Proceedings of the British Neurosurgical Research Group, hosted by Newcastle Spring neurosurgical unit, 2009

Click to increase image sizeClick to decrease image size

Association between cerebral chemistry and outcome after traumatic brain injury: microdialysis study in 223 patients

I. Timofeev¹, P. Al-Rawi¹, J. Nortje², K. Carpenter¹, M. O'Connell¹, M. Czosnyka¹, P. Smielewski¹, D. K. Menon², A. K. Gupta², J. D. Pickard¹, P. J. Kirkpatrick¹ & P. Hutchinson¹ (¹Academic Neurosurgery Unit and ²Division of Anaesthesia, University of Cambridge/Addenbrooke's Hospital, Cambridge, UK)

Introduction: Secondary insults adversely influence outcome following severe traumatic brain injury (TBI). Monitoring cerebral extracellular chemistry with microdialysis has the potential for early detection of metabolic derangements. An association between common biochemical markers and neurological outcome following TBI in a large cohort of patients has not been conclusively established, and evaluating such a relationship was the aim of this study.

Methods: Prospectively collected observational neuromonitoring data from 223 patients with TBI were analysed. All patients were managed in a single tertiary neurosurgical centre and required intensive care management and neuromonitoring. Monitoring modalities included intracranial pressure (ICP), cerebral perfusion pressure (CPP), cerebrovascular pressure reactivity (PRx) and microdialysis markers (glucose, lactate, pyruvate, glutamate, glycerol, lactate/pyruvate ratio (L/P ratio). Outcome was assessed using the Glasgow Outcome Scale (GOS). Patient averaged values of parameters were analysed by univariate non-parametric methods and multivariate logistic regression.

Results: Microdialysis monitoring commenced on median [IQR] day 1[1;2] from injury and median [IQR] duration of monitoring was 4[2;7] days. During the initial 72 hours of monitoring median L/P ratio (p = 0.026) and lactate (p = 0.033) levels were significantly lower in patients with favourable outcome (GOS good recovery, moderate disability). During the total monitoring period levels of glutamate (p = 0.048), L/P ratio (p = 0.044), ICP (p = 0.006) and PRx (p = 0.01) were significantly higher in patients who died. In a multivariate logistic regression model, glucose, pyruvate, L/P ratio and PRx were significant independent predictors of mortality, together with age and GCS (p < 0.0001).

Conclusion: The results of this study suggest that extracellular metabolic markers are independently associated with outcome following TBI, with L/P ratio being the most consistent predictor. Whether treatment related improvement in biochemistry translates into better clinical outcome remains to be established.

Normobaric hyperoxia alters the pattern of cell death on an in vitro model of traumatic brain injury

A. Kumaria^1,2, J. Feng³ & C. M. Tolias¹ (¹Department of Neurosurgery, King's College Hospital, ²King's College London School of Medicine and ³Clinical Neuroscience Department, Institute of Psychiatry, Kings College London, London, UK)

Introduction: Several clinical studies have shown that administration of normobaric hyperoxia improves cerebral metabolism following traumatic brain injury (TBI) [1]. In this study we explore its effects on an in vitro model of neurotrauma.

Methods: To model glutamate neurotoxicity in TBI, primary rat cortical neurons (DIV 7) were exposed to glutamate followed by 60% oxygen in a hyperoxia chamber or room air in an incubator for 24 hours. Next, to model mechanical injury in TBI, primary neurons were exposed to a sublethal mechanical insult using a Flexcell FX-4000 cell stretch apparatus [2] alongside varying concentrations of glutamate, followed by 60% oxygen or room air for 24 hours. Cell viability was assessed using the lactate dehydrogenase (LDH) release assay of cell membrane lysis: a marker of cell death.

Summary of results: Glutamate caused neuronal cell death in a dose dependent manner and the mechanical injury model synergised glutamate neurotoxicity. Hyperoxia significantly decreased neuronal necrosis in the glutamate injury model as demonstrated by decreased LDH release. Preliminary data suggest similar hyperoxia-mediated effect in the mechanical insult model as well, as suggested by LDH assays.

Conclusion: This study suggests a neuroprotective effect of hyperoxia on an in vitro model of neurotrauma by attenuating cell necrosis. We discuss possible mechanisms underlying these observed therapeutic effects.

The influence of peripheral inflammation on brain temperature and outcome after traumatic brain injury (TBI)

R. H. Sacho^1,3, S. J. Hopkins², T. Rainey³, M. Hoadley², A. Vail⁴, A. T. King^1,3 & C. Childs³ (¹Department of Neurosurgery and ²Clinical Neurosciences, Salford Royal NHS Foundation Trust ³Brain Injury Research Group and ⁴Health Methodology Research Group, School of Medicine, University of Manchester, UK)

Background: Raised temperature after TBI is thought to be associated with worsened outcome in patients. The mechanisms linking temperature to brain damage are unclear but may be via inflammatory pathways, as has been suggested in clinical studies of stroke [1].

Aim: To determine the association of IL-6 concentration with brain temperature and outcome.

Methods: Recruitment: prospective, consecutive adults with severe TBI admitted to ICU.

Monitoring: intracranial temperature monitoring and arterial blood sampling at 10 am during days 1-5 after admission, for determination of plasma IL-6 concentration.

Statistics: logistic regression with both unifactorial and multifactorial analysis to account for known predictors of outcome after TBI.

Primary outcome: 30 day mortality.

Results: 52 patients aged 16 to 75 (median 34) years were recruited. The range of peak plasma IL-6 was 25.0 to 26635.5 (median 335.69) pg · ml⁻¹. There was no significant association between IL-6 concentration and brain temperature. Peak IL-6 was significantly related to 30 day mortality (OR = 1.79; p = 0.03). This relationship was independent of brain temperature and other known predictors of outcome after TBI.

Conclusion: Raised brain temperature is not driven by plasma IL-6 concentration. Peripheral inflammation is significantly associated with increased mortality. This effect is independent of elevations in brain temperature.

Non-invasive assessment of brain compliance

M. Drinnan¹, C. Griffiths¹, D. Martin², A. D. Mendelow², R. Bergman¹, I. Chambers¹ & P. Mitchell² (Departments of ¹Medical Physics and ²Neurosurgery, Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle, UK)

Introduction: Brain compliance, the effect of volume change on intra-cranial pressure, probably reflects the function of ICP regulatory mechanisms; a compliant brain would accommodate significant changes in volume with relatively small changes in pressure. Small-volume compliance can be measured using an air-filled balloon introduced via a cranial bolt, but craniectomy offers the opportunity for measurement over much larger volumes. We have developed an instrument for non-invasive assessment of large-volume brain compliance.

Methods: The probe comprises a flat surface to palpate the craniectomy site:

Force: A strain gauge measures the force with which the probe is applied.

Contact area: The contact surface forms one plate and the dielectric of a capacitor. By measuring the capacitance between probe and skin, one can estimate the contact area.

Displacement: The magnetic coupling between a pair of coils follows a 1/r³ relationship with separation. We thereby measure probe movement relative to a stationary coil on the opposite side of the head.

Analysis of data: Pressure is force per unit contact area; volume is the integral of contact area, with respect to displacement. Compliance is estimated from the gradient of the volume-pressure graph.

Results: To date we have studied four subjects. In one, we had simultaneous invasive ICP measurements. The peak pressure estimated by the probe (approximately 20 mm Hg) corresponds well to that recorded by ICP (18 mm Hg). Brain compliance was estimated in all, and ranged from 1 to 10 cm³ per mm Hg. These values are higher, though comparable in magnitude, with invasive data from the literature.

Discussion: We have presented the first results from a new instrument for objective assessment of ICP and brain compliance in craniectomy. Estimated pressures were very similar to those obtained from direct measurement of ICP. Since the contents of the skull are essentially incompressible, volume changes reflect the shunting of CSF and blood between the intracranial and extracranial compartments. This new measurement might have value in the non-invasive assessment of progress following a craniectomy, but might also give us new insights into the regulatory mechanisms for ICP.

Glial Astrocytic Fibrillary Protein and S100B as prognostic and neuromonitoring markers of acute brain injury: a translational study

H. Ragab¹, V. Di Pietro¹, J. Hillman², L. Khang Yean¹, A. Pringle¹, B. Tavazzi³, G. Lazzarino⁴, A. Kolias⁵, A. Petzold⁵ & A. Belli¹ (¹Division of Clinical Neurosciences, University of Southampton, UK, ²Neurosurgical Department, University Hospital Linköping, Sweden, ³Institute of Biochemistry and Clinical Biochemistry, Catholic University of Rome, Italy, ⁴Department of Chemical Sciences, University of Catania, Italy, ⁵Department of Neuroimmunology, Institute of Neurology, London, UK)

Introduction: Markers of astroglial damage such as S100B and GFAP are released into the brain extracellular fluid (ECF) following acute brain injury (ABI) and can be sampled with cerebral microdialysis (MD). In order to investigate the role of these markers in neuromonitoring and prognostication, we analysed S100B and GFAP in MD samples of ABI patients and in an in vitro model of traumatic brain injury (TBI).

Methods: 40 patients with traumatic and vascular brain injuries were studied. MD samples were analysed by ELISA. Using a Linear Mixed Model analysis, S100B and GFAP concentrations were correlated with:

1. cerebral blood flow velocities

2. Xenon-CT values

3. clinical deterioration by GCS

4. outcome on discharge and at 6 months

5. MD cytokines, lactate/pyruvate ratio, glutamate and glycerol

In the in vitro model, different grades of stretch injury were delivered to rat organotypic hippocampal cultures. This produces damage equivalent to diffuse axonal injury in human TBI. S100B and GFAP ELISA concentrations and mRNA expression were correlated with rate of cell death at different time intervals over 5 consecutive days as measured by propidium iodide fluorescence. ANOVA with Tukey's test and correction for multiple comparisons was used for statistical analysis.

Results: In MD samples, median S100B was 2.79 ng/ml, while median GFAP was 26.26 ng/ml. Cconcentrations of both decreased with time from injury; there were several secondary peaks for S100B. GFAP was shown to be a strong predictor of 6-month outcome (p < 0.001). S100B showed a significantly different day-to-day progression (p < 0.001) between outcome groups but it was a poorer predictor of outcome. There was significant negative correlation between mean Xenon-CT hemispheric flow and initial GFAP and S100B levels (r = −0.52 and r = −0.54 respectively; p < 0.001). GFAP and S100b had significant correlations with several other MD biomarkers and clinical parameters.

In vitro, GFAP and S100B concentrations were significantly higher in severe injuries compared to both mild injuries and controls (p < 0.0001); levels correlated with cell death (p < 0.0001). GFAP mRNA expression peaked in the first few hours, reflecting the findings of the clinical samples.

Conclusions: GFAP and S100B correlate with severity of injury and appear to have prognostic and clinical value.

The burden of motorcycle related neuro-trauma in Ireland and associated helmet usage

R. Murphy, L. McEvoy, M. O. Kelleher, C. Bolger & J. Phillips (Neurosurgical Research and Development Unit, Beaumont Hospital, Dublin 9, Ireland)

Introduction: Motorcycles represent less than 2% of the licensed vehicles but motorcyclists account for 12% of road deaths in Ireland (RSA 2006). According to OECD figures a motorcyclist is 2–3 times more likely to be killed in Ireland than on the continent (RSA 2007).The most successful proven injury prevention method to reduce the severity of head injury and fatalities after a crash are motorcycle helmets (Cochrane 2008).

The COST 327 European research brought together experts from across Europe to investigate in detail, motorcyclists' head and neck injuries and recommend preventative strategies (Cost 327 1999). In light these recommendations the British Road Safety Authority has introduced the Sharp programme which hopes to save 50 lives in the U.K. each year alone by helping riders to choose the best fitting and safest helmets.

Methods: We evaluated the pattern of head injuries sustained by motorcyclists referred to the two neurosurgical centres Beaumont Hospital and Cork University Hospital in Ireland and ascertained if the new SHARP guidelines could be of benefit in reducing the burden of motorcycle related neurotrauma and disability in Ireland.

Results: Of the 62 patients registered 24% were not wearing helmets. Mild TBI (GCS 15–13) accounted for 40% of all new referrals, Moderate TBI (12–9) 6% and Severe TBI (8–3) 51%. CT Brain imaging of patients : Contusions 63%, Subdural haemorrhages 15% traumatic subarachnoid haemorrhage 15%, extra-dural haemorrhages 4%, and skull fractures 27%. The mortality for the patients registered after a motor cycle accident was 18% (11 patients).

Conclusions: Despite Ireland having mandatory helmet laws almost a quarter of our motorcyclists with traumatic brain injury were unhelmeted. Up to 20% reduction in mortality is predicted if all motorcyclists in Ireland were to wear helmets that satisfied the SHARP criteria.

Randomised controlled trial of the use of drains versus no drains after burr hole evacuation of chronic subdural haematoma

T. Santarius, P. J. Kirkpatrick, D. Ganesan, H. L. Chia, I. Jalloh, P. Smielewski, H. K. Richards, H. Marcus, S. J. Price, R. W. Kirollos & P. J. Hutchinson (Addenbrooke's Hospital and University of Cambridge, Hills Road, Cambridge, UK)

Introduction: Chronic subdural haematoma (CSDH) is a cause of significant morbidity and mortality which is often amenable to surgical drainage. The majority of patients recover rapidly following surgical intervention, but 5 to 30% suffer recurrence. Although use of drains after burr hole evacuation may lower recurrence, they are not used routinely for lack of evidence showing efficacy and a concern for increasing surgical morbidity. The aim of this study was to investigate the use of drains in reducing recurrence rates, and their influence on clinical outcome for those presenting with a CSDH.

Methods: Between November 2004 and 2007, 215 patients were randomised; 108 were assigned to drain, and 107 to no drain. The primary outcome was recurrence requiring re-drainage. The secondary outcomes included mortality at 30 days and 6 months, modified Rankin scale (MRS) score at discharge and at six months, and the duration of hospital stay.

Results: The CSDH recurrence rate was 10/108 (9.3%) in the drain group, and 26/107 (24%) in the non-drain group (χ², p = 0.003). At six months the mortality was 10/106 (9.4%) and 19/105 (18.1%) respectively (χ², p = 0.042). A greater proportion had favourable MRS (0–3) in the drain group at discharge and at six months [81/97 (84%) vs 64/95 (67%), χ², p = 0.009; and 64/76 (84%) vs 60/85 (71%), χ², p = 0.040]. In addition, a significantly greater proportion of patients with a drain were discharged with a GCS of 15 [76/94 (81%) vs 62/97 (63%), χ², p = 0.007] and without a neurological deficit [47/93 (51%) vs 63/96 (66%); χ², p = 0.036]. Although a trend advantage to the drain group was apparent, the overall hospital stay did not reach statistical significance. There was no difference in the rate of inpatient medical or surgical complications between the study groups.

Conclusion: The use of drain with burr hole drainage reduces the recurrence rate for CSDH, and is associated with a lower mortality at six months.

Safety of spinal cord repair utilizing autologous olfactory mucosal grafts and bone marrow stromal stem cells in patients with clinically complete chronic spinal cord injuries

A. N. Halaka¹, O. A. El-Banhawy², H. Ayad¹, F. El-chennawy³, M. El-Sherbiny⁴ & S. Khamis¹ (Department of Neurosurgery, EI-Hikmah Hospital¹, Mansura, Department of ENT surgery, El Menoufyia University², Department of Clinical Pathology and Immunology, Mansura University³, Department of Anesthesiology, Banha University⁴ Egypt)

Objectives: Currently we are running a prospective controlled study at EI-Hikma Hospital, Mansura, Egypt, for spinal cord repair utilizing autologous full-thickness nasal septal olfactory mucosa (NSOM) grafts and bone marrow stromal (BMS) stem cells in patients with clinically complete chronic spinal cord injuries (SCI). The objective of this work is to determine the safety of this procedure.

Methods: The study group included 40 patients who had suffered a SCI 6 months or longer before surgery, with complete motor, sensory and sphincteric dysfunction below the cord injury level. There were 6 patients with cervical cord lesions, and 34 with dorsal or dorsal-lumbar cord lesions. BMS stem cells were obtained overnight from the posterior iliac crest marrow aspirate, taken under local anaesthesia the day before surgery. Full-thickness NSOM was endoscopically identified and harvested via a nostril under hypotensive general anesthesia. The harvested NSOM was sharply cut into multiple cylindrical cables and cubes. They were immersed into cerebrospinal fluid (CSF) obtained intra-operatively via lumbar puncture until the time of implantation. The damaged spinal cord was approached via laminectomy. After dural opening, radical circumferential cord untethering was performed and central myelotomy was carried out. Excision or lysis of glial adhesions with removal of any bone fragments or calcification was carried out. Cables and cubes of the NSOM were used to bridge the gap between the proximal and distal normal looking cord adjacent to the damaged cord segment. Micro-sutures were used to close the pia mater of the myelotomy. Half the volume of BMS stem cells was injected directly in the repaired area, and the other half locally into CSF. All participants of this study will be assessed regularly for 3 years after enrolling in the trial.

Results: Patients of this study were 33 males and 7 females. Their ages ranged from 4 to 40 years (median 21, mean 25). The period since their SCI ranged from 6 to 110 months (median 10, mean 12 months). The length of injured spinal cord segment varied from 3 cm to 6 cm (median 4, mean 5 cm). There were no intra-operative complications. Two patients developed staphylococcal wound infection that responded to appropriate antibiotic therapy. Two patients developed CSF leakage and collection under the skin sutures, one resolved with aspiration/dressing and the other needed the insertion of cysto-subarachnoid tube placement. There was no neurological deterioration as a result of the procedures and most cases showed a variable degree of neurological recovery. There were no iatrogenic rhinological complications and re-epithelialization of the olfactory cleft mucosa was visualized at delayed postoperative nasal endoscopy carried out in some cases.

Conclusions: This study suggests that spinal cord repair after clinically complete chronic spinal cord injuries utilizing autologous full-thickness NSOM and BMS stem cells is both feasible and safe and is showing promising early results.

The AVERT-IT Project: Purpose and Progress

I. Piper¹, R. Donald², R. Sinnott³, A. Stell³ & J. Jipu³ on behalf of the AVERT-IT consortium (Southern General Hospital, Glasgow, UK¹ on behalf of the BrainIT Group, C3-Amulet Ltd, Dingwall, UK², National eScience Centre, Glasgow University, Glasgow, UK³)

Abstract: The ability to predict arterial pressure hypotensive adverse events would be of major benefit to the fields of secondary health-care especially to the traumatic brain injury population. One of the aims of the BrainIT group (www.brainit.org) is to promote new analysis methods from studies using the high resolution BrainIT database. Towards that end the “AVERT-IT” consortium was formed to assess the feasibility of using physiological time-series data from the database to train a Bayesian neural network to predict arterial hypotension. This is the basis for the Avert-IT project (http://www.avert-it.org), a three year collaborative EU-funded project involving six BrainIT clinical centres, a software development company (http://c3amulet.com) and eScience Grid technology experts in secure access to distributed medical datasets (http://www.nesc.ac.uk). We report on the project progress after the first 12 months which includes creating a standardised definition for hypotension events based upon both systolic and mean arterial pressure. Using this definition, approximately 2000 arterial hypotensive events were identified forming the basis of the training set. We report on early training results (sensitivity & specificity) from using a selection of time-series signals and their derived functions over a series of time-windows obtained before the hypotension events occurred. Multi-centre research ethics committee approval has now been obtained to conduct a prospective observational study across 6 centres recruiting 160 patients with traumatic brain injury to assess the prediction engines performance with actual live intensive care data. To acquire this data with a minimum of data double entry, middleware software systems are being deployed interfacing with local hospital information systems to automatically extract hospital based data elements required by the project schema. Data is automatically anonymised and pushed every 6 hours to the multi-centre database designed to support remote monitoring of the trial progress. Implications for future clinical trials of this approach of trial data capture will be discussed.

Study of severe head injuries admitted to the northern regional neurosurgery department in 2007 and changes over time

L. Chilton, H. Hastie, T. J. Jones & A. D. Mendelow (Department of Neurosurgery, Regional Neurosciences Centre, Newcastle General Hospital, Newcastle upon Tyne, UK)

Objective: In the Northern Region there are 13 Primary Care Trusts which fund care at over 32 hospitals. This covers the area from the Scottish boarders into North Yorkshire and the breadth of the UK at this point with a mix of urban and rural inhabitations, providing healthcare to just over 3 million residents. In this area Newcastle General Hospital is the only tertiary centre with specialist staff and equipment to treat neurological and neurosurgical cases. The objective was to audit ongoing admissions with severe head injury.

Method: Data on cases before 1950 are provided in Acute Injuries of the Head (1949). Data from 1980 to 2000 were collected as part of a formal audit of Head Injuries. Data from 2007 were collected as a specific project using coding logs, patient notes and admission records and the Traumatic Brain Injury Dataset. Data was analysed to assess the prevalence of severe head trauma and intracranial haemorrhages for the full year and compared with the historical data to assess the effects of the guidelines issued by the National Institute of Clinical Excellence (NICE) and the Brain Trauma Foundation. 26 patients who had an admission GCS of between 3 and 8 were analysed in greater detail.

Results: The average number of head injuries between 1980 and 2000 was 627.3 p.a., ranging from 486 to 714 cases. The case mix routinely included 20% (11% to 29%) intracranial haemorrhages. In 2007 there were 539 patients admitted with 27% suffering from intracranial haemorrhages. These were <38% of the cases that were coded as head injuries according to ICD-10 and were predominantly adult males with a mean age of 44 years however the average age of severe head injury patients was lower at 33 years old. Within the subset of patients analysed 92% of the patients were still alive ∼6 months after their initial admission.

Conclusion: The Newcastle Neurosurgery Unit has consistently admitted about 500 patients per year for more than 50 years. The introduction of the new guidelines does not appear to have changed the number of cases presenting locally but may have resulted in a shift to a greater number of intracranial haemorrhages being diagnosed.

STITCH(TRAUMA) – A new randomised controlled trial in traumatic intracerebral haemorrhage

B. A. Gregson¹, P. Mitchell², I. R. Chambers³ & A. D. Mendelow¹ on behalf of the STITCH(TRAUMA) principal investigators (¹Newcastle University, Newcastle upon Tyne, ²Newcastle General Hospital, Newcastle upon Tyne, ³James Cook University Hospital, Middlesbrough, UK)

Introduction: This new randomised controlled trial has been established to investigate the use of early surgery in the treatment of traumatic intracerebral haemorrhage and contusion. It will include a health economics component and will carry out a subgroup analysis of patients undergoing invasive monitoring. Funding has been sought from the NIHR HTA programme and the study will be undertaken alongside the STICH II.

Methods: Adults with CT scan evidence of traumatic supratentorial ICH of 25 ml or contusion with an area of raised attenuation above that of the background white and grey matter greater than 10 ml and who are within 12 hours of head injury will be eligible. Patients will be excluded if 1) there is evidence of a surface haematoma (EDH or SDH), 2) surgery can not be performed within 24 hours of injury or 12 hours of randomisation (whichever is the shorter), 3) there is severe pre-existing physical or mental disability/severe co-morbidity which might lead to a poor outcome even if the patient made a full recovery from the head injury. Patients will be randomised via a 24 telephone randomisation service. Patients randomised to early surgery should be taken to theatre to undergo a craniotomy, decompressive contusionectomy or minimal surgery as soon as possible. The appropriate surgical technique will be decided by the treating surgeon. The study will include both patients who are receiving invasive monitoring of ICP and those who are not. Outcome will be measured at 6 and 12 months by postal questionnaire dispatched by the STICH Office.

Results: The study protocol has been drafted and the UK ethics committee application has been prepared. Once this approval has been obtained, participating UK sites will need to complete the SSI and R&D approvals for their own centres and non-UK sites will need to prepare ethics committee applications. Study agreements will be prepared and circulated. The study team will be in post on 1 September 2009 and patient recruitment is due to commence on 1 October 2009. Data collection forms are in preparation.

Conclusion: Although the study specific team will not be in post until September it is vital to have undertaken the regulatory phase for as many sites as possible before this date so that they are able to commence recruitment as soon as possible. Members of the existing STICH II team will be happy to advise and assist with these processes. Centres who are interested in taking part in this study should email [email protected].

An interim analysis of a prospective, randomised and controlled trial investigating the use of lumbar cerebrospinal fluid drainage following aneurysmal subarachnoid haemorrhage

Y. Z. Al-Tamimi¹, D. Bhargava¹, G. Hall¹, R. Feltbower², A. C. Quinn³ & S. A. Ross¹ (Department of Neurosurgery, Leeds General Infirmary¹, Centre for Epidemiology and Biostatistics, University of Leeds², Department of Anaesthesia, Leeds General Infirmary³, Leeds, UK)

Objective: There is some evidence to suggest that lumbar drainage of cerebrospinal fluid (CSF) following aneurysmal subarachnoid haemorrhage (aSAH) reduces the incidence of delayed ischaemic neurological deficit (DIND) and improves patient outcome. A single-centre prospective randomised and controlled trial has been conducted in order to test this hypothesis. Results of an interim analysis are presented.

Method: Patients with World Federation of Neurosurgeons Grade (WFNS) 1–3 aSAH are randomised to either the study group of standard therapy plus insertion of a lumbar drain, or the control group of standard therapy alone. This interim analysis consists of 103 patients (82 females, 21 males) randomised to the study group (n = 51) and the control group (n = 52). There was no significant difference in age, sex, WFNS grade and Fisher grade between the study and control groups.

Results: There is a significant difference in the incidence of DIND (37% control v 18% study, p = 0.046) and the proportion of patients with a poor MRS (3, 4, 5 or 6) at day 10 (69% control v 44% study, p = 0.021). There is one case of delayed meningitis and two cases of superficial drain exit site infections in patients within the study group.

Conclusions: This interim analysis has demonstrated a promising difference in the incidence of DIND between the study and control group. Lumbar drain usage in this setting is safe. The trial will continue until 150 patients are recruited (expected completion June 2009).

A prospective study of intraocular haemorrhage in aneurysmal subarachnoid haemorrhage

V. S. Alg, S. Pushpananthan, O. Comyn, D. Oladiwura, M. C. Papadopoulos & B. A. Bell (Academic Neurosurgery Unit, St George's, University of London, London, UK)

Introduction: Intraocular haemorrhage (IOH) associated with aneurysmal subarachnoid haemorrhage (SAH) has been well described in largely retrospective studies [1], which confirm that SAH patients with IOH have a poor outcome [2]. Our study aimed to i) determine the incidence of IOH prospectively, and ii) the prognostic value of the presence of IOH.

Methods: Direct fundoscopy was performed in 197 patients with aneurysmal SAH to determine the incidence and type of IOH (subhyaloid, retinal or vitreous). The site of the aneurysm and the WFNS clinical grade was recorded and outcome at 6 months was assessed on the Glasgow Outcome Scale (GOS). Statistical analysis was carried out using the chi-squared test for categorical data.

Results: Of the 197 patients, 21 (10.7%) had IOH on fundoscopy and 9 (42.7%) died. In the 176 patients without IOH, 16 (9.1%) died and this difference was highly significant (p = 0.0001). A greater number of patients (47.6% v 15.9%) had a poor outcome (dead or dependant at 6 months) in the presence of IOH (p = 0.0014) and this affected both good (WFNS 1–2) and poor (WFNS 3–5) grade patients. Good (WFNS 1–2) grade patients with IOH were 6.8 times more likely to be dead or dependent than those without IOH and poor (WFNS 3–5) grade patients were 3.6 times more likely to be dead or dependent. There was no significant association between IOH and site of aneurysm, and whether the IOH was bilateral.

Conclusion: Our substantial prospective study confirms that IOH occurs in around 10% of SAH patients and indicates a significantly poorer prognosis. The presence of IOH increases the chance of a bad outcome in both both good (WFNS 1–2) and poor (WFNS 3–5) grade patients.

Are inflammatory markers predictive of delayed cerebral ischaemia after subarachnoid haemorrhage? A prospective study

C. J. McMahon^2,3, P. Tyrrell^1,2, A. T. King^1,2, A. Vail², S. Hopkins² & N. J. Rothwell² (Greater Manchester Neurosciences Centre, Salford Royal Hospitals NHS Foundation Trust¹, University of Manchester, Oxford road, Manchester² UK, Leeds General Infirmary, Leeds Teaching Hospitals NHS Trust³ UK)

Introduction: Despite advances in the treatment of ruptured intracranial aneurysms, subarachnoid haemorrhage (SAH) is recognised to cause death in 30–50% of its victims with 30–40% of survivors suffering significant neurological disability. The commonest cause of morbidity and mortality after SAH is delayed cerebral ischaemia (DCI). Ischaemic and haemorrhagic strokes induce a significant peripheral and central inflammatory response. These responses may be important in the exacerbation of ischaemic damage and in the development and exacerbation of DCI after SAH. The aim of this work was to investigate whether circulating markers of inflammation, (specifically, C-reactive protein (CRP), Interleukin-6 (IL-6), interleukin-1 receptor anatagonist (IL-1RA), white cell count (WCC) and erythrocyte sedimentation rate (ESR) predict the development of DCI after SAH.

Methods: This was a single-centre case-control study nested within a prospective cohort. 216 patients were recruited between January 2004 and August 2007. Blood samples were obtained and concentrations of IL-1RA, IL-6, WCC, CRP, and ESR were measured within 7 days of SAH prior to angiography and alternate days post-angiography (days 1, 3, 5, 7, 9 and 11) until the development of DCI or 15 days post-symptom onset. 46 patients developed DCI according to predefined criteria. 2 controls were targeted per case from the study population. Initial, peak, average, final and rate of change of each inflammatory marker were assessed. The association of inflammatory markers with the development of DCI was expressed as odds ratios (OR) derived from conditional logistic regression (controlling for age and sex).

Results: The development of DCI was positively associated with initial ESR (OR 2.4 (1.3 to 4.6); p0.006) average ESR (OR 2.3 (1.3 to 4.2); p0.006), peak ESR (OR 2.1 (1.1 to 3.9); p0.02) and final ESR (OR 2.0 (1.2 to 3.3); p0.009). Rate of change of IL-6 (OR 2.3 (1.1 to 5); p0.03), final WCC (OR 1.2 (1 to 1.3); p0.01) and rate of change WCC (OR 1.3 (1 to 1.6) p0.05) also showed an association with the subsequent development of DCI.

Conclusion: Elevated ESR in addition to leucocytosis and increasing IL-6, prior to DCI, appears to reflect impending cerebral ischaemia. Further study is required to identify the implications of this response in the pathophysiological process of DCI and the potential for preventative therapies.

Treatment of subarachnoid haemorrhage in a medium sized neurosurgical unit post International Subarachnoid Aneurysm Trial

M. Choo, S. Basu, V. Botting, J. Norris & G. Critchley (Hurstwood Park Neurological Centre, Lewes Road, Haywards Heath, UK)

Introduction: The aim of this study was to review the trend in treatment changes and assess the outcome for patients presenting with aneurysmal subarachnoid haemorrhage in a regional medium sized neurosurgical unit in the UK, following the International subarachnoid aneurysm trial (2002)¹, and in response to the Short Life Working Group report (2003).

Methods: All patients presenting with a subarachnoid haemorrhage to our neurosurgical unit between November 2002 and October 2007 were entered prospectively into a database. We assessed the patient age, sex, coexisting pathology, length of stay as well as aneurysm location and characteristics according to mode of treatment. WFNS and CT fissure grading were recorded. Patient's Glasgow Coma Score was measured at discharge and six month outcome was assessed via the extended Glasgow Outcome Scale (GOSE) and EuroQol.

Data was analysed using Microsoft Access.

Results: The total number of patients admitted to our unit with subarachnoid haemorrhage between 1.11.02 and 31.10.07 was 272, of which 188 were aneurysmal (mean 38 aSAH/year). The number of patients undergoing endovascular coiling was 95 (55%) and surgical clipping was 77 (45%) of treated aneurysms. Pre ISAT the clipping ratio was 66% (n = 33), coiling was 34% (n = 17). For the first two years post ISAT (2003–2004) the coiling rate increased, however, the following year (2005), the clipping rate was high (n = 20, 69%). The next two years showed an increase in coiling (2006 –n = 17, 74%, 2007 –n = 20, 56%).

The total number of patients who died during the admission to our unit was 26: 9 (35%) coiled; 9 (35%) clipped; 8 (31%) no treatment.

Of those patients with treated aneurysms who were alive at six months, favourable outcome was achieved in 72 (63%) patients: 40 (35%) coiled; 32 (28%) clipped. Unfavourable outcome occurred in 42 (37%) patients: 19 (17%) coiled; 23 (20%) clipped.

Conclusions: The neurosurgical outcomes in our unit post ISAT in terms of clipping and coiling are favourable and comparable to published data². Despite the early predictions following ISAT of considerable growth in treatment by endovascular coiling, our findings, in line with recent publications³, support the need for a multidisciplinary neurovascular team and a continued role for surgical intervention.

Effects of lipid profile following aneurysmal subarachnoid hemorrhage

M. Y. Tseng, P. J. Hutchinson & P. J. Kirkpatrick (Addenbrooke's Hospital, University of Cambridge, Cambridge, UK)

Introduction: Unfavourable lipid profile (high triglyceride, low high-density lipoprotein (HDL) cholesterol) is associated with development of atherosclerosis and ischaemic heart disease. We aimed to investigate effects of the unfavourable lipid profile on outcome following aneurysmal subarachnoid haemorrhage (aSAH).

Methods: Data of 160 aSAH patients were prospectively collected as part of two randomized controlled trials (pravastatin, erythropoietin). The lipid profile included cholesterol of total, low-density lipoprotein (LDL), and HDL, triglyceride, and LDL/HDL ratio (LHR). Because statins modify the lipid profile, fifty-four patients were excluded from analyses. Association between the lipid profile and adverse events were sought by using repeated measurement ANOVA and Dunnett's correction.

Results: Patients with favourable outcome (modified Rankin Scale 0–3) at discharge had higher HDL cholesterol levels on day 3 (0.99+/−0.34 vs. 0.78+/−0.30 mmol/l, p = 0.027), 6 (0.96+/−0.42 vs. 0.67+/−0.25 mmol/l, p < 0.001), 9 (0.92+/−0.39 vs. 0.64+/−0.26 mmol/l, p < 0.001), 12 (0.91+/−0.32 vs. 0.68+/−0.28 mmol/l, p = 0.027) and 15 (0.85+/−0.28 vs. 0.62+/−0.27 mmol/l, p = 0.043) following aSAH. Those with favourable outcome at 6 months also had higher HDL cholesterol levels on day 3 (0.98+/−0.33 vs. 0.76+/−0.30 mmol/l, p = 0.027), 6 (0.93+/−0.41 vs. 0.66+/−0.25 mmol/l, p = 0.015), and 9 (0.90+/−0.38 vs. 0.64+/−0.28 mmol/l, p = 0.025). Patients who developed delayed ischaemic deficits (DID) had higher triglyceride levels on day 0 (1.53+/−1.15 vs 1.07+/−0.48 mmol/l, p = 0.027) and those who died in-hospital had even higher triglyceride levels on day 0 (1.74+/−1.30 vs. 1.11+/−0.58 mmol/l, p = 0.018) and 3 (1.87+/−0.97 vs. 1.32+/−0.60 mmol/l, p = 0.025).

Conclusion: Unfavourable lipid profile during the acute phase of aSAH seems to correlate with unfavourable outcome, DID and early mortality. Immediately optimizing the lipid profile may reduce these adverse events.

The effect of the results of the STICH trial on the management of spontaneous supratentorial intracerebral haemorrhage in Newcastle

M. A. Kirkman^1,2, W. Mahattanakul¹, B. A. Gregson¹ & A. D. Mendelow¹ (Department of Neurosurgery, Regional Neurosciences Centre, Newcastle General Hospital, Newcastle upon Tyne¹ and The Medical School, Newcastle University, Newcastle upon Tyne², UK)

Objective: The role of surgery for spontaneous supratentorial intracerebral haemorrhage (ICH) remains controversial. Recently, the Surgical Trial in IntraCerebral Haemorrhage (STICH) showed no overall benefit from early surgery when compared with initial conservative treatment for ICH. Our aim was to evaluate the impact of the STICH trial results on management of ICH in the Newcastle upon Tyne Hospitals.

Method: The STICH results were released to the trial's coordinating centre, the Department of Neurosurgery at Newcastle General Hospital, in November 2003. Using ICD-10 hospital admissions data, we analysed ICH admissions before (2002) and after (2004, 2006, 2007) this. We assessed numbers of Neurosurgery and Stroke Unit admissions, numbers of clot evacuation procedures, and 30-day mortality rate (Neurosurgery versus Stroke Unit admissions). Subarachnoid haemorrhage (SAH) admissions data were also collected to corroborate our findings.

Results: There were 478 spontaneous supratentorial ICH admissions in total; 156 in 2002, 120 in 2004, 106 in 2006 and 96 in 2007. SAH admissions remained remarkably constant over this period. Neurosurgery admissions for ICH decreased significantly across the four time periods, from 71% of total ICH admissions (n = 156) in 2002 to 55% (n = 96) in 2007, and Stroke Unit admissions increased significantly from 8% (n = 156) in 2002 to 30% (n = 96) in 2007 (χ² = 20.968, p < 0.001, df = 3). Clot evacuation procedures also decreased significantly from 32% (n = 111) of Neurosurgery admissions in 2002 to 17% (n = 53) in 2007 (χ² = 11.919, p = 0.008, df = 3). 30-day mortality increased in Neurosurgery, from 14% of Neurosurgery admissions (n = 111) in 2002 to 26% (n = 53) in 2007, and decreased in the Stroke Unit, from 42 % of Stroke Unit admissions (n = 12) in 2002 to 17% (n = 29) in 2007.

Conclusion: The STICH results have significantly impacted the management of ICH in Newcastle, with a trend towards fewer Neurosurgery admissions and clot evacuations, and increased Stroke Unit admissions. Randomisation continues in STICH II for patients with superficial lobar ICH.

Could vascular neurosurgeons be trained in endovascular techniques by interventional neuroradiologists? – An observational study

D. Holliman¹, A. Gholkar² & P. Mitchell¹(Departments of Neurosurgery and Neuroradiology, Newcastle upon Tyne, UK)

Objective: The advent of interventional endovascular techniques, such as “coiling”, for treating neurovascular disorders has significantly altered the management of these conditions. In several countries, including the United States and Japan, neurosurgeons are able to perform both surgical and endovascular therapies. A study aimed at assessing vascular neurosurgeons' exposure to such techniques was undertaken. In addition, neuroradiologists' opinion regarding this was surveyed.

Methods: Questionnaires sent to neurosurgical and neuroradiology consultants in all units in the United Kingdom and Ireland. 43% (n = 75) of neurosurgical and 67% (n = 63) of neuroradiology respondees declared an interest in management of neurovascular disorders. These completed questionnaires were subjected to sub-group analysis.

Results: Vascular neurosurgeons were undertaking an average of 19 procedures/year (range 0–100). 56% (n = 42) stated an interest in interventional endovascular techniques though only one undertook such procedures. With regard to learning endovascular techniques, 20% (n = 15) would like to have the opportunity and 59% (n = 44) felt it was reasonable for neurosurgical trainees to be taught such procedures. 60% (n = 45) of vascular neurosurgeons indicated they provided some form of on-call neurovascular service of which 17% (n = 13) was 24/7 cover. Sufficient interventional neuroradiology cover to meet demand was reported by 66% (n = 50) of vascular neurosurgeons. Interventional neuroradiologists were performing an average of 85 procedures/year (range 0–350). 77% (n = 49) felt that there were sufficient interventional neuroradiologists and 24/7 cover for emergency procedures was indicated by 25% (n = 16). It was thought to be reasonable to train neurosurgeons in interventional endovascular techniques by 39% (n = 25) and 43% were willing to be involved in such training though 84% (n = 53) felt that a one year placement was unreasonable mainly because it was too short.

Conclusions: This study suggests that vascular neurosurgeons in the United Kingdom and Ireland are interested in interventional endovascular techniques though less keen to learn them. In addition, it would appear that interventional neuroradiologists are not strongly in favour of neurosurgeons learning such techniques but would be willing to teach them. The implications of these findings are unclear.

Imaging Neuronal Activity Using Intracranial Electrodes and Electrical Impedance Tomography

A. Ghosh^1,3, O. Gilad^1,2, D. Oh¹, A. W. McEvoy³ & D. S. Holder¹ (¹Departments of Clinical Neurophysiology and Medical Physics, University College London, UK, ²Abramson Center for Medical Physics, Tel-Aviv University, Tel-Aviv, Israel, ³National Hospital for Neurology & Neurosurgery UCLH, UK)

Objective: Electrical Impedance Tomography (EIT) has the potential to achieve non-invasive functional imaging of neuronal activity in the human brain [1], due to opening of ion channels during neuronal depolarization. This could be achieved due to the electrical resistance decreasing whilst ion channels open during neuronal depolarisation. The purpose of the presented studies was to determine if reproducible resistance changes could be recorded invasively with electrodes on the cerebral cortex during evoked responses.

Method: Low frequency recordings were made with an applied 1 Hz square wave current whilst evoking compound action potentials on a crab nerve. This was repeated on the cerebral cortex of anaesthetised rats and on epilepsy patients with implanted intracranial electrodes during somatosensory evoked responses.

Results: Decreases in resistance of 0.6 ± 0.1% were seen in the crab nerve during the compound action potentials. In the cerebral cortex of the anesthetised rat, no changes greater than 0.08% were observed during somatosensory evoked responses. Decreases of 0.5 ± 0.3% were observed in epilepsy patients with recording from intracranial electrodes during somatosensory stimulation.

Conclusion: These changes are encouraging and may enable invasive imaging with intracranial electrodes, which would constitute a revolutionary advance in neuroscience technology. Work in progress includes cortical imaging with improved electrodes and expanding the intracranial human dataset.

Intra-operative Optical Spectroscopy for Delineation of Eloquent Cortex

D. O. Bulters¹, P. R. Hoy², W. P. Gray¹ & H. N. Rutt² (Wessex Neurological Centre, Southampton General Hospital, Southampton, UK¹, Optoelectronics Research Centre, University of Southampton, Southampton, UK²)

Objective: The safety of surgical excision of intrinsic tumours is limited by the inability to visualise eloquent cortex adjacent to or invaded by tumours. We have therefore set out to try to use optical spectroscopy to identify eloquent cortex in patients undergoing awake surgery. In order to achieve this it was proposed that with spectroscopy we could identify the same changes in blood flow that are seen on blood oxygen level dependant BOLD MRI images.

Methods: We developed a small camera that attaches to the third port on a Zeiss operating microscope allowing surgery to be performed unhindered. This camera is sensitive to 610 nm with a 10 nm bandwidth. The camera operates at frame-rates high enough to recover the pulse rate and therefore immune to the degradation of the signal to noise ratio from this source.

Results: Initial results taken at frame rates between 2 and 10 frames per second in four patients have shown changes in signal with task execution by the patient. The areas of signal change have shown good correlation with cortical stimulation.

Conclusions: It is possible to detect changes in blood flow, which correlate with cortical stimulation and eloquent cortex using optical spectroscopy. This is the first step in generating a spectroscopic system to aid the safe and accurate excision of intrinsic tumours.

Stable Xenon CT measurement of regional cerebral blood flow in glioblastoma multiforme patients

M. Crocker, M. C. Papadopoulos & B. A. Bell (Academic Neurosurgery Unit, St George's University of London, UK)

Objective: Tumour angiogenesis plays a key role in the growth and spread of glioblastoma and is the target of new trials of vascular endothelial growth factor (VEGF) inhibitors. Different subtypes of glioblastoma can be identified on imaging characteristics including dynamic contrast enhancement and MR spectroscopy. We have investigated quantitative regional cerebral blood flow (CBF) in glioblastoma patients using Xenon CT.

Methods: In an ongoing pilot study patients presenting to St George's with suspected glioblastoma since August 2008 have had stable Xenon CT scan measurement of CBF in cortex and white matter adjacent and contralateral to the tumour, and in the tumour itself.

Results: In 14 patients studied so far there was no significant difference between normal cerebral hemisphere CBF and that containing tumour. Cortex directly overlying tumour had reduced CBF and within tumours there was a wide variability around the mean flow of 35 ml/100 g/min. High flow was measured in 4 tumours (mean 63 ml/100 g/min) and low flow in 7 tumours (mean 19 ml/100 g/min).

Conclusions: Results from this pilot study suggest that some tumours have high tissue flows whereas others have lower flows around the levels of normal white matter. Our hypothesis is that high flow glioblastomas may prove to have a poorer outcome but may respond to anti-angiogenesis chemotherapy.

A Novel MR-DTI Segmentation Technique for Tumour Boundary Delineation

T. L. Jones, B. A. Bell & T. R. Barrick (Biomedical Imaging Group & Academic Neurosurgery Unit, St George's University of London, London, UK)

Objective: Despite advances in imaging, accurate identification of brain tumour boundaries remains a difficult challenge that has implications for radiotherapy planning, monitoring of treatment response and image guided cytoreductive tumour surgery. We have developed a novel diffusion tensor imaging (DTI) segmentation technique capable of delineating space occupying lesions by automatically clustering voxels with similar isotropic and anisotropic diffusion characteristics in and around the tumour. This allows investigation of tumour infiltration patterns at the tumour boundary.

Methods: Pre-operative diffusion tensor MR scans were acquired from 59 patients (histologically confirmed as 24 high grade glioma, 11 low grade glioma, 13 metastasis and 11 meningioma) using a 1.5T General Electric Signa MR scanner with 2.5 mm × 2.5 mm × 2.8 mm resolution (TE =88 ms, TR = 8 s). Maps of isotropic (p) and anisotropic (q) diffusion characteristics were computed [1] and then segmented using a k-means clustering algorithm [2] which separated each voxel into one of sixteen segments based on its p and q characteristics. This generates an image with clearly delineated tumour and oedema boundaries that can be further investigated using a geometrical model comparing the orientation of the tumour surfaces with local white matter tract orientation (computed from the diffusion tensor) [3] as a potential measure of infiltration.

Results: Using k-means segmentation we have delineated the tumour edge and extent of oedema for each identified tumour. Having defined the tumour edge we are able to model the pattern of local infiltration and white matter tract deformation for each tumour type. Validation of this technique is in progress using an in vivo animal model of tumour infiltration.

Conclusion: Our application of k-means segmentation will be potentially useful in studying glioma infiltration patterns at the tumour boundary over time. It provides a novel approach to analysis of tumour DTI parameters and may improve definition of tumour margins for radiotherapy and radiosurgery planning and image guided surgical resection.

Multidisciplinary management of metastatic brain tumours

N. Mukerji & P. J. Kane (James Cook University Hospital, Middlesbrough, UK)

Objective: Multidisciplinary management of metastatic brain tumours has been well established for nearly a decade. We reviewed all the patients with a metastatic brain tumour treated at out centre over the past 10 years and attempted to study whether survival matched the increasing aggression in treatment.

Methods: 100 patients treated for metastatic brain tumours were identified from the histopathology records. Detailed case note review was undertaken. All data was analysed using SAS V9.1 and SPSS v16.0. Aggressive treatment was defined as administration of both chemoradiotherapy after surgery for brain metastasis.

Results: Presence of a single brain metastasis had no association with organ of origin/histology, and sex but had a significant association with aggressive Rx (Chi sq, p = 0.02). Significantly more metastatic lesions had craniotomy in the post 2003 era (Chi sq p = 0.02). This coincided with the move to a new hospital and arrival of neuronavigation. Aggressive treatment had significant association with single met (Chi sq p = 0.02) and post 2003 era (Chi sq p = 0.02) and no association with histology or organ of origin of tumor, age of patient or location of the tumor or presence or severity of co-morbidities or complications. No significant difference in median survival pre and post 2003 era (Log rank test p = 0.49) was found.

Conclusions: Single metastases are more aggressively treated, more so after 2003. Survival is poor and has not changed with evolving roles of the MDTs. Age, number of metastases and the primary organ are important predictors of survival (greater the age and number of metastases, poorer the survival); lung cancer is associated with the greatest hazard of death. Increased aggression in treating metastases has not translated into better survival.

On-call cranial neuro-oncology referrals: are they really appropriate?

H. Shekhar & P. J. Kane (James Cook University Hospital, Middlesbrough, UK)

Objective: A small proportion of referrals discussed with the on-call neurosurgery registrars are concerning patients with brain tumours. We undertook this study to evaluate the appropriateness of their urgency.

Methods: Neurosurgery referrals were analysed over a one year period. Out of a total of 1289 referrals, 169 were about brain tumours.

The parameters evaluated were the time of referral, source of referrals, patient's presentation, diagnosis mentioned and the outcome of discussion.

Results: Amongst the 169 referrals, 117 were about newly diagnosed brain tumours. Majority of the referrals were discussed in the routine hours. In spite of a robust in-house fax referral system, forty percent of the referrals were from the parent hospital.

At the time of the referral, 94% of the patients had a GCS of 14 or 15. Neurological deficit, seizure and headache were the commonest presentations. Solitary brain lesion tops the list of the diagnosis mentioned in the referrals and this includes both benign as well as malignant tumours.

Ten percent of these referrals resulted in emergency transfers. Of them, only 1 patient was transferred across for urgent surgery.

Conclusion: Majority of referrers seek advice about initial management of brain tumours. Better inter-disciplinary coordination and education could address this issue effectively.

Fluorescence image guided surgery in intracranial tumours? A prospective study of 114 consecutive lesions

M. S. Eljamel (Department of Neurosurgery, Ninewells Hospital and Medical School, Dundee, Scotland, UK)

Objective: ALA-induced fluorescence guided surgical resection (FGR) is becoming an important adjunct during malignant glioma-surgery. FGR is based on the discovery that tumor cells preferentially uptake and retain 5-aminolevulenic acid (ALA) or its byproducts. Optical technology therefore was developed to detect the generated fluorescence during surgery. ALA is a natural precursor for heme synthesis in all living mammalian cells and Protoporphyrin- IX (PpIX) is produced along the way. The conversion of PpIX to heme is a rate-limiting step in tumor cells leading to accumulation of PpIX in cancer cells. PpIX is a photoactive compound that absorbs violet-blue light and emits red fluorescence. This study reports the fluorescence outcome in 114 consecutive patients with 114 intracranial lesions to determine the potential usefulness of FGR.

Methods: A prospective observational study design. Patients were given 20 mg/kg body weight ALA mixed in non-fizzy orange juice to take orally three hours before surgery. Surgery was performed under image guidance and surgical microscopy. To detect PpIX-induced fluorescence we used a modified endoscope system and OPMI™ Pentero® surgical microscope with fluorescence module (photodiagnostic systems, PDS). The light source of both systems housed photodiagnostic filters that allowed switching light from standard to blue at the flick of a button and a camera system contained a longpass filter that allowed the fluorescence to be visualised easily. We have also used an optical biopsy system (OBS). The OBS consisted of a gallium-nitride laser emitting 405 nm blue laser light and a compact spectrometer that analysed the fluorescence and displayed its results on a laptop screen. 114 consecutive patients were studied, 47 had primary brain tumours, 22 brain metastases, 27 pituitary lesions and 18 others. The primaries consisted of 39 glioblastomas (GBM), 4 anaplastic astrocytomas (AA), 2 ependymomas and 2 lymphomas (PCL). The metastases were from lung cancer in 12, malignant melanoma in 4, colonic cancer in 5, and breast cancer in one patient. The miscellaneous group consisted of four glomus juglare tumours, two meningiomas, one chordoma, two myelomas and nine non-tumours. We have also examined dura matter, normal brain tissue, carotid arteries and optic nerves for comparison whenever possible.

Results: 94 of 105 tumors were fluorescent under the blue light (sensitivity 89.5%), All non-tumors, adjacent brain, optic nerves, vessels and dura matter showed no fluorescence under the blue light (specificity 100%). Fluorescence was observed with 100% sensitivity in some types of tumors, but the numbers were relatively small; AA, PCL, colonic metastases, choriocarcinoma metastasis, meningiomas and multiple meyloma. In tumors where there was sufficient numbers, the sensitivity varied from 83.34% in pulmonary metastases, to 85.7% in GBM, and to 87.5% in pituitary adenomas.

Conclusion: ALA – induced fluorescence was very specific and highly sensitive in a wide variety of intracranial tumors and can therefore be used to localise the lesion and guide surgical resection (FGR). However, the sensitivity was not 100% in all types of tumours because of tumour necrosis and photo-bleaching.

Effects of Lipopolysaccharide-induced inflammation on CNS regeneration following lumbar dorsal root transaction in adult rats

M. K. Hossain-Ibrahim^1,2, K. Rezajooi^1,3 & P. N. Anderson¹ (¹University College London, Gower St., ²University Hospital Coventry, Walsgrave and ³Royal National Orthopaedic Hospital, Stanmore, UK)

Objective: It has been postulated that inflammation near neuronal cell bodies enhances axonal regeneration. Injured dorsal root axons do not regenerate into the spinal cord, but stop at the dorsal root entry zone (DREZ) – i.e. at the junction with the growth-inhibitory CNS. However, injection of Corynebacterium into the DRG prior to crushing the dorsal spinal root results in a marked increase in regenerating fibres¹. The injured CNS has a much delayed macrophage/microglial response to degenerating axons than the injured PNS. We therefore investigated the possibility that LPS-induced inflammation around sites of dorsal root transection – and subsequent enhancement of the macrophage/microglial response – would enhance regeneration of their axons in the CNS.

Methods: Adult rats had a L5 dorsal root transection (controls, n = 3) or L5 transection plus LPS placement on the DREZ either at time of injury (n = 3) or a week later (n = 4). Axons were labelled with transganglionic CT-HRP and sections through the L5 DREZ analysed with immunohistochemistry 28 days after injury.

Results: The great majority of CT-HRP labelled axons appeared to stop at the DREZ in controls or LPS-treated animals. The addition of intraperitoneal LPS injections (to induce a systemic inflammatory response, n = 4) did not affect regeneration of injured dorsal root axons at the DREZ.

Dorsal root transection caused influx of macrophages to the central portion of the dorsal root, with minimal effect on macrophages and microglia in the lumbar spinal cord adjoining the DREZ and no effect on cervical DREZs and spinal cord. Addition of 2 intraperitoneal LPS injections caused dense accumulation of macrophages in the dorsal root, DREZ and adjacent lumbar spinal cord, with presumed microglial activation locally, but no effect on the spinal cord.

However, despite this microglial activation, the finding that almost all axons stopped at the DREZ was consistent in all groups analysed: controls, LPS, delayed LPS and delayed + I-P LPS.

Conclusion: Inflammation at the DREZ does not stimulate regeneration of injured dorsal roots across the DREZ into the spinal cord. Further experiments, altering the physical and/or chemical nature of the DREZ, may allow successful regeneration into the CNS.

Chronic injury reduces the production of new neurons from endogenous stem cells in the adult human brain – a microenvironment effect

M. J. Zaben¹, L. E. Sundstrom³ & W. P. Gray^1,2(¹Division of Clinical Neurosciences; University of Southampton, ²Wessex Neurological Centre; Southampton General Hospital, and ³Capsant Neurotechnologies, Southampton, UK)

Objective: Despite overwhelming evidence for the presence of stem cells in the adult brain, these cells do not produce neurons after brain injury. This is true even in the permissive niche of the hippocampus where new neurons, important for memory consolidation, are normally produced throughout life. We wished to examine the hypothesis that the injured environment reprograms these stem cells and see if this was reversible.

Methods: We isolated neural stem cells from normal cortex and from the sclerotic hippocampus in 5 patients undergoing epilepsy surgery. Stem cells were either grown as free-floating spheres in an ideal environment or within a 3D tissue (Hi-Spot®) generated from the hippocampus or cortex, and their ability to generate neurons quantified using BrdU and Neu-N immunohistochemistry.

Results: We show that stem cells from the uninjured cortex and the sclerotic (injured) hippocampus generate new neurons with equal efficiency when grown under ideal free-floating conditions. Interestingly, the numbers of newly-born neurons generated by matched stem cells grown in sclerotic hippocampal Hi-Spots were significantly lower (5.9 ± 1.4 cells/mm²) than those of the cortical Hi-Spots (23.7 ± 3.9 cells/mm²) (p < 0.001).

Conclusions: These results strongly implicate the microenvironment in defective hippocampal stem cell function in areas of chronic brain injury, and importantly from the viewpoint of brain repair, show that the effect is reversible.

Scaffolds of biomaterials to build brain neuronal aggregates and axonal connections in experimental models

J. A. Barcia, U. Gómez-Pinedo, S. Vidueira, C. Martínez, R. Prieto, A. Moreno, M. González-Hidalgo, J. Matías-Guiu, P. Simal, J. García-Verdugo & M. Monleón (Instituto de Neurociencias, Hospital Clínico San Carlos de Madrid and Centro de Investigación Príncipe Felipe, Valencia, Spain)

Introduction: Biomaterials implanted into the brain permit the growth and differentiation of neural precursors, their connection, and the long-distance axonal projections. They do so by providing a scaffold where preimplanted or local neural precursor cells grow, and by avoiding the presence of a glial scar or other local inhibitory factors. We aim to illustrate the use of implanted biomaterials in normal animals and in models of Parkinson's disease and focal ischaemia.

Methods: We used wild-type and transgenic rats expressing the green fluorescent protein (GFP). Grafts of biomaterials were implanted between the SNc and the striatum in order to reconstruct the nigrostriatal pathways in normal and lesioned rats, using peripheral nerve grafts as controls. We also implanted biomaterials, either void or prefilled with neural stem cells, into the brain of rats subjected to a focal ischaemic lesion and of normal rats. Integration and growth evidence was obtained by immunohistochemistry and electron microscopy.

Results: Axons grow from the SNc into the biomaterial towards the striatum, where signs of synaptogenesis may be found. Also, neural precursors previously implanted into the biomaterials survive up to six months, while void biomaterials are colonized by neural precursors, many of them from the ipsilateral subventricular zone. Also, the grafts are invaded by a local vascular neoformation.

Conclusions: Implanted biomaterials are invaded by local neural cells and axonal projections, thus providing a possibility for the reconstruction of central neural structures lost due to ischaemic or neurodegenerative diseases.

Poster Abstracts British Neurosurgical Group Newcastle March 2009

Seizure/Kainate enhances the proliferation and survival of Hippocampal radial glial precursor cells via AMPA receptors in-vitro and in-vivo

A. Shtaya¹ & W. P. Gray^1,2 (¹Division of Clinical Neurosciences, and ²Department of Neurosurgery, Wessex Neurological Centre, Southampton University Hospital Trust, Tremona Road, Southampton, UK)

Objective: Kainate-induced seizures and status epilepticus transiently enhances hippocampal neurogenesis but the mechanisms are not fully understood. To determine the underlying mechanisms, we examined the effects of Kainate on hippocampal precursors in-vitro and on pre-labelled and un-labelled clones of proliferating hippocampal precursors in-vivo.

Methods: Cultured hippocampal cells were prepared from rats P7-10 and exposed to 5 μM Kainate. BrdU and Ki-67 were used to measure cell proliferation while, caspase-3 and Time-lapse microscopy were used to study cell survival. Nestin, GFAP were used to label radial glial precursors. TuJ1 and Prox1 were used to stain neuroblasts and dentate gyrus cells, respectively. To examine Kainate effects in-vivo, a clone of proliferating cells in the dentate gyrus was pre-labelled with BrdU 24 hours before Kainate-induced status epilepticus and examined 6–72 h later.

Results: In-vitro, we found that Kainate increased the proliferation and survival of radial glial stem cells (nestin and GFAP positive cells), via AMPA receptors. In contrast, Kainate decreased non-radial glial stem cells (nestin only positive cells). It also enhanced the survival of radial glial precursors and TuJ1 cells with a proportional increase in neurogenesis. Consistently, Kainate/seizures in-vivo increased cell proliferation of both pre-labelled and un-labelled radial glial clones of precursors in the subgranular zone (SGZ) with increased cell cycle re-entry of the pre-labelled clone. In the granule cell layer (GCL) there was increased preferential proliferation of the pre-labelled clone in addition to increased cell cycle exit, without enhancing cell death. Kainate/seizures increased doublecortin positive cells in the GCL by 72 h.

Conclusions: We conclude that Kainate/seizure predominantly enhances hippocampal radial glial precursor proliferation via AMPA receptors without increasing cell death, and that it has a differential effect on the proliferation kinetics and fate choice of precursors in the SGZ and GCL. This may suggest targeting AMPA receptors in the treatment of memory deficits in mesial temporal lobe epilepsy.

Cerebral Cytokine Expression following Human Traumatic Brain Injury

A. Helmy¹, K. L. H. Carpenter^1,2, P. J. Kirkpatrick¹, D. K. Menon^2,3, J. D. Pickard^1,2 & P. J. A. Hutchinson^1,2 (¹Division of Neurosurgery, Department of Clinical Neurosciences, University of Cambridge, Addenbrooke's Hospital, Hills Road, Cambridge, UK, ²Wolfson Brain Imaging Centre, Department of Clinical Neurosciences, University of Cambridge, Addenbrooke's Hospital, Hills Road, Cambridge, UK, ³Division of Anaesthesia, Department of Medicine, University of Cambridge, Addenbrooke's Hospital, Hills Road, Cambridge, UK)

Objective: Traumatic brain injury causes neuronal injury via numerous mechanisms. One such mechanism, receiving increased interest, is the role of inflammation. Cytokines, a range of small proteins with pleiotropic functions, are implicated in many animal models of neuronal injury. Interest has mainly focussed on a small group of cytokines, namely, interleukin-1beta (IL-beta), tumour necrosis factor-alpha (TNF-alpha) and interferon-gamma (IFN-gamma).

Method: The purpose of this study is to characterise the cytokine response following traumatic brain injury, by utilising intracerebral microdialysis (molecular weight cut-off 100 kDa). We have collected microdialysis samples from patients following traumatic brain injury and carried out a simultaneous screen for 42 cytokines, using Luminex xMAP technology. We have identified a number of cytokines that have not previously been recovered from the human brain using this methodology, including fractalkine, Flt-3 ligand, IP-10, MCP-1, and MCP-3, as well as better-studied cytokines such as TNF-alpha, IFN-gamma, IL-1alpha, IL-1beta, IL-1receptor antagonist (IL-1ra) and IL-6.

Results and Conclusions: This study demonstrates the feasibility of the methodology, and also provides a catalogue of cytokines that are produced following traumatic brain injury. This supports and extends evidence from our previous work on cytokines in human brain microdialysates [1] as well as many previous reports of cytokine production in animal models of traumatic brain injury. This study also highlights potential targets for therapeutic intervention in man.

Handling on-calls in Neurosurgery – Lessons learnt from the maintenance of an on-call referrals database

N. Mukerji, H. Shekhar, K. S. Manjunath Prasad & P. J. Kane (James Cook University Hospital, Middlesbrough, UK)

Objective: Neurosurgical on-calls have traditionally been busy. With increasing pressures from EWTD the demand for reducing the actual working hours is increasing. A detailed analysis of the referrals received would be a way ahead in trying to identify areas where the workload can be reduced.

Methods: Our unit regularly maintains a MS Access database to record all the referrals made to the on call neurosurgical registrar. This has been in place since the end of 2006. All the details of every referral are recorded onto the database. The data recorded includes demographic details, clinical details and disposal. This database is periodically analysed and the information obtained is used to streamline the referrals process.

Results: 2157 referrals over a period of one year were analysed. Analysis of activity revealed a peak of referrals at 1700 hrs. The maximum referrals were found between 1300 and 1900 hrs. The major number of out of hours referrals were from the departments of medicine and A&E in the parent hospital. The number of referrals had no specific association with the day of the week but had peaks in months when new medical staff joined. The major diagnosis for out of hours referrals were sub-arachnoid haemorrhage (SAH), intra-cranial space occupying lesion (ICSOL), intra-cerebral haematomas (ICH) and shunt related problems.

Conclusions: This information can be used to educate the relevant departments and streamline the referrals accordingly. This would help to reduce the number of referrals and reduce actual working hours and rationalize the out of hours work. Having a two tier system for busy periods or SHO on calls with registrars as second on call during less busy times may be means of providing better training and exposure for trainees.

Investigating the expression pattern of AQP1 in a variety of CNS tumours by using tissue micro-array

K. Collinson³, A. Chakrabarty², D. Marples³, P. Chumas¹ & S. Surash¹ (Departments of Neurosurgery¹ and Neuropathology², The General Infirmary at Leeds, Great George Street, Leeds, UK and Institute of Membrane and Systems Biology³,University of Leeds, Leeds, UK)

Objective: Aquaporins (AQP's) are a family of transmembrane water channels. Within the central nervous system, AQP1 is expressed abundantly on the choroid plexus, and up regulated in glial tumours [1]. The objective of this study was to investigate the expression of AQP1 in a range of human CNS tumours.

Method: A tissue micro array (TMA) was constructed using a variety of CNS tumours, which included; glial tumours (grades 1 to 4), ependymomas, meningiomas, choroid plexus tumours, medulloblastomas and craniopharyngiomas. Immunohistochemistry was used to investigate the expression pattern of AQP1 in these tissue specimens.

Results: Quantitative analysis of AQP1 expression is being analysed by a double-blind method. Full results will be available shortly.

Conclusion: We propose that AQP1 remains a potential therapeutic target within the CNS for site-directed therapy, continuing from previous research presented at the BNRG [2].

Inflammation in delayed ischaemic neurological deficit following aneurysmal subarachnoid haemorrhage

Y. Z. Al-Tamimi¹, D. Bhargava¹, A. C. Quinn², R. Feltbower³, N. M. Orsi⁴, S. Homer-Vanniasinkham⁵ & S. A. Ross⁶ (Department of Neurosurgery, Leeds General Infirmary¹, Department of Anaesthesia, Leeds General Infirmary², Centre for Epidemiology and Biostatistics, University of Leeds³, Leeds Institute of Molecular Medicine, Welcome Trust Brenner Building⁴, Department of Vascular Surgery, Leeds General Infirmary⁵, Leeds, UK)

Objective: Evidence suggests that a microvascular and parenchymal inflammatory reaction is associated with delayed ischaemic neurological deficit (DIND) following aneurysmal subarachnoid haemorrhage (aSAH). This study investigated the level of key inflammatory and other un-investigated mediators within the plasma and cerebrospinal fluid (CSF) of aSAH patients.

Method: Plasma and CSF samples were collected from 36 aSAH patients at days 3, 5, 7 and 9 following ictus and 10 control patients undergoing elective orthopaedic surgery requiring spinal anaesthesia. DIND developed in 12 patients with aSAH. Levels of interleukin (IL)-1α, IL-1β, IL-4, IL-6, IL-8, IL-10, IL-15, IL-17, IL-18, tumour necrosis factor (TNF)-α, vascular endothelial growth factor (VEGF) and monocyte chemotactic protein (MCP)-1 were measured by multiplex immunoassay. Statistical analysis was conducted by Mann-Whitney-U tests following normality testing (Shapiro-Wilks test).

Results: Day 3 CSF concentrations were significantly higher than control CSF for all mediators except for IL-1α (p < 0.001). Day 3 CSF concentrations were significantly higher in DIND patients for all mediators apart from IL-1α, IL-18 and MCP-1, although MCP-1 levels increased significantly by day 7 (p = 0.05). Plasma IL-6 and MCP-1 levels were also higher on day 7 in DIND patients (p < 0.05). CSF levels of IL-6, IL-8, IL-10, IL-15, IL-18, VEGF and MCP-1 were significantly higher than in plasma at days 3, 5 and 7 (p < 0.05).

Conclusion: In this small pilot study, evidence suggests that an inflammatory process occurs early after aSAH, peaking at days 5–7, and is compartmentalised within the central nervous system. Early CSF levels were higher in patients subsequently developing DIND highlighting a possible role for these mediators in predicting insipient DIND. This warrants further investigation in a larger study.

Validating SF-36 following traumatic brain injury

M. R. Guilfoyle, H. Seeley, C. Harkin, H. Richards & P. J. A. Hutchinson (Academic Department of Neurosurgery, Addenbrooke's Hospital, Cambridge, UK)

Objectives: The Medical Outcomes Survey Short-Form 36-Item Health Questionnaire (SF-36) is a popular quality of life measure that may provide improved outcome assessment following traumatic brain injury (TBI). This study investigated whether the fundamental psychometric properties that underpin SF-36 score calculations are maintained in the TBI population and consequently if SF-36 scores can be validly reported for these patients.

Methods: Patients reviewed in the neurotrauma outpatient clinic following TBI completed the SF-36 questionnaire. Item responses were linearly transformed into the eight SF-36 health domain scales and the physical and mental component summary measures as per the published instructions [1]. The statistical relationships between item responses and domain scales defined by the original developers of SF-36 were examined in the study sample [1],[2]. Analysis was carried out in SPSS 15.0.

Results: A total of 370 patients were reviewed 1–24 (median 6) months after TBI. The median age was 29.0 and 75% were male. The SF-36 questionnaire data were 99.6% complete. Responses to the individual items were mostly symmetrical in distribution, with similar means and standard deviations. Item-own scale correlations were strong and all exceeded 0.4, indicating convergent validity. Further, for all 36 items the item-own scale correlation was greater than all item-other scale correlations, confirming the discriminant validity of each item. Cronbach's Alpha coefficients for the eight domain scales were all more than 0.8, and in each case exceeded the correlations between domains, demonstrating that the domain scores are reliable estimates and measure distinct constructs from one another. Five domains had significant floor and/or ceiling effects with over 15% of responses at one or both extremes of the scale. Principal components analysis of the eight domain scores found only a single component with an eigenvalue greater than unity, explaining 62.8% of the variance.

Conclusion: SF-36 responses from patients recovering from TBI fulfil the basic requirements to derive meaningful health domain scale scores. However, physical and mental component summary scores may not be valid in this context and should be interpreted with caution. SF-36 is a useful adjunct measure of outcome in TBI alongside traditional disease-specific assessments.

Training neurosurgeons in interventional endovascular techniques – the trainees view

D. Holliman¹, C. Barrett¹, A. Gholkar² & P. Mitchell¹ (Newcastle upon Tyne, UK)

Objectives: The development of interventional endovascular techniques for treating neurovascular conditions has significantly altered neurosurgical trainees' exposure to the management techniques involved. An observational study was undertaken to assess exposure to such techniques and whether aspirations existed to learn interventional endovascular techniques amongst neurosurgical trainees or to teach them amongst neuroradiology trainees.

Methods: The study involved completion of questionnaires sent to neurosurgical and neuroradiology trainees in all units in the United Kingdom and Ireland. Replies from neurosurgical registrars in 28 of 36 units (78%) were received with a response rate of 29% (n = 67). For neuroradiology trainees, 8% (n = 12) replied from 11 of 33 contacted units (33%).

Results: 43% (n = 29) of neurosurgical trainees declared an interest in vascular neurosurgery, being involved in an average of 13 neurovascular procedures/year (range 0–70). With regard to endovascular techniques, 42% (n = 28) had an interest but none undertook such procedures. Sufficient interventional neuroradiology cover to meet demand was reported by 52% (n = 35) while 21% (n = 14) indicated their units provided 24/7 vascular neurosurgical cover. With regard to an opportunity to learn endovascular procedures, 73% (n = 49) of neurosurgical trainees wanted this though 28% (n = 19) felt this should be restricted to trainees with an interest in vascular neurosurgery. With regard to neuroradiology trainees, 83% (n = 10) had an interest in endovascular techniques undertaking an average of 55 procedures/year (range 20–90). Sufficient interventional neuroradiology provision was reported by 67% (n = 8) though emergency interventional endovascular cover was indicated to be available 24/7 by 33% (n = 4). Of neuroradiology trainees, 33% (n = 4) thought it reasonable to train neurosurgical trainees in interventional endovascular techniques and were willing to be involved in such training.

Conclusions: This study suggests that neurosurgical trainees in the United Kingdom and Ireland are keen to learn interventional endovascular techniques. In addition, it would appear that neuroradiology trainees are not strongly in favour of neurosurgeons learning such techniques though the number of respondees was low. These findings may have implications for the future training of neurosurgeons.

STICH II – An update on recruitment centres, screening activity and characteristics of patients randomised to date

E. N. Rowan, B. A. Gregson, P. Mitchell & A. D. Mendelow on behalf of the STICH II investigators (STICH Office, Newcastle University, Ward 31 North Wing, Newcastle General Hospital, Newcastle upon Tyne, UK)

Objective: The MRC funded STICH II trial will establish whether early surgery is better than initial conservative treatment for patients with superficial (<1 cm) spontaneous supratentorial lobar haematomas. This subgroup was identified as having a better outcome in the first STICH trial [1].

Methods: The STICH team are registering collaborating centres to assist with recruitment. At least 15 patients per month are required to achieve our interim target of 250–290 patients by December 2009. Participating centres are required to complete monthly screening logs. Patients are randomised to either initial conservative treatment or early surgery and receive CT scans at pre-randomisation and 5 days. Following this, a 2 week/discharge form is completed by the medical team and a 6 month postal questionnaire is completed by each patient and/or their relatives/carers.

Results: At the time of writing (6th January 2008), 68 centres from 18 countries have registered and 36 (53%) of these have recruited at least one patient. Screening logs are submitted one month in arrears and therefore 61 centres have been established long enough to have submitted at least one log to date. Completion of screening logs is variable with 41/61 established centres (67%) completing and submitting at least one log so far. Nevertheless, of the 20 established centres who have not yet submitted their expected screening logs, 7 have recruited at least one patient and therefore 48/61 (79%) centres have actually demonstrated some evidence of recruitment-related activity. 116 patients have been randomised and the 6 month outcomes including deaths are known for 63/69 patients who have reached the 6month follow up time point to date; 15 (22%) have died. Updated versions of these figures will be presented at the meeting.

Conclusion: More patients are needed to attain our target recruitment. Work is underway to engage more centres and support existing centres which have not yet recruited patients. The STICH II team can be contacted through the website (www.ncl.ac.uk/stich) or by emailing: [email protected].

Simvastatin in aneurysmal subarachnoid haemorrhage – the STASH trial: an update

C. L. Turner¹, M.-Y. Tseng¹, A. D. Mendelow², P. J. A. Hutchinson¹ & P. J. Kirkpatrick¹ (¹Academic Dept of Neurosurgery, Addenbrooke's Hospital, Cambridge, ²Dept of Neurosurgery, Newcastle General Hospital, Newcastle upon Tyne, UK)

Objective: Pilot studies have demonstrated that acute statin therapy following aneurysmal subarachnoid haemorrhage reduces the incidence of vasospasm-related delayed ischemic deficits, reduces frequency and intensity for rescue therapy and improves long-term outcome at 6 months [1],[2]. We are conducting a phase III multi-centre randomized-controlled trial to establish if this holds true in a larger population.

Methods: 1600 aneurysmal subarachnoid haemorrhage patients (age 18–65 years, onset < = 96 hours) will be randomized to receive daily oral Simvastatin 40mg or placebo for up to 21 days. Primary outcome measure is the modified Rankin Disability Score at 6 months. Secondary outcome measures are the SF-36 questionnaire at 6 months, the incidence, duration and need for delayed ischemic deficit (DID) rescue therapy, incidence and severity of sepsis, length of intensive care and total acute hospital stay and percentage of patients discharged directly home.

Results: In 24 months, 334 patients (106 male, 228 female), have been recruited from 23 centres worldwide, (mean age 50 years, range 21–65 years). 247 patients (75%) were WFNS grades 1–2 on arrival, 84 (25%) were grades 3–5. Mean length of overall stay was 19 days (range 5–120 days). 142 patients (43%) were admitted to the intensive care unit, mean length of ITU stay 12 days (range 1–42 days). 156 patients (56%) were mRS 1–2 on discharge and 25 (7%) have died, of which 3 were post discharge. There have been no unexpected adverse events. We have 6 month outcome data on 242 patients (97%). Both UK and overseas Centres will be welcomed to participate, we have funding in place until July 2011.

The RESCUEicp decompressive craniectomy trial

I. Timofeev¹, P. J. Hutchinson¹, E. Corteen¹, S. Grainger¹, M. Czosnyka¹, D. Menon², J. Pickard¹ & P. Kirkpatrick¹ on behalf of the RESCUEicp investigators www.RESCUEicp.com(¹Academic Neurosurgery Unit and ²Division of Anaesthesia, University of Cambridge/Addenbrooke's Hospital, Cambridge, UK)

Objective: The RESCUEicp study (randomised evaluation of surgery with craniectomy for uncontrollable elevation of ICP) aims to provide Class I randomised evidence as to whether decompressive craniectomy is effective for the management of patients with raised and refractory intra-cranial pressure following severe traumatic brain injury.

Methods: An international multi-centre randomised trial comparing decompressive craniectomy with medical management. Patients (with traumatic brain injury and raised ICP (>25 mmHg) refractory to initial treatment measures are eligible for the study. The total number of patients will be 600 (300 in each arm of the study) for a 10% difference in outcome (increase in favourable outcome from 45% to 55%) (power = 80%, p = 0.05).

Patients are randomised to one of two arms: continuation of optimal medical management (including barbiturates) versus surgery (decompressive craniectomy). The inclusion criteria are: traumatic brain injury, age 10–65 years, abnormal CT scan. The exclusion criteria: bilateral fixed and dilated pupils, bleeding diathesis, devastating injury when patient is not expected to survive 24 hours, follow up not possible, primary decompression, barbiturates prior to randomisation, brainstem involvement. Patients treated on the Lund protocol are also not eligible.

Outcome is assessed using the extended Glasgow Outcome Score and SF-36 quality of life questionnaire at 6 months 1 year and 2 years, with additional surrogate endpoints (ICP control, length of stay in ITU, health economics analysis).

Results: The pilot phase of the study (n = 50) has been completed. Over 170 patients have been recruited to date. The study is ongoing with growing international participation. An up to date progress of the trial will be presented.

Conclusion: Randomising patients with traumatic brain injury to decompressive craniectomy versus optimal medical management is feasible. Whether this operation is effective and safe remains to be seen. We would welcome more centers participation.

British Neurosurgical Research Group In association with the Neurosurgical Department, Kings College Hospital, London Annual Meeting Abstract Booklet 2008

The role of neuropilin-1 in benign and malignant neuroepithelial tissue

des Etages (A Wessex Neurological Centre, Southampton General Hospital)

Introduction: To investigate the role of neuropilin-1 in benign as well as malignant neuroepithelial tissue by (1) demonstrating the presence of neuropilin-1 in benign and malignant neuroepithelial tissue using immunohistochemistry, (2) investigating differences in staining intensity amongst neuroepithelial tissue from different locations, different histological type and grade, different gender and (3) testing for any significant associations between age and staining intensity of neuropilin-1.

Methods: A commercial brain tissue microarray consisting of nineteen (19) core biopsy specimens of normal brain tissue and forty-eight (48) specimens of neuroepithelial tumours was de-paraffinised and subjected to DAB immunohistochemistry to demonstrate the presence of cytoplasmic polyclonal rabbit neuropilin-1. Staining intensities were quantified using optical density analysis with the “Image J” image analysis Java platform. The resultant optical densities were compared using Kruskal Wallis ANOVA and the Mann-Whitney U test, to detect any significant differences. Linear regression was performed with the S-curve to test for any association between age and staining intensity of benign neuroepithelial tissue.

Results: Neuropilin-1 was demonstrated in all neuroepithelial tissue. No significant differences in optical density were detected between optical density and neuroepithelial tissue type, gender or the location from which the biopsies were taken. No significant association was found between age and optical density of neuropilin-1 staining.

Conclusions: Neuropilin-1 is found in the cytoplasm of benign brain tissue and in the cytoplasm of the neuroepithelial tumours investigated. The ubiquitous presence of neuropilin-1 throughout the neuroepithelial tissues suggest that the tumours themselves may be aberrations of normal developmental processes along the neuroepithelial cell line. The findings suggest that neuropilin-1 plays a biological role in neuroepithelial tissue but is unlikely to play a significant role in the aetiology of neuroepithelial tumours and may serve only a facilitatory function to survival.

Improving accuracy in the characterisation of primary brain tumours with methionine PET: Combining measurement of tumour amino-acid uptake with evaluation of changes in the adjacent brain

D. J. Coope^1,2, C. Eggers³, J. Cížek^3,4, S. Vollmar³, W.-D. Heiss³ & K. H. Wolfson¹ (Molecular Imaging Centre, The University of Manchester, UK¹; Department of Neurosurgery, Salford Royal NHS Foundation Trust, Salford, UK²; Max-Planck-Institute for Neurological Research, Cologne, Germany³ and Centre for Applied Computer Science, University of Cologne, Germany⁴)

Introduction: Accurate pre-operative characterisation of primary brain tumours is essential to identify the most appropriate approach and timing of any intervention. Peak tumour uptake of labelled amino-acids as measured with PET has been shown to correlate with tumour grade and other features of malignancy such as microvascular proliferation. However, overlap in the uptake characteristics of the different histological sub-types of gliomas makes interpretation of the results difficult at initial presentation. Identification of other features of malignancy such as the effects on the brain adjacent to the tumour may be beneficial in supporting this assessment.

Methods: Thirty-nine 11C-methionine PET scans in patients with gliomas (13 WHO grade II, 11 grade III and 15 grade IV) were identified from the database at the Max-Planck-Institute for Neurological Research. Eighteen of these scans had been performed prior to any surgical intervention. Scans were analysed using the ratio to a normal uptake map as previously described [1] and this was extended to demonstrate statistical variability from the mean at each point in the scan. An automated method was developed to define and measure the tumour volume using a constrained region-growing technique. Texture analysis parameters were calculated using three-dimensional spatially invariant co-occurrence matrices for whole tumour and tumour with adjacent brain.

Results: Tumour grade was found to correlate with tumour volume (p < 0.01), peak methionine uptake (p < 0.05) and the texture measure of correlation (p < 0.05). Receiver operator curve (ROC) analysis demonstrated these three parameters to be the most effective in distinguishing low-grade and high-grade tumours (AUC for tumour volume = 0.87). Texture measures of entropy and co-occurrence mean on tumour with adjacent brain, reflecting loss of the normal distribution in adjacent brain, were similarly effective in separating low and high-grade tumours (ROC AUCs of 0.84 and 0.83 respectively). Discriminant analysis using peak uptake versus the product of the texture measures of entropy and co-occurrence mean gave 94.9% accuracy in distinguishing low and high-grade tumours from a single scan.

Conclusion: Combining functional assessments of tumour activity with evaluation of the surrounding brain may provide an objective measure to improve accuracy in distinguishing low-grade and high-grade tumours for treatment planning.

How can we improve preoperative diagnosis of brain tumours?

R. A. Corns, T. Rittman & K. Ashkan (Dept of Neurosurgery, King's College Hospital, London)

Introduction: It can be very difficult for referring hospitals and indeed specialist neurosurgical units to distinguish between brain abscesses and high-grade brain tumours. We conducted a review of notes of all patients diagnosed with a high-grade primary brain tumours, brain metastasis or brain abscess in the year 2006. We reviewed what diagnosis is suspected in the patient's local hospital and in our neurosurgical unit and compared that with the histopathological diagnosis. We also looked at which of the investigations performed were most helpful.

Methods: 10 of 14 (71.4%) abscesses were correctly identified at the patient's local hospital. One case was referred as an uncertain diagnosis and 3 were misidentified as tumours. On review at the tertiary referral centre by the registrar on-call, 3 of these abscesses were correctly identified and one was misidentified as a tumour. One abscess was correctly referred from the local hospital and misdiagnosed as a tumour prior to operation at the tertiary referral centre. Therefore, at the tertiary referral centre 12 of 14 (85.7%) abscesses were correctly diagnosed correctly.

Results/Conclusion: 114 of 122 (93.4%) tumours were correctly identified at the patient's local hospital. The six other tumours were referred as an uncertain diagnosis. Three of these were correctly identified as tumours at the tertiary referral centre and the other three were of uncertain diagnosis pre-operatively, giving an overall correct diagnosis in 117 of 122 (95.9%) cases.

There was a statistically significant difference between the CRP values in abscesses and tumour (p = 9.06×10-5). We used a number of cut-off points (>5 mg/l to >50 mg/l in units of 5) to try and identify a difference between tumour and abscess. Sensitivity reached 91.1% at >20 mg/l, however sensitivity was only 66.7%. Therefore the difference in CRP values is not clinically significant. There was a statistically significant difference between WCC in tumours and abscesses (p = 4.1×10-5), however the difference in means is too small to be clinically significant (mean 18.5×10-9/l in abscesses and 11.3×10-9 in tumours). When CT was used, 3 abscesses were incorrectly diagnosed as tumours. When MRI was used, no abscesses or tumours were misdiagnosed.

Assessment of malignant glioma biochemistry in vivo: a microdialysis study of energy-related molecules, growth factors and cytokines

H. J. Marcus¹, K. L. H. Carpenter^1,2, S. J. Price¹ & P. J. A. Hutchinson¹ (Academic Neurosurgery Division¹ and Wolfson Brain Imaging Centre², Department of Clinical Neurosciences, University of Cambridge, Addenbrooke's Hospital, Cambridge)

Introduction: Microdialysis is a technique that enables the measurement of the chemistry of the cerebral extracellular fluid. The objective of this study was to utilise microdialysis to monitor levels of glucose metabolites, glutamate and glycerol in patients undergoing surgery for intrinsic brain tumours, and to assess the concentration of growth factors and cytokines involved in the proliferation, invasion and angiogenesis of high grade gliomas in vivo.

Materials and Methods: Six patients with suspected high-grade gliomas were studied. Five of these underwent resection with one catheter placed at the tumour resection margin and a second in macroscopically normal peritumour tissue. The remaining glioma patient had an image-guided biopsy with a single catheter inserted stereotactically at the tumour margin. Of the tumour patients recruited, histology confirmed the presence of a WHO IV glioblastoma in five cases and one cerebral lymphoma. A patient with traumatic brain injury (TBI) undergoing routine monitoring with microdialysis was also included, as a non-cancerous control.

Results: In high-grade gliomas, tumour margin samples consistently showed significantly higher lactate, higher L/P, higher glutamate and higher glycerol, relative to peritumour samples (p < 0.05). This was in contrast to lymphoma and TBI samples. Growth factor and cytokine concentrations showed great variability between samples. Though differences in concentrations between tumour margin and peritumour samples did not reach statistical significance, IL-8 appeared higher, and TIMP-1 lower, within tumour margin samples when compared to surrounding tissue. One of the known properties of IL-8 is its ability to promote angiogenesis, whilst TIMP-1 is an endogenous inhibitor of matrix metalloproteinase, therefore lowering of TIMP-1 may promote tumour invasion.

Conclusion: Overall, the results indicate that this tissue of the glioma tumour margin is metabolically extremely active, with the biochemistry suggesting a glycolytic environment favouring angiogenesis and tissue invasion. Further research is needed to elucidate the biochemistry of glioma metabolism.

Visual field deficits after selective Trans-Sylvian Amygdalohippocampectomy

A. A. Khan & W. P. Gray (Wessex Neurosciences Centre, University of Southampton)

Introduction: To identify and evaluate the incidence of Visual Field Deficit (VFD) after selective Trans-Sylvian Amygdalohippocampectomy (TAH) performed consecutively by one surgeon and its clinical significance.

Methods: A retrospective analysis of Goldmann Perimetry charts using the I4e stimulus as standard.Total 27 patients, 13 right sided operation and 14 left sided operation; Age range 5–55 yrs. Mean follow up 2 years 7 months. All patients underwent TAH for epilepsy with a histological diagnosis of mesial temporal sclerosis (MTS).

Results: Of 27 patients 9 (33%) had no significant VFD in either eye, 18 (67% compared to 77%(2)) had VFD. The eye ipsilateral to the operation was more likely to have deficits closer to the central point of fixation than that contralateral to the operation. Of 54 eyes (27 patients), 24 eyes (44%) had a deficit below 20 degrees. These defects arose in 9 patients in both eyes and in 6 patients in the ipsilateral eye only. The likely explanation of this incongruency is that the fibres from ipsilateral and contralateral retinas extend equally anteriorly but fibres from the ipsilateral retina lie more laterally in the temporal lobe. Comparison of pre-op Visual Fields (VFs) showed no significant difference between patients later undergoing left or right surgery. 22 of 27 (81%) patients were seizure free (Engel 1), (compared to 85%(1)). Patients with VFD were more likely to have good seizure control, 16 of 18 patients (89%) Engel 1, compared to those with no VFD, 6 of 9 patients (67%) Engel 1. 15 Engel 1 patients (56% of all patients compared to 38%(2)) could have obtained a driving license using the smaller I4e stimulus, a less lenient criteria than the III4e stimulus stipulated by the DVLA.

Conclusions: Studying post-operative VFs provides an insight into the anatomy of resection involving the optic pathway and its effect on patients' ability to drive. Awareness of the potential effect of resection on VFs could guide in future surgical planning. Further work is in progress using higher resolution MRI imaging to determine any correlation between VFD and resection characteristics.

Memory and neurogenesis in chronic Temporal Lobe Epilepsy

L. Barkas¹, E. Redhead² & W. P. Gray^1,2 (Division of Clinical Neuroscience, University of Southampton¹ and Wessex Neurological Centre Southampton²)

Introduction: Patients with Temporal Lobe Epilepsy (TLE) often have untreatable progressive memory deficits. This may be due to damage to the hippocampus, a temporal lobe structure which has been consistently associated with memory. The hippocampus is uniquely responsible for adult neurogenesis, the division of adult stem cells to produce new neurons. Hippocampi from patients with intractable TLE and animal models of chronic epilepsy show reduced neurogenesis. This is suggestive of a relationship between levels of neurogenesis and the memory deficits found in TLE. Therefore we are investigating dentate neurogenesis and spatial memory impairment in chronic TLE.

Methods: Twelve Patients who have had a selective hippocampalamgydalaectomy to treat chronic TLE were recruited for this study. They completed spatial memory tasks using a Virtual Morris Water Maze, and their performance was compared to that of drug-controls and healthy age-matched controls. Furthermore rats with kainate induced TLE and sham injected controls were tested on a variety of spatial memory tasks at one, three and five month's post kainate injection. The rats were sacrificed on completion of the testing and the hippocampi quantitively analysed for neurogenesis using Bromodeoxyuridine (BrdU) as a marker of proliferation, and NeuN to identify neurons.

Results: Patients who had right sided surgery perform significantly worse in complex spatial memory tasks, and show pronounced deficits in spatial tasks which require the use of abstract visual cues. Animals with TLE perform significantly worse in complex spatial memory tasks in all time point groups but decrease and stabilize over time. However the animals' performance in memory retention tasks gets worse over time, and is significantly decreased at five months post-kainate. Immunohistological analysis has shown that levels of neurogenesis were unchanged between controls and kainate rats in one month post-kainate, decreased at three months post-kainate, and were significantly reduced five months post kainate.

Conclusion: Both patents who had right sided surgery and animals with kainate induced TLE have a pronounced and similar pattern of spatial memory deficits. Our results suggest that impairments in the retention of spatial memory in chronic epilepsy may be due to reduced or altered neurogenesis.

Control and recruitment of hippocampal stem cells in epilepsy

A. B. Y. Shtaya¹, A. K. Pringle¹ & W. P. Gray^1,2(Division of Clinical Neuroscience, University of Southampton¹ and Wessex Neurological Centre²)

Introduction: It is widely accepted that ongoing neurogenesis occurs in the adult hippocampus throughout life, and is important for learning and memory. Deficits in learning and memory are common in patients with temporal lobe epilepsy. Animal models show that kainate induced seizures and epilepsy alter hippocampal neurogenesis, but the mechanisms are unknown. In order to understand the effects of both seizures and kainate on hippocampal stem cells we first investigated kainic acid effects on hippocampal stem cells in cultures. Then, we used an animal model of kainic acid induced epilepsy to explore the effects of kainate/seizures in vivo.

Methods: Primary hippocampal cell cultures were prepared from Wistar rat Pups 7–10d and maintained in vitro for 4h–7d. Cells were exposed to 5 μM kainate for 4h–7d. We measured cell death, proliferation, and total cell numbers by counting Propidium Iodide (PI), Caspase-3-positive cells and Mitotracker, Ki67, BrdU incorporated cells, and DAPI stained cells, respectively. Cell phenotype was determined using immunohistochemistry against nestin and TuJ1. Survival and proliferation of neurons and stem/progenitor cells were also investigated. AMPA, NBQX and GYKI52466 were used as agonist and/or antagonists to study receptor modulation for kainate effects.

Results: Kainate was both trophic and proliferative to hippocampal stem cells in-vitro, acting on both nestin positive stem cells and more differentiated neuronal precursors. The effects of kainate were AMPA receptor mediated, as they were blocked by AMPA receptor antagonists and replicated by AMPA. In-vivo results examining a clone of pre-labelled stem cells, showed that acute seizures did not kill stem cells but resulted in altered proliferation kinetics and disruption of the neurogenic niche. Our data suggests that status epilepticus increase neurogenesis by enhancing the survival and proliferation of stem/progenitor cells and that the AMPA glutamate receptor is important for the proliferation and survival of adult stem cells.

Decompressive craniectomy following traumatic brain injury (TBI) – the RESCUEicp study

E. Corteen on behalf of the RESCUEicp investigators (Academic Unit of Neurosurgery Addenbrookes hospital, Cambridge)

Introduction: The RESCUEicp study (randomised evaluation of surgery with craniectomy for uncontrollable elevation of ICP) aims to provide Class I randomised evidence as to whether decompressive craniectomy is effective for the management of patients with raised and refractory intra-cranial pressure following traumatic brain injury.

Methods: A multi-centre randomised trial comparing decompressive craniectomy with medical management. Patients (n = 50 for the pilot phase, n = 600 for the main study) with traumatic brain injury and raised ICP (>25 mmHg) refractory to initial treatment measures are eligible for the study. Patients are randomised to one of two arms: continuation of optimal medical management (including barbiturates) versus surgery (decompressive craniectomy). The inclusion criteria are: traumatic brain injury, age 10–65 years, abnormal CT scan and the exclusion criteria: bilateral fixed and dilated pupils, bleeding diathesis, devastating injury not expected to survive 24 hours. Outcome is assessed using the extended Glasgow Outcome Score and SF-36 quality of life questionnaire at 6 months with additional surrogate endpoints (ICP control, length of stay in ITU).

Results: The pilot phase of the study has been completed; a summary of these results will be presented. The study is on going. The background, protocol, progress of the study will be presented.

Conclusion: Randomising patients with traumatic brain injury to decompressive craniectomy versus optimal medical management is feasible. Whether this operation is effective and safe remains to be seen.

Towards a human cellular model of traumatic brain injury for therapeutic discovery

O. N. Pathmanaban^1,2,3, A. Ekonomou¹, S. L. Minger¹ & C. Tolias² (Stem Cell Biology Laboratory, Wolfson Centre for Age-Related Diseases, Guy's Hospital, London¹, Department of Neurosurgery, King's College Hospital, London²,Department of Neurosurgery, Greater Manchester³)

Traumatic brain injury (TBI) is the leading cause of morbidity and mortality in young people worldwide. In vitro research relies on rodent primary cultures, but human cells would be more clinically relevant. Until now, limited tissue availability has made this unworkable. We aim to solve this problem by using human neural stem cells, which might be propagated in culture for extended periods and differentiated as required to provide unlimited neural cells for in vitro disease modelling. We have retrieved stem-cell rich tissue from the brains of 5 patients undergoing endoscopic neurosurgery, with no operative complications or additional tissue loss. Suspensions of these cells and also human embryonic stem cells have then been cultured as adherent mono-layers in medium selective for neural stem cell growth. Ultimately, human stem cell-derived neurons will be integrated into a novel in vitro model of TBI. Therefore a prototype of this model has been developed. Cells are seeded onto stretchable silicone membranes, which are subsequently clamped and sealed above a vacuum chamber. On application of a variable vacuum, the membrane is rapidly pulled down into the chamber and stretched over a platform, inflicting stretch injury to the adherent cells. For combined stretch-hypoxia insults, this is performed within an oxygen-controlled chamber.

The role of gamma knife surgery in the management of trigeminal neuralgia

J. R. Panchmatia¹, P. R. Bullock¹ & C. Lindquist¹ (Cromwell Hospital, London¹)

Introduction: Trigeminal neuralgia (TN) is managed medically at the outset and surgery considered if medical interventions fail. In the fifty years since Leksell's groundbreaking studies we have become increasingly aware of the role gamma knife surgery (GKS) has to play in the management TN.

Materials & Methods: A retrospective analysis of patients who underwent GKS, at a national centre, to treat TN.

Results: Our cohort (N = 90) have all been treated at a single institution in central London. Our preliminary results are based on our first 23 patients and have shown a mean age of 66.4 years. Patients suffered with TN for a mean of 9.8 years before being referred. General Practitioners were the commonest referral source (39.1%). In addition to medical management 11 had undergone previous surgery. 43.5% suffered with concurrent conditions such as diabetes mellitus. 1 patient opted for GKS as he was the sole carer for his wife and did not wish to be admitted to hospital. All patients received a treatment dose of between 70 Gy and 90 Gy with a modal value of 80 Gy. Patients were followed up for 13.0 months. 26% reported a reduction in pain at their latest follow up. 35% reported that their pain had resolved. There were 3 cases of reduced sensation post-GKS and 1 of these patients reported a subcutaneous haematoma. All patients were discharged within 1 night of GKS.

Conclusions: 61% reported a symptomatic improvement post- GKS and complications were both minor and infrequent. GKS is particularly beneficial in those with concurrent medical conditions making a general anaesthetic inadvisable and those that do not wish to be admitted to hospital for social reasons. The profile of GKS must be raised as it is unsatisfactory that patients are waiting for 9.8 years before being referred.

Epidemiology of head injury in England and Wales

S. Robinson¹, S. Thornton¹ & F. Lecky² (Royal Bolton Hospital, Bolton, Lancashire¹, Hope Hospital, Salford, Greater Manchester²)

Introduction: To identify the most common traumatic head injuries sustained within England and Wales and to compile a demographic profile of those most likely to present with such injuries.

Methods: Using data from the Trauma Audit and Research Network (TARN) database we selected all entries between 1996 and 2003 of patients sustaining serious head injuries. 14,760 patients met the inclusion criteria. These were then used to create frequency tables relating to the 10 most common head injuries.

Results: The most common head injury sustained is a subdural haematoma. Men were the most likely to sustain all of the 10 most common head injuries and road traffic accidents were the most frequent cause in all incidences except subdural haematoma which were most likely to be caused by falls. Logistical regression analysis identified GCS <9, ISS >16, age >55 and male gender as independent predictors of mortality.

Conclusions: Significant Head injuries most commonly occur to men during a RTA. Mortality is predicted by a GCS of <9, ISS>16, age > 55 and male gender.

The role of a Neurotrauma clinic

H. M. Seeley¹, S. Kirker², C. Dias³, C. Harkin¹ & P. J. Hutchinson¹ (Department of Neurosurgery, Addenbrooke's Hospital, Cambridge¹, Department of Rehabilitation Medicine, Addenbrooke's Hospital, Cambridge², Headway Cambridge³)

Introduction: Survivors of head injury are often left with varying degrees of disability and complex and varied needs, necessitating prolonged periods of rehabilitation and continuing care. Advances have been made in the acute management of these patients, but continuing management in terms of rehabilitation remains deficient with lack of specialist resources and a fragmented service. For head-injured patients, lack of access to appropriate ongoing rehabilitation may have profound effects on personal outcome and social re-integration. There are also considerable economic implications for planning and provision of services [1]. We describe, review and assess the role of a Neurotrauma clinic within the Head Injury Service at Addenbrooke's Hospital.

Methods: The multidisciplinary Neurotrauma clinic commenced in 2003 as a result of a collaborative research programme to assess current regional head injury service provision, and address deficiencies and management issues [2]. The clinic acts as a ‘gateway’ to access appropriate ongoing rehabilitation and a source of information and support. Patients also complete an SF-36 and GOSE questionnaire and data is entered on a database. This enables tracking of individual patient progress and outcome and provides an information resource for further research.

Results: 319 patients of all ages with ongoing problems following varying severity of head injury are included in this study. Patients were followed up at 3 months + post injury as appropriate. The findings highlight deficiencies in rehabilitation both in general service provision and specific patient need. 65% of patients were receiving adequate rehabilitation at 6 months post injury. Also only 25% were back to work at 24 months + post injury. Mean Physical and Mental Health SF-36 scores remained stable with time.

Conclusions: Evidence in support of demand, need and effectiveness of rehabilitation for head injury is particularly relevant within the limited resources of the NHS. Early indications show that the clinic can assist in providing continuity of patient care, improving coordination of services, and act as a resource for further research on epidemiology, outcome and impact of rehabilitation.

A microdialysis study of the effect of oral vigabatrin administration on the brain extracellular fluid composition in head injury patients: a preliminary evaluation

K. L. H. Carpenter^1,2, I. Timofeev^1,2, J. Nortje^2,3, J. D. Pickard^1,2 & P. J. Hutchinson^1,2 (Academic Neurosurgery Division¹ and Wolfson Brain Imaging Centre², Department of Clinical Neurosciences, and Division of Anaesthesia³, University of Cambridge, Addenbrooke's Hospital, Cambridge)

Introduction: Microdialysis continuously samples the chemistry of a small focal volume of the cerebral extracellular space. This study's objective was to assess the feasibility of administering a neuroprotective drug, vigabatrin (VGB; gamma-vinyl-gamma-aminobutyric acid) in conjunction with multimodality monitoring, including microdialysis, to patients with severe head injury, to determine the effect of VGB on surrogate endpoints and penetration across the blood-brain barrier (BBB). The anticonvulsant VGB is an irreversible inhibitor of gamma-aminobutyric acid-transaminase, the enzyme responsible for the degradation of the neurotransmitter gamma-aminobutyric acid (GABA).

Methods: Head-injury patients were recruited and randomised to VGB (0.5 g twice daily, orally; 10 patients) or control (no VGB; 10 patients). Microdialysis catheters (CMA71) were inserted into the cerebral cortex either via a cranial access device or placed under direct vision following craniotomy. Catheters were perfused at 0.3 microlitre/min with CNS perfusion fluid (CMA) and microdialysate collection vials were changed hourly. Microdialysate analysis for glucose, lactate, pyruvate, glutamate and glycerol was performed on a CMA600 analyser using enzymatic colorimetric assays. Microdialysate analysis for amino acids, including VGB and GABA, was performed on orthophthalaldhyde-derivatised samples by reversed-phase HPLC (Agilent 1100) with fluorescence detection. Plasma samples were also analysed similarly by HPLC.

Results: A preliminary evaluation of 5 patients who received VGB showed that VGB could be detected in their brain microdialysates, and that the rise in VGB was followed by a modest increase in GABA. VGB and GABA increases were more pronounced in microdialysates from an abnormal region of brain, as opposed to sites not close to focal lesions. VGB and GABA increases were more pronounced after two or more VGB doses than after one. The highest VGB and GABA levels encountered in microdialysates were 75 and 4 micromol/litre respectively. In some cases, decreases in microdialysate glucose and glycerol, and increases in glutamate and tyrosine followed VBG administration, although causation is unproven. VGB showed no overt effect on microdialysate levels of lactate, pyruvate or the lactate/pyruvate ratio in these patients.

Conclusion: These preliminary results suggest that VGB crosses the BBB, leading to modest increases in GABA levels in brain microdialysates. The analysis of further patients' samples is ongoing. Microdialysis has the potential to assist in the development of putative neuroprotective agents by determining penetration into the brain extracellular fluid.

Aquaporin-4 deletion in mice greatly improves outcome after spinal cord injury

M. C. Papadopoulos, S. Saadoun & B. A. Bell (Academic Neurosurgery Unit, St. George's University of London)

Introduction: Aquaporin-4 (AQP4) is a water channel protein expressed in astrocytes throughout the central nervous system. In brain, AQP4 facilitates water balance and glial scar formation, which are important determinants of outcome after injury [1]. Recent experiments showed increased AQP4 expression in injured spinal cord [2], but it is not known whether this is a beneficial or an adverse effect. Here, we provide evidence for AQP4-dependent spinal cord swelling following compression injury, resulting in remarkably improved outcome in AQP4-null mice.

Methods/Results: Two days after transient T6 spinal cord compression injury, wildtype mice developed more severe hind limb weakness than AQP4null mice, as assayed by the Basso open-field motor score, inclined plane method and hind paw print analysis. Basso motor scores were 1.3 ± 0.5 (wildtype) vs. 4.9 ± 0.6 (AQP4-null) (S.E., p < 0.001). Improved motor outcome in AQP4-null mice was independent of mouse strain and persisted for at least three weeks. AQP4-null mice also had improved sensory outcome at 2 d, as assessed by spinal somatosensory evoked responses, with signal amplitudes ∼10 microV (uninjured), 1.7 ± 0.7 microV (wildtype) and 6.4 ± 1.3 microV (AQP4-null) (p < 0.01). The improved motor and sensory indices in AQP4-null mice corresponded to remarkably less neuronal death and myelin vacuolation, as well as reduced spinal cord swelling and intraparenchymal spinal cord pressure measured at T6 at 2 d after injury.

Conclusions: Together our findings indicate that AQP4 provides a major route for water entry into the injured spinal cord, which is responsible for spinal cord swelling and elevated intraparenchymal spinal cord pressure. Our data support AQP4 inhibition as a novel neuroprotective treatment in spinal cord injury.

The relationship between cortical blood flow and spreading depolarisations in brain injured patients as measured with laser Doppler flowometry and Electrocorticography

A. D. Manning¹, A. J. Strong¹, J. P. Dreier², J. Hartings³, M. Fabricius⁴, R. Graf⁵ & M. Boutelle⁶ (Department of Neurosurger, Kings College Hospital, London,UK¹ Department of Neurology, Charité Campus Mitte, Berlin, Germany² Division of Psychiatry and Neuroscience, Walter Reed Army Institute of Research, Richmond, VA, USA³ Department of Clinical Neurophysiology, Glostrup University Hospital, Copenhagen, Denmark⁴ Max Planck Institute, Cologne, Germany⁵ Department of Bioengineering, Institute of Biomedical Engineering, Imperial College London, UK⁶)

Introduction: Two patterns of transient depression of electrocortical activity spreading across the penumbral cortex in patients with acute brain injury have been reliably detected in both animal & human studies. Peri-Infarct Depolarisations (PIDs) propagate at a similar speed to Cortical Spreading Depression (CSD), 3–5 mm/min, and have been associated with increases in infarct volume [1]. The characteristics that can be used to differentiate CSDs from the more harmful PIDs are currently under investigation and can be assessed by their electrical or vascular signatures. The relationship between PIDs and loss of perfusion has been shown in cats to cause secondary deterioration in the penumbra by propagation of ischaemia [2].

Methods: In the following case studies, the dynamic electrophysiological environment in the penumbra of two brain injured patients was monitored using electrocorticography (ECoG) and laser Doppler flowometry (LDF). A six contact subdural electrode strip with four laser Doppler fibre-optic probes built into the strip at 1 cm intervals was placed at craniotomy onto the penumbral cortical surface. Intracranial pressure (ICP) and microdialysis (MD) probes were also placed into the same region of cortex. Monitoring continued for 3–5 days on the intensive care unit before the probes were removed. In both cases, slow potential (DC) changes, characteristic of spreading depolarisations (SDs-incorporating CSD & PID) were seen. Cortical blood flow changes during these events were measured by LDF.

Results/Conclusions: Throughout the recordings in both patients, spreading depolarisations were accompanied by marked changes in LDF. The precise vascular response (hypo or hyper- perfusion) to SDs varied depending on the relative severity of compromised cortical tissue. In each event, the cortical perfusion changes were always preceded by the SD seen on ECoG curve. This ‘coupling’ between electrical event and vascular response appears to confirm what has already been seen in animal models and could provide further evidence to support this mechanism of secondary brain injury.

Detecting evidence of spreading depolarisations using microdialysis: on-line monitoring of extracellular potassium

R. Bhatia¹, P. Hashemi², H. Nakamura³, R. Graf³, M. G. Boutelle² & A. J. Strong¹ (Department of Clinical Neurosciences, Kings College, London, UK¹ Department of Bioengineering, Imperial College, London, UK² The Max Planck Institute, Cologne, Germany³)

Introduction: Waves of spontaneous depolarisation spreading across the cerebral cortex (SD) occur in patients with acute traumatic or ischaemic brain injury [1],[2]. During clinical monitoring with rapid-sampling microdialysis, a marker of depolarisation in the dialysate stream is required as verification of the depolarisation event affecting the volume of brain being sampled. SD is invariably associated with a transient increase in extracellular potassium: we therefore developed a flow-through sensor to monitor dialysate potassium.

Methods: A valinomycin-based potassium electrode was built and incorporated into a perspex cell (internal volume: 0.7 L). In vitro testing examined selectivity, stability and flowcell responses to transient and sustained K+ changes and in vitro recovery. In vivo work examined dialysate K+ changes associated with probe insertion, terminal ischaemia and SD events in healthy, or marginally ischaemic cerebral feline cortex after middle cerebral artery occlusion (MCAO). Cortical events were verified electrophysiologically. The dialysate stream was then passed into a rapid-sampling glucose and lactate assay, allowing serial online ionic and metabolic monitoring.

Results: The flowcell displayed adequate selectivity in the required physiologic range, responsiveness to flowstream K+ changes (t90% = 50 ± 5s), and stability over 70 hrs of continuous use (potential drift = 0.1 mV/hr). In vitro relative recovery of potassium was 65%. Insertional microtrauma was associated with a stereotypical and transient rise in potassium (5 mM, range: 3.5-8 mM, n = 5), followed by a fall phase characterised as a clearance rate, which was significantly reduced in post-MCAO cortex compared to healthy cortex (p = 0.01). Terminal ischaemia was associated with a delay of 3.0 ± 1.5 min before a rise in dialysate K+ (6.5 ± 0.5 mM, n = 6). SD events were accompanied by transient changes in dialysate potassium. Changes in dialysate glucose and lactate were described in conjunction with K+ associated with CSD.

Conclusion: The potassium flowcell is sufficiently selective, responsive and stable to monitor dialysate changes associated with probe insertion, terminal ischaemia and SD. We have also described, for the first time, online analysis of ionic and metabolic markers in microdialysate associated with SD in vivo. Further clinical development work is now required.

ECF GFAP: a candidate biomarker for cerebral vasospasm?

A. G. Kolias¹, A. Petzold¹, J. Sen¹, S. Sen¹, J. Hillman², G. Keir¹ & A. Belli³ (Dept. of Neuroimmunology, Institute of Neurology, University College London, UK¹, Dept. of Neurosurgery, University Hospital in Linköping, Sweden², Division of Clinical Neurosciences, University of Southampton, UK³)

Introduction: Cerebral vasospasm is the leading cause of poor outcome and death in patients who have survived SAH. Our hypothesis is that the release of astrogial protein GFAP (41 kDa) into brain extracellular fluid (ECF), could correlate with or even precede clinical manifestations of vasospasm. We chose to focus on GFAP, as the astrocytes are particularly vulnerable to ischaemia, play a critical role in the regulation of the brain microenvironment, and are immediately activated after the initial ictus.

Methods: This longitudinal study involved 15 patients (mean age 49, range 29–66) admitted due to SAH to NICU of the University Hospital in Linköping. RLS-85 (Swedish 8-point coma scale; 1 is best and 8 the worst possible score), and Fisher grade were recorded upon admission. Median RLS-85 was 3 (patient very drowsy or confused; responsive only to strong stimulation). 12 patients had a Fisher grade of 3, 2 of 4, and 1 of 2. TCD was performed at least once per day. 100 kDa molecular weight cut-off microdialysis catheters (CMA Microdialysis, Solna, Sweden) were used. GFAP was measured according to a sandwich ELISA technique [1].

Results: Median ECF GFAP levels were 138.81 ng/ml (Q1-Q3: 30.13-319.42 ng/ml). ECF GFAP was found to correlate with mean (p < 0.01, r = 0.327, n = 99) and maximum TCD flow velocities (p < 0.001, r = 0.421, n = 99). Moreover, mean (p = 0.046, r = 0.521, n = 15) and maximum (p = 0.012, r = 0.626, n = 15) ECF GFAP per patient correlated with Fisher grade on initial CT scan.

Discussion: In the current study, ECF GFAP was evaluated as novel biomarker after SAH. ECF GFAP has been quantified before [2], but the microdialysis methodology used is not clear. CSF GFAP has been previously shown to be able to predict secondary brain injury after SAH [1]. To our knowledge, this is the first demonstration of GFAP quantification in brain ECF samples obtained with commercially available 100 kDa cut-off catheters. Our results suggest that ECF GFAP is a candidate biomarker for cerebral vasospasm. Further studies are warranted in order to appraise the value of ECF GFAP in a clinical setting.

Hyperosmolarity induced by 23.5% hypertonic saline therapy is associated with stimulation of hypothalamus-pituitary axis in patients with poor-grade aneurysmal subarachnoid haemorrhage

M. Y. Tseng, P. J. Hutchinson & P. J. Kirkpatrick (Dept Neurosurgery, Addenbrooke's Hospital, Cambridge)

Introduction: Integrity of the hypothalamus-pituitary axis is associated with favourable long-term outcome after aneurysmal subarachnoid haemorrhage (aSAH). However, tests of the integrity of the axis by osmotic stimulation in these patients have not been reported.

Methods: Thirty-five poor-grade aSAH patients received intravenous infusions of 23.5% hypertonic saline, 2 mL/Kg. Full blood cell counts, coagulation profiles, and serum biochemistry were checked before and at first hour and then every 6 hours, for a total of 24 hours, after the infusion. Serum levels of adrenocortical tropic hormone (ACTH) and cortisol were checked before and at 1, 6 and 24 hours after the infusion. Intravenous physiological saline (0.9%) was given to keep the maximum rate of reduction in serum sodium <10 mmol/L per day in order to avoid potential central pontine myelinolysis.

Results: The average interval between aSAH and the treatment episodes was 4.6+/−3.2 days. Among the 15 patients who had baseline cortisol samples, 5 had levels below normal (33.33%). Among these 15 patients, 13 of them had baseline ACTH samples and 4 of them had levels below normal (30.77%). Changes in serum cortisol were not significant, but there were trends in increasing serum ACTH at 1 (+56.59%, p = 0.151) and 6 hours (+165.38%, p = 0.101) after the infusion, which correlated with serum osmolarity (at 24 hour, r = 0.782, p = 0.008). Favourable outcome at discharge measured with modified Rankin Scale (1–3) seemed to be associated with a higher cortisol level before the infusion (p = 0.093) and higher serum osmolarity at 24 hours (p = 0.130).

Conclusions: It seemed that a significant proportion of poor-grade aSAH patients had subnormal adrenal function during the acute phase. Hyperosmolality induced by hypertonic saline therapy appeared to be able to stimulate the hypothalamus-pituitary axis in patients with poor-grade aSAH. The association between a higher cortisol level before the hypertonic saline therapy and a favourable outcome at discharge implies that additional supplement of hydrocortisone may have positive effects on osmotherapy or may be able to compensate the abnormal hypothalamus-pituitary-adrenal axis.

Simvastatin in aneurysmal subarachnoid haemorrhage (STASH); a phase 111 randomised placebo controlled trial

C. L. Turner¹, M.-Y. Tseng¹, A. D. Mendelow², P. J. A. Hutchinson¹, G. Murray³, M. Brown⁴ & P. J. Kirkpatrick¹ (Academic Dept of Neurosurgery, Addenbrooke's Hospital, Cambridge¹, Dept of Neurosurgery, Newcastle General Hospital, Newcastle upon Tyne², Division of Community Health Sciences, University of Edinburgh³, Pharmacology Unit, Addenbrooke's Hospital, Cambridge⁴)

Introduction: A pilot study demonstrated that acute statin therapy following aneurysmal subarachnoid haemorrhage ameliorates vasospasm-related delayed ischaemic deficits, reduces frequency and intensity for rescue therapy and improves long-term outcome at 6 months [1],[2]. We are conducting a phase III multi-centre randomised-controlled trial to establish if this holds true in a larger population.

Methods: 1600 aneurysmal subarachnoid haemorrhage patients (age 18–65 years, onset < = 96 hours) will be randomised to receive daily oral Simvastatin 40mg or placebo for up to 21 days. Adverse clinical events will be recorded during the trial. Primary outcome measure is the modified Rankin Disability Score at 6 months. Secondary outcome measures are the Short Form 36 (SF-36 questionnaire) at 6 months, the incidence, duration and need for delayed ischaemic deficit (DID) rescue therapy, incidence and severity of sepsis, length of intensive care and total acute hospital stay and percentage of patients discharged directly home.

Results: 130 patients (44 male, 86 female), have been recruited from 11 centres worldwide in 10 months, (mean age 50 years, range 21 –65 years). 102 patients were WFNS grades 1–2 on arrival, 28 were grades 3–5. 49 patients were admitted to the intensive care unit, mean length of stay 12 days (range 2 – 42 days). 7 patients have died, 2 from rebleeds, 3 from DIDs, and 2 from DIDs with sepsis. There have been no unexpected adverse events. We have 6 month outcome data on 50 patients (98%).

Conclusions: The trial has recently received favourable response from the British Heart Foundation which will allow recruitment to accelerate. Both UK and overseas Centres will be welcomed to participate and will receive appropriate funding support.

Development of a computational model to determine the risk of aneurysm rupture

Alberto Marzo², Alan Waterworth¹, Rod Hose², Alejandro Frangi³, Pat Lawford², Keith McCormack², Stuart Coley¹ & Umang Patel (Royal Hallamshire Hospital, Sheffield, UK¹, University of Sheffield, Sheffield, UK², Universitat Pompeu Fabra, Barcelona, Spain³)

Introduction: @neurIST is a major multidisciplinary European initiative that brings together 32 European institutions, to improve the diagnosis and treatment of cerebral aneurysm and subarachnoid haemorrhage. The project is developing an IT infrastructure and computational tools to assist the collection, processing and integration of data from all data sources on cerebral aneurysms. The aim is to develop a usable interface for personalised risk assessment and treatment of patients affected by the disease. The University of Sheffield and the Sheffield Teaching Hospitals NHS Trust have a major role in the development of image processing and computational tools to provide new, non-observational measures that in the literature have been associated with aneurysmal growth or risk of rupture [1].

Methods and Results: Using input parameters such as neuroradiological studies (e.g. CTA), blood density, haematocrit, and heart rate, the team in Sheffield uses advanced computational models to extract haemodynamic and structural data. Examples of these measures include wall shear stress (i.e. viscous shear stress exerted by blood and experienced by endothelial cells) and strain of the aneurysmal wall. Fig. 1 and 2 illustrate some of the variables computed by these tools for a PComA aneurysm. The epidemiologists in @neurIST will then use these data as the basis for statistical association studies.

Conclusions: The goal of @neurIST is for all data, computed and data-mined, to be integrated and made available to medical practitioners through an intuitive interface in the clinic. The computational tools being developed within @neurIST will permit additional measures of haemodynamic and structural data of patient-specific cerebral aneurysms to be obtained without additional intervention. These tools will be applied to several hundred real aneurysms, computing in each case a comprehensive set of haemodynamic and structural variables. The epidemiologists in @neurIST will use advanced statistical tools to find an association between these data and risk of evolution and rupture of the processed cases. Its impact, in both health and economic terms, is likely to be profound.

Modelling turbulence using computational fluid dynamics in models of a partially occluded large basilar tip aneurysm

M. Foroughi, J. Bannister, D. M. O'Doherty, T. J. O'Doherty, R. Nannapaneni & R. H. Hatfield (University Hospital of Wales Cardiff & Cardiff School of Engineering)

Introduction: The haemodynamics of aneurysms has a governing effect on the rupture or growth, with high levels of turbulence being shown to increase chances of aneurysm growth and rupture [1]. To identify and analyse turbulence and blood flow regimes induced within a partially occluded large basilar tip aneurysm.

Methods: Computational fluid dynamics (CFD) modelling was carried out taking 3D co-ordinates from DICOM images. Various degrees of occlusion of the aneurysm were simulated by removing the appropriate percentage of the upper part of the modelled aneurysm. The CFD modelling was done using Fluent 6.2, and the computation was run using unsteady inlet flow with a standard k-ϵ turbulence model. Assumptions made were that the arterial walls are solid, non-elastic entities and that blood is an incompressible Newtonian fluid. To highlight the flow regimes, path line plots were constructed which track virtual particles released into the flow. An anonymous DICOM image set was provided of a large basilar tip aneurysm. A 2D model was used to test for the most suitable turbulence model in this application. An untreated aneurysm was modelled in 3D to assess that the modelling process was not unreasonable.

Results: The 70% treated model showed re-circulation similar to that in the untreated model. In the 80 and 90% treated models, the re-circulation was reduced; hence turbulent effects in the remnant were significantly lowered. In the 95% treated model the blood flow direction changed rapidly and induced higher levels of turbulence within the aneurysm.

Conclusions: This model would suggest that in a large basilar tip aneurysm a partial embolisation/occlusion of 80%−90% would results in less turbulence and therefore possibly less risk of growth than a greater level of occlusion.

Aquaporin-4 deletion in mice increases brain oedema after subarachnoid haemorrhage

M. J. Tait, B. A. Bell & M. C. Papadopoulos (Academic Neurosurgery Unit, St George's University of London)

Introduction: Approximately 8% of patients with subarachnoid haemorrhage (SAH) have global cerebral oedema at presentation. This is associated with poor clinical grades of SAH and is an independent negative prognostic factor [1]. The molecular mechanisms of SAH-associated brain oedema are not understood. We developed a new model of mouse SAH to investigate the role of the astrocyte water channel protein aquaporin-4 (AQP4) in SAH- induced brain oedema.

Methods: Autologous whole blood (30 HL) was stereotactically infused into the basal cisterns of CD1 male mice. A comparison was made between wild type (WT) and AQP4 knockout (KO) mice at 6 and 24 hours after surgery. Outcome measures include ICP, brain water content and a neurological outcome score (3–27 scale) as well as Evans blue dye extravasation (an indicator of blood-brain barrier disruption, the hallmark of vasogenic brain oedema). Data are mean + SEM.

Results: WT and KO mice had comparable baseline brain water contents, ICPs, neurological scores, and Evans blue dye extravasation, as well as comparable increases in ICP during subarachnoid blood injection. After SAH, KO mice developed more marked brain swelling and worse neurological score than WT mice. At 6 hours after SAH haemorrhage, brain water content was 79.3 + 0.13 vs. 78.3 + 0.18% (p = 0.0003), ICP was 42.4 +4.74 vs. 28.8 + 2.7 mmHg (p = 0.023) and neurological score was 18.1 + 1.4 vs. 22.8 + 1.4 in KO vs. WT mice. At 24 hours after SAH, brain water content was 79.1 + 0.2 vs. 77.9%+ 0.2% (p = 0.01) and ICP was 38.9 + 2.9 vs. 26.8 + 2.5 mmHg (p = 0.026), and neurological score was 17.9 + 1.9 vs. 24.4 + 0.9 in KO vs. WT mice. Compared with baseline values, there was markedly increased Evans blue dye extravasation at 24 hours in KO and WT mice. Histological examination revealed no hydrocephalus and no vasospasm.

Conclusions: We conclude that AQP4 deletion increases vasogenic brain swelling in mice after SAH and therefore AQP4 activators or upregulators may improve clinical outcome.

Electrical impedance tomography of acute stroke in the anaesthetized rat

B. Sokol ¹, A. Romsauerova², L. Fabrizi², A. McEwan², C. Yeo³ & D. Holder² (Hurstwood Park Neurological Centre, Haywards Heath¹, Department of Medical Physics, University College London, London², Department of Anatomy, University College London, London³)

Introduction: Electrical Impedance Tomography is a recently developed medical imaging technique with which tomographic images of the electrical impedance of a subject may be made with a system similar in size to an EEG machine and ECG type electrodes. It is fast, safe, portable and low cost. It would provide an invaluable rapid and low-cost method for imaging in acute stroke and so permitting rapid employment of thrombolysis. Preliminary studies in humans had revealed a poor image quality; the purpose of this work was to refine the method in anaesthetized rats.

Methods: A 3 female rats (Sprague-Dawley) were anaesthetized with halothane. 24 miniature screw electrodes were inserted into the skull so their tips lay in the subdural space. The UCH Mk 2.5 multifrequency EIT system, with applied current between 100 Hz and 1.6 MHz at 66 uA was employed to collect EIT data. After obtaining a reference data set at rest, the medial cerebral artery was surgically exposed and coagulated to produce local ischemia. Further measurements were obtained after the death of the animal, in order to record impedance changes occurring in the brain similar to the global ischemia.

Results/Conclusions: Analysis of the raw impedance data indicated a significant increase at low frequencies after stroke and a greater increase after death. Work is in progress to reconstruct this data into EIT images. If the recordings with subdural electrodes are successful, further recording will be undertaken with scalp electrodes to determine how accurate the method could be with non-invasive recording.

Selection of patients for STA – MCA bypass using CT perfusion scanning with acetazolomide challenge to establish cerebral hypoperfusion

A. A. Mazumder & C. Tolias (Department of Neurosurgery, King's College Hospital, London, UK)

Introduction: The superficial temporal to middle cerebral artery (STA-MCA) bypass aims to reduce the risk of stroke by increasing intracerebral blood flow. There is no current established method for selecting these patients. The current standards to assess cerebral blood flow include positron emission tomography and xenon-enhanced CT. However these modalities can be difficult to access. One alternative selection method for patients involves the dynamic CT perfusion (CTP) imaging, augmented by acetazolomide administration to test cerebrovascular reserve. We report a prospective case series of five patients with recurrent ischaemic symptoms who were assessed with CTP ± acetazolomide to assess cerebrovascular reserve.

Methods: Seven patients were referred between June 2006 and October 2007 with recurrent (5) or persistent disabling (2) neurological symptoms of unilateral cerebral ischaemia. This was either confirmed with diffusion-weighted MRI or CT imaging. All patients were assessed further using CTP scanning before and after administration of acetazolomide (1000 mg). A neuro-radiologist familiar with the technique subsequently reported the scans.

Results: Seven patients demonstrated significant unilateral reduction in CTP values. After administration of acetazolomide, two patients demonstrated no change in cerebral blood flow. Three patients in the symptomatic hemisphere had decreases of 15%, 24%, and 18% comparing scan pre- and post- vasodilator administration. Four patients underwent STA-MCA bypass and one had multiple burr holes. Follow-up perfusion imaging in three patients (up to the time of the report) demonstrated significant CTP values improvement.

Conclusion: CTP may offer a simpler alternative to the selection process of patients with unilateral cerebral ischaemia. A prospective study is warranted.

Neurotransmitter regulation of adult neural stem cells

M. Zaben¹ & W. P. Gray^1,2 (Division of Clinical Neuroscience, University of Southampton¹, Wessex Neurological Centre²)

Introduction: There is increasing evidence that neuronal activity regulates the generation of new neurons in the adult hippocampus, a phenomenon that appears important for certain forms of hippocampal learning and the maintenance of normal mood. Altered hippocampal neurogenesis after brain injury may explain cognitive and mood disorders seen in patients after head trauma. The mechanisms by which neuronal activity controls neurogenesis are largely unknown. We have previously shown that a peptide neurotransmitter (NPY) released by interneurons is proliferative for hippocampal stem cells. Currently, we are investigating how VIP, another peptide neurotransmitter, modulates hippocampal neurogenesis alone and in combination with other neurotransmitters.

Methods: We have investigated the effect of VIP alone and in conjunction with NPY in hippocampal neuronal cultures from postnatal rats (P7–9). Bromodeoxyuridine (BrdU) and Ki-67 were used to measure cell proliferation. Quantification of cell death was achieved using DAPI and propidium iodide (PI). Immunohistochemistry was used to determine cell-specific phenotypes. PCR assay was carried out to study the expression of VIP receptors. In-vivo studies were performed in germ-line VPAC2 knockout mice.

Results: Using Immunohistochemistry and PCR techniques, we have demonstrated the expression of VIP receptors and their mRNAs by hippocampal cells. We have also shown that VIP at nanomolar concentrations is trophic to hippocampal progenitor cells and their progeny. Our data suggest that the effect is mainly through enhancement of symmetrical cell division. Pharmacologically, our results indicate that VIP effects are mediated by the VPAC2 receptor. We confirm the effect of VIP in-vivo, by demonstrating significantly reduced survival of newly born neurons in germline VPAC2 receptor knockout mice. Interestingly, when cells were exposed to VIP in conjunction with NPY, VIP abolished the NPY-proliferative effect, but enhanced neurogenesis.

Conclusions: We conclude that VIP by itself and through interaction with NPY is an important trophic factor for hippocampal neurogenesis. We suggest that these neuropeptides provide a novel control system for hippocampal neurogenesis depending on their differential release. This control system may provide therapeutic targets for altered neurogenesis after brain injury.

Pharmacometric analysis informing the design of neurosurgical clinical trials – a case in point

J. Galea^1,3, K. Ogunbenro², L. Aarons², B. Houston², S. Hopkins³ & N. Rothwell¹ (Faculty of Life Sciences, Michael Smith Building, University of Manchester, Manchester¹, School of Pharmacy and Pharmaceutical Studies, University of Manchester, Manchester², Brain Injury Research Group, Salford Royal NHS Foundation Trust, Salford³)

Introduction: Pharmacometrics involves the analysis and interpretation of data produced in pre-clinical and clinical trials. It is being increasingly used to inform the design of protocols and studies relating to investigational medicinal products. Its use in neurosurgical trials is especially important given the relative small number of patients eligible and consenting for participation in clinical trials.

Methods: In designing a dose-ranging study of interleukin-1 receptor antagonist (IL-1RA) given intravenously in patients with subarachnoid haemorrhage, data from a previous pilot study done by our group were used to model the pharmacokinetics of the drug in a three-compartment model. Using commercially available software (NonMem, Mathlab) or freeware (R), it was possible to simulate expected concentrations of the drug in blood and cerebrospinal fluid (CSF) using these modelling data. This allowed optimisation of the design of the dose-ranging study by reducing sample size, minimising the sampling points and optimising the sampling times needed. Prediction interval planning was implemented as a measure to check whether the concentration curves of any of the dose regimes would fit the predicted model.

Results/Conclusions: The resulting better informed trial design has translated into clinical benefit for patients participating in the study by reducing sampling times and allowing for a predictable stepwise increase in plasma and CSF concentrations. Pharmacometric analysis should be strongly considered in assisting and informing the design of clinical drug trials.

Methodological considerations for the use of high molecular weight cut-off microdialysis membranes for the recovery of cytokines

A. Helmy, K. H. L. Carpenter & P. J. A. Hutchinson (Academic Unit of Neurosurgery, University of Cambridge)

Introduction: Microdialysis is an established technique for the recovery of molecules of interest following a range of insults such as traumatic brain injury. There is increasing interest in the use of high molecular weight cut-off microdialysis membranes (e.g. 100 kDa) for recovering other molecules of interest such as cytokines. Although cytokines have been successfully recovered from the human brain using this technique there are many considerations specific to higher molecular weight molecules that impact on the efficiency of recovery by microdialysis. These include, low concentrations of cytokine, small volumes recovered by microdialysis due to loss of fluid across the high molecular weight cut-off membrane and drop off in the relative recovery over time due to encrustation of the microdialysis membrane. We sought to investigate these methodological constraints and, in particular, investigate the effects of colloids such as albumin in the perfusing solution.

Methods/Results: Initially we sought to identify the percentage of fluid recovered by the microdialysis catheter over a range of albumin concentrations in the perfusing fluid. A 0.1% solution of Human Albumin Solution (Sigma) was used as the external solution. Four concentrations of Human Albumin Solution (0.1%, 1.0%, 3.5%, 5%) made up in CNS perfusion fluid were used across multiple time points. The fluid recovery across the range of perfusing solutions was calculated by weighing the microdialysis vials before and after the experiment. The fluid recovery was calculated as 85% (0.1% HAS), 172% (1.0% HAS), 176% (3.5% HAS), 183% (5% HAS). Many authors argue that addition of colloid to the perfusing solution is required to avoid fluid loss across the microdialysis membrane. This experiment would suggest that although this is true, addition of high concentrations of colloid may have the opposite effect and dehydrate the extracellular space.

We have also explored the effect of albumin in perfusate on the relative recovery of a range of cytokines. We have used 1% albumin in the perfusing solution and calculated the relative recovery of IL-1beta, IL2, IL4, IL6, IL17, TNFalpha, MIP-1beta. The addition of albumin to the perfusate improved relative recovery by 2-3 fold though recovery was still reliable without albumin. No drop off in recovery was found over a 48-hour period.

Conclusions: These results suggest that cytokines can recovered reliably with existing microdialysis methods.

On custom development of neurosurgery software for multiple platforms without the help of IT professionals

G. Narenthiran (Department of Neurosurgery, Wessex Neurological Centre, Southampton General Hospital, Southampton)

Introduction: Neurosurgeons regularly use computers in the course of their work. This may be in clinical settings or for purely administrative purposes. Software used by Neurosurgeons are usually complex, expensive and are created by IT professionals. However, recent advances in software development technology are making development of software for multiple platforms easier. Our aim was to custom develop neurosurgery software that could run on hand-held computers, desktop computers and over network or the internet, without the help of IT professionals.

Methods: We used Microsoft Visual Basic 2005 running on a Windows laptop computer (Sony VGN-B1VP; Intel Pentium M725 1.60 GHz; 40 GB hard drive; 1024 MB RAM; Microsoft Windows XP) and Macromedia Flash 8 running on a Power Mac (Intel Pentium Duo-Core Processor 2.33 GHz; 1 Gb RAM; 180 GB hard drive; Mac Os 10.4 (Tiger)). The hand-held computer we used was HPiPAQ running Windows Mobile 2005. From a commercial internet service provider we had subscribed memory for web-hosting (Unix based) and database (Microsoft-SQL). At no time was help from IT professionals obtained.

Results: We successfully developed the software for multiple platforms. For hand-held computer: ‘Electronic Visual Analogue Scale (eVAS)’– it has an electronic slider, the patient can move with a stylet on the screen and the result can be read with the click of a button. For desktop computers (both PC and Mac): electronic interactive versions of the original and the modified Japanese Orthopaedic Association scale. For network: electronic calculator to compute the American Spinal Injury Association (ASIA) score. For web: on-line database for abstracts (abstract archive); AVM calculator to estimate the risk of haemorrhage, morbidity and mortality from arteriovenous malformations.

Conclusion: Recent advances in technology have now enabled Neurosurgeons to customs develop simple and useful software for multiple computer platforms. Developing software is creative and enjoyable. Besides being useful in day-to-day practice, such developed software could have commercial potential.

The effect of the STICH trial on neurosurgical treatment of intracerebral haemorrhage

L. Chilton, B. Gregson & A. D. Mendelow(Department of Neurosurgery, University of Newcastle upon Tyne, Newcastle General Hospital, Newcastle upon Tyne)

Introduction: The surgical trial into intracerebral haemorrhage (STICH) was the first large-scale international multicentre randomised controlled trial comparing a policy of early surgery with that of initial conservative treatment for intracerebral haemorrhage. The trial recruited 1033 participants and was reported in the Lancet [1] in 2005. The overall finding of the STICH Trial was neutral. Following this publication various papers have made recommendations for the treatment of intracerebral haemorrhage, several stating that there is no role for surgery. The aim of this study is to investigate the effects of the STICH Trial data and how it has been interpreted by the neurosurgical community in order to put the STICH II trial in context.

Methods: This study used a written questionnaire which was sent to neurosurgeons at all centres who have expressed an interest in STICH II.

Results: To-date 120 questionnaires have been sent and 30 returned. 45% of these respondents have changed their surgical management of intracerebral haemorrhage patients as a result of the STICH Trial data. These changes are related to both the physiology of the haematoma itself and its location as well as the symptomology of the patients at presentation. The full dataset will be presented at the meeting. Comparisons will also be made with similar data collected prior to the Lancet publication.

Conclusion: The meta-analysis of the STICH data has highlighted the need for further study and that in certain subgroups there may be distinct differences between the two possible treatments. It is therefore advisable that neurosurgeons participate in the STICH II study and await the results of this trial before opting not to surgically treat all intracerebral haemorrhages.

Surgical trial in lobar intracerebral haemorrhage (STICH II) – an update

B. A. Gregson¹, L. Chilton¹, P. Mitchell² & A. D. Mendelow¹ (Newcastle University¹ and Newcastle General Hospital², Newcastle upon Tyne)

Introduction: Spontaneous superficial intracerebral haemorrhage (ICH) accounts for 20% of all stroke-related sudden neurological deficits and has the highest morbidity and mortality of all strokes. The role of surgery remains controversial following the report of the international STICH trial which was neutral. Further analyses and meta-analysis have suggested that patients with lobar haemorrhage may benefit from early surgery. STICH II aims to establish whether a policy of earlier surgical evacuation in selected patients with spontaneous lobar ICH will improve outcome compared to a policy of initial conservative treatment.

Methods: STICH II is an international multicentre randomised parallel group trial funded by the MRC (ISRCTN22153967) with a sample size of 600. Patients for whom the treating neurosurgeon is in equipoise about the benefits and risks of early craniotomy are eligible for the trial. Inclusion criteria include superficial, spontaneous lobar ICH on CT scan, best motor score on the Glasgow Coma Score (GCS) of 5 or 6, best eye score on the GCS of 2 or more, and volume of haematoma between 10 and 100m. Patients are ineligible if the haemorrhage is due to an aneurysm or angiographically proven arteriovenous malformation, is secondary to tumour or trauma or extends into the basal ganglia, thalamic, cerebellar or brain stem. Patients are also ineligible if there is any intraventricular haemorrhage or hydrocephalus or if there is severe, pre-existing physical or mental disability or severe co-morbidity which might interfere with assessment of outcome. The trial intervention is early evacuation of the haematoma within 12 hours of randomisation combined with appropriate best medical treatment versus best medical treatment, combined with delayed evacuation only if it becomes necessary later.

Outcome is measured at six months via a structured postal questionnaire including the Glasgow Outcome scale, Modified Rankin Scale and EuroQol and analysis will be on an “intention to treat” basis.

Results/Conclusion: Regulatory approvals have been obtained in the UK and are being sought in other countries around the world. Centre and patient recruitment are ongoing. Current details at 31 January 2008 of recruiting centres and recruited patients will be presented.

Further information available at: www.ncl.ac.uk/stich