The Role of ECMO in Neonates with CDH

Abstract

Source: The Congenital Diaphragmatic Hernia Study Group. Does extracorporeal membrane oxygenation improve survival in neonates with congenital diaphragmatic hernia? J Ped Surg. 1999;34(5):720–724; discussion 724–725.While the benefit from extracorporeal membrane oxygenation (ECMO) in improving survival of neonates with congenital diaphragmatic hernia (CDH) has never been clearly demonstrated with a randomized controlled trial, the lack of alternative therapy and the high mortality rate (approximately one-third) has led to widespread use of this aggressive and costly therapy. This study analyzes data from a CDH registry involving 65 centers with and without ECMO capability to determine if ECMO improved survival after adjusting for overall risk of mortality. Data were available for 632 neonates with birth weight (BW) >2kg, of whom 289 received ECMO. Patients were divided into quintiles based on their predicted risk of death (using BW and 5 minute Apgar). After adjusting for risk of death, ECMO was associated with significantly higher mortality in the 2 quintiles with the lowest risk of death, and lower mortality in the quintile with the highest risk of death, ≥80%. The first 2 groups accounted for 68% of the patients while the latter group for 5%. There was no significant association between ECMO and mortality in the middle 2 quintiles. The authors conclude that ECMO makes a difference in survival for infants at high risk of death. The apparent association of ECMO with lower survival in infants in the 2 lowest quintiles is unexplained, but may be due to a failure of the predictive model to adequately adjust for differences between infants receiving and not receiving ECMO in these groups.This article by the Congenital Diaphragmatic Hernia Group is important because it is one of the first publications from this large, voluntary collaboration of physicians who follow children with diaphragmatic hernia. The important message in this study was that ECMO has value in improving survival for infants with high-risk diaphragmatic hernia. With that in mind, some consideration should probably be given to delivery of infants with congenital diaphragmatic hernia close to an ECMO center.The ability to tell the effect of ECMO on improved survival rests on whether or not the infants with CDH who received ECMO and those who did not were comparable. The authors attempt to do so by stratifying on predicted risk of death (based on birth weight and Apgar), and then comparing survival within those strata. The authors and the commentator point out that ECMO appears to be associated with better survival in the highest risk group. However, survival in the two lowest risk groups was actually poorer in those infants receiving ECMO. The problems in sorting out the effect of ECMO are precisely those seen in studies which are not prospective, controlled clinical trials. In all likelihood, there are underlying differences (eg, degree of congenital lung dysplasia) between the two groups which were not adequately adjusted for in the model. Furthermore, even if the results in the highest risk group were real, and those in the lowest risk groups were not, the overall survival would only increase by 1.4% since the high risk group only accounts for 5% of CDH patients. The optimal use of ECMO in CDH is not yet apparent and almost certainly can only be ascertained by further detailed, analytic reports from this multi-center collaborative study group.