Clinical predictors of prognosis in patients with traumatic brain injury combined with extracranial trauma

Objective: The purpose of this study was to investigate whether routine blood tests on admission and clinical characteristics can predict prognosis in patients with traumatic brain injury (TBI) combined with extracranial trauma. Methods: Clinical data of 182 patients with TBI combined with extracranial trauma from April 2018 to December 2019 were retrospectively collected and analyzed. Based on GOSE score one month after discharge, the patients were divided into a favorable group (GOSE 1-4) and unfavorable group (GOSE 5-8). Routine blood tests on admission and clinical characteristics were recorded. Results: Overall, there were 48 (26.4%) patients with unfavorable outcome and 134 (73.6%) patients with favorable outcome. Based on multivariate analysis, independent risk factors associated with unfavorable outcome were age (odds ratio [OR], 1.070; 95% confidence interval [CI], 1.018-1.124; p<0.01), admission Glasgow Coma Scale (GCS) score (OR, 0.807; 95% CI, 0.675-0.965; p<0.05), heart rate (OR, 1.035; 95% CI, 1.004-1.067; p<0.05), platelets count (OR, 0.982; 95% CI, 0.967-0.997; p<0.05), and tracheotomy (OR, 15.201; 95% CI, 4.121-56.078; p<0.001). Areas under the curve (AUC) of age, admission GCS, heart rate, tracheotomy, and platelets count were 0.678 (95% CI, 0.584-0.771), 0.799 (95% CI, 0.723-0.875), 0.652 (95% CI, 0.553-0.751), 0.776 (95% CI, 0.692-0.859), and 0.688 (95% CI, 0.606-0.770), respectively. Conclusions: Age, admission GCS score, heart rate, tracheotomy, and platelets count can be recognized as independent predictors of clinical prognosis in patients with severe TBI combined with extracranial trauma.

Multiple trauma was defined as the injury associated with more than one body cavity or organ,

Ivyspring
International Publisher which was considered to be the major cause of death and morbidity [21]. As showed in previous clinical studies, TBI was often accompanied by injuries of extracranial sites [22]. In patients with multiple trauma, prognostic factors included age, sex, ISS score, GCS score, injury mechanism, and systolic blood pressure [23][24][25]. Condition of TBI patients with multiple injuries involves multiple parts, which are often severe and difficult to diagnose and treat, and may influence their prognosis. Hypotension and coagulopathy resulting from an extracranial injury were important factors for secondary injury to TBI [26]. The above predictive factors should be evaluated thoroughly in patients with TBI and extracranial trauma.
Therefore, to study the influence of extracranial trauma on the prognosis of TBI, we analyzed patients with TBI combined with extracranial trauma and evaluated whether blood analysis at the admission and clinical characteristics can be used to predict the patients' prognosis.

Material and methods
This retrospective study included patients with traumatic brain injury combined with extracranial trauma who were admitted to our department between April 2018 and December 2019 in Tongji Hospital. The study was approved by the local Ethics Committee. The diagnosis was based on a history of injury, clinical manifestations, and radiological examination. The inclusion criteria were as follows: (1) traumatic brain injury with head AIS ≥ 3, (2) at least one other body region with AIS ≥ 2, (3) admission Injury Severity Score (ISS) > 15, (4) interval from injury to hospital admission < 24 h, and (5) age ≥ 18 years. Patients with missing or incomplete data and isolated TBI were excluded.
Data regarding age, gender, injury mechanism, time interval from injury to admission, blood pressure, heart rate, pupil size, pupillary reflex, admission GCS score, ISS score, head AIS score, injury position, use of tracheal intubation and tracheotomy, and complications during hospitalization were collected from electronic medical records in Tongji Hospital. Patients were followed up and evaluated for Glasgow Outcome Scale-extended (GOSE) score 1 month after discharge.

Statistical analysis
Continuous data were expressed as mean ± standard deviation (SD), and compared using Student's t test. Categorical variables were analyzed by Pearson Chi-square or Fisher's exact probability test. To identify predictors independently associated with the prognosis, we performed multivariate logistic regression analysis with variables that had shown a statistical trend (p < 0.1) in univariate analysis. The corresponding 95% confidence interval (95% CI) was used to calculate and represent the odds ratio (ORs). The receiver operating characteristic (ROC) curve was used to show sensitivity and specificity of independent predictors to calculate the optimal cutoff points. The area under the ROC curve (AUC) was also calculated to judge the existence of discriminative ability. AUC > 0.5 indicated discriminative ability, AUC = 1 indicated complete discriminative ability, while AUC < 0.5 indicated no discriminative ability. All tests were two-sided, and p < 0.05 was considered statistically significant. Statistical analysis was conducted using SPSS 22.0.

Results
In total, 534 patients with traumatic brain injury were collected and assessed in this study. Among 182 patients finally included in the present study, unfavorable and favorable groups comprised 48 (26.4%) and 134 (73.6%) patients, respectively. Age distribution of the patients is shown in Figure 1. The mean age was 50.41 ± 14.42 years (range 18-84 years). There were 27 (14.8%) older patients (>64 years) with 15 (11.2%) patients in favorable groups and 12 (25.0%) in unfavorable groups, which were statistically different.   Table 1.

Discussion
This study showed the predictive value of blood routine tests at admission and clinical characteristics for neurological functional outcome in patients with TBI combined with extracranial trauma at one month after discharge in level I trauma center. The purpose of this study was to analyze the clinical characteristics and routine blood test results that may predict the prognosis in patients with TBI combined with extracranial trauma. Moreover, although some blood parameters were predictive in univariate analysis, they did not remain significant after multivariate correction. Therefore, our study focused on the possibility to use blood values at admission and clinical characteristics as prognostic markers in patients with TBI combined with extracranial trauma.
In terms of clinical characteristics, our logistic regression model showed that for each additional year and point of age and heart rate, the risk of unfavorable prognosis increased by 7% and 3.5%, respectively. For each point reduction in initial GCS and platelets at admission within 24 hours after injury, the risk of unfavorable outcome increased by 19.3% and 1.8%, respectively. In addition, anyone undergoing tracheotomy had 15.201-fold higher risk for poor prognosis. The AUC of age, admission GCS score, heart rate, tracheotomy, and platelets count was 0.678, 0.799, 0.652, 0.776, and 0.688, respectively.
Our study indicated a significant correlation between age and outcome at 1-month after discharge in patients with TBI and extracranial trauma. The age distribution histogram revealed that the age of patients was mainly between 40 and 70 years. In the unfavorable group, there were 25% of individuals with older age compared with 11.2% in the favorable group. Older age has been linked to worse outcomes, although possible reasons are still under discussion [32,33]. Among TBI patients, older patients more commonly use anticoagulants and antiplatelet agents [34,35]. Older individuals are more likely to experience domestic falls, which is associated with worse outcome compared with young individuals [1]. Moreover, we found that admission GCS score was an independent predictor of prognosis in TBI patients with extracranial trauma, which is consistent with some previous studies in TBI patients [3,7,36]. We did not confirm the ability of ISS to predict the prognosis of TBI patients after correcting for multiple blood indicators in a multivariate logistic regression model. However, our study confirmed that age and admission GCS can be used as remarkable predictors of the prognosis in TBI patients with extracranial trauma.
Heart rate is a vital sign and could be influenced by some pathological states, such as pain, shock, and intracranial hypertension. Reverse shock index, systolic blood pressure lower than the heart rate, also indicated an unfavorable outcome in patients with severe isolated TBI [37]. Moreover, another study revealed that cardio-cerebral network imbalance might influence the relationship of mean arterial pressure, intracranial pressure, and heart rate in sTBI patients [38]. As our study showed, heart rate was also an independent factor predicting the outcome in our patient population.
Tracheotomy is a common clinical procedure in patients with sTBI, which provides a stable and tolerated airway to ensure oxygen supply, despite some complications that may accompany the procedure [39]. Here we found that tracheotomy was an independent factor for unfavorable outcome. Some studies have shown increased survival in patients with tracheostomy compared with patients who remained intubated after severe TBI [40,41]. However, the optimal time to perform tracheostomy remains a highly controversial topic. Early tracheostomy within 72 hours of admission reduced the duration of mechanical ventilation and length of stay in intensive care unit (ICU) in 120 patients with TBI [42]. A randomized trial and a retrospective meta-analysis indicated that early tracheostomy increased risk for hospital death and did not decrease ventilator-associated pneumonia rates [39,43]. In contrast, several studies reported that early tracheostomy could improve prognosis in TBI patients [4].
The majority of patients had normal platelet count, and 29.7% of patients had low platelets in this study. We found that the platelets count was the only significant blood parameter in multivariate logistic regression model, apart from age, admission GCS score, and tracheotomy. Acute coagulopathy of trauma (ACT) is caused by tissue injury and tissue perfusion [44]. Coagulopathy is defined as low platelet count or elevated INR or prolonged APTT [45]. The platelet dysfunction, increased level of platelet distribution width (PDW), and low platelet count were associated with unfavorable outcome in TBI patients [46][47][48][49]. The fresh frozen plasma (FFP) resuscitation attenuated platelet dysfunction and improved survival in animals with multiple trauma [50]. Furthermore, the importance of ACT in TBI patients has been increasingly recognized. In a large population of patients with TBI, INR and APTT were recognized as independently related to in-hospital mortality [7,51]. High PT was also recognized as a predictor of mortality in patients with trauma [49,52]. However, PT, INR, and APTT were not independent predictors in this study.
Inflammation also plays an important role in TBI and multiple trauma [53,54]. Mass release of proinflammatory factors caused by trauma stimulation or tissue necrosis can lead to leukocyte activation and lymphocytes deficiency [55][56][57]. Neutrophil activation has dual effects on TBI, which might contribute to repair mechanisms or aggravate the pathophysiology of trauma [58,59]. The neutrophil-to-lymphocyte ratio (NLR) is associated with unfavorable outcomes in sTBI patients [8]. The majority of our patients had higher leukocyte and neutrophil counts and lower lymphocytes. The inflammatory stimulation caused by multiple injuries may cover the inflammatory manifestations caused by isolated traumatic brain injury, which may explain why inflammatory cell prediction was not significant.
Hypernatremia and hyperglycemia were associated with poor outcome after severe TBI [49,[60][61][62][63][64]. The main point is that stress response activates the hypothalamic-pituitary-adrenal axis and sympathetic nervous system, leading to elevated levels of neurohormones and insulin resistance [65]. Intensive insulin therapy was considered as a therapeutic strategy to treat cerebral metabolic distress in a previous study [66]. However, to prevent hypoglycemia, strict blood sugar control was not recommended [67]. Nevertheless, we did not find that high Na+ concentration, hyperglycemia, and high BUN were independent predictors of outcomes in our study. In addition, prognosis of patients with TBI was worse in patients with anemia [49,68,69]. Multiple trauma was also related with injury-associated anemia [70]. More than 50% of the patients in our study population had low hemoglobin levels. However, hemoglobin level was not an independent predictor of neurological prognosis in this study.
We applied the ROC curve and different combination models to evaluate the accuracy of some variables for predicting the outcome in patients with TBI combined with extracranial trauma. Our results showed that total combination models displayed proper accuracy to assess prognosis of patients, with 91.7% sensitivity and 83.5% specificity. The AUC of the model 2 was 0.903 and Youden index was 0.752. The model 2 is useful for predicting the clinical outcome because these variables are easy to obtain. To predict the early prognosis of patients, we excluded tracheotomy as a variable and established the model 1. The AUC of the model 1 was 0.851 and Youden index was 0.536, with 91.7% sensitivity and 61.9% specificity. The AUC of the model 2 was the largest, indicating the highest prognostic accuracy.
This study had several limitations. First, this was a single-center retrospective study, which is why selection bias may have existed. Second, this study investigated a relatively small cohort of patients that may not be significantly representative of a TBI population with severe multiple injuries. Some patient data about pre-injury drug use and more comorbidities were incomplete. In addition, the effects of different parts of multiple injuries on TBI may be different. Finally, the prognosis was assessed with a short-term outcome. Therefore, further randomized studies based on large populations and appropriate follow-up times are needed to provide stronger evidence for predicting patient outcomes.

Conclusion
We further confirmed that age, admission GCS score, pupillary reflex, tracheotomy, and platelets count can be used as independent predictors of clinical prognosis in patients with severe TBI with extracranial trauma.