Patients with repaired tetralogy of Fallot have a reduced percentage of predicted peak oxygen consumption (VO 2 ) and percentage of oxygen pulse (O 2 P%) compared to healthy controls. Because data regarding the progression of exercise intolerance in these patients is limited, we sought to analyze the serial exercise data from patients with Tetralogy of Fallot to quantify the changes in their exercise capacity over time and to identify associations with clinical and cardiac magnetic resonance imaging variables. The data from cardiopulmonary exercise tests (CPXs) from 2002 to 2010 for patients with repaired tetralogy of Fallot with ≥2 CPXs separated by ≥12 months were analyzed. Tests occurring after interventional catheterization or surgery were excluded. A total of 70 patients had 179 CPXs. They had a median age at the initial study of 23.6 years and an interval between the first and last CPX of 2.8 years. At the initial CPX, the peak VO 2 was 27.6 ± 8.8 ml/kg/min (78 ± 19% of predicted), and the peak O 2 P% was 89 ± 22% of predicted. At the most recent study, the peak VO 2 averaged 25.0 ± 7.4 ml/kg/min (73 ± 16% of predicted), and the peak O 2 P% averaged 83 ± 20% (p <0.01) for each versus the initial CPX. The decrease in the peak VO 2 was strongly associated with a decrease in O 2 P% and an increase (worsening) in the slope of the minute ventilation-versus-carbon dioxide production relation. Changes in the peak VO 2 did not correlate with concomitant changes in any other CPX variable. The rate of decrease was not related to a history of shunt palliation, age at CPX, or any other baseline clinical parameter, including cardiac magnetic resonance measurements. In conclusion, the exercise capacity of patients with repaired tetralogy of Fallot tends to decrease over time. This deterioration is variable and unpredictable and is primarily related to a decrease in the forward stroke volume at peak exercise.
Currently, information regarding the change in the exercise capacity over time in patients with congenital heart disease in general, and those with repaired tetralogy of Fallot (rTOF) in particular, is limited. In this population, several cross-sectional studies have reported an average peak oxygen consumption (VO 2 ) of 51% to 95% predicted. Given the greatly increased hemodynamic demands imposed on the right ventricle during exercise, the low peak VO 2 of these patients with residual right-sided heart disease is not surprising. The potential factors responsible for the depressed exercise capacity of patients with rTOF are numerous; past studies have implicated residual pulmonary regurgitation, pulmonary artery distortion, impaired lung function, chronotropic impairment, and ventricular dysfunction. However, a study of the natural history of the exercise function of patients with rTOF, according to assessments using modern cardiopulmonary exercise testing (CPX) technology, has not been undertaken. The purpose of the present study was to analyze the serial CPX data from patients with rTOF to quantify the changes in their exercise capacity over time. We also sought to identify the clinical and cardiac magnetic resonance imaging variables associated with any observed changes in exercise function.
We identified all patients with rTOF who had undergone CPX testing at our institution from 2002 to 2010. The patients were included if they had undergone ≥2 symptom-limited, progressive CPX tests, separated by ≥12 months, without an intervening cardiac surgical operation or interventional cardiac catheterization procedure. Patients with TOF and pulmonary atresia and patients with significant residual right ventricular outflow tract obstruction (mean gradient >30 mm Hg on the most recent echocardiogram) were excluded. To avoid confounding the present study of the natural history of the patient with rTOF with the complicated natural history of artificial conduits, we excluded the studies from patients with right ventricle to pulmonary artery conduits. To minimize the potential confounding effects of inadequate patient effort, we excluded the data from studies in which the patient did not achieve a respiratory exchange ratio at peak exercise of ≥1.09.
Doppler estimates of right ventricular outflow tract obstruction and data from cardiac magnetic resonance studies performed within 12 months of the first CPX, when available, were included in the present analysis. Additionally, co-morbidities were recorded from the closest clinic visit records.
During the exercise tests, electrocardiographic monitoring and breath-by-breath expiratory gas analysis were performed using the CardiO2 exercise testing system (Medical Graphics, Minneapolis, Minnesota). Cuff blood pressure determinations and complete 12-lead electrocardiograms were obtained at 2- to 3-minute intervals during exercise, at peak exercise, and at 1, 3, and 5 minutes after exercise. Pulse oximetry oxygen saturation was monitored throughout the study. Immediately before each exercise test, spirometric measurements of the patients’ forced vital capacity and volume of air exhaled in the first second of forced expiration were also obtained.
The temporal changes in the peak VO 2 (the most widely used index of exercise function) and the oxygen pulse at peak exercise (O 2 P, a surrogate for the forward stroke volume at peak exercise ) were the primary outcome variables for the present study. Because of the variation in patient age, size, and gender in our cohort, and because many of our subjects grew significantly and underwent pubertal-related changes in stature and body habitus during the study period, our analyses focused on the changes in the percentage of the predicted values (VO 2 % and O 2 P%), rather than the changes in the absolute magnitude or weight-normalized values of these variables. The secondary outcome parameters of exercise performance included changes in the percentage of predicted peak heart rate, the slope of the minute ventilation-versus-carbon dioxide production relation (V E /VCO 2 slope, an index of the efficiency of gas exchange during exercise ), oxygen saturation, and spirometric measurements. We also calculated the body mass index (weight in kilograms divided by the height in square meters) at each exercise test.
The clinical variables included gender, anatomic diagnoses, type of previous surgical procedures (e.g., early shunt vs primary intracardiac repair), age at repair and at CPXs, and Doppler-estimated right ventricular outflow tract gradients. When available, the pulmonary regurgitation fraction, indexed right ventricular end-diastolic and systolic volume, right ventricular ejection fraction, indexed left ventricular end-diastolic and systolic volumes, and left ventricular ejection fraction were collected from the cardiac magnetic resonance studies.
Continuous variables are presented as the mean ± SD and the categorical variables as the counts and percentages. For continuous variables with non-normal distributions, we report the medians and ranges. Paired t tests were used to compare the initial and final values for each CPX variable. To identify the factors associated with a steeper decrease in exercise function, the linear regression line of each exercise test variable against time served as each patient’s measure of the rate of change over time. One-sample t tests were used to test for significant changes over time. In this and subsequent analyses, observations were weighted by the interval between the first and last exercise tests to account for the varying lengths of follow-up. We used Pearson’s correlation coefficient to estimate the association between the rate of change in the VO 2 % (ΔVO 2 %) and the concomitant change in other exercise variables, the initial values of each exercise variable, and the cardiac magnetic resonance variables. We used Spearman’s rank correlation coefficient to estimate the association between ΔVO 2 % and age at initial CPX, age at surgery, and right ventricular outflow tract gradient. Comparisons of the mean ΔVO 2 % by gender, transannular patch status, and previous shunt palliation were made with 2-sample t tests. Multivariate regression analysis was used to identify the independent predictors of ΔVO 2 %. We compared the clinical characteristics of the patients with high rates of decrease in the peak VO 2 % (>4% point decrease/year) to those whose demonstrated a >1% point increase/year using chi-square tests and Wilcoxon rank sum tests. The cardiac magnetic resonance variables in these 2 groups were compared using 2-sample t tests. Analyses were performed using SAS software, version 9.2, SAS System for Windows (SAS Institute, Cary, North Carolina).
We identified 70 patients (53% male) with a total of 179 CPXs (mean 2.6 studies/patient) who met the inclusion criteria. The age at the first CPX was 27.8 ± 15 years (range 8.2 to 61.4). The interval between the first and last CPX was 2.7 ± 1.5 years (range 1.0 to 7.2). The patients’ initial TOF repairs were undertaken at a median age of 2.3 years (range 0.1 to 21.6). Of the 70 patients, 44 had undergone transannular repairs and 26 had nontransannular right ventricular outflow patches. The mean residual right ventricular outflow tract gradient was 8.6 ± 13.0 mm Hg. Although most patients had undergone primary complete TOF repair, 17 (24%) had had a palliative shunt placed before the full repair.
The CPX data from the first and last tests are summarized in Table 1 . The peak VO 2 on the initial CPX was mild to moderately depressed (27.6 ± 8.8 ml/kg/min; 78 ± 19% of predicted). The peak VO 2 on the patients’ final CPX averaged 25.0 ± 7.4 ml/kg/min (73 ± 16% of predicted; p ≤0.01 compared to the initial CPX). The decrease in the O 2 P was of a similar magnitude: 89 ± 22% of predicted at the initial CPX and 83 ± 20% of predicted at the final CPX (p <0.01). Statistically significant changes over time were not observed in the heart rate at peak exercise, the V E /VCO 2 slope, or the baseline spirometric measurements. A small, but statistically significant, increase in the body mass index was observed during the study course; however, the body mass index Z score did not change ( Table 1 ). The mean annual change in VO 2 % and O 2 P% (ΔVO 2 % and ΔO 2 P%) was −1.4 ± 9.2% and −1.8% ± 11.4% points annually, respectively ( Table 2 ). However, a wide variation was seen in the response over time ( Figure 1 ). In 20 patients (29%), the peak VO 2 decreased by >4% annually. In contrast, the peak VO 2 increased by ≥1% annually in 23 patients (33%).
|Variable||First Test||Last Test||Difference||p Value ⁎|
|Peak oxygen consumption (ml/kg/min)||27.6 ± 8.8||25.0 ± 7.4||−2.5 ± 4.5||<0.0001|
|Peak percentage of predicted oxygen consumption (%)||78 ± 19||73 ± 16||−4 ± 14||0.01|
|Oxygen pulse (ml/beat)||10.6 ± 3.3||10.8 ± 3.1||0.1 ± 2.4||0.65|
|Peak percentage of predicted oxygen pulse (%)||89 ± 22||83 ± 20||−6 ± 18||0.01|
|Slope of minute ventilation vs carbon dioxide production relation||28.2 ± 4.6||27.7 ± 4.1||−0.5 ± 4.3||0.37|
|Oxygen saturation at rest (%)||98 ± 2||98 ± 1||−0.5 ± 0.2||0.59|
|Peak oxygen saturation (%)||97 ± 2||97 ± 3||−0.6 ± 0.2||0.46|
|Percentage of predicted forced vital capacity (%)||82 ± 17||81 ± 17||−1 ± 9||0.50|
|Percentage of predicted volume of air exhaled in first second of forced exhalation (%)||81 ± 16||81 ± 16||0 ± 7||0.99|
|Heart rate at peak exercise (beats/min)||160.9 ± 21.1||160.0 ± 24.4||−0.9 ± 18.2||0.67|
|Body mass index (kg/m 2 )||24.0 ± 5.6||25.2 ± 6.0||1.3 ± 2.0||<0.0001|
|Body mass index Z score||0.05 ± 1.12||0.11 ± 1.18||0.06 ± 0.44||0.26|
|Variable||Rate of Change per Year of Follow-Up (Mean ± SD)||p Value ⁎|
|Peak oxygen consumption (ml/kg/min/year)||−0.8 ± 3.1||<0.001|
|Peak percentage of predicted oxygen consumption (% points/year)||−1.4 ± 9.2||0.04|
|Peak percentage of predicted oxygen pulse (% points/year)||−1.8 ± 11.4||0.02|
|Slope of minute ventilation vs carbon dioxide production relation (change/year)||−0.2 ± 3||0.38|
|Percentage of predicted forced vital capacity (% points/year)||−0.3 ± 5.5||0.49|
|Percentage of predicted volume of air exhaled in first second of forced exhalation (% points/year)||−0.1 ± 4.6||0.86|
|Heart rate at peak exercise (beats/min/year)||−0.4 ± 11.3||0.60|
|Peak percentage of predicted heart rate (% points/year)||0.3 ± 5.8||0.40|
|Body mass index (kg/m 2 per year)||0.4 ± 1.4||<0.001|
|Body mass index Z score (change/year)||0.02 ± 0.3||0.37|
The bivariate correlation analysis revealed a strong association between the ΔVO 2 % and the ΔO 2 P% ( Figure 2 and Table 3 ). A more rapid decrease in peak VO 2 was also strongly associated with a concurrent increase (worsening) in the V E /VCO 2 slope. Changes in the peak VO 2 did not correlate with concomitant changes in any other CPX variable. Patients with the greatest initial peak VO 2 and greatest initial O 2 P tended to have a steeper decrease in the peak VO 2 during the follow-up period ( Table 3 ). However, no association was found between the rate of decrease and age at CPX, age at reparative surgery, gender, use of a transannular patch, history of previous shunt palliation, or any other baseline clinical parameter, including heart rate, V E /VCO 2 , forced vital capacity, volume of air exhaled in the first second of forced expiration, or body mass index ( Table 4 ). Multivariate analysis revealed that only the ΔO 2 P% (p <0.001) and, to a lesser extent, the initial peak VO 2 % (p = 0.03) correlated significantly with the ΔVO 2 %.
|Rate of change in exercise test parameter|
|Peak percentage of predicted oxygen pulse (%)||0.79||<0.0001|
|Slope of minute ventilation vs carbon dioxide production relation||−0.38||0.001|
|Percentage of predicted forced vital capacity (%)||0.02||0.88|
|Percentage of predicted volume of air exhaled in first second of forced exhalation (%)||0.04||0.72|
|Percentage of predicted heart rate (%)||0.14||0.26|
|Body mass index (kg/m 2 )||0.13||0.27|
|Body mass index Z score||0.15||0.21|
|Initial exercise test parameter|
|Peak percentage of predicted oxygen consumption (%)||−0.49||<0.0001|
|Peak percentage of predicted oxygen pulse (%)||−0.38||0.001|
|Slope of minute ventilation vs carbon dioxide production relation||0.2||0.09|
|Percentage of predicted forced vital capacity (%)||−0.09||0.46|
|Percentage of predicted volume of air exhaled in first second of forced exhalation (%)||−0.11||0.38|
|Percentage of predicted heart rate (%)||−0.19||0.11|
|Body mass index (kg/m 2 )||−0.01||0.95|
|Body mass index Z score||−0.08||0.52|
|Characteristic||Correlation or Mean ± SD||p Value|
|Age at initial exercise test (years)||0.09||0.44|
|Age at surgery (years)||−0.16||0.18|
|Right ventricular outflow tract gradient (mm Hg)||0.06||0.63|
|Female (% point change in peak percentage of predicted oxygen consumption/year)||−0.9 ± 9.4|
|Male (% point change in peak percentage of predicted oxygen consumption/year)||−1.8 ± 9.1|
|No (% point change in peak percentage of predicted oxygen consumption/year)||−2.6 ± 9.1|
|Yes (% point change in peak percentage of predicted oxygen consumption/year)||−0.7 ± 9.2|
|Previous shunt palliation||0.05|
|No (% point change in peak percentage of predicted oxygen consumption/year)||−2.2 ± 9.2|
|Yes (% point change in peak percentage of predicted oxygen consumption/year)||0.7 ± 9.0|
A subset of 37 patients had cardiac magnetic resonance studies within 12 months of their initial CPX ( Table 5 ). No cardiac medication changes were made between the magnetic resonance study and the first CPX. The mean right ventricular end-diastolic volume Z score was 4.3 ± 2.4; the right ventricular ejection fraction was 51 ± 7%. The pulmonary regurgitation fraction averaged 35 ± 18% (range 1% to 67%). The mean left ventricular end-diastolic volume was normal, and the left ventricular ejection fraction was low-normal at 59 ± 7%. Changes in the peak VO 2 during the study course did not correlate with any of the baseline cardiac magnetic resonance measurements.
|Parameter||Patients (n)||Mean ± SD||Correlation With ΔVO2%||p Value|
|Right ventricular end-diastolic volume (ml/m 2 )||36||254.1 ± 80.9||0.04||0.81|
|Right ventricular end-diastolic volume Z score||36||4.3 ± 2.4||0.09||0.60|
|Right ventricular end-systolic volume (ml/m 2 )||36||127.1 ± 50.6||0.04||0.83|
|Right ventricular ejection fraction (%)||36||51 ± 7||−0.04||0.83|
|Left ventricular end-diastolic volume (ml/m 2 )||36||150.6 ± 52.2||−0.17||0.33|
|Left ventricular end-diastolic volume Z score||35||0.4 ± 1.2||−0.10||0.58|
|Left ventricular end-systolic volume (ml/m 2 )||36||64.3 ± 26.0||−0.07||0.69|
|Left ventricular ejection fraction (%)||36||59 ± 7||−0.01||0.95|
|Pulmonary regurgitation fraction (%)||37||35 ± 18||0.26||0.12|