130 Development of an Integrated Risk Scale for Prediction of Shunt Placement After Neonatal Intraventricular Hemorrhage. (1st April 2022)

• Record Type: Journal Article
• Authors: Jin, Michael C.; Parker, Jonathon J.; Rodrigues, Adrian; Colon, Gabriela R.; Garcia, Cesar A.; Mahaney, Kelly B.; Grant, Gerald A.; Prolo, Laura M.
• Is Part Of: Neurosurgery. Volume 68(2022)Supplement 1
• Page Start: 37
• Page End: 38
• Publication Date: 2022-04-01
• Subjects: Nervous system -- Surgery -- Periodicals; 617.48005
• DOI: 10.1227/NEU.0000000000001880_130
• Languages: English
• ISSNs: 0148-396X

Abstract

Abstract : INTRODUCTION: Neonatal intraventricular hemorrhage (IVH) is a major source of mortality and morbidity in infants, particularly following premature birth. Even after the acute phase, post-hemorrhagic hydrocephalus is a long-term complication, frequently requiring permanent ventriculoperitoneal shunt placement (VPS). Currently, there are no risk classification methods integrating the constellation of clinical data to predict short- and long-term prognosis in neonatal IVH. METHODS: Neonates with IVH were identified from the Optum Clinformatics DataMart administrative claims database using ICD-9/10 codes. Matched maternal-childbirth characteristics were obtained. Primary endpoints were short-term (30-day) mortality and long-term VPS placement. Classification of short-term mortality risk was evaluated using five different machine learning approaches and the best-performing method was validated using a withheld validation subset. Prediction of long-term shunt risk was performed using a multivariable Cox regression model with stepwise variable selection, which was subsequently converted to an easily applied integer risk scale. RESULTS: A total of 5, 926 neonate-maternal pairs were included. Empiric 30-day neonatal mortality risk was 10.9% across all IVH grades and highest among grade IV IVH (34.3%). Among those surviving beyond 30 days, actuarial 12-month risk of shunt placement was 5.4% across all IVH grades and 31.3% for grade IV IVH. The optimal short-term risk classifier was a random forest model achieving an AUROC of 0.882 with important predictors ranging from gestational age to diverse comorbid medical conditions. The final multivariable Cox regression model for long-term shunt risk stratification following stepwise feature selection included IVH grade, respiratory distress syndrome, disseminated intravascular coagulation, and maternal preeclampsia/eclampsia. An integer risk scale termed the Shunt Prediction After IVH in Neonates (SPAIN) scale was developed from these four features which, evaluated on withheld validation cases, demonstrated improved risk stratification compared to IVH grade alone (Harrell's concordance index 0.869 vs 0.852). CONCLUSION: We developed an integrated risk classification approach to anticipate short-term mortality and long-term shunt risk. Application of such approaches may improve outcome prognostication and identification of higher risk individuals warranting careful surveillance and early intervention.