334 - Use of Oxygen Saturation Index in predicting need for ECMO and death in Congenital Diaphragmatic Hernia

Poster Number: 334
Publication Number: 334.345

Srirupa Hari Gopal, Baylor College of Medicine, Houston, TX, United States; Kelly F. Martinek, Baylor College of Medicine, Houston, TX, United States; Amanda Holmes, Texas Children's Hospital, Brookshire, TX, United States; Joseph Hagan, Baylor College of Medicine, Houston, TX, United States; Caraciolo J.. Fernandes, Baylor College of Medicine, Houston, TX, United States

Background: The oxygenation index (OI) is used as a metric to quantify the severity of hypoxic-respiratory failure in neonates, including those with congenital diaphragmatic hernia (CDH). It requires invasive arterial blood sampling to determine the partial pressure of oxygen to be used in the formula. The oxygen saturation index (OSI) uses oxygen saturation (SpO2), removes the dependency on an arterial access and can monitor a neonate respiratory status on a continuous basis.

Objective: Our objective with the current study was to investigate specific oxygen saturation index-OSI (calculated as the FiO2 × mean airway pressure (MAP) × 100)/O2 saturation) values that predicted need for ECMO and death in the CDH population.

Design/Methods: This is a single center retrospective study of infants with CDH admitted to a tertiary-care level 4 NICU from January 2011 to December 2021. Demographic, maternal and birth data, vital sign, ventilator data in the first 120 hours of life and post-natal outcomes for need for ECMO and mortality were extracted for 180 patients.  Logistic regression analysis was used to investigate the association of maximum pre-ductal OSI values (OSI_max) at 6, 12, 24 and 48 hours of life (HOL) with death and ECMO.  Receiver operating characteristic (ROC) curve analysis was used to identify the OSI_max cutoff at each HOL that yielded the best predictive accuracy to predict death or ECMO.

Results: Patient characteristics are summarized in Table 1. OSImax values were statistically significant at 6-,12-,24- and 48 HOL for predicting need for ECMO and death.  An OSImax cut off of >13 at 6 HOL most predicted the need for ECMO and death in patients with CDH. The OSI_max cut off values for each time point and their corresponding sensitivity and specificity in predicting need for ECMO and death are shown in Table 2.

Conclusion(s):

We report that an OSI_max cut-off > 13 in the first 6 HOL predicted the need for ECMO and death in the CDH population. Monitoring OSI_max could assist clinicians in smaller NICUs, or in the absence of an arterial access, to identify CDH infants for transfer to a higher level of care.