Background/purpose: The incidence of congenital diaphragmatic hernia (CDH) approximates 1 in 3000 births, with mortality rates up to 50%. The ability to accurately and easily predict the outcomes of these infants could be a valuable management tool. The purpose of this study was to develop and validate a simplified clinical method for predicting survival outcomes in infants born with CDH.
Methods: The Wilford Hall/Santa Rosa clinical prediction formula (WHSR(PF) = highest PaO2 - highest PCO2) was generated from arterial blood gas values obtained during the initial 24 hours of life, but before surgical repair or extracorporeal membrane oxygenation, in a local group of infants with CDH identified by prospective and retrospective review. The WHSR(PF) was validated using a comparative group from the Congenital Diaphragmatic Hernia Study Group (CHDSG). Bivariate, multivariable, and area under the receiver operating curve (AUC) analysis was performed using SigmaStat and SPSS statistical programs (SPSS, Chicago, Ill).
Results: As initially developed from the local data, the WHSR(PF) had a positive predictive value (PPV) of 82%, a negative predictive value of 88% and AUC of 0.87. When validated against the CDHSG data, the positive predictive value was 83%, negative predictive value was 66%, and AUC 0.79. Area under the receiver operating curve analysis by the previously published CDHSG predictive equation was 0.76.
Conclusion: This novel formula is an easy to apply clinical tool with similar or better predictive abilities compared to previous methods of predicting survival in infants born with CDH. Currently, no method appears to have sufficient clinical accuracy for predicting the outcome of an individual infant with CDH. Further studies are indicated.