Exercise Capacity in Patients with Pulmonary Atresia with Intact Ventricular Septum: Does the Type of Surgical Repair Matter?

Two standard surgical palliative options for neonates born with pulmonary atresia and intact ventricular septum (PA/IVS) include uni-or biventricular repair. Whenever feasible, the biventricular repair is considered to have better exercise capacity (XC) and outcomes. However, there is a paucity of data comparing objective XC between these two surgical techniques. Our aim was to compare XC, including longitudinal changes in patients with PA/IVS following uni-biventricular repair. We performed a single-center retrospective study of survivors with repaired PA/IVS who underwent comprehensive treadmill cardiopulmonary exercise testing. Initial and latest exercise parameters were compared for longitudinal analysis. Demographic and exercise parameters were collated. Peak oxygen uptake (VO2 in ml/kg/min), an indicator of maximal aerobic capacity, peak heart rate, and other measures of spirometry performed at the same time were collected. Recorded parameters included, (a) Percentage of predicted VO2 (% VO2) normalized for age, weight, height, and gender, (b) % oxygen (O2) pulse, (c) anaerobic threshold (AT), (d) Chronotropic index (CI), (e) % Breathing reserve, (f) Forced vital capacity (FVC), (g) % Forced Expiratory volume in 1 s (FEV1), (h) Maximum voluntary ventilation (MVV), and (i) VE/VCO2. Appropriate statistical tests were performed, and a p value < 0.05 was considered significant. A total of 35 patients (43% male, 57% univentricular repair) were included, with a mean (SD) age of 20.1(7.5) years. Patients with univentricular palliation demonstrated significantly impaired peak heart rate, chronotropic index (0.50 ± 0.2 vs. 0.90 ± 0.1, p = 0.02), VE/VCO2 (35.4 ± 5.0 vs. 30.2 ± 2.8, p = 0.001), and %FVC (78.3 ± 8.3 vs. 88.6 ± 15.1, p = 0.02). There was a trend towards reduction in % VO2 in the Fontan patients though it was statistically similar between the groups (68.4 ± 21.4 vs. 81.2 ± 18.9, p = 0.07). Longitudinal data were available for 11 patients in each group, and there was no longitudinal decline in their exercise parameters over similar intermediate follow-up duration [6.8 (UV) vs. 5.3 (BV) years]. We conclude that young survivors with PA/IVS with prior univentricular palliation demonstrated an objective impairment in their chronotropic parameters compared with the biventricular repair. However, this did not translate into a significant difference in their exercise capacity. There was no longitudinal decline in exercise capacity or other parameters over intermediate follow-up.


RV
Right ventricle RVOT Right ventricular outflow tract SD Standard deviation SpO 2 Oxygen saturation TOF Tetralogy of fallot UV Univentricular VAT Ventilatory anaerobic threshold VCO 2 Carbon dioxide elimination VE Pulmonary ventilation VO 2 Peak oxygen uptake XC Exercise capacity

Introduction
Pulmonary atresia with intact ventricular septum (PA/IVS) is a rare congenital heart disease with a reported incidence of 0.6/10,000 live births [1]. There is a wide morphological spectrum from membranous pulmonary valve atresia to muscular right ventricular outflow tract (RVOT) obstruction. The latter is usually associated with severe RV hypoplasia. The surgical palliation approach is contingent upon the size of the RV, the morphology of the tricuspid valve, and the presence of RV-dependent coronary artery circulation [1]. The surgical options include a biventricular repair with RVOT reconstruction, 1.5 ventricle repair with a Glenn shunt, or Fontan/univentricular palliation. Biventricular repair is favored if RV is deemed to be of adequate size and devoid of RV-dependent coronary circulation [2]. In addition, the biventricular repair is 'anticipated' to engender improved long-term outcomes. Thus, there is a proclivity for persevering to achieve this 'goal' of biventricular repair whenever deemed feasible. Exercise stress testing is the established standard for assessing functional status and prognosis in patients with repaired congenital heart defects [3,4]. Patients with Fontan palliation have a decrease in exercise capacity over time [5]. However, data on the comparison of exercise capacity in patients with PA/IVS who underwent univentricular vs. biventricular repair, is limited, and the results are conflicting. In addition, data on the longitudinal changes in exercise capacity in this particular population is lacking [6][7][8][9]. Romeih et al. and Ekman-Joelsson et al. reported a better exercise capacity (XC) with biventricular repair in patients with PA/IVS [7,8]. Sanghavi et al. reported no difference between the two surgical techniques [6]. A multicenter study reported higher VO 2 in patients who had undergone 1.5 ventricle repair compared to uni/biventricular repair [9].
The primary aim of our study was to compare XC in patients with PA/IVS who underwent univentricular vs. biventricular repair. The secondary aim was to explore longitudinal changes in the exercise capacity in this patient population.

Study Population
This is a single-center retrospective chart review of patients with PA/IVS following uni or biventricular repair that successfully completed at least one comprehensive cardiopulmonary exercise treadmill stress test(s) over the last 20 years. Institutional Review Board (IRB) approval was obtained. Our exclusion criteria included patients who (i) had 1.5 ventricular repair or a pacemaker, (ii) were unable to complete the stress test, or (iii) had incomplete data.
The first exercise stress test available after the completion of the surgical palliation, and the latest tests were analyzed. Demographic parameters including age at exercise, height, weight, BMI, and gender were collated. Additional relevant clinical information, including medications, was obtained.

Cardiopulmonary Exercise Test
The exercise stress test was performed using a treadmill (Track master Treadmill 425 series) and modified Bruce Protocol [10]. Prior to the test, baseline blood pressure (BP), heart rate (HR), and oxygen saturation (SpO 2 ) were obtained. A baseline pulmonary function test (PFT) was obtained by attaching a flow meter through a microfilter (Carefusion) to the patient's mouth.
During the exercise testing, the speed of the treadmill was increased every minute in a stepwise manner to obtain the optimal exercise in 10-12 min. A continuous 12 lead electrocardiogram (EKG) and SpO 2 monitoring were performed during the exercise. BP was obtained every 3 min during the exercise. Peak VO 2 , carbon dioxide elimination (VCO 2 ), and minute ventilation (MV) were measured continuously via a mouthpiece. The test was concluded if the patient was exhausted, if the patient experienced symptoms like palpitations, chest pain, dyspnea, or an abnormal response noted in HR or BP.
The VMAX Encore 29 metabolic cart (CareFusion, California, USA) was used for the measurement of Maximum Voluntary Ventilation (MVV), VO 2, and end-tidal CO 2 . Additional variables recorded included the ventilatory anaerobic threshold (VAT), oxygen pulse, % oxygen pulse, peak VO 2, and % VO 2 .

Peak VO 2
An indicator for maximal aerobic exercise capacity and was defined as the highest VO 2 obtained during the exercise for ≥ 20 s. VO 2 was indexed as % VO 2 (percentage of predicted VO 2 , normalized for age, sex, weight, and height matched controls. An abnormal exercise capacity was defined as % VO 2 < 85.

Oxygen Pulse
This was calculated as VO 2 (mL/min) per heartbeat (VO 2 /HR). %Oxygen pulse was defined as the oxygen pulse expressed in reference to published normative data (matched to age, sex, weight, and height).

VE/VO 2 and VE/VCO 2
These were derived parameters by calculating the ratio between the pulmonary ventilation and the volume of oxygen consumed or CO 2 eliminated, respectively.

Ventilatory Anaerobic Threshold (VAT)
This was defined as the VO 2 at which the rate of increase in carbon dioxide production exceeds the uptake of oxygen. VAT was estimated via the slope method. VO 2 was represented on the abscissa/X-axis, and VE/ VO 2 and VE/ VCO 2 were plotted on the ordinate/Y-axis. VAT was defined as the value of VO 2 at which point VE/ VO 2 begins to increase without a corresponding increase in VE/ VCO 2 . This parameter was also expressed as %VAT with respect to the peak VO 2 .

Peak Heart Rate
This was defined as the maximal heart rate achieved during incremental exercise. It was also expressed as %HR (% of predicted peak HR). The predicted peak heart rate is calculated as 220-age in years.

Spirometry Parameters
Forced Expiratory Volume in 1 s or FEV 1 -Defined as the volume of air (measured in mL) exhaled in the first second of forceful expiration after a maximal inspiration.
Forced Vital Capacity or FVC-Defined as the total volume of air (measured in mL) that can be forcefully exhaled after maximal inspiration.
Maximal Voluntary Ventilation or MVV (L/min)-Defined as the volume of air inspired/expired during rapid and deep breathing for 12 s (L/min). This was also calculated as follows: MVV = 40*FEV1.

Breathing Reserve
This was derived as follows: (MVV-VE max) *100/MVV. A value < 20 was considered to be abnormal.

Statistical Analysis
SPSS version 20 for Windows was used for statistical analysis. All continuous variables were expressed as mean ± SD and categorical variables as percentages. The entire cohort was divided into two groups biventricular vs. univentricular repair. Various demographic, clinical, and exercise parameters were compared between the two groups using the student t test and Chi-square when appropriate. For longitudinal analysis, the paired t test and chi-square test were used for comparing various exercise parameters. A p value of less than 0.05 was considered statistically significant.

Results
Exercise data was available for 40 patients with a diagnosis of PA/IVS. However, five patients had undergone 1.5 ventricular repair and were excluded. Therefore, 35 patients were enrolled for further analysis. (Table 1) Thirty-five patients (43% male, univentricular repair 57%) underwent an exercise stress test at a mean (SD) age of 20.1 ± 7.5 years. Amongst patients with a univentricular/ Fontan palliation [n = 20 (75%)], the majority underwent a lateral tunnel iteration (n = 18; 90%) and only 2 (10%) underwent extra-cardiac type of repair. All patients with a biventricular repair (n = 15) underwent a transannular patch. There was no enrolled patient in the biventricular repair group who had surgical palliation using an RV to PA conduit.

Demographic and Clinical Characteristics at Latest Test
There was no significant difference in the baseline demographic/clinical parameters between patients with univentricular vs. biventricular repair (Table 1). Concurrent treatment with an atrioventricular (AV) nodal blocking agent (beta-blocker, calcium channel blocker, or digoxin) was also not significantly different (25 vs. 0%, p = 0.06) between patients with uni vs. biventricular repair, respectively.

Longitudinal Follow Up: Initial vs. Latest Test (Tables 4, 5, 6, 7)
Data for at least two exercise stress tests were available in 11 patients in each group. There was no significant longitudinal change in their cardiopulmonary profile. Fontan group had a similar % VO 2 and Chronotropic index at initial and latest test (% VO2: 73.6 ± 13.4 vs. 63.3 ± 21.8, p = 0.20, CI 0.6 ± 0.2 vs. 0.6 ± 0.2, p = 0.80) as well as the biventricular repair group (% VO2: 78.6 ± 16.2 vs.  71.0 ± 14.8, p = 0.26, CI 0.8 ± 0.1 vs. 0.9 ± 0.1, p = 0.11). When data were dichotomized, there was no significant difference in the number of patients with abnormal VO 2 , CI, and PFT at initial and latest tests in either of the groups.

Discussion
Our cohort of children and young adult survivors with PA/ IVS showed an overall reduced exercise capacity late into their second decade, and almost 3/4th of them had abnormal % VO 2 . The cardiopulmonary profile was further impaired in those with univentricular repair compared to the biventricular repair. This was particularly pronounced for various chronotropic parameters and reflected as a high prevalence of abnormal CI (90% vs. 10%) in the univentricular palliation group. Significant reductions in their mean VO 2 (ml/Kg/ min) and chronotropic indices did not translate into a statistically relevant decrease in the indexed measure of exercise capacity (% VO 2 ) (68.4 ± 21.4 vs. 81.2 ± 18.9, p = 0.07). However, there were no significant differences in VO 2 < 85% and PFT between the two groups. There was no significant longitudinal decline in cardiopulmonary exercise profile for a limited subset of eleven patients within each group over intermediate follow-up duration. This trend of a directionally, but not a statistically discernible reduction in % VO 2 noted in our univentricular patients is also reported by our predecessors. Sanghavi et al. reported similar % VO 2 (76.0 ± 17.5 vs. 83.5 ± 21.1, p = NS) amongst patients with PA/IVS who underwent uni (n = 10) or biventricular (n = 19) repair, respectively. They reported that older age correlated with a reduced XC [6]. Another multicenter study reported comparable % VO2 (63.96 ± 10.45 vs. 75.78 ± 15.98, p = NS) amongst patients following uni (n = 22) or biventricular (n = 45) repair respectively. There was a positive correlation between % VO 2 and initial tricuspid valve z score across all groups (p < 0.001) [9]. In contrast, Romeih et al. reported a better exercise capacity, % VO 2 (92.7 ± 6.7 vs. 55.7 ± 10.8, p < 0.001) amongst patients with PA/I. VS who underwent BV repair (n = 7) as compared to Fontan (n = 7) repair, respectively [7].
The chronotropic impairment noted in our patients with the univentricular repair is not only consistent with previous studies in this specific subset [6,7], but also generalizable for a Fontan population [12][13][14]. Traditionally, this is attributed to sinus node dysfunction secondary to multiple surgical procedures, mainly atriotomy. However, Claessen et al. reported an appropriate increase in HR relative to the intensity of exercise in patients with Fontan palliation. They speculated that an adaptive mechanism might operate to optimize heart rate and cardiac output in such patients by preventing excessive tachycardia related ventricular under filling [15]. Abnormal pulmonary function was seen in 24 (68%) patients in our cohort, with no difference between the two groups. Specifically, patients with univentricular palliation demonstrated restrictive lung physiology. There is a paucity of comparable data other than a small study that showed no difference in spirometry parameters in patients with PA/IVS following uni or biventricular repair [6]. Nevertheless, abnormal lung function is well described in patients with repaired congenital heart defects [16,17]. Restrictive pulmonary function, in particular, is prevalent in Fontan as well repaired TOF population, where it is associated with reduced aerobic capacity [18][19][20]. Abnormal lung development and multiple sternotomies/ thoracotomies have been implicated for the same. Matthews et al. documented a negative correlation between the number of thoracotomies and FVC [21]. This, in addition to chronic pulmonary vascular changes in Fontan patients, could explain a poorer lung function as compared to the BV repair.
Our study demonstrated a significantly impaired VE/ VCO 2 ratio in Fontan patients as compared to the BV repair (35.4 ± 5.0 vs. 30.2 ± 2.8, p = 0.002). An elevated VE/VCO 2 ratio has been shown in Fontan patients [22]. It reflects an abnormal cardiac output response to exercise, leading to a mismatch between ventilation and perfusion in the lungs.
To our knowledge, this is the first study to explore longitudinal changes in XC in patients with PA/IVS. In a subset of our patients (11 within each group) with longitudinal data, there was no demonstrable change in their exercise parameters over intermediate follow-up. This is in contrast to published data in patients with Fontan or BV repair [23][24][25][26]. In a multicenter study by Atz et al. Fontan patients showed a decrease in % VO 2 (n = 95), AT% (n = 196), and % O 2 pulse (n = 95) over a mean time interval of 9.5 years, but their chronotropic index remained similar [23]. Kipps et al. noted a longitudinal decrease in % VO 2 (78 ± 19 vs. 73 ± 16, p = 0.01) in patients with repaired Tetralogy of Fallot (n = 70) over a mean duration of 2.8 years [24]. This was associated with a concurrent decrease in % O 2 pulse. More long-term data is needed in a larger cohort of patients with PA/IVS for further validation.
The findings based on analysis of data in young survivors with PA/IVS who were mainly late in their second decade should be interpreted cautiously. Taken together (Table 8), hitherto available data in PA/IVS population may share a subtle bias towards a numerically better value for exercise capacity with a biventricular repair. However, robust statistical validation of a tangible improvement in an indexed parameter (% VO 2 ) is still lacking. This poses some unique questions regarding the robustness of an approach geared towards the 'quintessential' quest for a biventricular repair in this subset. However, our study schema is not intended to comment on the superiority of one stratagem over another.
The anticipated improved outcome of a biventricular repair is rooted in the deep-seated notion of its physiological superiority due to its retention of a pulsatile subpulmonary RV. Lack of this mechanism may be associated with pulmonary vascular endothelial dysfunction  [27]. The mechanics of passive Fontan circulation may promulgate further, leading to impaired left ventricular filling with elevated filling pressures, limited augmentation of preload during stress, and thus a blunted cardiac output [28]. This construct has heralded an era of aggressive pursuit of biventricular repair for PA/IVS, circumventing an alternative approach even when the physiognomy of the RV may be somewhat sub-optimal. However, this remains an area of clinical equipoise, and biventricular repair is not an automatic guarantor of superior outcomes, let alone better exercise capacity. A case in point is a study of patients (n = 41) with repaired Tricuspid Atresia who underwent the historical Fontan-Bjork modification to incorporate the RV within the circuit. The intended outcome of pulsatile pulmonary blood flow was achieved only in a minority (n = 11, 26.8%). The majority (n = 30, 74.2%) still retained a passive pulmonary circulation without any tangible benefit of improved exercise capacity compared to those with pulsatile pulmonary circulation [29]. Similarly, another study by Rhodes et al. noted no significant improvement in exercise function in patients with Fontan circulation who had an RV incorporated within the pulmonary circulation [30]. There is also data to suggest that in patients with left ventricular outflow tract obstruction and sub-optimal ventricle, a vociferous pursuit of biventricular repair is detrimental [31].

Study Limitations
This study is limited by its single-center retrospective design and intermediate follow-up. Only survivors with PA/IVS who had undergone a uni-or biventricular repair were included. Our study was likely underpowered to detect differences in key parameters such as % VO 2 and % VO 2 < 85%. Since pre-surgical clinical/echocardiographic data was unavailable, key clinical correlations such as that between tricuspid Z valve and XC could not be performed. There were not enough patients with 1.5 ventricle repair, and therefore they were excluded from our study.

Conclusions
Young survivors with PA/IVS with prior univentricular (Fontan) palliation demonstrated an objective impairment in their chronotropic and pulmonary function test parameters compared to those with the biventricular repair. However, this did not translate into a significant difference in their exercise capacity. There was no longitudinal change in exercise capacity or chronotropic index within a smaller subset of each group over intermediate follow-up.
Author Contributions Anusha Konduri: Acquisition of data, analysis and interpretation of data, drafting the manuscriptSanjeev Aggarwal: Conception and design of study, analysis and interpretation of data, drafting the manuscriptChenni Sriram: Analysis and interpretation of data, drafting the manuscriptDeemah Mahadin: Analysis and interpretation of data, drafting the manuscriptAll authors reviewed the manuscript.