Author + information
- Received January 23, 2006
- Revision received May 17, 2006
- Accepted May 22, 2006
- Published online September 19, 2006.
- Gerhard-Paul Diller, MD⁎,†,
- Konstantinos Dimopoulos, MD⁎,1,
- Darlington Okonko, BSc, MRCP†,1,
- Anselm Uebing, MD⁎,
- Craig S. Broberg, MD⁎,1,
- Sonya Babu-Narayan, MRCP⁎,1,
- Stephanie Bayne, BSc⁎,
- Philip A. Poole-Wilson, MD, FRCP†,
- Richard Sutton, DScMed‡,
- Darrel P. Francis, MA, MRCP§,1 and
- Michael A. Gatzoulis, MD, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Prof. Michael A. Gatzoulis, Adult Congenital Heart Program, Royal Brompton Hospital, Sydney Street, SW3 6NP London, United Kingdom.
Objectives To assess the prognostic value of heart rate response to exercise in adult congenital heart disease (ACHD) patients.
Background An abnormal heart rate response to exercise is related to autonomic dysfunction and may have prognostic implications in ACHD.
Methods We identified 727 consecutive ACHD patients (mean age [± SD] 33 ± 13 years) with varying diagnoses and without pacemakers. Peak oxygen consumption (peak Vo2), resting heart rate, and the increase in heart rate from resting level to peak exercise (“heart rate reserve”) were measured. We also quantified the decrease in heart rate (“heart rate recovery”) after cessation of exercise.
Results During a median follow-up of 28 months, 38 patients died. Lower values of heart rate reserve, peak heart rate, heart rate recovery, and peak Vo2(p < 0.01 for each) were associated with increased mortality in univariate analysis. Furthermore, heart rate reserve predicted mortality independently of antiarrhythmic therapy, functional class, and peak Vo2. Stratifying patients by diagnostic groups revealed that a lower heart rate reserve was also associated with a greater risk of death in patients with complex anatomy, Fontan circulation, and tetralogy of Fallot (p < 0.05 for each).
Conclusions An abnormal heart rate response to exercise identifies ACHD patients with a higher risk of mortality in the midterm, even after accounting for antiarrhythmic medication and exercise capacity. Heart rate reserve is a simple and inexpensive way to identify ACHD patients at higher mortality risk.
With advances in surgical management, an increasing number of patients with congenital heart disease reach adulthood (1,2). Such patients have a higher mortality over the medium and long term compared with healthy individuals with similar demographic characteristics (3–7). Development of simple risk stratification methods would permit resources to be directed to patients with adult congenital heart disease (ACHD) at greatest risk. Cardiopulmonary exercise testing with measurement of peak oxygen consumption is increasingly used in ACHD patients because it may provide similar prognostic information in ACHD as it does in patients with acquired heart disease. Measuring peak oxygen consumption, however, requires expensive equipment and specific expertise and is therefore not widely available.
Chronotropic incompetence—a blunted increase in heart rate during exercise—is an established predictor of mortality in patients with coronary artery disease and in healthy populations (8–10). Little is known about its prevalence and prognostic implications across the spectrum of ACHD. Attenuation of heart rate recovery—the rate of decrease in heart rate after cessation of exercise—also is associated with increased mortality in patients being assessed for coronary artery disease (11). Because cardiac autonomic dysfunction is common in ACHD patients (12,13), we hypothesized that abnormal heart rate response to exercise may also be common in ACHD and could be a simple means of risk stratification.
The aims of this study were: 1) to evaluate the prevalence of an abnormal heart rate response to exercise (chronotropic incompetence) in ACHD patients; 2) to assess the relationship between heart rate response and exercise capacity; and 3) to evaluate whether chronotropic incompetence is a prognostic marker in ACHD patients after accounting for exercise capacity and use of antiarrhythmic medication.
This was a retrospective study. We analyzed data from all cardiopulmonary exercise tests performed in ACHD patients at our institution between February 1999 and April 2005. Patients were referred for exercise testing as part of a protocolized clinical follow-up for ACHD patients. This study was approved by the local ethics committee. Almost all patients underwent only 1 test during the study period. Any repeat test was not included in the analysis. A main diagnosis was determined for every patient from the hospital records. If more than 1 cardiac lesion was present, the lesion considered hemodynamically most important was recorded as the main diagnosis. Patients with multiple complex cardiac lesions substantially affecting hemodynamics were classified as complex anatomy. The New York Heart Association (NYHA) functional class was determined by physician assessment of patients’ self-reported symptoms before the date of the exercise test. Antiarrhythmic drug use was recorded at the time of exercise testing.
Cardiopulmonary exercise testing
Cardiopulmonary exercise testing was performed on a treadmill according to a modified Bruce protocol (14,15) with the addition of a “stage 0” in which the patient walks at a velocity of 1 mile/h and a gradient of 5% for 3 min. All subjects were encouraged to exercise to exhaustion regardless of the maximal heart rate achieved. Ventilation, oxygen uptake, and carbon dioxide production were measured continuously using a respiratory mass spectrometer (Amis 2000; Innovision, Odense, Denmark) as described previously (16). Heart rate was assessed by continuous electrocardiography, and arterial blood pressure was recorded manually by sphygmomanometry. Resting heart rate was measured after at least 30 s in a seated position, and peak heart rate was defined as the maximal heart rate achieved during exercise. Predicted maximum heart rate was estimated according to the Astrand formula (220-age) (17), and percentage of maximum age-predicted heart rate was calculated as the ratio between peak heart rate and age-predicted maximum heart rate (220-age).
Calculation of heart rate reserve
Heart rate reserve was calculated as the difference between peak and resting heart rates. The chronotropic index, (peak heart rate − resting heart rate)/(220-age − resting heart rate) (9), is derived by applying the chronotropic metabolic relationship concept introduced by Wilkoff et al. (18) to a symptom-limited exercise test as described previously (10). This allows definition of the normal chronotropic response independently of age, resting heart rate, and functional state (18). In a group of 410 healthy adults,Wilkoff et al. (18) reported 95% limits of normality of chronotropic index to be 0.8 to 1.3. Based on this finding, we defined chronotropic incompetence as failure to achieve a chronotropic index of 0.8 (i.e., falling below 97.5% of healthy adults).
Calculation of heart rate recovery
Heart rate was also recorded 1, 2, 3, and 5 min after the cessation of exercise, and heart rate recovery was calculated as the difference between peak heart rate and the heart rate at these recovery time points. In addition, the relative decrement in heart rate was calculated as heart rate recovery divided by the heart rate at peak exercise.
Follow-up was complete for all patients. Survival status and time to death was assessed through the health service computer system, which is linked to a national database held by the Office of National Statistics. We planned the study to use all-cause mortality as the end point to eliminate any possibility of bias arising from incorrect classification of cause of death.
All values are presented as mean ± standard deviation. Comparisons between groups were made using the Student ttest, Mann-Whitney Utest, or chi-square test as appropriate. Variables were assessed on univariate analysis. Significant parameters were subsequently included into a multivariate regression model in a stepwise forward procedure. Univariate Cox proportional hazards analysis was used to assess the association between variables and the end point of all-cause mortality. Parameters significantly predicting prognosis in univariate analysis were subsequently tested in a multivariate Cox proportional hazards analysis by the stepwise forward method to assess the independent effect of these variables. Areas under curve (AUC) for sensitivity and specificity were calculated using receiver-operating characteristic (ROC) analysis to compare prognostic accuracy of different parameters. Statistical analyses were performed using the StatView 5.0 (Abacus Concepts, Berkeley, California) and MedCalc 8.2.1 (MedCalc Software, Mariakerke, Belgium) software packages.
Characteristics of the 727 consecutive ACHD patients (mean age 33 ± 13 years, 52% male) included in this analysis are presented in Tables 1 and 2.⇓⇓Patients with chronotropic incompetence were more likely to be cyanotic, female, or treated with antiarrhythmic drugs. In addition, by definition, such patients had lower values of peak heart rate and heart rate reserve, as shown in Table 1.
Prevalence of chronotropic incompetence
Chronotropic incompetence was present in 62% of the patients. The prevalence was highest in patients after Fontan palliation (84%) and in patients with Eisenmenger physiology (90%), and it was lowest in patients with repaired tetralogy of Fallot (52%), repaired ventricular septal defect (52%), and isolated valvar disease (47%), as shown in Table 2. On considering only those patients who reached anaerobic threshold (i.e., those with a respiratory quotient below 1.0), the frequency of chronotropic incompetence was found to be 54% overall. Once again it was most frequent in the Eisenmenger patients (96%) and least frequent in patients with repaired ventricular septal defects (35%).
Relationship to symptoms and exercise capacity
Patients with chronotropic incompetence were more likely to be in a higher NYHA class (34% NYHA class I, 47% NYHA class II, 19% NYHA class III) than the remaining patients (64% NYHA class I, 30% NYHA class II, 6% NYHA class III) (p < 0.0001). Patients with chronotropic incompetence also had lower peak oxygen consumption (20.4 ± 8.2 ml/kg/min vs. 28.0 ± 9.9 ml/kg/min; p < 0.0001) and shorter exercise duration (541 ± 196 s vs. 732 ± 196 s; p < 0.0001). In addition, heart rate reserve (r = 0.53; p < 0.0001) and peak heart rate (r = 0.49; p < 0.0001) correlated with peak oxygen consumption.
Prognostic value of parameters of chronotropic incompetence
During a median follow-up of 851 days after cardiopulmonary exercise testing (range 60 to 2,254 days), 38 patients died. The patients who died had the following diagnoses: Fontan physiology (n = 7), complex anatomy (n = 10), congenitally corrected transposition of the great arteries (n = 2), atrial switch procedure for transposition of the great arteries (n = 1), tetralogy of Fallot (n = 5), isolated valvar disease (n = 2), single ventricle physiology (n = 1), Eisenmenger syndrome (n = 1), aortic coarctation (n = 1), Ebstein anomaly (n = 3), repaired atrial (n = 1) and atrioventricular (n = 2) septal defects, and others (n = 2).
On univariate analysis, heart rate reserve, chronotropic index, peak heart rate, and percentage predicted heart rate predicted survival (Table 3).The other univariate predictors of survival were use of antiarrhythmic drug therapy, peak oxygen consumption, and NYHA functional class. In addition, amiodarone or digoxin use was related to survival, as shown in Table 3. Age, gender, cyanosis, and treatment with sotalol, beta-blocker, calcium antagonist, class I antiarrhythmic drugs, or angiotensin-converting enzyme inhibitors were not related to survival.
Measures of chronotropic response correlated strongly with each other (r value between 0.81 and 0.96; p < 0.001 for each). Therefore, for multivariate analysis we chose the parameter with the highest predictive value on univariate Cox proportional hazard analysis and the greatest AUC on ROC analysis. In both analyses, heart rate reserve (chi-square = 26.1; AUC = 0.74) and chronotropic index (chi-square = 26.9; AUC = 0.74) were superior to peak heart rate (chi-square = 20.9; AUC = 0.72) and percentage age-predicted peak heart rate (chi-square = 21.0; AUC = 0.72) in predicting prognosis. As a consequence heart rate reserve, representing a much simpler parameter than chronotropic index, was used in subsequent analyses.
On multivariate analysis, heart rate reserve, NYHA functional class, and therapy with antiarrhythmic drugs jointly predicted mortality, independently of peak oxygen consumption, as shown in Table 4.These results remained unchanged when patients who did not reach the anaerobic threshold during exercise (i.e., those with a respiratory quotient below 1.0) were excluded from the analyses. Figure 1illustrates the relationship between heart rate reserve and death from any cause among adult congenital heart disease patients stratified by quartiles of heart rate reserve.
Prognostic value of heart rate reserve in individual diagnostic groups
Heart rate reserve predicted mortality in patients after Fontan palliation (hazard ratio [HR] = 0.65 per 10 beats/min; 95% confidence interval [CI] 0.48 to 0.88; p < 0.05), complex anatomy (HR = 0.81 per 10 beats/min; 95% CI 0.65 to 0.998; p < 0.05), and repaired tetralogy of Fallot (HR = 0.66 per 10 beats/min; 95% CI 0.48 to 0.91; p < 0.05) on univariate analysis. No significant association between low heart rate reserve and mortality was found in patients with simple lesions, systemic right ventricles, or Ebstein anomaly of the tricuspid valve.
Comparative prognostic value of heart rate reserve and peak oxygen consumption
Heart rate reserve was at least as good as peak oxygen consumption in predicting mortality, both on univariate Cox analysis (chi-square = 26.1 vs. 19.3) and on ROC analysis (AUC = 0.74 vs. 0.68) (Fig. 2A).Combining these 2 variables was also helpful: Patients with both heart rate reserve and peak oxygen consumption within the lowest quartile (<51 beats/min and <16.7 ml/kg/min, respectively) had the worst prognosis, patients with only one in the lowest quartile had an intermediate prognosis, and those with neither in the lowest quartile had the best prognosis (p < 0.0001) (Fig. 1).
Prognostic value of heart rate recovery
Data on heart rate recovery was available in 505 patients (those patients who underwent exercise testing after March 2001). Of these, 16 patients died during follow-up. Heart rate recovery at 1, 2, 3, and 5 min was significantly lower in patients who died than in surviving patients (p < 0.05 for each), as was heart rate recovery expressed as percentage of peak heart rate (Fig. 2B). Heart rate recovery at 1, 2, 3, and 5 min after exercise was significantly related to mortality on univariate Cox proportional hazards analysis (p < 0.05 for each). After adjustment for antiarrhythmic drug therapy heart rate recovery at 1, 2, 3, and 5 min after exercise remained independently predictive of survival in bivariate Cox analysis (p < 0.05 for each).
This study demonstrates that a blunted heart rate response to exercise (chronotropic incompetence) is prevalent across the spectrum of ACHD and predicts an enhanced mortality risk independently of antiarrhythmic medication. Even an attenuated rate of recovery of heart rate after exercise testing carries important prognostic information. Moreover, a simple combination of heart rate reserve and peak oxygen consumption identifies a subpopulation of ACHD patients with a 3.8-fold increase in mortality.
Chronotropic incompetence was found to be prevalent in ACHD, affecting 62% of patients. In other cohorts, the prevalence of chronotropic incompetence ranges between 30% in patients with chronic heart failure (19) to 60% in patients with chronic atrial fibrillation (20). In the present study, the prevalence of chronotropic incompetence was lowest in patients with simple lesions, such as repaired ventricular septal defect, Ebstein anomaly, or palliated transposition of the great arteries, and was highest in patients with complex, uncorrected, and cyanotic lesions. This increase in prevalence of chronotropic incompetence parallels the decline in peak oxygen consumption across the spectrum of ACHD. It has been suggested that a blunted heart rate response may in part account for the diminished exercise capacity seen in these patients (21). The results of our study support this notion. We found that patients with chronotropic incompetence had poorer exercise capacity compared with patients without chronotropic incompetence. In addition, change in heart rate correlated with peak oxygen consumption. However, in this cohort heart rate explains only a quarter of the variation in peak oxygen consumption. Therefore, other parameters, such as age, gender, pulmonary function, cyanosis, and level of fitness may play an important role in determining exercise capacity in ACHD patients. We also found a relationship between chronotropic incompetence and symptomatic state. Patients with chronotropic incompetence were more likely to be in a higher NYHA functional class than patients with a normal heart rate response to exercise. Whether this is a causal relationship remains to be determined.
Heart rate reserve, though a simple and easily obtained marker, turned out to be a powerful prognostic marker in ACHD independently of antiarrhythmic medication and exercise capacity. Stratifying patients by diagnostic groups revealed that a lower heart rate reserve was also associated with a greater risk of death in patients with complex anatomy, Fontan circulation, and tetralogy of Fallot. Interestingly, despite their poor exercise capacity, Eisenmenger patients did not have a correspondingly poor survival, and, therefore, neither peak oxygen consumption nor heart rate reserve failed to predict prognosis in this population. We speculate that in the Eisenmenger patients the limitation to exercise does not arise from the usual broad constellation of ominous pathophysiologic abnormalities (including poor ventricular function, vascular remodeling, autonomic dysfunction, etc.) but rather more specifically from exercise-induced increase in right-to-left shunting. Thus there is a “cap” on exercise capacity and therefore on heart rate reserve. This cap may be far below that which would have been set by the usual pathophysiologic abnormalities which in turn are responsible for the impaired prognosis. As a consequence, their survival is nowhere near as poor as would be predicted from the exercise capacity. Further work will be needed to identify the subset of patients within the Eisenmenger cohort who are at highest risk of mortality.
In acquired heart disease, both exercise capacity and chronotropic incompetence are known to be predictors of poor prognosis (8,9,22,23). Of these, only peak oxygen consumption is routinely used for risk stratification, and even this is largely limited to acquired chronic heart failure. In ACHD patients, the prognostic value of both peak exercise and chronotropic incompetence is far from established. The present study shows in a large cohort of ACHD patients that both parameters of chronotropic incompetence and peak oxygen consumption are of prognostic value and that heart rate reserve is at least as good a predictor of mortality in ACHD patients as peak oxygen consumption. Combining parameters of chronotropic response with peak oxygen consumption enables further risk stratification.
Underlying mechanisms responsible for chronotropic incompetence in ACHD patients are not fully understood. It appears likely that chronotropic incompetence is a multifactorial phenomenon resulting from the confluence of several factors which themselves are associated with poor prognosis. Colluci et al. (24) reported that impaired chronotropic response to exercise in patients with chronic heart failure is, at least in part, due to postsynaptic desensitization of beta-adrenergic receptors. In addition, it has been demonstrated that heart rate variability is significantly decreased in patients with acquired heart disease who are chronotropically incompetent (25).
It remains to be elucidated whether the prognostic power of heart rate reserve results from its dependence on mechanisms such as exercise capacity, neurohormonal activation, autonomic dysfunction, and hemodynamic compromise. This study identifies heart rate reserve as a physiologically important piece of information to extract from a cardiopulmonary exercise test alongside the usual measurements. Whether specifically targeting abnormal heart rate reserve could improve prognosis remains unknown. There are no data to suggest that directly intervening on heart rate (e.g., inserting a rate-responsive pacemaker) would improve prognosis. Rather, for now, we believe these data indicate the potential utility of this additional information in selecting patients for special medical or further surgical attention because they are at greater risk of death than the clinician might otherwise predict.
Cardiopulmonary exercise testing in this study was performed as part of routine evaluation of patients in the ACHD clinic. All patients were at a tertiary ACHD center and, therefore, it is possible that they may not represent the pattern of ACHD that may exist in the community. Nevertheless, the patients were not restricted to any particular narrow diagnostic group but rather covered the entire spectrum of ACHD diagnoses and included all segments of the population regardless of age, gender, history, and nature of surgery. The number of deaths forming the basis of this report is limited and further prospective studies with a longer period of observation and a higher number of clinical events are desirable to validate the results reported in this cohort and to provide information on the potential response to different therapies. This study cannot identify why a blunted heart rate response to exercise in ACHD patients predicts poor prognosis. Indeed, even the mechanisms of depressed heart rate responses in ACHD remain unclear. Now that the prognostic value of heart rate responses are emerging, further research may be stimulated that could elucidate the mechanisms responsible.
An abnormal heart rate response to exercise is prevalent across the spectrum of adult congenital heart disease and is associated with a greater risk of death. Even on its own, heart rate reserve is potentially a simple means of identifying ACHD patients at elevated risk. In combination with formal measurement of peak oxygen consumption, it identifies a subpopulation with a 3.8-fold elevated risk of death in the mid term. Exercise testing should be considered as part of the routine assessment of adults with congenital heart disease.
We wish to acknowledge our exercise laboratory staff for their ongoing support.
↵1 Dr. Dimopoulos is supported by the European Society of Cardiology, Drs. Francis, Okonko, and Babu-Narayan by the British Heart Foundation, and Dr. Broberg by the Waring Trust.
The Royal Brompton Adult Congenital Heart Programme and the Department of Clinical Cardiology have received support from the British Heart Foundation and the Clinical Research Committee, Royal Brompton Hospital, London.
Presented as part of the 2005 Outstanding Research Award in Pediatric Cardiology at the American Heart Association Scientific Sessions, Dallas, Texas, November 13, 2005.
- Abbreviations and Acronyms
- adult congenital heart disease
- area under curve
- New York Heart Association
- receiver-operating characteristic
- Received January 23, 2006.
- Revision received May 17, 2006.
- Accepted May 22, 2006.
- American College of Cardiology Foundation
- Perloff J.K.,
- Warnes C.
- Engelfriet P.,
- Boersma E.,
- Oechslin E.,
- et al.
- Boneva R.S.,
- Botto L.D.,
- Moore C.A.,
- Yang Q.,
- Correa A.,
- Erickson J.D.
- Nieminen H.P.,
- Jokinen E.V.,
- Sairanen H.I.
- Lauer M.S.,
- Okin P.M.,
- Larson M.G.,
- Evans J.C.,
- Levy D.
- Bruce R.A.,
- Blackman J.R.,
- Jones J.W.
- Gatzoulis M.A.,
- Clark A.L.,
- Cullen S.,
- Claus G.H.,
- Redington A.N.
- Francis D.P.,
- Shamim W.,
- Davies L.C.,
- et al.
- Cohn J.N.,
- Johnson G.R.,
- Shabetai R.,
- et al.,
- V-HeFT VA Cooperative Studies Group
- Azarbal B.,
- Hayes S.W.,
- Lewin H.C.,
- Hachamovitch R.,
- Cohen I.,
- Berman D.S.
- Colucci W.S.,
- Ribeiro J.R.,
- Rocco M.B.,
- et al.