All head injury patients who presented at the emergency room triage of Chi-Mei Medical Center between 2010 and 2019 were retrospectively included in this study. The inclusion criteria were as follows: patients with head trauma who were admitted to the ER and had the following diagnostic codes—ICD-9: 800*-804*, 850*-854*, 959.0, 959.01, 959.8–959.9, ICD-10: S00*-T07*. Patients with missing or ambiguous values were excluded. Other patients and the public were not involved in any way in this research. Figure 1 displays the study flow diagram.

The twelve feature variables including patients’ age, gender, body mass index (BMI), Taiwan Triage and Acuity Scale (TTAS), heart rate, body temperature, respiratory rate, GCS, left and right pupil size, and light reflex were enrolled due to their wide availability in the triage setting. Dilated pupils were defined as ≥4 mm in diameter (11). The presence of unresponsive pupil is defined as unresponsive (<1 mm) to a light stimulus (19, 24). The pupillary light reflex was divided into three categories including both unreactive, one unreactive and both reactive. All patients in our hospital received the standard management protocol for TBI patients. Specifically, following surgery, patients underwent cerebral perfusion-guided management with the goal of maintaining a cerebral perfusion pressure of 60 mmHg or higher and ensuring intracranial pressure remained at 20 mmHg or lower. The recorded outcome was categorized as either mortality or survival.

Significant testing was conducted using the t-test for numerical variables and the Chi-square test for categorical variables. Additionally, data were analyzed using Spearman’s correlation method to demonstrate the strength of the correlation between each feature and mortality. Multiple logistic regression models were used for correlation between two variables and defined the independent risk factors for poor outcome. Using the Receiver Operating Characteristic (ROC) and the area under curve (AUC), variables cutoff value and its reliability in prognosis could be estimated (25). Commercial statistical software (SPSS for Windows, Version 15, SPSS Inc., Chicago, IL, United States) was utilized for this analysis. p-values less than 0.05 were considered statistically significant.

This study obtained ethics approval (IRB no. 10911-006) from the Institutional Review Board of Chi Mei Medical Center in Tainan, Taiwan. The authors conducted all methods in compliance with applicable guidelines and regulations. The Ethics Committee waived the need for informed consent due to the retrospective nature of the study.

This study conducted a review of relevant literature and introduced an innovative approach to early mortality risk prediction in the emergency triage setting, without the need for brain CT scans or blood sampling. We offer a more precise method for assessing a patient’s risk by considering various variables, including the size and reactivity of the left and right pupils, which are combined to provide a comprehensive assessment. We have also established specific thresholds for age (58.55 years), left pupil size 3.75 mm and right pupil size 3.25 mm that are associated with predicting the risk of death in TBI patients with GCS 3 and unresponsive pupil light reflex. These distinctions can aid physicians in accurately evaluating the risk of TBI patients and in developing more effective treatment plans. Nonetheless, it is advisable for physicians to consider the results of multiple studies when making clinical decisions.

A GCS score of 3 at the initial assessment has been associated with high mortality rates: 49.2% (12), 64% (10), 80.6% (13), and 89% (9). In our current study, the mortality rate was 44.10%, similar to Chamoun RBs’ reported 49.2% and better than the others (Table 8). This discrepancy may be related to changes in healthcare practices (e.g., ICP-guided management), patient characteristics (such as age distribution), and the timing of the observation endpoint (whether in-hospital or 6 months later).

Our results (Table 4) suggest that when patients have a GCS score of 3 and both pupils are reactive, the mortality rate is 23.08% (12/52), similar to Chamoun RBs’ reported 23.5% (12). Additionally, we observed that patients with a GCS score of 3 and both unreactive and dilated pupils (≥4 mm) have a significantly higher mortality rate of 69.39%, which is lower than the 79.7% reported by Chamoun et al. (12). In contrast, patients with a GCS score of 3 and both unreactive pupils with both pupils measuring <4 mm have a lower mortality rate (32.26%) compared to the GCS of 3-only group (44.10%). These statistics indicate that the combination of a low GCS score and pupil reactivity has varying impacts on the likelihood of patient survival, and this information is crucial for understanding patient outcomes in cases of severe neurological impairment. Therefore, we want to emphasize the importance of measuring pupil size in predicting mortality.”

Numerous studies consistently underscore the pivotal role of age in predicting prognosis. For instance, in patients with a Glasgow Coma Scale (GCS) score of 3 suffering from traumatic brain injury (TBI), earlier research found that all survivors were under 50 years old (9), while individuals above 55 faced a notably higher mortality risk (10). In our current study, non-survivors had a mean age of 62.32 years, whereas survivors averaged 50.37 years (Table 1). This discrepancy may be attributed to the relatively advanced age of our patient cohort compared to that in previous studies (Table 8). It is plausible that older patients may have pre-existing chronic health issues that negatively impact their recovery ability (26), whereas younger individuals may possess more resilient brains better able to withstand damage (27). Additionally, treatment complications and effectiveness may vary between younger and older individuals, potentially influencing treatment outcomes (28).

We also reaffirm that age is an independent predictive factor for mortality (as shown in Tables 3, 5–7). Furthermore, we identified specific age thresholds of 41.11, 58.55, and 43.87 years in three categories with varying feature combinations (as presented in Table 8). In all three categories, age emerges as the most clinically significant feature, exhibiting the highest AUC and statistical significance. Notably, pupil size, especially right pupil size, holds less significance in Category III. These findings suggest that age outperforms pupil size as a predictor of patient condition in these categories. Importantly, age is a non-invasive and easily obtainable parameter, making it a practical clinical tool. The age threshold can be clinically applied to assist in decisions regarding TBI conditions.

Lieberman et al. reported 100% mortality in cases with fixed, dilated pupils (29). In the current study, both the size of the left (r = 0.357, p < 0.001**) and right (r = 0.357, p < 0.001**) pupils showed a significant correlation with mortality. This suggests that both left and right pupil size measurements appear to have potential clinical significance, and that patients with larger pupils have a higher risk of mortality (Table 2). Additionally, we observed that left pupil size had an odds ratio of 1.712 (p < 0.001**) for independently predicting mortality (Table 3). Furthermore, left pupil size, with a slightly higher area under the curve (AUC) value than the right pupil, was found to be a significant predictor of prognosis in different feature combinations (Tables 5–7).

The probable reasons for the left pupil size being a more meaningful determinant of prognosis compared to the right pupil size may relate to (1) left-sided dominance brain injuries could affect functions that are more closely linked to the left eye, which might be reflected in left pupil size, (2) specific injuries or conditions on the left side of the brain that we were not considered in this study.

In the current study, the setting was the triage of the ER, and no imaging studies had been conducted yet. Therefore, left hemisphere compression by hematomas, or cases with greater midline shift, were not considered as predictors. However, it’s possible that left pupil size was associated with more cases of left hemisphere compression by hematomas, or those with greater midline shift. In the future, it will be necessary to include the aforementioned parameters as predictors of mortality. Pupil size measurements can be valuable in assessing and stratifying the risk of TBI patients. Further research and clinical validation may be necessary to determine the most effective clinical applications.”

Recently, several studies have shown that both age and pupil conditions are among the top predictors of in-hospital mortality after TBI (30–32). These studies also indicate that brain CT findings serve as important predictors. Our results are consistent with these findings. However, even without CT examination, we identified age and pupil size as key predictors of mortality in TBI patients with GCS 3 in the triage setting.

Our study underscores the significance of early mortality assessment in patients with a GCS score of 3, a group already recognized for its challenging prognosis. The recognition that age and pupil size serve as robust mortality predictors equips healthcare providers with valuable tools for making well-informed decisions in TBI patient management. This knowledge facilitates a more precise and individualized approach to care. For instance, patients exhibiting identified risk factors, such as advanced age or specific pupil size, can be closely monitored, and timely interventions can be initiated to enhance their chances of survival and minimize unfavorable outcomes while simultaneously improving their overall quality of life.

Furthermore, our study illuminates the necessity for additional research. While age and pupil size have emerged as prominent predictors, other variables have shown varying degrees of significance. This highlights the potential for further investigations into additional factors that might contribute to TBI patient outcomes, thus leading to a more comprehensive understanding of prognosis and the development of more effective treatment strategies.

Our study possesses several strengths. Firstly, the data in the current study is the most recent, covering the period from 2010 to 2019, thus reflecting the latest advancements in modern medical practices and technologies. Secondly, unlike other studies, our research includes a broader age distribution, encompassing older patients (22% over 70 years), which enhances our ability to gain a more comprehensive understanding of prognosis across various age groups. Thirdly, our study meticulously records both pupil size and reactivity, both of which are critical prognostic factors. In contrast, some other studies either lack this data or do not provide corresponding information.

However, it’s important to acknowledge several limitations in our study. Firstly, as a retrospective observational study, there is the possibility of miscoded feature variables. Researchers have limited control over the data collection process, which may introduce biases or confounding factors. Secondly, in comparison to some other studies, the current study has a relatively smaller sample size (161 cases), potentially constraining its statistical power and limiting the generalizability of findings to larger and more diverse patient populations. Thirdly, the accuracy of GCS scores and the consistency in measuring pupil size and reactivity may vary among different healthcare providers, potentially leading to measurement errors. Fourthly, the study, conducted at the ER triage, did not account for the existence of uncontrolled or unmeasured confounding factors that could influence the results. For instance, other medical conditions, various treatments, surgical procedures, and complications that could impact the outcome after TBI were not considered in the analysis. Finally, variations in study endpoints, with some studies focusing on in-hospital outcomes while others use a 60-day endpoint, may affect the comparability of study results.

Consequently, there is a need for larger prospective studies with more comprehensive data collection and the inclusion of additional variables to be considered in future research.