Traumatic brain injury (TBI) is a common disease with high rates of disability and mortality, which is one of the leading causes of death and disability in adults (1, 2). Moderate TBI is defined by a Glasgow Coma Scale (GCS) score of 9–12 at admission, with reported incidence rates ranging from 4 to 28% (3). Due to its relatively mild condition, little attention has been paid to moderate TBI in clinical practice. Currently, unlike mild or severe TBI, there are no guidelines providing recommendations for the management of moderate TBI (4). Indeed, patients with moderate TBI are in a dangerous state and at high risk of early neurological deterioration (END). Many of them who “talk and die” after trauma belong to this category (5, 6). END is a common complication after TBI, and is characterized by a sudden decline of GCS more than 2 or conversion to surgical treatment within 72 h after admission (7). Previous studies have shown that the prevalence of END in moderate TBI is as high as 33.3% and END serves as a prognostic factor for unfavorable outcomes, potentially resulting in severe disability or death in patients with moderate TBI (8, 9).

Patients at risk of END often require early, high-level neurological monitoring and interventions, such as intracranial pressure monitoring (10, 11). However, the risk factors for the occurrence of END have not been fully elucidated, which makes its early diagnosis and treatment challenging. Therefore, it is essential to explore its mechanism and the related risk factors after TBI. There is urgent clinical demand for an effective model to predict END in patients with moderate TBI. A nomogram is a visualization tool that integrates indicators from different dimensions, to predict the risk of disease occurrence in a graphical and individualized manner (12). Nowadays, there are few studies reporting risk factors or prediction models for END in patients with moderate TBI (13). Thus, the aim of this study was to explore the risk factors associated with END. Finally, we developed and validated a nomogram to assist in predicting the risk of END in moderate TBI patients.

Early neurological deterioration was reported in quite a large proportion of patients with moderate TBI, and it is associated mainly with unfavorable outcomes. In order to identify and prevent the occurrence of END as early as possible, it is important and urgent to explore predictors of END. This study found that the five variables collected at admission (GCS score, epidural hematoma, intracerebral hemorrhage, fibrinogen, and D-dimer) were significantly related to END, based on the logistic regression analysis. To our best knowledge, this is the first study to develop a nomogram for evaluating the risk of END in patients with moderate TBI.

The GCS score is the simplest and most reliable method to evaluate the severity of injury in TBI patients. Initial GCS score has been identified as a predictor of poor prognosis in patients with moderate TBI (18, 19). In the present study, GCS score was an independent predictor of END of moderate TBI, and we found that patients with lower GCS scores tended to have a higher risk of END. Nagesh et al. reported that in patients with lower GCS scores, repeat CT scans could exhibit abnormalities during hospitalization (20). Besides, a lower GCS score was found to be associated with an increased risk of contusion expansion (21), which is a major cause of END.

EDH is a collection of blood clots between the outer layer of dura and inner table of skull, which typically occurs from tearing of the middle meningeal artery or dural venous sinuses (22). It is not stable and has a tendency for rebleeding, resulting in sudden changes and deterioration of the patient’s condition, with a life-threatening brain herniation occurring within a few hours. Wang et al. found that EDH is most likely to enlarge in the first 6 h but rarely does so after 24 h (23). It has been reported that up to 10.1% of EDHs still tend to grow within 5–6 h after injury (24). In a study by Xiao et al. it was found that acute EDH without surgical intervention in the early stage may be associated with hematoma enlargement when CT time was added (24). The present study demonstrated that enlarged epidural hematoma was an independent risk factor for END, which was consistent with the previous reports (25). Thus, an aggressive CT repeat strategy should be employed in EDH to assess the progression of hematoma, especially in the early stages after TBI, and these patients often require surgical intervention because of the higher possibility of END during follow-up.

The ICH was also found to be a predictor of END in patients with moderate TBI. TBI is frequently accompanied by intracranial hemorrhage, with progressive ICH being the most common type of progressive hemorrhagic injury (23). The findings of this study indicate that another cause of END is delayed intracerebral hemorrhage, making it necessary for patients with moderate TBI to undergo repeat head CT scans. Studies have shown that ICH can contribute to cerebral edema and increase intracranial pressure (26), additionally, ICH has been proven to be associated with the development of acute hydrocephalus, which can increase intracranial pressure, potentially leading to brain injury, herniation and END (27). In patients with ICH, the volume was greater than 30 mL, and patients were at a high risk of hemorrhagic progression (28), therefore, it is recommended that patients with moderate TBI accompanied by ICH receive immediate surgical treatment to prevent the occurrence of END.

Fibrinogen, a glycoprotein mainly synthesized by the liver, is the major structural component of a clot (29). During the coagulation process, fibrinogen is converted into fibrin by thrombin activation, and participates in platelet aggregation through a cross-linked fibrin network (30). Many studies have revealed that patients with TBI have a significant decrease in fibrinogen levels in the early stages, which is correlated with the severity of brain injury (31). Peng et al. reported that lower fibrinogen levels were predictive for progressive hemorrhagic injury after TBI (32). In another retrospective study involving 106 patients with severe TBI, authors found that fibrinogen at admission was independently associated with the occurrence of progressive hemorrhagic injury and was the main cause of neurological deterioration (33). Moreover, fibrinogen level is also considered to be an independent risk factor for the enlargement of intracranial hemorrhage. One such study reported a negative correlation between plasma fibrinogen levels and the volume of hematoma in TBI patients (34). In the present study, low fibrinogen level was also identified as a predictor of END. Therefore, patients with moderate TBI with lower fibrinogen levels may be at a higher risk of hematoma expansion, which can further explain the association between low fibrinogen levels and END, and poor prognosis.

High D-dimer levels were also demonstrated to predict END. D-dimer is a soluble fibrin degradation product deriving from the plasmin-mediated degradation of cross-linked fibrin, and is one of the most important indicators of coagulation function. Its presence reflects concomitant activation of both coagulation and fibrinolysis (35). Selladurai et al. reported that the incidence of abnormal coagulation following TBI was 40% (36), and patients with high D-dimer levels experienced a greater frequency of unfavorable outcomes after TBI (37, 38). This study demonstrated a significant association between elevated D-dimer levels and the occurrence of END in patients with moderate TBI. Although the mechanism of the impact of D-dimer on END was not well defined, a recent study confirmed that D-dimer was a significant independent predictor of hemorrhage progression (39). Tong et al. stated that an increase in D-dimer correlated with progressive hemorrhagic injury (40). Similarly, Peng et al. suggested that higher D-dimer at admission could predict progressive hemorrhage after head trauma (32). In addition, higher D-dimer levels were found to be associated with a higher risk of contusion progression (41, 42), which is a major cause of END. With regard to logistic regression analysis of multiple factors, D-dimer level was strongly related to the occurrence of END, which indicates that the higher D-dimer level, the greater likelihood of END.

Early detection is a key to preventing the occurrence of END. This is best accomplished by identifying high-risk patients. Consequently, these patients need to be under observation, followed up, and should undergo dynamic-CT review. Early intervention and treatment are needed after END is found. The nomogram is one of the methods to present the predicted model as the graphic scoring. After performing the multivariate regression analysis, we included GCS score, epidural hematoma, intracerebral hemorrhage, fibrinogen and D-dimer, a total of five factors, as nomogram score points. These variables represent the central nervous system function, severity of injury and coagulation status of TBI patients, which basically encompass the degree of anatomical damage and pathophysiological alterations in the body following trauma. Furthermore, the results can be obtained within a relatively short period of time after patient admission, making it convenient, fast, and easy to use. The nomogram prediction model had good discrimination and clinical utility. The AUC for predicting the risk of END occurrence in the training group was 0.901, and also reached 0.927 in the validation group. The calibration plot almost perfectly fit the actual situation. DCA analysis showed that the nomogram exhibited excellent clinical applicability in predicting the risk of END, and the prediction model also performed well in the goodness-of-fit test. The above indicates that the nomogram prediction model may play a positive role in predicting the risk of END in patients with moderate TBI. Compared with traditional clinical prediction methods, the nomogram model can more intuitively show the relationship between various risk factors and prognosis, providing a quantitative prediction tool for clinicians (43). In addition, the model helps to improve the accuracy of prediction and provides the basis for patients to develop individual treatment plans. In patients with moderate TBI, targeted treatment measures can be formulated according to the relevant risk factors to adjust the treatment regimen, thereby effectively improving the prognosis of the patient.

This study has several limitations. First, this was a retrospective study and may therefore have a certain degree of selection bias. Second, a limited number of patients were recruited from a single institution. The prediction model was only subjected to internal validation, lacking external validation, it is necessary to conduct large sample and multicenter studies to prove the feasibility of the nomogram and increase the possibility of extensive popularization of the model. Third, blood parameters were recorded only once on admission and were not monitored dynamically. Last, our model only focuses on the END of patients with moderate TBI, which, although important for clinical practice, does not assess long-term prognosis. To overcome these limitations and further improve the predictive accuracy of the model, we plan to adopt a multi-center, multi-regional approach in future studies to collect more diverse patient demographic data. This will help us build a more comprehensive and accurate model to better predict the risk of END or prognosis of moderate TBI patients.

In this study, we found that GCS score, epidural hematoma, intracerebral hemorrhage, fibrinogen, and D-dimer were predictors of END in patients with moderate TBI. A prediction nomogram for assessing the risk of END was developed, and the receiver operating characteristic curve, calibration plot, and DCA were used to demonstrate that the nomogram had a good predictive performance, calibration, and clinical utility. It might be a reliable and easy-to-use tool to predict END in moderate TBI patients.