Discussion

Article Outline

• Copyright

Dr Edward L. Bove (Ann Arbor, Mich). I would like to thank Dr Hickey and his coauthors from the CHSS for an excellent study with a new, at least to me, statistical evaluation tool. In this presentation, the authors have asked the age-old question: “Is a high-risk BVR better than a low-risk UVR approach?” Despite the substantial decline in early mortality for univentricular approaches, most centers still desire to push borderline patients to a BVR, believing that it is safer and affords better long-term outcome. The authors have documented that the first assumption is not always correct, although the second still remains uncertain. Further complicating this decision process is the realization that, at least for the types of patients in this particular analysis, namely, those with critical LVOT obstruction, the single-ventricle approach perhaps offers the best outcomes among many variants of hypoplastic left heart syndrome.

The authors performed an exhaustive analysis of multiple risk factors for death for both UVR and BVR and then formed a UVR-SA tool that allowed them to reanalyze predicted survival for those whose repair was either concordant or discordant using predictive formulas previously published. This model was further refined and used to predict an optimal survival path for the patients in the analysis.

Although there are multiple intriguing findings among these data, the authors found that discordant decision-making, namely, pursuing BVR when UVR had a higher predicted survival or vice versa, costs lives. More important, a discordant BVR decision is more costly than a UVR one. The inference is that the marginal left heart, for whatever reason, has little or no room to compensate and transfers all the risk up front. Equally important, although not addressed here, is the potential that many BVR survivors may also face repeated valve replacement procedures and pulmonary hypertension in their later years. I have 2 questions for Dr. Hickey.

It seems from the analysis of their paper that a smaller mitral valve annulus was a risk factor only for UVR. Others have reported that the mitral valve size was strongly predictive of outcome for BVR. Could the authors provide some insight regarding the influence of the mitral valve, either size or function, on survival for BVR patients?

Second, do the authors believe that the more liberal use of Ross–Konno procedures, either with or without arch repair as needed or even resection of EFE, might improve survival for BVR?

I thank Dr. Hickey for an excellent presentation.

Dr. Hickey. Thank you very much, Dr. Bove, for your comments and your questions. The issue with the mitral valve annulus is intriguing, because, as you point out, it was identified as a risk factor for death following univentricular repair, which is clearly a surprise. We explored this further, and our conclusions are as follows.

The mitral annular Z scores as a risk factor for univentricular repair do not correlate with either mitral stenosis or mitral regurgitation in this cohort. Therefore, this variable, although representing the mitral annulus, is not necessarily a functional variable. It is a morphological dimension that correlates strongly with overall left-sided structural hypoplasia. So it correlates with dimensions of the LV outflow tract and the ventricular size and ventricular hypoplasia. So it is in some ways misleading that it is the mitral annulus. It is, we feel, a variable representing the overall dimension and degree of left-sided hypoplasia.

So why is it not a risk factor for biventricular repair? Well, the answer to that is that we obviously looked at and explored the potential for both functional regurgitation or stenosis as a risk factor after biventricular repair, including mitral morphology and dimensions. All of these variables were overshadowed in our statistical analysis by more robust variables—namely, the degree of LV dysfunction—and particularly the severity of endocardial fibroelastosis or thickening, which is an extremely strong predictor of poor prognosis after biventricular repair. So although we don’t deny that mitral valve variables have been reported as important in other series, here they have been overshadowed by other functional and morphologic left-sided variables.

In regard to your second question and the use of neonatal Ross–Konno procedures, we have been looking at the index procedure as the intention to treat. Invariably in our series this was either through balloon aortic valvotomy or surgical valvotomy, with only a minority of more complex repairs (namely, 2 Yasui and 5 index Ross–Konno procedures). A number of other Ross–Konno procedures were undertaken subsequently as repeat interventions—in fact, a total of 30. Overall outcome for these was 66% survival at 5 years. Now, of course there are reports of further improved survival with Ross–Konno procedures, and this may translate into a further bias towards management. However, at the same time as that survival will have improved with Ross–Konno procedures, survival also will have improved over the intervening decade with staged Norwood palliation, for example. So the only way of actually determining whether index Ross–Konno results in improved BVR survival is to undertake a contemporary investigation in which we enroll a higher proportion of Ross–Konno procedures and complex repairs. In fact, that is an investigation that the CHSS are looking at now with our latest LV outflow tract series.