Early Blood Biomarkers and MRI Injury After Cardiac Arrest: Secondary Analysis of the 2017-2020 "Personalized Outcomes After Child Cardiac Arrest" Study.

Objectives: Brain MRI is used to inform prognosis of pediatric cardiac arrest (CA). We analyzed the association between early levels of four brain injury biomarkers and pattern of brain injury on MRI.

Design, setting, and patients: This secondary analysis of a multicenter prospective cohort study in 14 U.S. hospitals (from May 16, 2017, to August 19, 2020) recruited children 48 hours to 17 years old who were resuscitated after CA and had a brain MRI within 14 days postarrest.

Interventions: None.

Measurements and main results: Brain MRI injury score was calculated as a sum of T2- and diffusion-weighted imaging lesions. We used the Kruskal-Wallis test to compare maximum biomarker values on days 1-3 between three categories of MRI injury severity (i.e., no injury, mild-moderate injury, and severe injury). Maximum neurofilament light chain (NfL), tubulin-associated unit, glial fibrillary acidic protein, and ubiquitin C-terminal hydrolase L1 levels were associated with severity of total injury, gray matter injury, and white matter injury. Using logistic regression, individual biomarker levels were associated with presence of injury on MRI after adjusting for age, presence of congenital heart disease, and severity of illness using Pediatric Index of Mortality 3 score. Of 40 patients with injury on MRI and 1-year outcome data, median (interquartile range [IQR]) NfL levels were higher in the 15 patients who died compared with the 21 patients with favorable outcome (7.10 pg/mL [IQR, 5.94-7.51 pg/mL] vs. 5.10 pg/mL [IQR, 4.10-5.94 pg/mL]; log transformed; p < 0.001), but we failed to identify a difference in levels between those with unfavorable outcome (Vineland Adaptive Behavior Score < 70, n = 4) vs. favorable outcome.

Conclusions: Blood biomarkers measured early after injury are associated with MRI injury and may provide additional information for prognostication when incorporated in a multimodal evaluation.