Decreased Brain Volumes and Infants With Congenital Heart Disease Undergoing Venoarterial Extracorporeal Membrane Oxygenation.

Objectives: The aims of this study were to: i) determine the spectrum of brain injury and ii) compare brain volumes between pre- and postoperative brain MRI in the infants receiving extracorporeal membrane oxygenation compared with those who did not require extracorporeal membrane oxygenation.

Design: Cohort study of infants with D-transposition of the great arteries or single ventricle physiology. Brain volume (cm) was measured using a segmentation of a volumetric T1-weighted gradient echo sequence. Brain imaging findings (intraventricular hemorrhage, white matter injuries, and stroke) were analyzed with respect to known clinical risk factors for brain injury and adverse neurodevelopmental outcomes. Clinical factors were collected by retrospective chart review. The association between brain volume and extracorporeal membrane oxygenation was evaluated using generalized estimating equations to account for repeated measures.

Setting: Prospective and single-centered study.

Patients: One hundred nine infants (median gestational age, 39.1 wk) with D-transposition of the great arteries (n = 77) or single ventricle physiology (n = 32) were studied pre- and postoperatively with MRI as per clinical protocol.

Interventions: None.

Measurements And Main Results: Of the 28 infants (26%) receiving extracorporeal membrane oxygenation, 19 (68%) were supported with extracorporeal membrane oxygenation once, and nine (32%) were supported 2-4 times. On postoperative MRI, new white matter injury was found in only five (17%) of the extracorporeal membrane oxygenation infants versus 40 (49%) in the non-extracorporeal membrane oxygenation group (p = 0.073). The rate of stroke (9% vs 10%), intraventricular hemorrhage (24% vs 29%), and hypoxic ischemia (3% vs 14%) did not differ between the non-extracorporeal membrane oxygenation and extracorporeal membrane oxygenation groups (all p > 0.5). Accounting for D-transposition of the great arteries or single ventricle physiology diagnosis, infants requiring extracorporeal membrane oxygenation had slower brain volume with single (β = -1.67) or multiple extracorporeal membrane oxygenation runs ([β = -6.54]; overall interaction p = 0.012).

Conclusions: Patients with d-transposition of the great arteries or single ventricle physiology undergoing extracorporeal membrane oxygenation at our center have a similar incidence of brain injury but more significant impairment of perioperative brain volumes than those not requiring extracorporeal membrane oxygenation.