305 - Risk of Intracranial Hemorrhage or Death during ECMO in Neonates with Hypoxic-ischemic Encephalopathy

Poster Number: 305
Publication Number: 305.3129

Background: Neonatal hypoxic-ischemic encephalopathy (HIE) is an important cause of morbidity and mortality, and about 4-9% will require extracorporeal membrane oxygenation (ECMO) for respiratory and circulatory failure. Neonates with HIE requiring ECMO have a theoretically increased risk of intracranial hemorrhage (ICH) due to liver dysfunction, coagulopathy during therapeutic hypothermia (TH), and differences in cerebral autoregulation and perfusion. Limited data currently exists to support this theory.

Objective: Our objective is to determine whether a diagnosis of neonatal HIE increases the risk of ICH and mortality during ECMO. Our primary outcome is brain ultrasound or MRI findings of significant ICH and in-hospital death.

Design/Methods: We conducted a retrospective chart review of all neonates admitted to our hospital who required ECMO within the first 7 days of life in a 10-year period. Exclusion criteria were neonates with congenital diaphragmatic hernia and abnormal cranial ultrasound before ECMO initiation. We defined significant ICH as Grade III or IV on brain ultrasound or MRI. A logistic regression analysis of additional clinical risk factors for ICH or death related to ECMO was performed.

Results: A total of 82 neonates were included in the study, 30 with HIE and 52 without HIE. The proportion of neonates with HIE requiring ECMO cannulation varied over a ten-year period (Figure 1).
Significant ICH or death occurred more frequently in neonates with HIE, however, this was not statistically significant. Combined outcomes of ICH and death occurred more frequently in neonates with HIE, this was also not statistically significant (Table 1).
Reduced multivariable regression analysis of a variety of clinical factors was performed (Table 2). Younger gestational age and higher pre-ECMO INR value increased the odds of ICH. Older age at cannulation, and greater peak INR value during ECMO increased the odds of in-hospital death. Later gestational age at cannulation showed increased odds of in-hospital death (OR 4.2, [95% CI 1.05-16], p=0.04), but not death during ECMO (OR 0.85 [95% CI 0.5-1.3], p=0.5). Increased duration of ECMO showed decreased odds of death.

Conclusion(s): Although HIE trended towards worse outcomes on ECMO, this was not statistically significant. Gestational age did not have a significant impact on survival during ECMO, but later gestation increased the odds of in-hospital death. This may be due to other causes not related to the patient’s ECMO course. Clinical factors such as degree of coagulopathy and age at the time of cannulation were better indicators of morbidity and mortality on ECMO.