Correlation Between Heart Rate Control During Exercise and Exercise Capacity in Patients With Chronic Atrial Fibrillation

Abstract

Background:

Rate control is an acceptable alternative to rhythm control in patients with chronic atrial fibrillation (AF).

Hypothesis:

The aim of this study of AF patients was to understand the correlation between their exercise capacity and both heart rate (HR) and HR variation index during exercise.

Methods:

The exercise capacity of 85 male patients with chronic AF was measured using a cardiopulmonary exercise test (CPX). Within this population, we compared the exercise tolerance of patients with a normal chronotropic response (maximal HR 85%–115% that of the maximal age‐predicted HR during CPX) to those whose HR response exceeded this range. Two similar comparisons were made by dividing the subject population according to (1) whether or not their HR variation index (HRVI) during CPX exceeded 10 bpm/min, and (2) whether their HR during the 6‐minute walk test exceeded 110 bpm.

Results:

Patients with an HRVI not over 10 bpm/min showed higher maximal oxygen uptake compared to patients with a higher HRVI (26.7 ± 6.1 vs 22.8 ± 4.8 mL O₂/kg/min, P = 0.002) and a longer distance walked during CPX (705.6 ± 200.3 vs 520.9 ± 155.5 m, P<0.001). No other significant influence on exercise capacity was seen. Multivariate regression analysis revealed that both the body mass index and the HRVI during CPX were independent predictors of the maximal oxygen uptake.

Conclusions:

Better HRVI control on CPX was correlated with better exercise capacity in patients with chronic AF. © 2011 Wiley Periodicals, Inc.

Jefferson Jaber, MD was supported by a fellowship grant from CNPq, Brazil. The authors have no other funding, financial relationships, or conflicts of interest to disclose.

Introduction

Heart rate (HR) control in patients with atrial fibrillation (AF) is a common therapeutic goal.1, 2, 3, 4 It is desirable to attain control of the HR both at rest and during exercise. According to the RACE II (Rate Control Efficacy in Permanent Atrial Fibrillation: a Comparison between Lenient versus Strict Rate Control II) trial, lenient HR control (HR at rest <110 bpm) was noninferior in terms of major clinical events difference compared with strict HR control (HR at rest <80 bpm and HR during moderate exercise <110 bpm).5 However, in this trial the correlation between HR and oxygen uptake during exercise was not studied. For patients in sinus rhythm, HR during exercise is linearly related to oxygen uptake. A normal chronotropic response has been defined as HR between 85% and 115% of the maximum age‐predicted HR at peak exercise. This range encompasses 2 standard deviations from the expected mean, based on patients in normal sinus rhythm.6

The correlation between HR and oxygen uptake in AF is complex due to the hemodynamic changes that accompany arrhythmia. Elevated ventricular rate increases cardiac output during exercise. In this context, strict HR control might lead to chronotropic incompetence, compromising the cardiac output.

Despite being a primary target in AF, HR control might not lead to an improvement in aerobic capacity. In fact, we believe that a disproportional increase in HR during maximal or submaximal efforts could limit oxygen uptake. Based on this hypothesis, the aim of this study was to assess the possible correlation between HR at rest, the peak HR, and the HR variation index during exercise testing with aerobic capacity in patients with chronic AF.

Methods

We studied a consecutive series of 85 sedentary male patients presenting with chronic AF. The Joint Research Ethics Board of São Paulo Federal University approved the study, and written informed consent was obtained previously from all patients. We excluded patients with a resting HR above 90 bpm (HR was assessed by means of an electrocardiographic measurement), and those with a history of previous ischemic exercise testing, limited mobility, exertional angina, heart failure (New York Heart Association [NYHA], class III‐IV), pulmonary disease, or those who used a pacemaker.7, 8

The baseline characteristics are shown in Table 1. Primary pharmacological therapy in use to achieve HR control in this population included β‐blockers, calcium‐channel blockers, and digoxin, alone or in combination.9, 10 All patients underwent a 6‐minute walk test (6MWT) according to the standards of the American Thoracic Society.11 They were encouraged to walk as far as they could in 6 minutes. HR was measured by apical auscultation for 1 minute immediately after the test.12

Table 1.

Baseline Characteristics

Age (y)	62.4 ± 10.7
LA (mm)	49.7 ± 8.5
BMI (kg/m2)	26.4 ± 3.9
LVEF	0.56 ± 0.13
Underlying cardiovascular disease (%)
Coronary artery disease	7.0
Valvular heart disease	16.5
Chagas disease	7.0
Dilated cardiomyopathy	26.0
Hypertension	23.5
None	20.0
Functional class (%)
NYHA I	47.0
NYHA II	53.0

Abbreviations: BMI, body mass index; LA, left atrium; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association.

The cardiopulmonary exercise test (CPX) was performed according to a symptom‐limited, modified Bruce protocol, which began at 1.7 mph and 0%, and thereafter followed the standard protocol. Respiratory gas exchange analysis was performed using a Vista Cx1 17560 system (VacuMed, Ventura, CA). Oxygen uptake was assessed at peak exercise (VO_{2 PEAK}) and at the 1st and 2nd ventilatory thresholds (AT₁ and AT₂, respectively) according to the American Thoracic Society.13 Peak HR during exercise was determined by means of a frequency monitor (Polar S‐410; Polar, Helsinki, Finland). Maximal age‐predicted HR was calculated using the Karvonen formula, subtracting patient's age in years from 220.14, 15, 16 Blood pressure was measured by a trained nurse using a properly calibrated and validated sphygmomanometer. Total distance walked and total work rate were also recorded.

A normal chronotropic response during CPX was considered to be an HR between 85% and 115% of the maximal age‐predicted HR, with either higher or lower values considered to be abnormal.6 HR variation index (HRVI) at peak exercise was calculated as the difference between maximal HR at peak exercise and HR at rest, divided by the exercise duration. CPX was repeated twice in 10 patients to assess test reproducibility.

On the basis of CPX, the patients were divided into 2 groups: subjects with normal chronotropic response (peak HR between 85% to 115% of the maximal age‐predicted HR) (n = 36) and subjects with abnormal chronotropic response (peak HR >115% of the maximal age‐predicted HR) (n = 46). Three patients with peak HR below 85% were excluded.

To determine the metabolic value of HRVI, those patients were also divided into 2 groups according to whether their HRVI exceeded 10 bpm/min on CPX (n = 55) or did not (n = 30) on CPX.

Based on data from the 6MWT, these patients were also divided into 2 further groups: those with HR not greater than 110 bpm (n = 35) and those with HR exceeding 110 bpm (n = 42).2 In 8 patients, technical limitations prevented measurement of HR immediately after 6MWT.

Statistical Analysis

Data are presented as mean and standard deviation for continuous variables, and as frequencies and percentages for categorical variables. Group comparison of continuous variables was performed using the unpaired Student t test.

To determine the reproducibility of CPX, the interclass coefficient correlation was calculated, using an analysis of variance for each repeated variable (VO_{2 PEAK}, maximal walking distance, and peak HR). Data were analyzed to disclose the variables linked to an increase in VO_{2 PEAK} (left ventricular ejection fraction, left atrium diameter, age, body mass index, use of β‐blocker, HR at rest, peak HR, and HR variation index on CPX). A univariate regression analysis was carried out, with VO_{2 PEAK} as the dependent variable. Multivariate analysis, based on stepwise multiple regression analysis, was then used to assess independent determinants of better VO_{2 PEAK} (>18 mL O₂/kg/min) among measurements that showed significant correlations with VO_{2 PEAK} in the univariate analysis. A P value <0.05 was considered significant.

Results

The data are presented in Tables 2 through 5. On the 6MWT, the walked distance was 544 ± 61 m, whereas the HR achieved was 114.6 ± 18.3 bpm at the end of the test. On CPX, the main parameters measured were: walking distance (586 ± 193 m), VO_{2 PEAK} (24.2 ± 5.6 mL O₂/kg/min), AT1 (20.1 ± 4.8 mL O₂/kg/min), total work (3931 ± 1805 kpm), and HR variation index (11.3 ± 3.5 bpm/min). The indication for CPX termination in all cases was the patient's request to stop due to exhaustion.

Table 2.

Exercise Capacity in Patients With Atrial Fibrillation According to Peak Heart Rate Expressed as Percentage of Maximal Age‐Adjusted Heart Rate Achieved on Cardiopulmonary Exercise Test

Variable	Group 1, Peak HR between 85–115% n=46	Group 2, Peak HR >115% n=46	P
VO2 PEAK (mL O2/kg/min)	23.6 ± 5.8	24.6 ± 5.4	0.42
AT1 (mL O2/kg/min)	19.7 ± 4.7	20.4 ± 5.0	0.48
CPX walked distance (m)	561.2 ± 175.9	602.5 ± 210.0	0.34
Total work (kpm)	3739 ± 1554	4099 ± 2028	0.38

Abbreviations: AT1, first ventilatory threshold; CPX, cardiopulmonary exercise test; HR, heart rate; VO_{2 PEAK}, peak oxygen uptake.

Table 5.

Univariate Analysis of Peak Oxygen Uptake >18 mL O₂/kg/min in Patients With Atrial Fibrillation Undergoing Cardiopulmonary Exercise Test

Variable	Odds Ratio	P	95% CI
Age	0.98	0.570	0.92–1.05
Left atrium diameter	0.97	0.334	0.90–1.04
LVEF (≥or ≤ 0.50)	2.71	0.163	0.67–10.99
Body mass index	0.75	0.006	0.62–0.92
Peak HR	1.01	0.432	0.98–1.04
HR variation index	0.68	0.006	0.52–0.90
HR at rest	0.99	0.644	0.96–1.05
Use of β‐blocker	1.72	0.520	0.33–8.87

Abbreviations: CI, confidence interval; HR, heart rate; LVEF, left ventricular ejection fraction.

The reproducibility of CPX was acceptable. All variables presented good interclass correlation coefficients: maximal HR (0.86), walking distance (0.75), and VO_{2 PEAK} (0.81). At the end of CPX, 36 patients (42.4%) had a normal chronotropic response, whereas 46 patients (54.1%) had a peak HR >115% maximal age‐predicted HR. Only 3 patients (3.5%) had a HR <85% maximal age‐predicted HR.

There was no significant difference in VO_{2 PEAK}, AT₁, walked distance, or total work between the groups with normal and elevated chronotropic response during CPX, or between the groups divided on the basis of whether HR was >110 bpm on the 6MWT (Tables 2 and 3). However, patients with an HRVI not exceeding 10 bpm/min presented a better VO_{2 PEAK}, total work, and total distance walked during CPX, as well as a trend to better AT₁, when compared with the subjects whose HRVI exceeded 10 bpm/min (Table 4).

Table 3.

Exercise Capacity in Patients With Atrial Fibrillation and Peak Heart Rate ≤110 bpm or >110 bpm on 6‐Minute Walk Test

Variable	HR ≤110 bpm, n=35	HR >110 bpm, n=42	P
VO2 PEAK (mL O2/kg/min)	24.0 ± 5.0	24.5 ± 6.4	0.71
AT 1 (mL O2/kg/min)	19.8 ± 4.4	20.3 ± 5.5	0.65
Distance walked on CPX (m)	621.3 ± 180.6	556.7 ± 213.5	0.16
Total work (kpm)	4216 ± 1674	3659 ± 1991	0.19

Abbreviations: AT₁, first ventilatory threshold; CPX, cardiopulmonary exercise test; HR, heart rate; VO_{2 PEAK}, peak oxygen uptake.

Table 4.

Exercise Capacity in Patients With Heart Rate Variation Index ≤10 bpm/min or >10 bpm/min in Patients With Atrial Fibrillation on Cardiopulmonary Exercise Test

Variable	HRVI ≤10 bpm/min, n=30	HRVI >10 bpm/min, n=55	P
VO2 PEAK (mL O2/kg/min)	26.7 ± 6.1	22.8 ± 4.8	0.002
AT1 (mL O2/kg/min)	21.4 ± 5.4	19.3 ± 4.4	0.065
Walked distance on CPX	705.6 ± 200.3	520.9 ± 155.5	<0.001
Total work (kpm)	4756 ± 1818	3480 ± 1645	0.001

Abbreviations: AT₁, first ventilatory threshold; CPX, cardiopulmonary exercise test; HR, heart rate; VO_{2 PEAK}, peak oxygen uptake.

Univariate analysis revealed a significant correlation between VO_{2 PEAK} and both body mass index and HR variation index on CPX (Table 5). Both parameters appeared as independent variables of VO_{2 PEAK} through multiple stepwise regression analysis: body mass index (odds ratio [OR], 0.80; 95% confidence interval [CI], 0.65–0.98; P = 0.035) and HRVI (OR, 0.71; 95% CI, 0.53–0.95; P = 0.02).

Discussion

HR control is frequently employed as a first‐choice therapy in AF patients. Although recent studies have established that morbidity and mortality are comparable between rate‐ and rhythm‐control therapies, exercise tolerance in AF patients is worse when compared to sinus‐rhythm subjects.4, 17 Moreover, reestablishing normal sinus rhythm has been shown to improve cardiac output.18, 19 HR during exercise is linearly related to oxygen uptake in sinus rhythm. However, the relationship between HR and exercise capacity is not clear in AF, due to the complex interactions between the reduction of atrial function, ventricular diastolic filling time, and HR irregularity.

In the present study, 54% of the patients studied had a chronotropic response to exercise with a peak HR exceeding 115% of the maximal predicted age‐adjusted HR. However, we observed no difference in the exercise capacity of this group when compared with those whose peak HR was in the normal range. Therefore, in AF patients an elevated peak HR cannot be considered an exaggerated chronotropic response, but rather an adequate response to maintain a proper cardiac output. The fact that the vast majority of patients exceeded the maximal age‐predicted heart rate before reaching exhaustion during exercise suggests that the their peak oxygen uptake would have been underestimated had the test been prematurely terminated on the basis of heart rate.

This finding is in disagreement with a previous study, in which Ueshima et al reported that peak HR was a predictor of peak oxygen uptake in AF patients.20 Similarly, Atwood et al demonstrated that there was a lower maximal and anaerobic threshold in patients using celiprolol to achieve HR control when compared to patients using placebo.21 However, in the present study neither peak HR nor the use of β‐blocker therapy were determinants of peak oxygen uptake (Table 6).

Our results are in agreement with those of a study using data from AFFIRM (The Atrial Fibrillation Follow‐Up Investigation of Rhythm Management), which found no difference in terms of distance walked among the quartiles of HR obtained at the end of the 6MWT.12 When we analyzed 1 of the criteria used by the AFFIRM study to achieve HR control in exercise (HR exceeding 110 bpm on 6MWT), we also found no difference in exercise capacity or distance walked between groups.2 It must be noted that a limitation of the 6MWT to analyze HR control is that patients determine their own speed, and a difference in work intensity over the 6 minutes can influence not only HR but also the total distance walked. However, when we employed the Bruce modified protocol in CPX, we established the same workload for all patients in each stage of the test.

Stepwise multiple regression analysis demonstrated that only the body mass index and the HR variation index were predictors of peak oxygen uptake. Furthermore, an HRVI not exceeding 10 bpm/min during CPX was correlated with a better peak oxygen uptake and longer walking distance on CPX. Therefore, we believe that clinicians should not target the peak HR at the end of the exercise, because it is not an independent predictor of maximal oxygen uptake. The ideal is to achieve gradual HR control according to workload, because this parameter was responsible for a better exercise capacity in this population.

We were not able to establish whether better HR variation index control is a cause or a consequence of the altered hemodynamic status of these patients. However, all patients were in NYHA functional class I and II, and their average left ventricular function (0.56 ± 0.13) was preserved. Therefore, it is unlikely that these patients presented higher HR variation index due to clinical decompensation.

Study Limitations

A limitation of this study is that the analyzed groups were not randomized, as the group assignments depended on the HR response obtained in each test. Another limitation lies in the diversity of drugs that had been prescribed by the patients' various physicians to achieve HR control. The great majority of patients (68%), however, used β‐blockers (isolated or in association), most likely because these are considered effective and safe medications and are also of benefit for other concomitant cardiovascular disease.9 The finding in this study that β‐blockers did not appear to have any influence in the peak oxygen uptake should remove 1 potential contraindication to their use in AF patients. Another potential limitation of our study is that to keep the population as homogeneous as possible, and in view of the demonstrated differences in the exercise performance between males and females, only male patients were studied.

Conclusion

Our results indicate that among AF patients, an HR variation index that does not exceed 10 bpm/min is associated with a better peak oxygen uptake and total distance walked during cardiopulmonary exercise testing. Improved HR variation index may be a useful therapeutic target when the objective is to improve the exercise capacity of patients with AF.