Author + information
- Received June 25, 2013
- Revision received December 28, 2013
- Accepted January 6, 2014
- Published online April 1, 2014.
- Jian Yang, MD, PhD∗∗ (, )
- Lifang Yang, MD∗,
- Shiqiang Yu, MD, PhD∗,
- Jincheng Liu, MD, PhD∗,
- Jian Zuo, MD∗,
- Wensheng Chen, MD, PhD∗,
- Weixun Duan, MD∗,
- Qijun Zheng, MD∗,
- Xuezeng Xu, MD∗,
- Jun Li, MD∗,
- Jun Zhang, MD, PhD∗,
- Jian Xu, MD∗,
- Lijun Sun, MD, PhD∗,
- Xiuling Yang, MD∗,
- Lize Xiong, MD, PhD∗,
- Dinghua Yi, MD, PhD∗∗,
- Lei Wang, MD†,
- Qingchun Liu, MD‡,
- Shuping Ge, MD∗,§∗ and
- Jun Ren, MD, PhD∗,‖
- ∗Department of Cardiovascular Surgery, Xijing Hospital, Fourth Military Medical University, Xi'an, China
- †Department of Cardiology, Xi'an Children Hospital, Xi'an, China
- ‡Department of Cardiovascular Surgery, Hanzhong Central Hospital, Hanzhong, China
- §Heart Center at St. Christopher's Hospital for Children and Drexel University College of Medicine, Philadelphia, Pennsylvania
- ‖College of Health Sciences, University of Wyoming, Laramie, Wyoming
- ↵∗Reprint requests and correspondence:
Drs. Jian Yang, Dinghua Yi, or Shuping Ge, Xijing Cardiovascular Hospital, Xi'an, China, 710032.
Objectives The objective of this study was to evaluate the safety and efficacy of the surgical versus transcatheter approach to correct perimembranous ventricular septal defects (pmVSDs) in a prospective, randomized, controlled clinical trial.
Background pmVSD is a common congenital heart disease in children. Surgical closure of pmVSD is a well-established therapy but requires open-heart surgery with cardiopulmonary bypass. Although the transcatheter approach is associated with significant incidence of complete atrioventricular block, it may provide a less invasive alternative. Critical comparison of the safety and efficacy of the 2 interventions necessitates a prospective, randomized, controlled trial.
Methods Between January 2009 and July 2010, 229 children with pmVSD were randomly assigned to surgical or transcatheter intervention. Clinical, laboratory, procedural, and follow-up data over a 2-year period were compared.
Results Neither group had mortality or major complications. However, statistical analysis of the 2 groups demonstrated significant differences (p < 0.001) in minor adverse events (32 vs. 7), quantity of blood transfused, duration of the procedure, median hospital stay, median intensive care unit stay, median hospitalization cost, and median blood loss. During a median follow-up of 2 years, the left ventricular end-diastolic dimension of both groups returned to normal and there was no difference in closure rate, adverse events, and complications between groups.
Conclusions Transcatheter device closure and surgical repair are effective interventions with excellent midterm results for treating pmVSD in children. Transcatheter device closure has a lower incidence of myocardial injury, less blood transfused, faster recovery, shorter hospital stay, and lower medical expenses. (Transcatheter Closure Versus Surgery of Perimembranous Ventricular Septal Defects; NCT00890799)
Ventricular septal defect (VSD) is the most common type of congenital heart disease, in which perimembranous ventricular septal defect (pmVSD) is 1 of the most common subtypes (1–3). Treatment of pmVSD has improved dramatically over the last 50 years (4–6). Although open-heart surgery is considered a standard treatment for pmVSD, catheter-based intervention is a promising alternative (7–9). Transcatheter device closure of pmVSD is not currently approved in the United States because of an unacceptably high rate of post-procedural and late-onset heart block (2.9% to 5.7%) (5,7,10). Nevertheless, this procedure is widely used in developing countries such as China and India, with a low mortality and morbidity rate using modified nitinol shape-memory alloy devices (11,12). Nonetheless, data comparing the clinical outcomes of the 2 procedures are minimal (13). In 2009, we performed retrospective studies to examine the early results and complications of percutaneous compared with surgical closure of VSDs (14–16). On the basis of these preliminary findings, we designed a randomized, controlled trial to prospectively compare the safety and efficacy of the 2 interventions for pmVSD in children.
Between January 2009 and July 2010, 465 children age 3 to 12 years with pmVSD from 3 major medical centers in northwest China (Xijing Hospital, Xi'an; Xi'an Children's Hospital, Xi'an; and Hanzhong Central Hospital, Hanzhong) were enrolled in the study. After clinical and transthoracic echocardiographic (TTE) assessment for eligibility, 236 patients were excluded from the study and the other 229 were randomly allocated to either the surgical or transcatheter group (Fig. 1). See details in the Online Appendix.
A comprehensive TTE study, which included M-mode, 2-dimensional, continuous-wave, pulsed-wave, and color-flow Doppler echocardiography, was performed prior to operation or intervention and at each follow-up visit (Fig. 2) (see Online Videos 1, 2, 3, and 4 of echocardiography screening and follow-up of transcatheter pmVSD closure). See details in the Online Appendix.
Each patient underwent catheterization before surgery or transcatheter intervention. Catheterization was done in the catheterization laboratory with the patients under conscious sedation and local anesthesia. Hemodynamic parameters, including pulmonary arterial, aortic, atrial, and ventricular pressures, were recorded, and oximetric values were measured in the cardiac chambers and vessels, based on which pulmonary-to-systemic flow ratio (Qp/Qs) was calculated.
Transcatheter device implantation
Transcatheter intervention was described in previous reports (15,16) (see Online Videos 5, 6, 7, 8, 9, 10, 11, and 12 of transcatheter pmVSD closure of different kinds of pmVSD). See details in the Online Appendix.
The surgical closure was performed with the patient under general anesthesia the second day after catheterization. The chest was opened through a standard median sternotomy. Depending on the size of the pmVSD, direct-suture, Dacron, or pericardium patch was used to repair the defect. All patients were transferred to the intensive care unit for further treatment. See details in the Online Appendix.
All patients were followed up for 2 years. Each patient underwent serial follow-up at 3 days, 3 months, 6 months, 1 year, and 2 years following intervention or surgery. See details in the Online Appendix.
The biochemical assays included measurement of levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), blood urea nitrogen (BUN), creatinine, cardiac troponin I (cTnI), and creatine kinase-MB (CK-MB) at the induction of anesthesia and 1, 3, 6, 12, 24, and 72 h after pmVSD occluder deployment/aorta cross-clamp removal. See details in the Online Appendix.
See details in the Online Appendix.
The sample size needed for this study was calculated, assuming the largest difference between treatment groups. With a significance level of 0.05 and a power of 80%, we calculated the needed number of events to be 35 in each group. We also determined that a minimum of 43 patients in each group was necessary to conduct the present study.
The patients were followed up until June 2012, with a median follow-up of 2 years (2 to 40 months). During this period, 29 patients withdrew from the study. A total of 101 patients in the transcatheter group and 99 patients in the surgery group completed follow-up examinations. The trial profile is shown in Figure 1. The 2 groups were well matched for baseline characteristics (Table 1).
The 2 groups had comparable baseline characteristics (Table 1) with regard to age, sex, weight, clinical manifestations, echocardiographic data, and invasive data. Twenty-five patients in the surgery group and 28 in the transcatheter group had Qp/Qs values below 1.5:1. Among them, 17 patients in the surgery group and 21 in the transcatheter group had no left ventricular enlargement (Z-score <+2). These patients were enrolled mainly because they had symptoms such as refractory pneumonia, congestive heart failure, delayed growth, exercise intolerance, frequent colds, or unyielding parental requests. The average age at closure was 5.8 ± 2.4 years and 5.5 ± 2.6 years, respectively (p = 0.398). Most patients (56.6% in the surgery group and 65.3% in the transcatheter group; p = 0.531) were treated because of hemodynamic changes. The mean defect size was 5.9 ± 5.3 mm and 5.2 ± 6.1 mm in the surgical and transcatheter groups, respectively (p = 0.866).
Transcatheter versus surgical treatment
The operative characteristics and outcomes in patients according to treatment are summarized in Table 1. In both groups, no mortality, stroke, or neurological deficit was observed. According to ventriculographic examination of the left ventricle, pmVSD was classified as tubular (n = 24), window-like (n = 5), aneurysmal (n = 33), or infundibular (n = 37) in the surgery group, a distribution that was comparable to that in the transcatheter group. The average procedural time in the surgery group was 180.5 ± 66.1 min, with an average cardiac arrest time of 24.2 ± 15.4 min and an average cardiopulmonary bypass time of 55.1 ± 32.0 min. In the group receiving surgery, 75 patients were repaired with a Dacron or pericardium patch and 24 patients with primary closure. One major adverse event occurred in the surgery group (i.e., bleeding occurred following operation in a 7-year-old girl that required thoracic re-exploration). Several minor adverse events were reported. Blood loss that required blood transfusion occurred in 23 patients. Prior to discharge, 2 patients experienced a residual shunt through the rim of the VSD patch, as demonstrated by color-flow Doppler mapping. Furthermore, 2 patients developed new or increased tricuspid valvular regurgitation <2 grades and were followed without further intervention. Other minor adverse events in the surgery group included 3 right bundle-branch blocks and 1 partial left bundle-branch block (Table 2). Two patients were found to display a prolonged PR interval (>0.02 s) after surgical closure and before discharge.
In the transcatheter group, device deployment was successful in all patients. The mean defect size measured by TTE was 5.2 ± 6.1 mm, and the mean size of the occluder was 7.4 mm. The Shanghai pmVSD device was used for all patients. The mean fluoroscopy time was 8.2 ± 3.4 min, and the total procedural time was 38.2 ± 24.6 min. A mean of 42.7 ± 6.6 ml of contrast was used. Based on ventriculographic results (12,16), pmVSD was classified as tubular (n = 27), window-like (n = 2), aneurysmal (n = 26), or infundibular (n = 46). The transcatheter procedure failed with the first occluder in 3 patients, but a second device was deployed successfully in all of them. No major adverse event was observed in the transcatheter group, and none of the patients required a transfusion as a result of blood loss. There were 7 minor adverse events, including 1 hematoma of the groin, 1 right bundle-branch block, 2 junctional rhythms, 1 partial left bundle-branch block, 1 residual shunt before discharge, and 1 new case of mild tricuspid valve regurgitation. Five patients had a prolonged PR interval (>0.02 s) following transcatheter closure before discharge, which was not significantly different than that of the surgery group (p = 0.260). All of the patients were asymptomatic and did not require further treatment.
Overall, there were 32 minor adverse events in the surgery group (32.3%) and 7 in the transcatheter group (6.9%). The Fisher exact test showed a significant difference in minor adverse events between the 2 groups (p < 0.001).
The levels of ALT in the surgery group began to increase 1 h after removal of the aortic cross-clamp and peaked at 24 h. ALT levels in the transcatheter group remained unchanged after the procedure. A comparison of ALT levels between the 2 groups displayed a significant difference at 1 and 24 h following pmVSD closure (p < 0.01). The trends of changes in AST levels were the same as those for ALT. There were also significant differences in the AST levels at 1 and 24 h following pmVSD repair (p < 0.01). BUN levels in the surgery group rose 1 h after removal of the aortic cross-clamp. The BUN level in the transcatheter group remained constant. There was no significant difference in BUN level alteration between the 2 groups at any time after pmVSD closure. Creatinine levels showed a more significant increase in the surgery group than in the transcatheter group during the 24 h after the procedure (p < 0.01) (Fig. 4).
The concentrations of CK-MB and cTnI before intervention were normal in both groups. In the surgery group, CK-MB levels peaked after 24 h, whereas cTnI levels began to rise after aortic clamping and peaked 12 h later. The levels of cTnI did not return to normal even 72 h after declamping. In the transcatheter group, both the CK-MB and cTnI levels remained unchanged after the deployment of the device. The 0- to 72-h area under the curve of CK-MB and cTnI release was significantly higher in the surgery group than in the transcatheter group (3,415.8 vs. 425.1 and 597.6 vs. 4.7, respectively; p < 0.001) (Fig. 5). No hemoglobinuria was detected in either group.
TTE examination revealed that 97 of the 99 pmVSDs in the surgery group were completely closed 3 days following the operation and before discharge (Fig. 6). Two patients had residual shunts at the rim of the patch, measuring 1 and 1.5 mm by color-flow Doppler mapping. Two patients in the surgery group had mild tricuspid insufficiencies.
In the transcatheter group, a shunt that measured <2 ml by ventriculography through the occluder was noted in 28 of the 101 patients (27.7%). One patient got a residual shunt before discharge, and all shunts diminished at the 3-month follow-up. One patient developed new tricuspid insufficiency (3-mm vena contracta by color-flow Doppler mapping). The patient was asymptomatic and was followed closely.
Two-year follow-up and clinical outcomes
All of the 101 patients in the transcatheter group and 99 patients in the surgery group were followed for a median duration of 2 years. No mortality, neurological deficits, thromboembolism, or endocarditis were reported in either group. All patients displayed New York Heart Association functional class I to II with a normal sinus rhythm. Echocardiographically, the Z-scores of the left ventricular end-diastolic dimension decreased from 1.8 ± 1.5 to 0.8 ± 0.6 in the surgery group (p < 0.001) and from 1.7 ± 1.4 to 0.7 ± 0.8 (p < 0.001) in the transcatheter group at the 2-year follow-up. Comparison of Z-scores showed no statistical significance between the catheter and surgical groups either before (p = 0.626) or after the procedure (p = 0.319) (Fig. 7).
In the surgery group, 1 patient with a residual shunt experienced spontaneous pmVSD closure at the 2-year follow-up. Another patient had a small residual VSD shunt (1.5 mm) but was asymptomatic. No further intervention was indicated. The 2 patients with tricuspid valve insufficiency still had mild regurgitation but did not require intervention. Six patients had arrhythmias at the 2-year follow-up, including 4 with right bundle-branch block and 2 with partial left bundle-branch block. In the transcatheter group, all pmVSDs were closed at the 2-year follow-up, which was confirmed echocardiographically by the stable position of the device and the lack of a residual VSD shunt. At the 1-year follow-up, the patient who had a residual shunt prior to discharge exhibited no left-to-right shunt across the interventricular septum. Overall, 5 patients had arrhythmias, including 1 with right bundle-branch block, 3 with junctional rhythms (1 additional patient developed junctional rhythm at the 3-month follow-up), and 1 complete left bundle-branch block. The patient with mild post-interventional tricuspid regurgitation was asymptomatic and required no further treatment. During a median follow-up of 2 years, no differences were observed in the closure rate, adverse events, or complications between the 2 groups.
In developing countries, many children with congenital heart disease are waiting to be treated. Among congenital heart diseases, pmVSD is perhaps the most common, accounting for almost one-fifth of all defects. There is an immediate need for an economical, effective, less-invasive technique. In 2002, Hijazi et al. (17) first closed pmVSDs using an Amplatzer membranous VSD occluder. Whether this less-invasive approach may replace open-heart surgery as the treatment of choice is debatable. Over the past decade, the Amplatzer occluder has been proven to be superior to previous devices, such as the Gianturco coils (Cook Inc., Bloomington, Indiana), the Rashkind device (USCI Angiographics, Billerica, Massachusetts), and Sideris immediate release patch (Custom Medical Devices, Athens, Greece) (18–20), in closing congenital heart defects. However, the Amplatzer pmVSD occluder has been associated with a relatively high risk of complete atrioventricular block; consequently, the sponsor voluntarily stopped the U.S. clinical study (5,10,21,22). In recent years, similar devices from China have been adopted in developing countries with good results (23,24). Interest is growing as to whether this new technique can replace traditional open-heart surgery as the “gold standard” for treatment of pmVSD. Therefore, we designed this randomized, controlled trial as a prospective comparison of the 2 treatments for children with pmVSD.
Both transcatheter device closure and surgical repair have achieved excellent midterm results for children with pmVSD. Safety and efficacy are considered in terms of success of the procedure, complications, cost, hospital stay, and time to return to normal activities. There was no mortality or complete atrioventricular block in either group. Minor complications, such as residual shunt, arrhythmia, and valve insufficiency, were minimal in both groups. Our results showed that the Z-score of the left ventricular end-diastolic dimension of both groups returned to normal at the 2-year follow-up, and compared with open-heart surgical repair, transcatheter device closure possesses several advantages, including no requirement for intensive care unit stay, shorter procedural time, less blood loss, lower medical expenses, and a much faster recovery time (Table 1).
We found that tubular, window, and infundibular types of pmVSD were technically easier to repair using transcatheter intervention and had few complications. Several technical issues were crucial to the overall success of transcatheter pmVSD closure, especially for the aneurysmal type. First, crossing the pmVSD was the most crucial step toward successful transcatheter closure and was especially challenging for the aneurysmal type of pmVSD. In such circumstances, selection of different types of catheters (e.g., right Judkins, 3DRC, or partly-cut pigtail catheter) and different wires may be helpful. Second, for the aneurysmal type of pmVSD, placing the occluder inside the aneurysmal sac at all times was an important step for minimizing heart rhythm disturbances and complications (Online Video 10). In our series, occluders were placed away from the ventricular septum and inside the aneurysmal tissue in 23 of 26 patients with aneurysmal pmVSD. Meanwhile, given that aneurysmal tissues were often adjacent to or even part of the tricuspid valve, caution should be taken when the operator passes the wire and catheter through the tricuspid valve to establish the arteriovenous circuit. Furthermore, detection of new tricuspid regurgitation using TTE was routinely carried out before the occluder was released. Third, selection of the size of the device is important for a successful transcatheter pmVSD intervention. We recommended previously that the relationship between occluder size (y) and pmVSD size (x) as measured by ventriculography was y = 0.966x + 2.073, R² = 0.967 (17). An oversized device (150% over the calculated size) should be avoided to prevent possible complete atrioventricular block and valve impingement. Finally, strict adherence to the inclusion and exclusion criteria for transcatheter pmVSD closure is crucial to the overall success of the intervention. We excluded patients with low body weight (<10 kg) and those younger than age 3 years to ensure higher success rates and fewer complications.
We chose cTnI to evaluate myocardial injury. Although cTnI has been used routinely to assess myocardial injury in open-heart surgery, little information is available regarding the changes in cTnI levels in transcatheter pmVSD closure. The results of our study have contributed to the current literature and indicate the remarkable advantage of transcatheter intervention over surgical repair in preventing myocardial injury. By obviating the need for the ischemia-reperfusion process and diminishing the side effects of cardiopulmonary bypass, the transcatheter closure technique remarkably decreased myocardial injury associated with open surgical repair of pmVSD.
Although our 2-year follow-up study showed excellent midterm outcomes in treating pmVSD in children, indications for intervention in children are based on clinical, echocardiographic, and catheterization data. We excluded patients who presented with small defects without hemodynamic changes or symptoms due to the possibility of spontaneous closure and good life-long survival. We also excluded patients whose anatomy was not suitable for transcatheter closure. These patients were referred directly for surgical repair. In addition, a Qp/Qs ratio of 1.5:1 to 2.5:1, or evidence of increased left ventricular overload, is an indication for closure. The benefits of avoiding cardiopulmonary bypass, reducing the psychological impact, and offering the less painful and less uncomfortable percutaneous intervention are extremely attractive for both patients and physicians (3,7,25). Recently, clinical trials with the new Amplatzer membranous VSD occluder 2 and Amplatzer duct occluder II device (AGA Medical Corporation, Plymouth, Minnesota) have demonstrated promising results in initial studies (26–28). With more refinement to the devices and new clinical trials for patients with pmVSD, the treatment of choice is likely to be updated frequently.
First, as in our previous study and in accordance with the consensus of Chinese experts (29), we followed strict inclusion/exclusion criteria. Only children >3 years of age were enrolled. Also, because of the design of this study, we only evaluated the results of children <12 years of age; adult patients were excluded. Second, according to the design of the study, patients with small defects without hemodynamic changes or symptoms and patients whose anatomy was not suitable for transcatheter closure were excluded from the current study (i.e., the results of the current study cannot be applied to all kinds of patients with pmVSD). Third, the patients were enrolled from 3 large-volume centers in northwest China and may not be representative with respect to ethnicity and other factors. Last, the number of patients was relatively small, and the 2-year follow-up period was short. Additional data are required to evaluate the 2 procedures longitudinally.
We demonstrated that both transcatheter device closure and surgical repair are effective treatments, with excellent midterm outcomes, for pmVSD in children who met our inclusion criteria. In our study, compared with surgical repair, transcatheter device closure had fewer myocardial injuries, fewer transfusions, shorter hospital stays, reduced medical costs, and faster recovery times. Therefore, transcatheter device closure is the treatment of choice for pmVSD in children in this cohort.
The authors would like to dedicate this paper to Professor Weiyong Liu (Xijing Hospital), who participated in the first case of open-heart surgery (pmVSD closure) in China in 1958, for his 85th birthday on November 23, 2013. The authors would also like to thank Dr. Zhaoming He (Department of Mechanical Engineering, Texas Tech University, Lubbock, Texas) and Dr. Heywood R. Sawyer (University of Wyoming, Laramie, Wyoming) for their help with the preparation of this manuscript and Dr. Yi Wan (Department of Health Statistics, Fourth Military Medical University, Xi'an, China) for the help with statistical analysis.
This trial was supported by funding from the National Natural Science Foundation of China (grants 30900462 and 81170213). All authors have reported that they have no relationships relevant to the contents of this paper to disclose. The first 5 authors contributed equally to this paper.
- Abbreviations and Acronyms
- alanine aminotransferase
- aspartate aminotransferase
- blood urea nitrogen
- creatine kinase-MB
- cardiac troponin I
- perimembranous ventricular septal defect
- transthoracic echocardiography/echocardiographic
- ventricular septal defect
- Received June 25, 2013.
- Revision received December 28, 2013.
- Accepted January 6, 2014.
- 2014 American College of Cardiology Foundation
- Lock J.E.,
- Block P.C.,
- McKay R.G.,
- Baim D.S.,
- Keane J.F.
- Hoffman J.I.E.,
- Kaplan S.
- Sousa J.E.,
- Costa M.A.,
- Tuzcu E.M.,
- Yadav J.S.,
- Ellis S.
- Fu Y.-C.,
- Bass J.,
- Amin Z.,
- et al.
- Butera G.,
- Carminati M.,
- Chessa M.,
- et al.
- Knauth A.L.,
- Lock J.E.,
- Perry S.B.,
- et al.
- Tucker E.M.,
- Pyles L.A.,
- Bass J.L.,
- Moller J.H.
- Holzer R.,
- Balzer D.,
- Cao Q.L.,
- Lock K.,
- Hijazi Z.M.
- Koneti N.R.,
- Sreeram N.,
- Penumatsa R.R.,
- Arramraj S.K.,
- Karunakar V.,
- Trieschmann U.
- Xian-yang Z.