Congenital Heart Disease

Transcription

1 Congenital Heart Disease Clinical and Hemodynamic Outcomes up to 7 Years After Transcatheter Pulmonary Valve Replacement in the US Melody Valve Investigational Device Exemption Trial John P. Cheatham, MD; William E. Hellenbrand, MD; Evan M. Zahn, MD; Thomas K. Jones, MD; Darren P. Berman, MD; Julie A. Vincent, MD; Doff B. McElhinney, MD Background Studies of transcatheter pulmonary valve (TPV) replacement with the Melody valve have demonstrated good short-term outcomes, but there are no published long-term follow-up data. Methods and Results The US Investigational Device Exemption trial prospectively enrolled 171 pediatric and adult patients (median age, 19 years) with right ventricular outflow tract conduit obstruction or regurgitation. The 148 patients who received and were discharged with a TPV were followed up annually according to a standardized protocol. During a median follow-up of 4.5 years (range, years), 32 patients underwent right ventricular outflow tract reintervention for obstruction (n=27, with stent fracture in 22), endocarditis (n=3, 2 with stenosis and 1 with pulmonary regurgitation), or right ventricular dysfunction (n=2). Eleven patients had the TPV explanted as an initial or second reintervention. Five-year freedom from reintervention and explantation was 76±4% and 92±3%, respectively. A conduit prestent and lower discharge right ventricular outflow tract gradient were associated with longer freedom from reintervention. In the 113 patients who were alive and reintervention free, the follow-up gradient (median, 4.5 years after implantation) was unchanged from early post-tpv replacement, and all but 1 patient had mild or less pulmonary regurgitation. Almost all patients were in New York Heart Association class I or II. More severely impaired baseline spirometry was associated with a lower likelihood of improvement in exercise function after TPV replacement. Conclusions TPV replacement with the Melody valve provided good hemodynamic and clinical outcomes up to 7 years after implantation. Primary valve failure was rare. The main cause of TPV dysfunction was stenosis related to stent fracture, which was uncommon once prestenting became more widely adopted. Clinical Trial Registration URL: Unique identifier: NCT (Circulation. 2015;131: DOI: /CIRCULATIONAHA ) Key Words: heart diseases heart valves pediatrics pulmonary heart disease tetralogy of Fallot Transcatheter pulmonary valve (TPV) replacement (TPVR) with the Melody valve was first performed by Bonhoeffer et al 1 in 2000 and was introduced in the United States in 2007 through a prospective, multicenter Investigational Device Exemption (IDE) trial. 2 4 A number of single-center and multicenter reports have confirmed that TPVR is technically reproducible and results in excellent early and short-term outcomes, with reduction of right ventricular outflow tract (RVOT) obstruction, elimination of pulmonary regurgitation (PR), and clinical improvement. 2 9 At a functional level, TPVR is followed by improved circulatory efficiency and ventricular strain, enhanced cardiopulmonary exercise function in some patients, and better quality of life Although these and other reports helped substantiate TPVR as a valuable tool in the management of patients with postoperative RVOT dysfunction, the lack of longer-term follow-up data remains an important limitation in the Melody valve literature. To date, the studies with the longest follow-up were a 2008 report of 155 patients from the Bonhoeffer group and a 63-patient series collected through an Italian registry that was published in 2013, which reported outcomes at a median of 2.3 and 2.5 years after TPVR, respectively. 5,6 To address this deficiency, we analyzed midterm hemodynamic and clinical outcomes in the IDE trial patients, who were all at least 4 years out from Melody valve implantation. Editorial see p 1943 Clinical Perspective on p 1970 Received October 2, 2014; accepted March 23, From Division of Cardiology, Nationwide Children s Hospital, Ohio State University School of Medicine, Columbus (J.P.C., D.P.B.); Division of Pediatric Cardiology, Yale University, New Haven, CT (W.E.H.); Division of Cardiology, Cedars-Sinai Medical Center, Los Angeles, CA (E.M.Z.); Division of Cardiology, Seattle Children s Hospital, University of Washington School of Medicine (T.K.J.); Division of Cardiology, Miami Children s Hospital, FL (D.P.B.); Division of Pediatric Cardiology, Columbia University Medical Center, New York, NY (J.A.V.); and Department of Cardiothoracic Surgery, Stanford University School of Medicine, Palo Alto, CA (D.B.M.). Guest Editor for this article was William T. Mahle, MD. The online-only Data Supplement is available with this article at /-/DC1. Correspondence to John P. Cheatham, MD, Nationwide Children s Hospital, 700 Children s Dr, Columbus, OH John.Cheatham@ NationwideChildrens.org 2015 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 Cheatham et al Transcatheter PVR Outcomes to 7 Years 1961 Methods Patients and Study Protocol The US Melody valve IDE trial was a nonrandomized, prospective study sponsored by Medtronic, Inc (Minneapolis, MN) that enrolled patients at 5 sites. The study was conducted under IDE G and was registered in ClinicalTrials.gov ( gov; identifier, NCT ). The protocol was approved by the US Food and Drug Administration and by the Institutional Review Board at each institution. The trial was initially designed to follow patients for 5 years after implantation or until explantation, but was modified in 2011 to allow follow-up out to 10 years in patients who provided supplemental written informed consent. The trial design, protocols, and procedural and short-term results were described previously. 2 4,14,15 Follow-Up Evaluation Follow-up consisting of echocardiography, cardiopulmonary exercise testing, chest radiography, and clinical history and examination was conducted annually, within predetermined windows, at the implanting center or elsewhere after 5 years if more convenient for the patient. Details of follow-up evaluation protocols were summarized previously. 2 4,14,15 Radiographic studies were performed to assess for Melody valve stent fractures (MSFs), which were classified according to a previously reported system. 19 Stent fracture assessment did not include examination of conduit stents (ie, prestents) placed at a prior procedure (existing) or during the TPVR procedure (new). The protocol for coronary artery compression testing was discussed in detail in a prior report. 20 Endocarditis was defined and reported as described in a prior report. 21 All data, including findings on nonprotocol evaluations, reinterventions, and adverse events, were recorded on prespecified case report forms and entered into the Web-based data collection system maintained by the sponsor. The database was locked for this analysis on May 16, Statistical Analysis Procedural results were described previously, but, given the reported relationships between residual obstruction and adverse outcomes, 4,18,21,22 a new analysis was conducted of factors associated with a higher early postimplantation gradient, which included all patients in whom TPVR was attempted. Follow-up analyses included only patients discharged with a Melody valve in place. The primary outcome was TPVR reintervention over time; secondary outcomes included Melody valve obstruction or regurgitation, functional status, exercise function, MSF (any MSF or a type II MSF, defined as compromise of stent integrity), 4,19 endocarditis, and any other TPV-related adverse events. RVOT gradients were presented as the mean Doppler gradient across the conduit or valve. TPV reinterventions were categorized as dilation (angioplasty without placement of an additional stent or a new valve), TPV-in-TPV implantation (placement of a second TPV within the original Melody valve), 22 or explantation (surgical removal of the TPV with placement of a new conduit or valve). According to the protocol, patients were removed from the trial at the time of death, TPV explantation, or designation as lost to follow-up. Freedom-from-event estimates and curves were generated with the Kaplan-Meier method and were presented as rate±standard error, which was determined by the Greenwood method. Factors associated with shorter freedom from events were assessed by Cox regression, with variables significant to P 0.05 on univariable analysis considered for inclusion in a forward stepwise multivariable model. Additional details concerning these analyses are presented in the online-only Data Supplement. To illustrate a hypothetical scenario in which technology and practices allow the prevention of a major MSF, freedom from TPV reintervention was also estimated for the cohort of patients who did not develop a type II MSF. Hazard ratios (HRs) were presented with 95% confidence intervals (CIs). The Wilcoxon rank-sum test was used to compare continuous data between groups, and the signed-rank test was used to evaluate the change in continuous paired data. For selected analyses, continuous variables (eg, RVOT gradients and the ratio of angiographic to nominal conduit diameter) were dichotomized, with threshold values within a reasonable clinical range determined from receiver-operating characteristic curves. Categorical variables were compared between groups with the use of χ 2 analysis or the Fisher exact test. Exploratory multiple linear regression analysis was performed to evaluate factors associated with a higher post-tpvr RVOT gradient. Data were presented as median (minimum maximum) or frequency (%). The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written. Results Patients and Procedural Outcomes The US IDE trial enrolled 171 patients with postoperative RVOT conduit dysfunction from January 2007 through January Of these, 167 underwent catheterization, and 150 had a Melody valve implanted (Figure 1), which was deployed in the intended location in all cases. The cohort was almost evenly divided between pediatric patients (age 18 years) and adults, and there were several notable diagnostic differences between these populations (Table 1). As described previously, 2,3 1 patient had emergent pulmonary valve replacement for conduit rupture after TPVR, and 1 died of cerebral hemorrhage after a course complicated by coronary artery dissection before Melody valve implantation and subsequent mechanical circulatory support. The remaining 148 patients were discharged with the valve in place and constitute the study cohort. Acute Hemodynamic Results TPVR provided a competent pulmonary valve in all patients. On the discharge echocardiogram, PR was absent or trivial in 140 patients and mild in 5 (valve not visualized in 3), which was a significant improvement from before implantation, when 79% of patients had moderate (n=45) or severe (n=69) PR and 9% (n=14) had trivial or no PR (P<0.001). TPVR was also followed by a significant reduction in the mean Doppler RVOT gradient, from a median of 33 mm Hg (range, 5 97 mm Hg) to 17 mm Hg (range, 3 51 mm Hg) on the discharge echocardiogram (P<0.001). However, a subset of patients had discharge gradients >20 mm Hg (n=49, 33%) or >25 mm Hg (n=17, 12%). Diagnostic and procedural factors associated with a higher discharge gradient are summarized in Table 2. On multivariable linear regression, a preimplantation mean RVOT gradient >35 mm Hg and the absence of a new prestent were associated with higher discharge gradient but provided only modest explanatory power (model R 2 =0.08, P<0.001). In general, pediatric patients had higher gradients both before and early after Melody valve implantation than adults. Follow-Up Survival During a median follow-up of 4.5 years (range, years) with a total of 651 patient-years of observation, 4 patients died, 1 of multisystem failure resulting from sepsis/endocarditis and 3 of other causes with no apparent relationship to the Melody valve (1 of respiratory failure and 2 of unknown causes). Estimated survival was 98±1% at 5 years (Figure 2). After exclusion of 6 patients classified as lost to follow-up

3 1962 Circulation June 2, 2015 Figure 1. Flow diagram depicting patients enrolled, catheterized, and discharged with the Melody valve in place and by subsequent outcomes. *Four patients were not catheterized after enrollment for the following reasons: 1 withdrew consent, 2 did not meet echocardiographic inclusion criteria (patients were enrolled before the screening echocardiogram), and 1 had better-than-anticipated function on magnetic resonance imaging and the primary cardiologist elected not to perform the planned catheterization. Of the 17 patients catheterized who did not undergo implantation, 6 had documented coronary artery compromise on compression testing, 20 6 had a conduit that was too large for Melody valve implantation according to protocol criteria, 3 had conduit function that was better than expected or did not meet the protocol criteria for implantation, 1 had an adequate hemodynamic result with angioplasty alone, and 1 had pulmonary artery rather than conduit obstruction and needed surgery on the left ventricular outflow tract. Does not depict 2 patients who underwent transcatheter pulmonary valve (TPV) dilation or a second TPV-in-TPV after an initial TPV-in-TPV implantation. or withdrawn, the median follow-up was 4.9 years. When the database was locked, all patients were at least 4 years out from TPVR. However, complete follow-up was not available in some patients, most of whom were not local to the implanting center and had missed the prior protocol evaluation but were known to be alive and receiving care from another physician. If such patients had indicated an intention to continue in the protocol, they were not classified as lost to follow-up and were analyzed only through the most recent protocol testing. RVOT Reintervention In addition to the patient who experienced an acute conduit rupture, 32 patients underwent reintervention on the Melody valve: the first reintervention consisted of TPV redilation in 6 patients, TPV-in-TPV implantation in 19, and TPV explantation in 7 (Figure 1 and Table 3). Three other patients had a bare stent placed in the RVOT conduit proximal or distal to the TPV for stenoses that were recurrent or not fully treated in the TPV implantation procedure; in all 3 cases, the TPV was well functioning and not affected by the reintervention. Details of and indications for reinterventions are summarized in Table 3. Freedom from reintervention on the Melody valve was 76±4% at 5 years and on multivariable Cox regression was shorter in patients with a homograft conduit, a conduit that was not prestented, a discharge RVOT mean Doppler gradient >25 mm Hg, and preimplantation moderate to severe TR (Table 4). There were several notable interactions between conduit stenting and pre- and post-tpvr RVOT gradients. The longest freedom from Melody valve reintervention was in patients who received a new prestent or who did not receive a new prestent but had a preimplantation mean Doppler gradient <35 mm Hg. Compared with that group, patients with no new prestent, a preimplantation gradient 35 mm Hg, and a discharge gradient >20 mm Hg had the shortest freedom from reintervention (HR, 3.8; 95% CI, ; P=0.002). Patients with no new prestent, a preimplantation gradient 35 mm Hg, and a discharge gradient 20 mm Hg were intermediate (HR, 2.6; 95% CI, ; P=0.035 versus new prestent group). Freedom from RVOT reintervention in the cohort of patients who did not develop a type II MSF was 98±1% at 3 years and 91±3% at 5 years and was shorter in those with a higher discharge RVOT gradient (HR, 2.6 per 10 mm Hg; 95% CI, ; P=0.03). Transcatheter Melody Valve Reintervention Twenty-five patients underwent transcatheter TPV reintervention, only 3 of whom had a new prestent placed before the original Melody valve implantation. In 6 of these patients, the initial reintervention was redilation of the Melody valve. Five of these patients had the valve implanted on the smallest (18 mm) delivery system; 4 were among the first 50 implants overall; 4 were among and the first 10 implants at a given center; and all had relatively high residual gradients early after TPVR, with a mean Doppler gradient on the discharge echocardiogram >20 mm Hg 5 of the 6. An MSF was diagnosed before the redilation in only 2 of these 6 patients. The initial reintervention was a TPV-in-TPV implantation in 19 patients. Including 1 other patient whose first reintervention was TPV dilation, 20 patients ultimately underwent implantation of a second Melody valve for TPV dysfunction, which consisted of RVOT obstruction associated with MSF in all cases. All patients who underwent TPV-in-TPV implantation had 1 (n=1) or more (2 in 14, 3 in 5) prestents placed before implantation of the second Melody valve. After the second TPVR, the mean Doppler gradient decreased in all patients, from 40 mm Hg (range, mm Hg) to 17 mm Hg (8 37 mm Hg; P<0.001). Seven of the 25 patients whose initial RVOT reintervention was redilation (3 of 6) or TPV-in-TPV implantation (4 of 19) underwent another intervention during the study period, as detailed in Figure 1 and summarized in Table 3. Freedom from a second reintervention after transcatheter reintervention (dilation or a second TPVR) was 70±11% at 3 years and was significantly shorter if the first reintervention was

4 Cheatham et al Transcatheter PVR Outcomes to 7 Years 1963 Table 1. Details of the Cohort of Implanted Patients Overall and by Age Variable Total (n=150) Pediatric Patients 18 y of Age (n=71) Adults >18 y of Age (n=79) Demographic, diagnostic, and procedural data Age, y 19 (7 53) 14 (7 18) 26 (19 53) Diagnosis, n (%) Tetralogy of Fallot 77 (51) 33 (47) 44 (56) Left heart disease, prior Ross procedure 31 (21) 24 (34) 7 (9) Transposition of the great arteries 16 (11) 3 (4) 13 (17) Truncus arteriosus 15 (10) 8 (11) 7 (9) Double-outlet right ventricle 7 (5) 2 (3) 5 (6) Isolated valvar pulmonary stenosis 3 (2) 1 (1) 2 (3) Other 1 (0.1) 0 (0.1) 1 (1) Conduit type, n (%) 0.17 Homograft 111 (74) 56 (79) 55 (70) Bioprosthetic valve/conduit 32 (21) 14 (20) 18 (23) Synthetic 7 (5) 1 (1) 6 (7) Conduit size Nominal surgical implant diameter, mm 21 (16 28) 20 (16 26) 22 (16 28) Narrowest angiographic diameter at TPVR, mm 13 (5 21) 12 (5 21) 13.4 (5.8 20) 0.02 Angiographic/nominal diameter ratio 0.62 ( ) 0.59 ( ) 0.64 ( ) 0.12 Age of conduit, y 9.5 (0.2 38) 7.2 ( ) 13.4 (0.4 38) <0.001 Primary implant indication, n (%) PR 80 (53) 29 (41) 51 (65) RVOT obstruction 39 (26) 23 (33) 16 (20) Mixed PR and obstruction 31 (21) 19 (27) 12 (15) Delivery system balloon size, n (%) mm 23 (15) 14 (20) 9 (11) 20 mm 41 (27) 23 (32) 18 (23) 22 mm 86 (57) 34 (48) 52 (66) Exercise and pulmonary function Peak oxygen consumption, ml kg 1 min 1 24 (8 59) 26 (9 59) 20 (8 39) <0.001 Predicted peak oxygen consumption, % 58 (21 122) 58 (21 122) 57 (26 101) 0.18 Oxygen consumption at the AT, ml kg 1 min ( ) 16.6 ( ) 13.0 ( ) <0.001 Maximum workload, W 104 (3 338) 98.0 ( ) (15 250) 0.48 V e/v co 2 ratio 30 (21 48) 31 (25 48) 30 (21 46) 0.58 Predicted FEV 1, % 75 (23 144) 83 (25 128) 69 (23 109) <0.001 Predicted FVC, % 77 (21 139) 85 (21 139) 72 (26 111) Mean Doppler RVOT gradient, mm Hg Preimplantation echocardiogram 33 (5 97) 37 (5 69) 29 (8 97) 0.03 Discharge echocardiogram 17 (3 36) 18 (4 51) 16 (3 36) 0.01 AT indicates anaerobic threshold; FEV 1, forced expiratory volume in 1 second; FVC, forced vital capacity; PR, pulmonary regurgitation; RVOT, right ventricular outflow tract obstruction; TPVR, transcatheter pulmonary valve replacement; and V e/v co 2, ratio of minute ventilation to carbon dioxide production at the anaerobic threshold. *Comparison of pediatric and adult cohorts. Minimum conduit diameter measured by angiography at the implantation catheterization. Implantation indication was defined as obstruction, PR, or mixed (met criteria for both obstruction and PR), as described previously. 2,3 P Value* redilation rather than TPV-in-TPV implantation (P=0.048; Figure 2). Freedom from subsequent RVOT reintervention after TPV-in-TPV implantation (either as an initial reintervention or after prior Melody valve dilation) was 87±9% at 3 years after the second Melody valve was implanted (Figure 2). Melody Valve Explantation A total of 11 patients underwent Melody valve explantation during follow-up, including the 7 whose initial reintervention was Melody valve explantation and 4 who had the valve explanted after prior transcatheter reintervention. At 5 years, freedom from explantation was 92±3% (Figure 2). The limited

5 1964 Circulation June 2, 2015 Table 2. Diagnostic and Procedural Factors Associated With a Higher Discharge RVOT Gradient Variable Mean Doppler RVOT Gradient at Discharge, mm Hg P Value Preimplantation mean Doppler 0.04 gradient, mm Hg >35 (n=70) 19 (3 51) 35 (n=78) 15.5 (4 33) Primary indication for TPVR RVOT obstruction or mixed 19 (3 51) disease (n=69) PR (n=79) 16 (4 33) Procedure order within a given trial center First 10 patients (n=49) 20 (7 34) After first 10 patients (n=99) 17 (3 51) Delivery system size, mm (n=23) 21 (4 36) 20 or 22 (n=125) 17 (3 51) Includes only the 148 patients discharged with a transcatheter pulmonary valve in place. PR indicates pulmonary regurgitation; RVOT, right ventricular outflow tract obstruction; and TPVR, transcatheter pulmonary valve replacement. number of events precluded multivariable analysis, but on univariable Cox regression, a primary implantation indication of RVOT obstruction or mixed disease (ie, not primary PR; HR, 5.2; 95% CI, ; P=0.038), a smaller ratio of angiographic to nominal conduit diameter (HR, 0.47 per 0.1 difference in the ratio; 95% CI, ; P=0.001), and moderate or severe tricuspid regurgitation on the discharge echocardiogram (HR, 5.4; 95% CI, ; P=0.008) were associated with shorter freedom from explantation. Other Reinterventions Reinterventions during follow-up that did not involve the RVOT are listed in Table I in the online-only Data Supplement. Melody Valve Stent Fracture An MSF was diagnosed in 50 patients and a type II MSF was diagnosed in 25 patients (Figure I in the online-only Data Supplement). At 5 years, freedom from any MSF was 61±5% and from a type II MSF was 83±3%. Freedom from a type II MSF after the first MSF diagnosis was 50±8% at 3 years (Figure 3). Of 7 patients who were diagnosed with a type II MSF 2 years after the initial MSF diagnosis, most (n=5) did not have a prestent placed, and all had mean Doppler gradients at discharge <25 mm Hg. Freedom from reintervention after diagnosis of a stent fracture was 51±8% at 3 years (Figure 3) and in multivariable Cox regression was shorter in patients with a higher discharge RVOT gradient (HR, 1.07 per 1 mm Hg; 95% CI, ; P=0.003) and a ratio of angiographic to nominal conduit diameter <0.65 (HR, 4.2; 95% CI, ; P=0.008). For the 25 patients with MSF who had not undergone reintervention, the RVOT gradient ranged from 4 to 39 mm Hg (median, 16 mm Hg) at the most recent follow-up a median of 2.9 years later. Figure 2. Top, Kaplan Meier curves depicting cumulative survival, freedom from Melody valve explantation, and freedom from reintervention on the Melody valve over time. Bottom, Kaplan Meier curves depicting freedom from a subsequent reintervention after an initial transcatheter reintervention (Any) or after transcatheter pulmonary valve (TPV) in TPV implantation (redo TPV replacement [TPVR]). Other Adverse Events A total of 14 trial patients were reported to have definite or presumed endocarditis or a bloodstream infection; at 5 years, freedom from endocarditis was 89±3%. Five of these patients underwent reintervention on the Melody valve within 3 months of the diagnosis (1 explantation, 2 TPV-in-TPV implantations, 1 dilation, and 1 dilation followed shortly by explantation; Figure 4). Endocarditis was the indication for reintervention in 2 patients who were found to have new TPV obstruction or PR, but the other 3 patients had TPV obstruction before the diagnosis of infection, and reintervention was indicated on the basis of the existing stenosis. The other 9 patients improved on medical therapy with no change in Melody valve function; 2 of them underwent TPV reintervention remote from the diagnosis of infection (1.8 and 2.2 years later). One patient developed endocarditis for the first time 16 months after RVOT reintervention, and another had a second episode 2.7 years after TPV-in-TPV implantation. Melody Valve Function Among the 113 patients who were alive and free from reintervention a median of 4.5 years after implantation, the most recent RVOT gradient was unchanged from early after TPVR (median, 15 mm Hg [range, 4 45 mm Hg] compared with 17 mm Hg [range, 3 36 mm Hg] on the discharge echocardiogram; P=0.7). On the most recent echocardiogram, PR was

6 Cheatham et al Transcatheter PVR Outcomes to 7 Years 1965 Table 3. Indications for and Details of RVOT Reinterventions n RVOT reintervention 35 First TPV reintervention 33* TPV dilation without replacement 6 TPV-in-TPV implantation 19 Surgical PVR 8* Second TPV reintervention 7 TPV dilation after TPV-in-TPV implantation 1 TPV-in-TPV implantation after TPV dilation 1 TPV-in-TPV after TPV-in-TPV implantation 1 Surgical PVR after TPV dilation 2 Surgical PVR after TPV-in-TPV implantation 2 RVOT conduit stent proximal or distal to TPV 3 Surgical PVR after non-tpv RVOT stent 1 Indication for first TPV reintervention RVOT obstruction 27 Associated with Melody valve stent fracture 22 Recurrent stenosis 2 Concomitant TV repair for progression of TR 1 Moderate obstruction, progressive RV dysfunction 1 Endocarditis 2 With consequent TPV regurgitation 1 With consequent TPV obstruction 1 RV dysfunction, concomitant RVAD implantation and mitral 1 valve repair Pulmonary hypertension, increasing RV pressure 1 Conduit rupture at time of TPV implant 1 Conduit regurgitation without endocarditis 0 Mean Doppler RVOT gradient before first TPV reintervention 37 (11 72) Data presented as frequency or median (minimum maximum). PVR indicates pulmonary valve replacement; RV, right ventricle; RVAD, right ventricular assist device; RVOT, right ventricular outflow tract obstruction; TPV, transcatheter pulmonary valve; TR, tricuspid regurgitation; and TV, tricuspid valve. *These values include the patient who underwent emergent surgical PVR as a result of conduit rupture at the time of implantation; in the text, this acute reintervention is presented separately from reinterventions that occurred during follow-up. One of the patients who had a stent placed distal to the Melody valve subsequently underwent TPV explantation. Does not include 3 patients who underwent reintervention within 3 months of being diagnosed with endocarditis but had pre-existing stenosis as the reported primary indication for reintervention. absent/trivial PR in 97 of these patients, mild in 15 patients, and moderate in 1 patient. Functional and Cardiopulmonary Status Before TPVR, 14% of patients were in New York Heart Association (NYHA) class I and 17% were in class III or IV. At every postimplantation annual evaluation, at least 74% of patients were in class I and most others were in class II, with no more than 1% to 2% in class III or IV (Figure 5). Patients who were in NYHA class III or IV before TPVR differed from those in class I or II in several diagnostic indexes, as summarized in Table II in the onlineonly Data Supplement. Table 4. Result of Multivariable Cox Regression Analysis of Factors Associated With TPV Reintervention Variable HR (95% CI) P Value Homograft conduit 5.2 ( ) 0.01 Conduit angiographic/nominal 2.6 ( ) diameter ratio <0.65 Any unfractured prestent (new 0.29 ( ) or existing) Discharge RVOT mean gradient 5.9 ( ) <0.001 >25 mm Hg Moderate or severe TR before implantation 2.3 ( ) CI indicates confidence interval; HR, hazard ratio, and TPV, transcatheter pulmonary valve. Cardiopulmonary exercise testing and spirometry data at baseline are summarized in Table 1. Baseline pulmonary function metrics demonstrated obstructive and restrictive patterns in the majority of patients, with an abnormal (<80% of predicted) forced expiratory volume in the first second of expiration (FEV 1 ) in 58% of patients, an abnormal forced vital capacity (FVC) in 53%, and abnormalities of both in 50%. Severely abnormal (<50% of predicted) FEV 1 and FVC were present in 15% and 14% of patients, respectively. Percent predicted FEV 1 and FVC were both significantly higher at baseline in pediatric patients than adults. There were no significant Figure 3. Top, Kaplan Meier curves depicting cumulative freedom from a diagnosis of any Melody valve stent fracture and from a type II fracture. Bottom, Kaplan Meier curves depicting freedom from diagnosis of a type II Melody valve stent fracture and from Melody valve reintervention after the first diagnosis of any stent fracture.

7 1966 Circulation June 2, 2015 Figure 4. Flow diagram depicting the pattern of reinterventions in the 14 patients who were diagnosed with endocarditis or a bloodstream infection. Two patients developed endocarditis after transcatheter pulmonary valve (TPV) in TPV implantation (redo TPV), 1 as an initial episode and 1 who had a recurrent episode after the initial episode before a second TPV was implanted. The time frame for reinterventions after a diagnosis of endocarditis is summarized in the text. changes in FEV 1 from before implantation to 1 year (median, 75% of predicted to 79%; P=0.4) or FVC (median, 77% of predicted to 80%; P=0.5). The maximum workload achieved increased from a median of 105 W (range, W) before TPVR to 121 W (range, W) at the 1-year evaluation (P=0.01) and remained stable at 2 and 3 years. In the overall population, however, there were no significant changes in the percent of predicted peak oxygen consumption (percent V o 2 max). As reported previously, 14 the ratio of minute ventilation to carbon dioxide production at the anaerobic threshold (V e/v co 2 ) improved early after TPVR and remained stable thereafter throughout followup (median, at all follow-up evaluations; P 0.001). Other than comparison of exercise measures according to age (Table 1), NYHA class (Table II in the online-only Data Supplement), and baseline spirometry (below), no subgroup analyses of exercise function were performed for this study. Figure 5. Bar graph depicting New York Heart Association functional class before transcatheter pulmonary valve replacement and at each annual follow-up interval through 6 years as a percentage of patients with evaluation at each time point. The distribution changed significantly from before implantation to 1 year (P<0.001), but there was no difference from 1 year to subsequent follow-up time points. The 68 patients in whom both FEV 1 and FVC were abnormal at baseline had lower percent V o 2 max (median, 56% of predicted [range, 21% 89%] versus 61% [range, 26% 122%]; P=0.005) and maximum work (91 W [range, W] versus 119 W [range, W]; P=0.002) than those in whom 1 or both were normal. The percent V o 2 max was normal in only 3 of these 68 patients (4%) and <50% of predicted in 26 (38%). Likewise, among the 15 patients with severely depressed FEV 1 and FVC, the median V o 2 max was 34% of predicted (range, 21% 73%). Baseline respiratory abnormalities were not associated with changes in percent V o 2 max after TPVR, but there was greater improvement in maximum work among patients with a normal FEV 1 or FVC compared with those in whom both were abnormal (P<0.03). Changes in Practice and Outcome Over Time There were several changes in practice over the course of the IDE trial. Conduit prestenting was not permitted during the first 35 implantations, after which the protocol was modified to allow concomitant procedures. Accordingly, placement of a new prestent was less common in the first 50-patient tertile than in the second or third tertile (10% versus 54% and 42%; P<0.001). Overall, only 36% of patients underwent prestenting (Figure 1). In addition, there were changes in implantation practice that were not associated with protocol modifications. For example, 32% of patients in the first tertile underwent TPVR with the largest 22-mm delivery system, but this increased to 56% during the second tertile and 84% during the third (all P 0.008), whereas use of the smallest 18-mm system decreased from 36% to 6% to 4% (the nominal conduit diameter was 1.5 mm smaller in patients in the first tertile than in those treated later; P<0.001). In contrast, postimplantation dilation of the Melody valve was more common in the first 50 patients (66%) than the second (38%) or third (36%; P=0.004). There were also differences in outcomes between the first and second 50 patients, including a higher acute post-tpvr RVOT gradient (median, 14 mm Hg [range, 3 37 mm Hg] versus 11 mm Hg [range, 0 24 mm Hg]; P=0.006), shorter freedom from a diagnosis of any MSF (HR, 2.8; 95% CI, ; P=0.004) or a type II MSF (HR, 3.3; 95% CI, ; P=0.025), and shorter freedom from TPV reintervention (HR, 2.9; 95% CI, ; P=0.021). As depicted in Figure II in the online-only Data Supplement, 5-year freedom from Melody valve reintervention was 68±7% in patients in the first tertile compared with 87±5% in patients in the second tertile (P=0.02). The differences in freedom from MSF and reintervention were not significant on multivariable Cox regression because implantation tertile was collinear with prestenting. Discussion Intermediate-Term Melody Valve Function and Clinical Outcome One of the most notable findings of this study was that the Melody TPV continued to function well, with little progression of RVOT obstruction and excellent valve competence in the majority of patients, out to at least 7 years after implantation. A subset of patients developed RVOT obstruction, almost always associated with an MSF or inadequate gradient relief

8 Cheatham et al Transcatheter PVR Outcomes to 7 Years 1967 at the time of implantation. However, in the 113 patients who were alive and free from reintervention a median of 4.5 years after TPVR, Melody valve function was unchanged from the early postimplantation echocardiogram. Short-term follow-up from prior studies reported similarly excellent valve function, 2,3,5 7,9 but no data on longer-term valve function have been published. In that regard, the findings of this study are encouraging. Clinical improvement in this cohort paralleled the hemodynamic benefits of TPVR, with almost all patients in a better NYHA class after implantation than before and few patients worse than class II at any point during follow-up. There was also a significant increase in the maximum work achieved on exercise testing, which remained higher throughout followup. However, as in prior studies, most measures of cardiopulmonary function at peak exercise were not improved at 1 year after TPVR. 2,14 Similarly, there was no change in submaximal indexes, namely, V o 2 at the anaerobic threshold. This discordance between improvement in subjective functional status and objective measures of exercise cardiopulmonary function is one of the vexing issues in this population and may be related to the high prevalence of pulmonary dysfunction, as evidenced by frequent abnormalities of FEV 1 and FVC in this cohort and proposed by others. 23 Notably, patients with normal spirometry before TPVR experienced greater improvement in peak work after implantation. In addition, because NYHA class tends to reflect perceptions of limitation at submaximal exercise, this discrepancy may indicate a difference between experiential and peak exercise derived cardiopulmonary limitations. This hypothesis is supported by the higher maximum work rate after TPVR in this series and by the significant cardiac, exercise, and respiratory differences between patients in lower and higher NYHA classes. Prior studies found inconsistent physiological effects of TPVR, with improved exercise function in patients with RVOT obstruction as an indication for implantation but not in those with primary PR. 10,11,16 Although the present study did not examine exercise function in depth, the long-term relationship among post-tpvr changes in hemodynamics, symptomatic status, and exercise physiology deserves further detailed analysis. Stent Fracture, RVOT Obstruction, and Melody Valve Reintervention Another notable finding of this study was the ongoing risk beyond 4 years after TPVR of MSF and associated RVOT obstruction. With the additional follow-up duration and greater number of outcome events than in prior studies, this analysis also provided more robust estimates of progression of MSF from minor to major and of freedom from RVOT reintervention after diagnosis of an MSF. These observations provide further support for the previously advanced argument that an MSF per se is not necessarily a critical adverse outcome; rather, the relevant outcome is recurrent RVOT obstruction that results from loss of TPV integrity, which is a frequent but not inevitable outcome of an MSF. The most common indication for reintervention after TPVR was MSF with recurrent RVOT obstruction, interrelated outcomes that prior studies have shown to be associated with a constellation of factors. 4,19 In an effort to understand better the interaction between TPV protection and gradient relief, we performed a focused analysis to assess the relationship between Melody valve reintervention and combinations of prestenting status and preimplantation and postimplantation RVOT gradients. In that analysis, prestenting was associated with a low risk for reintervention regardless of preimplantation or postimplantation gradient (too few prestented patients underwent reintervention to determine the importance of residual gradient), but so was the combination of low preimplantation and discharge gradients regardless of prestent status. Patients at highest risk for reintervention were those who did not receive a prestent and had high RVOT gradients before implantation and at discharge. This analysis indicates that prestenting is protective against reintervention but also suggests that more thorough reduction of RVOT obstruction in patients will result in better long-term outcomes, and it identifies patients with minor conduit obstruction as a subset in whom the benefits of prestenting are less clear. TPV implanters are often faced with the dilemma of how aggressively to treat RVOT stenosis, whether to leave a mild residual gradient or to expand the conduit further. We and others have argued that thorough gradient reduction is important for a combination of reasons. 4,21,24 The above analysis supports that contention in patients without a prestent, and we expect that it will apply in prestented patients, although there were too few reinterventions in that cohort to determine risk factors. This is a difficult question to address retrospectively and is beset by one of the major challenges in studying TPV therapy, namely heterogeneity of the target population and implantation environment, and variability of the procedure itself. This diversity, in a small sample, makes it difficult to adjust adequately for potential confounders, whereas blanket evaluation of the full cohort may obscure important differences between subgroups. Accordingly, as TPVR therapy expands to younger and smaller patients or those with dysfunction of a native or patched RVOT, 25,26 efforts to dissect the population with subgroup or interaction analyses will be critical if we are to achieve an incisive understanding of best practices for Melody valve therapy. The aforementioned analysis was one attempt to drill down, and substantiated the benefit of thoroughly relieving RVOT obstruction independently from prestenting. As our understanding of the benefits of prestenting and risk factors for MSF improves, it is plausible that major MSF after TVPR will be nearly eliminated. Along these lines, it seems reasonable to propose that the patients in this trial who were not presented, but then developed MSF with recurrent RVOT obstruction, and subsequently underwent TPV reintervention, may not be representative of current practice. To provide a more realistic, albeit hypothetical, estimate of the durability of a protected Melody valve, we simulated a scenario in which the problems of major MSF and consequent RVOT obstruction are overcome by analyzing freedom from reintervention in patients who did not develop a type II MSF. In that speculative analysis, 5-year freedom from TPV reintervention was 91±3% compared with 76±4% in the full cohort, and was shorter in patients with a higher discharge RVOT gradient. These observations are

9 1968 Circulation June 2, 2015 consistent with our previously advanced supposition about the importance of residual obstruction in prestented patients and support the recommendation for thorough relief of RVOT obstruction at the time of TPVR. Other Adverse Melody Valve-Related Outcomes Consistent with recent reports focused on the issue of endocarditis after TPVR, 21,27 29 infection was a notable adverse outcome in this series. At the time the database was locked for this analysis, only basic data were available for most of the reported cases of endocarditis because details were not included among routine follow-up data collection elements, so only preliminary analysis was performed. However, endocarditis or sepsis was a contributing factor in 1 of 4 late deaths and was the proximate indication for reintervention in 2 of 32 patients (1 was the patient who died) and a mitigating factor in 3 others. Thus, infection and endocarditis, which were explored in the IDE population in combination with 2 other prospective Melody valve trial cohorts in a prior report, 21 are clearly a high-priority issue for ongoing investigation and will continue to be a focal point in analyses of this study cohort and this technology in general. Outcomes of Transcatheter Melody Valve Reintervention TPV-in-TPV implantation provided good relief of recurrent RVOT obstruction and freedom from surgical pulmonary valve replacement for at least several years in most patients. All of the patients who underwent TPV-in-TPV implantation had developed RVOT obstruction in association with MSF, and additional prestents (typically 2) were placed at the time of reintervention. The durability of the second valve in a protected environment, after the failure of a first unprotected valve, supports the utility of prestenting and the general hemodynamic benefit of TPVR with the Melody valve. The role of Melody valve redilation for obstruction without associated MSF is less clear. Redilation was performed in 6 patients, most of whom underwent TPVR early in the trial, had the valve implanted on the smallest delivery system, and had a relatively high discharge gradient, possibly indicating suboptimal initial therapy. There were too few patients to estimate freedom from further reintervention in this cohort, but 3 of the 6 patients underwent another reintervention within a year, suggesting that the benefits of redilation may be limited. Nevertheless, it is possible that isolated TPV redilation will be effective in some patients such as those who undergo TPVR at a small size and develop RVOT obstruction as a result of somatic growth. Learning Curve There were several noteworthy changes in implantation practices and outcomes over time, which likely reflected learning curve effects, protocol modifications, and an evolving understanding of better practices. 3,5,19,22 Patients implanted during the first third of the trial were less likely to have a new prestent placed and more likely to undergo TPVR with the smallest delivery system (18 mm) and to have the implanted Melody valve postdilated. These differences corresponded to less complete relief of RVOT obstruction and shorter freedom from MSF and reintervention in the first 50 patients than in those enrolled later. Because this series represents the first prospective multicenter trial and the first experience with TPVR in the United States, the results in the earliest patients, marked by a relatively high rate of MSF and consequent reintervention, reflect a naïve experience. Thus, although the data provide important insights, they are not representative of contemporary best practices, which include maximal relief of RVOT obstruction and, in most cases, implantation of 1 conduit prestents. Once the trial protocol allowed prestenting, >50% of patients received a prestent, but still substantially fewer than in more recent series. 6,7,30 Accordingly, although these data will necessarily serve as a landmark, given that they represent the longest reported follow-up, the freedomfrom-event analyses of this cohort most likely overestimate the risk of MSF, reintervention, and explantation relative to contemporary practice. Limitations Although this study is unique and provides important new data, it has important limitations. The population was heterogeneous, and several procedural parameters were not strictly guided by the trial protocol, such as whether to prestent, how many or what type of stents to implant, and how aggressively to reduce the RVOT gradient before or after Melody valve implantation. The latitude that these discretionary factors provided to the operator was clinically necessary but potentially confounded the analysis of outcomes. Similarly, although there was general consensus among investigators about a reasonable threshold for reintervention (a mean Doppler RVOT gradient >35 40 mm Hg), criteria for and methods of reintervention were not specified by the protocol, and practices varied. The statistical analyses in this study were descriptive with no prespecified hypotheses and therefore should be considered exploratory. There were no multiple testing adjustments for controlling type I error. A number of patients (who typically lived elsewhere) did not attend the most recent prescribed protocol evaluation at the implanting center, although most indicated an intention to continue in the study; thus, the total available follow-up for this analysis was less than anticipated. This may have introduced limited bias, although it was known for most of these patients that no reintervention or serious adverse events had occurred (still, only data collected per protocol were included). Conclusions TPVR with the Melody valve continues to provide good hemodynamic and clinical results 4 to 7 years after implantation. Primary valve failure was rare. The main cause of TPV dysfunction was stenosis related to MSF, which was less common once prestenting became more widely adopted, as other reports also demonstrated. 4,6,7,31 Efforts to prevent or reduce hemodynamically important MSF are likely to yield even better long-term outcomes than documented in this series and should be an ongoing focus. Similarly, endocarditis was an important adverse outcome during follow-up, and continued efforts to understand the pathophysiology of and risk factors for endocarditis after TPVR are warranted.

10 Cheatham et al Transcatheter PVR Outcomes to 7 Years 1969 Acknowledgments We acknowledge Jessica Dries-Devlin, Kristin Lawman, and Te-Hsin Lung for assistance with this manuscript. Source of Funding This study was sponsored by Medtronic, Inc. Disclosures Drs Cheatham, Zahn, Hellenbrand, Jones, Vincent, and McElhinney serve as consultants to Medtronic, Inc, the manufacturer of the Melody valve, and all authors act as investigators or proctors. Dr Jones has received research funding from Medtronic, Inc. Dr Cheatham serves as a consultant for NuMED. References 1. Bonhoeffer P, Boudjemline Y, Saliba Z, Merckx J, Aggoun Y, Bonnet D, Acar P, Le Bidois J, Sidi D, Kachaner J. Percutaneous replacement of pulmonary valve in a right-ventricle to pulmonary-artery prosthetic conduit with valve dysfunction. Lancet. 2000;356: doi: / S (00) Zahn EM, Hellenbrand WE, Lock JE, McElhinney DB. Implantation of the Melody transcatheter pulmonary valve in patients with dysfunctional right ventricular outflow tract conduits: early results from the U.S. clinical trial. J Am Coll Cardiol. 2009;54: doi: /j.jacc McElhinney DB, Hellenbrand WE, Zahn EM, Jones TK, Cheatham JP, Lock JE, Vincent JA. Short- and medium-term outcomes after transcatheter pulmonary valve placement in the expanded multicenter U.S. Melody valve trial. Circulation. 2010;122: doi: /CIRCULATION AHA McElhinney DB, Cheatham JP, Jones TK, Lock JE, Vincent JA, Zahn EM, Hellenbrand WE. Stent fracture, valve dysfunction, and right ventricular outflow tract reintervention after transcatheter pulmonary valve implantation: patient-related and procedural risk factors in the US Melody Valve Trial. Circ Cardiovasc Interv. 2011;4: doi: / CIRCINTERVENTIONS Lurz P, Coats L, Khambadkone S, Nordmeyer J, Boudjemline Y, Schievano S, Muthurangu V, Lee TY, Parenzan G, Derrick G, Cullen S, Walker F, Tsang V, Deanfield J, Taylor AM, Bonhoeffer P. Percutaneous pulmonary valve implantation: impact of evolving technology and learning curve on clinical outcome. Circulation. 2008;117: doi: / CIRCULATIONAHA Butera G, Milanesi O, Spadoni I, Piazza L, Donti A, Ricci C, Agnoletti G, Pangrazi A, Chessa M, Carminati M. Melody transcatheter pulmonary valve implantation: results from the registry of the Italian Society of Pediatric Cardiology (SICP). Catheter Cardiovasc Interv. 2013;81: doi: /ccd Eicken A, Ewert P, Hager A, Peters B, Fratz S, Kuehne T, Busch R, Hess J, Berger F. Percutaneous pulmonary valve implantation: two-centre experience with more than 100 patients. Eur Heart J. 2011;32: doi: /eurheartj/ehq Gillespie MJ, Rome JJ, Levi DS, Williams RJ, Rhodes JF, Cheatham JP, Hellenbrand WE, Jones TK, Vincent JA, Zahn EM, McElhinney DB. Melody valve implant within failed bioprosthetic valves in the pulmonary position: a multicenter experience. Circ Cardiovasc Interv. 2012;5: doi: /CIRCINTERVENTIONS Armstrong AK, Balzer DT, Cabalka AK, Gray RG, Javois AJ, Moore JW, Rome JJ, Turner DR, Zellers TM, Kreutzer J. One-year follow-up of the Melody transcatheter pulmonary valve multicenter post-approval study. JACC Cardiovasc Interv. 2014;7: doi: /j. jcin Coats L, Khambadkone S, Derrick G, Sridharan S, Schievano S, Mist B, Jones R, Deanfield JE, Pellerin D, Bonhoeffer P, Taylor AM. Physiological and clinical consequences of relief of right ventricular outflow tract obstruction late after repair of congenital heart defects. Circulation. 2006;113: doi: /CIRCULATIONAHA Coats L, Khambadkone S, Derrick G, Hughes M, Jones R, Mist B, Pellerin D, Marek J, Deanfield JE, Bonhoeffer P, Taylor AM. Physiological consequences of percutaneous pulmonary valve implantation: the different behaviour of volume- and pressure-overloaded ventricles. Eur Heart J. 2007;28: doi: /eurheartj/ehm Romeih S, Kroft LJ, Bokenkamp R, Schalij MJ, Grotenhuis H, Hazekamp MG, Groenink M, de Roos A, Blom NA. Delayed improvement of right ventricular diastolic function and regression of right ventricular mass after percutaneous pulmonary valve implantation in patients with congenital heart disease. Am Heart J. 2009;158: doi: /j.ahj Lurz P, Puranik R, Nordmeyer J, Muthurangu V, Hansen MS, Schievano S, Marek J, Bonhoeffer P, Taylor AM. Improvement in left ventricular filling properties after relief of right ventricle to pulmonary artery conduit obstruction: contribution of septal motion and interventricular mechanical delay. Eur Heart J. 2009;30: doi: /eurheartj/ehp Batra AS, McElhinney DB, Wang W, Zakheim R, Garofano RP, Daniels C, Yung D, Cooper DM, Rhodes J. Cardiopulmonary exercise function among patients undergoing transcatheter pulmonary valve implantation in the US Melody valve investigational trial. Am Heart J. 2012;163: doi: /j.ahj Brown DW, McElhinney DB, Araoz PA, Zahn EM, Vincent JA, Cheatham JP, Jones TK, Hellenbrand WE, O Leary PW. Reliability and accuracy of echocardiographic right heart evaluation in the U.S. Melody Valve Investigational Trial. J Am Soc Echocardiogr. 2012;25: e4. doi: /j.echo Lurz P, Nordmeyer J, Giardini A, Khambadkone S, Muthurangu V, Schievano S, Thambo JB, Walker F, Cullen S, Derrick G, Taylor AM, Bonhoeffer P. Early versus late functional outcome after successful percutaneous pulmonary valve implantation: are the acute effects of altered right ventricular loading all we can expect? J Am Coll Cardiol. 2011;57: doi: /j.jacc Müller J, Engelhardt A, Fratz S, Eicken A, Ewert P, Hager A. Improved exercise performance and quality of life after percutaneous pulmonary valve implantation. Int J Cardiol. 2014;173: doi: /j. ijcard Harrild DM, Marcus E, Hasan B, Alexander ME, Powell AJ, Geva T, McElhinney DB. Impact of transcatheter pulmonary valve replacement on biventricular strain and synchrony assessed by cardiac magnetic resonance feature tracking. Circ Cardiovasc Interv. 2013;6: doi: / CIRCINTERVENTIONS Nordmeyer J, Khambadkone S, Coats L, Schievano S, Lurz P, Parenzan G, Taylor AM, Lock JE, Bonhoeffer P. Risk stratification, systematic classification, and anticipatory management strategies for stent fracture after percutaneous pulmonary valve implantation. Circulation. 2007;115: doi: /CIRCULATIONAHA Morray BH, McElhinney DB, Cheatham JP, Zahn EM, Berman DP, Sullivan PM, Lock JE, Jones TK. Risk of coronary artery compression among patients referred for transcatheter pulmonary valve implantation: a multicenter experience. Circ Cardiovasc Interv. 2013;6: doi: /CIRCINTERVENTIONS McElhinney DB, Benson LN, Eicken A, Kreutzer J, Padera RF, Zahn EM. Infective endocarditis after transcatheter pulmonary valve replacement using the Melody valve: combined results of 3 prospective North American and European studies. Circ Cardiovasc Interv. 2013;6: doi: /CIRCINTERVENTIONS Nordmeyer J, Coats L, Lurz P, Lee TY, Derrick G, Rees P, Cullen S, Taylor AM, Khambadkone S, Bonhoeffer P. Percutaneous pulmonary valve-in-valve implantation: a successful treatment concept for early device failure. Eur Heart J. 2008;29: doi: /eurheartj/ ehn Sterrett LE, Ebenroth ES, Query C, Ho J, Montgomery GS, Hurwitz RA, Baye F, Schamberger MS. Why exercise capacity does not improve after pulmonary valve replacement. Pediatr Cardiol. 2014;35: doi: /s Hasan BS, Lunze FI, Chen MH, Brown DW, Boudreau MJ, Rhodes J, McElhinney DB. Effects of transcatheter pulmonary valve replacement on the hemodynamic and ventricular response to exercise in patients with obstructed right ventricle-to-pulmonary artery conduits. JACC Cardiovasc Interv. 2014;7: doi: /j.jcin Berman DP, McElhinney DB, Vincent JA, Hellenbrand WE, Zahn EM. Feasibility and short-term outcomes of percutaneous transcatheter pulmonary valve replacement in small (<30 kg) children with dysfunctional right ventricular outflow tract conduits. Circ Cardiovasc Interv. 2014;7: doi: /CIRCINTERVENTIONS Meadows JJ, Moore PM, Berman DP, Cheatham JP, Cheatham SL, Porras D, Gillespie MJ, Rome JJ, Zahn EM, McElhinney DB. Use and performance of the Melody Transcatheter Pulmonary Valve in native and postsurgical, nonconduit right ventricular outflow tracts. Circ Cardiovasc Interv. 2014;7: doi: / CIRCINTERVENTIONS