Article Text

PDF

Glasgow Coma Scale is unreliable for the prediction of severe head injury in elderly trauma patients
  1. A Kehoe1,
  2. S Rennie1,
  3. J E Smith1,2
  1. 1Emergency Department, Derriford Hospital, Plymouth, UK
  2. 2Academic Department of Military Emergency Medicine, Royal Centre for Defence Medicine (Research & Academia), Medical Directorate, Joint Medical Command, Birmingham, UK
  1. Correspondence to Dr J E Smith, Emergency Department, Derriford Hospital, Derriford Road Plymouth PL6 8DH, UK; jasonesmith{at}doctors.org.uk

Abstract

Objectives and background Elderly patients comprise an ever-increasing proportion of major trauma patients. The presenting GCS in elderly patients with traumatic brain injury (TBI) may not reflect the severity of injury as accurately as it does in the younger patient population. However, GCS is often used as part of the decision tool to define the population transferred directly to a major trauma centre. The aim of this study was to explore the relationship between age and presenting GCS in patients with isolated TBI.

Methods A retrospective database review was undertaken using the Trauma Audit and Research Network database. All patients presenting to Derriford Hospital, Plymouth, between 1 January 2009 and 31 May 2014 with isolated TBI were included. Demographic, mechanistic, physiological, resource use and outcome data were collected. Abbreviated injury scale (AIS) was recorded for all patients. Patients were categorised into those older and younger than 65 years on presentation. Distribution of GCS, categorised into severe (GCS 3–8), moderate (GCS 9–12) and mild TBI (13–15), was compared between the age groups. Median GCS at each AIS level was also compared.

Results The distribution of GCS differed between young and old patients with TBI (22.1% vs 9.8% had a GCS 3–8, respectively) despite a higher burden of anatomical injury in the elderly group. Presenting GCS was higher in the elderly at each level of AIS. The difference was more apparent in the presence of more severe injury (AIS 5).

Conclusions Elderly patients who have sustained isolated severe TBI may present with a higher GCS than younger patients. Triage tools using GCS may need to be modified and validated for use in elderly patients with TBI.

  • aged
  • triage
  • clinical assessment
  • Trauma, head

Statistics from Altmetric.com

Introduction

Emerging evidence demonstrates that the elderly comprise an ever-increasing proportion of the major trauma population, with low falls becoming the predominant mechanism of injury.1 The implications of this, and challenges to the traditional approach to trauma management, are yet to be fully defined.

Derriford Hospital is the major trauma centre (MTC) for the Peninsula Trauma Network. The population it serves contains a high proportion of elderly patients, and the impact of these demographic changes may therefore become apparent here first.

Many trauma triage tools include the Glasgow Coma Scale (GCS) when defining those patients requiring direct transfer to an MTC rather than the nearest hospital. The reliability of the GCS to predict significant brain injury has been used as part of several triage tools in the last few decades, from the original triage and trauma scores,2 to the triage-revised trauma score.3 Another triage tool (the Wessex trauma triage tool) is currently in use in the Peninsula Trauma Network, which uses a combination of anatomical and physiological criteria (including the motor score of the GCS) to identify those patients suitable for transfer directly to an MTC.4

It has been suggested that the presenting GCS in patients with isolated traumatic brain injury (TBI) may not reflect the severity of injury as accurately in the elderly as it does in the young.5 ,6

The aim of this study was to explore the relationship between age and presenting GCS in patients with isolated TBI in our population.

Methods

The Trauma Audit and Research Network (TARN) database was interrogated to identify patients attending Derriford Hospital, Plymouth, between 1 January 2009 and 31 May 2014 with significant, isolated TBI (abbreviated injury scale (AIS) head of 3, 4 and 5, with no other TARN-qualifying injury with AIS >2). The AIS gives an indication of the severity of the anatomical injury and can be broadly thought of as representing minor injury (1), moderate injury (2) and serious to critical (3–5), with a maximal injury (6) thought to be unsurvivable.7 For example, a base of skull fracture is graded as AIS 3, a small subdural haematoma (≤50 cm3) is AIS 4, a massive extradural haematoma is AIS 5 and transection of the brain stem is AIS 6.

TARN eligibility includes trauma patients of any age who are admitted to hospital for 72 h or more, are admitted to a critical care unit, die in hospital or are transferred to another hospital for specialist care. Specific patient groups, such as those aged over 65 years, who are admitted due to a fracture of the neck of femur, are excluded.8 ,9 Patients who did not meet TARN eligibility criteria were not included in this study.

Entries with missing data were also excluded. Demographic, mechanistic, physiological, resource use and outcome data were collected. Patients were categorised into those older and younger than 65 years on presentation. Presenting GCS values were not normally distributed. The difference in distribution of presenting GCS between age groups was therefore considered by comparing medians and IQRs for the whole study population and subgroups defined by AIS score and also by categorisation into severe (GCS 3–8), moderate (GCS 9–12) and mild TBI (GCS 13–15) groups according to Teasdale and Jennett's original classification.10–12

Categorical data were compared using χ2 analysis or Fisher's exact test for two-by-two tables. Continuous data were not normally distributed and were therefore compared using the Mann–Whitney U test. Significance was set at p<0.05. No adjustment was made for multiple comparisons.

The Health Research Authority regulates the use of patient information in the UK and has given ethical approval for research using anonymised TARN data (approval number: ECC 7-05(g)/2011).

Results

During the study period, 3249 cases were entered onto the local TARN database. TARN data completeness has averaged 98% over the last 3 years. In total, 874 cases met the inclusion criteria. In 313 cases, initial GCS was not recorded, leaving 561 cases for analysis, whose data are presented in table 1. Outcomes and resource use differed significantly between the two age groups (table 2). The distribution of presenting GCS was significantly higher in elderly patients than in younger patients overall, and at each AIS severity (table 3). Median presenting GCS was higher in elderly patients at AIS 5, for patients who underwent a procedure and those who ultimately died (table 3).

Table 1

Characteristics of patients with isolated traumatic brain injury categorised by age group

Table 2

Outcomes of patients with isolated traumatic brain injury categorised by age group

Table 3

Presenting GCS of patients with isolated traumatic brain injury categorised by age group

Discussion

This study shows that for a given anatomical severity of head injury, older patients present with a higher GCS. This has implications for the initial assessment and identification of elderly patients with potentially significant intracranial injury, but with little evidence of this on initial clinical examination.

In our study, anatomical injury severity between the two age groups was broadly similar. Injury severity score (ISS) was identical, but more elderly patients had sustained the highest AIS category of head injury. Despite this, the distribution of presenting GCS in elderly patients was higher than that in the younger group. This difference remained when subgroups at each AIS level were compared and appeared more marked at AIS 5. This supports the findings of previous studies that have found a higher presenting GCS in elderly patients with TBI, although in these studies AIS differences between young and elderly groups were not described.5 ,6

While our study cannot explain this phenomenon, demographic differences between the two groups may signpost productive avenues for further research. In the overwhelming majority of cases (86.4%), head injury in our elderly group was caused by a low-energy fall, whereas in the younger group this mechanism was responsible for only 37.7% of injuries. The AIS is a relatively crude quantification of the severity of anatomical injury, frequently reflecting the size of haematoma detected on CT. It may well be the case that in higher-energy transfer mechanisms, as well as visible haematoma, a greater degree of parenchymal injury and subsequent inflammation occurs that results in increased brain dysfunction. Alternatively, it may be the case that cerebral atrophy in the elderly allows a greater amount of intracranial haematoma and oedema to be accommodated before intracranial pressure rises and GCS falls. Given the difference in the proportion of women in each group, it is also possible that female gender somehow confers a degree of neuroprotection in isolated TBI, although the mechanism by which this might occur in elderly patients is difficult to imagine and a previous meta-analysis suggests the opposite.13

Several studies have previously described increased mortality in elderly patients with TBI despite apparently lower injury burden, and the difference in mortality between young and older patients in our study (6.4% vs 22.4%, respectively) is therefore unsurprising.12 ,14 Median GCS of elderly patients who died was significantly higher than younger patients, confirming that presenting GCS is a poor predictor of outcome in this group. Perhaps, it is more surprising that the rate of neurosurgical intervention in the elderly group in our study was as high as 28.4%. This may reflect local practice in an institution serving a population with a high proportion of elderly patients and is consistent with the intervention rate reported in the recent RAIN study.15 Neurosurgical intervention in our study seemed reserved for those with a higher presenting GCS, suggesting that other factors such as comorbidity and premorbid function are equally influential in neurosurgical decision making.

The overall length of hospital stay was increased for elderly patients. Although we were unable to determine comorbidities from the TARN database, it is reasonable to assume that the prevalence of complicating comorbidity would be higher in the elderly group, as previously described.16 Previous studies have also reported worse functional outcomes for elderly patients following even mild TBI, which also might account for the increased length of stay.11

Our findings challenge the notion that GCS can be used to define the severity of TBI across all age groups. Elderly patients may have severe anatomical TBI with high ensuing mortality despite presenting with a near-normal GCS. There are also implications for prehospital bypass tools that use GCS to determine which patients should be transported directly to an MTC.4 This is particularly relevant given the high rate of neurosurgical intervention in our elderly group despite a median initial GCS of 14. Reworking of emergency department triage systems and national TBI guidelines may also be required to improve identification of elderly patients with TBI who need urgent CT imaging.17 ,18

Conclusion

Elderly patients who have sustained isolated severe TBI may present with a higher GCS than younger patients. Further work is required to confirm this observation in the wider UK population and to explore the underlying pathophysiology. Triage tools using GCS may need to be modified and validated for use in elderly patients with TBI.

Acknowledgments

The authors are grateful to the Trauma Nurse Coordinators at Derriford Hospital who have coordinated the data collection for submission to TARN, and are also grateful to TARN for their assistance with this study.

References

View Abstract

Footnotes

  • Collaborators Trauma Audit Research Network.

  • Contributors AK and JES initiated the study. SR collected the initial data and drafted the initial manuscript. AK analysed the data and revised the manuscript. JES drafted subsequent versions of the manuscript. All authors contributed to and have approved the final manuscript.

  • Competing interests None.

  • Ethics approval The Health Research Authority regulates the use of patient information in the UK and has given ethical approval for research using anonymised TARN data (approval number: ECC 7-05(g)/2011).

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

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.