Author + information
- Received September 25, 2007
- Revision received February 4, 2008
- Accepted February 12, 2008
- Published online July 8, 2008.
- Stephen M. Paridon, MD⁎,⁎ (, )
- Paul D. Mitchell, MS†,
- Steven D. Colan, MD⁎⁎,
- Richard V. Williams, MD‡,
- Andrew Blaufox, MD§,
- Jennifer S. Li, MD∥,
- Renee Margossian, MD⁎⁎,
- Seema Mital, MD¶,
- Jennifer Russell, MD#,
- Jonathan Rhodes, MD⁎⁎,
- Pediatric Heart Network Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Stephen M. Paridon, Cardiology Division, Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, Pennsylvania 19104.
Objectives The aim of this study was to describe exercise performance during the first 2 decades of life in Fontan survivors by a cross-sectional study and to identify factors that influence exercise performance.
Background Exercise performance after the Fontan procedure is reduced relative to performance in healthy subjects. Data on pre-adolescents are limited, and the patterns of exercise performance in different ages are unexplored.
Methods Ramp cycle ergometry was performed with expired gas. Data were analyzed for the entire study population and for subpopulations that did and did not achieve a maximal aerobic capacity.
Results Of 411 subjects tested (12.4 ± 3.2 years of age), 166 achieved a maximal aerobic capacity. Peak oxygen consumption (VO2) was 26.3 ml/kg/min (65% of predicted for age and gender [% predicted]) for the entire population and was lower in the submaximal capacity subgroup compared with the maximal capacity subgroup (63% predicted and 67% predicted, respectively; p = 0.02). Oxygen consumption at ventilatory anaerobic threshold (VAT) was better preserved (78% predicted for the total population) than peak VO2. Higher % predicted O2 pulse at peak exercise was associated with greater % predicted peak VO2, work rate, and VAT. Adolescence and male gender were associated with decreased % predicted peak VO2. The relationship between echocardiographic indexes of ventricular function and exercise function was surprisingly weak.
Conclusions In Fontan patients, maximal aerobic capacity is reduced compared with healthy subjects, with better preservation of submaximal performance. Higher O2 pulse is associated with better exercise performance, whereas adolescence and male gender are associated with decreased performance compared with healthy subjects.
The Fontan procedure has become the most widely used palliation for single ventricle physiology (1). As patients undergo the Fontan procedure at younger ages, it becomes increasingly important to define functional status and long-term outcome and to investigate factors that may influence outcome.
The Pediatric Heart Network, consisting of 7 centers in the U.S. and Canada and a Data Coordinating Center at the New England Research Institutes, is funded by the National Heart, Lung, and Blood Institute of the National Institutes of Health to develop evidence-based recommendations for children with heart disease. In planning for interventional studies of the Fontan population, the network was confronted with a lack of normative data on the functional and clinical status of this population. Therefore, a cross-sectional study to evaluate functional and clinical status in Fontan survivors was undertaken. As part of this study, cardiopulmonary exercise testing was performed to assess exercise performance, an important component of a child's quality of life and functional status (2).
Previous studies in patients with the Fontan procedure report poor exercise performance, decreased peak oxygen consumption (VO2), and decreased VO2 at ventilatory anaerobic threshold (VAT) (2–10). Although likely multi-factorial, the impaired exercise performance of this population is in part explained by an inability to increase or maintain stroke volume during exercise (11). Previous findings are limited by small numbers of patients and single-center retrospective designs. Most past studies were performed in post-pubertal subjects. There are limited data regarding younger subjects and changing patterns of exercise performance during adolescence (7).
The purposes of this investigation were to define exercise performance during the first 2 decades of life in a large cohort of Fontan survivors participating in a multicenter cross-sectional study and to determine the association between defined clinical factors and exercise performance.
The Fontan cross-sectional study design
The primary aim of the main study was to assess the correlation between functional health status and measures of ventricular performance in children ages 6 to 18 years who have undergone a Fontan procedure. Consent and assent were obtained for all subjects according to local guidelines. Subjects were ≥115 cm tall and able to perform cycle ergometry.
Medical records of 1,078 children who had undergone a Fontan procedure were screened, and 831 (77%) were identified as potentially eligible for the study. Of 831, 644 (77%) were fully eligible to participate in the complete Fontan protocol, on the basis of their ability to be contacted and perform the necessary testing at participating centers. Details regarding the Pediatric Heart Network and the methods of the Fontan cross-sectional study have been reported previously (12). A total of 546 (85% of 644) children were enrolled; of these, 411 (75% of 546) underwent exercise testing. Reasons for not testing were height (39%), refusal (29%), and inability to cooperate or confounding medical condition (22%).
After consent, prospective maximal exercise testing was performed with a ramp protocol on an electronically braked cycle ergometer. Equipment was calibrated to manufacturers' specifications, and testing was performed with standard protocols previously used in children, adolescents, and subjects with Fontan physiology (7,12–14).
Resting pulmonary measurements
After familiarization with the equipment, subjects underwent inspiratory and expiratory flow volume loops. Subjects performed hyperventilation for 10 s to measure maximal voluntary ventilation (MVV) (l/min). Testing was repeated at least 3 times for reproducibility.
Subjects pedaled in an unloaded state for 3 min. Workload was then increased continuously with a slope chosen to achieve each subject's predicted maximal work rate (W) after 10 to 12 min of cycling.
Expired gases were measured for 3 min of quiet rest and throughout the exercise protocol. Oxygen consumption, carbon dioxide production (VCO2), and minute ventilation (VE) were measured on a breath-by-breath basis. Peak VO2 was defined as the highest VO2 achieved by the subject during the test. Ventilatory anaerobic threshold (VAT) was measured by V-slope method when it could be accurately determined. Values for VO2 were indexed to body weight and expressed as percentage of predicted values for healthy age- and gender-matched subjects as reported by Cooper and Weiler-Ravell (14), with a similar protocol. The ventilatory equivalents of carbon dioxide (VE/VCO2) were measured at VAT. The respiratory exchange ratio (RER) (VCO2/VO2) was measured continuously. The O2 pulse (VO2/heart rate [HR]) was measured at peak VO2 and indexed to body surface area. The O2 pulse is equal to the product of stroke volume and the arterial-venous O2 content difference. Because the arterial-venous O2 content difference at peak exercise varies little among subjects, the O2 pulse might be used as a surrogate for stroke volume at peak exercise (13).
Breathing reserve was calculated by the following formula:
Resting 12-lead electrocardiograms were performed in the supine, sitting, and standing position and during brief hyperventilation. Heart rate was monitored continuously. Chronotropic index, a measure of chronotropic response that is independent of resting HR and stroke volume (15), was defined as:
Arterial O2 saturation was measured continuously with an ear or finger pulse oximeter.
Echocardiography was performed with standard techniques to obtain ventricular mass and volumes, peak early and peak late atrioventricular valve inflow velocities (16), and the mean rate of pressure change during isovolumic contraction (mean dP/dtic) (a geometry-independent index of ventricular function that approximates and closely correlates with peak dP/dt) (17) as:
Isovolumic contraction time was determined from the atrioventricular valve inflow and aortic valve outflow Doppler tracings. Tei index was calculated from systemic ventricular inflow and outflow Doppler tracings (18,19). Composite indexes of diastolic function were derived as (early inflow velocity/peak late inflow velocity) and (early inflow velocity/peak early diastolic tissue velocity). Diastolic function was classified according to 2 systems: restrictive pattern present or absent, and diastolic dysfunction grade (20). Studies were evaluated at a core laboratory as described in the design article (12).
Subject characteristics and exercise performance measures are presented for the 411 subjects who underwent exercise testing and were stratified according to whether or not subjects achieved their maximal aerobic capacity (defined as achieving a peak RER ≥1.1). These groups are subsequently referred to as the “maximal capacity and submaximal capacity” groups. Data are described as mean ± SD for quantitative variables and n (%) for qualitative variables. Comparisons between subgroups for continuous data were made with the Student t test if normally distributed and the Wilcoxon rank sum test otherwise. Tests between subgroups for discrete data were made with the Pearson chi-square test and the Fisher exact test. All tests of significance were 2-sided.
The coefficient of determination (R2) was used to assess the relative contributions of O2 pulse, chronotropy, and resting arterial O2 saturation to exercise performance. Subjects with submaximal capacity and subjects with pacemakers were excluded, to eliminate confounding effects of rate responsive pacemakers on the chronotropic index and O2 pulse. The percent of predicted for age and gender (% predicted) O2 pulse was used as a surrogate for stroke volume, chronotropic index was used as a measure of chronotropic ability, and resting arterial O2 saturation was used as a measure of arterial O2 content. Exercise performance measures included % predicted peak VO2, % predicted maximal work rate, and % predicted VO2 at VAT.
Multiple linear regression was used to investigate the independent association of the variables listed in Table 1 with each dependent variable (% predicted peak VO2 and % predicted maximum O2 pulse). An interaction term was included in each model to test for differential effects of age by gender on the outcomes and was subsequently dropped when found to be not statistically significant. The final models were informed by stepwise selection (21). Logistic regression was used to assess the association of maximal effort and Fontan type after adjustment for age and race. All data analysis was performed with SAS/STAT software (SAS Institute, Cary, North Carolina).
Characteristics for the entire study sample that underwent exercise testing are summarized in Table 2. Subjects achieving maximal capacity were on average older and larger but had a body mass index (BMI) Z score similar to those with submaximal capacity. There were statistically significant differences between the groups for race and type of Fontan operation but not gender, anatomical diagnosis, pacemaker use, or use of angiotensin-converting enzyme inhibitors. Fatigue was reported as the reason for submaximal capacity in 58%, insufficient effort in 25%, and inability to cooperate or other reasons in 18%.
Exercise performance measures for the entire population and the 2 subpopulations are summarized in Table 3. Of the variables listed in Table 1, only older age at exercise testing and an intracardiac lateral tunnel independently predicted achieving maximal capacity. A multivariate model demonstrated that the association between Fontan type and maximal capacity status was weak (p = 0.054) after adjustment for age and race.
Peak VO2 was depressed compared with healthy subjects. Average peak VO2 for the total population was 65% predicted for age and gender. Only 113 (28%) subjects achieved a peak VO2 within the normal. Although VO2 at VAT was impaired when compared with normal predicted values (78% predicted), it was better preserved than peak VO2 with the majority (63%) of subjects in the normal range. This pattern was also seen in 162 of 166 patients in the maximal capacity subgroup in whom a VAT could be identified, indicating that the disproportionately low peak VO2 was not a consequence of inadequate effort.
Physical working capacity
Maximal work rate was decreased similarly to peak VO2 (61% predicted). Work rate was greater in the maximal capacity subgroup compared with the submaximal capacity subgroup (66% predicted and 57% predicted, respectively; p < 0.0001).
Resting and exercise pulmonary functions are summarized in Table 3. The VE/VCO2 at VAT was elevated (43 ± 10) among the 320 (78%) subjects in whom it could be measured, and the forced vital capacity, forced expiratory volume at 1 s, and forced expiratory volume at 1 s/forced vital capacity was decreased, with no differences between the maximal capacity and submaximal capacity subgroups. Despite these abnormalities, breathing reserve for the total study group was in the normal range (26 ± 21%). However, the SD was large, and breathing reserve was below 15%—considered the lower limit of normal—in 34% of all subjects (data not shown). The mean values for both subgroups were also in the normal range, although the value of the maximal group was significantly less than the submaximal group.
Chronotropic impairment was prominent with a mean peak HR of 74 ± 11% predicted for the entire study population. This value was similar in subjects attaining maximal capacity. The O2 pulse index was also decreased and differed between subgroups (5.7 ± 1.4 ml O2/beat/body surface area in the maximal capacity subgroup, 5.4 ± 1.5 ml O2/beat/body surface area in the submaximal capacity subgroup; p = 0.03). Mild arterial O2 desaturation (94 ± 4%) was present at rest, and further desaturation (91 ± 6%) occurred with exercise and was similar in both subgroups.
The relative contribution of O2 pulse, chronotropy, and arterial O2 saturation to exercise performance measures in the maximal capacity group is presented in Table 4. Although chronotropic impairment and mild arterial O2 desaturation are prominent in this population, they are responsible for an insignificant amount of the variance (<5%) in measures of exercise performance. Maximum % predicted O2 pulse, as a surrogate of stroke volume, accounts for 73% of the variation in % predicted peak VO2, 26% of the variation in % predicted maximum work rate, and 35% of the variation in % predicted VO2 at VAT.
Determinants of exercise performance
Results of the multiple linear regression used to investigate the independent association of the variables listed in Table 1 with % predicted peak VO2 and % predicted maximum O2 pulse for the maximal capacity group without pacemakers are displayed in Table 5. These 2 dependent variables were chosen as the primary measures of aerobic capacity and stroke volume, respectively. Male gender, post-pubertal age, and increased BMI Z score were negatively associated with both variables. A weak negative correlation was found between D-looping and % predicted peak VO2. A weak positive correlation also existed between mean dP/dtic and peak O2 pulse. No significant association was found for the other anatomical, surgical, and resting measures of cardiac function.
This study provides a unique opportunity to characterize the exercise function of a large group of pediatric subjects who have undergone a Fontan procedure. It demonstrates that the exercise function of these patients varies greatly. In some, exercise capacity is well preserved, with normal and even above-average peak VO2 and work rate (Table 3). However, in most cases peak exercise capacity is significantly depressed. For the entire population, only 28% had a peak VO2 within the normal range. These results are similar to previous reports (3–6).
A unique finding in this study is that VAT was better preserved than peak VO2. In fact, it was in the normal range for the majority of subjects. This finding, to our knowledge not previously reported, suggests that Fontan subjects can tolerate a higher level of submaximal activity than might be anticipated from measurement of peak VO2. It may also indicate that absence of a subpulmonary ventricle may limit the Fontan patients' ability to accommodate hemodynamic burdens associated with levels of exercise above anaerobic threshold.
It is clear from this study that subjects with Fontan physiology often have exercise performance that is limited by factors unrelated to their cardiovascular system. Fifty-seven percent of subjects failed to achieve an RER ≥1.1 and presumably did not exhaust their cardiovascular capacity. Other potential causes of exercise limitation include pulmonary, musculoskeletal deconditioning, and motivational factors.
Although pulmonary functions were abnormal throughout the study population, they were similar in both the maximal and submaximal populations. The significantly higher breathing reserve in the submaximal subgroup indicates that pulmonary abnormalities, although common, did not limit exercise performance to a greater degree in the submaximal group compared with the maximal capacity group. It is worth noting that the standard deviation of the breathing reserve is high across the subgroups. In fact, despite an overall normal mean, one-third of the subjects had a breathing reserve below the normal limit of 15%. This suggests a degree of pulmonary limitation in some subjects. However, as seen in Table 3, breathing reserve is negative in some subjects. This is the result of a greater peak VE than MVV and indicates an inadequate effort in these subjects when performing the highly effort-dependent MVV. The lower breathing reserve in the maximal subgroup is likely due to the higher VE that would be achieved at peak exercise. Although breathing reserve was low in some subjects, their high RERs reflect lactic acidosis. Thus, their exercise limitation was a consequence of exhaustion of their cardiovascular reserve rather than primarily a pulmonary limitation.
The VE/VCO2 at VAT and peak respiratory rate were equally high in all study groups (Table 3). Recent studies in adults with congenital heart disease suggest that decreased ventilatory efficiency is associated with poor survival in acyanotic but not cyanotic subjects (8). The implication of these findings for the current study group is unclear. Although only mildly desaturated at rest, our population has significant desaturation at peak exercise. This would suggest that the ventilatory inefficiency was a consequence of intracardiac and extracardiac shunting rather than a marker of ventricular dysfunction. As such, in this younger and healthier population, it probably will not have a significant prognostic usefulness.
Fatigue was the major reported reason for exercise termination in the submaximal group. In a young population, this might indicate muscular deconditioning but is also the usual reason given by the subject for exercise termination when expending a submaximal effort. Thus, motivation might be a major reason for exercise limitation in the submaximal group. The group is significantly younger than the maximal capacity group. Younger subjects may often have more difficulty performing formal exercise testing and continuing to exhaustion. This finding suggests that the aerobic capacity in the younger Fontan population may often be underestimated due to submaximal effort.
Cardiovascular response to exercise
In assessing the cardiovascular factors limiting exercise performance, we focused only on subjects who expended a maximal aerobic effort (RER ≥1.1; the maximal capacity subgroup) (i.e., those whose exercise capacities were limited by their cardiovascular system). Overall, this population had an aerobic capacity as measured by peak VO2, VO2 at VAT, and maximal working capacity similar to that reported in previous studies (5,7,8,11,22–24). These previous studies have shown that chronotropic impairment, arterial O2 desaturation, and decreased stroke volume response to exercise are common in Fontan physiology. To our knowledge, this report is the first to attempt to assess the relative contributions of each of these factors to the variance in aerobic exercise performance observed among patients with Fontan physiology. Although common, chronotropic incompetence and arterial desaturation were minimally responsible for the variance in aerobic performance (both at peak VO2 and at VAT) and physical working capacity in the subgroup of subjects who achieved maximum effort (Table 4). In contrast, O2 pulse at peak exercise was, by far, the most important factor accounting for the variance in aerobic performance.
The explanation for the difference in the results of the current study from previous studies regarding the effects of chronotropy on exercise performance may be related to the fact that we restricted our analyses to patients who achieved a true maximal aerobic capacity. In contrast, most previous studies have included at least some subjects with submaximal efforts (on the basis of the reported RERs and the SD of these values). Subjects with submaximal effort achieve HRs and VO2 at peak exercise below their true maximal VO2 and maximal HR. Consequently, any study that includes submaximal tests will inevitably find a significant positive relationship between HR response and “maximal” VO2. By restricting analyses to subjects who achieved a true maximal aerobic capacity, this potential pitfall was avoided. When these measures were undertaken, we found that differences in chronotropic response accounted for very little of the variance in maximal VO2.
Percent of predicted O2 pulse index was used as a surrogate for stroke volume. The O2 pulse is equal to stroke volume times the arterial-venous O2 content difference. The high coefficient of determination between O2 pulse and % predicted peak VO2 seen in Table 4 is not surprising, because the 2 variables are mathematically related. However, the coefficients of determination for O2 pulse at peak exercise to both % predicted VO2 at VAT and % predicted maximal work rate are also high. Neither of these variables is mathematically related to the O2 pulse at peak exercise. These data suggest that, of the 3 factors responsible for O2 delivery during exercise (HR, arterial O2 content, and stroke volume), stroke volume reserve is almost exclusively responsible for the variation in aerobic performance (as measured by peak VO2, VO2 at VAT, and peak physical working capacity) in Fontan patients.
Determinants of exercise performance
To better understand variation in cardiovascular performance during exercise, we assessed the relationship of a large number of anatomical, surgical, and demographic variables on the primary variable of cardiac performance (% predicted O2 pulse index) and the primary measure of aerobic capacity (% predicted peak VO2) for the maximal capacity group (Table 5). The negative relationship between these 2 dependent variables with puberty, male gender, and BMI Z score is interesting. As the male Fontan subject goes through puberty, the ability of his cardiovascular system to meet the metabolic demands of exercise becomes more compromised (relative to healthy male subjects). This observation is not unexpected, because the increase in muscle mass during puberty is much greater in males than females. With Fontan physiology, the cardiovascular system is unable to meet the metabolic demands of this increased muscle mass during exercise. Consequently, the exercise capacity of the post-pubertal male Fontan patient declines disproportionately compared with the healthy adolescent male.
Data from the adult population suggest that this trend might continue as these patients move into their third decade of life. In a recent study of adults with congenital heart disease, Diller et al. (9) reported a peak VO2 of 19.8 ml/kg/min in 34 subjects (mean age 26.7 years). However, this group was much older at the time of Fontan and might not be comparable to the cohort in this study. These differences emphasize the need for longitudinal studies of this population.
Despite the negative relationship to BMI Z score, poor exercise performance was not secondary to obesity in the pubertal male subjects, because this population had no evidence of being overweight on the basis of the normal BMI Z score distribution (Table 2). This suggests that increased muscle mass, inadequately served by the cardiovascular system, rather than obesity is responsible for this negative relationship.
Mean dP/dtic was the only echocardiographic index related to % predicted O2 pulse in this patient population, suggesting that mean dP/dtic might be a useful geometry-independent index of cardiac performance in Fontan physiology. However, the correlation between % predicted peak O2 pulse index and mean dP/dtic was weak, and no measure of resting systolic or diastolic function was a significant determinant of % predicted peak VO2. More importantly, these observations support the hypothesis that the limited exercise capacity that is typical of the Fontan population is primarily mediated by inadequate stroke volume secondary to an inability to achieve higher levels of transpulmonary flow (i.e., reduced preload) rather than as a consequence of intrinsic limitations of the systemic ventricle.
This study is limited to subjects who survived beyond early childhood and could perform exercise tests and does not reflect the entire spectrum of patients with Fontan procedures. As a cross-sectional rather than a longitudinal study, it is difficult to determine whether some of the observed differences between older and younger subjects are related to the passage of time or to changes in management strategies that have evolved since the subjects of this study began to undergo Fontan procedures. It is also possible that subjects who underwent exercise testing might not be representative of all Fontan survivors.
Although a number of factors might influence a Fontan patient's peak VO2, superior O2 pulse at peak exercise (probably due to higher stroke volume at peak exercise) seems to be the most important factor that distinguishes the high-functioning Fontan patient. Contributions of chronotropic incompetence and arterial desaturation to the Fontan patient's exercise dysfunction seem to be relatively small. The ability of the Fontan circulation to accommodate the metabolic demands of the post-pubertal male's increased skeletal musculature is particularly limited. Resting echocardiographic indexes of ventricular function correlate poorly with exercise function. Other factors, such as increased body mass associated with puberty and male gender, may be more important determinants of exercise function in Fontan physiology.
This work was supported by grants U01-HL068270, U01-HL068269, U01-HL068292, U01-HL068288, U01-HL068285, U01-HL068281, U01-HL068279, and U01 HL068290 from the National Heart, Lung, and Blood Institute, National Institutes of Health/Department of Health and Human Services.
- Abbreviations and Acronyms
- body mass index
- heart rate
- rate of pressure change during isovolumic contraction
- maximal voluntary ventilation
- % predicted
- percent of predicted for age and gender
- respiratory exchange ratio
- ventilatory anaerobic threshold
- minute carbon dioxide production
- minute ventilation
- oxygen consumption
- Received September 25, 2007.
- Revision received February 4, 2008.
- Accepted February 12, 2008.
- American College of Cardiology Foundation
- Fontan F.,
- Baudet E.
- Nir A.,
- Driscoll D.J.,
- Mottram C.D.,
- et al.
- Harrison D.A.,
- Liu P.,
- Walters J.E.,
- et al.
- Durongpisitkul K.,
- Driscoll D.J.,
- Mahoney D.W.,
- et al.
- Troutman W.B.,
- Barstow T.J.,
- Galindo A.J.,
- Cooper D.M.
- Mahle W.T.,
- Wernovsky G.,
- Bridges N.D.,
- Linton A.B.,
- Paridon S.M.
- Dimopoulos K.,
- Okonko D.O.,
- Diller G.,
- et al.
- Diller G.,
- Dimopoulos K.,
- Okonko D.,
- et al.
- Gewillig M.H.,
- Lundstrom U.R.,
- Bull C.,
- Wyse R.K.,
- Deanfield J.E.
- Wasserman K.,
- Hansen J.E.,
- Sue D.Y.,
- Casaburi R.,
- Whipp B.J.
- Gerhard-Paul D.,
- Konstantinos D.,
- Darlington O.,
- et al.
- Sluysmans T.,
- Colan S.D.
- Rhodes J.,
- Udelson J.E.,
- Marx G.R.,
- et al.
- Nikitin N.P.,
- Loh P.H.,
- Silva R.,
- et al.
- Nishimura R.A.,
- Tajik A.J.
- Neter J.,
- Wasserman W.,
- Kutner M.