Article Text

Download PDFPDF
Is direct transport to a trauma centre best for patients with severe traumatic brain injury? A study in south-central Taiwan
  1. Kuang-Yu Hsiao1,2,3,
  2. I-Chuan Chen1,2,3,
  3. Chia-Jung Yang1,
  4. Cheng-Ting Hsiao1,2,
  5. Kai-Hua Chen2,4
  1. 1Department of Emergency Medicine, Chang Gung Memorial Hospital, Chiayi, Taiwan
  2. 2College of Medicine, Chang Gung University, Taoyuan, Taiwan
  3. 3Chang Gung Institute of Technology, Chiayi, Taiwan
  4. 4Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, Chiayi, Taiwan
  1. Correspondence to Dr Kai-Hua Chen, Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, No. 6, West Sec., Chia-Pu Road, Putz City, Chiayi County 613, Taiwan; karenkhchen{at}


Objective This study attempted to identify any differences between the outcomes of patients with severe traumatic brain injury (TBI) who were directly transported to Chang Gung Memorial Hospital and those who were stabilised initially at other hospitals in south-central Taiwan.

Methods A retrospective review of the records of 254 patients with isolated severe TBI who visited this hospital's emergency department from July 2003 to June 2008, of whom 167 were referred from other hospitals. Logistic regression was used to assess the effects of transfer and its components on mortality.

Results Transfer from another hospital was not significantly correlated with mortality in this study (OR 0.513, 95% CI 0.240 to 1.097). Moreover, neither intubation (OR 1.356, 95% CI 0.445 to 4.133) nor transfer time over 4 h (OR 0.549, 95% CI 0.119 to 1.744) had a significant effect on mortality.

Conclusion No differences in outcome were found between patients with isolated severe TBI who were directly transported and those transferred to this hospital's emergency room.

  • Emergency care systems
  • trauma

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Traumatic brain injury (TBI) constitutes a major proportion of trauma injuries. The damage to the brain from trauma is caused not only by direct impact, but also secondary insults, such as hypoxia and hypotension.1 Although it seems logical that the correction of such reversible insults should be started as early as possible, even during prehospital care, such measures may cause delays before the patient arrives for definitive treatment; and the benefits of such early treatment remain a matter of debate. For example, studies have shown that aggressive airway management in the prehospital setting did not absolutely lead to better outcomes.2–4 In a recent study, Berlot et al5 demonstrated that patients with severe TBI had better outcomes when they were sent by helicopter and received more aggressive prehospital treatment, although this involved more transportation time, when compared with those who were sent by ground transport. Svenson et al6 found 44% of CT scans could be omitted because patients received the same studies after transfer and spent more time at their initial hospitals if they received CT scans. Sampalis et al7 found that directly transporting severely injured patients from the scene to level 1 trauma centres could improve patient outcomes. Similar results were also found for patients with TBI.8

In our region, advanced airway management (eg, orotracheal intubation or laryngeal mask) and intravenous fluid resuscitation are rarely performed by emergency medical technicians (EMT). Patients transported by emergency medical system (EMS) usually get oxygen supplement, immobilisation and basic life support as needed. Patients with respiratory arrest usually receive bag–valve-mask ventilation during resuscitation, but sometimes laryngeal mask instead (if the EMT has been trained and personal preference). In our region, the EMS usually take the patient to the nearest hospital. However, bypassing the nearest hospital to another hospital occurs as a result of the request of the patient or their family. Patients with spontaneous breathing will not be intubated (with orotracheal tube or laryngeal mask) until they arrive at the hospital. Because the delivery of prehospital care in our region is different from that described in previous studies, we designed this retrospective study to compare the outcomes of patients with severe TBI who were directly transported to our hospital with those of patients who were stabilised initially at other hospitals in our region.

Materials and methods

Patient selection

Our hospital is a university-affiliated general hospital in south-central Taiwan. Approximately 15% of the 65 000 patients who visit our emergency department (ED) annually are trauma patients. There are 1000 beds in the acute care ward and 100 beds in the intensive care unit of the hospital, and neurosurgeons and operating rooms are available on a 24 h basis. Currently, the hospital is classified as ‘severe’ class for emergency care, which is similar to a level 1 trauma centre in some other trauma systems. The rating system for emergency care facilities (including trauma care) among hospitals began after the period of this study; therefore, no good guides were available when patients sought help for their injuries during our study (July 2003 to June 2008). In our hospital, patients with head injury with a Glasgow Coma Scale (GCS) lower than 9 will be intubated whether they will receive further treatment (such as operation) or not. However, the policy of definitive airway creation in other hospitals could be based on either compromised airway or GCS less than 9. The criterion for inclusion was the diagnosis of severe TBI: GCS score 3∼8 after initial resuscitation at the ED. Patients with any of the following criteria were excluded: loss of vital signs at the scene or any time before arrival at hospital; multiple traumas; penetrating brain injury; less than 18 years of age; and GCS score greater than 8 after drugs (including alcohol) were eliminated. We defined multiple traumas as either any abbreviated injury scale score other than to the head of more than 2 or a sum of abbreviated injury scale scores other than to the head of more than 3. A total of 254 patients was enrolled, of whom 167 were referred from other hospitals. Clinical data were collected retrospectively from the charts and analysed.

Data analysis

Age, sex, trauma mechanism, temperature, blood pressure, GCS score, airway management before arrival at the hospital, laboratory data, surgical intervention, and mortality or survival on discharge were recorded as variables for analysis. We defined these variables as follows: hypotension, systolic blood pressure less than 90 mm Hg; hypertension, systolic pressure greater than 180 mm Hg; acidosis, pH less than 7.35 at arterial blood gas; hypothermia, body temperature less than 36.0°C; hyperthermia, body temperature 38.0°C or greater; hyperglycaemia, blood sugar more than 200 mg/dl. Brain CT reports were not included for analysis because some were performed only at the initial hospital, and no official reports created by the radiologist could be collected at the chart review.

Differences between patients directly transported (direct group) or transferred from other hospitals (transferred group) to our hospital were assessed using the χ2 test for categorical variables. The independent t test was used to test the differences for continuous variables. The logistic model was applied for variables with p values less than 0.1 for the differences between the direct and transferred groups to identify the possible effects on mortality. Differences between mortality and survival were also assessed using the χ2 and independent t test as described previously. Factors with p values less than 0.1 for the difference between the groups were also put into the logistic regression model. Because transfer is the major factor we investigated, it was included in the logistic regression model. The logistic regression model was using the enter procedure. The OR for factors influencing death was estimated by logistic regression. Furthermore, we also created a logistic regression model to assess the effects of variables on mortality for subgroups of the transferred patients. Time from arrival at the first ED to ours and airway management at the initial hospital were considered important factors that might influence mortality, so both were included in the logistic regression model for the subgroups of transferred patients. Otherwise, all factors included in the logistic regression model for the transferred subgroups were the same as those in the first model, and equivalent procedures were used.

All statistical assessments are two-sided and evaluated at the 0.05 level of significant difference. Statistical analyses were performed using SPSS 17.0 statistics software.


There was no significant difference between the two groups with regard to sex, trauma mechanism, number of patients with hypotension during their ED stay, GCS scores, or rate of mortality (table 1). However, patients in the direct group were significantly older, a higher percentage had hypertension and a lower percentage had hyperglycaemia.

Table 1

Patients' characteristics

The significant differences between patients who survived and those who did not included age, GCS score, percentage with hypotension, hypertension, hyperthermia, hyperglycaemia and surgical treatment (table 2).

Table 2

Variables according to survival and mortality

After controlling for the factors significantly related to mortality, the effect of transfer was not significantly correlated with mortality (table 3).

Table 3

Logistic regression analyses identifying the effects of variables on mortality

Among the patients in the transferred subgroup, neither intubation nor transfer time greater than 4 h had a significant effect on mortality according to the logistic regression model (table 4).

Table 4

Effects of variables on mortality for transferred patients


In our study, we found no differences in mortality between patients with TBI who were directly transported and those transferred from other hospitals to our hospital. Our hospital is located in a rural area of Taiwan. However, because of the high density of hospitals and small geographical area of Taiwan, most traumas occur at locations within 30 min of a hospital. Studies showed that trauma patients' outcomes improved after a regional trauma system was established;9–12 this was also observed among patients with TBI.13 Direct transport of patients with severe injuries to higher level trauma centres has been suggested to improve outcomes, but there have also been conflicting results reported for patients transferred directly to trauma centres in rural settings. Sampalis et al7 demonstrated that the direct transport of patients to tertiary trauma centres in rural trauma systems improved patient survival, but another study focusing on a rural area showed no reduction in mortality among patients taken to a trauma centre instead of the nearest hospital.14 In our previous study (one of the few studies regarding our area), we found a higher mortality and need for care in the intensive care unit among patients with head injuries who were transferred from other hospitals, but these differences did not reach statistical significance.15 Once again in the current study, we found no better outcome when patients were sent directly to our hospital.

Another important variable in the trauma system is the type of prehospital care administered. In this study, we found no differences in mortality among transferred patients with or without intubation. Although only basic airway management was provided in the prehospital setting in our area, it is important to consider the effect of airway management in more detail because results from other regions might not be suitable for ours. Some studies that focused on patients with TBI found that aggressive prehospital treatment increased patient survival,3 5 but conflicting results were demonstrated by other studies that focused on all patients with TBI or those in whom prehospital paramedics performed rapid sequence intubation.2 16 One study also suggested that blood carbon dioxide should be maintained at a level between 30 and 35 mm Hg to achieve better outcomes.17 Although early correction of hypoxia may improve the outcomes of patient with TBI, it is not easy to maintain adequate ventilation during transportation; thus, the benefit may possibly be reduced. In another study regarding prehospital ventilation, Stockinger et al18 showed no survival advantage for intubation over bag–valve–mask ventilation. It seems that airway management may be associated with mortality; but the effect is more complicated than simply intubation for all patients with severe TBI, and additional studies focusing on airway management are still needed. We also found that hypotension was significantly associated with poor survival. However, fluid resuscitation is rarely if ever initiated by the EMT in our area. Therefore, the fact that hypotension could not be corrected in a prehospital setting supports the argument in favour of taking patients with severe TBI to the nearest hospital.

We also assessed the effect of delay during transfer on mortality, but our study revealed no differences in outcomes between transfers that took less than or more than 4 h. Harrington et al19 demonstrated similar results in a study focused generally on patients with trauma. Our study differed in that it focused on patients with isolated severe TBI; whereas patients with multiple traumas, which are more time consuming, were excluded. Furthermore, in most cases, transfers took less than 1 h because of the geography of Taiwan; therefore, a long delay before patients receive definitive care is not the usual case. However, given the retrospective nature of our study, further assessment of factors regarding delayed transfer was not possible. In addition, there were only small numbers of patients who underwent transfer that lasted over 4 h in our study, which may have influenced the outcome in the logistic regression model.

In our study, we found a higher mortality rate in our patients compared with other hospitals in general but similar to those with less aggressive treatment.20 A possibility leading to a higher mortality rate is that the families tend to decline aggressive treatments for patients with very poor prognosis in our culture, especially when the patient is elderly, although improvement of survival by surgical decompression has been found in our study. The differences between the EMS should be considered as a possible factor of such a result. Again, we emphasise that EMS provides only basic and less invasive prehospital care in our region, and it is quite different from western countries. These basic differences may cause different results, and results from other systems may not be applied perfectly to ours.

Most of the limitations of this study arose from its relatively small sample size and its retrospective nature. We excluded patients who sustained multiple traumas in order to eliminate the possible confounding effects of other injuries. However, selection bias may exist because we did not exclude patients who died soon after being directly transported to our hospital. In contrast, severely injured patients with no hope of survival might not have been transferred from their initial hospitals. This could generate an apparent worse outcome among the direct group during statistical analysis. The lack of brain CT results in our study may also have introduced some bias in our logistic regression model. Besides, bypass during EMS transport could be another possible selection bias. Although bypass (requested by family) in severe trauma patients is rare in our system, it could not be completely eliminated by this chart review based study. Therefore, one should be careful to consider our results in context and recognise they cannot be applied to all trauma patients, because other factors may play significant roles in mortality among patients in other settings.

In conclusion, we found no differences in mortality between directly transported and transferred patients with isolated severe TBI. Neither intubation before transfer nor delayed transfer was significantly associated with inhospital mortality. However, to achieve the best outcomes for patients with severe TBI, more prospective studies focusing on the components of the trauma system in our region are warranted.



  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Chang Gung Memorial Hospital, Taiwan.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • Primary survey
    Darren Walter