Factors affecting worsening intracranial injuries in pediatric patients with mild traumatic brain injury

Article information

Pediatr Emerg Med J. 2026;13(2):58-64

Publication date (electronic) : 2026 March 4

doi : https://doi.org/10.22470/pemj.2026.01564

Nu Ri Baik, Jin Seong Cho, Jae-Hyug Woo, Jae Ho Jang, Woo Sung Choi, Yong Su Lim, Jea Yeon Choi

Department of Emergency Medicine, Gachon University Gil Medical Center, Gachon University College of Medicine, Incheon, Republic of Korea

Corresponding author: Jea Yeon Choi, Department of Emergency Medicine, Gachon University Gil Medical Center, Gachon University College of Medicine, 21 Namdong-daero 774beon-gil, Namdong-gu, Incheon 21565, Republic of Korea Tel: +82-32-460-3901; E-mail: chjy14664@gilhospital.com

Received 2026 January 28; Revised 2026 February 12; Accepted 2026 February 15.

Abstract

Introduction

Traumatic brain injury (TBI) is common in the pediatric population and may result in serious neurological morbidity or death if not appropriately diagnosed and managed (1). Computed tomography (CT) is widely used as an essential imaging modality for evaluating intracranial injuries (ICIs) (2). Repeat CT (RCT) can facilitate early detection of newly developed lesions or progressive hemorrhagic injury, potentially prevent neurological deterioration and improve clinical outcomes (3).

In patients with severe TBI, defined as a Glasgow Coma Scale (GCS) score <8 (4), routine RCT is often recommended due to the difficulty of reliable neurological assessment. However, it remains controversial to apply same strategy to pediatric patients with mild TBI (GCS score 13–15) (5,6). Some studies suggested the utility of routine RCT (7,8,9), while others reported limited benefit in the absence of changes in clinical status (3,10). Furthermore, considering the potential risks of radiation exposure in children, unnecessary CT should be minimized (11). Overall, insufficient evidence exists regarding appropriate indications in applying RCT on pediatric patients with mild TBI.

Therefore, we aimed to identify clinical and radiologic factors associated with worsening ICI on RCT in pediatric patients with mild TBI. Identifying such risk factors may provide evidence for more selective and efficient use of RCT.

Methods

Results

Among 224 eligible pediatric patients, 212 were enrolled after excluding 10 patients for unavailable pre-transfer imaging and 2 for concomitant severe injuries. The median age of the study population was 9.0 years (interquartile range, 4.0–16.0). Based on RCT findings, 48 patients (22.6%) were classified the worsening group. The worsening group showed an older median age (13.0 vs. 8.5 years), and higher percentages of initial GCS score of 14 (29.2% vs. 11.6%) and motorcycle/bicycle as the injury mechanism (52.1% vs. 20.7%) than the stable group (Table 1). On initial CT findings, the worsening group showed higher percentages of skull fracture (79.2% vs. 56.1%), EDH (52.1% vs. 24.4%), SDH (50.0% vs. 23.8%), and pneumocephalus (37.5% vs. 15.9%), compared with the stable group. A median time to RCT was shorter in the worsening group (189.0 minutes [172.0–228.0] vs. 229.0 minutes [193.0–269.0]). The worsening group showed a more frequent hospitalization to the ICU (79.2% vs. 35.4%). Median values of overall and ICU LOS were longer in the worsening group. Five patients underwent surgical interventions due to increased hemorrhage or clinical deterioration after RCT. In addition, 2 patients underwent elective surgery for skull fractures.

Table 1.

Baseline characteristics of the study population

Variables included in the regression model were age, sex, initial GCS score, motorcycle/bicycle as the injury mechanism, skull fracture, SDH, EDH, and pneumocephalus. Among these variables, independent predictors of worsening ICI were SDH (odds ratio, 4.99 [95% confidence interval, 2.08–11.96]), EDH (4.04 [1.73–9.44]), and motorcycle/bicycle (2.94 [1.14–7.59]; Table 2).

Table 2.

Multivariate logistic regression model for predicting worsening intracranial injury

Discussion

This study, which included 22.6% of the study population with worsening ICIs on RCT, showed SDH, EDH, and motorcycle or bicycle accidents as the independent predictors of radiologic worsening. These findings indicate that among pediatric patients with mild TBI with positive initial CT findings, a subset remains at risk for progression of ICI. Hence, early identification of patients at higher risk of such radiological deterioration may allow more focused clinical monitoring while avoiding unnecessary repeat implementation of CT. The percentage of radiological worsening in this study was comparable to or slightly higher than the equivalent findings in previous studies, which ranged from 11% to 19.8% (6,12,13). Although male patients constituted the majority of the study population, sex was not a predictor, consistent with previous reports. In addition, age was not significant in the regression model.

High-energy injury mechanisms, such as traffic accidents, are known to raise the risk of delayed or progressive hemorrhagic injury, even in patients with initially preserved neurological status (6). Di et al. (6) demonstrated that traffic accidents were associated with a higher risk of progressive hemorrhagic injury compared with low-energy mechanisms such as ground-level falls, emphasizing the importance of injury mechanism over initial GCS score alone. Similarly, Hollingworth et al. (14) identified traffic accident-related mechanisms as predictors of new or worsening intracranial findings on RCT in pediatric head injury. In our study, motorcycle or bicycle accidents were associated with a nearly 3-fold increase in the risk of radiologic worsening.

In this research, both EDH and SDH identified on initial CT scan were predictors of worsening ICI. EDH is consistently reported as a major reason for delayed neurosurgical intervention in pediatric TBI (10,13). In children, the elasticity of the skull allows dural separation even without an obvious fracture, and arterial bleeding may rapidly expand despite minimal initial hemorrhage. Regarding SDH, Shafiei et al. (15) reported an association between SDH and radiologic progression following head injury. Mizu et al. (16) also noted that patients younger than 15 years with radiologic deterioration on RCT frequently had SDH or EDH on initial imaging. Because children have limited intracranial reserve due to minimal brain atrophy, even small increases in SDH volume can lead to clinically relevant progression. In support of this concept, Hanalioglu et al. (12) included SDH as a weighted variable in the IniCT score for predicting clinically important TBI in pediatric patients.

Skull fracture showed a borderline association with progression of ICI in the regression model (Table 2). Earlier studies have suggested that isolated skull fractures in neurologically intact children carry a favorable prognosis and do not routinely require RCT (17). However, in our study, skull fractures were frequently accompanied by ICIs, particularly SDH or EDH. Durham et al. (18) reported that skull fractures associated with intracranial hemorrhage were linked to an increased risk of progression, suggesting that skull fractures might serve as a marker of injury severity rather than an independent predictor of worsening.

Radiologic worsening on RCT does not automatically indicate clinical decline or the need for neurosurgical intervention. Stippler et al. (19) showed that routine follow-up CT scans in neurologically stable individuals with mild TBI rarely changed management, emphasizing that neurological changes should primarily direct decisions for repeat imaging. Aziz et al. (10) suggested that routine RCT could be replaced by careful neurological examination in pediatric mild-to-moderate TBI. However, imaging findings should not be overlooked, as radiologic progression may indicate delayed hemorrhage or identify patients who require closer observation. Hill et al. (3) and Hollingworth et al. (14) reported that although the majority of cases of radiologic progression were treated conservatively, RCT occasionally provided clinically meaningful information. The study findings support a risk-stratified approach in pediatric mild TBI. Identifying patients with high-risk features may allow selective use of RCT while prioritizing close clinical and neurological monitoring, thereby reducing unnecessary radiation exposure without compromising patient safety.

This study has limitations. First, given the single-center retrospective design, selection bias cannot be excluded. Patients who were clinically stable and discharged without undergoing RCT were not included, potentially overestimating the incidence of radiological worsening. Second, the relatively limited sample size reduced statistical power for analyzing a rarer injury pattern, such as subarachnoid hemorrhage. In addition, the number of variables included in the regression model was relatively large compared to the sample size of the worsening group, which might raise the risk of overfitting. Third, only 5 patients younger than 2 years were included in the worsening group, limiting the feasibility of age-stratified analyses. Considering the unique anatomical and physiological features of this age group, it may be difficult to generalize our findings to children younger than 2 years. Finally, the timing of RCT was not standardized and was determined by the physician’s judgment. This variability in timing and clinical context might influence the detection of radiological worsening and subsequent management decisions. Future studies applying standardized imaging protocols would be essential to validate our results.

In conclusion, motorcycle or bicycle accidents and the presence of SDH or EDH on initial CT scans were independent predictors of worsening ICI in pediatric patients with mild TBI. These results indicate that RCT should be selectively evaluated in patients with high-risk characteristics, avoiding universally routine follow-up CT scans. A risk-based approach combining clinical evaluation and focused imaging could reduce unnecessary radiation exposure while allowing timely detection of neurological deterioration.

Article information Continued