|
|
||||||||
Ann Thorac Surg 2006;81:678-684
© 2006 The Society of Thoracic Surgeons
Division of Cardiothoracic Surgery, Department of Surgery, University of California-Los Angeles (UCLA) School of Medicine, Los Angeles, California
Accepted for publication July 13, 2005.
* Address correspondence to Dr Odim, Division of Cardiothoracic Surgery, UCLA School of Medicine, 10833 Le Conte Avenue, CHS 62-182, Los Angeles, CA 90095 (Email: jodim{at}mednet.ucla.edu).
| Abstract |
|---|
|
|
|---|
METHODS: We reviewed the medical records for 106 patients with pulmonary atresia-intact ventricular septum (PA-IVS) treated between 1982 and 2001. Over this period, children were assigned to mild (>2/3 normal size, 23.7% of patients), moderate (1/3 to 2/3, 41.2%), or severe (1/3, 35.1%) right ventricular hypoplasia, and this grouping, along with severity of coronary anomalies (45% right ventricle to coronary fistulae, 16% with right ventricle dependent coronary circulation [RVDCC]), triaged children to eventual single ventricle (severe) or two-ventricle (mild or moderate) repair.
RESULTS: Actuarial 10-year survival was 86.3% with mortality predicted by severe hypoplasia (odds ratio [OR] 12.9, p < 0.001), RVDCC (OR 15.0, p < 0.001), and non-Caucasian race (OR 10.7, p < 0.001). Multivariate analysis with a Cox proportional hazards model confirmed only RVDCC (risk ratio [RR] 10.9, p = 0.0009} and non-Caucasian race (RR 6.9, p = 0.007) as significant. Although not an independent risk factor for survival, the degree of hypoplasia was the most important determinant for definitive repair. Severe hypoplasia virtually precluded two-ventricle repair (OR 33.1, p < 0.001 by
2 analysis) and was the strongest risk factor for a one-ventricle system (OR 78.7, p < 0.001). Actuarial survival after either repair was 91%, and no biventricular repair later converted to a Fontan system.
CONCLUSIONS: Surgical management of patients based on this three tier grade for right ventricular hypoplasia results in excellent survival and correctly predicts patients destined for eventual Fontan and biventricular repair.
| Introduction |
|---|
|
|
|---|
| Patients and Methods |
|---|
|
|
|---|
The average number of palliative operations was 1.6 ± 0.1 and median age at initial palliation was 4 days (range, 1 to 216 days). The most common palliative procedure was a systemic to pulmonary artery communication performed in 100 of the 106 patients (94.3%) (see Table 1 for systemic to PA shunting in 97 children assigned a right ventricle [RV] size). Sixteen of this group required additional systemic to pulmonary artery shunting for persistent cyanosis. Twenty-four patients underwent a cavopulmonary anastomosis. Fifty-six of the 109 patients (52.8%) received relief of right ventricle outflow tract obstruction (RVOTO) by pulmonary valvotomy (32.1%), transannular patch (13.2%), or both (7.5%). Only five patients in this study presented with additional Ebstein's malformation of the tricuspid valve. In this subset, one patient achieved a complete biventricular repair and another patient successfully underwent partial biventricular correction. A third patient died after initial palliation and two remaining patients await further intervention after recent palliation.
|
The qualitative estimates of right ventricular dimensions are based on preoperative two-dimensional, transthoracic echocardiograms and were available for 97 of the 106 patients. In PA-IVS, the accompanying left ventricle (LV) is usually of normal size and this is the yardstick used to assess RV hypoplasia. In 60 children, tricuspid valve orifice diameters were quantified, and were used in conjunction with the patient's body surface area to determine tricuspid Z scores from published nomograms [3].
Right Ventricle to Coronary Fistulae and RVDCC Dependence
Right ventricle to coronary fistulae were identified by coronary angiograms and ventriculograms for all patients. A spectrum of fistulae covering a broad range of physiologic significance existed. For the purpose of statistical analysis, their mere presence was treated identically among all patients. Right ventricle dependent coronary circulation (RVDCC) was deemed present when a major portion of LV myocardial perfusion was supplied exclusively from the right ventricle cavity. This circumstance occurred with proximal left coronary stenoses with concomitant right ventricle to left coronary fistulae, extensive right ventricle to coronary fistulae, or myocardial perfusion scans indicative of reversible defects.
Definitive Repair
A one-ventricle repair was synonymous with Fontan completion. A complete two-ventricle repair occurred when each ventricle ejected into its respective great artery without patent systemic pulmonary artery shunts. For most of these repairs an adjustable interatrial communication was created. Fifty-seven of the 106 children underwent biventricular repairs (complete or partial). In 30 children (52.3%) the atrial communication was completely snared closed after operation. The mean and median number of days toward ASD snare closure after biventricular repair was 474 ± 155 and 67 days, respectively. Of the remaining 25 patients, the ASD remained partially snared in 9.1%, spontaneously closed in 6.8%, and remained open permitting right to left shunting in 32%. The mean follow-up for this analysis was 351 ± 221 days. The concept of an adjustable ASD allows for a restrictive interatrial communication that can be narrowed or closed postoperatively by a snaring device left underneath the linear alba. This technique was universally applied to all candidates for biventricular repair to reduce right atrial pressure below 15 mm Hg, thereby preserving LV preload and cardiac output in the midst of acceptable oxygen saturations in the mid 80s. For statistical analysis, patients with partial biventricular repair had neither a one-ventricle nor complete two-ventricle system.
Study Design and Statistical Analysis
All patients were assigned one of six outcomes: completed two-ventricle repair, partial two-ventricle repair, one-ventricle repair, transplantation, death prior to definitive repair, or awaiting definitive repair. To evaluate the influence of right ventricle hypoplasia, right ventricle to coronary fistulae, and RVDCC on achieving these outcomes, each of these variables was used to segregate patients for the construction of 2 x 2 contingency tables. The
2 analysis was used to calculate the p value.
The influence of race, gender, right ventricular hypoplasia, right ventricle to coronary fistulae, and RVDCC on time-dependent variables of survival, freedom from one-ventricle repair, and probability of two-ventricle repair was determined using Kaplan-Meier actuarial analysis, with p values calculated by log rank statistics [4]. The Cox proportional hazards regression model [5] was used for the multivariate analysis. In arriving at a summarizing model, each significant risk factor for mortality, one-ventricle repair, or two-ventricle repair was subjected to multivariable linear regression analysis. Death prior to definitive repair received a score of zero, while one-ventricle, partial two-ventricle, and completed two-ventricle repairs received scores of 1.0, 1.5, and 2.0, respectively. Significance was determined by the analysis of the one way analysis of variance method. To refine the model further, multivariable linear regression was repeated after exclusion of insignificant coefficients.
| Results |
|---|
|
|
|---|
RVDCC and Non-Caucasian Race Predict Poor Survival in Patients With PA-IVS
The actuarial survival of patients at 10 years was 86.3% (Fig 1). Twelve of the 16 deaths occurred within the first year of life, with seven within the first month. Although all patients received palliative intervention, 9 of these 16 patients died before definitive repair. These early deaths after palliation comprise the vast majority of deaths through the first 8 months of life (Fig 1). However, after this period the survival curve for the patient population mirrors that for those children reaching definitive repair. Actuarial survival after definitive repair was 91% over 10 years regardless of type of repair. Analysis of risk factors for death in this critical interval between palliation and definitive repair identified one demographic and two anatomic predictors of mortality (Table 2). A severe grade of right ventricular hypoplasia increases the odds of death by 12.9-fold, while RVDCC causes a 15-fold elevation in risk. Kaplan-Meier analysis confirms these two parameters as risk factors for death. One year actuarial survival in children with severe RV hypoplasia is 79.3% but 94.7% with mild or moderate hypoplasia (log-rank p = 0.049). Similarly, one year actuarial survival in patients with RVDCC is 56.3% and 93.8% without this coronary anomaly (log-rank p < 0.001). For both anatomic factors, risk is accrued entirely within the first year of life.
|
|
|
Male Gender and Severe RV Hypoplasia Predict One-Ventricle Repair in Patients With PA-IVS
Severe right ventricular hypoplasia carries the greatest weight in determining whether a child with PA-IVS will progress to Fontan. Among 33 patients with severe hypoplasia, 19 (51.6%) received a one-ventricle repair. The odds of yielding to a one-ventricle system with severe hypoplasia were 78.7-fold greater than in children with either mild or moderate RV hypoplasia (p < 0.001). The median duration to Fontan repair with severe RV hypoplasia was 4.2 years. No more than 7% of children with mild or moderate hypoplasia progressed to Fontan over 17 years of follow-up (log-rank p < 0. 001, Fig 3A). A weaker anatomic risk factor predicting single ventricle repair was the presence of right ventricle to coronary fistulae. Thirty-five percent of patients with fistulae required one-ventricle repair compared with 14% without these communications (OR 3.3%, p = 0.032).
|
p = 0.002). In this model, the presence of right ventricular to coronary fistulae is not an independent risk factor for eventual one-ventricle repair (p = 0.57).
|
|
|
|
|
|
Race: 1 = Caucasian, 0 = non-Caucasian
RVDCC: 1 = Present, 0 = Absent
This model encompasses many of the results from this study. For example, a severe grade of right ventricular hypoplasia means the best outcome score possible is 1.28, a value intermediate between one-ventricle repair and a mixed circulation. These patients are not amenable to two-ventricle repair (outcome of 2.0). Additionally, the best score possible for a patient with RVDCC is 1.27, again intermediate between one-ventricle repair and a mixed circulation. However, since no patient with mild hypoplasia had RVDCC in this cohort, the best outcome is more reasonably estimated as 0.94 or lower.
| Comment |
|---|
|
|
|---|
The ten-year survival of 86.3% in this study mirrors the steady improvement in outcomes reported at multiple centers over the past 25 years [612]. In an early multiinstitutional prospective study, Hanley and colleagues [3] reported a 64% four-year actuarial survival for infants treated between 1987 and 1991. Similar to our study and others [9, 13], their patients experienced the greatest hazard to survival in the first year of life. A later analysis between 1987 and 1997 from the Congenital Heart Surgeons Society analysis advocated a morphologically driven institutional protocol emphasizing both biventricular and Fontan pathways may mitigate the negative effect of unfavorable morphology. Ashburn and colleagues [14] demonstrated in this analysis that 85% of neonates were likely to reach a definitive surgical end point with a biventricular repair attained in up to 50%. Among a group of 47 patients at Children's Hospital in Boston with PA-IVS operated on between 1991 and 1998, the actuarial survival was 98% at 7 years [9], and in large single center series similar to our own with twenty-year patient accruals, long-term survival has ranged from 68 to 76% [10, 15].
Although it is tempting to attribute the successful outcomes to time-related refinements in surgical techniques and management strategies, evaluation of the data shows that virtually equal numbers of patients died in the first and last decade of this study. Our group has relied on this three tier grading scale for right ventricular hypoplasia and presence or absence of RVDCC to guide pursuance of either a one-ventricle or two-ventricle repair. A patient's right ventricular size contributes to his risk for early mortality, whether defined as tricuspid Z score [3], muscular pulmonary atresia [15], or presence of a "well-formed" infundibulum [7]. This influence of hypoplasia on survival was not altered by stratifying patients into mild, moderate, and severe groups in lieu of a continuous measure like the Z score or binary measure as the presence of a "well-formed" infundibulum. Among our patients, severe hypoplasia increased the odds of death by 12.9, but was a much smaller risk factor after accounting for RVDCC. As suggested by others [16], the real risk of death in patients with severe right ventricular hypoplasia may reside in coexistent RVDCC [2, 3].
Although subtle differences in the incidence of different congenital heart defects are noted from epidemiologic databases between Caucasian and non-Caucasian populations [1719], no previous reports on PA-IVS have noted a racial predilection for early mortality. However, this study identifies the potential importance of non-Caucasian background when correcting for hypoplasia and RVDCC. The significance of this finding is unknown, and at best speculative as most large clinical series neglect racial composition in the analysis [3, 9, 10]. Greater focus on this issue in the future may provide explanations for its influence. Of note, when overall outcome for the entire cohort is modeled through multivariable linear regression, non-Caucasian background plays a small but significant role, especially among patients with unfavorable cardiac anatomy, and male gender is not a significant influence on outcome.
In conclusion, the grade of RV hypoplasia guides surgical management and predicts outcomes in children with PA-IVS. The use of a three variable linear regression model incorporating the degree of hypoplasia, the presence or absence of RVDCC, and race gives a quantitative measure of whether a patient will die prior to definitive repair. Additionally, the score obtained with this model provides an estimation of whether a patient will receive a one-ventricle, partial two-ventricle, or complete two-ventricle repair and provides treating physicians with a prognostic measure that can be communicated to the child's parents and used to direct optimal initial palliation [1922].
| Member and Individual Subscriber Access to the Online Annals |
|---|
|
|
|---|
Society members and non-members alike who do not know their CTSNet user name and password should follow the link &x201C;Forgot your user name or password?&x201D; that appears below the boxes where you are asked to enter this information when you try to gain full-text access. Your user name and password will be e-mailed to the e-mail address you designate.
In lieu of the above procedure, if you have forgotten your CTSNet username and/or password, you can always send an email to CTSNet via the feedback button from the left navigation menu on the homepage of the online Annals or go directly to http://ats.ctsnetjournals.org/cgi/feedback .
We hope that you will view the online Annals and take advantage of the many features available to our subscribers as part of the CTSNet Journals Online. These include inter-journal linking from within the reference sections of Annals' articles to over 350 journals available through the HighWire Press collection (HighWire provides the platform for the delivery of the online Annals). There is also cross-journal advanced searching, eTOC Alerts, Subject Alerts, Cite-Track, and much more. A listing of these features can be found at http://ats.ctsnetjournals.org/help/features.dtl .
We encourage you to visit the online Annals at http://ats.ctsnetjournals.org and explore.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
R. Bryant III, E. R. Nowicki, R. B.B. Mee, J. Rajeswaran, B. W. Duncan, G. L. Rosenthal, U. Mohan, M. Mumtaz, and E. H. Blackstone Success and limitations of right ventricular sinus myectomy for pulmonary atresia with intact ventricular septum J. Thorac. Cardiovasc. Surg., September 1, 2008; 136(3): 735 - 742. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. E. Foker, S. P. Setty, J. Berry, P. Jain, K. Catton, A. C. Gittenberger-de-Groot, and L. A. Pyles Treatment of right ventricle to coronary artery connections in infants with pulmonary atresia and intact ventricular septum J. Thorac. Cardiovasc. Surg., September 1, 2008; 136(3): 749 - 756. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. W. Salvin, D. B. McElhinney, S. D. Colan, K. Gauvreau, P. J. del Nido, K. J. Jenkins, J. E. Lock, and W. Tworetzky Fetal Tricuspid Valve Size and Growth as Predictors of Outcome in Pulmonary Atresia With Intact Ventricular Septum Pediatrics, August 1, 2006; 118(2): e415 - e420. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. P. Graham Jr The Year in Congenital Heart Disease J. Am. Coll. Cardiol., June 20, 2006; 47(12): 2545 - 2553. [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 |