Long-term biventricular systolic performance is a key determinant of clinical outcomes late after tetralogy of Fallot (TOF) repair. A need exists for early indices of both left ventricular (LV) and right ventricular (RV) compromise in this population.
Twenty-nine children (age range, 5–18 years) with repaired TOF and 44 healthy controls were prospectively evaluated. M-mode and tissue Doppler data were obtained for each ventricle and the RV outflow tract at rest and during semisupine bicycle exercise. By making measurements of myocardial acceleration during isovolumic contraction during exercise, at increasing heart rates, LV force-frequency curves were constructed. Patients also underwent cardiac magnetic resonance imaging, cardiopulmonary exercise testing, and measurement of serum neurohormonal markers.
Children with repaired TOF had dilated right ventricles (RV end-diastolic volume index = 153 ± 37.3 mL/m 2 ) but normal ejection fractions as measured on magnetic resonance imaging (LV ejection fraction = 59.3 ± 6.2%, RV ejection fraction = 50.2 ± 8.5%) and normal serum neurohormonal markers. Detailed resting echocardiography detected abnormal ventricular function, worst in the right ventricle and RV outflow tract. Exercise exacerbated these findings and provoked significant decline in LV indices. The LV force-frequency curves of patients were attenuated, with an early plateau and inadequate increase of isovolumic contraction. Correlations were seen between peak exercise LV isovolumic contraction and percentage predicted peak oxygen uptake ( r = 0.51, P = .02), LV and RV ejection fractions ( r = 0.41, P = .03), and RV and LV long-axis fractional shortening ( r = 0.44, P = .02).
The postsurgical pathophysiology of TOF begins early after repair. At a time when clinically well and while routine indices of heart function remain normal, children with repaired TOF exhibit RV dilatation and subtle, interlinked biventricular abnormalities on resting echocardiography. Exercise echocardiography provides additional information and reveals abnormal LV excitation-contractile coupling that may be linked to impaired exercise capacity.
Although tetralogy of Fallot (TOF) is traditionally considered a right-sided disease, it has become increasingly apparent that the postoperative pathophysiology of TOF affects both ventricles and that late clinical outcomes relate to the durability of left ventricular (LV), as well as right ventricular (RV), systolic performance. Although half of all patients with repaired TOF will undergo reintervention targeted at preserving ventricular function, selection criteria and optimal timing of these procedures remain unclear. Current guidelines place weight on the appearance of symptoms and known risk factors, such as RV dilatation. Because population-specific early indices of ventricular compromise are yet to be identified, little emphasis has been given to the recognition and management of subclinical ventricular impairment.
Exercise stress echocardiography is emerging as an important tool for cardiac assessment and has been shown to be of prognostic value in patients with left bundle branch block, coronary artery disease, and atrial fibrillation. Our group previously found that exertional stress provokes biventricular mechanical dyssynchrony in children with repaired TOF and has also demonstrated that exercise echocardiography can afford insights into LV contractile reserve in children. The aims of the present study were to (1) characterize biventricular systolic function during the early years after TOF repair, using exercise echocardiography to identify early hallmarks of compromise, and (2) study the myocardial response of these children to increased physical demands.
Study Participants and Design
This prospective study was conducted with institutional ethics approval, and written consent was provided for all participants. Children with repaired TOF aged 5 to 18 years were identified from outpatient schedules. Records were reviewed against the inclusion and exclusion criteria and suitable patients invited to participate. Children with symptoms, pulmonary atresia, prior pulmonary valve replacement, contraindications to magnetic resonance imaging (MRI), or inability to pedal a bicycle were excluded. Controls of a similar age range were recruited from children with normal hearts undergoing echocardiographic screening and the siblings of patients with TOF. Children with TOF attended the hospital for two study visits in random order, within a 4-week period. One visit included 12-lead electrocardiography, echocardiography, and a standard bicycle metabolic exercise test. The other visit was for cardiac MRI and serum measurement of N-terminal prohormone brain natriuretic peptide, renin, aldosterone, norepinephrine, and epinephrine levels. Control subjects attended the hospital only once, for echocardiography. Results from this study cohort relating to dyssynchrony assessment have previously been published elsewhere.
A single observer (L.G-W.), blinded to all other results, protocolled and analyzed the results of MRI. Ventricular volumes, ejection fractions, and pulmonary regurgitation were quantified on a 1.5-T scanner (Avanto; Siemens Medical Solutions, Erlangen, Germany). For volumetry, a short-axis cine stack was acquired during breath-hold, using a steady-state free precession gradient-echo sequence with minimum echo and repetition times, a flip angle of 70°, bandwidth of 89 kHz, number of signals acquired of 1, views (lines) per segment to allow 20 true reconstructed phases per cardiac cycle, slice thickness of 5 to 7 mm, 10 to 12 slices, gap adjusted to cover both ventricles, and in-plane spatial resolution of 1.5 to 2.5 mm. Pulmonary flow phase-contrast velocity mapping was performed perpendicular to the right and left pulmonary arteries during free breathing with slice thickness of 5 mm and in-plane resolution of 1.5 to 2 mm. A dedicated workstation (Medis Medical Imaging Systems; Leiden, The Netherlands) was used for volumetric and flow analysis. For pulmonary blood flow and regurgitation volumes right and left pulmonary artery net forward flow volumes and reverse flow volumes, respectively, were measured individually then added, as is our clinical practice.
A single observer (S.L.R.), blinded to all other results, collected data from echocardiographic images. A Vivid 7 system (GE Vingmed Ultrasound AS, Horten, Norway) was used with simultaneous electrocardiographic recording. Frame rates for Doppler tissue imaging were maintained at >120 frames/sec. A dedicated workstation (EchoPAC version 6.0.1; GE Vingmed Ultrasound AS) was used for offline analysis.
Children lay on the bed of a Lode Angio Echocardiac Stress Table (Lode BV, Groningen, The Netherlands) adjusted to be semirecumbent and tilted to the left. A full baseline study was performed, including Doppler assessment of pressure gradients across the RV outflow tract (RVOT) and tricuspid valve. Resting image were recorded in the following views: apical four chamber, an apical view angled to include the RVOT, parasternal long axis (LAX), parasternal short axis, and subcostal.
Children pedaled against incremental resistance at a rate permitting a slow, steady increase in heart rate without excessive chest movement. At each 10 beats/min increase in heart rate, color-coded Doppler tissue images focused on the LV free wall were recorded in five- to 10-beat digital loops for the measurement of myocardial acceleration during isovolumic contraction (IVA). After reaching a pedaling velocity of 60—70 rpm, resistance was gradually added until heart rate plateaued or the child showed signs of exhaustion; then “peak exercise” apical four-chamber and anteriorly angled apical images were recorded.
Echocardiographic Indices of RV Systolic Function
Each RV functional index was measured at rest and peak exercise. Peak systolic myocardial tissue Doppler velocities (Sm) were measured at the lateral tricuspid annulus and averaged over three beats. Indices usually measured as absolute lengths (e.g., tricuspid annular plane systolic excursion) are affected by age and body surface area in childhood. To account for variable heart sizes, we expressed such measurements as percentage changes in length from diastole to systole. Tricuspid annular plane systolic excursion was normalized to cardiac dimensions and thereby expressed RV LAX fractional shortening (FS). Using a method similar to that described by Uebing et al ., RVOT LAX FS was calculated from M-mode tracings recorded from an anteriorly angled apical view, with the sample volume line positioned from the apex to the lateral aspect of the pulmonary valve annulus ( Figure 1 ). The following measurements were then made from the M-mode tracing: distance from apex to pulmonary valve at end-diastole ( D dias RVOT ) and distance from apex to pulmonary valve at peak systole ( D syst RVOT ). RVOT FS was calculated as follows: RVOT FS = ( D dias RVOT − D syst RVOT )/ D dias RVOT × 100.
Echocardiographic Indices of LV Systolic Function
LV FS was measured at rest in children without paradoxical septal motion (22 of 29 subjects). It was measured from an M-mode trace in a parasternal LAX view. End-diastole was measured at the onset of the QRS complex on electrocardiography. Peak myocardial mitral annular tissue Doppler velocity (Sm), mitral annular plane systolic excursion normalized to heart size and expressed as LAX FS, and LV IVA measured in the myocardium at the basal third of the LV free wall and averaged from five consecutive heartbeats were measured at rest and on peak exertion. Assessment of LV IVA at each 10 beats/min increase in heart rate permitted the development of force-frequency curves by plotting LV IVA against heart rate.
Statistical analysis was performed with a commercially available package, SAS version 9.1 (SAS Institute Inc, Cary, NC). Descriptive statistics of skewed continuous data are presented as medians and interquartile ranges. Approximately normal data are expressed as mean ± SD. Characteristics expressed as proportions were compared by using Fisher exact tests. Mean indices of ventricular function were compared between children with TOF and controls using unpaired, two-tailed Student t tests. Paired two-tailed Student t tests were used for intragroup comparisons between rest and exercise. Associations between functional indices and continuous variables were assessed by linear regression. Locally weighted scatterplot smoothing moving-average curves were drawn for visual comparison of force-frequency relationships between children with TOF and controls. Repeated-measures analysis was used to inspect the curves at different heart rates to look for statistical differences in mean IVA between the two groups.
Twenty-nine children with repaired TOF participated in the study. One was unable to tolerate MRI because of claustrophobia. The remainder completed the entire protocol. All patients underwent repair between 1991 and 2001. Other clinical characteristics are detailed in Table 1 . Forty-four healthy controls were studied at rest, and 27 of these were also studied at peak exercise. Children with TOF and controls were of similar ages (12.1 vs 12.4 years, P = .64), sex distribution (56% vs 55% male, P = .90), heights (146.8 vs 150.2 cm, P = .22), weights (43.2 vs 43.9 kg, P = .82), and body surface areas (1.3 vs 1.3 m 2 , P = .59). There were no differences in age or sex between the control subjects who did and did not exercise.
|Characteristic||Number (%) or median (IQR)|
|Patients with BT shunt before surgical repair||5.0 (17.2)|
|Age at complete surgical repair (mo)||9.6 (5.6–16.5)|
|Patients requiring ventriculotomy during repair||11.0 (37.9)|
|Patients with transannular patches||13.0 (44.8)|
|Duration of follow-up since repair (y)||11.2 (9.5–12.6)|
|Peak VO 2 (mL/min/m 2 )||29.5 (24.9–33)|
|Percentage of predicted peak VO 2 (%)||63.7 (54.3–73.3)|
|QRS duration (msec)||130.0 (120.0–146.7)|
|Echocardiographic RVOT gradient (mm Hg)||15.7 (8.4–20.6)|
|Indexed pulmonary regurgitant volume (L/min/m 2 )||2.1 (1.4–2.8)|
|Pulmonary regurgitant fraction (%)||38.0 (30–46)|
|RVEDVi (mL/m 2 )||145.2 (123.0–185.7)|
|MRI-measured LVEF||66.0 (56.2–62.3)|
|MRI-measured RVEF||51.3 (44.8–53.9)|
|Renin (U/L) (reference range, 0–17 U/L)||0.6 (0.5–0.9)|
|Aldosterone (pmol/L) (reference range, 28–444 pmol/L)||277.0 (190.5–411.5)|
|Norepinephrine (nmol/L) (reference range, 0.8–3.4 nmol/L)||1.3 (0.7–1.9)|
|NT-proBNP (pmol/L) (reference range, <30 pmol/L)||19.7 (10.6–24.7)|
Ventricular Function at Rest
All patients had a normal MRI-measured LV ejection fractions (LVEFs) and normal or low-normal RV ejection fractions (RVEFs) ( Table 1 ). Most of the echocardiographic indices of resting ventricular function were lower in patients than in controls ( Table 2 ). The differences were most apparent in RV indices. There were positive linear associations between RVEF and LVEF and between RV LAX FS and LV LAX FS ( Figure 2 ).
|Variable||Controls ( n = 44)||Patients with TOF ( n = 29)||P||Percentage of patients with TOF with measurements ≥2 SDs below controls’ mean|
|RVOT FS (%)||32.6 ± 5.5||18.9 ± 8.1||<.0001||69|
|RV LAX FS (%)||30.9 ± 6.4||16.8 ± 5.4||<.0001||59|
|RV IVA (m/sec 2 )||2.2 ± 0.7||0.7 ± 0.3||<.0001||62|
|RV Sm (cm/sec)||9.6 ± 1.4||6.1 ± 1.3||<.0001||66|
|LV FS (%)||35.7 ± 4.6||31.1 ± 7.7||.01||18|
|LV LAX FS (%)||20.2 ± 4.5||18.7 ± 5.3||NS||10|
|LV IVA (m/sec 2 )||1.3 ± 0.5||1.0 ± 0.4||.02||0|
|LV Sm (cm/sec)||6.8 ± 1.3||5.0 ± 1.4||<.0001||10|
Influence of Age at Surgical Repair
Although LVEFs were within the normal range for all children, an inverse relationship was seen between LVEF and age at the time of surgical repair (r = −0.50, P = .007), driven in part by a single distant outlier repaired at 70 months of age. To explore this finding further, patients were divided according to those aged ≤12 months at the time of repair ( n = 18) and those aged >12 months at the time of repair ( n = 11). LVEFs were lower in those aged >12 months (62% vs 56%, P = .03). No echocardiographic LV index was related to age at surgical repair.
Ventricular Function during Exercise
All echocardiographic indices of function increased with exercise in both children with TOF and controls ( Table 3 ). The magnitude of this increase was smaller in patients with TOF, such that between-group differences became more pronounced with exercise, particularly with regard to LV indices ( Tables 3 and 4 ).
|RV indices in controls|
|RVOT FS (%)||32.8 ± 5.8||37.7 ± 7.8||NS|
|RV LAX FS (%)||31.2 ± 6.8||34.6 ± 7.4||NS|
|RV IVA (m/sec 2 )||2.1 ± 0.7||10.3 ± 2.0||<.0001|
|RV Sm (cm/sec)||9.5 ± 1.2||13.1 ± 2.5||<.0001|
|LV indices in controls|
|LV LAX FS (%)||19.8 ± 4.7||22.3 ± 4.6||NS|
|LV IVA (m/sec 2 )||1.2 ± 0.4||9.9 ± 2.6||<.0001|
|LV Sm (cm/sec)||6.9 ± 1.4||9.9 ± 1.7||<.0001|
|RV indices in patients with TOF|
|RVOT FS (%)||18.9 ± 8.1||25.3 ± 7.1||.03|
|RV LAX FS (%)||16.8 ± 5.4||21.2 ± 6.1||.02|
|RV IVA (m/sec 2 )||0.7 ± 0.3||3.5 ± 1.4||<.0001|
|RV Sm (cm/sec)||6.1 ± 1.3||9.2 ± 2.5||<.0001|
|LV indices in patients with TOF|
|LV LAX FS (%)||18.7 ± 5.3||23.0 ± 6.3||.006|
|LV IVA (m/sec 2 )||1.0 ± 0.4||3.8 ± 1.6||<.0001|
|LV Sm (cm/sec)||5.0 ± 1.4||8.1 ± 2.4||<.0001|
|Variable||Controls ( n = 27)||Patients with TOF ( n = 29)||P||Percentage of patients with TOF with measurements ≥2 SDs below controls’ mean|
|Peak heart rate||164 ± 15.1||137.2 ± 15.6||.0005||41|
|RVOT FS (%)||37.7 ± 7.8||25.3 ± 7.1||<.0001||32|
|RV LAX FS (%)||34.6 ± 7.4||21.2 ± 6.1||<.0001||48|
|RV IVA (m/sec 2 )||10.3 ± 2.0||3.5 ± 1.4||<.0001||100|
|RV Sm (cm/sec)||13.1 ± 2.5||9.2 ± 2.5||<.0001||34|
|LV LAX FS (%)||22.3 ± 4.6||23.0 ± 6.3||NS||7|
|LV IVA (m/sec 2 )||9.9 ± 2.6||3.8 ± 1.6||<.0001||75|
|LV Sm (cm/sec)||9.9 ± 1.7||8.1 ± 2.4||.001||25|