Abstract

Abstract INTRODUCTION Prognostic factors in paediatric traumatic brain injury (TBI) remain unclear. This study assessed the prognostic value of leucocytosis in paediatric isolated TBI METHODS A total of 106 paediatric isolated TBI patients treated at our institution between June 2008 and June 2016 were retrospectively analysed. Data collected included initial blood leucocyte count (WCC), presenting Glasgow Coma Score (GCS), computed tomography (CT), hospital stay, and Paediatric Cerebral Performance Category Scale (PCPCS), and their relationship with WCC were statistically evaluated. RESULTS A total of 39, 37, and 30 patients had severe, moderate, and mild TBI, respectively. For patients with GCS 3-8, 9-13, and 14-15, WCC was 20, 15.9, and 10.7 × 109/L, respectively. Differences in WCC were significant between the different GCS groups (P < .01). WCC was greatest and smallest in patients with CT findings demonstrating significant mass effect (Marshall grading V-VI) and minimal mass effect (Marshall grading I-II), respectively (P < .05). Length of hospital stay and extent of midline shift on CT each significantly correlated with WCC (P < .05). In addition, higher WCC counts were associated with a poorer 6-mo PCPCS rating (P < .05). Multivariate regression analysis revealed a cut-off leucocyte count of 16.1 × 109/L, and Neutrophil-to-Lymphocyte ratio (NLR) of 5.2, above which GCS, CT findings, length of hospital stay and PCPCS were less favourable. Application of the International Mission on Prognosis and Analysis of randomized Controlled Trials in TBI (IMPACT) adult TBI prediction model to our paediatric cohort, using area under the operating curve (AUROC) and coefficient analyses, demonstrated increased accuracy with incorporation of WCC count as a risk factor. CONCLUSION High leucocyte count (>16.1 × 109/L) and NLR > 5.2 each have a predictive value for poor GCS, severe CT findings, lengthy hospital stay and poor PCPCS in isolated paediatric TBI. Incorporating initial leucocyte count into TBI prediction models may increase the accuracy of prognostication.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call