|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2005; 46:1891-1899, doi:10.1016/j.jacc.2005.07.040
(Published online 19 October 2005). © 2005 by the American College of Cardiology Foundation |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio, USA
Axio Research Corporation, Seattle, Washington, USA
Department of Medicine, Division of Neurology, University of Texas, San Antonio, Texas, USA
Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada
|| Department of Medicine, University of Calgary, Calgary, Alberta, Canada
¶ Montefiore Medical Center, Bronx, New York, USA
# Kaiser Permanente Mid-Atlantic Region, Washington, DC, USA
** National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland, USA.
Manuscript received April 25, 2005; revised manuscript received July 5, 2005, accepted July 11, 2005.
* Correspondence to: Dr. Mina K. Chung, Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Desk F-15, Cleveland, Ohio 44195. Reprint requests to: Axio Research Corporation, 2601 4th Avenue, Suite 200, Seattle, Washington 98121. (Email: chungm{at}ccf.org).
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
BACKGROUND: Randomized studies, including the AFFIRM study, have failed to demonstrate survival benefits between rate-control and rhythm-control strategies for atrial fibrillation (AF). However, AF may cause functional capacity or cognitive impairment that might justify maintenance of sinus rhythm.
METHODS: Investigators of the AFFIRM study enrolled 4,060 patients with AF who required long-term therapy and who were 65 years of age or older or who had another risk factor for stroke or death. New York Heart Association functional class (NYHA-FC) and Canadian Cardiovascular Society Angina Classification were assessed at initial and each follow-up visit. From 22 randomly chosen functional status substudy sites, 245 participants underwent 6-min walk tests and Mini-Mental State Examination (MMSE) at initial, two-month, and yearly visits. Patients were assigned randomly to rate-controlling drugs, allowing AF to persist, or rhythm-controlling antiarrhythmic drugs, to maintain sinus rhythm.
RESULTS: The NYHA-FC worsened with time in both rate-control and rhythm-control groups, with no differences between groups. Presence of AF was associated with worse NYHA-FC (p < 0.0001). No differences were observed in Canadian Cardiovascular Society Angina Classification or MMSE scores. Six-minute walk distance improved over time in both study arms. On average, walk distance was 94 feet greater in the rhythm-control group (adjusted p = 0.049).
CONCLUSIONS: Modest improvement in 6-min walk distance was noted in the rhythm-control arm. Presence of AF was associated with worse NYHA-FC. No difference in cognitive function was detected.
| ||||||||||
Functional status may be assessed by standard indices, including the New York Heart Association classification of functional capacity (NYHA-FC) (12,13) or Canadian Cardiovascular Society angina classification (CCS-AC) (13,14). Six-minute walk tests have been used to assess functional status in patients with cardiovascular and pulmonary diseases, including congestive heart failure (CHF) (15–17) and AF (18). The Mini-Mental State Examination (MMSE), a standardized test of cognition (19–21), has been used in intervention studies to monitor treatment effects and in AF studies to evaluate cognition (4–8,22).
The Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study (23,24), which assigned patients with AF randomly to rate-control versus rhythm-control strategies, showed no significant difference in mortality between these approaches for AF. In light of this finding, morbidity or other clinical factors, such as functional status, may determine which strategy is most appropriate for patients with AF.
Functional status was assessed in a prospective, prespecified substudy of AFFIRM study patients to test the hypothesis that functional status is similar in rate-control and rhythm-control strategies. The substudy aimed to determine whether differences in functional status outcome measurements were associated with randomized strategy or rhythm present at testing. The NYHA-FC and CCS-AC were assessed in all patients at enrollment and all follow-up visits. Serial 6-min walk tests and MMSEs were performed in a subgroup of patients.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Patients included in the study were 
65 years of age or had another risk factor for stroke or death. Documentation was required that sufficient AF had been present to require long-term therapy. Patients were candidates for anticoagulation therapy and eligible to undergo trials with at least two drugs in both treatment strategies. The protocol was approved by each site’s institutional review board. All patients gave informed, written consent to participate in the main study and, if participating in the functional status substudy, also for this substudy.
In the rate-control strategy, beta-blockers, calcium channel blockers (verapamil or diltiazem), and/or digoxin were used at the treating physician’s discretion. Heart rate control was assessed at rest and exercise, with target heart rates of 
80 beats/min at rest and 110 beats/min during the 6-min walk tests (23–26).
In the rhythm-control strategy, antiarrhythmic drug use was selected by the treating clinician with specific guidelines for the use of antiarrhythmic drugs in the protocol (25). Patients in the rhythm-control arm underwent cardioversion as needed. Anticoagulation was used in both strategies according to published guidelines.
Functional status substudy populations. The NYHA-FC, CCS-AC, and cardiac rhythm were assessed on all AFFIRM study patients at an initial and every follow-up visit. Prospective substudies of functional status and quality of life were conducted. Of 213 clinical sites participating in the AFFIRM study, 25% were selected randomly to participate in the quality of life substudy (27). Of the quality of life substudy sites, 22 sites randomly were selected to participate in the functional status substudy. Functional status substudy patients underwent the 6-min walk test and MMSE at the initial, 2-month, 1 year, and yearly follow-up visits thereafter up to 5 years.
Measurements.
CHF and angina status
The NYHA-FC, assessing CHF symptoms of dyspnea and fatigue, and CCS-AC, assessing anginal chest pain, were defined by standard scales, with the addition of a class 0 in each scale to signify no symptoms of CHF or angina (12–14,25,28). Initial assessments were made at the time of randomization or during an initial visit within 14 days of randomization. In patients who were randomized while hospitalized, status was assessed at discharge. Both the NYHA-FC and CCS-AC were assessed at each follow-up visit (two months, four months, and every four months thereafter up to five years).
Cognitive function
The MMSE, a brief, quantified assessment of cognitive state, evaluates orientation, registration, attention, memory, and the ability to name, follow verbal and written commands, write a sentence, and copy a complex figure (19–21). The examination is scored on a 30-point scale. Higher scores indicate better performance. Tasks that could not be attempted because of physical impairment (e.g., vision, hearing, severe arthritis) were not added to the total score. Tasks that could not otherwise be attempted or completed were scored 0.
6-min walk test
Six-minute walking tests were conducted in the standardized fashion (1,29–34). Distance walked during 6 min of self-paced walking and standing heart rates at baseline and at the end of the walk were recorded.
Sample size. Sample size calculations for the functional status subpopulation were computed based on the MMSE with 80% power to detect a 10% difference in scale (three points) and on the 6-min walk test with 80% power to detect a 25 meter (82 feet) difference. For both calculations, the required sample size to detect differences of this magnitude was fewer than 50 patients.
Statistical analysis.
The primary analysis was an unadjusted intention-to-treat comparison between rate-control and rhythm-control strategies for NYHA-FC, CCS-AC, 6-min walk test distance, and MMSE scores. Additionally, intention-to-treat comparisons were performed adjusting for imbalances of baseline covariates. Data also were analyzed by the actual rhythm present at the time of evaluation in adjusted analyses not including treatment arm. Covariates used in adjusted models included prespecified subgroups (age, qualifying AF episode being the first episode of AF, gender, rhythm at randomization, duration of qualifying episode 2 days, normal left ventricular ejection fraction [LVEF], presence of coronary artery disease [CAD], history of CHF or hypertension), as well as other variables that could be related to functional status (history of stroke or transient ischemic attack, diabetes, smoking, and college education). Data are reported as mean ± standard deviation unless otherwise noted. Differences were considered statistically significant at p values <0.05.
Baseline characteristics were compared with the chi-square and Student t tests. Analyses using all visits were based on repeated measures models. These statistical analyses used generalized estimating equations, a statistical technique that combines all longitudinal data into one model and accounts for the nonindependence of within-patient observations (35). The correlation structure for within-patient observations was assumed to be exchangeable (all off-diagonal elements equal). Unadjusted analyses of all variables but NYHA-FC and CCS-AC included visit as a categorical variable so we could analyze group differences. Because of the greater number of visits with assessments of NYHA-FC and CCS-AC, unadjusted analyses of these variables used time as a continuous variable, based on years since randomization. Adjusted analyses used a similar continuous variable for the time effect and assessed change over the course of time by including two-way interaction terms.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Patient characteristics. Functional status substudy subjects were similar to the rest of the AFFIRM study population (Table 1), except that compared with subjects not in the substudy, the substudy group had a lower percentage of females (33% vs. 40%, p = 0.03) and ethnic minorities (7% vs. 12%, p = 0.04) and a higher proportion with warfarin use at baseline (93% vs. 84%, p = 0.0003), previous antiarrhythmic drug failure (25% vs. 17%, p = 0.002), and class IA antiarrhythmic drug use (14% vs. 7%, p = 0.0007). No significant differences between rate and rhythm control groups were present in the total AFFIRM study group (24) nor in the substudy (Table 1, p values all >0.05). Mean follow-up time was 3.60 ± 1.2 years for patients in the substudy and 3.48 ± 1.3 years for the entire AFFIRM study. The number and percentage of substudy patients in sinus rhythm at their final visit were 45 of 125 (36%) in the rate-control and 78 of 120 (65%) in the rhythm-control arms (chi-square p < 0.0001).
|
|
65 years (p = 0.0005), diabetes (p = 0.0017), abnormal LVEF (p < 0.0001), and CAD (p < 0.0001). History of CHF was significantly associated with worsening NYHA-FC over the course of time (p = 0.0045).
Adjusted analysis by current rhythm
Adjusted analysis of NHYA-FC across all visits including current rhythm (AF or sinus rhythm) at each visit demonstrated that current AF was significantly associated with worse NYHA-FC (p < 0.0001, 95% confidence interval 0.0770 to 0.1218) (Fig. 1B). Other significant covariates included CHF (p < 0.0001), age 65 years (p = 0.0074), male gender (p < 0.0001), diabetes (p = 0.0019), abnormal LVEF (p < 0.0001), and CAD (p < 0.0001). History of CHF was the only covariate significantly associated with worsening NYHA-FC over the course of time (p = 0.0055).
CCS-AC.
Intention-to-treat unadjusted analysis
In the AFFIRM study, only 9% of rate-control and 10% of rhythm control patients had anginal symptoms at the initial visits. There were no significant differences between the two treatment arms across all visits or at any specific visit.
Intention-to-treat adjusted analysis
Covariates significantly associated with worse CCS-AC included history of CAD (p < 0.0001), duration of qualifying episode 2 days (p < 0.0001), diabetes (p = 0.0093), CHF (p < 0.0001), and prior stroke or transient ischemic attack (p = 0.024). The CCS-AC did not significantly change over the course of time, and no variables were significantly associated with change over the course of time. Treatment arm was not significantly associated with CCS-AC across all visits or over the course of time.
Adjusted analysis by current rhythm
In adjusted models including current rhythm at each visit, presence of AF was not significantly associated with CCS-AC.
Cognitive function: MMSE.
Intention-to-treat unadjusted analysis
At the initial visit, the mean MMSE scores were higher for rate-control (28.3 ± 2.2) compared with rhythm-control (27.3 ± 2.6; p = 0.0054) (Table 2). There were no significant differences between arms at any follow-up visit or across all visits.
|
65 years old experienced a decrease of 0.28 points per year more than patients <65 years of age.
Adjusted analysis by current rhythm
Current AF at each visit was not a significant covariate (Table 2). Warfarin use was not significantly associated with MMSE scores.
Functional capacity: 6-min walk test distance.
Intention-to-treat unadjusted analysis
Average walk distance increased with time in both treatment strategies (Fig. 2A). Rhythm-control patients walked significantly more distance on average, compared with rate-control patients with differences of 118 feet more at the initial visit, 124 feet at two months and 124 feet at one year. Averaging across all visits, differences between the rate-control and the rhythm-control arms were borderline significant (p = 0.06), with rhythm-control patients walking on average 100 feet (8.5%) more than rate-control patients.
|
Adjusted analysis by current rhythm
Walk distance at each visit by current rhythm is shown in Figure 2B. In adjusted models, AF at each visit was not significantly associated with distance walked.
Functional capacity: 6-min walk test heart rates.
Intention-to-treat unadjusted analysis
Mean pre-walk heart rates were significantly higher in the rate-control compared with rhythm-control group across all visits (average difference 3.6 beats/min, p = 0.0058). Pre-walk heart rates in both groups decreased over the course of time (Table 3).
|
Intention-to-treat adjusted analysis
In adjusted analyses, pre-walk heart rates were significantly higher in the rate-control compared with rhythm-control group (average difference 3.6 beats/min, p = 0.004), but there were no significant differences between groups in post-walk heart rates.
Adjusted analysis by current rhythm
Including current rhythm in adjusted models, AF was associated with higher pre-walk (mean 11.5 beats/min higher, p < 0.0001) and post-walk (mean 12.7 beats/min higher, p < 0.0001) heart rates.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
NYHA-FC. The NYHA-FC slightly increased over time in both rate- and rhythm-control patients. Although no significant differences between treatment groups were observed in NYHA-FC, the presence of current AF at follow-up visits was significantly associated with worse NYHA-FC. Part of this paradox between the goal of maintaining sinus rhythm and the actual presence of sinus rhythm may be related to the inefficacy of antiarrhythmic drugs to achieve their goal of maintaining sinus rhythm.
CCS-AC. Angina pectoris was uncommon in AFFIRM, consistent with other studies showing chest pain to be an uncommon symptom in AF (36). Few patients in the AFFIRM study had angina at the beginning of the study. No differences in angina classification were detected between treatment groups.
Cognitive function. Adjusted MMSE scores did not differ significantly between treatment groups, in analyses based on actual rhythm present at follow-up, or after adding warfarin use to the analyses.
Age, although not previous CVA, was associated with change in scores over time. In a previous study of elderly patients with nonvalvular chronic AF, age and number of lacunar lesions were associated with lower MMSE scores (22). Previous reports of cognitive dysfunction associated with AF studied patient cohorts in which the use or dose of anticoagulation was low (22,37–39). Although there was a difference in anticoagulation use between the randomized arms of the AFFIRM study, anticoagulation use was high overall, with no significant difference in stroke rates between groups, which may explain the lack of detection of differences in cognitive function in AF patients, if these differences depended upon the relative incidences of cerebrovascular events.
Walk testing. Functional status, as measured by 6-min walk test distance, improved in both treatment arms. Walk distance across all visits was modestly although significantly longer in the rhythm-control group. These results are similar to those reported in the Pharmacological Intervention in Atrial Fibrillation (PIAF) study (18) of 252 patients randomized to rate versus rhythm-control. Walk distances were on average shorter in the AFFIRM study, but the AFFIRM study population was older (mean age 70 years) than in PIAF (mean age 60 years). Although there appeared to be a trend toward longer walk distances in patients in sinus rhythm compared with AF up through year 2, this difference was not sustained at years 3 and 4. The reason for this is unclear, particularly as prewalk and postwalk heart rate control did not appear worse on average at later years, although differences in the type of medications used in the rate- and rhythm-control groups could have been contributory.
Whether the mean difference of approximately 100 feet (8.5%) between rate-control and rhythm-control walk distance is clinically significant is uncertain. In a study of 45 older patients with CHF, the mean difference in 6-min walk distance associated with clinical differences in patient perception of worsening was 43 meters (141 feet) (17). In studies of biventricular pacing in heart failure patients, clinical benefit was reported with 6-min walk distance differences of 29 meters (95 feet) in the Multicenter InSync Randomized Clinical Evaluation (MIRACLE) trial (40) and 38.5 meters (126 feet) at three months and 42 meters (138 feet) at six months in the Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) trial (41). Among studies reporting differences as a percentage of walk distances, benefits with biventricular pacing range from 17% to 32%, corresponding to walk distance differences of 29 to 93 meters (95 to 305 feet) (39,42–44). Detection of changes in 6-min walk distances may be restricted by limited responsiveness of the measure (45).
Pre-walk heart rates were significantly higher in the rate-control arm. Moreover, current AF was strongly associated with both higher pre-walk and post-walk heart rates compared with sinus rhythm. These differences persisted over time and could have contributed to the shorter walk distances observed in patients in the rate-control group. It is unclear whether targeting to lower heart rate levels would have improved walk distances or achieved other functional status benefits in the rate control strategy. Moreover, it is uncertain whether higher heart rates or AF itself led to the differences observed in walk distance or functional class.
Study limitations. Several factors may have affected the impact of treatment strategies on functional status. Patients with frequent or severe AF symptoms may not have been enrolled, and it is unknown how more symptomatic patients would have responded to treatment strategies. It is also possible that more-stringent heart rate control in the rate control arm would have achieved better functional status.
True baseline assessment of functional status before randomization was not available in all patients because initial visits for functional status assessments were dissimilar to randomization dates in 34% of patients. At initial functional status visits, 52% of rate-control and 35% of rhythm-control patients were in AF (p = 0.012), suggesting an effect of prescribed treatment before initial functional status assessment. Thus, changes in functional status parameters from initial assessments cannot necessarily infer change from a baseline prerandomization state. This limitation may particularly be true for walk test results.
Relative insensitivity of the MMSE to detect changes in cognitive function, especially near the higher end of the scale, may have limited these results. However, this instrument was a standard measure of cognitive function at the time when the AFFIRM study was designed. Finally, in the AFFIRM study, assessments were not blinded. It is possible that knowledge of the randomized strategy or the presence of AF resulted in biased functional status evaluations.
Conclusions. The NYHA-FC gradually worsened with time in both rate and rhythm-control groups and was significantly worse if AF was present at follow-up. Six-min walk test distance improved over the course of time in patients treated with either strategy and was modestly longer in the rhythm-control patients. Higher resting pre-walk heart rates were observed in rate-control strategy patients; pre-walk and post-walk heart rates were higher in patients with current AF at testing. It is unclear whether lower heart rate targets in the rate control strategy would have improved walk distances. Cognitive function, as measured by MMSE scores, did not significantly differ over the course of time in patients treated with either strategy. Although a rhythm-control approach used to treat AF does not improve survival, this strategy might be appropriate for patients with AF who remain symptomatic despite rate control, as some incremental improvement in functional status may be achievable.
|
Footnotes |
|---|
E-mail: mailto:elainen{at}axioresearch.com
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. G. Cosio, E. Aliot, G. L. Botto, H. Heidbuchel, C. J. Geller, P. Kirchhof, J.-C. De Haro, R. Frank, J. P. Villacastin, J. Vijgen, et al. Delayed rhythm control of atrial fibrillation may be a cause of failure to prevent recurrences: reasons for change to active antiarrhythmic treatment at the time of the first detected episode Europace, January 1, 2008; 10(1): 21 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Zimetbaum Amiodarone for Atrial Fibrillation N. Engl. J. Med., March 1, 2007; 356(9): 935 - 941. [Full Text] [PDF]
|
|
|
|
 |
|
 W. J. Hall Update in geriatrics. Ann Intern Med, October 3, 2006; 145(7): 538 - 543. [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. Falk Atrial Fibrillation or Sinus Rhythm?: Controversy and Contradiction in Quality of Life Outcomes J. Am. Coll. Cardiol., August 15, 2006; 48(4): 731 - 733. [Full Text] [PDF]
|
|
|
|
|
|
E. P. Gerstenfeld Does Rhythm Control Improve Functional Status in Patients With Atrial Fibrillation? J. Am. Coll. Cardiol., November 15, 2005; 46(10): 1900 - 1901. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
