 100 patients included several who died after foxglove exposure, including an alcoholic with cirrhosis and oliguria who developed “fatal syncope” after digitalis infusion (1). In medical school and residency, considerable time is expended on recognizing the harms of digoxin, including electrocardiogram findings of digoxin effect versus toxicity, indications for digoxin antigen-binding fragments, and vignettes of little old tea-drinking ladies presenting with nausea and bradycardia. Digoxin is cardiology’s “great masquerader”—our version of syphilis: it can conceivably cause almost any arrhythmia. And even when all appears well, a patient may report green-tinted vision or weight loss from intractable nausea. The observational association of harm in heart disease is historically well founded; 37 years ago, Moss et al. (3) reported an increase in cardiac mortality associated with digoxin when used after myocardial infarction.
What, then, is our obsession in cardiology with this not-so-wonder-drug, particularly in the absence of any compelling data demonstrating benefit in AF? Perhaps it is our attraction to the many mechanisms of digoxin, ranging from parasympathetic and vagolytic effects that slow atrioventricular node conduction to improving myocardial contractility under favorable energetics. In the DIG (Digitalis Investigation Group) trial, a historical relic that enrolled patients 26 years ago with left ventricular systolic dysfunction and sinus rhythm, digoxin reduced the need for hospitalization without affecting survival. Importantly, the DIG trial excluded patients with AF and also predated other life-saving heart failure therapies, including beta-blockers, angiotensin blockers, and aldosterone antagonists, among others. A post hoc analysis of this trial identified an association between digoxin level and mortality. Still, American and European professional society guidelines for AF give digoxin a class I indication for use as a rate control agent (4,5).
In the absence of trial data, dozens of observational studies and meta-analyses have attempted to answer this question. Although several meticulous, large observational studies have shown an association of digoxin use in AF with mortality (6,7), many have not, and meta-analyses have sometimes obscured the clarity of individual, well-done studies. The Achilles’ heel for all of these nonrandomized studies has been the significant potential for confounding and bias, with a seemingly intractable tradeoff between granularity in small, potentially biased and nongeneralizable data sources versus limited granularity in larger but more generalizable data.
Lopes et al. (8) have put their stake in the ground on the topic in this issue of the Journal. Using data from the ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) trial (9), the authors systematically and comprehensively assessed the association of digoxin with mortality in AF. The nature of their data has numerous strengths over previous research: systematic rather than indication-driven data collection as part of a controlled clinical trial, adjudicated cause of death, biomarker assessments with high explanatory power for mortality, and changes to variables captured over time. The authors performed 2 parallel assessments in prevalent (more prone to survival bias) and incident (new user) analyses. The latter was rigorously performed with sophisticated but judicious propensity matching across a dense set of covariates (including biomarkers) and state-of-the-art methods for accounting for time-dependent covariates.
In 17,897 participants representing 98% of the ARISTOTLE study cohort, prevalent digoxin use (5,824 participants) was not associated with all-cause, cardiovascular, or sudden cardiac death (8). In a prevalent user analysis according to digoxin concentration, there was a linear, dose-dependent increase in the adjusted hazard of all-cause death and cardiovascular death (but not sudden death). In the incident (new user) cohort, limited by a much smaller digoxin-exposed sample size of 873 participants, digoxin initiation was associated with an increased risk of all-cause death and sudden cardiac death, with early separation of sudden death survival curves. Importantly, none of these effects was modified by the presence of heart failure. However, new digoxin users had an increased risk of heart failure hospitalization. The authors appropriately concluded that although the risk of digoxin is related to serum concentration, initiation of digoxin is independently associated with high mortality, regardless of heart failure status.
Ultimately, no amount of statistical machination, however brilliant, can overcome the problem of unidentified confounders, particularly those that are time varying, which the authors fully and humbly acknowledge (8). Their analytical rigor and range of sensitivity analyses (note the 11 supplemental tables and figures) provide clarity but not certainty. If the trial had ascertained other outcomes that would be associated with confounders but not necessarily with digoxin (e.g., pneumonia), then a falsification analysis could certainly bolster the results. However, the results are biologically plausible. In the MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy) study of cardiac resynchronization therapy with defibrillation, patients taking digoxin had a higher risk of ventricular tachycardia/ventricular fibrillation (10). In the PALLAS (Permanent Atrial Fibrillation Outcome Study Using Dronedarone on Top of Standard Therapy) trial of AF with randomization to treatment with dronedarone, which increases digoxin levels, there was an interaction with digoxin and death: all arrhythmic deaths in digoxin-treated patients were among those taking dronedarone (11).
Should these new data therefore serve as the final nail in the coffin for digoxin? Also, what are the global implications for a low-cost drug still prescribed in up to 40% of the world’s AF population? The RATE-AF (Rate Control Therapy Evaluation in Permanent Atrial Fibrillation) trial, planned for completion in 2019, will compare digoxin versus bisoprolol as a first-line rate control therapy in permanent AF for the outcome of quality of life (12). However, with only 160 patients, the study is not powered to safety at effect sizes estimated from observational studies.
Therefore, for the majority of patients with AF, there is no reason to believe that there is any benefit but rather possible or probable harm of digoxin compared with other atrioventricular nodal agents. The temptation to preferentially reach for digoxin in heart failure is also not supported by these data. Cautious initiation (or discontinuation) guided by digoxin levels seems biologically rational and empirically supported. However, a more pragmatic solution is to simply stop using it.
Perhaps it’s time leave foxglove in the garden and in the history books—and out of the medicine cabinet.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
The content and opinions expressed are solely the responsibility of the author and do not necessarily represent the views or policies of the Department of Veterans Affairs. Dr. Turakhia has received research grant support from Janssen Pharmaceuticals, Medtronic Inc., AstraZeneca, Veterans Health Administration, and Cardiva Medical Inc.; other research support from AliveCor Inc., Amazon, Zipline Medical Inc., iBeat Inc., and iRhythm Technologies Inc.; and honoraria from Abbott, Medtronic Inc., Boehringer Ingelheim, Precision Health Economics, iBeat Inc., Akebia, Cardiva Medical Inc., and Medscape/theheart.org.
- ↵Withering W. An Account of the Foxglove, and Some of its Medical Uses with Practical Remarks on Dropsy, and Other Diseases. Birmingham Print. by M Swinney 1785:207.
- ↵World Health Organization. WHO Model List of Essential Medicines—19th List (April 2015). Essent Med. 2015:1–45. Available at: http://www.who.int/medicines/publications/pharmacopoeia. Accessed January 4, 2018.
- Moss A.J.,
- Davis H.T.,
- Conard D.L.,
- DeCamilla J.J.,
- Odoroff C.L.
- January C.T.,
- Wann L.S.,
- Alpert J.S.,
- et al.
- Turakhia M.P.,
- Santangeli P.,
- Winkelmayer W.C.,
- et al.
- Lopes R.D.,
- Rordorf R.,
- De Ferrari GM,
- et al.
- Lee A.Y.,
- Kutyifa V.,
- Ruwald M.H.,
- et al.
- Hohnloser S.H.,
- Halperin J.L.,
- Camm A.J.,
- Gao P.,
- Radzik D.,
- Connolly S.J.
- ↵United States National Library of Medicine. Rate Control Therapy Evaluation in Permanent Atrial Fibrillation (RATE-AF). Available at: https://clinicaltrials.gov/ct2/show/NCT02391337. Accessed January 9, 2018.