Prehospital Predictors for Urgent Neurosurgical Intervention in the Head Trauma Patient: A 2-Year Multicenter Retrospective Study

Background Traumatic brain injury (TBI) is the main cause of disability in the world. Prehospital diagnosis of patients requiring rapid neurosurgical intervention and the earliest possible introduction of procedures preventing secondary brain injuries (SBI) are crucial. Methodology and Study Population. The authors of this paper assumed that certain age groups with specific injuries are more likely to require urgent neurosurgical intervention compared with patients who did not require such an intervention. Out of 54,814 head CT scans, based on the inclusion criteria, 7,864 were selected for the study. Data such as sex, age, the mechanism of injury, comorbid trauma, and abnormal findings in the examination of patients qualified for urgent neurosurgical intervention were analyzed in order to find statistically significant factors through a comparison with all head trauma patients. Results Patients qualified for urgent neurosurgical intervention were significantly older compared with the others (63 years vs. 49 years). Patients transferred from the emergency department directly to the operating room were more often admitted to the hospital due to the fall (64.1% vs. 45.1%, p = 0.004). The following were observed much more commonly among the patients qualified for urgent neurosurgical intervention than in the entire study group of subjects with traumatic brain injury (TBI), e.g., calf deformity (2.2% vs. 0.1%, p = 0.019) and bleeding from the mouth (4.3% vs. 0.0%, p < 0.001). On the other hand, superciliary arch wounds were observed much less commonly than in the entire group (0.0% vs. 5%, p = 0.221). Conclusion Patients admitted directly to the operating neurosurgical room from emergency departments constitute a small percentage of TBI patients, and their prognosis for normal performance status upon discharge is poor. Maximum efforts should be made to distinguish these patients and to start proper treatment even during prehospital care.


Introduction
Head trauma is an important and growing public health problem.Sometimes it is even referred to as a silent pandemic [1].Traumatic brain injuries (TBIs) are the main cause of disability among healthy men around the world.In the USA alone, there were over 64,000 TBI-related deaths in 2020, which equates to 176 TBI-related deaths every day [2].It is estimated that every year, 500-800 out of every 100,000 European residents will sufer from head trauma.As these data come primarily from developed countries, researchers suppose that these numbers may be three times higher in China and other developing countries [3].Aging population is another problem.Until recently, TBI patients were mostly young men involved in road trafc collision (RTC), while the current data show that almost one-third of TBI patients are people aged over 75, who sufer a fall.Tis age group also accounts for 32% of TBI-related hospitalizations and 28% of deaths [4].Terefore, each year an estimated 69 million individuals will sufer a TBI, the vast majority of which will be mild (81%) or moderate (11%) in severity.Per capita, the highest annual incidence of all-cause TBIs is observed in the North America and Europe (1,299 and 1,012 cases per 100,000 people, respectively) [5].It seems that knowledge about how to stop a pathophysiological trauma-related reaction in the brain and especially an efective and evidencebased medicine (EBM) treatment is disproportionately low.According to data from the current Brain Trauma Foundation guidelines, none of the work has produced even one valid conclusion, pharmacotherapy, or procedure improving TBI management in the last 20 years.Since most work does not provide any evidence, interest in research on this issue is declining among sponsors and pharmaceutical companies.For this reason, ongoing assessment of TBI patients admitted for neurosurgical treatment seems to be crucial.Te authors decided to examine the demographic and clinical characteristics and the mechanism of trauma, which could indicate the need for urgent neurosurgical intervention at the stage of prehospital medicine.

Study Population.
We retrospectively evaluated data from two clinical university hospitals from head trauma patients in 2018-2019.As the authors were aware of the risk that many CT results could be omitted if an appropriate ICD-10 S code was missing, the decision was made to request a list of all head CT scans ordered at the ED level from the authorities of both hospitals.In total, 54,814 results were received (29,141 CT scans from a trauma centre and 25,673 head CT scans from another university hospital, which is also the only thrombectomy center within a 100-km radius).All 54,814 hospital records of patients admitted to both EDs were analyzed.In any case where basic data on the reason for CT were missing, data were excluded (Figure 1).A total of 7,930 ED visits were obtained after applying the inclusion criteria: age >18 years, closed head trauma, patient delivered by an emergency medical team, referred, or presenting on their own due to head trauma, mention of head trauma in the emergency medical records or in the referral, signs of head trauma on physical examination, and preserved vital functions upon admission.Te exclusion criteria were no signs of trauma on examination, no mention of head trauma in the medical record, stroke, and patients who underwent cardiopulmonary resuscitation before reaching hospital.

Statistical Analysis.
Te Mann-Whitney U test, Kruskal-Wallis test (with Dunn's post hoc test and correction for multiple Bonferonni testing), chi-square, and Fisher tests were used in the study.Te Mann-Whitney U test is a nonparametric test used to compare numerical variables between two groups of observations.Statistically signifcant results obtained on its basis prove that there is a diference in the distribution of a given variable between these groups.
Te Kruskal-Wallis test is also a nonparametric test for comparing the distribution of a variable between multiple groups.In order to examine the relationship between categorical variables, the chi-square test or Fisher's test was used.Te signifcance level was p � 0.05; however, statistically signifcant results were also indicated for the levels of p � 0.01 and p � 0.001.

Results
Table 1 shows general patient characteristics.Males comprised slightly more than 60% of the patients (n � 4753, 60.3%), while females nearly 40%.Te median age was 49.3 years; the standard deviation was over 42% of the mean value, which indicates signifcant age diversity.On an average, women were 55 years of age, while the mean age of men were 45.5 years.Te minimum age was identical in both groups (18 years), while the maximum age was slightly higher in women (104 years).Based on the analysis, it can be stated that there was a statistically signifcant age diference between men and women (p < 0.05).Te analysis of the weekly distribution of patients revealed that in general, most patients were admitted on Mondays (15.0%),Sundays (14.9%), and Tursdays (14.9%).Te fewest were admitted on Wednesdays (13.0%).No statistically signifcant diferences were observed between individual years and months.After implementation of inclusion criteria, all 7,864 patients were divided into 5 groups: (1) those discharged directly from ED or admitted to nonneurosurgical ward, (2) C1-qualifed for exigent neurosurgical intervention-direct transfer from ED to operating room, (3) C2-qualifed for neurosurgical intervention <24 hours, (4) C3-admitted to the Department of Neurosurgery or observed in ED for 24 hours in order to perform a control head CT scan and assess the progression/regression of posttraumatic lesions, (5) C4-no neurosurgical indication for hospitalization but patients were on anticoagulants, dangerous mechanism of trauma, were severely vomiting, confused, or/and without good family care.
For each patient, the main mechanism of trauma was determined and assigned to one of the 13 categories.Patients were most commonly hospitalized because of a fall (46.4%),RTC (16.9%), and beating (13.2%).Less common causes included hitting (9.4%), fainting (5.7%), epilepsy (5.7%), and pedestrian hit by a car (1.9%).Te other categories accounted for less than 1% of trauma mechanisms.All individual cases were classifed into the category "other" (rape was an exception distinguished separately).Almost 21% of the patients did not present at the hospital at the day of injury.Te majority presented between one and four days following the injury.However, some patients waited much longer, which can be seen by a large diference between the mean (about 4.5 days) and median (2 days) as well as a wide spread of observations (SD of about 12 days).Te median Glasgow Coma Scale (GCS) score was 15.Loss of consciousness was reported in slightly over 42% of the patients, and 0.5% recalled the event only partially.Over 21% of the patients were admitted under the infuence of alcohol, with a mean level of about 304 (±107) mg/dL.Neurological 2 Emergency Medicine International abnormalities were found in 7.1% of the patients, with anisocoria being the most common (19.7%).Confusion (13%), paresis of the limb (8.1%), and amnesia (7.3%) were less common.Complaints were reported in 48% of the patients.Among them, the average number of complaints was two (±1).Physical examination showed abnormalities in over 65% of the patients.One or two abnormalities were usually found, although the record number was 10 in some cases.Other radiological examinations (apart from head CT that qualifed patients for the study) were performed in almost 60% of the patients and included CT angiography (0.2%), X-ray (44%), CT (28.8%), and ultrasound (13.6%).It is worth noting that trauma CT was performed in 1.3% of the patients.Radiological examinations revealed 74 diferent abnormalities.Te most common were nasal fracture (4.6%), orbital fracture (2.9%), maxillary sinus fracture (2.6%), and fracture of at least one rib (2.4%).Detailed analysis is shown in Table 2.
Patients qualifed for urgent neurosurgery (C1) were signifcantly older compared with the others (mean of about 63 years vs. 49 years).However, no signifcant diferences were found between the C1 and C2 groups, meaning that patients with delayed surgery were not signifcantly younger than those qualifed for urgent neurosurgical intervention (63.31 vs. 58.76).Tere is a visible trend of a higher risk of more severe abnormalities in older patients; mean age changes depending on the group (C1: 63 years, C2: 59 years, others with diagnosed injury: 53 years, and no injury: 48 years).Age diferences between the groups are presented in Figure 2.However, it is worth noting that SD values of the age are quite similar in these groups.C1 patients who were admitted to the hospital on a day other than the day of injury presented signifcantly later than patients in the other groups (mean, 5.31 days vs. 4.67 days; median, 3.5 days vs. 2 days), including those in the C2 group (mean, 5.31 days vs. 2.68 days; median, 3.5 days vs. 2 days).Te long period before arrival at the hospital from the time of the injury in patients qualifed for urgent neurosurgical intervention resulted from the patient's refusal to transfer to the hospital immediately after the injury, explaining drowsiness/decrease in GCS with possible alcohol intoxication or loneliness and rare contact with the family.Which was easy to predict, GCS in C1 patients group had signifcantly lower scores than patients in the other groups (mean, 10.68 vs. 14.83;median 13 vs. 15), including those in the C2 group (mean, 10.68 vs. 12.99; median 13 vs. 15).Te proportion of patients who lost consciousness was higher in the C1 group than in the other groups (67.7% vs. 42%) with no signifcant diferences between the C1 and C2 groups.Data are summarized in Table 2.
Patients qualifed for instant transfer to the operating room were more often admitted to the hospital due to the fall (64.1% vs. 45.1%).Less common causes were RTC (4.7% vs. 16.5%) and beating (4.7% vs. 13.0%).Signifcance results for individual trauma mechanisms are shown in Table 3.To verify between which mechanisms of trauma the groups difered signifcantly, tests for diferences in proportions were carried out.Such an analysis showed that statistical signifcance (p value <0.05) was achieved for the fall and road trafc accident.Te former was more common in the     Emergency Medicine International C1 group than in the other groups (64.1% vs. 45.1%), while the latter less common (4.7% vs. 16.5%).To verify diferences between all patients who had neurosurgery intervention (<24 from transfer to hospital) with patient without neurosurgery during hospital stay, between which mechanisms of trauma and the groups difered signifcantly with tests for diferences in proportions were carried out again (C1 + C2 group vs. rest of the patients).Tis analysis showed that statistical signifcance (p value <0.05) was achieved for the fall and RTC.Te former was more common in the C1 and C2 groups than in the other groups (61.5% vs. 46%), while the latter less common (10.4% vs. 17.1%).Diferences were also observed for beating (6.5% vs. 13.5%) and hitting (3.9% vs. 9.6%), which occurred less frequently in the C1 and C2 groups than in the others.In contrast, epilepsy-related injuries were more common in the C1 and C2 groups (10.8% vs. 5.5%).

Emergency Medicine International
Te number of patients under the infuence of alcohol between C1 and all other head trauma patients' groups was comparable and with no statistically signifcant diference.However, C1 patients under the infuence of alcohol had higher mean blood alcohol levels than patients from other groups (mean, 476 mg/dL vs. 303.98md/dL; median 476 mg/ dL vs. 306 mg/dL) (Figure 3).Te alcohol level was associated with a signifcantly lower risk of death.Using the test for diferences in proportions, the occurrence of neurological abnormalities was compared between C1 patients and all the other patients.Signifcant diferences (p value <0.05) are summarized in Table 4.
Neurological abnormalities were signifcantly more commonly reported in C1 group patients (71.9% vs. 39.9%).Te C1 and C2 groups did not difer signifcantly in specifc neurological abnormalities (p value >0.05).Every patient admitted for urgent neurosurgical intervention had a concurrent injury of another body part.Compared with other patients, a signifcantly higher percentage of patients admitted for urgent neurosurgical intervention (C1) had the following: deformity of the lower leg, bleeding from the mouth, bleeding from the nose, hematoma in the parietal region, wound of the occiput, wound in the temporal region, contusion of the shoulder, contusion of the occiput, and palpable skull fracture.Among the complaints, signifcant diferences were found for pain in the cervical spine/neck which was uncommon in the C1 group (0.0% vs. 12.2%).Te following Table 5 shows signs and symptoms classifed into those found on physical examination and those reported by the patients.Tere were a lot of complaints signifcantly diferentiating the C1 and C2 groups from all the other head trauma patients.Tese included the following: Among the complaints, signifcant diferences were found for the following sings and/or symptoms: Pain in the cervical spine/neck-less common in C1 and C2 patients than in the other patients with any signs and/or symptoms (3.7% vs. 12.3%, p < 0.001), vomiting-more common in C1 and C2 patients than in the other patients with any signs and/or symptoms (7.9% vs. 3.7%, p � 0.006).Death occurred more frequently in C1 patients than the rest of the patients (32.8% vs. 0.6%) and more often in the C1 or C2 groups than the rest of the patients (19.5% vs. 0.2%).

Discussion
Te primary assumption of this article was, based on the most signifcant possible number of subjects, to extract new data on prehospital abnormalities in physical examination, which may indicate the need for urgent neurosurgical intervention.TBI remains the leading cause of death, longterm disabilities in previously healthy adults.Head trauma is one of the main causes of death in trauma patients.In North America, TBI kills 7 patients every hour [2,4].Although exact number is hard to determine, it is assessed that TBI afects 27-69 million people and their families every year [6,7].Due to the poorly explored efective treatments for TBI (most guidelines are based on weak data), surviving patients leave the hospital requiring many years of intensive rehabilitation, especially when they are in a minimally conscious state or in a persistent vegetative state [7].According to a study conducted in 16 European countries, the number of years of life lost due to head trauma in Europe in 2013 was 1.3 million overall (1.1 million in males and 271,000 in females) [8].From the perspective of head trauma patients, it is extremely important to initiate appropriate procedures to prevent secondary brain injury as early as 8 Emergency Medicine International possible.Maintaining normoxia and normotension are crucial.Pathophysiological reactions leading to posttraumatic coagulopathy (SWINE), fever, and glucose level fuctuations are other processes that threaten the development of secondary brain injury and decrease the chances of successful treatment completion.A single drop of systolic blood pressure <100 mmHg doubles mortality, and if SBP drops below 70 mmHg, it increases the death risk 6-fold [9].Te authors focused on the most commonly reported patient's complaints and on abnormal physical examination fndings-apart from evident external symptoms of head trauma and a low GCS score-that may raise suspicion of head trauma requiring urgent neurosurgical intervention.Our analysis revealed few interesting insights, a lot corresponds to the data from most European countries, where a shift in the age of head trauma patients from young patients to those >65 years old has been observed in recent years [10].In contrast with a paper by Bossers et al. [7], the most common cause of injuries in our study was falls on the same level.Te incidence of serious injuries in the elderly is increasing, and the most common cause of injury is falls; falls are also the main cause of death in this group of patients [11].Motor vehicle accidents were the most common cause of head trauma in a Dutch study but not even ranked among subjects who were qualifed for urgent neurosurgical intervention in our study.All kinds of vehicles accidents were preceded by falls and resulted from epilepsy to beating.An aging society, the growing population of people who use various types of medications which afect blood clotting, polypharmacy, and the resulting falls all combine with poorly explored treatments for TBI to create a situation in which nearly one in fve patients does not survive until discharge.
Currently, we do not have any efective prehospital tool for assessing patients' eligibility for urgent neurosurgical intervention.Te GCS, as a simple and well-known scale, can yield widely varying scores and is useless after the administration of sedatives and analgesics-which is invariably part of the treatment in posttrauma patients-as well as in patients under the infuence of alcohol.Te pupillary assessment remains important, though 20% of the population has physiologic anisocoria and the clinical response to light may be disturbed by the administration of analgesics and sedatives.Also, the majority of people with posttraumatic anisocoria show no changes in CT.Norwegian HEMS project to implement CT scanner into their aircrafts, assays of SB100 in the patient's saliva or of the blood markers ubiquitin C-terminal hydrolase-L1 (UCHL-1) and glial fbrillary acidic protein (GFAP), or IT programs for evaluating indications for neurosurgical intervention, are unlikely to be introduced into the daily practice of prehospital teams for years to come [12][13][14].Nonetheless, studies suggest that even basic prehospital emergency procedures in head trauma patients, such as ensuring appropriate oxygenation

Emergency Medicine International
or maintaining normal pressure, are not performed correctly despite explicit guidelines [15].In many countries, farreaching attempts to prevent secondary brain injury are undertaken, such as smooth driving on the way to the hospital due to the likely negative impact of braking on intracranial pressure [16].
3.1.Future Research and Limitations.Our study on "Prehospital predictors for urgent neurosurgical intervention in the head trauma patient" yielded several intriguing fndings.Confrmation of these fndings through prospective studies conducted by other centres could potentially lead to the implementation of early response systems during the anamnesis collection stage.With the aid of computer systems, we can access information on the mechanism of injury, external injuries, and the patients' age, even at the dispatcher stage, which allows the neurosurgical facilities to be available when the patient arrives in the hospital.We do not believe the addition or introduction of a new scale or tool would beneft these patients.Our model demonstrated the feasibility of establishing an out-of-hospital early warning system.Te frst limitation is that this is a retrospective study.Te environment in the emergency department is one typically characterized by urgency, and so, medical personnel may not be able to identify all signs and symptoms of trauma during the data collection stage, especially during intensive and life-threatening situations, where time constraints would pose a threat to the patient's life.Although it is a retrospective study, we thoroughly reviewed each patient's record to ensure that no signs of trauma were missed.All data collected including ambulance cards, ED cards, and all neurosurgical records were carefully checked not to miss any signs of trauma, that could impact the results of our study, but of course, it can have some biases.Another limitation of this study was that it did not account for potential confounding variables that may have infuenced the outcome, such as the severity of the trauma, comorbidities, or medications taken by the patient.Defnitely, the lack of data on the medication taken is an important limitation of the study, and the intake of anticoagulants can certainly afect the prognosis.Undoubtedly, a signifcant limitation of the study stems from the inadequate access to data concerning the use of medications (specifcally anticoagulants), chronic illnesses, surgical procedures, and the presence of cancer in the participants' medical histories..However, from our experience in prehospital medicine, we often do not know the list of medications taken by unconscious patients.Are we able to determine the severity of the injury (how many stairs the patient fell down) of an unconscious patient found by his family at home?We are almost certain that at the prehospital and early-hospital level, it is necessary to simplify similar examinations as much as possible and the mechanism of injury and visible injuries must be enough for us to assess the patient because the rest can only be guessed at.Furthermore, this study was limited by the fact that is based only on two hospitals, although the two largest in our region.One was a large trauma centre and the second an emergency department which serves 60,000 patients/year.Moreover, the treatment strategies used in this study refect the convention and tendencies used in our centres.

Conclusion
TBI remains a condition that devastates the lives of patients and their families.Efective treatments for TBI are disproportionately poorly explored.Tis study indicates many statistically signifcant demographic, mechanism of trauma, and abnormal examination fndings in head trauma patients, which may suggest that transfer to a facility with a neurosurgery department is required as early as the prehospital care stage.While patients admitted to neurosurgery departments constitute a small percentage of TBI patients, this group is characterized by high mortality and their prognosis for normal performance status upon discharge is poor.

Figure 1 :
Figure 1: Patients admitted to Emergency Medicine Departments of two clinical hospitals who underwent head CT scan.

Table 1 :
General characteristics of all head trauma patients.

Table 2 :
Comparison of patient characteristic variables in relation to the variable: classifcation Te bold values indicate signifcance.

Table 3 :
Detailed summary of the mechanisms of trauma in relation to patient classifcation.
* p value of signifcance test for diferences in proportions.Te bold values indicate signifcance.

Table 4 :
Detailed summary of neurological abnormalities in relation to patient classifcation.

Table 5 :
Detailed summary of signs and symptoms in relation to patient classifcation.
Te bold values indicate signifcance.