Conference Coverage

EEG asymmetry predicts poor pediatric ECMO outcomes



– Children who have background EEG asymmetry while on extracorporeal membrane oxygenation (ECMO) have worse outcomes even after adjustment for recent cardiac arrest and EEG suppression, according to a review of 41 children treated at Washington University, St. Louis.

Dr. Kristin Guilliams, assistant professor of pediatric critical care medicine, Washington University, St. Louis

Dr. Kristin Guilliams

ECMO is a last-ditch heart/lung bypass for patients near death, be it from infection, trauma, cardiac abnormalities, or any other issue. Children can be on it for days or weeks while problems are addressed and the body attempts to recover. Sometimes ECMO works, and children make a remarkable recovery, but other times they die or are left with severe disabilities, and no one really knows why.

Because of this, the investigators in this review sought to identify predictors of poor outcomes with an eye toward identifying modifiable risk factors, said senior investigator Kristin Guilliams, MD, an assistant professor of pediatric critical care medicine.

“We are trying to figure out why some kids do fantastically, and others don’t. We were looking at whether EEG can give us any clues and new ways to think about modifiable risk factors so that every kid rescued by ECMO can go back to their normal life,” she said at the American Neurological Association annual meeting.

The 41 children had an EEG within a day or 2 of starting ECMO; 22 did well, but 19 had bad outcomes, defined in the study as either dying in the hospital or being discharged with a Functional Status Score above 12, meaning mild dysfunction across six domains or more severe disability in particular ones.

The finding that all four children with EEG suppression – overall low brain activity – did poorly was not surprising, but the fact that EEG background asymmetry – one side of the brain being much less active than the other or giving different signals – in five children predicted poor outcomes, even after adjustment for cardiac arrest and overall suppression, was “a big surprise,” Dr. Guilliams said (odds ratio, 29.3; 95% confidence interval, 2.2-398.3; P = .003).

“The asymmetry tells me that we need to look more closely into brain blood flow patterns on ECMO,” she said. There might be a way to change delivery that could help, but “it’s not obvious right now.” The issue warrants further investigation, Dr. Guilliams said.

Twelve children had ECMO during chest compressions for cardiac arrest, which as expected, also predicted poor outcomes (OR, 9.5; 95% CI 1.6-58.2; P = .008).

Neuroimaging was available for 34 children. Abnormalities (n = 13; P = .2), including ischemia (n = 8; P = .1), hemorrhage (n = 8; P = .06), and seizures (n = 4; P = .2) did not predict poor outcomes, nor did sex, age, and mode of ECMO delivery (veno-arterial versus veno-venous).

As of about a year ago, EEGs at the university are now standard for children on ECMO, with special software to pick out asymmetries. “We are paying more attention to” EEGs, Dr. Guilliams said.

Children were a median of about 10 years old, and subjects were at least 1 year old. There were about equal numbers of boys and girls; 25 children were alive at discharge.

There was no external funding, and Dr. Guilliams didn’t have any disclosures.

Next Article: