Does Digoxin Really Increase Mortality Among Outpatients With Atrial Fibrillation?

When association is association and causation is causation.

Integrated health care systems with large patient databases offer a way to look at broad populations like never before. Rather than observational studies with a few thousand participants, a pair of studies that associated azithromycin with an increased risk of cardiovascular death had populations of more than 2 million patients and more than 7 million patients, respectively.^1,2 In similar methodological fashion, the TREAT-AF study followed suit.³

Digoxin is currently recommended as a pharmacological rate control strategy in atrial fibrillation (AF) patients with heart failure or left ventricular dysfunction, as well as sedentary patients.⁴ Digoxin can also be used alone or in combination with beta-blockers or nondihydropyridine calcium-channel blockers in patients with AF whose ventricular rate is not adequately controlled either at rest or during exercise.⁴

However, these recommendations rely on low-quality evidence, and the available evidence is decades old. So, in an effort to re-evaluate the use of digoxin in contemporary management of AF, the TREAT-AF investigators set out to examine the association between digoxin therapy and mortality in newly diagnosed AF patients treated in the US Department of Veterans Affairs (VA) health care system.

Using data from multiple VA centralized datasets, the investigators identified 122,465 patients who were newly diagnosed with nonvaluvlar AF, were seen within 90 days in an outpatient care setting, and received any outpatient prescription within 90 days after index AF diagnosis. Patients were excluded if they had any prior diagnosis of AF, history of valve disease or replacement, thyroid disease, kidney transplant, or cardiac surgery within 30 days. The patient populations compared were those who received a digoxin prescription within 90 days of the index AF diagnosis, and those who did not receive digoxin in the same time period. The primary outcome was time to death, beginning 90 days after index AF diagnosis.

The authors demonstrated a statistically significant association between digoxin and increased risk of unadjusted mortality in a 90-day period (hazard ratio [HR]: 1.26, 95% confidence interval [CI]: 1.23 to 1.29, P < 0.001). The unadjusted mortality rates, which were made available in an online-only table, were 10.2% (8724 of 85,259) in the digoxin group, compared with 7.4% (19,999 of 267,909) in the non-digoxin group.

Investigators matched patients with similar propensity scores to account for unidentified confounding variables (HR: 1.21, 95% CI: 1.17 to 1.25, P <.001), and the rates of propensity matched mortality were 10.5% (8439 of 79,886) among those who received digoxin, compared with 8.7% (6641 of 76,246) among those who did not receive digoxin. The results were consistent across all subgroups, including drug adherence, kidney dysfunction, heart failure, or concomitant therapy with beta-blockers or amiodarone.

The results of this study add to the available evidence, albeit observational in nature, that digoxin may have a negative impact on patient survival.^5-6 However, before digoxin is relegated to antiquity, it is critical to point out that these studies demonstrate association, not causation. While the statistical methods used in this study diligently attempt to account for all identified and unidentified confounders, it cannot replace a prospective, randomized study design.

Furthermore, interesting phenomena of unusual associations with large data sets such as this has been discussed on numerous media sources, including a blog called Spurious Correlations (via The Skeptics Guide To Emergency Medicine).^7,8 Such statistically significant correlations, or associations, that can be made include how the number of people who drowned by falling into a swimming pool correlated with the number of films Nicolas Cage appeared in (correlation: 0.666004).^7,8

While it may be silly to make a comparison between this association and the one observed in the TREAT-AF study, the approach to this type of data should include the same amount of critical skepticism and spark discussions among colleagues and cardiologists regarding these findings.

References:

1. Ray WA, Murray KT, Hall K, Arbogast PG, Stein CM. Azithromycin and the risk of cardiovascular death. N Engl J Med. 2012 May 17;366(20):1881-90

2. Svanström H, Pasternak B, Hviid A. Use of azithromycin and death from cardiovascular causes. N Engl J Med. 2013 May 2;368(18):1704-12

3. Turakhia MP, Santangeli P, Winkelmayer WC, et al. Increased mortality associated with digoxin in contemporary patients with atrial fibrillation: findings from the TREAT-AF study. J Am Coll Cardiol. 2014 Aug 19;64(7):660-8

4. Anderson JL, Halperin JL, Albert NM, et al. Management of patients with atrial fibrillation (compilation of 2006 ACCF/AHA/ESC and 2011 ACCF/AHA/HRS recommendations): a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2013;61:1935—44

5. The Digitalis Investigation Group. The effect of digoxin on mortality and morbidity in patients with heart failure. N Engl J Med. 1997; 336:525—33

6. Whitbeck MG, Charnigo RJ, Khairy P, et al. Increased mortality among patients taking digoxin—analysis from the AFFIRM study. Eur Heart J. 2013;34:1481—8

7. Vigen T. Spurious Correlations [Internet]. —cited 2014 October 14. Available from: http://www.tylervigen.com/

8. Rezaie S, Milne K. The Skeptics Guide to Emergency Medicine [Internet]. SGEM#76: And the Beat Goes On (Azithromycin and Risk of Cardiovascular Death). 2014 May —[cited 2014 October 14]. Available from: http://thesgem.com/2014/05/sgem76-and-the-beat-goes-on-azithromycin-and-risk-of-cardiovascular-death/