Impaired chronotropic response to physical activities in heart failure patients

Selection of patients

Forty patients were enrolled between June 2014 and April 2015. The inclusion criteria included: (1) patients had heart failure based on ESC 2012 heart failure guidelines; (2) patients’ NYHA classification was ranged from I to III and left ventricular ejection fraction (LVEF) was more than 35%; (3) patients were capable of performing moderate exercise. Patients were excluded if they (1) received a pacemaker, implantable cardioverter-defibrillator or cardiac resynchronization therapy device; (2) had persistent or permanent atrial fibrillation; (3) were incapable of exercise due to angina, heart failure decompensation, ST deviation >2 mm, or active pericarditis and myocarditis; (4) had acute myocardial infarction <45 days; (5) had uncontrolled hypertension; or (6) had other medical issues that would compound the present study.

Study procedures

All enrolled patients received 24-h Holter ECG (NorthEast Monitoring, Inc., Maynard, MA, USA) for HR and ActiGraph GT3 accelerometer (ActiGraph, Shalimer, FL, USA) monitoring for determining the magnitude of physical activities. The clinically validated Actigraph accelerometer determined counts of 10-s acceleration activities in X-, Y-, and Z- vectors with each count equal to 16 m-g per second where the g is 9.825 m•s⁻² [15–17]. The vector magnitude (VM) was calculated as the square root of the sum of the second power of X-, Y-, and Z-vector counts. Once the Holter-Actigraph monitoring was in place, patients underwent a 6 min hall walk (6MHW) and a symptom-limited treadmill exercise test. Some patients, based on physicians’ clinical judgment and patients’ willingness, did not participate in treadmill exercise testing. All enrolled patients received stable medication and there was no requirement to withhold any medication. Patients’ demographic information, echocardiography, blood pro-NT-BNP, and medications were collected at enrollment.

The symptom-limited maximal treadmill exercise test adopted the modified Naughton-Balke method [18]. Briefly, the total time of treadmill exercise testing was 14 min with two minutes at base speed with slope 0 (no inclination), two minutes at slope 1 level and speed 2, and two minutes at each of next five slopes with no change in speed. During exercise testing, 12-lead electrocardiogram and blood pressure were recorded in addition to Holter-Actigraph recording.

Assessment of chronotropic incompetence

The assessment of chronotropic incompetence was based on the achieved maximal HR during peak treadmill exercise. CI is defined as maximal HR at peak treadmill exercise that fails to reach an arbitrary percentage of APMHR, for which 70% was chosen as the criterion for diagnosis of CI in heart failure patients [12, 19, 20]. In the present study, as most patients took beta-blockers, a newly proposed equation based on HR reserve for determining chronotropic impairment was also used to assess CI, e.g., chronotropic index for patients with beta blockers (called chronotropic index-β) = (HR at peak exercise – resting HR) / (119 + (resting HR / 2) – (age / 2) – resting HR) [11].

Chronotropic response during daily living was assessed by 24-h Holter-Actigraph recording. An event of physical activity was defined once Actigraph recording reported 10 or more than 10 s of activities. Maximal Actigraph VM value in counts and HR were determined for each physical activity. Resting HR was determined by averaging HR values in the time window from 10 min after the previous activity event to the time right before the incoming activity. Thus, for each physical activity, three parameters, e.g., resting HR, maximal HR and maximal Actigraph VM, were obtained. The regression between HR values and corresponding Actigraph activity levels was performed for 24-h daily living. The overall intensity of daily physical activities was expressed as the averaged maximal VM of all physical activities during the recording window.

Statistical analysis

Continuous variables are expressed with mean ± standard deviation. The categorical variables are used for number of patients. Comparisons for continuous variables were made using t test whenever appropriate. Categorical variables were compared using the Chi square test. The correlation between the CI values by exercise test reports and those by Holter-Actigraph recording was determined using Pearson correlation coefficients with corresponding correlation equations. The simple linear regression between HR and physical activities for 24-h Holter-Actigraph recording was performed. To minimize the impact of variations among individuals on regression analysis results, HR was taken as a function of the percentage of the maximal daily physical activity. Binary logistic regression was used to detect the effect of variables in patient characteristics (age, weight, gender, coronary heart disease, cardiomyopathy, LVEF, NYHA, β-blockers) on the results (CI or no-CI) of treadmill exercise testing. A two-tailed P value of ≤0.05 was considered significant. The statistical tool SPSS was used for the statistical analyses.

Treadmill exercise testing

When β-blocker use was taken into account in the analysis using chronotropic index-β, 38.5% patients (10 / 26) failed to achieve 60% chronotropic index-β. While there were no significant differences in age and corresponding APMHR between CI and non-CI patients, CI patients achieved significantly lower maximal HR, HR reserve, and chronotropic index-β than non-CI patients (Table 1).

Figure 1a shows an example of HR and physical activity levels recorded by the Holter-Actigraph devices during treadmill exercise testing in a patient. There was a significant correlation between HR and activity levels, e.g., HR = (0.0629 × VM) + 80.999 (R = 0.7322, P < 0.0001, Fig. 1b), demonstrating an increase in HR in response to the increase in physical activity levels. The magnitude of physical activity level (VM) during peak exercise appeared lower in CI patients (VM: 380.3 ± 194.5 unit counts) than in non-CI patients (VM: 475.5 ± 142.0 unit counts, P = 0.183 vs. the CI patients).

HR analysis during 24-h daily living

An example of the relationship between HR and the VM of physical activities from a patient is presented in Fig. 2a, showing chronotropic response. The overall relationship between HR and all physical activities over the 24-h time window in all patients is presented in Fig. 2b, in which HR is significantly correlated with VM levels of physical activities. The correlation coefficients were at the same level between non-CI patients (Fig. 2c) and CI patients (Fig. 2d). However, the frequency of HR over 100 bpm during activity events was significantly higher (P = 0.0006) in non-CI patients (20% of all activity events, Fig. 2c) than in CI patients (8.8% of all activity events, Fig. 2d).

Maximal HR achieved during daily physical activities was 106.7 ± 11.8 bpm (equal to 63.9 ± 6.8% APMHR) in CI patients, which was lower than 121.8 ± 17.3 bpm (73.0 ± 9.6% APMHR) in non-CI patients (P = 0.020 vs. CI patients). The maximal HR reserve achieved during the 24-h window was 31.9 ± 7.0 bpm in CI patients, significantly smaller than that in non-CI patients (40.9 ± 12.2 bpm P = 0.035 vs. CI patients). The overall intensity (VM) of daily maximal physical activities was 333.2 ± 87.6 counts in CI patients, which was smaller than in non-CI patients (390.0 ± 118.8 counts, P = 0.178 vs. CI patients).

HR analysis during 6-min hall walk

Maximal HR achieved during 6MHW (100.1 ± 19.6 bpm) was significantly lower than maximal HR achieved during the peak of treadmill exercise testing (112.0 ± 14.8 bpm, P = 0.012). However, there was a significant correlation between the 6MHW distance and the corresponding percent APMHR achieved during treadmill exercise testing (Distance = 571.7 × %APMHR + 68.6, R = 0.615, P = 0.001). The percent APMHR achieved during 6MHW was 57.5 ± 6.3% in CI patients which was significantly smaller than that in non-CI patients (63.5 ± 6.9%, P = 0.038 vs. CI patients) and the 6MHW distance was significantly shorter in CI patients (416.7 ± 37.1 m) than in non-CI patients (486.2 ± 62.0 m, P = 0.003, Table 1).

CI assessment in patients without treadmill exercise testing

Fourteen patients did not take treadmill exercise testing based on patients’ willingness and physicians’ clinical judgment. In these 14 patients, the achieved percent APMHR was 60 ± 6.3% during daily physical activities and 56.4 ± 7.9% during 6MHW, both values of the percent APMHR were significantly lower than in non-CI patients (73.0 ± 9.6% APMHR during daily activities and 63.5 ± 6.9% during 6MHW, both P values <0.05 vs. corresponding values in 14 patients), but comparable to the values of the percent APMHR in CI patients (63.9 ± 6.8% APMHR during daily activities and 57.5 ± 6.3% during 6MHW, both P values >0.05 vs. the corresponding values in 14 patients). Moreover, the overall intensity (VM) of daily maximal physical activities was 315.5 ± 107.8 counts in VM, which appeared lower than in non-CI patients (390.0 ± 118.8 counts, P = 0.101) but not significantly different compared with CI patients (333.2 ± 87.6 counts, P > 0.05).

Clinical perspectives

Recent studies have demonstrated a high incidence of CI in the heart failure population and CI has been recognized as an independent risk factor for cardiovascular morbidity and mortality. Daily exercise is recommended for patients with chronic heart failure. However, the exercise intensity in terms of magnitude and time interval still remains uncertain and the intensity of exercise is significantly influenced by the capability of chronotropic response. Furthermore, use of β-blockers complicates chronotropic response, leading to a reduced chronotropic response [8, 9, 11, 21–23]. On the other hand, β-blockers have been a standard therapy that prolongs survival of heart failure patients in several clinical trials [24–26]. There are approximately 5.8 million patients with heart failure in the United States and 23 million worldwide [27]. It can be assumed that the majority of heart failure patients would not receive an assessment of their chronotropic response. As shown in the present study, of 40 heart failure patients, 14 (35%) patients who declined treadmill exercise testing did have low %APMHR achieved during 6MHW and daily physical activities. Thus, if a simple, ease-of-use tool can be used to screen for CI and assess impaired chronotropic response in heart failure population, more patients with impaired chronotropic response could be identified and better managed during heart failure treatment including exercise-based cardiac rehabilitation and treatment with β-blockers. Rate-responsive pacing is an effective method to relieve symptoms caused by CI [8, 28]. However, such a therapy may be underutilized in the heart failure population because exercise intolerance-related symptoms are more often recognized as heart failure symptoms instead CI-caused symptoms. Exertional dyspnea and weakness-asthenia are common symptoms in both chronic heart failure and CI. If drug therapy, such as loop diuretics, does not effectively relieve exertion-related symptoms in heart failure patients with CI, rate-responsive pacing can be considered. In addition, CI can be a manifestation of sick sinus dysfunction whose symptoms include dyspnea, asthenia, fatigue, frequent dizziness, and possible fainting or syncope with the potential of precipitating or aggravating a state of heart failure [29]. Thus, rate-responsive pacing can be used if these patients receive a pacemaker. The present study clinically not only confirmed a high CI incidence in heart failure patients, but also investigated the feasibility of using Holter-Actigraph recording as an alternate and simpler tool to identify heart failure patients with CI and impaired chronotropic response. Prospective studies are needed to evaluate clinical benefits that can be provided by the assessment of chronotropic impairment with the use of tools as the Holter-Actigraph recording.

Limitations

The present study used the symptom-limited maximal exercise test that did not have respiratory monitoring or measurement of peak oxygen consumption. Thus, the present study did not incorporate the information of metabolic and oxygen demand during exercise. Second, the present study had a relatively small sample size and enrolled heart failure patients with LVEF ranging from 35% to 53%. Moreover, exercise intolerance is one of the clinical manifestations of heart failure and use of β-blockers further complicates chronotropic response [8, 11, 21–23]. Thus, the variables for determining CI or chronotropic impairment derived from the present study need to be further confirmed in a clinical study with a large population including patients with all levels of LVEF and different medications.