4.1 Principal findings
In this study of children and adolescents with repaired TOF, important reductions in workload and oxygen consumption were observed. The cohort’s average VO2peak (%predicted = 74%) is consistent with previous studies1 5 10 17,32,33, including a systematic review of repaired TOF encompassing 2262 patients in which % predicted VO2peak = 68%20. However, moderate to severe exercise limitation, as defined by VO2peak < 65% predicted, was found in 26% of patients. The mechanisms contributing to the reduced exercise capacity included arterial desaturation, chronotropic insufficiency and stroke volume increment abnormalities as identified through the heart rate-VO2 relationship.
4.2 CPET challenges in childhood
In children aerobic capacity is strongly influenced by body size and pubertal stage 23. In deriving CPET measurements, differences in body size, sex and age need to be accounted in order to distinguish expected anthropometric maturation from pathological response. In this study parameters were normalized using the formulae and coefficients provided by Blanchard et al26. This group have developed a calibration algorithm incorporating height, weight, age and sex using a healthy Canadian children population aged 12–17 years as reference. In our cohort the patients were UK, Scottish residents, where the age ranged from 6 to 17 years. These demographic differences may potentially render incongruent results particularly in the younger age. For example, we observed a modest negative correlation between age and oxygen consumption, and workload, which may relate to a true age-related decline in exercise capacity or an insufficient correction for age and body size. Despite these caveats we considered the Blanchard equations provided meaningful and consistent results and is now adopted in our laboratory. Ideally, we’d define our laboratory reference normal values by testing healthy children from our resident population.
A submaximal exercise is a test that doesn’t achieve the required intensity of exercise, defined by final RER or peak heart rate targets. Early termination can arise from insufficient motivation 34 or where ongoing strenuous exercise maybe harmful, both of which can apply in children with heart disease. Relying on VO2peak from a test that ends prematurely will underestimate the patient’s functional capacity reducing it prognostic power 35. Normally an adequately performed test is accepted when RER exceeds 1.1, or peak heart rate is greater than 85% of the predicted maximum36,37. Submaximal parameters can provide a useful complement particularly when these targets are not achieved. The OUES has been proposed as a reliable measure of cardiovascular reserve that can be obtained from a submaximal exercise27,38. It measures the absolute increase in VO2 in relation to the log transformed increase in ventilation during exercise. Previous studies have confirmed that this relation is linear throughout the exercise period and that the OUES derived from an exercise that ends prematurely reliably predicts VO2peak 27,38. The OUES reflects cardiovascular, musculoskeletal and respiratory function in a single parameter and has been found to significant predictor for mortality in adult heart failure 39.This study confirmed previous work by finding that (normalized) OUES did not vary with exercise intensity, as quantified by peak RER 27, and was highly correlated with VO2peak 40, a stronger correlation than existed with other submaximal parameters 38. These findings that would support the use of OUES as effort independent measure of functional capacity in similar cohorts.
Although there was a considerable range of exercise duration between patients, from 5 to 12.45 mins (variation of 150%), all tests achieved an RER of 1 or greater, and can be considered as near maximal tests. Therefore, the OUES variation observed with exercise duration is interpreted as a meaningful quantification of the patients’ true functional capacity rather than an indication of OUES effort dependency.
4.3 Causal mechanisms of exercise limitation
4.31 Breathing reserve, arterial desaturation and ventilation response
Patients with repaired TOF can have abnormal baseline spirometry and pulmonary dysfunction that may contribute to exercise intolerance and clinical outcome41,42. Impaired ventilation mechanics can arise as a sequelae of surgery42: pleural adhesions, thoracotomies, phrenic nerve injury and scoliosis 8. Furthermore, previous pathological studies in TOF have demonstrated a deficiency in the number of alveoli which may contribute to hypoplastic lung development43. In this study 40% of patients had evidence of abnormal ventilation response to exercise manifest as breathing reserves < 20% or with peak VE either less than 80% or exceeding 100% of predicted values at termination of exercise. Lower oxygen consumption and workloads occurred in patients with higher breathing reserve. These findings are expected where there’s a dominant cardiovascular cause limiting exercise, as only those patients with less cardiovascular restriction are capable of exercising beyond breathing reserves of less than 20% 30. In the current cohort the values of maximum voluntary ventilation, derived from FEV1 at baseline, fell within the expected range for an age-equivalent health population44. This suggests that the low breathing reserves results from excessive ventilatory response to exercise rather than reduced vital capacity.
Nine percent of patients experienced arterial desaturation during exercise, most likely from increasing right to left shunt via residual intracardiac communication. Patients who desaturated had lower exercise performance, although the findings were not significant. Furthermore, when the presence of desaturation was included in the regression models of exercise capacity, it was also found to be a non-significant predictor. It is possible that the small number of patients available to test this finding was insufficient.
Peak minute ventilation measures the ventilatory response at the termination of exercise. In this study VE varied widely within the cohort, and exceeded the maximum predicted values in one-third of patients. Higher VE was associated with greater exercise capacity and indices of cardiac function: heart rate and O2 Pulse. These findings indicate, that when measured at the termination of exercise (i.e., peak VE), an excessive ventilation response is most apparent when cardiac function is least impaired. Abnormal ventilatory responses to exercise can also be determined over the period of the exercise test by quantifying the slope of the VE/VCO2. In this study inefficient ventilation (VE/VCO2 slope > 34) was apparent in 30% of tests. These results suggest that the high ventilation rates result from an inefficient VE response related to cardiac dysfunction mechanisms, e.g., V/Q mismatch or early onset of decompensated acidosis45, rather than primary lung disease. The OUES can also help to discriminate between cardiovascular and respiratory causation of reduced exercise tolerance 19, and given the reduced OUES found in this cohort further supports a cardiovascular mechanism as the principal cause of exercise impairment in this group.
4.32 Chronotropic incompetence
Chronotropic response to exercise has been previously reported as an important influence on exercise capacity in child-adolescent rTOF population 46 21 31,47. Sinus node dysfunction and abnormal sympathetic cardiac autonomic activity have been identified as mechanisms of blunted HR responses in post-operative CHD 8. In this study, chronotropic incompetence was present in 26% of patients and was associated reduced exercise performance. When analyzed as a continuous variable (i.e., chronotropic index) a greater heart rate increment was associated with higher levels of exercise performed, significantly predicting peak oxygen consumption and workload. Similarly, Diller et al in adult rTOF reported that the heart rate response explained 29% of the variation in VO2peak 13,16. We did not recognize a pattern of negative deflection/flattening in the HR-VO2 slope where VO2 continues to increase without the corresponding linear increase in heart rate within individual studies. This suggests that heart rate limitation had an overall cohort effect and was not limited to a particular subgroup of patients. It is important to note that in the three patients who died, the HR-VO2 gradient and the % increment in O2 Pulse during exercise was lower compared with those alive. Furthermore, these patients had reduced peak heart rates and chronotropic incompetence. This suggests that presence of impaired stroke volume with a coexisting inability to compensate with an adequate heart rate response renders a highly vulnerable clinical situation.
4.33 Oxygen uptake kinetics
Ventricular dysfunction is likely to be another important mechanism for exercise intolerance in repaired TOF 48 33,49,50. Stroke volume, as a measure of ventricular function, can be indirectly assessed using the heart rate–VO2 relationships conditional to the arterial venous oxygen content difference increasing linearly over the exercise period 29. Arterial-Venous O2 content is not routinely measured at CPET and potentially it may be affected by arterial desaturation occurring with right to left shunting in rTOF. Arterial desaturation will tend to reduce HR-VO2 relationship and underestimate the derived stroke volume; as noted, 9% patients experienced arterial desaturation during exercise testing.
Given these caveats, the stroke volume response to exercise was explored using O2 Pulse measured at rest and peak exercise and by the gradient of the heart rate –VO2 relation measured throughout exercise. The HR-VO2 relation (the reciprocal of O2Pulse) as it is obtained over the entire period is less susceptible to error and outlier values compared with point estimates such as rest and peak O2Pulse. Furthermore, it can provide insights into the mechanisms of abnormal response particularly when a non-linear relation manifests.
The magnitude of exercise performed, as quantified by peak workload and oxygen consumption was positively correlated with both the increment in O2 Pulse during exercise and negatively correlated with the gradient of the HR-VO2 relation during the second period of exercise. This would indicate that with reduced exercise capacity an insufficient forward stroke volume response is likely to be an important contributing factor. This is consistent with study by Meadows et al who reported that both reduced VO2 (at anerobic threshold and peak) and reduced O2 Pulse were predicted by impaired RV ejection fraction determined by CMR in young adults with repaired TOF 48. Reduced forward stroke volume response to exercises can also result from severe RV obstructive lesions, severe pulmonary and tricuspid valvular regurgitation following repaired TOF 4 5 9,31.
Under exercise conditions, the heart may respond to compromised stroke volume by an accelerated heart rate as a compensating mechanism to maintain cardiac output and continue exercise. This was evident by the finding of an increase gradient in the HR-VO2 slope during the second period of exercise in a proportion of patient tests. However, with greater restrictions of stroke volume exercise will terminate early despite an increased heart rate response as it is insufficient to compensate. In the current study the HR-VO2 relation became non-linear, with an upward deflection in ~ 40% of patients. These patients had significantly lower exercise capacity compared with patients in whom the HR-VO2 relation remained linear throughout exercise. This suggests that stroke volume impairment was most implicated in this ‘non-linear’ sub-group rather than a phenomenon affecting the entire cohort 16. However, this may ignore patients where exercise terminates early and where the HR-VO2 relation remains linear during that period. This may occur in the situation of impaired stroke volume with minimal capacity to provide compensatory increase in heart rate i.e., with co-existing chronotropic insufficiency, a situation described as ‘pseudo-normalization’ of the HR versus VO2 relationship 31
4.4 Clinical implications
Reliably quantifying functional capacity in children in the clinic can be challenging due to patients’ self-perception of ‘normal’ physical capacity, communication barriers and life style adjustments 13 2. Important decline in cardiopulmonary functioning can be anticipated in rTOF 7 51 and may occur prior to the overt expression of symptoms delaying treatment and risking potentially avoidable morbid events 21 47.
Previous studies have shown that CPET can predict morbidity and mortality in rTOF in adult14 and child-adolescent populations 52 41. In this study univariable analysis identified several uncorrected CPET indices, including workload, VO2peak, O2Pulse and peak ventilation, that were associated with adverse clinical status as defined by symptoms or death. By contrast only reduced workload (% pred), as a normalized parameter, predicted adverse clinical status, while other normalized indices including VO2peak did not. This difference between the predictability of absolute and corrected indices might arise because when normalizing for sex, age and body size the range of variation within each individual CPET parameter tends to reduce in magnitude, which may decrease the sensitivity of the analysis. Also, our study was based on retrospective data where symptoms were informally reported and also there may have been insufficient patient-events to confidentially confirm relationships between CPET and clinical outcome. Ideally the use of standardized questionnaires to gauge symptomatic status and relate to CPET measurements would be preferred. Given these caveats this study suggests that CPET may support the clinical assessment and provide objective confirmation of functional status33.
Reintervention in rTOF is common particularly on the RVOT to relieve obstruction or pulmonary regurgitation and prevent further RV dilation and decline 53. In this study increased gradient of the HR–VO2 slope predicted future RVOT intervention. As discussed, an increase in the HR-VO2 gradient can indicate an impaired stroke volume response to exercise and given that reduced RV and LV stroke volume is a known consequence of RVOT obstruction and regurgitation 4, it seems plausible that the HR-VO2 relation is providing an important insight. In this study, decision-making on RVOT intervention was based on RV pressure/volume loading criteria and not CPET data a priori. Nonetheless, CPET has provided additional functional data that may support the decision for RVOT intervention specifically by identifying patients with impaired stroke volume response54.
The three patients who died had a significantly higher VE/VCO2 slope compared with those alive (absolute values 39.2 vs. 31.8). The association of higher VE/VCO2 slope with event free survival has been a consistent finding in previous non-congenital heart failure studies39,45,55,56 and adult TOF populations41,54,57. Given the findings of trend to lower heart rates and O2 Pulse increment in those that died, suggest that patients with a combination of impaired stroke volume response and reduced exercise heart rates are particularly vulnerable.
4.5 Limitations
This is an observation study based on retrospective data therefore constitutes an exploratory analysis. Further work is needed to reproduce the current findings in a larger cohort and in particular to compare the sensitivity and specificity of exercise indices as predictors of clinical outcome. This cohort may not represent the general population of repaired TOF as the patients were selected for CPET at the cardiologists’ discretion based on a variety of reasons: assessment for PVR, symptomatic and benchmark prior to ACHD transfer. Within the cohort there existed a range of morphologies that fell under the umbrella term of repaired TOF, for example pulmonary atresia with MAPCA, and with this variation in complexity an impact on exercise tolerance would be anticipated. The application of Blanchard calibration for paediatric CPET indices may be vulnerable to error as the study cohort differed in age distribution, country of origin and contained patients with BMI-for-age +- 2 z when compared to their healthy Canadian reference population.
4.6 Conclusions
From this study, overall children and adolescents with repaired TOF demonstrated low-normal exercise capacity as quantified by peak oxygen consumption. However, within the cohort 25% had moderate to severe impairment of functional capacity due predominately to cardiovascular mechanisms of heart rate limitation and insufficient stroke volume response with exercise. Although abnormalities of ventilation and arterial desaturation occurred less frequently these may be important additional contributors influencing exercise capacity on an individual basis. CPET may provide additional collaborative data to quantify clinical status and guide decision-making on RVOT re-intervention.