Usefulness of Magnetic Resonance Imaging to Guide Referral for Pulmonary Valve Replacement in Repaired Tetralogy of Fallot Matthew J. Lewis, MD, MPHa,*, Daniel S. O’Connor, MDa, Anna Rozenshtien, MDb, Siqin Ye, MDa, Andrew J. Einstein, MD, PhDa, Jonathon M. Ginns, MDa, and Marlon S. Rosenbaum, MDa The aim of this study was to determine if adult patients with repaired tetralogy of Fallot are being referred for pulmonary valve replacement (PVR) earlier on the basis of cardiac magnetic resonance imaging (CMR) parameters despite the absence of CMR-based recommendations in the American College of Cardiology and American Heart Association joint guidelines. Variables defined by the guidelines were analyzed in conjunction with CMRbased parameters across 3 groups defined by the release of the guidelines: (1) patients referred before the guidelines, (2) patients referred 0 to 3 years after the guidelines, and (3) patients referred ‡3 years after the guidelines. Seventy-nine patients were identified. No significant trend was observed in guideline-defined variables. Significant trends in indexed right ventricular end-diastolic volume (p [ 0.034), indexed right ventricular end-systolic volume (p [ 0.001), and the right ventricular ejection fraction (p [ 0.005) were observed across groups. By multivariate regression, patients who underwent PVR ‡3 years after the release of the guidelines had a 29 ml/m2 smaller indexed right ventricular end-diastolic volume (p [ 0.01) and a 33 ml/m2 smaller indexed right ventricular end-systolic volume (p 50 mm HG Symptoms Peak VO2, median, (cc*kg/min) Time from PVR to CMR, median, (months)
All Patients (n ¼ 79)
CMR (n ¼ 71)
No CMR (n ¼ 8)
44 (56%) 36 1.2 11 (14%) 6 (8%) 2 (3%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) 1 (1%)
39 (55%) 35 1.2 10 (14%) 6 (8%) 2 (3%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) 1 (1%)
5 (63%) 41 5.1 1 (13%) 0 0 0 0 0 0 0
26 11 3 3 2 2 1 1
26 10 2 3 2 2 1 1
(37%) (14%) (3%) (4%) (3%) (3%) (1%) (1%)
0 1 (13%) 1 (13%) 0 0 0 0 0
21 (30%) 9 (13%) 0 (0) 31 (44%) 26.7 (8) 8.6 (15)
2 (25%) 2 (25%) 1 (13%) 5 (63%) 23.7 (11) —
(33%) (14%) (4%) (4%) (3%) (3%) (1%) (1%)
23 (29%) 11 (14%) 1 (1%) 36 (46%) 26.7 (9) 8.6 (15)
Values are mean SD, n (%), or median (interquartile range). ASD ¼ atrial septal defect; CMR ¼ cardiac MRI; NYHA ¼ New York Heart Association; PFO ¼ patent foramen ovale; PVR ¼ pulmonary valve replacement; RV-PA ¼ Right ventricle-pulmonary artery; VSD ¼ ventricular septal defect.
timing of each patient’s initial surgical repair and any subsequent procedure, including PVR, were recorded. Patients were considered symptomatic if they reported dyspnea, decreased exercise tolerance, or exertional chest pain before PVR. The presence of clinically significant arrhythmias was defined as sustained, nonsinus supraventricular tachycardia or sustained ventricular tachycardia documented on 12-lead electrocardiography, Holter monitoring, or electrophysiology. Right ventricular outflow tract obstruction was defined as a gradient 50 mm Hg by preoperative cardiac catheterization. We reviewed reports from preoperative 2-dimensional color Doppler transthoracic echocardiograms for each patient. All studies were performed and interpreted at the Schneeweiss Adult Congenital Heart Center by cardiologists with years of expertise in congenital cardiac echocardiography. The severity of PR and tricuspid regurgitation was graded visually. Patients were classified as having mild or less tricuspid regurgitation or moderate or greater tricuspid regurgitation by transthoracic echocardiography. Right ventricular function was assessed and defined as normal or abnormal (including mild, moderate, and severe dysfunction) on the basis of echocardiographic appearance. CMR studies were performed with breath holding and electrocardiographic gating using a Signa 1.5-T magnetic
1407
Table 2 Demographic and clinical characteristics of participants, by timing of pulmonary valve replacement
Age at PVR, mean, (years) Echocardiographic data Moderate TR Elevated RVOT gradient Decreased RV function Symptoms Dyspnea Chest Pain Any Arrhythmia Any Symptom NYHA Class 2 Peak VO2†, median (cc*kg/min)
Group 1*
Group 2*
Group 3*
p-value for trend
37 1.7
35 2.2
36 2.6
0.73
6 (17%) 4 (11%) 22 (61%)
2 (8%) 1 (4%) 6 (25%)
3 (17%) 1 (5%) 7 (39%)
0.29 0.36 0.17
10 (28%) 2 (6%) 9 (25%) 12 (33%) 14 (39%) 24.4 (10)
12 (50%) 2 (8%) 8 (33%) 12 (50%) 8 (33%) 26.3 (8)
5 (28%) 3 (17%) 2 (11%) 12 (67%) 1 (5%) 28.5 (6)
0.18 0.29 0.20 0.16 0.17 0.09
Values are mean SD, n (%), or median (interquartile range). NYHA ¼ New York Heart Association; PVR ¼ pulmonary valve replacement; RV ¼ right ventricular; RVOT ¼ right ventricular outflow tract; TR ¼ tricuspid valve regurgitation. * Group 1 represents patients referred before the guidelines, Group 2 represents patients referred 0e3 years after the release of the guidelines, and Group 3 represents patients released 3 years after the release of the guidelines. † Peak VO2 was assessed in 50 patients.
resonance imaging scanner (GE Healthcare, Milwaukee, Wisconsin) and an 8-channel phased array. Before June 2003, short axis cine gradient echocardiographic images were obtained with the following parameters: repetition time 8.8 ms, echo time 15.2 ms, flip angle 15 , 8 views per segment, field of view 30 cm, acquisition matrix 256 128, slice thickness 8 mm with no gap, and receiver bandwidth 31.25 kHz. From June 2003 onward, short-axis cine images were acquired using a steady-state free precession pulse sequence with the following parameters: repetition time 3.6 ms, echo time 11.5 ms, flip angle 45 , 24 views per segment, field of view 35 cm, acquisition matrix 192 160, slice thickness 8 mm with no gap, receiver bandwidth 125 kHz. Images were reviewed and analyzed using ReportCARD software (GE Healthcare). A single reader with training and years of expertise in CMR imaging who was blinded to clinical status and the results of echocardiography performed CMR image analysis. Cine loops were used to select images at enddiastole and end-systole. End-diastole and end-systole were defined independently for the right and left ventricles as the phases with the largest and smallest volumes, respectively. Endocardial segmentation was performed by manual tracing of each end-diastolic and end-systolic short-axis view; areas were multiplied by slice thickness and summed to calculate right and left ventricular volumes. By convention, trabeculations and papillary muscles were considered part of the ventricular cavity in systole and diastole. Ejection fractions were calculated using the end-diastolic and end-systolic values. Testing for a significant trend across groups was conducted using Goodman and Kruskal’s gamma statistic for categorical variables. For continuous variables, we used an extension of Wilcoxon’s rank-sum test developed by Cuzick20 to test for trend across groups. To assess between-group differences in
1408
The American Journal of Cardiology (www.ajconline.org)
Figure 1. Box plot of RVEDVi by time from guidelines. Median is indicated by the line in the middle of each box, first and third quartiles by the ends of the box, and minimum and maximum values by the ends of the whiskers. There was a significant trend (p ¼ 0.034) in RVEDVi across groups. The median RVEDVi for group 1 was 191 ml/m2, compared with 189 ml/m2 for group 2 and 154 ml/m2 for group 3.
Figure 3. Box plot of RVEF by time from guidelines. Median is indicated by the line in the middle of each box, first and third quartiles by the ends of the box, minimum and maximum nonoutlier values by the ends of the whiskers, and outlier points are plotted individually. There was a significant trend (p ¼ 0.005) in RVEF across groups. The median RVEF for group 1 is 37%, compared with 45% for group 2 and 43% for group 3.
performed using Stata version 13.1 (StataCorp LP, College Station, Texas). Results
Figure 2. Box plot of RVESVi by time from guidelines. Median is indicated by the line in the middle of each box, first and third quartiles by the ends of the box, and minimum and maximum values by the ends of the whiskers. There was a significant trend (p ¼ 0.001) in RVESVi across groups. The median RVESVi for Group 1 was 112 ml/m2, compared with 102 ml/m2 for group 2 and 86 ml/m2 for group 3.
CMR parameters of interest, including indexed right ventricular end-diastolic volume (RVEDVi), indexed right ventricular end-systolic volume (RVESVi), and the right ventricular ejection fraction (RVEF), we first constructed univariate linear regression models with individual predictors. On the basis of the univariate analysis, we constructed multivariate linear regression models. Age at PVR and moderate or greater tricuspid regurgitation were prespecified for inclusion in the model, and any additional covariates reaching p 50 Any clinical symptom RVESVI/(ml/m2) Time of PVR Pre-Guidelines 0e3 years After 3 years After Moderate TR Age at PVR/year Male RV-PA gradient>50 Any clinical symptom RVEF/(%) Time of PVR Pre-Guidelines 0e3 years After 3 years After Moderate TR Age at PVR/year Male RV-PA gradient>50 Any clinical symptom
Multivariate Predictors p
0.8 28 30 0.7 4 22 9
Ref (21, 19) (51, 6.3) (4, 56) (2, 0.2) (21, 15) (62, 18) (27, 9)
Ref 0.93 0.01 0.03 0.14 0.70 0.27 0.32
12 33 29 0.2 8 8 7
Ref (28, 4) (50, 16) (8, 50) (1, 0.5) (23, 7) (41, 25) (22, 8)
Ref 0.13
All Patients (n ¼ 79)
CMR (n ¼ 71)
No CMR (n ¼ 8)
44 (56%) 36 1.2 11 (14%) 6 (8%) 2 (3%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) 1 (1%)
39 (55%) 35 1.2 10 (14%) 6 (8%) 2 (3%) 1 (1%) 1 (1%) 1 (1%) 1 (1%) 1 (1%)
5 (63%) 41 5.1 1 (13%) 0 0 0 0 0 0 0
26 11 3 3 2 2 1 1
26 10 2 3 2 2 1 1
(37%) (14%) (3%) (4%) (3%) (3%) (1%) (1%)
0 1 (13%) 1 (13%) 0 0 0 0 0
21 (30%) 9 (13%) 0 (0) 31 (44%) 26.7 (8) 8.6 (15)
2 (25%) 2 (25%) 1 (13%) 5 (63%) 23.7 (11) —
(33%) (14%) (4%) (4%) (3%) (3%) (1%) (1%)
23 (29%) 11 (14%) 1 (1%) 36 (46%) 26.7 (9) 8.6 (15)
Values are mean SD, n (%), or median (interquartile range). ASD ¼ atrial septal defect; CMR ¼ cardiac MRI; NYHA ¼ New York Heart Association; PFO ¼ patent foramen ovale; PVR ¼ pulmonary valve replacement; RV-PA ¼ Right ventricle-pulmonary artery; VSD ¼ ventricular septal defect.
timing of each patient’s initial surgical repair and any subsequent procedure, including PVR, were recorded. Patients were considered symptomatic if they reported dyspnea, decreased exercise tolerance, or exertional chest pain before PVR. The presence of clinically significant arrhythmias was defined as sustained, nonsinus supraventricular tachycardia or sustained ventricular tachycardia documented on 12-lead electrocardiography, Holter monitoring, or electrophysiology. Right ventricular outflow tract obstruction was defined as a gradient 50 mm Hg by preoperative cardiac catheterization. We reviewed reports from preoperative 2-dimensional color Doppler transthoracic echocardiograms for each patient. All studies were performed and interpreted at the Schneeweiss Adult Congenital Heart Center by cardiologists with years of expertise in congenital cardiac echocardiography. The severity of PR and tricuspid regurgitation was graded visually. Patients were classified as having mild or less tricuspid regurgitation or moderate or greater tricuspid regurgitation by transthoracic echocardiography. Right ventricular function was assessed and defined as normal or abnormal (including mild, moderate, and severe dysfunction) on the basis of echocardiographic appearance. CMR studies were performed with breath holding and electrocardiographic gating using a Signa 1.5-T magnetic
1407
Table 2 Demographic and clinical characteristics of participants, by timing of pulmonary valve replacement
Age at PVR, mean, (years) Echocardiographic data Moderate TR Elevated RVOT gradient Decreased RV function Symptoms Dyspnea Chest Pain Any Arrhythmia Any Symptom NYHA Class 2 Peak VO2†, median (cc*kg/min)
Group 1*
Group 2*
Group 3*
p-value for trend
37 1.7
35 2.2
36 2.6
0.73
6 (17%) 4 (11%) 22 (61%)
2 (8%) 1 (4%) 6 (25%)
3 (17%) 1 (5%) 7 (39%)
0.29 0.36 0.17
10 (28%) 2 (6%) 9 (25%) 12 (33%) 14 (39%) 24.4 (10)
12 (50%) 2 (8%) 8 (33%) 12 (50%) 8 (33%) 26.3 (8)
5 (28%) 3 (17%) 2 (11%) 12 (67%) 1 (5%) 28.5 (6)
0.18 0.29 0.20 0.16 0.17 0.09
Values are mean SD, n (%), or median (interquartile range). NYHA ¼ New York Heart Association; PVR ¼ pulmonary valve replacement; RV ¼ right ventricular; RVOT ¼ right ventricular outflow tract; TR ¼ tricuspid valve regurgitation. * Group 1 represents patients referred before the guidelines, Group 2 represents patients referred 0e3 years after the release of the guidelines, and Group 3 represents patients released 3 years after the release of the guidelines. † Peak VO2 was assessed in 50 patients.
resonance imaging scanner (GE Healthcare, Milwaukee, Wisconsin) and an 8-channel phased array. Before June 2003, short axis cine gradient echocardiographic images were obtained with the following parameters: repetition time 8.8 ms, echo time 15.2 ms, flip angle 15 , 8 views per segment, field of view 30 cm, acquisition matrix 256 128, slice thickness 8 mm with no gap, and receiver bandwidth 31.25 kHz. From June 2003 onward, short-axis cine images were acquired using a steady-state free precession pulse sequence with the following parameters: repetition time 3.6 ms, echo time 11.5 ms, flip angle 45 , 24 views per segment, field of view 35 cm, acquisition matrix 192 160, slice thickness 8 mm with no gap, receiver bandwidth 125 kHz. Images were reviewed and analyzed using ReportCARD software (GE Healthcare). A single reader with training and years of expertise in CMR imaging who was blinded to clinical status and the results of echocardiography performed CMR image analysis. Cine loops were used to select images at enddiastole and end-systole. End-diastole and end-systole were defined independently for the right and left ventricles as the phases with the largest and smallest volumes, respectively. Endocardial segmentation was performed by manual tracing of each end-diastolic and end-systolic short-axis view; areas were multiplied by slice thickness and summed to calculate right and left ventricular volumes. By convention, trabeculations and papillary muscles were considered part of the ventricular cavity in systole and diastole. Ejection fractions were calculated using the end-diastolic and end-systolic values. Testing for a significant trend across groups was conducted using Goodman and Kruskal’s gamma statistic for categorical variables. For continuous variables, we used an extension of Wilcoxon’s rank-sum test developed by Cuzick20 to test for trend across groups. To assess between-group differences in
1408
The American Journal of Cardiology (www.ajconline.org)
Figure 1. Box plot of RVEDVi by time from guidelines. Median is indicated by the line in the middle of each box, first and third quartiles by the ends of the box, and minimum and maximum values by the ends of the whiskers. There was a significant trend (p ¼ 0.034) in RVEDVi across groups. The median RVEDVi for group 1 was 191 ml/m2, compared with 189 ml/m2 for group 2 and 154 ml/m2 for group 3.
Figure 3. Box plot of RVEF by time from guidelines. Median is indicated by the line in the middle of each box, first and third quartiles by the ends of the box, minimum and maximum nonoutlier values by the ends of the whiskers, and outlier points are plotted individually. There was a significant trend (p ¼ 0.005) in RVEF across groups. The median RVEF for group 1 is 37%, compared with 45% for group 2 and 43% for group 3.
performed using Stata version 13.1 (StataCorp LP, College Station, Texas). Results
Figure 2. Box plot of RVESVi by time from guidelines. Median is indicated by the line in the middle of each box, first and third quartiles by the ends of the box, and minimum and maximum values by the ends of the whiskers. There was a significant trend (p ¼ 0.001) in RVESVi across groups. The median RVESVi for Group 1 was 112 ml/m2, compared with 102 ml/m2 for group 2 and 86 ml/m2 for group 3.
CMR parameters of interest, including indexed right ventricular end-diastolic volume (RVEDVi), indexed right ventricular end-systolic volume (RVESVi), and the right ventricular ejection fraction (RVEF), we first constructed univariate linear regression models with individual predictors. On the basis of the univariate analysis, we constructed multivariate linear regression models. Age at PVR and moderate or greater tricuspid regurgitation were prespecified for inclusion in the model, and any additional covariates reaching p 50 Any clinical symptom RVESVI/(ml/m2) Time of PVR Pre-Guidelines 0e3 years After 3 years After Moderate TR Age at PVR/year Male RV-PA gradient>50 Any clinical symptom RVEF/(%) Time of PVR Pre-Guidelines 0e3 years After 3 years After Moderate TR Age at PVR/year Male RV-PA gradient>50 Any clinical symptom
Multivariate Predictors p
0.8 28 30 0.7 4 22 9
Ref (21, 19) (51, 6.3) (4, 56) (2, 0.2) (21, 15) (62, 18) (27, 9)
Ref 0.93 0.01 0.03 0.14 0.70 0.27 0.32
12 33 29 0.2 8 8 7
Ref (28, 4) (50, 16) (8, 50) (1, 0.5) (23, 7) (41, 25) (22, 8)
Ref 0.13
