| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;84:907-912
© 2007 The Society of Thoracic Surgeons
a Department of Pediatrics, University of Utah and Primary Children’s Medical Center, Salt Lake City, Utah
b Department of Surgery, University of Utah and Primary Children’s Medical Center, Salt Lake City, Utah
c Department of Internal Medicine, University of Utah, Salt Lake City, Utah
Accepted for publication April 24, 2007.
* Address correspondence to Dr Puchalski, Primary Children’s Medical Center, 100 N Medical Dr, Salt Lake City, UT 84113 (Email: michael.puchalski{at}intermountainmail.org).
Presented at the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29–31, 2007.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
Methods: We reviewed the records of all patients undergoing RV-PA cryopreserved valved allograft placement from 1988 to 2006 for diagnosis, age, allograft type, time to valved conduit explant, and indication for surgery. Conduit size at the time of insertion was compared with pulmonary valve size normalized for body surface area (z score). Multivariate Cox regression models with cluster analysis were constructed to assess risk of allograft oversizing for conduit failure. Kaplan-Meier analysis was used to obtain median freedom from explantation time.
Results: A total of 140 cryopreserved valved allografts (z score, 1.8 ± 1.3; range, –1.5 to 4.9) were implanted in 99 patients (median age, 5.6 years). Reoperation was required in 66 (67%) of 99 patients during the study period. Modeling z scores as a dichotomous variable revealed that risk of allograft explantation increases 113% when allografts with z scores of 2.7 or higher are used compared with those with z scores of less than 2.7 (p < 0.01). Median adjusted freedom from explantation for this same grouping was 4.9 years versus 9.4 years. The presence of branch pulmonary artery stenosis shortens the conduit life (p < 0.001), whereas insertion of a pulmonary allograft may extend the conduit life (p = 0.13).
Conclusions: Cryopreserved valved allograft oversizing (z score 
2.7) in the pulmonary position results in decreased longevity in children. Presence of branch pulmonary artery stenosis is also associated with earlier conduit explantation.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
Conduit Characteristics
Each valve implant was analyzed as a separate event. Heterografts and valveless conduits were excluded from the analysis. The minimal follow-up time was 6 months. Conduit size was determined from the manufacturer’s implant form provided with the cryopreserved valve and compared with normal pulmonary valve size for BSA (z score) at the time of insertion or explantation [7]. Over-sizing was defined as z score of 2.0 or more. Allograft failure was defined as explant of the valve for any cause. Indications for allograft explantation were determined by reviewing preexplant echocardiograms, the clinical record, catheterization, and operative reports. Cardiac magnetic resonance imaging was not used owing to limited numbers (n = 5).
The indications were classified into three general categories: conduit stenosis, conduit insufficiency, or mixed stenosis and insufficiency. In general, conduits with a transvalvular peak instantaneous pressure gradient of greater than 50 mm Hg [8] along with a peak right ventricle (RV) pressure greater than 75% of systemic pressure were considered stenotic and were replaced. Branch pulmonary artery gradients were evaluated in the same manner. Valved conduits were replaced for conduit valve insufficiency if the insufficiency was at least moderate and associated with RV dilation, diminished RV function, or patient symptoms. Allograft insufficiency was classified as at least moderate by the presence of color Doppler flow reversal in the branch pulmonary arteries and a regurgitant jet width-to-annulus ratio exceeding 0.4 [9].
Statistical Analysis
Descriptive data are presented as mean values ± standard deviation or percentages. Allograft survival was determined using the Kaplan-Meier method, and comparisons were performed using the log-rank test. To assess risk of allograft oversizing for valve failure, Cox regression models were fitted to the data. Multiple conduit operations per patient were included in the data, so a robust variance estimate of the standard errors was used to account for the lack of independence among conduit operations this introduced into the model.
Collinearity among candidate predictor variables for the regression model was assessed by fitting regression models with each variable in turn as the dependent variable, with the remaining variables as the independent variables [10]. For collinearly related variables, the decision of which variable to retain in the final Cox model was determined by the clinical relevance of the retained variable. Statistical analysis was performed using Stata 9.1 software (StataCorp, College Station, TX), and the graphic presentations were done in GraphPad Prism 4 software (GraphPad Software Inc, San Diego, CA).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
|
|
|
2.7) relative to the first three quartiles (z score < 2.7). Modeling z scores as a dichotomous variable revealed that risk of conduit explantation increases 113% when allografts with z scores of 2.7 or more are used compared with z scores of less than 2.7 (HR, 2.13; 95% confidence interval [CI], 1.27 to 3.56; p < 0.01), after adjusting for diagnosis, branch PA stenosis and allograft type (Table 3). Using the same covariates, we generated predicted Kaplan-Meier curves that showed the adjusted median freedom from explantation was 4.9 years in the allografts with scores of 2.7 and higher and 9.4 years in those with scores of less then 2.7 (Fig 2).
|
|
In addition, there is a trend toward decreased risk of conduit explantation with pulmonary allograft use. To determine whether the largest allografts (z score 2.7, n = 35) are failing due to outgrowth, the z scores were calculated again at the time of their explantation using the current BSA and the original conduit size. Reoperations were done on 22 conduits of this size, and the median (interquartile ranges) z score at explantation was 0.11 (range, –0.6 to 1.1). Only one z score fell out of the normal range, and the conduit was replaced because of stenosis. The remaining conduits were replaced for a mixture of stenosis, insufficiency, or both. A review of all explanted conduits shows 19 (25%) of 77 had z scores of –2.0 or less, but only half were replaced for pure stenosis.
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
The effect of allograft diameter on failure has been extensively studied in the past, but no consensus has been reached on the role size plays in conduit failure in children. Previous reports have shown that smaller conduit diameter is associated with shorter longevity, but these studies failed to normalize allograft size to BSA [11, 12]. Other investigators looking exclusively at allografts found that absolute diameter had no relation to longevity [13–15].
Given the wide variation in patient size in a pediatric population, it is more appropriate to analyze the effects of allograft size by comparison with normal size for a given BSA (z scores). Using this approach, Caldarone and colleagues [16] retrospectively analyzed 945 prosthetic valve implantations in 726 patients across an age range of 3 months to 65 years. They found that smaller z scores predicted shorter conduit life in patients aged older than 3 months but not for those younger than 3 months. Similarly, Tweddell and colleagues [17] looked at the outcomes of 220 homografts implanted in 183 patients (mean age, 6.9 ± 7.6 years) and found that a z score of less than 2 was an independent risk factor for valve reoperation.
In contrast, a more recent study by Karamlou and colleagues [18] evaluated the effect of conduit oversizing, defined as a z score exceeding 1, in children and young adults requiring RV-PA reconstruction. They reported that oversized allografts did not improve durability in this cohort. In a large study by the Congenital Heart Surgeon’s Society [19], pulmonary conduit function and durability were found to decrease with conduit z scores exceeding 3. Our results support their findings. We have found that not only did oversizing an allograft with a z score of 2.7 or more fail to improve longevity, it actually decreased time to conduit explantation, independent of other factors in a multivariable analysis. Of the explanted conduits, 75% had an assumed (we did not factor in changes in the conduit over time) normal z score and lacked pure stenosis as a reason for explant. Furthermore, only one conduit (1/22), which had a z score exceeding 2.7 at implant, was considered small at explant, suggesting that these conduits were replaced for reasons other than pure outgrowth of the original conduit in most cases.
In addition to evaluation of oversizing, we found other risk factors for allograft failure. First, preoperative branch PA stenosis was a significant risk factor for earlier allograft failure. This is similar to the findings of Karamlou and colleagues [19] and emphasizes the role that arborization, branch PA stenosis, and PA pressure may play in conduit deterioration. Second, the pulmonary allografts appear more durable than aortic allografts. Some investigations have reported findings [17] of the superiority of pulmonary allografts and have postulated inherent differences in aortic allografts and their tendency towards calcification [11, 15], whereas others found no difference [3, 14].
This study has several limitations. It is a retrospective analysis with its inherent difficulties. We have included only variables that have been frequently mentioned previously in the literature as independent risk factors for allograft failure. We cannot exclude the possibility that the z score variable could be confounded by other factors we did not examine in our study, such as ABO compatibility or donor ischemia time. We also included 9 Ross patients, who may be less likely to have allograft failure because of the anatomic position of the valve [20]. The numbers of patients with preoperative branch pulmonary stenosis as well as those with aortic allografts were small.
In conclusion, we found that allografts with very large z scores of 2.7 or more are associated with earlier explantation compared with allografts with z scores of less than 2.7. Our data suggest that somatic outgrowth may be only a small part of the reason for conduit failure in children. Large allografts may fail more rapidly due to external compression of the conduit by the closed sternum, valve distortion, and insufficiency with sternal compression, or distortion of distal PAs from the oversized conduit. In addition, we found branch PA stenosis and aortic allografts to be associated with earlier conduit explantation. Avoidance of extremely oversized allografts, relief of all branch PA stenosis, and the use of pulmonary rather than aortic allografts may maximize durability of allografts in children.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
And my second question is whether you explored with some type of sensitivity analysis the optimum z score, or how you came to dichotomizing the z-score variable at a cutpoint of 2.7?
I ask this in light of the CHHS study we presented last year at the AATS, wherein there was actually a quadratic relationship, in other words, a very small or a very large homograft was actually a risk factor for conduit-related dysfunction. The CHSS study also found that there was actually a just right size, which is around a z score of plus 1.
DR ASKOVICH: Okay. Great questions. Thank you very much. So, yes, we realized that obviously there are differences, some differences were due to the age when we included all of the children. So the analysis basically averages out, and on average, oversizing was significant risk. However, we do feel that our data suggests that if you take the youngest children, you will see, and probably if we had more patients, that possibly you will get the significance in the other direction, that oversizing is actually good and protective. So I would say if we had larger numbers of kids of the smaller age, that could be significant.
The second question was about the cutoff values. Actually, when we first did the analysis and used z score as a continuous variable, we just saw a slightly larger than 1 hazard ratio. And that actually prompted us to try to look more carefully into whether the risks exist. So that is why we divided the z score in four quartiles. And we saw that the risk is increased in the last quartile where allografts are oversized, that is how we decided to use that. And that could be a possible explanation why, for instance, your study found no increased risk beyond a cutoff value of 1. So that could be a possible explanation why in your study you did not detect the significant risk with oversizing and we detected it.
DR RALPH S. MOSCA (New York, NY): That was a very interesting study. Your last slide depicts the probable causes of reduced longevity of the conduits. These included compression under the sternum in the heterotopic location combined with abnormal flow characteristics through the conduit. Given these factors, did you see any difference in longevity of conduits when used in a Ross procedure as opposed to others?
DR ASKOVICH: We used diagnosis as a multicategorical variable, and we modeled it with eight dummy variables. So when you do it that way, then one of your diagnoses becomes the reference. We saw some differences, for instance, with the Ross compared to the reference. And we are aware that is one of the limitations of the study, that we did include the Ross in our study population. But because we treated it as separate, and there are also a small number of patients with this diagnosis, that it did not make a huge impact on our conclusions.
DR HAWKINS: One quick thing. To answer Tara’s question, I think that in the recent study that the CHHS published in October and Tara presented at AATS last year, we did look at very small conduits in dividing these up into quartiles. We did not do it that way in this.
First of all, we had very, very few, less than a z score of 1. So there weren’t any to analyze down below 1 because I think that is probably an unacceptable size most of the time. And I think otherwise this dovetails very nicely, almost identical with the data out in the CHHS study, which said a z score greater than 3 was actually a risk factor for earlier explantation or earlier reintervention.
Our number happened to fall out at 2.8, which is where the quartiles came out, and that was the division between the third and the fourth quartile. So I think it actually is almost identical to that data in the recent CHHS study.
DR CARL L. BACKER (Chicago, IL): I think this is a great confirmatory study to the recent Congenital Heart Surgeons’ Society report. I have two questions, one relating to the paper from Switzerland. First, do you have any data on ABO compatibility?
DR ASKOVICH: We have not, unfortunately, collected the ABO data, but that would be something worthwhile.
DR BACKER: So when you implant these valves, you are not selecting them based on ABO compatibility? Dr Hawkins, do you want to answer that? When they ask you what valve you want, do you ask if it is ABO compatible, or do you just take what they offer you?
DR HAWKINS: Well, probably a third of the time we have no ABO information. For the other two thirds that I do have the information, I generally try to do match ABO if I can. I don’t know that it matters, but on the outside chance that it does, I want to be on the correct side. So I do try to do it if it’s at all possible. I would say probably overall, maybe 40% of our patients will receive an ABO compatible, a third I don’t know, and whatever the math is on the rest.
DR BACKER: John, one other question while you are still up. Do you use any postoperative therapy, aspirin, Indocin, cyclosporine, anything else in the postoperative period to treat these patients?
DR HAWKINS: No. No, we do not. We have done one randomized trial of azathioprine and another pilot study with CellCept, both of which suppressed the antibody formation, but we could not demonstrate any difference in valve durability.
DR BACKER: I would like to poll the audience. How many surgeons are putting their patients on some sort of antiinflammatory after placement of a homograft?
(A show of hands.)
And how many are not?
(A show of hands.)
So it looks like about 25% to 30% of the surgeons are putting their patients on some sort of antiinflammatory agent and 70% are not. Of those surgeons that are putting their patients on some form of antiinflammatory agent, does anyone want to come to the microphone? One of those hands that went up was Dr Brown.
DR JOHN W. BROWN (Indianapolis, IN): Aspirin and ibuprofen. The ibuprofen is at 10 mg/kg 3 times a day.
DR BACKER: Is that for life or for—
DR BROWN: For 3 months.
DR BACKER: For 3 months. And how about the aspirin?
DR BROWN: Usually a baby aspirin, depending on the size of the child, but again 10 mg/kg.
DR BACKER: And that’s for 3 months or for life?
DR BROWN: Three months.
DR BACKER: So 3 months of treatment with Indocin and aspirin, all your patients?
DR BROWN: Yes.
DR BACKER: And that’s the Contegra and the homograft?
DR BROWN: We don’t use the ibuprofen for the Contegra. We just use aspirin.
DR BACKER: Thank you very much. That was an excellent discussion.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. P.V. Zachariah, F. A. Pigula, J. E. Mayer Jr, and D. B. McElhinney Right ventricle to pulmonary artery conduit augmentation compared with replacement in young children. Ann. Thorac. Surg., August 1, 2009; 88(2): 574 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ando and Y. Takahashi Ten-year experience with handmade trileaflet polytetrafluoroethylene valved conduit used for pulmonary reconstruction. J. Thorac. Cardiovasc. Surg., January 1, 2009; 137(1): 124 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Lurz, L. Coats, S. Khambadkone, J. Nordmeyer, Y. Boudjemline, S. Schievano, V. Muthurangu, T. Y. Lee, G. Parenzan, G. Derrick, et al. Percutaneous Pulmonary Valve Implantation: Impact of Evolving Technology and Learning Curve on Clinical Outcome Circulation, April 15, 2008; 117(15): 1964 - 1972. [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 |
