Utility of Combining Frailty and Comorbid Disease Indices to Better Predict Outcomes following Craniotomy for Pediatric Primary Brain Tumors

Alan Nguyen; Michelot Michel; Shane Shahrestani; Andre Boyke; Catherine M Garcia; Simon Menaker; Moise Danielpour; David Bonda

doi:10.1159/000548771

Utility of Combining Frailty and Comorbid Disease Indices to Better Predict Outcomes following Craniotomy for Pediatric Primary Brain Tumors

Pediatr Neurosurg. 2025 Oct 3:1-9. doi: 10.1159/000548771. Online ahead of print.

Authors

Alan Nguyen^{1

2}, Michelot Michel^{2

3}, Shane Shahrestani², Andre Boyke², Catherine M Garcia², Simon Menaker², Moise Danielpour², David Bonda²

Affiliations

¹ College of Medicine, Charles R. Drew University of Medicine and Science, Los Angeles, California, USA.
² Department of Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, California, USA.
³ Department of Neurosurgery, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.

PMID: 41042723
DOI: 10.1159/000548771

Abstract

Introduction: There are no predictive outcome scales that have been validated in pediatric patients with brain tumors. An index can help identify children with increased risk for negative postoperative results. The Johns Hopkins Adjusted Clinical Groups (JHACG) frailty and the Elixhauser Comorbidity Index (ECI) have been used independently in adult brain tumor patients to identify patients who are at an increased risk for detrimental outcomes. We investigated whether JHACG and ECI can better predict hospital length of stay (LOS), nonroutine discharge, and 1-year readmission in pediatric patients undergoing craniotomy for primary brain tumors.

Methods: The Nationwide Readmissions Database was queried for pediatric brain tumor resections between 2016 and 2019. In total, 237 and 1,235 patients with benign and malignant tumors were identified, respectively. Frailty, ECI, and Frailty+ECI were assessed as predictors using generalized linear mixed-effects models. Receiver operating characteristic curves evaluated predictive performance.

Results: Frailty+ECI, frailty, and ECI scores similarly predicted hospital LOS, nonroutine discharge, and 1-year readmission in the benign tumor cohort. In the malignant cohort, Frailty+ECI (area under the curve [AUC] 0.895) outperformed frailty alone (AUC 0.742, p = 0.001) but performed similar to ECI score alone (AUC 0.893, p = 0.438) in predicting hospital LOS. Concerning nonroutine discharge prediction, Frailty+ECI (AUC 0.871) also outperformed frailty alone (AUC 0.744, p = 0.04) while performing similarly to ECI score alone (AUC 0.869, p = 0.871). All indices performed in a similar way to predict 1-year readmission in this cohort.

Conclusion: Our study showed that Frailty+ECI demonstrated a robust ability to predict hospital LOS and nonroutine discharge disposition in pediatric patients undergoing malignant brain tumor resection. These findings suggest that combining these indices may improve the prediction of postoperative outcomes in this population. While further studies are warranted, these findings can be used as a risk assessment index to coordinate care plans with the patient and their family after an operation.

Keywords: Elixhauser Comorbidity Index; Frailty; Glioma; Johns Hopkins Adjusted Clinical Groups; Risk assessments.