| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;75:412-421
© 2003 The Society of Thoracic Surgeons
a Department of Paediatric Cardiology, Royal Brompton and Harefield NHS trust, NHLI, Imperial College, Salmia, Kuwait
b Department of Occupational and Environmental Medicine, Royal Brompton and Harefield NHS trust, NHLI, Imperial College, Salmia, Kuwait
c Department of Cardiac Surgery, Royal Brompton and Harefield NHS Trust, NHLI, Imperial College, London, United Kingdom
Accepted for publication July 8, 2002.
* Address reprint requests to Dr Al-Hay, Chest Hospital, P.O. Box 4278, 22043 Salmia, Kuwait.
e-mail: amira_alhay{at}hotmail.com
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
METHODS: A total of 147 consecutive children underwent repair between January 1986 and December 1998. Of those, 106 had Down syndrome, 37% had normal chromosomes, and 4 had other syndromes; 108 underwent primary repair, 19 had prior pulmonary artery banding, and 20 had additional tetralogy of Fallot. The median weight at primary repair was 4.5 kg. A two-patch technique was used in 88%.
RESULTS: The 30-day mortality was 15% (70% confidence interval [CI] 12% to 19%). A double orifice atrioventricular valve was found to be a significant risk factor (p = 0.002), with 6 of 11 patients dying. If double orifice atrioventricular valve patients are excluded, the mortality rate falls to 12% (70% CI 9% to 15%). No difference in mortality was found between Down syndrome and chromosomally normal children but the latter more commonly required reoperation. Chromosomally normal children frequently have a dysplastic common atrioventricular valve (24% versus 3% in Down children, p < 0.001). In a multivariate Cox model including both variables, the presence of a dysplastic atrioventricular valve was a significant risk factor for reoperation. After controlling for the presence of a dysplastic atrioventricular valve, Down syndrome retained a significant protective effect but the upper limit of the confidence interval was close to 1.
CONCLUSIONS: The presence of a double orifice atrioventricular valve emerged as an unforeseen risk factor for death.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
Of the 147 children, 106 (72%) had trisomy 21 and 41 had normal chromosomes, 2 with Noonan and 2 with Smith-Lemli-Opitz syndrome. One hundred twenty-seven (86%) had unrestricted pulmonary blood flow, 83% of whom presented with heart failure or failure to thrive, 2% with cyanosis, and 1% with obstructive apnea; 14% were asymptomatic. Twenty patients (14%) had additional tetralogy of Fallot as evidenced by anterior deviation of the outlet septum resulting in subvalvar pulmonary stenosis. Of these 20 patients, 11 (55%) required early palliative systemic pulmonary shunts (7 unilateral and 3 bilateral modified Blalock-Taussig shunts and 1 Waterston shunt). Some of the children with Down syndrome were asymptomatic but were referred for echocardiographic screening because of the known high association of congenital heart disease. In all children, the diagnosis of a complete AVSD and additional anomalies was confirmed by cross-sectional echocardiography. These anomalies were as follows: ostium secundum atrial septal defect in 41 (28%), persistent arterial duct in 32 (22%), multiple ventricular septal defects in 7 (5%), left superior caval vein to coronary sinus in 4 (3%), relative ventricular hypoplasia (although still considered suitable for a biventricular repair) in 4 (3%; 2 left, 2 right), double outlet right ventricle in 3 (2%), common atrium in 2 (1%), coarctation of the aorta in 1, and subvalvar aortic stenosis in 1. The left atrioventricular valve was dysplastic in 13 (9%) and relatively hypoplastic in 8 (5%). Eleven patients had a double orifice atrioventricular valve, 9 left and 2 right. All these anomalies were diagnosed by echocardiography and confirmed during surgery except for 1 patient in whom the double orifice left atrioventricular valve was first identified at postmortem.
Of those with unrestricted pulmonary blood flow, 108 (85%) underwent primary repair; 29 pulmonary artery banding; 5 arterial duct ligation; and 1 repair of coarctation of the aorta. The reasons for pulmonary artery banding were heart failure in infants younger than 3 months of age (8), hypoplastic left atrioventricular valve (4), multiple ventricular septal defects (4), double outlet right ventricle (2), and miscellaneous (5). Nineteen banded patients have subsequently been repaired. The remaining 10 who were banded are all alive but have not yet undergone repair. They are not considered further in this paper. Thus there are three groups who underwent complete AVSD repair: 108 (73%) with primary repair, 19 (13%) who were previously banded, and 20 (14%) with additional tetralogy of Fallot. Demographic data for each group are presented in Table 1.
|
Operative technique
Repair was carried out using a standard cardiopulmonary bypass technique with moderate or profound hypothermia (mean 22.3°C, range 15°C to 32°C). Deep hypothermic circulatory arrest was used in 28 patients who weighed less than 3 kg or were younger than 3 months. All had a single period of aortic cross-clamping and cold cardioplegia solution, of which 81% was crystalloid and 19% blood cardioplegia. The median cardiopulmonary bypass time was 82 minutes (range 30 to 150). A two-patch technique was used in 129 patients (Dacron [C.R. Bard, Haverhill, PA] for interventricular and autologous pericardium for interatrial communication in 109, and two autologous pericardial patches in 20). A single patch of autologous pericardium was used in 8. Ten patients had direct closure of the interatrial communication and a patch for the interventricular communication (6 Dacron and 4 autologous pericardium).
After closure of the interventricular communication, the left atrioventricular valve was tested by injecting saline rapidly into the left ventricle after which the middle point of coaption of the bridging leaflets was marked. If the valve were shown to be competent, no stitches were placed. Sutures were placed in the septal commissure (line of apposition between left superior and inferior bridging leaflets) in 125 patients (85%) using buttressed or pericardial pledget stitches whereas the left atrioventricular valve was left as trileaflet in the remainder. In the 11 patients identified preoperatively as having a double orifice AV valve particular care was taken not to exacerbate or produce stenosis. Of these, 6 patients had two stitches, 3 had one stitch, and 2 had no stitches. Despite having a double orifice none of these patients were found to have reduplication of the subvalvar apparatus. The interatrial communication was closed leaving the coronary sinus draining to the left of the atrial patch in 120 cases (82%). When prior pulmonary artery banding had been performed, removal of the band and reconstruction of the pulmonary artery were undertaken after the intracardiac repair. In patients with additional tetralogy of Fallot the right ventricular outflow tract obstruction was relieved by excision of trabecular muscle with a transannular patch in 85% and a right ventricle to pulmonary artery homograft in 10%.
Statistical methods
Hospital death rates for various operation types together with 70% confidence intervals (70% CI) were calculated using the technique described in Kirklin and Barratt-Boyes [4]. Univariate analyses were then used to assess whether there was a relationship between outcome (hospital death or reoperation) and possible predictors. The Wilcoxon rank sums (two-sample) test was used for nonnormal data, Fisher’s exact test for categorical data, and the 
2 test for trend for ordinal data. Univariate hazard ratios and their 95% confidence intervals (95% CI) were estimated for variables found to have a statistically significant or borderline significant (p < 0.1) relationship with surgical outcome. Possible predictors were categorized as demographic, cardiac, and surgical. Multivariate Cox regression analyses were performed to assess the impact of selected variables on events (hospital death, reoperation) when controlling for potential confounders. The results quoted are for the hazard ratio with 95% CI. Kaplan-Meier actuarial survival curves were calculated for hospital and overall mortality and for freedom from reoperation. The log-rank test was used to assess whether there was a significant difference in survival between subject groups. All statistical analyses were conducted using the SAS software package (SAS Institute, Cary, NC) and Stata (versions 6 and 7; Stata Corporation, College Station, TX).
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
|
|
Demographic variables
There was no significant difference in age or weight between those who died (median age 5.5 months and weight 4.9 kg) and those who survived beyond 30 days (median age 5.2 and weight 5.3 kg, p = 0.82 and 0.77, respectively). Hospital death rates were examined according to whether the operation occurred before 1995 (14 of 83, 17%) or after (8 of 64, 13%) and again no significant difference was found during the study period (p = 0.49). The hospital mortality for children with Down syndrome (17 of 106, 16%) did not differ significantly from that for chromosomally normal children (5 of 41, 12%; p = 0.80).
Cardiac variables
The following were not significantly associated with a risk of hospital death: degrees of preoperative atrioventricular valve regurgitation (p = 0.86), the presence of other anomalies such as an ostium secundum atrial septal defect or patent arterial duct (p = 0.19), or previous shunt (p = 0.21). While the association between hospital death and the presence of a dysplastic or relatively hypoplastic valve had borderline significant (Fisher’s exact test p = 0.08, hazard ratio: 2.54, 95% CI 0.99 to 6.49), when dysplasia and hypoplasia were assessed individually neither was statistically significant (p = 0.41 and p = 0.10, respectively). An important risk factor for death was the presence of a double orifice atrioventricular valve (Fisher’s exact test p = 0.002, hazard ratio 6.68, 95% CI 2.60 to 17.17). Survival analyses showed that patients with double orifice atrioventricular valve had a significantly lower probability of survival than did patients without (log-rank test p < 0.001; Fig 2).
|
To analyze the impact of a double orifice atrioventricular valve further, the hazard ratio of hospital death was estimated controlling for whether the atrioventricular valve was dysplastic or relatively hypoplastic. Whether or not the valve was affected, the risk of hospital death was still more than six times greater for a double orifice atrioventricular valve than for other valve anatomy (hazard ratio 6.27, 95% CI 2.42 to 16.25). The presence of a dysplastic or relatively hypoplastic valve remained a nonsignificant risk factor after controlling for the presence of a double orifice atrioventricular valve (hazard ratio 2.27, 95% CI 0.88 to 5.86).
Overall mortality
Eight patients died later than 30 days after surgery giving a 13.5-year actuarial survival of 80%. There was no significant difference in death rates among the three operation types (p = 0.25). As with early mortality, there was a significantly higher death rate among those with a double orifice atrioventricular valve (Fisher’s exact test p = 0.01) and survival analyses showed that the probability of survival was significantly lower in this group compared with those without the condition (log-rank test p < 0.001; Fig 3).
|
|
|
The rate of reoperation increased with degree of preoperative left atrioventricular valve regurgitation (three levels) and the trend was borderline significant (p = 0.06). To satisfy the proportional hazards criteria when estimating the univariate hazard ratio, the data were collapsed to two levels (no or trivial regurgitation versus mild to severe regurgitation) and the resulting measure of risk was not statistically significant (hazard ratio 1.85, 95% CI 0.79, 4.32). Children with a dysplastic left atrioventricular valve were found to have a higher rate of reoperation (46% with dysplasia versus 14% without dysplasia; p = 0.01) and a significantly lower probability of survival (in terms of freedom from reoperation) than those without the condition (log-rank test p < 0.001, Fig 6). The univariate hazard ratio for requiring reoperation was more than six times higher for those with dysplasia compared with those without (hazard ratio 6.43, 95% CI 2.49 to 16.57). The rate of dysplasia was lower among Down’s syndrome patients (3%) than among chromosomally normal patients (24%, p < 0.001).
|
Follow-up
The median follow-up time after hospital discharge was 57.3 months (range 5.2 to 158.6). Clinically 56 (48%) patients had no evidence of left atrioventricular valve regurgitation, and 50 (43%) had mild, 7 (6%) had moderate, and 4 had severe regurgitation that required medical treatment. Mild left atrioventricular valve stenosis was present in 4 patients who were asymptomatic. Ten patients (7%) had a hemodynamically insignificant small residual ventricular septal defect. The clinical findings were confirmed by cross-sectional echocardiography. Three patients needed a permanent pacemaker for complete heart block.
|
Comment |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
In several studies, pulmonary artery banding, as a palliative procedure in early infancy is no longer recommended unless other associated abnormalities make primary repair a high-risk operation [5–9, 11, 13]. Our main indications for banding were severe heart failure in infants younger than 3 months (particularly earlier in the series), multiple ventricular septal defects, common atrioventricular valve dominantly committed to the right ventricle but with a left ventricle deemed of adequate size for biventricular repair, double outlet right ventricle, and infections such as severe bronchiolitis or myocarditis. In patients with marked atrioventricular valve regurgitation, pulmonary artery banding is contraindicated as an increase in ventricular afterload may exacerbate left ventricle to right atrial shunting (obligatory shunt) [16]. In contrast to other reports [7] none of our patients undergoing pulmonary artery banding died at palliation, although 1 of 19 died at subsequent repair. This patient had a severely dysplastic left atrioventricular valve and multiple ventricular septal defects. He died after a second bypass procedure to repair residual left atrioventricular valve regurgitation.
Outcome
Our overall hospital mortality of 15%, while within the range of many other studies (Table 4),
was not as low as reported by Weintraub and colleagues [3]. Our data include all children undergoing repair in a 13-year period with no omissions. The completeness of data were independently audited after allegations (subsequently shown to be unfounded) of discrimination against children with Down syndrome. We were still concerned, however, to determine any factors that might explain the mortality rate.
|
The Mayo Clinic group [19] reported double orifice left atrioventricular valve in 9 of 253 (3.6%) cases of complete AVSD with one early and two late deaths after repair. The early death was in a 5-month-old infant, but in general repair was carried out in an older age group. Nonetheless they did not find double orifice left atrioventricular valve to be a particular risk factor.
Patients with complete AVSD are at greatest risk of pulmonary vascular disease even in the first year of life [7, 11, 20]. They are also at risk of postoperative pulmonary hypertensive crisis, the main cause of death in 6 of our patients. In addition to hyperventilation and a high inspired FiO2, nitric oxide has also been used since 1993 in the treatment of postoperative pulmonary hypertension although its introduction has not had a major impact in preventing hospital death. In Down syndrome early pulmonary vascular disease may be the cause of death in the first year [17]. Haworth [21] found that lung biopsy of infants who had died may show either severe intimal fibrosis with minimal increase in muscularity or a great increase in muscularity. Haworth also emphasized that even potentially reversible abnormalities such as increased muscularity can prejudice intracardiac repair. Clapp and associates [18] found that 10 of 81 children (12%) with Down syndrome had fixed pulmonary vascular disease that developed well before 1 year of age. They suggested that cardiac catheterization with measurement of pulmonary vascular resistance should be performed in all children with a complete AVSD even younger than 6 months of age, especially in Down syndrome. Many groups, however, do not consider cardiac catheterization necessary for patients younger than 6 months.
We found, as have others, that age at operation is not a significant risk factor for hospital death although 93% of our patients with nonrestricted pulmonary blood flow were either repaired or banded under a year. Similarly hospital death was similar among Down syndrome and normal chromosome children [7, 9]. Reller and coworkers [14] performed a population-based study to analyze the potential role that Down syndrome plays in the surgical outcome of corrected congenital heart disease. They found that a complete AVSD but not other cardiac anomalies was associated with a significantly higher early (13% versus 5%) and late mortality (20% versus 5%, p < 0.05). In our study weight at operation, the degree of left atrioventricular valve regurgitation, additional anomalies, and cardiopulmonary bypass time were not significant risk factors. The particular surgical technique used also did not influence mortality.
Twenty-four of our patients (17% of those available; 70% CI 13% to 20%) required a second operation requiring cardiopulmonary bypass, 19 for left atrioventricular valve regurgitation, similar to other studies [3, 7, 12]. We also confirmed the finding [3, 6] that children with normal chromosomes had a statistically higher risk of requiring reoperation than did children with Down syndrome. That may be in part because dysplasia of the left atrioventricular valve is more common in children with normal chromosomes, but Down syndrome remained protective after controlling for dysplasia. While Pozzi and colleagues [7] found that preoperative regurgitation was a risk factor for the need for reoperation, we found a borderline significant (p < 0.1) trend in the need for reoperation with increasing regurgitation. After collapsing the variable to two levels (no or trivial regurgitation versus mild to severe regurgitation), however, the hazard ratio was not statistically significant. Others [12] consider their reduction in reoperation rate contributed to by the use of pledglets when suturing the zone of apposition of the inferior and superior bridging leaflets, selective use of left atrioventricular valve annuloplasty, and careful intraoperative assessment of valve repair for residual regurgitation. Suzuki and associates [13] and Michielon and colleagues [6] showed that primary repair at less than 6 months may prevent left atrioventricular valve regurgitation from annular dilatation and degenerative changes in the valve. In contrast Bonnet and colleagues [9] thought those with left atrioventricular valve anomalies who required surgery at less than 4 months were at higher risk for significant left atrioventricular valve regurgitation than those operated on later. In our study there was a trend, although not reaching statistical significance for repair at less than 3 months, to be a protective factor against the need for reoperation for left atrioventricular valve regurgitation (3% versus 16%).
It is clear that achieving a competent, nonstenotic left atrioventricular valve at the time of reconstruction is important for outcome. The morphology of the common atrioventricular valve is variable but usually allows for the creation of two adequate valves even if there is some degree of left atrioventricular valve incompetence or stenosis. The latter may be produced particularly when the mural leaflet is small. When a double orifice left atrioventricular valve is identified preoperatively, especially if either component is hypoplastic or dysplastic, the need for left atrioventricular valve replacement should be considered before repair. There is also a suggestion that postoperative His bundle tachycardia is more common with this anomaly. Parents should be informed about the higher risk of death when this complex valve anomaly is present. Because of the concern of placing a mechanical valve in a young infant, none of our patients had a valve replacement at initial repair. Thus accurate preoperative and postoperative evaluation of left atrioventricular valve morphology is important in the assessment of risk and outcome for mortality and reoperation rate.
It is now well recognized that repair of a complete AVSD is optimally carried out before the patient reaches 6 months or preferably 4 months to prevent the development of pulmonary vascular disease [3]. Operation before 3 months may also improve the long-term durability of atrioventricular valve function [12].
Study limitation
This was a retrospective study in which the median follow-up was relatively short. A retrospective study could have limitations in terms of how complete and correct the information is in the medical and surgical records. As our case records have been independently audited we do know that we are not missing cases with a poor outcome, the exclusion of which would give an apparently lower mortality.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Abadir, V. Fouilloux, D. Metras, O. Ghez, B. Kreitmann, and A. Fraisse Isolated cleft of the mitral valve: distinctive features and surgical management. Ann. Thorac. Surg., September 1, 2009; 88(3): 839 - 843. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Stulak, H. M. Burkhart, J. A. Dearani, H. V. Schaff, F. Cetta, R. D. Barnes, and F. J. Puga Reoperations After Initial Repair of Complete Atrioventricular Septal Defect. Ann. Thorac. Surg., June 1, 2009; 87(6): 1872 - 1878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. H. Shuhaiber, S. Y. Ho, M. Rigby, and B. Sethia Current options and outcomes for the management of atrioventricular septal defect Eur. J. Cardiothorac. Surg., May 1, 2009; 35(5): 891 - 900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Formigari, G. Michielon, M. C. Digilio, G. Piacentini, A. Carotti, A. Giardini, R. M. Di Donato, and B. Marino Genetic syndromes and congenital heart defects: how is surgical management affected? Eur. J. Cardiothorac. Surg., April 1, 2009; 35(4): 606 - 614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Dodge-Khatami, S. Herger, V. Rousson, M. Comber, W. Knirsch, U. Bauersfeld, and R. Pretre Outcomes and reoperations after total correction of complete atrio-ventricular septal defect Eur. J. Cardiothorac. Surg., October 1, 2008; 34(4): 745 - 750. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Locker, J. A. Dearani, P. W. O'Leary, and F. J. Puga Endoluminal Pulmonary Artery Banding: Technique, Applications and Results Ann. Thorac. Surg., August 1, 2008; 86(2): 588 - 595. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Lange, T. Guenther, R. Busch, J. Hess, and C. Schreiber The presence of Down syndrome is not a risk factor in complete atrioventricular septal defect repair J. Thorac. Cardiovasc. Surg., August 1, 2007; 134(2): 304 - 310. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-Q. Lai, Y. Luo, C. Zhang, and Z.-G. Zhang Utilization of Double-Orifice Valve Plasty in Correction of Atrioventricular Septal Defect Ann. Thorac. Surg., April 1, 2006; 81(4): 1450 - 1454. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. K. Kim, W.-H. Kim, S. W. Hwang, J. Y. Lee, J. Y. Song, S.-J. Kim, and K. Y. Jang Predictive Value of Intraoperative Transesophageal Echocardiography in Complete Atrioventricular Septal Defect Ann. Thorac. Surg., July 1, 2005; 80(1): 56 - 59. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Masuda, H. Kado, Y. Tanoue, K. Fukae, T. Onzuka, Y. Shiokawa, T. Shirota, and H. Yasui Does Down syndrome affect the long-term results of complete atrioventricular septal defect when the defect is repaired during the first year of life? Eur. J. Cardiothorac. Surg., March 1, 2005; 27(3): 405 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D. J. Ten Harkel, A. H. Cromme-Dijkhuis, B. C.C. Heinerman, W. C. Hop, and A. J.J.C. Bogers Development of Left Atrioventricular Valve Regurgitation After Correction of Atrioventricular Septal Defect Ann. Thorac. Surg., February 1, 2005; 79(2): 607 - 612. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A.A. Al-Hay, C. R. Lincoln, D. F. Shore, and E. A. Shinebourne The left atrioventricular valve in partial atrioventricular septal defect: management strategy and surgical outcome Eur. J. Cardiothorac. Surg., October 1, 2004; 26(4): 754 - 761. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Sekarski, P. Fridez, A. F. Corno, L. K. von Segesser, and E. J. Meijboom Doppler-guided regulation of a telemetrically operated adjustable pulmonary banding system J. Am. Coll. Cardiol., September 1, 2004; 44(5): 1087 - 1094. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Formigari, R. M. Di Donato, G. Gargiulo, D. Di Carlo, C. Feltri, F. M. Picchio, and B. Marino Better surgical prognosis for patients with complete atrioventricular septal defect and Down's syndrome Ann. Thorac. Surg., August 1, 2004; 78(2): 666 - 672. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
