Happy Christmas to all NASGBI members and in an attempt to maintain your CME over the festive period Elfyn Thomas, consultant intensivist and neuroanaesthetist from Plymouth, has selected the following articles on traumatic brain injury (TBI) for you. Some are in journals you would not necessarily screen as neuroanaesthetists and intensivists eg JAMA, JAMA Surgery and Emergency Medicine Journal. Happy reading and the opinions expressed are those of the author – not necessarily NASGBI’s GCS in the Elderly On the 40th anniversary of the Glasgow Coma Score (GCS) , the re-launch of which and be viewed from the news section of the homepage, it is interesting and probably important to remember that the original paper by Graham Teasdale and Bryan Jennet in 1974 was first developed and validated in a cohort of adults with a mean age of 33 years. In the RAIN study the mean age of the population admitted to British Neuro ICUs was 45 and a significant proportion were over the age of 65 as shown in this graph from the publication. 
In an ever increasingly aging population – is the GCS as accurate and predictive in the older patient as in the population it was validated? The effect of age on Glasgow Coma Scale score in patients with traumatic brain injury. Salottolo K, Levy S, Slone DS, Mains CW, Bar-Or D. JAMA Surg. 2014;149(7):727-734. In the July issue of JAMA Surgery, Kristin Salottolo and colleagues from Colorado looked at whether age affects the predictive ability of the GCS for severity of TBI, comparing two cohorts over and under the age of 65 over a 5 year period. When comparing physiological effects using the GCS and severity of anatomical injury using the Abbreviated Injury Score (AIS) their hypothesis was that the GCS in elderly would be different compared to younger patients with the same level of anatomic severity of TBI. This was a retrospective cohort study of prospectively collected data entered into the trauma registers (similar to TARN) of two level 1 trauma centres in Colorado, USA, using standard definitions of GCS and AIS. They identified 6710 patients with a mean age of 48 (similar to RAIN), 63.7% were male and falls were the commonest injury modality (38.76%) closely followed by RTC (30.52%); 1800 patients (26.83%) were over the age of 65. Not surprisingly there were significant demographic differences between the two groups ( see table below) with older patients having less extra-cranial injuries, a higher incidence of subdural haematomas, falls as a mechanism, greater female:male ratio but nearly half of the older cohort (48.89%) had an injury severity score (ISS) of greater than 16 – the official definition of ‘major trauma’. They also found that for every AIS category there were significant differences in the GCS scores with older patients having significantly higher GCS scores – at AIS 4 (severe) 83% of over 65s had a GCS of 13-15 compared to 55% of those <65 and at the AIS of 5 (critical) 56% of over 65s still had a GCs of 13-15 whilst 63% of younger patients has a GCS of 3-8. Even after correcting for extra-cranial injuries in the 5261 patients with isolated TBI the difference remained. 
At any level of anatomical TBI injury severity, as described by the AIS, patients under 65 were 2-6 times more likely to present with a severe deficit as described by the GCS (3-8) than those over 65 years – and this remained even after correcting for confounders. At any level of AIS ( i.e anatomical injury severity) there was no difference in mortality between the two cohorts however when mortality was stratified by GCS the mortality was significantly higher in those > 65 in all GCS categories. Whilst this study reinforces previous work that older people have a worse outcome after TBI – it hints that this is no different if you classify TBI based on anatomical injury (CT scan/AIS) rather than physiological derangement (GCS) and suggests that older patients have a better GCS than younger ones with similar TBI severity. As our population ages this has significant implications for head injury research in those over 65 and, as trauma networks and bypass protocols become established in the UK, neuroscience centres may need to reconsider historical approaches to transfer and intervention based on age and the GCS and take more account of the CT scan and the extent of the anatomical injury itself. Bilateral fixed dilated pupils Prognosis of patients with bilateral fixed dilated pupils secondary to traumatic extradural or subdural haematoma who undergo surgery : a systematic review and meta-analysis. Scotter J, Hendrickson S, Marcus HJ, Wilson MH. Emerg Med J 2014;0:1–6. doi:10.1136/emermed-2014-204260 Maintaining a TBI theme and challenging some surgical (and anaesthetic) dogma I came across a thought provoking systematic review and meta-analysis from Mark Wilson and his colleagues at St Marys. Studies reporting the outcomes of patients with bilateral fixed and dilated pupils (BFDPs) have generally concluded that there is no appreciable chance of meaningful survival, and that further interventions such as surgery and admission to an intensive care unit may not be warranted. The received surgical ‘wisdom’ is that patients with closed head injury and BFDPs should be managed conservatively with resuscitative efforts concentrating on salvaging the patient for organ donation. Some groups have argued, however, that selected TBI patients with extradural and subdural haematomas and BFDPs should have surgical evacuation and can make a good recovery; this was the question that Mark and his colleagues aimed to study. Their review methodology was robust and they identified a total of 82 patients with BFDPs who had surgical evacuation and follow-up data which were included in the meta-analysis, 25 had extradural haematomas (EDH) and 57 had acute subdural haematomas evacuated (ASDH). In EDH the overall mortality rate was 29.7% and the proportion of patients who had a favourable outcome (GOS 4 or 5) was 54.3% whilst in ASDH the overall mortality rate was 66.4% and the proportion of patients who had a good functional outcome was 6.6%. Despite the small numbers, heterogeneity of the groups and retrospective nature of the study the authors conclude that 54% of patients with EDH and BFDPs survive with a good outcome and that this is an underappreciated prognosis by many in our collective disciplines, the same cannot be said of ASDH. This study reminds us that EDH, unlike ASDH, is not primarily a brain/ parenchymal injury and that fixed dilated pupils should not necessarily stop us from aggressively decompressing and managing these patients. Too much of a good thing? Significance of arterial hyperoxia and relationship with case fatality in traumatic brain injury: a multicentre cohort study. Rincon F, Kang J, Vibbert M, Urtecho J, Athar MK, Jallo J. J Neurol Neurosurg Psychiatry 2014;85:799-805 Whilst you were probably thinking about your summer holidays and wondering if too much sun would be bad for you this paper was published in July’s JNNP with an accompanying editorial by Phil Hopkins and Peter Andrews which suggested that too much oxygen may not be beneficial in TBI. There are enthusiasts and early adopters who believe that normobaric hyperoxia is beneficial in TBI management, even being recommended by some societies – however there are no randomised controlled trials demonstrating benefit, only historic control data, and systematic reviews demonstrate the lack of evidence for tissue oxygen directed therapies in TBI (see editorial above). Some of you may recall the Project IMPACT based article published in JAMA in 2010 looking at patients admitted to the ICU following resuscitation from cardiac arrest who found that arterial hyperoxia was independently associated with increased in-hospital mortality compared with either hypoxia or normoxia. They have now done a similar analysis using the Project IMPACT database (a bit like ICNARC in the UK) to test the hypothesis that ‘hyperoxia was not associated with higher in-hospital case fatality in ventilated traumatic brain injury patients admitted to the ICU’. They performed a retrospective analysis of the prospectively collected data on the Project IMPACT database from 131 adult ICUs from across the US (DGC and teaching) and their cohort were ventilated patients with TBI consecutively admitted to ICU between 2003 and 2008, older than 17 with arterial blood gases available for the first 24 hours. 405,668 admissions became 4,176 TBI patients of which 1752 were ventilated but 540 omitted due to incomplete ABG data leaving a study cohort of 1212. This cohort came from 61 different hospital ICUs, mainly from academic, urban and extra-large community hospitals – none were from a neurological or neurosurgical ICU – possibly comparable therefore to UK DGHs. Hypoxaemia was defined as a PaO2 less than 60mmHg (7.99 kPa) or a PaO2/ FiO2 ratio <300, hyperoxia was defined as PaO2 > 300 mmHg (PaO2 39.99 kPa) – these are the same definitions used in the 2010 JAMA paper mentioned earlier. Of the 1212 patients 33% were ‘normoxic’ (mean PaO2: 28kPa ±6.5), 46% were ‘hypoxaemic’ (mean PaO2 16.8 ±11.46, mean PaO2/FiO2 182±74) and 21% were ‘hyperoxic’ (mean PaO2 56.39 ±11.06). There were no significant differences or associations at ICU admission between the three groups in baseline variables, hospital type or other initial physiological or biochemical values, GCS scores were well matched. 400 died in hospital (33%) with case fatality rate being 41% in the hypoxic group, 32% in the hyperoxic group and 23% in the normoxic group. The Kaplan-Meier curve below demonstrates the difference between hyperoxia and normoxia. As you can see the curves separate early and beyond day 4-5 appear parallel. 
After controlling for confounding variables in a robust multivariable analysis exposure to hyperoxia was an independent predictor on in hospital mortality. They also demonstrated that hypoxaemia remains bad for injured brains and would support that oxygenation should be kept within the ‘normal’ range – at least in the first 24 – 48 hours and that supra-physiological values may be harmful in an unselected brain injured population (in a non-neuroscience ICU in the USA). There are a number of potential confounders and weaknesses with this study which are explored in both the article and the editorial and there is no attempt to link cause and effect etc – however it should make us stop and think about what we do to our patients, what advice we give our colleagues in DGHs and that indiscriminate use of supra-normal levels of oxygen may be harming brain injured patients and that less may be more and more may be bad. Effect of erythropoietin and transfusion threshold on neurological recover after traumatic brain injury. A randomized clinical trial. Robertson CS, Hannay HJ, Yamal J-M, Gopinath S, Goodman JC, Tilley BC. JAMA 2014;312(1)36-47. There has been much talk recently of haemoglobin levels in general critical care with the publication of the TRISS trail in October which reaffirms the conclusions of the TRICC trial nearly 15 years ago. Despite the lack of evidence of benefit of higher haemoglobin levels in neurocritical care and recent evidence suggesting a higher VTE rate in transfused SAH patients, many clinicians aim for haemoglobin levels in excess of 10g/dL when managing brain injured patients. This important prospective randomised bi-factorial design trial of EPO and haemoglobin transfusion thresholds in TBI conducted by Claudia Robertson’s group in Texas should result in us questioning the wisdom of this liberal practice. The trial aimed to address two questions: does the early administration of erythropoietin (EPO) improve outcome in TBI and also does a transfusion threshold of 10 g/dL vs 7 g/dL improve outcome? They aimed to recruit adult intubated TBI patients with a motor score less than 5, who did not have bilateral fixed pupils and could be randomised within 6 hours of injury. The investigators were blinded to EPO or placebo but could obviously not be blinded to the transfusion threshold – which was allocated randomly, and outcome assessors were blinded to both. The primary outcome measure was the GOS dichotomised to favourable and unfavourable outcomes. 895 patients were assessed, 598 screened to produce a 200 randomised cohort who were allocated to the 4 treatment arms – so not a huge study. They found no interaction between the EPO and haemoglobin transfusion threshold groups and therefore groups were combined for their respective analyses. Erythopoietin Regimens – Results No difference in neurological outcome – this study would not support the use of early EPO in TBI – however it is a much smaller study than the EPO-TBI trial being run by the ANZICS clinical trials group which has just finished recruiting600 patinets and should report next year. The absence of even a signal in this trial though is discouraging. Haemoglobin transfusion thresholds – Results No difference in neurological outcome. The 6-month GOS was available for 87 in the 7g/dL group and 94 patients in the 10g/dL group. 43.5% in the lower threshold group had a favourable GOS compared to 33% in the 10g/dL group – no statistically significant difference. 
There was no statistical difference in the incidence of ARDS (7g/dL 16/99 vs 10g/dL 25/101) nor infections (7g/dL 27/99 vs 10g/dL 36/101) however there was a significant difference in incidence of at least one thromboembolic event (7g/dL 8/99 vs 10g/dL 22/101 p=0.009) favouring the lower transfusion threshold trigger (OR 0.32). Given the lack of clinical neurological benefit and apparent harm due to increased risk of VTE (as also demonstrated in SAH) this study does not provide evidence to support a haemoglobin transfusion threshold of 10g/dL in preference to 7 g/dL in closed TBI. In the absence of further evidence to the contrary we should all be reviewing and re-defining our transfusion thresholds in NeuroICUs towards 7 g/dL as our colleagues in general ICUs have done for a number of years. And on that bombshell….…….as all those sacred neurosurgical cows are led to slaughter……… Happy Christmas !!!
