Comparative study of different scoring system on the prognosis of infant traumatic brain injury

Objective To compare the predictive efficiency of different scoring systems including TraumaInfant Neurologic Score(TINS), Marshall CT score, Rotterdam CT score and Helsinki CT scorefor thelong⁃term prognosis of infant (≤3yearsold) traumatic brain injury (TBI). Methods Clinical records and imaging data of 37 cases with infant TBI (admitted to Department of Neurosurgery, Xiangya Hospital, Central South University from January 2012 to December 2017) were analyzed retrospectively. Glasgow Outcome Scale (GOS) was used to evaluate the prognosis 6 months after TBI. TINS score, Marshall CT score, Rotterdam CT score and Helsinki CT score were used to predict the prognosis. The Receiver Operating Characteristics Curve (ROC) comparing among TINS, Marshall CT, Rotterdam CT and Helsinki CT scores was performed to predict the prognosis of infant TBI. Results Among 37 cases, 19 cases (51.35%) had good prognosis (GOSgrade5), 11 cases (29.73%) had slight disability (GOSgrade4), 6 cases (16.22%) had severe disability (GOSgrade3), and one case (2.70%) died (GOSgrade1). The areas under the curve (AUC) of the TINS, Marshall CT, Rotterdam CT and Helsinki CT scores at 6⁃month after injury in GOS grade 1-3 group were 0.774 (95%CI: 0.539-1.000, P =0.026), 0.807 (95%CI: 0.668-0.946, P = 0.012), 0.748 (95%CI: 0.535-0.960, P=0.044) and 0.740 (95%CI: 0.533-0.948, P=0.050), respectively. Conclusions TINS, Marshall CT, Rotterdam CT and Helsinki CT scores have good evaluation efficiency on the prognosis of infant TBI in long⁃term. In comparison, Marshall CT score is more advantageous in evaluating the prognosis of infant TBI 6 months after discharge. DOI：10.3969/j.issn.1672⁃6731.2020.07.010