Author + information
- Received November 24, 2008
- Revision received February 17, 2009
- Accepted February 24, 2009
- Published online June 23, 2009.
- Ashish Y. Patwala, MB ChB⁎,
- Paul R. Woods, PhD†,
- Lisa Sharp, PhD†,
- David F. Goldspink, DSc†,
- Lip B. Tan, PhD‡ and
- David J. Wright, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. David J. Wright, The Liverpool Heart and Chest Hospital, Thomas Drive, Broadgreen, L14 3PE Liverpool, United Kingdom
Objectives We evaluated the benefits of additional exercise training after cardiac resynchronization therapy (CRT).
Background Cardiac resynchronization therapy results in improved morbidity and mortality in appropriate patients. We hypothesized that a structured exercise training program in addition to CRT would maximize the improvements in exercise capacity, symptoms, and quality of life (QOL).
Methods Fifty patients referred for CRT were recruited. Patients were assessed before and 3 and 6 months after CRT. Functional class and QOL scores were recorded, and exercise tests were performed with hemodynamic measurements. Peak lower limb skeletal muscle torque was measured during extension, and echocardiography was undertaken at each visit. At 3 months, patients were randomized with a simple sealed envelope method to exercise training (n = 25) or control group (n = 25). The exercise group underwent an exercise program consisting of 3 visits/week for 3 months. Paired sample ttests were used to look for in-group differences and independent sample ttests for between-group differences.
Results Three months after CRT there were significant improvements in all functional, exercise hemodynamic, and echocardiographic measures. After randomization the exercise group showed further significant improvements in functional, exercise hemodynamic, and QOL measures compared with the control group. There were also significant in-group improvements in peak skeletal muscle function and ejection fraction that did not reach statistical significance on intergroup analysis.
Conclusions Exercise training leads to further improvements in exercise capacity, hemodynamic measures, and QOL in addition to the improvements seen after CRT. Therefore, exercise training allows maximal benefit to be attained after CRT.
Chronic heart failure (CHF) is an increasingly common condition. In suitable patients cardiac resynchronization therapy (CRT) leads to an improvement in exercise capacity, functional class, peak oxygen consumption (VO2), hemodynamic measures, and quality of life (QOL) scores (1,2). More recently a significant improvement in mortality has also been demonstrated (3). These benefits primarily occur due to an improvement in the central cardiovascular function of the heart. The improvement in cardiac function leads to reverse remodeling with a reduction in the left ventricular size and an improvement in the ejection fraction (EF) (4). However, despite these improvements, all of the major trials have demonstrated a 20% to 30% nonresponder rate. This might reflect poor lead placement due to limited venous anatomy, poor pacing parameters due to proximity of myocardial damage, or diaphragmatic stimulation. We postulate that the nonresponder rate might also reflect the fact that CRT only improves the central hemodynamic status of the patient and a significant proportion of patient symptoms are due to peripheral factors.
Recognition of peripheral abnormalities has led to the “muscle hypothesis of cardiac failure.” This proposes that inadequate skeletal muscle perfusion generates symptoms and, via muscle ergoreceptors, leads to increased neurohormonal activation, peripheral vasoconstriction, and therefore disease progression (5). It is now widely accepted that in patients with chronic heart failure a period of exercise training can lead to improvements in exercise capacity, peak VO2, QOL, and in 1 trial, mortality (6,7). The mechanism behind the improvement in peak VO2is likely to be a combined improvement in central cardiac output and peripheral oxygen extraction. However, the majority of trials evaluating exercise training have focused on patients with mild heart failure (New York Heart Association [NYHA] functional class I to II). Most recently the HF-ACTION (Heart Failure and A Controlled Trial Investigating Outcomes of Exercise Training) trial showed that exercise training was safe in patients with more severe heart failure (NYHA functional class II to IV) and led to an improvement in peak VO2. Despite this there were no significant improvements in mortality or morbidity in this group (8). Patients suitable for CRT are of a comparable functional status to those in the HF-ACTION trial. The combination of exercise training and CRT has not been well-investigated, with only 1 small-scale study suggesting possible benefit (9).
Our hypothesis was that patients suitable for CRT would demonstrate both central and peripheral decompensation due to a prolonged inability to exercise. Although CRT would improve cardiac output and exercise capacity, it would not necessarily improve skeletal muscle performance. Therefore, we hypothesized that exercise training in addition to CRT would lead to further improvements in peak VO2by improving skeletal muscle performance.
Fifty patients awaiting CRT at our center between July 2004 and June 2006 were invited to participate in this study. Standard clinical guidelines were used to assess suitability for CRT (10). All patients were in NYHA functional class III to IV. All patients were receiving optimal medical therapy for CHF and had been in a stable condition for more than 1 month (no hospital stay for heart failure, no change in medication, and no change in NYHA functional class). All had a QRS width >120 ms and a left ventricular EF <35%. Exclusion criteria included noncardiac physical limitation (e.g., arthritis, pulmonary disease, myocardial infarction within the last 3 months), coronary artery bypass grafting or percutaneous transluminal coronary angioplasty within the last 3 months or planned for the future, and significant untreated valvular heart disease. The study was approved by the Liverpool local research ethics committee, and informed written consent was obtained from all patients.
All patients underwent cardiopulmonary exercise testing, a resting transthoracic echocardiogram, QOL assessment with the Minnesota Living With Heart Failure (MLWHF) questionnaire, and measurement of peak skeletal muscle torque during extension. All measurements were taken at baseline and repeated 3 and 6 months after CRT. Before the study a separate familiarization test was performed and the results discarded.
After the 3-month assessment patients were randomized (as described in the following text) into either an exercise group (n = 25) or a control group (n = 25). The exercise group underwent a program of physician-supervised exercise training consisting of 3 30-min visits/week. Each session consisted of 10 min treadmill walking followed by 10 min cycling and then a further 10 min treadmill walking. The intensity was 80% of the peak heart rate (HR) achieved at the 3-month test for the first 4 weeks, 85% for the next 4 weeks, and 90% for the final 4 weeks. This was performed at a nonclinical setting (Liverpool John Moores University) away from the base hospital. Patients were supervised by an advanced life support-trained physician not involved in the CRT implant or follow-up at the base hospital (Liverpool Heart and Chest Hospital). Clinical follow-up was organized at the base hospital for both groups. The exercise group was not provided with any specific instruction or guidance to perform exercise outside of the study. The control group was given no specific advice on exercise training and underwent no supervised training. No change in cardiac medication occurred as part of the study protocol.
Randomization occurred at 3 months. It was not performed at baseline, to eliminate the potential for bias due to pre-emptive training in the exercise group. Randomization was performed with a simple sealed envelope method. Fifty sealed envelopes, 25 with exercise training and 25 with control, were placed into a box. After the 3-month assessment for each patient, an envelope was picked by a person not involved in the study.
Cardiopulmonary exercise tests
These were conducted as described in previous reports (11,12). Exercise testing was conducted over 2 consecutive days. On the first day resting cardiac output (CO) was measured with the carbon dioxide (CO2) rebreathing technique of Collier (13) and calculated with the indirect Fick method. The concentrations of gases in the rebreathe mixture were 10% CO2, 35% O2, and 55% nitrogen, and the volume was twice that of the subject's resting tidal volume. Before each CO measurement systolic and diastolic blood pressure were measured with a manual sphygmomanometer. Mean arterial pressure (MAP) was calculated from the standard equation: MAP = (systolic pressure + 2 × diastolic pressure)/3 (14). Mean CORestand MAPRestwere calculated from triplicate measurements, separated by 4 min to allow for CO2washout from the circulation.
After this a symptom-limited incremental exercise test was performed on a Cosmos treadmill (Cosmos, Nussdorf-Traustein, Germany) with the modified Bruce protocol to measure the VO2, peak CO2production, respiratory exchange ratio (RER), HR, and exercise duration, with the MedGraphics CPX/D system (Medical Graphics Corporation, St. Paul, Minnesota). At peak exercise blood pressure was measured, a CO2rebreathing maneuver was performed to determine peak CO with the Defares (15) method, and gas concentrations of 5% CO2,35% O2, and 60% nitrogen were employed. Ventilatory anaerobic threshold was calculated with the V-slope method. On the second day a second exercise test was performed to the same VO2and peak CO2production as on the first day. At peak exercise a second peak CO measurement was made with the aforementioned protocol. The 2 peak CO measurements were averaged to obtain a mean COPeak. Cardiac power output (CPO) was calculated at rest and peak exercise from the averaged CO (l/min) and MAP (mm Hg) with the formula:where Kis the conversion factor (2.22 × 10−3) into Watts.
A resting transthoracic echocardiogram was performed at each visit after the patient had been resting for a minimum of 30 min. Left ventricular end-diastolic dimension (LVEDD) was measured in the parasternal long axis view. Ejection fraction was measured with the biplane Simpson's method via the apical 4- and 2-chamber views.
Skeletal muscle function was assessed in both legs during 2 sets of 5 knee extensions at 60°/s with a Biodex isokinetic dynamometer. Testing was performed after a minimum of 30 min rest.
Implantation and programming of atrio-biventricular pacemaker
Implantation was performed according to standard technique of atrio-biventricular pacemaker implantation. The pacemaker was programmed to a base rate of 60 beats/min and the upper limit 85% of the age–sex maximum predicted HR. Atrioventricular delays were programmed to standard settings (120 ms paced, 100 ms sensed). No V-V offset was used during this study. Biventricular VVIR pacemakers were implanted in patients with atrial fibrillation. For the purposes of this study patients in atrial fibrillation did not undergo atrioventricular node ablation before or during the study period. Patients received CRT-P and CRT-D devices as appropriate, and no distinction was made between them.
All continuous variables are presented as mean ± SD. Discrete variables are presented as modes. For in-group analyses paired sample ttests were used to look for statistical significance in continuous variables, and chi-square tests were used for discrete variables. To look for differences between the exercise and control groups at 6 months, change scores were calculated and independent sample ttests were used. Analyses were carried out with SPSS version 12.0.1 for Windows (SPSS Inc., Chicago, Illinois). Statistical significance was set at the 5% level.
At baseline the mean age was 64.4 years with a 92% male preponderance. Patients had a mode NYHA functional class of 3 and a mean peak VO2of 16.12 mls/kg/min. Mean electrocardiographic QRS duration was 160 ms, and echocardiography provided a mean LVEDD of 7.07 cm with an EF of 24.0%. Thirty-three of the 50 subjects were in sinus rhythm (66%). Patients were well-medicated at baseline. Full baseline data is shown in Table 1.
Three months after CRT
Results from baseline and 3 months are seen in Table 2.There was a significant increase in peak VO2from 16.12 ± 3.44 mls/min/kg to 18.41 ± 3.56 mls/min/kg (p < 0.001). There were also significant improvements in NYHA functional classification from a mode of 3 to 2 (chi-square p < 0.001) and exercise duration from 374 ± 188 s to 562 ± 220 s (p < 0.001).
During this period there were also significant improvements in exercise hemodynamic measures. Peak CPO increased from 2.48 ± 0.79 W to 3.20 ± 0.78 W (p < 0.001), and cardiac reserve improved from 1.85 ± 0.68 W to 2.54 ± 0.70 W (p < 0.001). The maximum RER achieved improved from 1.00 ± 0.12 to 1.08 ± 0.09 (p < 0.001); however, the percentage of peak VO2at the anaerobic threshold remained unchanged at this point (71.8 ± 17.2% vs. 69.1 ± 8.9%; p = 0.319).
Echocardiographic measures showed a significant reduction in the LVEDD (7.07 ± 0.87 cm vs. 6.64 ± 0.77 cm; p < 0.001) and an increase in the EF from 23.7 ± 8.7% to 32.4 ± 6.2% (p < 0.001). Quality of life, as measured by the MLWHF score, also improved significantly (61.8 ± 17.7 to 31.8 ± 19.5; p < 0.001).
Peak skeletal muscle function (Newton Meters [N-M]) showed no significant change in either leg at 3 months.
At 3 months the patients were randomized to either exercise training or control. The 2 groups were well-matched at the point of randomization. There were no significant differences in any variables as seen in Table 3.
Six months after CRT
After randomization the exercise group showed a further significant improvement in peak VO2(18.74 ± 3.40 mls/kg/min vs. 20.10 ± 3.84 mls/kg/min; p = 0.011) (Fig. 1),NYHA functional class (2 vs. 1; p < 0.001) (Fig. 2),exercise duration (581 ± 226 s vs. 752 ± 179 s; p < 0.001), and MLWHF score (34.6 ± 22.3 vs. 26.2 ± 20.5; p = 0.013). The control group showed a significant improvement in exercise duration (542 ± 215 s vs. 572 ± 220 s; p = 0.028) but no improvements in peak VO2, NYHA functional class, or MLWHF score. Intergroup analyses confirmed that the improvements in NYHA functional class (mean change exercise: 1 vs. control: 0; p = 0.001), exercise duration (mean change exercise: 171 ± 174 s vs. control: 30 ± 64 s; p < 0.001), peak VO2(mean change exercise: 1.37 ± 2.49 mls/kg/min vs. control: −0.01 ± 1.49 mls/kg/min; p = 0.022), and MLWHF scores (mean change exercise: −8.4 ± 15.6 vs. control: 1.6 ± 15.3; p = 0.026) were significantly better in the exercise group.
Peak CPO (3.27 ± 0.82 W vs. 3.76 ± 1.01 W; p < 0.001) and cardiac reserve (2.57 ± 0.72 W vs. 3.05 ± 0.93 W; p < 0.001) were both improved in the exercise group alone. The maximum RER achievable was significantly increased in the exercise group (1.10 ± 0.09 vs. 1.14 ± 0.08; p = 0.001), and there was a significant reduction in the percentage of peak VO2at the anaerobic threshold (67.5 ± 8.3% vs. 62.1 ± 10.0%; p = 0.010) (Fig. 3).The control group showed no change during the same period. Intergroup analysis confirmed that the improvement in peak CPO (mean change exercise: 0.49 ± 0.61 W vs. control: 0.06 ± 0.64 W; p = 0.019), cardiac reserve (mean change exercise: 0.48 ± 0.57 W vs. control: 0.01 ± 0.63 W; p = 0.008), and RER (mean change exercise: 0.04 ± 0.06 vs. control: −0.01 ± 0.07; p = 0.008) were significantly greater in the exercise group. There was a trend toward a difference between the 2 groups when assessing the change in the percentage of peak VO2at the anaerobic threshold, but this did not reach significance (mean change exercise: −5.4 ± 9.7% vs. control: −0.8 ± 10.3%; p = 0.109).
At 6 months the exercise group showed a trend toward an improvement in LVEDD, but this did not reach statistical significance (6.64 ± 0.89 cm vs. 6.40 ± 0.53 cm; p = 0.164). The improvement in the control group did achieve significance (6.57 ± 0.56 cm vs. 6.34 ± 0.57 cm; p = 0.006). Conversely the EF significantly improved in the exercise group (32.8 ± 6.2% vs. 37.3 ± 5.4%; p = 0.010) but not in the control group (32.6 ± 7.0% vs. 35.0 ± 7.2%; p = 0.132). Intergroup analyses of the changes showed no significant difference between the 2 groups in either LVEDD (mean change exercise: −0.20 ± 0.74 cm vs. control: −0.23 ± 0.38 cm; p = 0.963) or EF (mean change exercise: 4.5 ± 8.2% vs. control: 2.5 ± 8.0%; p = 0.371).
After the period of exercise training the peak skeletal muscle torque during extension (N-M) was significantly improved in the exercise group (right: 135.0 ± 50.3 vs. 144.8 ± 57.6; p = 0.009; left: 135.0 ± 50.4 vs. 143.6 ± 53.1; p = 0.004). The control group showed no significant change during this period. Intergroup analyses showed no significant difference between the 2 groups in either right-sided extension (N-M) (mean change exercise: 9.8 ± 17.1 vs. control: 3.8 ± 9.7; p = 0.133) or left-sided extension (N-M) (mean change exercise: 8.7 ± 13.8 vs. control: 5.0 ± 11.8; p = 0.309).
No patients in the exercise group had any complications from the exercise training, and there were no problems with cardiac arrhythmias. Full results are given in Table 3.
Three months after CRT we demonstrated significant improvements in all functional and hemodynamic measures. These improvements were of a similar magnitude and over a similar time scale as those seen in other multicenter trials (16,17). At 3 months there was a 14% improvement in the peak VO2and an approximately 30% improvement in the peak CPO and cardiac reserve. Ejection fraction also showed a 37% improvement. In addition to these improvements the MLWHF score improved by almost 50%. The baseline peak VO2was 16.12 mls/kg/min, which is slightly higher than has previously been reported in CRT trials (1,18). However, the correlation between peak VO2and NYHA functional class is known to be poor (19). Peak VO2showed a significant improvement at 3 months without a significant change in the percentage of VO2at the anaerobic threshold. This suggests that the improvement in peak VO2was due to an equal improvement in the capacity to perform aerobic and anaerobic exercise. After CRT the increased RER reflects an ability to exercise to a greater level of physiological stress. This was not a familiarization response because a separate exercise test had been performed before the study and the results discarded. The results at 3 months support the theory that CRT improves function due to enhanced central cardiac function alone. Our study confirms the absence of any peripheral skeletal muscle changes 3 months after CRT alone.
At the point of randomization there were no significant differences between the 2 groups. Randomization was delayed until the 3-month stage, to limit bias. Had patients been aware at an earlier stage that they were going to be in the exercise group, there would have been the potential for them to self train. This would have influenced the 3-month results and made any comparison between the 2 groups less reliable. The exercise training group underwent a graduated intensity exercise program over 3 months, whereas the control group was given no specific advice relating to exercise. We recognize that this exercise training program was more intensive than the usual cardiac rehabilitation offered to patients after myocardial infarction or coronary revascularization. Previous trials looking at exercise in CHF have had varying intensity and frequency. In the ExTraMATCH (Exercise Training Meta-Analysis of Trials in Patients With Chronic Heart Failure) meta-analyses, the trials varied from 2 to 7 times/week and 50% to 80% of the peak HR (7), and in the HF-ACTION trial the exercise program was 30 min 3 times/week in the supervised stage. However, by the end of the trial the average duration of exercise was only 50 min/week (8). Because our trial was based over a 3-month training period, it was necessary to have a higher intensity of exercise. Hence we started at 80% and went up to 90% of peak HR. Despite the high intensity it was well tolerated and no patients in the exercise group had any complications from the exercise training.
At 6 months, those patients undertaking exercise training demonstrated further significant improvements in NYHA functional class, exercise time, peak VO2, peak CPO, cardiac reserve, RER, and MLWHF score that were not mirrored in the control group. Although in-group analysis showed a significant improvement in skeletal muscle function with exercise training, this was not supported by significant intergroup changes compared with the control group. This might reflect the relatively small number of patients in each group. It might equally be due to the frequency or duration of the training program. Had the patients exercised more frequently or for a longer period, the differences in improvement in peak skeletal muscle function might have achieved significance. Nevertheless the overall changes at the end of the study were greater in the exercise group for most variables (Fig. 4).The overall improvement in exercise duration seen in the control group (53%) was comparable to previously published data (18). However, the introduction of exercise training increased the improvement to 101%. This level of improvement has never been described after CRT alone. Peak VO2, known as a powerful predictor of prognosis (20), was improved by 12% in the control group compared with 25% in the exercise group. Again, this level of improvement has never previously been reported. In all hemodynamic measures the addition of exercise training doubled the percentage improvement at 6 months when compared with the control group. The exercise group also showed more substantial improvements in both right- (13%) and left-sided (16%) peak skeletal muscle function. The control group showed only a 2.4% improvement in right-sided and a 0.2% improvement in left-sided extension during the same period.
The likely explanation behind these overall changes is that CRT alone improves functional capacity and QOL by enhancing cardiac function, and exercise training improves functional capacity further by enhancing skeletal muscle and cardiac function.
This study used a randomized, controlled protocol. A further control group suitable for CRT but randomized to exercise training alone would have improved the methodology. However, in view of the CARE-HF (Cardiac Resynchronization-Heart Failure) study (8)—which showed a mortality benefit from CRT—it was considered unethical to withhold CRT even temporarily. Overall the dataset was relatively small, and therefore some caution should be exerted when interpreting the results. However, all of the principle measures achieved highly significant improvements. In view of the relatively small number of patients involved in this study, it was not appropriate to look for benefits in mortality or CHF hospital stay. These end points could only be accurately assessed by larger-scale trials.
By the very nature of exercise training, it is impossible to blind the patient to whether they are in the exercise or control group. We attempted to limit any potential impact by delaying randomization until the 3 month stage. We also tried to minimize bias from actual training visits in the exercise training group by performing the exercise training in a nonclinical setting and by using a physician not involved in the pacemaker implant or follow up.
Clinical implications of this study
In view of the significantly improved outcomes in the exercise group, it would be reasonable to suggest that exercise training should be offered to all patients after CRT. This would ensure the best possible outcome from CRT. The exercise program used in this study was a practical regimen that could be used in clinical practice.
CRT is an effective treatment for suitable patients. The addition of exercise training significantly enhances the benefits seen by improving both the central cardiac function and the peripheral skeletal muscle function. Exercise training would provide only a small additional cost to the overall cost of CRT, and therefore we feel that it would be justified to offer this to all patients after CRT.
Dr. Tan has been an invited lecturer for and received educational sponsorship from pacemaker manufacturers. Dr. Wright has given educational lectures on behalf of Medtronic and Boston Scientific.
- Abbreviations and Acronyms
- chronic heart failure
- cardiac output
- cardiac power output
- cardiac resynchronization therapy
- ejection fraction
- heart rate
- left ventricular end diastolic dimension
- mean arterial pressure
- Minnesota Living with Heart Failure
- New York Heart Association
- quality of life
- respiratory exchange ratio
- oxygen consumption
- Received November 24, 2008.
- Revision received February 17, 2009.
- Accepted February 24, 2009.
- American College of Cardiology Foundation
- Bristow M.R.,
- Saxon L.A.,
- Boehmer J.,
- et al.,
- Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) Investigators
- St. John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Clark A.L.,
- Poole-Wilson P.A.,
- Coats A.J.
- Piepoli M.F.,
- Davos C.,
- Francis D.P.,
- Coats A.J.,
- ExTraMATCH Collaborative
- O'Connor C.M.,
- Whellan D.J.,
- Lee K.L.,
- et al.
- Gregoratos G.,
- Abrams J.,
- Epstein A.E.,
- et al.
- Cooke G.A.,
- Marshall P.,
- al-Timman J.K.,
- et al.
- Tan L.B.
- Collier C.R.
- Fox S.I.
- Defares J.G.
- Auricchio A.,
- Stellbrink C.,
- Sack S.,
- et al.
- Genth S.,
- Zotz R.,
- Darius H.,
- et al.
- Mancini D.M.,
- Eisen H.,
- Kussmaul W.,
- Mull R.,
- Edmunds L.H. Jr..,
- Wilson J.R.