Atrial fibrillation (AF) is the most common cardiac arrhythmia, affecting over 2 million Americans, and is the most frequent cause for hospitalization among disorders of cardiac rhythm (1). Atrial fibrillation is associated with an increased risk of stroke related to thromboembolism, which can be reduced with anticoagulation (2). Atrial fibrillation may also lead to tachycardia-induced cardiomyopathy with symptoms and signs of heart failure (3).

In the wake of recent neutral mortality impact of restoration and maintenance of sinus rhythm (SR), there has been an increasing focus on the control of ventricular rate combined with anticoagulation as an initial treatment approach to AF (4). On the other hand, patients remaining in AF often continue to experience disabling symptoms and poor exercise tolerance that may significantly compromise their quality of life (QOL) (5). Researchers and clinicians continue to recognize the importance of the patient’s perspective in the assessment of health care treatments (6). Thus, QOL is becoming an important focus for the assessment of treatments for AF (7- 9).

The present QOL evaluation is a substudy of the SAFE-T (Sotalol Amiodarone Atrial Fibrillation Efficacy Trial) study (10) to test the hypothesis that restoration and maintenance of SR in patients with persistent AF improves QOL and exercise performance (EP).

The details of the study (SAFE-T) design from which the data are derived have been reported previously (11). In brief, patients with persistent AF were randomly assigned double-blind to amiodarone, sotalol, or placebo after optimal anticoagulation (international normalized ratio 2 to 3). After 4 weeks of therapy, those not in SR had their AF electrically cardioverted. Once SR was achieved, anticoagulation was continued for at least 8 weeks thereafter. Recurrence of AF was documented by weekly transtelephonic monitoring. All patients were followed for at least 1 year. Patients remaining in SR were maintained on their originally assigned treatment. Those who remained in or reverted to AF after a second cardioversion were placed on open-label medications with continued follow-up to 1 year. During AF the objective was to slow the ventricular rate to 60 to 90 beats/min, which was achieved by using diltiazem, verapamil, and/or digoxin. For the present substudy, patients were classified into the SR group or the AF group according to their rhythm status at follow-up visits.

General health-related QOL was evaluated using the Medical Outcomes Study Short Form-36 (SF-36) questionnaire. It contains 8 subscales: physical functioning, physical role limitations, bodily pain, general health, vitality, social functioning, emotional role limitations, and mental health. Scores range from 0 to 100, with lower scores representing a poorer QOL. Disease-specific measures of QOL included the arrhythmia-related Symptom Checklist (SCL), which contains 16 questions and assesses both symptom frequency and severity with scores ranging from 0 to 64 for the former and 0 to 48 for the latter (12). The higher scores indicate more frequent and serious symptoms. The Specific Activity Scale (SAS) measures subjective functional capacity and consists of 20 items (13). The score of this scale ranges from 0 to 80 with a higher score representing more difficulty in carrying on physical activities. The Atrial Fibrillation Severity Scale (AFSS), a 12-item AF-specific scale, was included to evaluate subjective rating of AF symptom burden (14). The AF symptom burden score derived from AFSS is a summary score which averages the frequency, duration, and severity of AF episodes; a higher score indicates a greater AF symptomatology. The modified Naughton protocol was used for performing exercise treadmill tests as described previously (10- 11). The QOL evaluation and exercise treadmill tests were performed at baseline and at the 8 weeks and 1 year follow-up visits.

The present substudy focused on 2 groups of patients: 624 patients who completed 8 weeks of follow-up (4 weeks after cardioversion) and 557 patients who completed 1 year of follow-up. To evaluate the effect of rhythm on QOL and EP, the patients were classified into the SR group or AF group according to their rhythm status at 8 weeks or 1 year, regardless of the intermittent rhythm changes during the follow-up. The SR group included patients who converted to and remained in sinus rhythm; the AF group included patients who remained in or reverted to AF. The differences between the SR and AF groups in change of QOL scores and EP during treadmill testing from baseline to 8 weeks and from baseline to 1 year were compared using the 2-sample t test. The differences in such changes were also assessed using analysis of covariance controlling for baseline characteristics including age, body mass index (BMI), duration of AF, left ventricular ejection fraction (LVEF), left atrial dimension (LAD), presence of hypertension, study drug assignment, and success of cardioversion. Within-group changes were analyzed by the paired t test.

To determine whether maintaining SR exerted a greater effect on QOL and EP in patients with AF symptoms (defined as palpitations, syncope, lightheadedness, shortness of breath, chest pain, or fatigue) than in those without symptoms, we compared the changes from the baseline to 8 weeks and from baseline to 1 year between the SR and AF groups based on whether patients experienced these AF symptoms at baseline. The same statistical tests were performed for these analyses: The 2-sample t test was used to test differences in change between the AF and SR groups, and the paired t test was used to test within-group changes. The degree of association between QOL measures and clinical measures was calculated by Pearson correlation coefficient. All p values reported were based on 2-sided tests; p ≤ 0.05 was the criterion for statistical significance. Analyses were performed with the use of SAS software, version 8 (SAS Institute, Cary, North Carolina).

Effect size, defined as the difference between groups in the mean changes from baseline divided by the pooled standard deviation of the changes, was calculated to assess the clinical importance of differences in the general QOL measures (15). An effect size exceeding 0.2 was considered clinically meaningful (16- 19).

The baseline characteristics of patients in SR and those in AF are presented in (Table 1). Six hundred twenty-four patients were followed to 8 weeks, in which 305 were in the SR group and 319 in the AF group. Among the 557 patients who completed 1 year of follow-up, 347 were in the SR group and 210 were in the AF group. Overall, the mean age was ∼67 years and 99% were male. Patients in SR and in AF had a mean BMI of 30.5 kg/m² and 32.4 kg/m², respectively. The respective figures for the mean LAD in patients in SR and in AF were 47.0 and 49.1 mm (p < 0.001). A larger number of patients in the SR group were on the randomly assigned study drugs compared with those in the AF group (p < 0.001). On average, approximately 63% of patients experienced AF symptoms, 66% had hypertension, and 25% had ischemic heart disease. The clinical characteristics of patients followed for 1 year were similar to those followed for 8 weeks.

The QOL measurements were available for 89% of the patients (554 of 624) followed to 8 weeks and 89% (496 of 557) of those followed to 1 year. The mean changes in QOL scores from baseline to 8 weeks between patients in SR and those in AF are shown in (Table 2). At 8 weeks, there were significant improvements in physical functioning (p < 0.001), physical role limitation (p = 0.03), general health (p = 0.002), and vitality (p < 0.001); there were also significant improvements in patients with SR in symptom severity (p = 0.01), functional capacity (SAS; p = 0.003), and AF symptom burden (p < 0.001). The mean changes in QOL scores from baseline to 1 year between patients in SR and those in AF are shown in (Table 3). In those followed for 1 year, SR patients had significant improvements in general health (p = 0.007) and social functioning (p = 0.02) compared with the AF patients. Symptom frequency and severity was significantly reduced in the SR group compared with the AF group (p = 0.05 and p < 0.001, respectively), and AF symptom burden decreased remarkably among the patients who maintained SR versus those who were in AF (p < 0.001). These outcomes did not change after adjustments were made for baseline age, BMI, LAD, AF duration, presence of hypertension, the assignment of the randomized treatment, and success of cardioversion. The effect sizes comparing SR to AF on physical functioning, general health, and vitality were 0.32, 0.26, and 0.28 standard deviation units, respectively, at 8 weeks; the effect sizes for general health and social functioning were 0.27 and 0.23 standard deviation units, respectively, at 1 year.

(Figure 1) shows comparisons of the mean changes in QOL scores for SR versus AF within symptomatic and asymptomatic subgroups for patients followed to 8 weeks. For the symptomatic subgroup, patients in SR reported a significant improvement in physical functioning (p = 0.02), physical role limitation (p = 0.03), general health (p = 0.007), vitality (p = 0.003), symptom severity (p = 0.003), functional capacity (p = 0.03), and AF symptom burden (p < 0.001) compared with patients in AF. For the asymptomatic subgroup, improvement in QOL measures among patients in SR included physical functioning (p = 0.004), vitality (p = 0.05), functional capacity (p = 0.02), and AF symptom burden (p < 0.001) compared with patients in AF. Comparisons of the mean changes in QOL scores for SR versus those for AF within symptomatic or asymptomatic subgroups for patients followed to 1 year are shown in (Figure 2). The SR patients in the symptomatic subgroup had a greater improvement in physical functioning (p = 0.05), general health (p = 0.004), vitality (p = 0.05), and social functioning (p = 0.03), compared with their AF counterparts. It was noted that symptom severity and AF symptom burden decreased significantly in SR patients in the symptomatic subgroup (p < 0.001). In contrast, there were no differences in improvement in all QOL measures but 1 between SR patients and AF patients in the asymptomatic subgroup at 1 year.

For the symptomatic patients, the effect sizes comparing SR to AF on physical functioning, physical role limitation, general health, and vitality were 0.27, 0.24, 0.31, and 0.33 standard deviation units, respectively, at 8 weeks; the effect sizes for physical functioning, general health, vitality, and social functioning were 0.23, 0.36, 0.24, and 0.26, respectively, at 1 year. For the asymptomatic patients, the effect sizes for physical function and vitality were 0.41 and 0.27 standard deviation units at 8 weeks.

The EP measurements were available for 57% of the patients (358 of 624) followed to 8 weeks and 60% (332 of 557) of those followed to 1 year. (Figure 3) shows the comparisons of the mean changes in maximal exercise duration. Overall, the increase in exercise tolerance in the SR group was significantly higher than that in the AF group at 8 weeks (81.5 s vs. 33.5 s; p = 0.01) and at 1 year (74.6 s vs. 15.2 s; p = 0.02). The exercise duration was not significantly affected by amiodarone or sotalol. For the symptomatic subgroup, a greater increase in exercise duration was observed in SR patients (n = 123) versus AF patients (n = 102) at 8 weeks (80.1 s vs. 17.3 s; p = 0.006); there was a trend toward improvement in exercise performance among SR patients (n = 136) compared with AF patients (n = 69) at 1 year (72.2 s vs. 11.1 s; p = 0.08). However, significant differences in improvement of exercise tolerance between SR and AF groups were not observed in the asymptomatic subgroup at 8 weeks and 1 year. The observed differences between the SR and AF groups were not altered after the adjustment for a group of the baseline patient characteristics.

Correlations between exercise duration and QOL measures were examined to determine the nature of the association between objective measures and subjective measures of physical function in patients who maintained SR versus those who remained in or reverted to AF. The correlation coefficients (r) for exercise duration and QOL measures are presented in (Table 4). Exercise duration at 8 weeks correlated systematically with 5 SF-36 subscales in SR patients and with 3 SF-36 subscales in AF patients. At 8 weeks, EP had stronger correlation with patient-perceived physical functioning in both the AF and the SR groups (r = 0.28 and r = 0.36, respectively). At 1 year, EP was associated with all but 1 SF-36 subscale in patients remaining in SR (r = 0.19 to 0.47); however, none of the SF-36 subscales correlated with exercise duration in patients remaining in AF. There was no correlation between EP and the mental health subscale regardless of rhythm status and length of follow-up. Exercise capacity correlated well with functioning capacity, with the correlation coefficient ranging from −0.39 to −0.49 in all groups except the AF group at 1 year.

Among patients followed to 8 weeks and those followed to 1 year, approximately 93% of the patients in SR and 85% in AF completed QOL questionnaires at baseline and at follow-up visits. Seventy percent of the patients in SR and 46% in AF completed the exercise tests at baseline and at 8 weeks, and 63% of patients in SR and 54% in AF performed the exercise tests at baseline and at the 1-year visit. To assess the effect of missing data, the baseline values in patients who completed the QOL questionnaires were compared with those in patients who did not complete them. There was no statistical difference in demographic and clinical measures. Moreover, a similar comparison between patients who had the exercise test at follow-up visits and those who did not complete the test, revealed no statistical differences except for the mean BMI, which was 10% higher in patients who did not undergo the test compared with those who did complete it (p ≤ 0.01).

The data for the present substudy was derived from the SAFE-T study (10). The clinical correlates of sustained SR versus persistent AF relative to QOL, EP, and their interactions were defined. The results showed that restoration and maintenance of SR were associated with significant increases in a number of QOL measures as well as significant improvements in EP. The data also indicated that patients who were symptomatic at baseline were more likely to have improvement in QOL compared with those who were asymptomatic. A novel finding of the substudy was the strong correlation between exercise duration, an objective measure, and QOL, a subjective measure, of physical function in patients who maintained SR versus those who persisted in or reverted to AF. Thus, our data have relevance to the conversion and maintenance of SR in patients with persistent AF.

Across most QOL domains of the SF-36, patients with persistent or intermittent AF consistently experience a greatly worse QOL compared with that of healthy controls (7- 9,14,20). However, the clinical evidence that conversion and maintenance of SR in patients with AF may predictably result in improvement of QOL continues to be the subject of vigorous debate (21- 22). The debate may have stemmed largely from the results of the AFFIRM trial (4). Most of the patients in the AFFIRM trial had paroxysmal AF, and neither QOL nor EP based on the 6-minute walk test was found to be superior in patients with AF converted to SR (22). However, subsequent on-therapy analyses of the AFFIRM data in the setting of restoration and maintenance of SR revealed a mortality benefit that was deemed to be offset by the adverse effects of antiarrhythmic drugs (23) in the parent trial.

In the AFFIRM trial, 40% of the patients in the SR strategy arm were actually in AF, which may have influenced the trial outcomes. In contrast, in the PIAF (Pharmacologic Intervention in Atrial Fibrillation) study (n = 252), there was no difference in QOL between the 2 treatment groups, but EP did increase in patients restored to and maintained in SR compared with those remaining in AF (24). Nevertheless, there have been a number of recent clinical trials (10,14,20) that have demonstrated significant improvements in QOL in patients with AF converted to SR. The substudy of the RACE (Rate Control Versus Electrical Cardioversion) trial showed that, in patients with symptomatic persistent AF, restoration of SR increased QOL as measured by SF-36 questionnaires (25). The CTAF (Canadian Trial of Atrial Fibrillation), using the SF-36 and the same disease-specific instruments used in the present substudy in patients with symptomatic AF, found QOL improvements over time which were not significantly different across drug groups (amiodarone, sotalol, and propafenone) even though amiodarone was superior in maintaining normal SR (14).

The improvement in QOL on restoration and maintenance of SR is not confined solely to drug-induced maintenance of SR. Significant QOL enhancement also occurs when SR is restored and maintained after catheter ablation (12- 13,26) or surgery, as in the case of mitral stenosis (27). Therefore, our data indicate that the critical issue here is conversion and maintenance of SR from AF rather than the method whereby it is achieved in the pursuit of augmented QOL. The data from the substudy of the SAFE-T study have also indicated that patients with AF who had symptoms at baseline were more likely to have improvements in QOL on conversion to SR compared with those who were asymptomatic. At 8-week follow-up, patients who were in AF at baseline and who were now in SR had significant improvements in 7 QOL scores for the symptomatic subgroup and in 4 QOL scores for the asymptomatic subgroup. The precise basis for such a difference is unclear but may be related to the variability in the nature and severity of the underlying disease inducing the AF. Alternatively, it may possibly be due to the smaller population of subjects in the study exhibiting the phenomenon of “silent” AF. It should be emphasized that on active drug therapy 20%, and on placebo 32%, of the subjects failed cardioversion (10). The inability to be cardioverted may have been related to conditions that were responsible for the diminished performance and QOL in the fibrillation group independently of the presence of AF.

The exercise data presented here and obtained from a homogeneous population with persistent AF have not only confirmed that exercise duration increases significantly when AF patients have been converted to SR (28- 29) but also demonstrated a statistically significant association between QOL and EP for patients in SR versus those in AF. We correlated exercise duration and raw QOL scores at 8 weeks and at 1 year. It is noted that EP as an objective measure of physical function correlated not only with self-reported physical functioning, such as physical functioning, physical role limitation, general health, and functioning capability, but also with emotional and social function in patients with SR at 8 weeks and at 1 year. It bears emphasis that at 1 year the 8 measures that were significant in SR were moderately or strongly correlated with exercise performance. It is striking that exercise performance did not correlate with QOL in any patient groups with AF at 1 year. The fact that the 1-year SR data was an improvement over the degree of correlation between QOL and exercise duration at the 8-week follow-up indicates the importance of restoration and long-term maintenance of sinus rhythm in patients with persistent AF.

The SF-36 is a general QOL instrument that has been found to discriminate among various disease states but is less sensitive to changes within a particular disease state. Nevertheless, we found significant improvements in 4 of the 5 physical-related subscales at 8 weeks and in 2 of the subscales at 1 year among patients who maintained SR. In particular, significant improvements in 4 subscales at 8 weeks and in 2 subscales at 1 year were found in patients who had AF symptoms. Research in patients with chronic diseases indicates that a 3- to 5-point improvement in SF-36 subscales is considered to be clinically important (30). An effect size of at least 0.2 SD units is generally regarded as the threshold of clinical importance. Our data have shown that favorable changes in those SF-36 subscales in SR patients compared with those in AF patients were from 4.1 to 7.1 points; the effect sizes for maintaining SR on these subscales ranged from 0.23 to 0.41 standard deviation units. Thus, using either approach, the data on QOL presented here are both statistically significant and clinically relevant.

The major limitation of the substudy is that the groups compared here (AF, SR) were not constructed by randomization. However, the only fair way to determine the differences between SR and AF is to analyze the data according to rhythm status rather than using the intention-to-treat principle. Other unmeasured factors may have differed at baseline between those who were in SR at 8 weeks or 1 year and those who were in AF. It is possible that the inability to cardiovert a proportion, albeit relatively small, may have identified a different population, and this might have influenced the outcome. However, we conducted analyses which adjusted for known baseline differences and found the effects on QOL and exercise duration were not diminished after statistical adjustment. Moreover, patients who were able to complete exercise testing at baseline but not during follow-up were excluded from the analyses of changes in duration. Because patients in AF during follow-up were more likely not to complete follow-up exercise testing, our comparisons between the SR and AF group may have underestimated the benefits of SR on exercise duration. Adequate rate control in AF patients might not have been achieved in our study compared with others. This may have influenced our results. However, the optimal heart rate in AF relating to quality of life measures has not been defined.

The results of this substudy of the SAFE-T study have demonstrated that, in patients with persistent AF, restoration and maintenance of SR is associated with improvement in a number of QOL measures. These include physical functioning, general health, vitality, and social functioning as well as disease-specific QOL measures such as SCL, SAS, and AFSS. The data have indicated that patients having symptoms at baseline were more likely to derive improvement in QOL than those without symptoms. The improvements in QOL during SR were present at both short and long terms. There was a stronger correlation between QOL scores and exercise duration on treadmill in patients with SR than in those with AF. The data underscore the importance of restoration and maintenance of SR in patients with AF in the overall management of the arrhythmia.