Prognostic value of leukocytosis in pediatric traumatic brain injury

OBJECTIVE The purpose of this study was to assess leukocytosis and its prognostic value in pediatric isolated traumatic brain injury (TBI). METHODS Two hundred one children with isolated TBI admitted to the authors' institution between June 2006 and June 2018 were prospectively followed and their data retrospectively analyzed. Initial blood leukocyte count (i.e., white cell count [WCC]), Glasgow Coma Scale (GCS) score, CT scans, duration of hospital stay, and Pediatric Cerebral Performance Category Scale (PCPCS) scores were analyzed. RESULTS The mean age was 4.2 years (range 0.2-16 years). Seventy-four, 70, and 57 patients had severe (GCS score 3-8), moderate (GCS score 9-13), and mild (GCS score 14-15) TBI, respectively, with associated WCC of 20, 15.9, and 10.7 x 10(9)/L and neutrophil counts of 15.6, 11.3, and 6.1 x 10(9)/L, respectively (p < 0.01). Higher WCC and neutrophil counts were demonstrated in patients with increased intracranial mass effect on CT, longer hospital stay, and worse 6-month PCPCS score (p < 0.05). Multivariate regression revealed a cutoff leukocyte count of 16.1 x 10(9)/L, neutrophil count of 11.9 x 10(9)/L, and neutrophil-to-lymphocyte ratio (NLR) of 5.2, above which length of hospital stay and PCPCS scores were less favorable. Furthermore, NLR was the second most important independent risk factor for a poor outcome (after GCS score). The IMPACT (International Mission for Prognosis and Analysis of Clinical Trials in TBI) adult TBI prediction model applied to this pediatric cohort demonstrated increased accuracy when WCC was incorporated as a risk factor. CONCLUSIONS In the largest and first prospective study of isolated pediatric head injury to date, the authors have demonstrated that WCC > 16.1 x 10(9)/L, neutrophil count > 11.9 x 10(9)/L and NLR > 5.2 each have predictive value for lengthy hospital stay and poor PCPCS scores, and NLR is an independent risk factor for poor outcome. Incorporating the initial leukocyte count into TBI prediction models may improve prognostication.