Rotterdam computed tomography score as a predictor of early death among patients with traumatic brain injury at a tertiary hospital in Kampala, Uganda: A prospective study

Abstract

[Uncorrected Proof]

Background: Noncontrast computed tomography (CT) plays a crucial role in the assessment and triage of traumatic brain injured (TBI) patients. This study aimed to determine whether the Rotterdam CT score (RS) had good predictive value for short-term mortality among TBI patients in Uganda.

Methods: This was a hospital-based, prospective study of TBI treated in a tertiary, private hospital in Kampala, Uganda. A total of 108 TBI patients were consecutively enrolled from September 2017 through May 2018 and followed up from admission to either in-hospital death or discharge. Noncontrast CT brain imaging was conducted, and parameters of CT findings were used to calculate the RS. An area under the receiver operating characteristic (AUROC) plot of sensitivity vs specificity was generated to determine the discriminative power of the RS to predict in-hospital mortality.

Results: In total, 38% of TBI patients were aged 20 to 29 years, 25.9 % were aged 30 to 39 years, and 11% were over 50 years of age. Most patients were male (81.4%), while 19% were female. Six deaths occurred among study participants. Study mortality (5.6%) was lower than that predicted by the RS (11.1%). The AUROC for the RS was 68% (95% CI, 48% to 90%), indicating that it had poor to moderate power to predict short-term death in patients with TBI.

Conclusions: The Rotterdam CT score is a poor predictor for mortality of TBI in the short term.