EDITORIAL COMMENT
Predicting Thromboembolism and Selecting Patients for Anticoagulant Therapy in Atrial Fibrillation*
William S. Weintraub, MD, FACC*
Christiana Care Health System, Newark, Delaware.
* Reprint requests and correspondence: Dr. William S. Weintraub, Cardiology Section, Christiana Care Health System, 4755 Ogletown-Stanton Road, Newark, Delaware 19718. (Email: wweintraub{at}christianacare.org).
Atrial fibrillation remains a challenge to manage. Although there are issues concerning rate and rhythm control, prevention of thromboembolism remains the most important aspect of management. Untreated, strokes will occur in about 2% to 5% of patients per year, and the strokes caused by atrial fibrillation can be quite devastating (1,2). Although new pharmacologic and nonpharmacologic approaches to prevention of thromboembolism have been subjects of intense investigation, treatment with warfarin remains the cornerstone of therapy. This is true despite the well-known limitations of warfarin, and remains true because warfarin is effective and can be expected to reduce the incidence of thromboembolism by about two-thirds.
Warfarin has several problems that limit the enthusiasm of clinicians: there is a narrow therapeutic window, there are multiple drug and food interactions, there is genetic variability in response, the half-life is long, and it increases the incidence of bleeding. Therefore, it would appear wise to try to identify patients at sufficiently low risk of thromboembolism or sufficiently high risk of complications of warfarin so that warfarin would not be used. This was investigated by Fang et al. (3) using the ATRIA (AnTicoagulation and Risk Factors In Atrial Fibrillation) study cohort. This cohort is composed of 13,559 adults with atrial fibrillation, with a median follow-up of 6 years. The study considered 10,932 patients who were off of warfarin for some period of time and also 5,588 patients who were not on warfarin at baseline and had at least a 12-month fixed period off of warfarin. Events were adjudicated by committee. Five different stratification schemes were investigated: the Atrial Fibrillation Investigators, Stroke Prevention in Atrial Fibrillation, Congestive heart failure, Hypertension, Age >75 years, Diabetes mellitus, and prior Stroke or transient ischemic attack score, Framingham Score, and 7th American College of Chest Physicians Conference on Antithrombotic and Thrombolytic Therapy (ACCP) guidelines. For each scheme the patients were grouped as low risk, intermediate risk, or high risk. The prevalence of low risk varied from 11.7% with 7th ACCP to 37.1% with Framingham and high risk from 16.4% with Framingham and 80.4% with 7th ACCP. The annualized thromboembolism rate in the low-risk group varied from 0.13% with 7th ACCP to 0.81% with Framingham. The annualized thromboembolism rate in the high-risk group varied from 2.5% with 7th ACCP to 3.9% with Framingham.
The performance of the risk schemes was evaluated with the c index. The c index evaluates the discrimination of prediction models; thus, the c index is the faction of pairs of patients, 1 with and 1 without an event, where the one with the event will have a higher probability of the event occurring based on the model. A c index of 0.5 means that there is no ability to discriminate and a c index of 1 means perfect discrimination. This study showed c indexes of 0.60 to 0.69 in the group with no warfarin at baseline. This suggests that discrimination is fair at best. Nonetheless, the models could be viewed as calibrating better than discriminating. Calibration refers to how well a model’s prediction of risk correlates with observed risk. Here the results were not so bad, and risk varied from 0.13% in low-risk 7th ACCP to 2.5% in high-risk 7th ACCP. It should be noted that calibration and discrimination are not the same thing and that models with perfect calibration have an upper limit of discrimination (4).
Overall, this study is well done and carefully analyzed. There are several limitations that the investigators recognize. It is not possible or ethical to keep patients off of warfarin. Thus to assess natural history, it is necessary to either follow up patients during periods off of warfarin or follow up patients who do not or cannot take warfarin. This may provide a population that is not entirely characteristic of all patients with atrial fibrillation. Also, it was not possible to assess aspirin usage, which may have lowered the thromboembolism rate somewhat. There are also limitations to the database in that blood pressure and left ventricular function were not assessed. This resulted in several of the scores being approximations rather than exact assessments using the scores as published.
The primary finding of this study is that although models predict a spectrum of risk, it is hard to say in any one patient whether they will or will not suffer an event. Given the catastrophic outcome of stroke in patients with atrial fibrillation as well as the efficacy of warfarin, it remains the standard of care to treat almost everyone (1,5). What level of risk might be considered so low that warfarin might not be used? This is actually beyond the scope of the present study; this issue could, in principle, be approached with cost-effectiveness analysis. However, in all patients in whom there is net clinical benefit to warfarin, at least on clinical grounds it should be used. The low-risk group identified by Fang et al. (3) with ACCP, with a risk of 0.13% per year, might qualify. These are patients younger than age 65 with no identified risk factors, a reasonable group to consider not anticoagulating. However, this was only 11.7% of the population.
So, for now, most patients with atrial fibrillation will be anticoagulated, saving those at very low risk and patients who for one reason or another cannot or will not take warfarin. Is there any difference between permanent, persistent, and paroxysmal atrial fibrillation? The risk of stroke is increased in all 3 types of atrial fibrillation, and generally the type of atrial fibrillation is not considered (2). However, there are always exceptional cases. Does the patient presenting with a first-ever episode who spontaneously converts need to be anticoagulated? Does anticoagulation always need to be for life? If a patient has been in sinus rhythm for 2 years with no palpitations, can he or she come off warfarin? Although this may seem reasonable, it is not possible to be sure that the patient was not having silent episodes of atrial fibrillation. When does the risk of falls and head injury become too great and thus the competing risk of embolic stroke become acceptable?
The end result is that atrial fibrillation management will remain challenging, with difficult medical decision making that may not be adequately assessed by guidelines. Fang et al. (3) stress the need for more detailed models with better ability to discriminate which patients will suffer a thrombotic event. There is also a need for better pharmacotherapy, offering less complicated management than warfarin; this is a subject of intense interest by the pharmaceutical industry. Finally there is need for reliable, durable approaches to maintaining sinus rhythm. It is not a given that maintenance of sinus rhythm will translate into reduced stroke risk. The stroke risk was actually high in the rhythm control arm of the AFFIRM (Atrial Fibrillation Follow-up Investigation of Rhythm Management) study (6). Approximately one-half of the patients who are converted to sinus rhythm can be expected to have recurrent atrial fibrillation within 1 year (7). Thus, we still lack a reliable pharmacologic or procedural approach to establishing and maintaining sinus rhythm. When such an approach becomes available, it will still be necessary to prove that patients are not at a level of risk of thromboembolism that would still warrant anticoagulation.
So, the challenge of atrial fibrillation management continues. Most of our patients will be on warfarin, mostly for a long time.
 |
Footnotes
|
|---|
* 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. 
|
References
|
|---|
1. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillationAnalysis of pooled data from five randomized controlled trials. Arch Intern Med 1994;154:1449-1457.[Abstract/Free Full Text]2. Fuster V, Ryden LE, Cannom DS, et al. ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation) Circulation 2006;114e257–354. 3. Fang MC, Go AS, Chang Y, Borowsky L, Pomernacki NK, Singer DE, ATRIA Study Group Comparison of risk stratification schemes to predict thromboembolism in people with nonvalvular atrial fibrillation J Am Coll Cardiol 2008;51:810-815.[Abstract/Free Full Text] 4. Diamond GA. What price perfection?Calibration and discrimination of clinical prediction models. J Clin Epidemiol 1992;45:85-89.[CrossRef][Web of Science][Medline] 5. Lin HJ, Wolf PA, Kelly-Hayes M, et al. Stroke severity in atrial fibrillationThe Framingham Study. Stroke 1996;27:1760-1764.[Abstract/Free Full Text] 6. Wyse DG, Waldo AL, DiMarco JP, et al. A comparison of rate control and rhythm control in patients with atrial fibrillation N Engl J Med 2002;347:1825-1833.[Abstract/Free Full Text] 7. Madrid AH, Bueno MG, Rebollo JM, et al. Use of irbesartan to maintain sinus rhythm in patients with long-lasting persistent atrial fibrillation: a prospective and randomized study Circulation 2002;106:331-336.[Abstract/Free Full Text]
Related Article
-
Comparison of Risk Stratification Schemes to Predict Thromboembolism in People With Nonvalvular Atrial Fibrillation
- Margaret C. Fang, Alan S. Go, Yuchiao Chang, Leila Borowsky, Niela K. Pomernacki, Daniel E. Singer for the ATRIA Study Group
J. Am. Coll. Cardiol. 2008 51: 810-815.
[Abstract]
[Full Text]
[PDF]
This article has been cited by other articles:
|
|
|
|
 
M. M. Sood, P. Komenda, A. R. Sood, C. Rigatto, and J. Bueti
The Intersection of Risk and Benefit: Is Warfarin Anticoagulation Suitable for Atrial Fibrillation in Patients on Hemodialysis?
Chest,
October 1, 2009;
136(4):
1128 - 1133.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. N. DeMaria, O. Ben-Yehuda, J. J. Bax, G. K. Feld, B. H. Greenberg, W. Y.W. Lew, J. A.C. Lima, A. S. Maisel, S. M. Narayan, D. J. Sahn, et al.
Highlights of the Year in JACC 2008.
J. Am. Coll. Cardiol.,
January 27, 2009;
53(4):
373 - 398.
[Full Text]
[PDF]
|
|
|
|
Top
References