Congenital Heart Disease

Transcription

1 Congenital Heart Disease Randomized Trial of Cutting Balloon Compared With High-Pressure Angioplasty for the Treatment of Resistant Pulmonary Artery Stenosis Lisa Bergersen, MD, MPH; Kimberlee Gauvreau, ScD; Henri Justino, MD; Alan Nugent, MD; Jonathon Rome, MD; Jacqueline Kreutzer, MD; John Rhodes, MD; David Nykanen, MD; Evan Zahn, MD; Larry Latson, MD; Phillip Moore, MD; James Lock, MD; Kathy Jenkins, MD, MPH Background We sought to determine the safety and efficacy of Cutting Balloon therapy (CB) compared with conventional high-pressure balloon therapy (HPB) for the treatment of pulmonary artery stenosis. Methods and Results This prospective, randomized, multicenter, investigational device exemption trial compared CB with HPB. Patient eligibility was determined at the precatheterization assessment; vessel eligibility was determined at catheterization. In all vessels, low-pressure balloon dilation to 8 atm was performed, and if it was not successful, the vessel was randomized to CB or HPB. The primary efficacy outcome was percent change in minimum lumen diameter. A core laboratory performed all vessel measurements and angiographic assessment of vessel damage. The primary safety outcome was any serious adverse event attributable to vessel dilation as assessed by the Data and Safety Monitoring Board. Seventy-three patients from 8 institutions were enrolled between 2004 and In these patients, 72 vessels responded to low-pressure balloon dilation. Of the 173 vessels that met eligibility criteria, 107 were randomized to CB and 66 to HPB. In randomized vessels, CB therapy was associated with greater percent increase in lumen diameter (85% versus 52%; P 0.004). After crossover was introduced, 26 of 47 vessels treated with HPB underwent CB therapy and experienced an additional 48% increase in lumen diameter; the final diameter after CB was 99% greater than the initial diameter. There were no serious adverse events related to treatment in a study vessel. Conclusion CB therapy for pulmonary artery stenosis not responsive to low-pressure balloon is more effective than HPB therapy and has an equivalent safety profile. (Circulation. 2011;124: ) Key Words: angioplasty balloon dilation congenital abnormalities pulmonary artery randomized controlled trial Pulmonary artery (PA) stenosis is a form of congenital heart disease that occurs in isolation and as a complicating feature of more complex malformations with multiple obstructions in the distal vasculature, especially tetralogy of Fallot. 1 Transcatheter treatment has become the primary mode of treatment for these vascular obstructions, and is 1 of the most common interventional procedures in laboratories focused on the treatment of congenital heart disease. 1 5 The goal of therapy is to improve pulmonary blood flow, thus reducing the work (pressure) of the right ventricle in an effort to improve or preserve right ventricular function, to decrease cyanosis, or to mitigate against the risk of death. Limited surgical options exist, particularly in the distal vasculature. Clinical Perspective on p 2396 Beginning in the early 1980s, low-pressure balloon angioplasty was shown to be successful in increasing lumen diameter in about half the vessels treated. 6 These results improved with the development of balloons capable of higher inflation pressures; however, a subset of vessels, 30%, remained resistant to conventional balloon techniques. 7,8 For these vessels, treatment options did not exist until the offlabel use of the Cutting Balloon (CB) was reported to be effective for PA vessels resistant to conventional highpressure balloon (HPB) therapy, although the risk profile of CB therapy for primary treatment was unclear Given the important role this technology potentially offered for pediat- Received January 31, 2011; accepted September 19, From the Children s Hospital Boston, Boston, MA (L.B., K.G., J.L., K.J.); Texas Children s Hospital, Houston (H.J.); UT Southwestern Medical Center, Dallas, TX (A.N.); Children s Hospital of Philadelphia, Philadelphia, PA (J. Rome); Children s Hospital Pittsburgh, Pittsburgh, PA (J.K.); Duke University Medical Center, Durham, NC (J. Rhodes); Miami Children s Hospital, Miami, FL (D.N., E.Z.); Cleveland Clinic Foundation, Cleveland, OH (L.L.); and University of California San Francisco, San Francisco (P.M.). Guest Editor for this article was Carole A. Warnes, MD. The online-only Data Supplement is available with this article at /-/DC1. Correspondence to Lisa Bergersen, MD, MPH, Department of Cardiology, The Children s Hospital, 300 Longwood Ave, Boston, MA Lisa.Bergersen@cardio.CHBoston.org 2011 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 Bergersen et al Randomized Pulmonary Artery Cutting Balloon Trial 2389 ric patients with this disease, we sought to design a prospective randomized trial to compare the safety and effectiveness of the new CB technology with the conventional HPB in PA stenosis resistant to low-pressure angioplasty. We hypothesized that the treatments would have similar safety profiles, but that CB treatment would have superior efficacy, resulting in a greater percent increase in lumen diameter. Methods Patient Eligibility Patients referred for catheterization with suspected PA stenosis at 8 investigational sites were screened for patient-level eligibility, including at least 1 secondary effect from PA stenosis: (1) greater than half systemic right ventricular pressure, (2) regional decrease in pulmonary blood flow by lung scan, (3) elevated PA pressures ( 20 mm Hg mean), or (4) cyanosis due at least in part to PA obstruction. Patients were excluded if pregnant or if PA surgery had occurred in the prior 6 weeks. If the patient or parent or guardian agreed to participate in follow-up testing, informed consent was obtained. Enrollment in the trial occurred if both patient-level eligibility as determined at the precatheterization assessment was met and at least 1 eligible vessel was identified for randomization at the time of the catheterization procedure. All participating centers received Institutional Review Board approval. Figure 1. Cutting Balloons (CB; Boston Scientific, Inc, Natick, MA) are noncompliant balloons with atherotomes (microsurgical blades) mounted longitudinally on their outer surface. A, The Flextome CB. B, A magnified 2-cm-long Peripheral CB device deflated and inflated. Vessel Eligibility and Treatment Angiograms were obtained by investigators, and stenotic vessels were identified. Vessels were not eligible for randomized treatment if the area of obstruction was associated with an aneurysm ( 100% increase compared with adjacent normal vessel diameter), stent, or arterial collateral unifocalization site or was not amenable to delivery or safe position of the CB owing to unfavorable anatomy as determined by the physician after assessment of vessel, angles, size, length, and proximity to other vessels. If a vessel was determined to be eligible on the basis of these criteria, a balloon was selected on the basis of the maximum and minimum lumen diameter according to study balloon sizing charts (Table I in the online-only Data Supplement), positioned across the area of obstruction over a guidewire, and inflated to a pressure not exceeding 8 atm. If the waist in the balloon was eliminated, the dilation was considered successful, or in compliant lesions never formed in an appropriately sized balloon, then no further therapy was indicated, as indicated by the protocol. However, if a persistent waist remained, then the vessel was eligible for randomization if the waist diameter was 7.5 mm (ie, 0.5 mm less than the largest available CB). Pulmonary arteries were individually assessed for eligibility criteria, and multiple vessels could be randomized in each patient. In 1 treatment pathway, the vessel was dilated with a CB, followed by additional low-pressure dilation with a balloon 0 to 1.0 mm larger than the CB and not exceeding an inflation pressure of 8 atm. The size of the CB selected for therapy was determined from the maximum lumen diameter and size of the waist present at the initial low-pressure dilation, according to study balloon sizing charts (Table II in the online-only Data Supplement). In the second treatment pathway, the same balloon used for low-pressure dilation in the vessel was inflated to at least 15 atm but not 22 atm (or another balloon of identical diameter was used if the low-pressure angioplasty balloon was unable to reach the high pressures required), the HPB therapy. Investigators screened for vessel injury with posttreatment angiography. Postrandomized therapy vessel stenting was allowed only in the event of intraluminal damage resulting in obstruction to distal flow and accordingly required an adverse event report. Similarly, if vascular tears resulted in extravasation of contrast outside the lumen, additional angiography was required to determine whether the tear was contained locally or whether coil occlusion of the vessel would be necessary to control bleeding. The remainder of the catheterization procedure was at the discretion of the interventionalist; however, all eligible vessels were randomized. After the first 14 patients were enrolled, an amended investigational device exemption was approved by the Food and Drug Administration (FDA) that allowed crossover therapy. Crossover was permitted only in vessels that had a 50% increase in minimum lumen diameter, in which lumen diameter did not exceed 75% of the normal vessel diameter, and in which no aneurysm, defined as a focal enlargement of the vessel 100% greater than the unaffected vessel diameter, was present. Randomization Procedure Eligible vessels were randomized to the 2 treatment pathways with a randomly permuted blocks approach stratified by investigational site and prepared by the statistician. The permuted block approach (with random blocks of 2 or 4 vessels) was used to ensure blinding of the investigators to the treatment assignment before the site data coordinator opened the randomization envelope. For the first 32 patients enrolled, vessel treatment allocation was 1 to 1. For the remaining patients, the protocol was amended to modify treatment allocation with 2 vessels randomly assigned to the CB pathway for every 1 vessel assigned to HPB. Equipment CBs (Boston Scientific, Inc, Natick, MA) are noncompliant balloons with atherotomes (microsurgical blades) mounted longitudinally on their outer surface. When the CB is inflated, the atherotomes with a working height of mm score the vessel, creating initiation sites for vessel expansion. 14 This process allows dilatation of the target lesion along these linear incisions with less pressure. CBs 2 to 4 mm in diameter were approved for use in resistant coronary artery stenosis. 15 Larger-diameter (5 to 8 mm) balloons, referred to as peripheral, are indicated for treatment of obstructed dialysis fistulas. In 2005, Boston Scientific discontinued production of the 1-cm-long Peripheral CB developed by Interventional Technologies, Inc and replaced it with a 2-cm-long Peripheral CB device. The coronary device also underwent modifications during the study period, evolving from the Ultra to the Ultra 2 and finally to the Flextome CB device in 2007 (Figure 1). Standard angioplasty balloons without a labeled indication but used for pulmonary angioplasty are double-lumen catheters with a noncompliant balloon mounted at the distal tip. One lumen is used for inflation of the balloon with contrast medium; the other lumen

3 2390 Circulation November 29, 2011 permits the use of a guidewire to facilitate advancement of the catheter to and through the stenosis to be dilated. Dilatation balloon catheters are used to exert radial force to dilate narrow vessel segments by tearing the intima and media and allowing healing of the vessel with a larger-diameter lumen. 16,17 Time Interval and Data Collection Predilation and postdilation hemodynamic information, including right ventricular pressure, was obtained. After the procedure, patients were transferred to a clinical ward or to an intensive care unit in accordance with their clinical condition and usual institutional practice. Appropriate hemodynamic monitoring and nursing care were provided. The postprocedural testing was performed at hospital discharge or 48 hours after the procedure and included a physical examination with oxygen saturation and chest x-ray. Patients were typically discharged within 24 to 48 hours after their procedure. Follow-up testing was requested from the primary cardiologist 3 months after the procedure within a 2-month window and included interval history, physical examination, oxygen saturation, chest x-ray, and echocardiogram. The noninvasive tests were performed before significant invasive interventions such as surgery or catheterization procedures if they were scheduled before the clinical follow-up window. If catheterization was performed for clinical care, data for baseline hemodynamics and angiography were collected. Core Laboratory Assessment Investigators were required to obtain pretherapy and posttherapy angiography and to document the initial low-pressure balloon inflation used to determine eligibility. The core laboratory was blinded to treatment assignment and was responsible for obtaining pretreatment and posttreatment vessel measurements, assessing the low-pressure balloon inflation (to confirm vessel eligibility), and assessing the angiographic appearance of the vessel after therapy. The angiographic appearance of the vessel lumen and presence of a tear was classified according to prespecified definitions (Table III in the online-only Data Supplement). For patients referred for follow-up catheterization during the study period, further angiography was collected and measurements were made correlating to previous treatment sites. Adverse Events Patients were monitored for adverse events from the time of enrollment to follow-up assessment 2 to 4 months after the procedure. Adverse events included any event potentially related to the catheterization procedure. All events were individually reviewed and classified according to event relationship and degree of seriousness by the Data and Safety Monitoring Board (DSMB) according to prespecified definitions (Table IV in the online-only Data Supplement). Primary and Secondary Efficacy and Safety Outcomes The primary efficacy outcome was percent change in minimum lumen diameter from before to immediately after intervention. Sample size calculations were determined on the basis of the primary efficacy outcome. The secondary efficacy outcome was percent change in minimum lumen diameter from before intervention to follow-up angiography at subsequent catheterization if performed. Other secondary outcomes assessed at the patient level were change in oxygen saturation from before to after the procedure and follow-up for patients with intracardiac shunting, as well as changes in right ventricular pressure by both echocardiography and catheterization from before the procedure to the follow-up for patients without intracardiac shunting. The primary safety outcome was the proportion of vessel dilations in each treatment group resulting in any serious adverse event. Secondary safety outcomes were the proportion of vessel dilations resulting in any serious adverse event definitely attributable to the CB or HPB dilation, the proportion of vessel dilations resulting in any somewhat serious or serious adverse event, and the proportion of vessel dilations resulting in any adverse event. At the patient level, the proportion of patients exposed to the CB who experienced any serious adverse event was determined. Sample Size Sample size calculations were performed on the basis of the primary efficacy outcome percent change in minimum lumen diameter from before to immediately after intervention as assessed by angiography. Original calculations were for a comparison of the 1-cm CB and HPB. From the preliminary data, percent change in minimum lumen diameter was estimated to be 72 71% among vessels treated with HPB and % for those treated with CB. To have 90% power to detect an 81% difference in percent change with a 2-sided test conducted at the 0.05 level of significance, a sample of 34 vessels would be required in each group. Because of the correlation among multiple vessels within the same patient, we conservatively increased the number of vessels required by 25% for a total of 85 vessels. After 32 patients with 76 treated vessels were enrolled, the peripheral CB available in 5- to 8-mm diameters was changed from the 1-cm-long technology to a modified and improved 2-cm-long balloon. At this point, it was determined that, for the remainder of the study, 2 vessels would be randomly assigned to CB for every 1 vessel assigned to HPB. With the same assumptions described above and additional preliminary data for patients treated with the 2-cm balloon, it was determined that, if 30 additional patients were exposed to the 2-cm-long CB, each contributing 2 vessels on average, the power to detect a 36% difference in change in minimum lumen diameter between HPB and all CB vessels combined would be 91%. The power to detect a 50% difference between HPB and the 2-cm CB alone would be 89%. Thus, the target sample size was a minimum of 62 patients, as agreed on between the investigators and FDA. The investigational device exemption protocol specified that no more than two thirds of patients could be treated at any 1 investigative site. Statistical Methods Data from all 8 investigative sites were combined for analysis. Characteristics of the patients, catheterization procedures, and treated vessels were summarized. Primary analyses compared the 2 treatment strategies, CB versus HPB, on an intention-to-treat basis. Data from vessels treated with 1- and 2-cm-long CBs were pooled after testing for appropriateness. All vessel-level comparisons, including analyses of the primary efficacy and safety outcomes, were conducted by regression analysis with Huber-White estimates of the model coefficient SEs to adjust for the correlation among multiple vessels within the same patient. 18,19 Linear, logistic, and multinomial regression was used for continuous, dichotomous, and categorical variables, respectively. In addition to the intention-to-treat analyses, changes in minimum lumen diameter after crossover therapy were evaluated separately for vessels randomized to HPB that crossed over to CB. For secondary efficacy outcomes assessed at the patient level, comparisons from before the procedure to follow-up were assessed with the Wilcoxon signed-rank test. For patients with large unrestrictive ventricular septal defects resulting in shunting, changes in oxygen saturation from the preprocedural to the postprocedural and follow-up clinical assessments were evaluated, as well as Qp:Qs if an elective catheterization was performed. For patients without intracardiac shunting, the change in right ventricular pressure from before the procedure to follow-up was assessed by echocardiography (less than half systemic, half systemic, greater than half systemic, suprasytemic) and catheterization data (right ventricle systolic pressure to systemic systolic pressure ratio) when available. Numbers of adverse events were tabulated by their relationship to the treatment, severity, and timing in relation to the procedure. The proportion of patients exposed to a CB who experienced any serious event was determined, and a 95% confidence interval was calculated. Results Patient and Case Characteristics Seventy-three patients were enrolled at 8 centers between February 2004 and December 2008 with an investigational

4 Bergersen et al Randomized Pulmonary Artery Cutting Balloon Trial 2391 Table 1. Patient Enrollment by Investigational Sites Table 2. Patient Characteristics Site Patients, n (%) Children s Hospital Boston 44 (60) Cleveland Clinic Foundation 2 (3) Duke University Medical Center 2 (3) Children s Hospital of Philadelphia 12 (16) Miami Children s Hospital 1 (1) University of California San Francisco 3 (4) Texas Children s Hospital 8 (11) Children s Hospital Pittsburgh 1 (1) Total 73 (100) plan that no site would contribute 66% of the sample. Children s Hospital Boston contributed 44 patients (60%), and the remaining sites contributed 40% (Table 1). The majority of patients (56%) had tetralogy of Fallot; the next largest group had primary peripheral PA stenosis (Table 2). Cases were commonly performed under general anesthesia, were relatively long, and required contrast doses that exceeded 6 ml/kg in 50% of the cases (Table 3). Eighteen patients (24%) required additional procedures such as septal occlusion, collateral occlusion, or conduit interventions. The median number of vessels treated per patient was 2 (interquartile range, 1 3; maximum, 7). Vessel Treatment Among 331 stenotic vessels identified, 86 were not eligible because of 1 or multiple vessel exclusion criteria, and 245 were potentially eligible (Figure 2). In these vessels, a balloon was inflated to a pressure up to but not exceeding 8 atm. In 72 vessels (29%), the waist was eliminated, and the dilation was considered successful. In the remaining 173, a persistent waist was present, and the vessel met all eligibility criteria; 66 were randomized to HPB and 107 to CB. Sixty-eight vessels received HPB and 102 vessels received CB. Twenty-six of 73 patients (36%) received CB only; 8 (11%) received HPB only; and 39 (55%) received both (Figure 3). After crossover was permitted, 26 of 47 vessels randomized to HPB did not achieve a 50% increase in minimum lumen diameter and crossed over to CB at the discretion of the interventionalist, and 1 of 88 vessels treated with CB similarly crossed over to HPB. Three vessels randomized to CB did not receive treatment because wire position was lost and could not be reestablished in 1 vessel, the investigator was concerned about vascular damage in another vessel, and the procedure was terminated after randomization as a result of the patient s unstable hemodynamic situation in 1 case. Among the first cases performed, 2 vessels randomized to CB received HPB owing to insufficient CB size availability; when this was identified, the study was put on hold until all sites had full CB inventory. Finally, 2 vessels randomized to HPB were later determined to be ineligible as a result of significant changes with the low-pressure balloon inflated to 8 atm with near elimination of the waist, according to the core laboratory postprocedural assessment. Patient Characteristics (n 73) Sex, n (%) Male 40 (55) Female 33 (45) Age group, n (%) 1 y 6 (8) 1 4 y 32 (44) 5 10 y 21 (29) y 13 (18) 18 y 1 (1) Weight, kg (6 97) Cardiac diagnosis, n (%)* TOF with PS 8 (11) TOF with PA 33 (45) Primary PPS 22 (30) Other 10 (14) Previous catheterizations, n (0 14) History of previous surgery, n (%) 58 (79) Secondary effects from pulmonary stenosis, n (%) More than half systemic right ventricular 60 (82) pressure Regional decrease in pulmonary blood flow 58 (79) Elevated PA pressures (mean 20 mm Hg) 47 (64) Cyanosis resulting in part from pulmonary 19 (26) stenosis Intracardiac shunt present, n (%) 28 (38) TOF indicates tetralogy of Fallot; PS, pulmonary stenosis; TOF with PA, TOF with pulmonary atresia; PPS, peripheral pulmonary artery stenosis; and PA, pulmonary artery. Values are median interquartile (range) when appropriate. *Other cardiac diagnoses include double-outlet right ventricle with pulmonary stenosis, hypoplastic left heart syndrome, isolated left pulmonary artery, transposition of the great arteries with interrupted aortic arch, transposition of the great arteries with pulmonary atresia and multiple aorta pulmonary collaterals, tetralogy of Fallot with absent pulmonary valve, truncus arteriosus, unbalanced atrioventricular canal with pulmonary atresia, and hypoplastic left ventricle. Vessels were similar with respect to angioplasty site, size of the balloon that failed at 8 atm, and minimum and maximum vessel diameters (Table 4). The waist in the study balloon at the lesion in the study vessel was nearly always eliminated with CB but was eliminated in only 29% of the HPB vessels (P 0.001). Half of the vessels randomized to HPB received crossover therapy with CB when crossover was permitted. Efficacy Outcome Percent change in minimum lumen diameter was significantly greater in the CB group (n 106) compared with the HPB group (n 65) at % versus % (P 0.004). The absolute change in minimum lumen diameter was mm for the CB group and mm for the HPB group. A larger proportion of vessels treated with CB achieved a percent change in minimum lumen diameter 50% compared with HPB (Figure 4). Among the 26 vessels that underwent crossover therapy, the mean initial lumen

5 2392 Circulation November 29, 2011 Table 3. Case Characteristics Case Characteristics (n 75) Airway management and anesthesia, n (%) Conscious sedation, spontaneously 17 (23) breathing General anesthesia, intubated before 55 (73) access General anesthesia, intubated during 3 (4) the case Contrast dose, cm 3 /kg (1 15) Fluoroscopy time, min (24 192) Procedure time, min (72 551) Transfusion administered, n (%) 12 (16) Transferred, intubated at the end of the 11 (15) case, n (%) Time to hospital discharge (n 73), d* (1 19) Values are median interquartile (range) when appropriate. *For 2 patients, the case was staged and the patient received angioplasty treatment in 1 lung followed by treatment on the other side the next day. diameter measured 1.9 mm and increased to 2.6 mm after HPB. Further gains were achieved in some of these vessels with a final mean lumen diameter after CB of 3.5 mm (Figure 5). In these vessels, percent change in lumen diameter after HPB was 51% compared with 99% after CB. Among all vessels with follow-up measurements at subsequent catheterization (n 26), percent change in minimum lumen diameter was % for CB (n 45) compared with % for HPB (n 31; P 0.15) at a median follow-up of 3.4 months (range, months). Figure 2. Vessel treatment inclusion and exclusion. CB indicates Cutting Balloon; HPB, high-pressure balloon. Figure 3. Number and randomization assignment for vessels by patient. CB indicates Cutting Balloon; HPB, high-pressure balloon. Patient-Level Outcomes At clinical follow-up, no interval adverse events were identified in 69 study patients (3 patients withdrew participation after the catheterization and 1 patient was lost to follow-up). Assignments of secondary outcomes at the patient level were based on the underlying physiology; patients with intracardiac shunting were assessed on the basis of oxygen saturation. Among the 28 patients in this category, there was a small but statistically significant increase in oxygen saturation (median, 1%; range, 7% to 16%; P 0.031) from precatheterization to follow-up. However, in the 11 who underwent elective catheterization at follow-up, no significant change in Qp:Qs was detected. Patients without intracardiac shunts were assessed on the basis of right ventricle systolic pressure. Data on change in right ventricular pressure as assessed by echocardiogram were available in 27 of the 41 patients without an intracardiac shunt. In the remaining patients, noninvasive testing with echocardiography was not obtained by the primary cardiologist. In the patients with echocardiography follow-up data, right ventricular pressure increased a category in 3 (11%), stayed the same in 14 (52%), and decreased in 10 (37%; P 0.018). Among the 15 patients in this group who underwent cardiac catheterization, there was a statistically significant decrease in the ratio of the right ventricular systolic pressure to aortic systolic pressure from baseline to the follow-up catheterization (median, 0.11; range, 0.42 to 0.13; P 0.047). Patient-level outcomes are provided only in aggregate for all study patients, given that comparative data between the CB and HPB groups are not possible because any given patient could have vessels randomized to both treatment arms. Safety Outcomes Serious adverse events did not occur in a study vessel in either treatment group. There was no significant difference in the rate of somewhat serious or serious events, with events occurring in 3% of vessels randomized to CB and 2% of vessels randomized to HPB (P 0.85; Table 5). There were more events in the CB group than in the HPB group; however, 7 of 13 of these events were pulmonary edema potentially caused by reperfusion after a successful dilation and were classified as not serious in most cases (Table 6). Pulmonary edema was classified as somewhat serious in 2 patients who required prolonged mechanical ventilator support. The other somewhat serious event was a vessel that required stent placement because of flow obstruction after angioplasty. The remaining not serious events were related to

6 Bergersen et al Randomized Pulmonary Artery Cutting Balloon Trial 2393 Table 4. Vessel and Treatment Characteristics Characteristics Cutting Balloon (n 107) High-Pressure Balloon (n 66) P Angioplasty site, n (%) Proximal RPA/LPA 6 (6) 5 (8) 0.63 Lobe 50 (47) 26 (39) Sublobar segment 51 (48) 35 (53) Minimum vessel diameter, initial, mm ( ) ( ) 0.22 Minimum vessel diameter, final, mm ( ) ( ) 0.26 Maximum vessel diameter, initial mm ( ) ( ) 0.81 Maximum vessel diameter, final (n 105, 65), mm ( ) ( ) 0.48 Size of balloon that failed at 8 atm, mm (2.2 12) (2 14) 0.56 Size of study balloon (n 104, 66), mm (2 8) (2.5 14) Waist eliminated, n (%) 101 (97) 19 (29) Treatment crossover, n (%) 1 (1) 26 (54) RPA indicates right pulmonary artery; LPA, left pulmonary artery. balloon malfunction or other device problem: A CB was trapped on a stent and the removal was difficult in 1 case, 3 balloon ruptures were reported, 1 catheter shaft was fractured, and delivery of a CB out of the sheath was difficult in 1 case. The core laboratory assessed the angiographic appearance of all vessels after therapy. There were no significant differences in the angiographic appearance, presence of tears, or aneurysms after therapy (Table V in the online-only Data Supplement). In patients with follow-up angiography, no new aneurysms or vascular occlusions were reported. Among events not related to randomized therapy, 2 events were classified as serious; 1 patient developed hemoptysis and hypotension after low-pressure angioplasty, and another patient required a stent to treat vessel thrombosis in a nonstudy vessel. Other somewhat serious events included pulmonary edema, arrhythmia, heart block, pulse loss, arterial damage, acidosis, and hypotension (Table VI in the onlineonly Data Supplement). Discussion This study is the first randomized, blinded clinical trial evaluating a complex intervention for a congenital heart defect in children in which randomization was done after vessel eligibility was established during the procedure. The results indicate that CB therapy, followed by low-pressure dilation, is more effective than HPB therapy for PA obstructions resistant to initial low-pressure angioplasty. Furthermore, the safety profiles for the 2 techniques appear equivalent although the study was not powered to show a difference. We observed a significantly greater change in minimum lumen diameter for the CB group compared with the HPB (85% versus 52%; P 0.004). Only 29% of the dilations were successful at 8 atm, suggesting that alternative techniques must be available when treating patient with PA stenosis. In addition, 55% of vessels treated with HPB therapy crossed over to CB when permitted; therefore, CB technology has a Figure 4. Percent change in lumen diameter of Cutting Balloon (CB) and high-pressure balloon (HPB). Figure 5. Cross over therapy (dark circle) represents change in lumen diameter for vessels randomized to high-pressure balloon therapy that crossed over to Cutting Balloon treatment; open circle shows the change in lumen diameter for the one vessel which crossed over from Cutting Balloon to high pressure treatment. At each of the intervals, the mean SD is shown above the point estimate and SD summary bar for the CB group.

7 2394 Circulation November 29, 2011 Table 5. Primary and Secondary Safety Outcomes Primary and Secondary Safety Outcomes Cutting Balloon (n 107), n (%) High-Pressure Balloon (n 66), n (%) P Primary safety outcome Any serious event in a study vessel 0 (0) 0 (0) Secondary safety outcomes Any serious event definitely attributable to balloon 0 (0) 0 (0) angioplasty in a study vessel Any somewhat serious or serious adverse event resulting 3 (3) 1 (2) 0.85 from balloon angioplasty in a study vessel Any adverse event resulting from balloon angioplasty in a study vessel* 12 (11) 4 (6) 0.19 *Thirteen events in 12 vessels, including 1 patient with 2 events in 1 vessel (both device malfunction). One adverse event in a vessel randomized to high-pressure balloon therapy occurred after crossover (balloon rupture); here, it is attributed to the high-pressure balloon. Table 6. Types of Adverse Events Related to Randomized Vessel Treatment Events Related to Type of Study Treatment Caused by Cutting Balloon angioplasty in a study vessel Serious, n Somewhat Serious, n Not Serious, n Device malfunction Balloon rupture Balloon caught on stent Fracture of catheter shaft Pulmonary edema Intravascular tear Caused by high-pressure balloon angioplasty in a study vessel Pulmonary edema significant impact on the treatment of PA stenosis in vessels resistant to standard low- and high-pressure angioplasty. This study demonstrates reassuring safety information. First, the angiographic appearance of vessels and the occurrence of tears or aneurysms were similar between the treatment arms. Second, at follow-up, adverse events, angiographic appearance, and maintenance of luminal diameter were not different between the groups. However, there were technical problems in the delivery of the CBs worth consideration, with vigilance needed to avoid blade entrapment on stents or at the end of long sheaths. In this trial, we explored patient-level outcomes, knowing that we would not necessarily be able to attribute the results to a specific therapy because multiple vessels were randomized to both interventions in most patients. Nevertheless, we observed a higher incidence of pulmonary edema in lung segments treated with CB. And although we recorded these as adverse events, previous studies have shown that pulmonary edema is associated with successful dilation procedures and resolves within 72 hours. 20 In terms of overall patient outcomes, there appeared to be a hemodynamic benefit in the population undergoing interventional catheterization procedures involving PA dilations, with some patients experiencing an improvement in oxygen saturation or a decrease in right ventricular pressures. These clinical benefits were difficult to demonstrate as convincingly in prior retrospective reviews. 1,5 7,11 13 CB for Congenital Heart Disease The first experimental use of CB angioplasty in PAs was reported by Magee et al 9 in a piglet model of PA stenosis created by suture plication of the left PA. Thereafter, initial reports of the clinical use of CB to treat congenital or postoperative stenoses in PAs and aortopulmonary collateral arteries emerged ,21 CB was first used in PAs to treat resistant lesions, ie, those unable to be dilated despite high pressure. Indeed, in these early studies, CB was only used after failure of high-pressure angioplasty with pressures of 15 to 22 atm. 11,12 Such lesions were frequently undilatable with the HPBs available then and did not respond well to stenting because the stent could not be completely expanded. In the initial application of CB to treat PA stenotic lesions that were resistant to HPBs, the percent lumen gain was significantly greater, and this was sustained at follow-up angiography. 12,13 In addition, the complication rate compared favorably with standard balloon angioplasty, and all vascular complications proved to be apparent at the index procedure, with no new vascular complications discovered at follow-up. 13,22 The treatment of proximal PA lesions can be limited by the size of available CB technology. 23,24 Currently, the largest available CB is 8 mm in diameter; thus, vessels with a low-pressure balloon waist 7.5 mm in diameter are not amenable to CB therapy. Fortunately, proximal lesions are frequently compliant and amenable to stent placement when recoil characteristics are present. 25 Nevertheless, there would be a role of larger CBs in the treatment of intravascular obstructions in children and adults with congenital heart disease and vascular lesions. Study Design We had to overcome a number of challenges in the design and execution of this study. First, the equipment that had become the standard of practice in the treatment of congenital heart disease, balloons capable of low- and high-pressure angio-

8 Bergersen et al Randomized Pulmonary Artery Cutting Balloon Trial 2395 plasty, had not previously been approved for pulmonary angioplasty. A comprehensive contemporary review of performance was necessary to show the safety and efficacy of this technology before approval was granted for the investigational device exemption. 25 Second, although randomization is generally done at the patient level, vessel-level randomization was necessary to promote enrollment ethically because high-pressure dilation was known to not always be successful. Randomization of vessels required that efficacy and safety be assessed at the vessel level. Although this was easily achievable for efficacy, assessment of safety was more problematic. This was solved by using the DSMB to attribute events to specific aspects of the procedure. Because of this attribution requirement, crossover was not permitted initially. However, after the first DSMB meeting, it was determined that adjudication was not problematic and that attributability could be assigned if crossover was permitted. This determination was imperative at the time because the larger Peripheral CBs became available after the first phase of the study and therefore became available for the off-label use by the operators, potentially threatening ongoing enrollment. Finally, changes in balloon technology over the study period, with the coronary CB evolving to the Flextome and the Peripheral CB changing from 1 cm to 2 cm in length, required modifications of sample size and number of vessels assessed to allow adequate evaluation of the new technology. Labeling Indications in Children Children with congenital heart disease requiring interventional tools represent a small minority of the market share for companies producing intravascular treatment technologies. Thus, in most cases, interventional cardiologists use equipment developed for adult indications. Most of the balloons and stents used to treat congenital lesions are therefore being used for off-label indications. 26 It is imperative that interventional pediatric cardiologists partner with the FDA and industry to develop and evaluate the efficacy and safety of equipment for pediatric indications. This prospective randomized study is an example of how we can plan, execute, and complete an evaluation of equipment for a potential labeling indication in pediatric patients with congenital heart disease. To the best of our knowledge, this is only the second study in the field of pediatric and congenital heart disease in which therapy was randomized and the first one at multiple sites for a complex intervention in which, after randomized therapy, outcome assessment was performed by researchers blinded to treatment. 27 Limitations The 8-atm cutoff was chosen arbitrarily on the basis of the judgment of the investigators. More studies are needed to determine whether 8 atm is indeed the correct cutoff pressure at which a standard balloon angioplasty should be aborted in favor of a CB angioplasty. Incomplete follow-up on patientlevel outcomes, specifically changes in right ventricular pressure, multiple interventions in the same patient, and interval surgical interventions, limits our ability to make definitive conclusions regarding the benefits of angioplasty at follow-up. Furthermore, longer-term studies are required to confirm whether the luminal gains in both techniques are sustained over time. Finally, ongoing surveillance of the safety of the technology is required. Conclusion This multicenter randomized trial demonstrates that a strategy of initial CB with additional low-pressure balloon angioplasty is superior to a strategy of high-pressure angioplasty alone in the treatment of PA stenoses resistant to low-pressure dilation with equivalent safety profiles. Acknowledgments Richard E. Ringel, MD (chairman), Tom Doyle, MD, and Michael Gewitz, MD, served on the DSMB. Barry Keane, MD, ran the core laboratory. Robyn Morse, Lily Maltz, Dawn England, Denise Norton, Heidi Moses, Eunice Newbert, Linda Drake, Kathleen Gilmartin, Delores Standford, Kevin Stiegler Summer Roberts, Ruby Whalen, Pantelis Konstantinopoulos, and Elizabeth Tong were research study coordinators and study monitors. Sources of Funding This work was supported by Boston Scientific Corp. None. Disclosures References 1. Kreutzer J, Perry S, Jonas R, Mayer J, Castaneda A, Lock J. Tetralogy of Fallot with diminutive pulmonary arteries: preoperative pulmonary valve dilation and transcatheter rehabilitation of pulmonary arteries. J Am Coll Cardiol. 1996;27: Rocchini A, Kveselis D, MacDonald D, Crowley D, Snider R, Rosenthal A. Use of balloon angioplasty to treat peripheral pulmonary stenosis. Am J Cardiol. 1984;54: Zeevi B, Berant M, Blieden L. Midterm clinical impact versus procedural success of balloon angioplasty for pulmonary artery stenosis. Pediatr Cardiol. 1997;18: Bergersen L, Lock J. Pulmonary artery stenosis, angioplasty, stenting, or cutting balloon: what is the current treatment of first choice? A review. Cardiol Young. 2006;16: Hosking M, Thomaidis C, Hamilton R, Burrows P, Freedom R, Benson L. Clinical impact of balloon angioplasty for branch pulmonary arterial stenosis. Am J Cardiol. 1992;69: Rothman A, Perry S, Keane J, Lock J. Early results and follow-up of balloon angioplasty for branch pulmonary artery stenoses. J Am Coll Cardiol. 1990;15: Gentles T, Lock J, Perry S. High pressure balloon angioplasty for branch pulmonary artery stenosis: early experience. J Am Coll Cardiol. 1993; 22: Ettinger L, Hijazi Z, Geggel R, Supran S, Cao Q, Schmid C. Peripheral pulmonary artery stenosis: acute and mid-term results of high pressure angioplasty. J Interv Cardiol. 1998;11: Magee A, Wax D, Siki Y, Rebekya I, Benson L. Experimental branch pulmonary artery stenosis angioplasty using a novel cutting balloon. Can J Cardiol. 1998;14: Schneider M, Zartner P, Magee A. Images in cardiology: cutting Balloon for treatment of severe peripheral pulmonary stenosis in a child. Heart. 1999;82: Rhodes J, Lane G, Mesia C, Moore J, Nasman C, Cowan D, Latson L. Cutting Balloon angioplasty for children with small-vessel pulmonary artery stenoses. Catheter Cardiovasc Interv. 2002;55: Bergersen L, Perry S, Lock J. Effect of cutting balloon angioplasty on resistant pulmonary artery stenosis. Am J Cardiol. 2003;91: Bergersen L, Marshall A, Lang P, Perry S, Lock J. Follow-up results of cutting balloon angioplasty used to relieve stenoses in small pulmonary arteries. Cardiol Young. 2005;15: Barath P, Fishbein M, Vari S, Forrester J. Cutting balloon: a novel approach to percutaneous angioplasty. Am J Cardiol. 1991;68:

9 2396 Circulation November 29, Mauri L, Bonan R, Wiener B, Legrand V, Bassand J, Popma J, Niemyski P, Prpic R, Ho K, Chauhan M, Cutlip D, Bertrand O, Kuntz R. Cutting balloon angioplasty for the prevention of restenosis: results of the Cutting Balloon Global Randomized Trial. Am J Cardiol. 2002;90: Lock J, Niemi T, Einzig S, Amplatz K, Burk B, Bass J. Transvenous angioplasty of experimental branch pulmonary artery stenosis in newborn lambs. Circulation. 1981;64: Lock J, Niemi T, Burk B, Einzig S, Cataneda-Zuniga W. Transcutaneous angioplasty of experimental aortic coarctation. Circulation. 1982;66: Williams R. A note on robust variance estimation for cluster-correlated data. Biometrics. 2000;56: Froot K. Consistent covariance matrix estimation with cross-sectional dependence and heteroskedasticity in financial data. J Financial Quant Anal. 1989;24: Arnold L, Keane J, Kan J, Fellows K, Lock J. Transient unilateral pulmonary edema after successful balloon dilation of peripheral pulmonary artery stenosis. Am J Cardiol. 1988;62: Mertens L, Dens J, Gewillig M. Use of a cutting balloon catheter to dilate resistant stenoses in major aortic-to-pulmonary collateral arteries. Cardiol Young. 2001;11: Butera G, Antonio L, Massimo C, Mario C. Expanding indications for the treatment of pulmonary artery stenosis in children using cutting balloon angioplasty. Catheter Cardiovasc Interv. 2006;67: Sugiyama H, Veldtman G, Norgard G, Lee K, Chaturvedi R, Benson L. Bladed balloon angioplasty for peripheral pulmonary artery stenosis. Catheter Cardiovasc Interv. 2004;62: Sudu K, Matsumura M, Hayashi H, Nishimura K. Comparison of efficacy of medium-sized cutting balloons versus standard balloons for dilation of peripheral pulmonary artery stenosis. Am J Cardiol. 2006;97: Bergersen L, Gauvreau K, Lock J, Jenkins K. Recent results of pulmonary arterial angioplasty: the differences between proximal and distal lesions. Cardiol Young. 2005;15: Sutherell J, Hirsch R, Beekman R. Pediatric interventional cardiology in the United States is dependent on the off-label use of medical devices. Congenit Heart Dis. 2010;5: Justino H, Justo R, Ovaert C, Magee A, Lee K, Hashmi A, Nykanen D, McCrindle B, Freedom R, Benson L. Comparison of two transcatheter closure methods of persistently patent arterial duct. Am J Cardiol. 2001;87: CLINICAL PERSPECTIVE Children born with congenital heart disease not uncommonly require cardiac catheterization to treat congenital and acquired malformations such as pulmonary artery stenosis, replacing or complementing surgical techniques. In this study of Cutting Balloons compared with high-pressure balloon angioplasty for resistant pulmonary artery stenosis, we have shown superior efficacy with the Cutting Balloon technology and an equivalent safety profile. This finding has important clinical application for a population with previously untreatable disease. Although few studies have evaluated the performance of devices developed for adults but used in children, in this study, we have demonstrated that some of the unique study design and execution difficulties met in the pediatric population can be overcome. We hope that future studies will continue to rigorously evaluate the performance of devices used to treat children with rare diseases relative to more common adult indications.