Validation of a Clinical Tool to Predict Neurological Outcomes in Critically Ill Children-A Prospective Observational Study.

OBJECTIVES: To evaluate the performance of the empiric tool by Gupta et al. in predicting neurological outcomes in children admitted to the pediatric intensive care unit (PICU) and to evaluate the association of biomarkers S100B and NSE with neurological outcomes. METHODS: This prospective observational study was conducted in 163 critically ill children aged 2 mo to 17 y admitted to the PICU from June 2020 to July 2021. The authors used the prediction tool developed by Gupta et al.; the tool was applied at admission and at PICU discharge/death. Samples for NSE and S100B were collected at admission and discharge. The performance of the new tool was assessed through discrimination and calibration. Risk factors for "unfavorable outcomes" (decline in PCPC score by > 1) were evaluated by multivariate analysis. RESULTS: The PICU mortality was 28% (n = 45). When the tool developed by Gupta et al. was used at the time of admission, favorable neurological outcomes were predicted for 69% (112) children. The area under the curve for the new tool at admission was 0.72 and at discharge/death it was 0.99, and the calibration was excellent at both time points. Independent factors associated with unfavorable neurological outcomes were higher PCPC scores and organ failure. As the number of samples processed for NSE and S100B was less, statistical analysis was not attempted. CONCLUSIONS: The new tool by Gupta et al. has good discrimination, calibration, sensitivity, and specificity and can be used as a prediction tool. NSE and S100B are promising biomarkers and need further evaluation.