doi:10.1016/j.jacc.2003.11.032 2004;43;1201-1208 J. Am. Coll. Cardiol.

Michael A. Brodsky, H. Leon Greene, and AFFIRM Investigators Gearoid O'Neill, Arjun Sharma, Edward Platia, Gregory K. Feld, Toshio Akiyama,

Brian Olshansky, Lynda E. Rosenfeld, Alberta L. Warner, Allen J. Solomon, (AFFIRM) study: Approaches to control rate in atrial fibrillation

The Atrial Fibrillation Follow-up Investigation of Rhythm Management

This information is current as of May 24, 2011

http://content.onlinejacc.org/cgi/content/full/43/7/1201located on the World Wide Web at:

The online version of this article, along with updated information and services, is

by on May 24, 2011 content.onlinejacc.orgDownloaded from

TFRABAEMIW

T(c

cbAB

MMW¶DCRaLB

2

Journal of the American College of Cardiology Vol. 43, No. 7, 2004© 2004 by the American College of Cardiology Foundation ISSN 0735-1097/04/$30.00Published by Elsevier Inc. doi:10.1016/j.jacc.2003.11.032

Atrial Fibrillation

he Atrial Fibrillationollow-Up Investigation ofhythm Management (AFFIRM) Studypproaches to Control Rate in Atrial Fibrillation

rian Olshansky, MD, FACC,* Lynda E. Rosenfeld, MD, FACC,† Alberta L. Warner, MD, FACC,‡llen J. Solomon, MD, FACC,§ Gearoid O’Neill, MD, FACC,� Arjun Sharma, MD, FACC,�dward Platia, MD, FACC,¶ Gregory K. Feld, MD, FACC,# Toshio Akiyama, MD, FACC,**ichael A. Brodsky, MD, FACC,†† H. Leon Greene, MD, FACC,‡‡ and the AFFIRM Investigators§§

owa City, Iowa; New Haven, Connecticut; Los Angeles, Sacramento, San Diego, and Irvine, California;ashington, DC; Rochester, New York; and Seattle, Washington

OBJECTIVES We sought to evaluate approaches used to control rate, the effectiveness of rate control, andswitches from one drug class to another in the Atrial Fibrillation Follow-up Investigation ofRhythm Management (AFFIRM) study.

BACKGROUND The AFFIRM study showed that atrial fibrillation (AF) can be treated effectively with ratecontrol and anticoagulation, but drug efficacy to control rate remains uncertain.

METHODS Patients (n � 2,027) randomized to rate control in the AFFIRM study were givenrate-controlling drugs by their treating physicians. Standardized rate-control efficacy criteriadeveloped a priori included resting heart rate and 6-min walk tests and/or ambulatoryelectrocardiographic results.

RESULTS Average follow-up was 3.5 � 1.3 years. Initial treatment included a beta-adrenergic blocker(beta-blocker) alone in 24%, a calcium channel blocker alone in 17%, digoxin alone in 16%,a beta-blocker and digoxin in 14%, or a calcium channel blocker and digoxin in 14% ofpatients. Overall rate control was achieved in 70% of patients given beta-blockers as the firstdrug (with or without digoxin), 54% with calcium channel blockers (with or without digoxin),and 58% with digoxin alone. Adequate overall rate control was achieved in 58% of patientswith the first drug or combination. Multivariate analysis revealed an association between firstdrug class and several clinical variables. There were more changes to beta-blockers than to theother two-drug classes (p � 0.0001).

CONCLUSIONS Rate control in AF is possible in the majority of patients with AF. Beta-blockers were themost effective drugs. To achieve the goal of adequate rate control in all patients, frequentmedication changes and drug combinations were needed. (J Am Coll Cardiol 2004;43:1201–8) © 2004 by the American College of Cardiology Foundation

he general approaches used to manage atrial fibrillation blockers, and digoxin have been compared in small studies(pt

RtLtppaspot

MT

AF) include: 1) ventricular rate control, and 2) rhythmontrol (attempts to maintain sinus rhythm [SR]), both in

See page 1209

ombination with anticoagulation. Despite the large num-er of patients treated with rate-controlling therapies forF, the best pharmacologic approach remains unknown.eta-adrenergic blockers (beta-blockers), calcium channel

From the *University of Iowa, Iowa City, Iowa; †Yale University School ofedicine, New Haven, Connecticut; ‡West Los Angeles Veterans Administrationedical Center, Los Angeles, California; §Georgetown University Medical Center,ashington, DC; �Sutter Institute for Medical Research, Sacramento, California;

Washington Hospital Center, Washington, DC; #University of California, Saniego, California; **University of Rochester, Rochester, New York; ††University ofalifornia at Irvine, Irvine, California; and ‡‡University of Washington and Axioesearch Corporation, Seattle, Washington. §§The AFFIRM investigators and their

ffiliations are listed in reference 32. This study was supported by the National Heart,ung, and Blood Institute (contract N01-HC-55139), National Institutes of Health,ethesda, Maryland. Mel Scheinman, MD, acted as Guest Editor for this paper.Manuscript received August 12, 2003; revised manuscript received October 29,

r003, accepted November 20, 2003.

content.onlinejDownloaded from

1–29), but no large trial has determined the optimalharmacologic method to control the ventricular responseo AF in ambulatory patients.

The Atrial Fibrillation Follow-up Investigation ofhythm Management (AFFIRM) study was a large, mul-

icenter, randomized trial sponsored by the National Heart,ung, and Blood Institute. This trial compared rate-control

herapy with rhythm-control therapy to treat AF in aopulation at high risk of stroke or death (30–34). Therimary goal was to determine which approach was associ-ted with a better survival outcome. The AFFIRM studyhowed that rate control was an acceptable, if not thereferable, option (33). A substudy of this trial offered thepportunity to evaluate and compare several drug classes forhe long-term ventricular rate control of AF.

ETHODShis investigation assessed the effectiveness of ventricular

ate-control approaches in AF in a large, general population

by on May 24, 2011 acc.org

osvd

FoaBpDqRtwargawppwteriwaaaomant

wwppt

metmtcCAdneidDawdoptapc

dfopdcc

icntocStwdwcoce

b

TC

12

1202 Olshansky et al. JACC Vol. 43, No. 7, 2004Rate Control in AFFIRM April 7, 2004:1201–8

f predominantly elderly patients. Data on all AFFIRMtudy patients assigned to the rate-control arm were re-iewed to determine the effectiveness of rate control by therug classes chosen by the treating physician.The overall AFFIRM study was approved by the AF-

IRM Steering Committee, the Institutional Review Boardf the University of Washington (Clinical Trial Center),nd by each participating site’s local Institutional Reviewoard. All patients gave informed, written consent toarticipate in the AFFIRM study.efinition of rate control. Prespecified criteria for ade-

uate heart rate (HR) control in AF are outlined in Table 1.ate control was assessed only if the patient was in AF at

he time of evaluation. If a patient in the rate-control armas in SR at a specified visit, rate control was not assessed

nd not recorded at that visit. Adequate rate controlequired the initial resting HR to be �80 beats/min. If thatoal was achieved, a standard 6-min walk test (35) or 24-hmbulatory electrocardiographic recording (Holter monitor)as required to confirm adequate rate control. The ap-roach used was left to the discretion of the treatinghysician. If both Holter monitoring and a 6-min walk testere performed, rate-control criteria had to be met for both

ests. Overall rate control required evidence for resting andxercise rate control, as defined. Tests were repeated untilate control was achieved or when the patient had a changen status or change in drug dosage. Further drug titrationas based on the results of these studies. The drug dose was

djusted and optimized by the treating physician in anttempt to achieve the preselected rate-control guidelines. Ifdequate rate control could not be achieved at a low dosagef medication, the dose was titrated upward, or a differentedication was used and titrated in an attempt to achieve an

cceptable rate response. If adequate rate control could stillot be achieved despite a maximally tolerated drug dose, thereating physician could add another drug. A medication

Abbreviations and AcronymsAF � atrial fibrillationAFFIRM � Atrial Fibrillation Follow-up Investigation

of Rhythm Management studyAV � atrioventricularHF � heart failureHR � heart rateMI � myocardial infarctionSR � sinus rhythm

able 1. AFFIRM Definition of Adequate Ventricular Rateontrol

. Average heart rate at rest �80 beats/min, and

. Either (A) or (B)A. Heart rate maximum during a 6-min walk �110 beats/min, orB. Average heart rate during 24-h ambulatory Holter monitoring

(ECG) �100 beats/min (at least 18 h of interpretable monitoring)and no heart rate �110% maximum predicted age-adjustedexercise heart rate

T

content.onlinejDownloaded from

as stopped if the patient developed adverse effects, such asorsening congestive heart failure (HF) or hypotension. If aharmacologic approach was not effective, the treatinghysician could proceed with atrioventricular (AV) junc-ional ablation, as defined in the protocol (31).

Patients were evaluated at two months using the sameethods and criteria. The resting HR was determined at

ach routine clinic visit (every four months). A 6-min walkest or Holter monitoring was to be performed at a mini-um of once a year, even if no changes had been made in

he drug regimen and no changes were noted in the patient’slinical status.hoice of drugs. Patients in the rate-control arm of theFFIRM study were treated with an AV nodal blockingrug chosen by their treating physician. Drug choice wasot randomized and was not expected to be distributedqually across all patient types or conditions. Treatmentncluded beta-blockers, calcium channel blockers, origoxin, alone or in combination.ata collection. The primary end point of this study was

dequate “overall” pharmacologic rate control at rest andith exercise while taking the initial drug class chosen, afterose optimization by the treating physician. The adequacyf rate control was determined sequentially over the entireeriod of patient follow-up, but the specific end points ofhis substudy were: 1) achievement of adequate rate control,s defined earlier, with the first drug class chosen for eachatient; and 2) the rate of switching from the first drug classhosen to another drug class.

Only drug classes (not individual drugs) administereduring the last follow-up period were reported on the dataorms submitted to the Clinical Trial Center. For example,nly the category of beta-blockers would have been re-orted, not propranolol, metoprolol, and the like. Drugoses and start/stop dates were not reported. The rate-ontrol drug classes were digoxin, beta-blockers, and cal-ium channel blockers.

Patients were classified by the drug group initially admin-stered after randomization. Rate control for the initial druglass chosen for a patient was defined as the last determi-ation of rate control while taking that drug before discon-inuing the drug or adding a new drug. Patients who crossedver to rhythm control (248 of 2,027 patients) (33) wereensored from further analysis at the time of crossover.tatistics. Baseline comparisons among all four major drug

reatment groups (digoxin alone, beta-blockers with orithout digoxin, calcium channel blockers with or withoutigoxin, and beta-blockers plus calcium channel blockersith or without digoxin) were performed using two-way

hi-square tests for homogeneity for dichotomous variablesr analysis of variance for continuous variables. Drug dis-ontinuation was assessed using Kaplan-Meier time-to-vent analysis and log-rank statistics.

Multivariate analyses evaluated the association betweenaseline clinical factors and the initial choice of drugs.

hese analyses were performed using linear model tech-

by on May 24, 2011 acc.org

nCssmstcnsptbps

dsscasi

R

TtAPed4

recbhHf

bati2a1

a(pheP

ilpcwmwIa

T

AFP

H

D

FD

F

L

L

M

*(e‡ep

hm

1203JACC Vol. 43, No. 7, 2004 Olshansky et al.April 7, 2004:1201–8 Rate Control in AFFIRM

iques generalized to categorical dependent variables (PROCATMOD in SAS, Cary, North Carolina). For these analy-

es, only main effects were included, and a stepwise backwardelection method was used. The criteria for remaining in theodel was p � 0.01, as well as the factors associated with the

uccess or failure of the first drug class chosen. The factorsested included age �65 years, gender, ethnicity, primaryardiac diagnosis, duration of qualifying AF, history of pulmo-ary disease, coronary artery disease, congestive HF, hyperten-ion, diabetes, angina pectoris, myocardial infarction (MI),revious coronary artery bypass graft surgery or revasculariza-ion procedure, hepatic or renal disease, qualifying episodeeing the first episode of AF, frequency of episodes of AF,revious drug failure, left ventricular ejection fraction, left atrialize, and baseline HR.

Differences in mortality and long-term symptoms amongrug classes could not be compared because patients werewitched to different drug classes frequently throughout thetudy. Comparisons of rate control across different groupsould not be performed for the same reasons; thus, thisrticle is primarily a descriptive report. Furthermore, thisubstudy did not have sufficient power to detect differencesn mortality among the different drug treatment groups.

ESULTS

his evaluation included the 2,027 patients randomized tohe rate-control arm of the AFFIRM study (Table 2).verage follow-up was 3.5 � 1.3 years.atient demographics and clinical characteristics. Beforentering the AFFIRM study, 53% of patients had receivedigoxin, 40% had received a calcium channel blocker, and2% a beta-blocker.Clinical characteristics are listed in Table 3, as catego-

ized by the drug class chosen initially. Significant differ-nces existed among drug classes with regard to gender,oronary artery disease, angina pectoris, MI, coronary arteryypass graft surgery, interventional coronary procedure,ypertension, cardiomyopathy, pulmonary disease, baselineR, SR at randomization, previous drug use, and ejection

raction.Initially, 485 patients (24%) were treated with a beta-

locker alone, 344 (17%) with a calcium channel blockerlone, and 315 (16%) with digoxin alone. A drug combina-ion was used frequently—a beta-blocking drug and digoxinn 292 (14%) and a calcium channel blocker plus digoxin in87 (14%). Other drugs, other combinations, and no ther-py accounted for 154 (8%), and missing data accounted for50 (7%), yielding a total of 304 patients (15%).The multivariate analysis revealed the following items to be

ssociated with the first drug class used for rate control: genderp � 0.0001), history of coronary artery disease (p � 0.0001),ulmonary disease (p � 0.0001), congestive HF (p � 0.0001),ypertension (p � 0.0001), qualifying episode being the firstpisode of AF (p � 0.005), and baseline HR (p � 0.006).

atients were likely to be prescribed the drug being admin- 7

content.onlinejDownloaded from

stered before randomization. Beta-blockers were moreikely to be used in patients having coronary disease, angina,revious MI, and previous coronary interventions. Calciumhannel blockers were more likely to be used in patientsith pulmonary disease and in females. Digoxin was usedore often in patients with cardiomyopathy and in non-hites (Table 3).

nitial rate control at rest. Rate control at rest waschieved with a beta-blocker (with or without digoxin) in

able 2. Patient Demographics

AFFIRM StudyRate-Control Arm

(n � 2,027)

ge (yrs) 69.8 � 8.9emale gender 823 (40.6)rimary diagnosisNone apparent 265 (13.1)CAD 497 (24.5)Cardiomyopathy 99 (4.9)Hypertension 1,045 (51.6)Valvular heart disease 98 (4.8)Other 23 (1.1)istoryCAD 757 (37.4)MI 332 (16.4)Cardiomyopathy 174 (8.6)Hypertension 1,434 (70.7)CHF 475 (23.4)Angina 515 (25.4)uration of qualifying AF�6 h 158 (7.8)6–12 h 217 (10.7)12–48 h 246 (12.1)2–30 days 528 (26.1)30 days to 6 months 755 (37.3)�6 months 123 (6.1)

irst episode 700 (35.8)rug failures*Any 364 (18.0)1 248 (68.1)2 85 (23.4)�3 31 (8.5)

requency of episodesFrequent/very frequent† 354 (27.2)

A size‡Normal 549 (33.3)Mild enlargement 451 (27.3)Moderate to severe enlargement 557 (33.8)

VEF§Normal 1,131 (68.6)Mild dysfunction 192 (11.6)Moderate dysfunction 107 (6.5)Severe dysfunction 81 (4.9)ean maximum ventricular rate (beats/min) 108.4 � 32.5

Before randomization. †Meaning more than one symptomatic episode per monthpatients presenting with first episodes and patients with asymptomatic AF werexcluded). These data points were recorded only after July 1, 1997 (n � 2,611).N � 3,311 echocardiograms; left atrial size was unknown in 184. §N � 3,311chocardiograms; left ventricular function was unknown in 279. Data areresented as the mean value � SD or number (%) of patients.

AF � atrial fibrillation; CAD � coronary artery disease; CHF � congestiveeart failure; LA � left atrial; LVEF � left ventricular ejection fraction; MI �yocardial infarction.

5% of patients, with digoxin alone in 68%, and with a

by on May 24, 2011 acc.org

c(bbIo8wfica(Iw

ewwwbtaRancfific

T

AFNCAMBIHCPP

BSLL

*p

1204 Olshansky et al. JACC Vol. 43, No. 7, 2004Rate Control in AFFIRM April 7, 2004:1201–8

alcium channel blocker (with or without digoxin) in 66%Table 4). Similarly, rate control at rest was achieved with aeta-blocker alone in 68% and with a calcium channellocker alone in 60%.nitial rate control during exertion. Beta-blockers (withr without digoxin) provided rate control during exertion in5%, as compared with calcium channel blockers (with orithout digoxin) in 72% and digoxin alone in 70%. Similarndings were noted with beta-blockers alone (72%) andalcium channel blockers alone (58%), although digoxinlone was associated with good rate control with exertion70%).nitial overall rate control. The overall goal of rate controlas to achieve a satisfactory HR both during rest and with

able 3. Clinical Characteristics Associated With First Rate-Con

Digoxin Alone(n � 315)

Beta-Blocker*(n � 777)

ge (yrs) 70.1 � 9.1 69.2 � 8.7emales 98 (31.1) 315 (40.5)on-Caucasian 52 (16.5) 79 (10.2)AD 112 (35.6) 331 (42.6)ngina pectoris 74 (23.5) 226 (29.1)yocardial infarction 48 (15.2) 149 (19.2)

ypass graft surgery 42 (13.3) 105 (13.5)nterventional procedure 17 (5.4) 91 (11.7)ypertension 196 (62.2) 564 (72.6)ardiomyopathy 46 (14.6) 60 (7.7)ulmonary disease 51 (16.2) 62 (8.0)revious drug useDigoxin 239 (75.9) 350 (45.1)Calcium channel blocker 50 (15.9) 133 (17.1)Beta-blocker 32 (10.2) 565 (72.7)

aseline heart rate (beats/min) 94.9 � 29.1 98.1 � 28.3inus rhythm at randomization 149 (47.3) 427 (55.0)A size normal† 88 (33.7) 194 (31.4)VEF normal‡ 159 (60.9) 432 (69.9)

With or without digoxin. †N � 1,509; normal �4.0 cm. ‡By echocardiography; Natients.

Abbreviations as in Table 2.

Table 4. Cumulative Achievement of Adequat

Initial Drug TherapyOv

(n �

BB alone 51/CCB alone 20/Digoxin 33/BB plus digoxin 41/CCB plus digoxin 31/BB plus CCB 10/BB plus CCB plus digoxin 13/None of the above* 10/BB with/without digoxin 99/1CCB with/without digoxin 63/1BB plus CCB with/without digoxin 23/

*This group consists mostly of patients for whom drug use arecorded on initial versions of the data form). Also includedas initial therapy, despite being randomized to the rate-contropatient was counted as having unsuccessful rate control in bnumber of patients with atrial fibrillation with successfulNumbers represent the last rate evaluation before any drug c

BB � beta-blocker; CCB � calcium channel blocker.

content.onlinejDownloaded from

xertion. This goal was reached in 70% of patients treatedith a beta-blocker (with or without digoxin), as comparedith 54% treated with a calcium channel blocker (with orithout digoxin) and 58% with digoxin alone. The use of aeta-blocker alone showed overall rate control similar tohat of digoxin alone (59% vs. 58%), with only 38%dequately controlled with a calcium channel blocker alone.ate control over time. Adequate overall rate control was

chieved in 58% of patients with the first drug or combi-ation (Table 4). A total of 635 patients (37%) required ahange in or the addition of another drug over the course ofve years. Changes in rate-controlling drug therapy in therst year were common: 23% of patients switched fromalcium channel blockers to beta-blockers, 19% switched

Drug Used

Calcium ChannelBlocker* (n � 631)

Beta-Blocker and CalciumChannel Blocker Combined*

(n � 138) p Value

70.1 � 9.2 69.3 � 9.1 0.2452278 (44.1) 77 (55.8) �0.000174 (11.7) 16 (11.6) 0.0343

208 (33.0) 47 (34.1) 0.0015142 (22.5) 40 (29.0) 0.022987 (13.8) 15 (10.9) 0.012151 (8.1) 19 (13.8) 0.007844 (7.0) 9 (6.5) 0.0009

449 (71.2) 114 (82.6) �0.000145 (7.1) 6 (4.4) 0.0001

133 (21.1) 17 (12.3) �0.0001

344 (54.5) 65 (47.1) �0.0001469 (74.3) 114 (82.6) �0.000196 (15.2) 108 (78.3) �0.0001

104.3 � 28.7 103.1 � 29.9 �0.0001361 (57.2) 59 (42.8) 0.0015189 (36.6) 30 (26.6) 0.8665369 (71.4) 86 (76.1) �0.0001

09; normal �0.50. Data are presented as the mean value � SD or number (%) of

tricular Rate Control

)Resting

(n � 740)Exercise

(n � 361)

9) 108/159 (68) 62/86 (72)8) 58/96 (60) 30/52 (58)8) 92/135 (68) 40/57 (70)8) 100/133 (75) 52/61 (85)0) 78/119 (66) 41/52 (79)9) 22/33 (67) 12/17 (71)6) 20/27 (74) 15/17 (88)3) 29/38 (76) 12/19 (63)0) 245/327 (75) 121/142 (85)4) 166/252 (66) 84/116 (72)2) 44/67 (66) 29/37 (78)

tion of therapy was not recorded (this information was nots group are several patients who took no rate-control drugsIf a patient did not achieve the target resting heart rate, the

e exercise and overall categories. Data are presented as thel/number of patients with atrial fibrillation evaluated (%).or addition.

trol

� 1,5

e Ven

erall361

86 (552 (357 (560 (652 (617 (517 (719 (542 (716 (537 (6

t initiain thil arm.oth thcontrohange

by on May 24, 2011 acc.org

fbepcbroRs

�pdwwfbcaAweNr

Fd one.

Fd

TD

SPCSPGGCEI“O

1205JACC Vol. 43, No. 7, 2004 Olshansky et al.April 7, 2004:1201–8 Rate Control in AFFIRM

rom beta-blockers to calcium channel blockers, and 34%egan taking either beta-blockers or calcium channel block-rs after having taken only digoxin. Over time, moreatients taking digoxin or a calcium channel blocker werehanged to another drug, as compared with patients takingeta-blockers (p � 0.0001) (Fig. 1). Overall rate control,ate control at rest, and rate control with exertion improvedver time (Fig. 2).ate control by HF status. Among patients without HF

ymptoms and with a left ventricular ejection fraction

igure 1. Time to discontinuation of rate-control therapy. Patients are cigoxin), calcium channel blocker (with or without digoxin), or digoxin al

igure 2. Rate control throughout the AFFIRM study. Time zero is the

ay of randomization. Rate control improved throughout the study.

content.onlinejDownloaded from

0.40, adequate rate control at rest and exertion, as definedreviously, was 55% with a beta-blocker (with or withoutigoxin), 51% with a calcium channel blocker (with orithout digoxin), and 71% with digoxin alone. For patientsith a history of HF symptoms or a left ventricular ejection

raction �0.40, successful rate control was observed with aeta-blocker (with or without digoxin) in 81%, with aalcium channel blocker (with or without digoxin) in 59%,nd with digoxin alone in 54% at one year.dverse drug effects. All drug classes used for rate controlere tolerated reasonably well (Table 5). Serious adverse

ffects were uncommon.umber of patients in SR. There were 1,055 of 2,027

ate-controlled patients in SR at the time of randomization.

ed by whether they started therapy on a beta-blocker (with or withoutTime zero is the day of randomization into the main AFFIRM study.

able 5. Reasons for Discontinuation of Rate-Controllingrugs

Reason forDiscontinuation

Beta-Blockers(n)

Calcium ChannelBlockers (n)

Digoxin(n)

ymptomatic bradycardia 21 16 26roarrhythmia 0 1 0ongestive heart failure 14 22 1yncope 1 2 1ulmonary 16 0 0astrointestinal 1 18 7enito-urinary 3 0 1entral nervous system 8 10 0ndocrine 3 1 2

nefficacy 38 63 25No longer needed” 21 43 51ther reasons 54 57 21

lassifi

by on May 24, 2011 acc.org

At((t1fiNdpA0whbhrcwttp

f

D

Vmcdrpd

iutMdpDruwbasr6wlActf

bbutHaouDaspptrphAfpdtwi

ocpa

mttpdCnd(cshVmTAtitBmdbeo

1206 Olshansky et al. JACC Vol. 43, No. 7, 2004Rate Control in AFFIRM April 7, 2004:1201–8

mong these 1,055 patients, 427 (40%) were initiallyreated with a beta-blocker (with or without digoxin), 36134%) were initially treated with a calcium channel blockerwith or without digoxin), and 149 (14%) were initiallyreated with digoxin alone. On follow-up, 2,162 (58%),,349 (56%), and 244 (49%) were in SR at one, three, andve years, respectively.onpharmacologic therapy. A total of 108 patients un-

erwent ablation of the AV junction and insertion of aermanent pacemaker to control the ventricular rate duringF. Before ablation, these patients received a mean of 2.4 �.7 drug classes for rate control. Combination drug therapyas not individually reported. Furthermore, a patient mightave received many different drugs and doses in a classefore ablation, but would only have been reported asaving received that drug class. For example, if a patienteceived propranolol, metoprolol, and atenolol during theourse of the study before AV junctional ablation, he or sheould only have been reported as receiving beta-blocker

herapy. The reasons for undertaking nonpharmacologicherapy were individual but generally related to failure ofharmacologic therapy.An additional 147 patients had a pacemaker implanted

or symptomatic bradycardia.

ISCUSSION

entricular rate control is an important aspect of theanagement of AF. It can improve symptoms, exercise

apacity, and cardiac function (36–40). Although it was notesigned to assess these rate-control issues specifically in aandomized method, the AFFIRM study offered the op-ortunity to evaluate how physicians use rate-controllingrugs.Rate control can be difficult, and drugs had to be changed

n approximately one-third of patients. Drug therapy wasltimately successful by the AFFIRM definitions in two-hirds of the patients, and rate control improved over time.

ore patients were switched to beta-blockers than to otherrug classes. Often, combination therapy was needed. Fewatients underwent AV junctional ablation.efinition of adequate rate control. The best criteria for

ate control in AF are not well defined. Effectiveness criteriased in the AFFIRM study are based on a consensus ofhat was considered to be reasonable. The criteria may haveeen either too lenient or too strict, or the criteria may beddressing the wrong issue (ventricular rate, instead ofymptoms). In many practice settings, the effectiveness ofate control may not be tested at rest and with frequent-min walk tests or Holter monitoring. It is not knownhether rate control based on the AFFIRM study guide-

ines has any impact on symptoms or survival.dequate rate control at rest and with exertion. Rate

ontrol at rest does not ensure rate control with exertion, buthe AFFIRM study showed that use of a functional test was

easible. It is uncertain what rate control criteria provide the i

content.onlinejDownloaded from

est compromise between hemodynamic and symptomaticenefit and drug-induced adverse effects. No drug class wasniformly effective. Differences in patients’ clinical charac-eristics led to a selection bias of one therapy over another.

owever, patients often required multiple drug iterations tochieve adequate rate control, so initial drug selections basedn disease type and illness severity are unlikely to beniformly successful.igoxin. In the AFFIRM study, digoxin was used often assingle drug. Rate control with digoxin during exercise was

imilar to that with beta-blocker use. This finding iserplexing, but it is possible that digoxin alone was givenrimarily to patients who were chronotropically incompe-ent, as evidenced by a slower initial ventricular responseate to AF. Another possibility is that this group wasreselected. Patients whose rate was controlled with digoxinad often been started on this drug before enrollment inFFIRM. It is possible that these results do not portray a

air representation of the effects of digoxin in an unselectedopulation. On the other hand, these data do indicate thatigoxin can be associated with adequate rate control whenhe drug is used clinically and selected by clinical criteria. Itas not possible to know the relationship between these

ssues and apparent rate control with digoxin.Similarly, it appeared that combinations of digoxin with

ther drug classes were associated with improved rateontrol. Such results may not be achievable in an unselectedopulation, but as prescribed clinically, rate control can bechieved with these drug combinations without undue risk.

Drug titration occurred in some patients before enroll-ent in AFFIRM. The success of initial drug therapy was

hus partly the result of drug exposure before randomiza-ion. Because no placebo control was used in this trial, it isossible that no medication would have worked as well asigoxin did to control the rate.alcium channel blockers. Verapamil, alone or in combi-ation with digoxin, can be superior to digoxin alone inecreasing the resting and exercise rate (17). In two studies6,41), verapamil combined with digoxin improved exerciseapacity. Diltiazem, alone or combined with digoxin, wasimilar to digoxin for control of the resting HR (18), butigh-dose diltiazem can have frequent adverse effects (19).erapamil and diltiazem, when combined with digoxin,ay have similar effects in patients with chronic AF (9).heir use may be problematic after MI and with HF. In theFFIRM study, calcium channel blockers were less effective

han beta-blockers at rest and with exertion. More patientsn AFFIRM were switched from calcium channel blockerso beta-blockers than vice versa.eta-adrenergic blockers. By using nadolol plus digoxin,ean and exercise HRs can be reduced, as compared with

igoxin, and the exercise time can be reduced too (23). Theeta1-selective drug, celiprolol, can improve HR control toxercise at the expense of decreased exercise capacity andxygen consumption (24). Excessive dosing may result in an

mpaired exercise HR response (25).

by on May 24, 2011 acc.org

cemotDrpcrcsatcnRddoplAs

iicOobrciScdbdiarddtmpdcwpiCts

mmbf

RU5

R

1

1

1

1

1

1

1

1

1

1207JACC Vol. 43, No. 7, 2004 Olshansky et al.April 7, 2004:1201–8 Rate Control in AFFIRM

In the AFFIRM study, a beta-blocker, used alone or inombination with digoxin, or a calcium channel blocker wasffective, and their use increased with time. Beta-blockersay help to maintain SR better than other rate-control

ptions, but it was not possible to evaluate this hypothesis inhe AFFIRM study.

rug changes. In the AFFIRM study, changing from oneate-controlling drug class to another was common. Moreatients were switched to beta-blockers than to other druglasses. Drug combinations were associated with improvedate control. Allowing changes in rate-controlling drugs andombination therapy may explain the improved rate controleen in the AFFIRM study over time. The success ofchieving rate control in the AFFIRM study may hinge onhe flexibility of the investigators to use more than one druglass for rate control. It is possible that any drug class wasot fully tested and any drug titration was not complete.elationship among patient characteristics, initial drugosing, and rate control. The results from this substudyo not allow conclusions to be drawn regarding the successr failure of any specific rate-controlling drug class based onatient- or disease-specific characteristics. Initial drug se-

ection was partly based on drug use before enrollment inFFIRM and on the patient’s clinical characteristics. These

election biases could not be controlled in our analysis.Rate control was possible for the majority of patients, and

t improved over time. Initial rate control may have beennfluenced by previous drug testing, but long-term rateontrol was achievable in this elderly population with AF.

ther methods of rate control. Radiofrequency ablationf the AV junction (with implantation of a pacemaker) cane used for rate control in AF. This approach could beeserved for patients in whom the rate cannot otherwise beontrolled. In the AFFIRM study, this method was usednfrequently.tudy limitations. The treating physician was free tohoose the first drug administered. Drug use was partlyependent on historical and clinical variables. Drugs usedefore randomization in the AFFIRM study influenced firstrug use after randomization. Neither the patient nor

nvestigator was blinded to the study drugs. Patients in SRt any time point were not included in the percent success ofate control at that time point. Rate control was analyzed byrug class, not by individual drug, and doses and start/stopates were not reported. It is important to emphasize thathere was no placebo group in this study, and there was noandatory baseline period during which drug therapy was

rohibited. Drug selection, dose selection and titration, andrug changes were discretionary. We could not exclude theontribution of natural changes in AV nodal conductionith time. Without a placebo control group or washouteriod, it was not possible to assess the mechanism ofmproved rate control over time.

onclusions. Based on rate-control criteria developed forhe AFFIRM study, most patients have an adequate re-

ponse to the available drug classes, although frequent

content.onlinejDownloaded from

edication changes may be necessary and combinationsay be needed. Based on the results of this study, beta-

lockers tend to be used more commonly over time, andewer patients abandoned this drug class.

eprint requests and correspondence: Dr. Brian Olshansky,niversity of Iowa Hospitals, 200 Hawkins Drive, Iowa City, Iowa2242. E-mail: brian-olshansky@uiowa.edu.

EFERENCES

1. Segal JB, McNamara RL, Miller MR, et al. The evidence regardingthe drugs used for ventricular rate control. J Fam Pract 2000;49:47–59.

2. Koh KK, Kwon KS, Park HB, et al. Efficacy and safety of digoxinalone and in combination with low-dose diltiazem or betaxolol tocontrol ventricular rate in chronic atrial fibrillation. Am J Cardiol1995;75:88–90.

3. Koh KK, Song JH, Kwon KS, et al. Comparative study of efficacy andsafety of low-dose diltiazem or betaxolol in combination with digoxinto control ventricular rate in chronic atrial fibrillation: randomizedcrossover study. Int J Cardiol 1995;52:167–74.

4. Stafford RS, Robson DC, Misra B, Ruskin J, Singer DE. Rate controland sinus rhythm maintenance in atrial fibrillation: national trends inmedication use, 1980–1996. Arch Intern Med 1998;158:2144–8.

5. Schreck DM, Rivera AR, Tricarico VJ. Emergency management ofatrial fibrillation and flutter: intravenous diltiazem versus intravenousdigoxin. Ann Emerg Med 1997;29:135–40.

6. Pomfret SM, Beasley CR, Challenor V, Holgate ST. Relative efficacyof oral verapamil and digoxin alone and in combination for thetreatment of patients with chronic atrial fibrillation. Clin Sci (Colch)1988;74:351–7.

7. Beasley R, Smith DA, McHaffie DJ. Exercise heart rates at differentserum digoxin concentrations in patients with atrial fibrillation. BMJ(Clin Res Ed) 1985;290:9–11.

8. Lundstrom T, Moor E, Ryden L. Differential effects of xamoterol andverapamil on ventricular rate regulation in patients with chronic atrialfibrillation. Am Heart J 1992;124:917–23.

9. Lundstrom T, Ryden L. Ventricular rate control and exercise perfor-mance in chronic atrial fibrillation: effects of diltiazem and verapamil.J Am Coll Cardiol 1990;16:86–90.

0. The Digitalis in Acute Atrial Fibrillation (DAAF) Trial Group.Intravenous digoxin in acute atrial fibrillation: results of a randomized,placebo-controlled multicentre trial in 239 patients. Eur Heart J1997;18:649–54.

1. Goldman S, Probst P, Selzer A, Cohn K. Inefficacy of ‘therapeutic’serum levels of digoxin in controlling the ventricular rate in atrialfibrillation. Am J Cardiol 1975;35:651–5.

2. Redfors A. Digoxin dosage and ventricular rate at rest and exercise inpatients with atrial fibrillation. Acta Med Scand 1971;190:321–33.

3. Hornestam B, Held P, Edvardsson N, the Digitalis in Acute AtrialFibrillation (DAAF) Trial Group. Effects of digoxin on electrocardio-gram in patients with acute atrial fibrillation—a randomized, placebo-controlled study. Clin Cardiol 1999;22:96–102.

4. Ellenbogen KA, Dias VC, Cardello FP, et al. Safety and efficacy ofintravenous diltiazem in atrial fibrillation or atrial flutter. Am J Cardiol1995;75:45–9.

5. Ellenbogen KA. Role of calcium antagonists for heart rate control inatrial fibrillation. Am J Cardiol 1992;69:36B–40B.

6. Ellenbogen KA, Dias VC, Plumb VJ, Heywood JT, Mirvis DM. Aplacebo-controlled trial of continuous intravenous diltiazem infusionfor 24-hour heart rate control during atrial fibrillation and atrial flutter:a multicenter study. J Am Coll Cardiol 1991;18:891–7.

7. Lang R, Klein HO, Weiss E, et al. Superiority of oral verapamiltherapy to digoxin in treatment of chronic atrial fibrillation. Chest1983;83:491–9.

8. Roth A, Harrison E, Mitani G, Cohen J, Rahimtoola SH, Elkayam U.Efficacy and safety of medium- and high-dose diltiazem alone and incombination with digoxin for control of heart rate at rest and duringexercise in patients with chronic atrial fibrillation. Circulation 1986;

73:316–24.

by on May 24, 2011 acc.org

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

1208 Olshansky et al. JACC Vol. 43, No. 7, 2004Rate Control in AFFIRM April 7, 2004:1201–8

9. Farshi R, Kistner D, Sarma JS, Longmate JA, Singh BN. Ventricularrate control in chronic atrial fibrillation during daily activity andprogrammed exercise: a crossover open-label study of five drugregimens. J Am Coll Cardiol 1999;33:304–10.

0. Lewis RV, McDevitt DG. Factors affecting the clinical response totreatment with digoxin and two calcium antagonists in patients withatrial fibrillation. Br J Clin Pharmacol 1988;25:603–6.

1. Lewis R, Lakhani M, Moreland TA, McDevitt DG. A comparison ofverapamil and digoxin in the treatment of atrial fibrillation. EurHeart J 1987;8:148–53.

2. Lewis RV, McMurray J, McDevitt DG. Effects of atenolol, verapamil,and xamoterol on heart rate and exercise tolerance in digitalizedpatients with chronic atrial fibrillation. J Cardiovasc Pharmacol 1989;13:1–6.

3. DiBianco R, Morganroth J, Freitag JA, et al. Effects of nadolol on thespontaneous and exercise-provoked heart rate of patients with chronicatrial fibrillation receiving stable dosages of digoxin. Am Heart J1984;108:1121–7.

4. Atwood JE, Sullivan M, Forbes S, et al. Effect of beta-adrenergicblockade on exercise performance in patients with chronic atrialfibrillation. J Am Coll Cardiol 1987;10:314–20.

5. Falk RH. Pharmacologic control of heart rate in atrial fibrillation.Cardiol Clin 1996;14:521–36.

6. Brodsky MA, Chun JG, Podrid PJ, Douban S, Allen BJ, Cygan R.Regional attitudes of generalists, specialists, and subspecialists aboutmanagement of atrial fibrillation. Arch Intern Med 1996;156:2553–62.

7. Zarowitz BJ, Gheorghiade M. Optimal heart rate control for patientswith chronic atrial fibrillation: are pharmacologic choices really chang-ing? Am Heart J 1992;123:1401–3.

8. Lundstrom T, Ryden L. Ventricular rate control and exercise perfor-mance in chronic atrial fibrillation: effects of diltiazem and verapamil.J Am Coll Cardiol 1990;16:86–90.

9. Steinberg JS, Katz RJ, Bren GB, Buff LA, Varghese PJ. Efficacy of oraldiltiazem to control ventricular response in chronic atrial fibrillation atrest and during exercise. J Am Coll Cardiol 1987;9:405–11.

0. Wyse DG. The AFFIRM trial: main trial and substudies—what canwe expect? J Interv Card Electrophysiol 2000;4 Suppl 1:171–6.

content.onlinejDownloaded from

1. The Planning and Steering Committees of the AFFIRM Study for theNHLBI AFFIRM Investigators. Atrial fibrillation follow-up investi-gation of rhythm management—the AFFIRM study design. Am JCardiol 1997;79:1198–202.

2. The AFFIRM Investigators. Baseline characteristics of patients withatrial fibrillation—the AFFIRM study. Am Heart J 2002;143:991–1001.

3. The AFFIRM Investigators. Survival of patients presenting with atrialfibrillation in the Atrial Fibrillation Follow-up Investigation ofRhythm Management (AFFIRM) study. N Engl J Med 2002;347:1825–33.

4. The AFFIRM First Antiarrhythmic Drug Substudy Investigators.Maintenance of normal sinus rhythm in patients with a history of atrialfibrillation: the effects of the first antiarrhythmic drug administered inthe AFFIRM study. J Am Coll Cardiol 2003;42:20–9.

5. Lipkin DP, Scriven AJ, Crake T, Poole-Wilson PA. Six minutewalking test for assessing exercise capacity in chronic heart failure.BMJ 1986;292:653–5.

6. Hohnloser SH, Kuck KH, Lilienthal J. Rhythm or rate control in atrialfibrillation—Pharmacological Intervention in Atrial Fibrillation(PIAF): a randomised trial. Lancet 2000;356:1789–94.

7. Levy T, Walker S, Mason M, et al. Importance of rate control or rateregulation for improving exercise capacity and quality of life in patientswith permanent atrial fibrillation and normal left ventricular function:a randomised controlled study. Heart 2001;85:171–8.

8. Ostermaier RH, Lampert S, Dalla Vecchia L, Ravid S. The effect ofatrial fibrillation and the ventricular rate control on exercise capacity.Clin Cardiol 1997;20:23–7.

9. Verma A, Newman D, Geist M, et al. Effects of rhythm regularizationand rate control in improving left ventricular function in atrialfibrillation patients undergoing atrioventricular nodal ablation. CanJ Cardiol 2001;17:437–45.

0. Wattanasuwan N, Khan IA, Mehta NJ, et al. Acute ventricular ratecontrol in atrial fibrillation: IV. Combination of diltiazem and digoxinvs. IV diltiazem alone. Chest 2001;119:502–6.

1. Lang R, Klein HO, DiSegni E, et al. Verapamil improves exercisecapacity in chronic atrial fibrillation: double-blind crossover study. AmHeart J 1983;105:820–5.

by on May 24, 2011 acc.org

doi:10.1016/j.jacc.2003.11.032 2004;43;1201-1208 J. Am. Coll. Cardiol.

Michael A. Brodsky, H. Leon Greene, and AFFIRM Investigators Gearoid O'Neill, Arjun Sharma, Edward Platia, Gregory K. Feld, Toshio Akiyama,

Brian Olshansky, Lynda E. Rosenfeld, Alberta L. Warner, Allen J. Solomon, (AFFIRM) study: Approaches to control rate in atrial fibrillation

The Atrial Fibrillation Follow-up Investigation of Rhythm Management

This information is current as of May 24, 2011

& ServicesUpdated Information

http://content.onlinejacc.org/cgi/content/full/43/7/1201including high-resolution figures, can be found at:

References

http://content.onlinejacc.org/cgi/content/full/43/7/1201#BIBLfree at: This article cites 40 articles, 17 of which you can access for

Citations

rticleshttp://content.onlinejacc.org/cgi/content/full/43/7/1201#otheraThis article has been cited by 28 HighWire-hosted articles:

Rights & Permissions

http://content.onlinejacc.org/misc/permissions.dtltables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://content.onlinejacc.org/misc/reprints.dtl

Information about ordering reprints can be found online:

by on May 24, 2011 content.onlinejacc.orgDownloaded from