Introduction
Traumatic brain injury (TBI) is the leading cause of death and disability among the young around the world [
1]. The American College of Surgeons Committee on Trauma and Centers for Disease and Prevention have developed field triage guidelines for TBI patients, which are in conjunction with the Brain Trauma Foundation’s pre-hospital guidelines, internationally acknowledged cornerstones of pre-hospital TBI triaging and treatment [
2,
3]. The guidelines recommend direct transport of patients with TBI to hospitals with availability of neurosurgical care including computerized tomography (CT) scanning, neurosurgical care, intracranial monitoring and treatment [
3]. Accordance with the guidelines has been shown to improve outcome in TBI patients [
4‐
6]. Despite this nearly half of all TBI patients are initially transported to a hospital without neurosurgical expertise, before inter-transfer to a hospital with neurosurgical capability, which has been shown to increase mortality [
5,
7].
Accordingly, we sought to assess pre-hospital factors associated with an initial transport to a non-neurosurgical hospital in a region where all TBI care is centralized to one single trauma center.
Discussion
The centralization of TBI care to specialized trauma centers have improved outcome [
15,
16]. Direct transport of patients, from the scene of injury, to such specialized trauma centers has been shown to significantly increase survival [
5,
6,
15]. Thus, initial pre-hospital triage and subsequent transport is of major role in the handling of TBI patients. The purpose of this study was to investigate pre-hospital factors associated with an inappropriate transport of TBI patients in a region where all TBI care is centralized to one single trauma center. Our study showed that despite a regional trauma care protocol, mandating direct transport of TBI patients that are likely to specialized care to the TC, over one third are initially transported to a hospital without the capacity of handling these patients. We identified several factors associated with transport to another hospital than the designated trauma center, i.e. male gender, low energy of trauma, lack of pre-hospital physician involvement, absence of major extra cranial injuries, normal pupillary light reflex and severe alcohol intoxication. Our re-transfer rate is somewhat lower than previously described, indicating a well-organized system in the region [
7,
17]. A probable major factor influencing this is the availability of pre-hospital physicians in the study region.
Several studies have showed an improved outcome after direct admission to a neurosurgical trauma center, although there is no class I evidence to back this up. Härtl et al. showed that patients transported indirectly to a neurosurgical trauma center, via a lower level of care hospital, had a 50% higher risk of death than patients transported directly to a trauma center with neurosurgical expertise [
5]. Hannan et al. reported an almost twice as low risk for in patients with TBI treated in a regional trauma center compared to local non-trauma centers [
19]. In our study, we did not find any association between initial transport destination and outcome. Despite the fact that patients who were initially transported to a local hospital (delayed admission) were more severely injured than the patients in the direct admission groups. Further sub-group analysis showed that patients in the delayed admission group more often required acute craniotomy than patients in the direct admission group. As such one might expect a higher mortality among these patients, as it has been shown in numerous studies before [
20‐
23]. However, this as well was not evident in our study. The reason why we could not find any correlation between initial transport destination and outcome is probable due to lack of power. Another possible reason might be that even though the indirectly transported patients had a significantly longer time delay to trauma center admission than the directly transported patients the delay wasn’t long enough to affect outcome. This is supported by the fact that the majority of patients in both the direct and delayed admission group undergoing acute craniotomy was operated on well within acknowledged time limits [
18,
24]. Moreover, we found no significant difference in time to trauma center admission between survivors and non-survivors. This supports the earlier statement regarding a well working trauma system in the region. Another factor and major strength to the present study that has to be considered is that we used 6-month mortality as outcome measure whereas most other studies have used in-hospital mortality. In-hospital mortality is not an optimal marker for measuring TBI as it severely underestimates the number of deaths [
11]. Thus, it may simply be that indirect transport increases the risk of short-term but not long-term mortality. However, no statistical significant difference in 14-day mortality was noted between the direct and delayed admission groups in univariate analysis.
Not surprisingly a high field GCS and the absence of major extra cranial injuries associated with delayed trauma center admission. However, field assessment of injury severity was proven not to be a reliable marker of injury severity as post-admission investigations revealed that patients in the delayed admission group had higher injury severity scores and that there were no differences in TBI severity, despite contrary signs in the pre-hospital setting. The GCS has been shown to be of limited value in determining TBI severity in the field [
25]. Fourteen percent of patients with an initial GCS of 14 develops an intracranial lesion [
26]. Intracranial lesions in patients with an initial GCS of 15 are uncommon unless risk factors are present, such as high age, use of anti-coagulants, focal neurological symptoms and alcohol-intoxication [
27,
28]. Approximately 3-6% of all TBI patients who die have an field GCS of 13 to 15 [
29,
30]. This is of significance in pre-hospital care as lucid patients with a high initial GCS easily can be under triaged and the possibility of an intracranial lesion overlooked. This problem was noted in our study as 31% of all patients with a field GCS of 13 to 15 required acute craniotomy. Furthermore, 15% of all patients with a field GCS of 13 to 15 died (compared to 32% of those with a field GCS of 3 to 12). It should however be noted that field GCS was measured by a paramedic if a pre-hospital physician was not present (direct admission 35%, delayed admission 84%) but TC GCS was always measured by an emergency department physician (trauma surgeon, anesthesiologist or neurosurgeon). This potential confounding factor in determining accurately determining the GCS should be the focus of future studies.
Systemic hypotension and hypoxia in the pre-hospital setting are associated with poor outcome in TBI patients [
3]. Twenty percent of all patients had a hypoxic insult (SO2< 90%) and eleven percent a hypotensive insult (< 90 mmHg) during the pre-hospital transport. In univariate analysis both hypotension and hypoxia were significantly associated with direct admission. However, in multivariate analysis hypotension but not hypoxia was showed significant association with direct admission. To further investigate this we looked at the number of hypotensive and hypoxic patients accompanied by a pre-hospital physician, as the presence of a physician was the strongest independent predictor of direct admission in the logistic regression analysis. However, we found no significant difference in hypotensive respectively hypoxic patients accompanied by a physician (76%, 70%).
Securing the airway and adequate oxygenation and ventilation is of vital role in the pre-hospital management of patients with TBI [
3]. A significant higher proportion of patients in the direct admission group were intubated on the field compared to patients in the delayed admission group. Of all patients intubated in the field 82% were accompanied by a physician (3 out of 7 in the delayed admission group and 133 out of 158 in the direct admission group). However, a considerable proportion of patients in the delayed admission group were intubated at the local hospital before transfer to the TC. Notable is that there were no significant difference in time to TC admission between patients intubated and not intubated at the local hospital before transfer to the TC.
We found that severe alcohol intoxication was associated with delayed admission. Alcohol intoxication in TBI patients is an internationally acknowledged problem [
1]. Studies have shown that roughly half of TBI patients are alcohol-intoxicated at the time of injury and that alcohol intoxication increases the risk of TBI, especially as a result of binge drinking [
31‐
34]. Alcohol has also been shown to impede with the clinical TBI diagnosis by having a level of consciousness lowering effect [
25]. This makes it challenging for the EMS to judge whether a trauma patient’s altered level of consciousness (i.e. GCS) is caused by alcohol intoxication, intracranial lesion or other injuries. Ground level falls is the major cause of injury among alcohol intoxicated patients and the elderly [
35]. Our results show that low energy traumas, such as GLF, significantly increase the risk of delayed admission. Thus, GLF might be an underestimated cause of injury in TBI patients. Though, it has been shown that there is no difference in TBI severity caused by GLFs compared to other causes of injury, such as RTA [
36]. Hence, the elderly and alcohol-intoxicated patients may be at the core of this problem; pre-hospital mistriage leading to delayed life-saving treatment.
Altogether, there are many contributing factors contributing to difficult triage decisions with the risk of both under or over triage of these patients. Over triaging is not a viable option as this would lead to an excess trauma center burdening and under triage with the risks of having devastating effects on patient prognosis. It is in such cases that the importance of pre-hospital physician consultation is highlighted. We acknowledge the fact that not all regions have pre-hospital physicians available 24/7, however in areas of high trauma burdening this could be an opportunity worth exploring.
In conclusion our study shows that some patients with TBI are often misdiagnosed in the field due to a number of reasons and that there is a major need for further studies exploring diagnostic and educational means to improve rapid identification of those requiring neurosurgical treatment in patients whose initial clinical presentation is confounded by factors such as low level of initial symptoms and alcohol intoxication.
Study limitations
Our study has several limitations. Most important, due to the retrospective nature of this study we were unable to collect the number of patients with a TBI requiring intensive care initially transported to another hospital were the patient died before reaching the trauma center’s emergency department. Likewise we are unaware of the number patients dying at the scene of injury before reaching the TC, or at the local hospital prior to trauma center transfer. However, in a large prospective study by Myburgh et al. they found that only 3% of patients with TBI die during the pre-intensive period [
7]. We believe that these numbers are representative for our study as it is very rare that trauma patients requiring neurosurgical expertise are not referred to the designated TC (directly or indirectly) since the other hospitals do not offer any neurosurgical services. Second, BAL is not routinely tested for in the pre-hospital setting without indications, giving us a substantial amount of patients untested for BAL. It is clear that BAL testing is not indicated nor is it possible for everyone and this is probably the current situation in most countries where BAL is mainly tested when it might impact treatment decisions. Furthermore, measuring BAL in the pre-hospital setting using an alcohol breath-test is challenging in TBI patients considering the nature of the disease (potential lack of co-operation and unconscious patients). However, following TC admission a total of 48% of patients were tested for blood alcohol levels using standard laboratory blood samples. This was shown to correlate extremely well with the BAL measured in the pre-hospital setting. Third, in the present retrospective study we could not note any statistical significant association between initial transport location and long-term mortality, which is probably a consequence of lack of power.
Competing interests
The authors declare that they have no competing interests.
Authors’ contribution
RR, JS and MS designed the study. RR drafted the manuscript assisted by MS, JS and JL, RR, TB and RK performed the data collection. RR is responsible for integrity of the collected data. The statistical analysis of the data was performed and interpreted by RR, JS and MS. RK viewed and interpreted CT scans of the included patients. MS, JL contributed to the interpretation of the data and writing of the manuscript. All authors revised the manuscript and approved it in the final form.