Evidence Based Review Article

The Journal of Informed Pharmacotherapy 2003;12:202.

The Atrial Fibrillation Follow-up Investigation of Rhythm Management: Does this Study AFFIRM the Need for Rate or Rhythm Control?

Reviewer: Richard Slavik, Pharm.D.
Reviewer’s e-mail address: rslavik@interchange.ubc.ca 

Reviewer's profession/specialty: Pharmacotherapeutic Specialist -
Critical Care

Original Citation

Wyse DG, Waldo AL, DiMarco JP, Domanski MJ, Rosenberg Y, Schron EB, Kellen JC, Greene HL, Mickel MC, Dalquist JE, Corley SD; The Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) Investigators.  A Comparison of Rate Control and Rhythm Control in Patients with Atrial Fibrillation.  NEJM 2002:347:1825-33. PubMed Citation

Overall Study Question

The AFFIRM trial was a prospective, randomized, open-label, multicentre study of atrial fibrillation (AF) patients enrolled from November 1995 to October 1999 at 213 sites in the U.S. and Canada.  

The objective of the trial was to compare the long-term effects of "rate control" and "rhythm control" strategies. (1,2,3)  Patient enrollment required that all of the following inclusion criteria be met: 1) an EKG-documented AF episode (³ 1 hour) within the past 12 weeks; 2) at least 65 years old and at least one other risk factor for stroke (e.g. systemic hypertension, diabetes mellitus, congestive heart failure, transient ischemic attack, prior stroke, left atrium ³ 50 mm on echocardiograph, left ventricular fractional shortening < 25% on echocardiogram, or left ventricular ejection fraction < 40%); 3) duration of AF episodes in the previous 6 months must have totaled ³ 6 hours; 4) duration of continuous AF must have been < 6 months, unless normal sinus rhythm (NSR) could be restored and maintained ³ 24 hours; 5) patient must have been eligible to receive either a "rate control" or "rhythm control" strategy, and; 6) patients must have been eligible to receive ³ 2 rate controlling or antiarrhythmic drugs (including low and high dose amiodarone). (1,2,3) 

Exclusion criteria included failed cardioversion prior to randomization, contraindications to any of the potential study therapies, and other cardiac, medical, and non-medical contraindications that were not stated in the AFFIRM publication. (www.axioresearch.com/affirm/study)  

Patients were randomized to a strategy of either "rate control" or "rhythm control", with specific therapies chosen based on physicians' discretion but guided by a list of "acceptable" options from the study protocol.  Anticoagulation was mandated in the group assigned to the "rate control" strategy, but it could be withdrawn and replaced by ASA in patients in the "rhythm control" group if they were in continuous NSR for at least 4 (but preferably 12) weeks. 

The primary outcome for this study was all-cause mortality.  The secondary outcome was a composite of death, disabling stroke, disabling anoxic encephalopathy, major bleeding, and cardiac arrest.

What was the declared relationship between authors and sponsors for this study?

The AFFIRM Writing Group assumed overall responsibility for the content of the manuscript, although this trial was coordinated at the AFFIRM Clinical Trial Center by Axio Research (Axio Research, Seattle, WA).  The study was funded by a National Heart, Lung and Blood Institute grant.  Drs. Wyse, Waldo, DiMarco, and Green have declared relationships with several Pharmaceutical Manufacturers, and Wyeth-Ayerst Laboratories contributed supplies of amiodarone for the study.

Are the Results of the Study Valid?

1. Was assignment of patients randomized?

Yes.  Randomization was centrally performed using a permuted block design with equal allocation and stratification by clinical site.  The authors did not explicitly state what method was used to randomize patients, and if it was performed under concealed conditions.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Yes.  During a 48-month period, 7,401 patients were screened, 4,060 (55%) were enrolled, with 2,027 participants assigned to the "rate control" strategy and 2,033 assigned to the "rhythm control" strategy.  The results of the screening log and a study flow chart were not provided to account of all of the patients screened and enrolled in the trial. The authors do state that 71 patients (1.7%) withdrew consent, and that vital status was unknown in 26 patients (0.6%).  The primary outcome was time to all-cause mortality analyzed using intention-to-treat principles.

3. Were patients, their clinicians, and study personnel 'blind' to treatment?

No.  This was an open-label trial and both the investigators and patients were aware of the treatment assignment.

4. Were the groups similar at the start of the trial?

Yes.  The baseline characteristics of both treatment groups were described in the primary AFFIRM publication and a supplemental publication. (2,3) The mean age was 70 years old, with 39% female, and 89% Caucasian. Hypertension was present in 71%, coronary artery disease in 38%, mean left ventricular ejection fraction was 55%, and left atrial enlargement was presenting 65% of patients. (2,3)  There was a history of congestive heart failure in 24%, diabetes in 20%, myocardial infarction in 17%, coronary artery bypass graft (CABG) surgery in 13%, and a stroke or TIA in 13%. (2,3)  The study primarily enrolled patients with paroxysmal AF. Thirty-six percent were enrolled after their first episode of AF, 69% had a qualifying episode of AF for ³ 2 days and 18% had had a previous failure of an antiarrhythmic medication. The baseline rhythm was NSR in 54% of patients in the entire cohort. (2)  Overall, patients in both treatment strategy groups had similar baseline demographic, clinical, and prognostic characteristics. (2,3)

5. Aside from the experimental intervention, were the groups treated equally?

Unable to determine.  Given that this trial was designed to compare treatment strategies under real world conditions, there were many treatment options utilized within each of the groups.  

Acceptable agents for the "rate control" strategy included beta blockers, calcium channel blockers (verapamil and diltiazem), digoxin, or a combination of these agents to achieve a goal heart rate of < 80 bpm and exercise HR £ 110 bpm.  If pharmacological agents failed, AV nodal modification or ablation and the placement of a pacemaker was allowed. (1,3)  

Acceptable agents for the "rhythm control" strategy included amiodarone, disopyramide, flecainide, moricizine, procainamide, propafenone, quinidine, sotalol, dofetilide, or a combination of these agents based on specific guidelines. (1,3)  Co-interventional anticoagulation was mandated in the group assigned to the "rate control" strategy, but it could be withdrawn and replaced by ASA in patients in the "rhythm control" group if they were in continuous NSR for at least 4 (but preferably 12) weeks. Specific information on anticoagulation use in each group was not reported. As expected, anytime use of beta blockers was more common in the "rate control" group (68%) versus the rhythm control group (50%). 

There was no information concerning any other co-interventions that could have affected long-term mortality (e.g. ACE inhibitors), and whether they were mandated by the investigators. This may be important in an un-blinded trial due to potential inherent biases with open-label treatment regimens.

What were the Results?

1. How large was the treatment effect? 

As mentioned, 4,060 patients were enrolled with a mean and maximum follow-up of 3.5 years and 6 years, respectively.  In the entire cohort, 35% of patients were in NSR, and 80% of those in AF had adequate heart rate control at the 5-year visit.  The primary analysis of time to death from any cause as determined from the Kaplan Meier analysis showed a trend towards increased mortality in the "rhythm control" group.  Overall, there were 352 deaths (23.8%) in the "rhythm control" group and 306 deaths (21.3%) in the "rate control" group at 5 years (unadjusted HR 1.15 (95%CI 0.99 to 1.34), p=0.08). After adjustment for differences between the groups in pre-specified prognostic factors using the Cox proportional hazards model, the results were the same (adjusted HR 1.18 (95%CI 0.99 to 1.41), p=0.07). The frequency of the composite secondary endpoint was similar in the "rhythm control" and "rate control" groups (32.7% vs. 32.0%, p=0.33).  Due to lack of effectiveness or an adverse drug reaction (ADR) associated with the antiarrhythmic medications, a significantly higher proportion of patients crossed-over from the "rhythm control" arm to the "rate control" arm than vice versa (29.2% vs. 12.2%, p<0.001).

There was no difference in the incidence of ischemic stroke (7.1% and 5.5%, p=0.79) or hemorrhage not involving the central nervous system (6.9% vs. 7.7%, p=0.44) between the "rhythm control" and "rate control" groups. The majority of ischemic strokes occurred in patients who had stopped taking warfarin or whose INR was subtherapeutic.  Scores on the Mini-Mental State exam and selected measures of quality of life were reported as similar between the groups.

There was a significantly higher incidence of hospitalizations during follow-up in the rhythm control group than the "rate control" group (80.1% vs. 73.0%, p<0.001, NNH=14). There was also a higher incidence of bradycardic cardiac arrests (0.6% vs. 0.1%, p=0.01, NNH=200). There were more ADRs leading to discontinuation of study therapy in the "rhythm control" group than the "rate control" group: pulmonary event (7.3% vs. 1.7%, p<0.001, NNH=18); gastrointestinal event (8.0% vs. 2.1%, p<0.001, NNH=17); bradycardia (6.0% vs. 4.2%, p=0.001, NNH=56); QTc>520 msec (1.9% vs. 0.3%, p<0.001, NNH=63); other unspecified ADRs (25.4% vs. 14.0%, p<0.001, NNH=9).

2. How precise was the estimate of the treatment effect?

For the primary outcome of all-cause mortality at five years, the 95% CI for the unadjusted and adjusted hazard ratios crossed one, and thus the mortality difference is not statistically significant.  Unfortunately the 95% CI for many of the endpoints are not provided, and cannot be easily calculated from the data presented.  Thus, the precision of the treatment effects for most of the endpoints cannot be determined.

Will the Results Help Me in Caring for My Patients?

1. Can the results be applied to my patient care?

Yes.  The AFFIRM trial is the first large-scale clinical trial to compare long-term treatment strategies of "rate control" versus "rhythm control" in a generally representative population of AF patients.  The only subsets not represented in this trial are patients with chronic, persistent AF with frequent or severe symptoms, and younger patients with AF and no other risk factors for stroke.  For patients meeting the AFFIRM enrollment criteria, this study has shown that a "rhythm control" strategy is associated with a non-statistically significant trend towards higher mortality, more bradycardic cardiac arrests, more frequent hospitalizations, and a higher incidence of ADRs causing discontinuation of medications.

2. Were all clinically important outcomes considered?

Yes.  The most important outcomes of all-cause mortality, ischemic stroke, bleeding rates, and ADRs were reported, however quality of life data was not specifically shown.  A pharmacoeconomic analysis of this study would also be interesting.

3. Are the likely treatment benefits worth the potential harms and costs?

No. Based on the AFFIRM results, the theoretical benefits of employing a routine long-term strategy of "rhythm control" preferentially over "rate control" have not been realized, and are not worth the harms and costs of therapy.  Given the non-statistically significant trend towards higher mortality, lack of reduction in stroke risk, and higher hospitalization rates and ADRs with the "rhythm control" strategy, the preferred initial treatment strategy should be "rate control" until further data are available. 


AF is the most common sustained cardiac arrhythmia with an incidence and prevalence that increases with age.  With our aging population, the impact of AF on symptoms, morbidity, mortality, and costs will continue to rise.  Traditional pharmacological goals for AF have been to control ventricular response, convert to NSR, and prevent stroke. However, there has been much controversy over the ideal long-term treatment strategy - "rate control" or "rhythm control".

Recently published studies in post-CABG surgery patients with AF and outpatients with recurrent, persistent AF have suggested that maintenance of NSR with antiarrhythmic medications may not be the preferred treatment strategy, despite its theoretical benefits. (4-6)  A randomized open-label pilot study comparing a "rate control" strategy versus a "rhythm control" strategy in 50 post-CABG patients with AF showed no difference in time to conversion to NSR, relapse to AF at 1, 4, or 8 weeks, ADRs, or complications. (4)  There was a significantly shorter length of hospital stay in the "rate control" group (9.0 days vs. 13.2 days, p=0.05). (4)  The Pharmacological Intervention in Atrial Fibrillation (PIAF) trial was a randomized open-label trial of 252 patients with AF comparing a "rate control" strategy with diltiazem to a "rhythm control" strategy with amiodarone. (5)  Although significantly more patients were in NSR at 1 year in the "rhythm control" group (56% vs. 10%, p<0.05), significantly more patients had ³ 1 hospitalization during the 1-year follow-up (69% vs. 24%, p<0.05), and more patients stopped their medication due to ADRs (25% vs. 14%, p<0.05) in the "rhythm control" group. (5) Finally, a recently-published randomized, open-label study of 522 patients with persistent AF or atrial flutter showed the rates of a composite endpoint of death from cardiovascular causes, heart failure, thromboembolic complications, bleeding, pacemaker implantation, and serious ADRs were similar with "rhythm control" and "rate control" strategies (22.6% vs. 17.2%, ARR -5.4 (90% CI -11.0% to 0.4%), and that the "rate control" strategy was not inferior to "rhythm control". (6)

The AFFIRM trial confirms that despite theoretical pathophysiological benefits of maintenance of NSR, a "rate control" strategy should be preferred over a "rhythm control" strategy for the long-term management of AF.  Rhythm control strategies should only be employed in patients with intolerable symptoms despite pharmacological or non-pharmacological ventricular rate control.

Reviewer Competing Interests

None declared.


1. The Planning and Steering Committees of the AFFIRM Study for NHLBI AFFIRM Investigators. Atrial fibrillation follow-up investigation of rhythm management - the AFFIRM study design. Am J Cardiol 1997;79:1198-1202.

2. The AFFIRM Investigators. Baseline characteristics of patients with atrial fibrillation: The AFFIRM study. Am Heart J 2002;143:991-1001.

3. The Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Investigators. A comparison of rate control and rhythm control in patients with atrial fibrillation. N Engl J Med 2002;347:1825-1833.

4. Lee JK, Klein GJ, Krahn AD, Yee R, Zarnke K, Simpson C, et al. Rate-control versus conversion strategy in postoperative atrial fibrillation: a prospective, randomized pilot study. Am Heart J 2000;140:871-877.

5. Hohnloser SH, Kuck KH, Lilienthal J, for PIAF Investigators. Rhythm or rate control in atrial fibrillation - pharmacological intervention in atrial fibrillation (PIAF): a randomized trial. Lancet 2000;356:1789-1794.

6. Van Gelder IC, Hagens VE, Bosker HA, Kingma JH, Kamp O, Kingma T, et al for the Rate Control versus Electrical Cardioversion for Persistent Atrial Fibrillation Study Group. A comparison of rate control and rhythm control in patients with recurrent persistent atrial fibrillation. N Engl J Med 2002;347:1834-1840.

Copyright © 2003 by the Journal of Informed Pharmacotherapy. All rights reserved.