Traumatic Brain Injury: Management on the Neurointensive Care Unit
Clemens Pahl FRCA DICM
Consultant Intensivist King’s College Hospital
Focus on neurointensive care management
There are many components to the modern day neurointensive care unit (NICU) management of traumatic brain injury (TBI). Important aspects of NICU management include:
• Monitoring techniques
• intracranial pressure (ICP)-targeted therapies
• CPP targeted therapies
• Specific medical therapies
• Neurosurgical interventions
• Management of recognised complications
• General multi-organ supportive measures
Figure 1: A simplified, protocol-based approach to the management of intracranial hypertension.
Patient at risk of intracranial hypertension
1 Evacuate space occupying lesion
15-30ο head up
Ensure no venous obstruction
Ensure core temperature ≤37C
Initial sedation: Propofol and fentanyl
Consider phenytoin, especially if depressed skull fracture, witnessed seizure)
PaO2 ≥ 11 kPa
2 If ICP >20 mmHg:
? Aims of Box 1 achieved
Consider CT imaging to exclude new/expanding space occupying lesion
3 If ICP still >20 mmHg:
Sedation bolus (propofol)
20% mannitol 2ml/kg, repeated if plasma osmolality <320 mosmol/kg
4 If ICP still > 20mmHg:
Check ICP probe
Consider repeat CT imaging
Ensure 450 sitting up if CPP maintained
Consider following options:
Moderate targeted hypothermia (≤33ο C)
Focus on standard monitoring
On admission to the NICU, all patients with severe TBI are at risk of developing raised ICP, and must have standard systemic monitoring established: pulse oximetry, invasive arterial blood pressure with regular analyses of arterial blood gases and blood glucose and central venous access with central venous pressure monitoring. End-tidal carbon dioxide monitoring is invaluable in this group of patients because it enables early correction of hypercapnia-induced rises in ICP.
Focus on monitoring of ICP
Targeting ICP and CPP therapeutically mandates monitoring of ICP in patients with severe head injury (Glasgow Coma Scale [GCS] ≤8 and abnormal CT scan), but the precise indications vary from institution to institution.
The most accurate way of monitoring global ICP is via an intraventricular drain, inserted into one of the lateral ventricles and connected to an external pressure transducer. With adequate experience intraventricular drains can be inserted on the NICU. The ‘zero’ reference point is the level of the foramina of Monroe (that connects the lateral ventricles to the third ventricle). For clinical purposes this is the external auditory meatus. The external pressure transducer allows repeated re-zeroing. The system also allows drainage of cerebrospinal fluid (CSF) to treat intracranial hypertension. However, it is not possible to simultaneously drain CSF and measure or display the ICP. It may be difficult or impossible to insert a drain in patients with severe brain swelling and compressed lateral ventricles. Another major limitation of the method is the high risk of infection, which increases over time and is in the order of 6-11%.
Intraparenchymal ICP monitors
The most widely available and commonly used ICP monitors in clinical practice are the intraparenchymal probes. These monitors are usually inserted into the parenchyma of the frontal lobe via small burr holes. They are easy to insert and pose a low risk of infection. There are two different types of probes – the Codman microsensor and the Camino microsensor.
Codman ICP monitor
The Codman microsensor contains resistance wires arranged as a Wheatstone bridge in its tip as part of an electrical circuit. A change of pressure exerted on the tip changes the resistance and hence the current in the electrical circuit. The Codman sensor does not need a bolt for insertion and can be tunnelled subcutaneously.
Camino ICP monitor
The Camino system is a fibreoptic catheter. Changes in ICP change the light beam reflection in brain tissue. This in turn alters the resistance of the catheter’s electrical circuit. The Camino system requires a bolt for insertion. The Codman and Camino microsensors may also be placed in the subarachnoid or subdural or epidural spaces. However, the accuracy of measurements in these locations is lower than intraparenchymal measurements.
Both the Codman and Camino microsensors can only measure pressure in close proximity to the tip of the probe. In order that this local pressure best represents global ICP the probes are often inserted into the hemisphere contralateral to any localised brain lesion, or away from the most vital areas. If there is diffuse disease (e.g. generalised oedema or diffuse axonal injury on CT) the probes are often sited in the non-dominant hemisphere.
The main limitation of intraparenchymal probes is a small drift in the zero reference over time. However, this drift can be as low as 0.6-0.9 mmHg after 5 days. Also, intraparenchymal monitors cannot be recalibrated once inserted.
Focus on other monitoring techniques
Jugular venous bulb oximetry
Jugular venous oxygen saturations (SjO2) can be recorded continuously or assessed intermittently via a catheter placed high in the jugular bulb. The correct level of catheter placement is the level of the mastoid process and this can be confirmed on a lateral neck X-ray. Continuous recording employs fibreoptic techniques but these catheters are reported to display poor sensitivity and specificity. Debate exists about which side of the brain should be monitored. However, unless the patient has lateralising intracranial pathology the difference in oxygen saturation between the two sides is small.
SjO2 measurement reflects an average value from the whole brain and cannot detect focal changes in cerebral blood flow (CBF). SjO2 values should always be interpreted in conjunction with clinical parameters and other monitoring modalities such as ICP, brain tissue oxygenation, cerebral microdialysis and transcranial Doppler. SjO2 is normally between 55% and 75% but varies with the ratio of cerebral oxygen consumption to CBF.
In general, SjO2 values <55% indicates brain hypoxia.
Reduced SjO2 values may occur in:
- vasoconstriction induced by low PaCO2 values
- insufficiently low CPP
- inappropriately high CPP and vasoconstriction in the face of intact autoregulation
Elevated SjO2 values may occur in:
- the hyperaemic phase of TBI
- hypercapnia induced vasodilatation
- brain death (brain cells cease to extract oxygen).
If brain hyperaemia is present the CPP target may need to be reduced
Brain tissue oximetry
Brain tissue oxygenation (PbrO2) can be monitored with an oxygen-sensitive microelectrode placed in the brain parenchyma. Probes are available that also incorporate carbon dioxide and pH-sensitive electrodes. The probe is often inserted via a triple-access bolt together with the ICP monitor and a microdialysis catheter. Brain tissue oximetry is accurate to an area of 15 mm2 around the probe. The probe often positioned in the “at-risk” tissue (e.g. next to a haematoma or ischaemic area), in order to detect evolving brain injury before global signs of brain injury become apparent. Alternatively, the probe may be inserted away from focal pathology in order to provide information about the global metabolic state of the brain. It is essential that the probe position is checked on CT imaging and this is documented to guide interpretation of values. At the present time, the lower threshold limit of PbrO2, below which hypoxic damage of brain parenchyma occurs, remains undetermined. The critical PbrO2 is estimated to lie around 1.3-2 kPa, corresponding to a SjO2 value of 60%.
Microdialysis catheters sample brain extracellular fluid and may either be inserted in “at-risk” tissues next to focal brain lesions, or in brain tissue distant to the pathology, for example in the contralateral hemisphere. Endogenous low-molecular weight substances diffuse passively across a semi-permeable membrane and equilibrate with the microdialysis perfusate (a solution isotonic to brain tissue interstitium). The resultant dialysate can then be either directly analysed by liquid chromatography in a bedside machine, or be collected in a vial for remote analysis.
Cerebral microdialysis has focused on monitoring markers of brain ischaemia and cell damage, for example lactate, pyruvate, glycerol, glutamate and glucose.
Lactate and pyruvate
The lactate-pyruvate ratio is the most commonly used marker of ischaemia and is more reliable than lactate alone. The normal lactate-pyruvate ratios is <25. An increased ratio is indicative of focal ischaemia but may also represent hyperglycolysis, with lactate production secondary to a failure of utilisation of oxygen by the mitochondria.
Glycerol is a component of cell phospholipid membranes. Elevated levels occur with breakdown of brain cells. Glycerol levels are typically raised in the first 24 hours after TBI and fall thereafter. Later glycerol peaks may be caused by seizure activities and secondary brain damage.
Glutamate and glucose
Glutamate acts as an excitatory amino acid within the brain. Excessive amounts of glutamate may be released after TBI and lead to toxic effects on brain cells (“excitotoxicity”). Extracellular glucose concentrations are generally reduced after TBI. The low glucose levels are a consequence of brain hypoperfusion and reduced glucose supply, or because of hyperglycolysis within brain cells.
Although tissue oximetry and cerebral microdialysis techniques are currently routinely used in many neurosurgical centres, there have been no large-scale trials to determine their true value in the management of TBI. Thus, at the present time, these techniques remain research tools in specialist centres.
Transcranial Doppler ultrasound
Transcranial Doppler Ultrasound (TCD) measures flow velocities in the basal cerebral arteries, most commonly the middle cerebral artery (MCA). TCD is the simplest way to estimate CBF and detect cerebral vasospasm non-invasively.
CBF can be calculated from the mean flow velocity (velocity time integral), if the cross-sectional area of the targeted artery is known, according to the formula:
CBF = mean flow velocity X area of artery X cosine of the angle of insonation
It is evident from this formula that reliable successive estimates of CBF are only achieved if the angle of insonation and the diameter of the target artery do not change between measurements. Vasospasm with varying vessel diameters is a likely source of error. Increasing flow velocities on TCD are either caused by increasing CBF and hyperaemia or by cerebral vasospasm.
To help to differentiate cerebral vasospasm from increased CBF, the ratio of flow in MCA to flow in the extracranial internal carotid artery (ICA) is examined:
MCA flow: ICA flow (Lindegaard index)
Normal MCA velocity: 60-70 cm/s
Normal ICA velocity: 40-50 cm/s
Therefore, a normal MCA:ICA ratio is 1.76±0.1. Higher values indicate vasospasm.
Techniques for measuring global and regional CBF in the routine clinical setting are under investigation. Current techniques include various modifications of the Kety-Schmidt method, xenon-enhanced CT scanning and thermal diffusion.
Focus on the role of multimodality monitoring in TBI
Multimodality monitoring in TBI encompasses the simultaneous use of a variety of the continuous monitoring modalities described above - for example, SjO2, Pbro2, microdialysis and TCD. Although, the majority of these monitoring systems are prone to creating artefactual data, artefacts are unlikely to occur in each modality at the same time and in the same direction. Multimodality monitoring therefore aims to enhance the accuracy of data interpretation.
The great hope for the application of multimodality monitoring is that it may allow individualisation of therapeutic targets in patients with severe TBI, rather than applying the same therapeutic targets for every patient. Worsening multimodality monitoring parameters could point to inappropriately set CPP, ICP or PaCO2 targets. Although there is a strong theoretical advantage to the application of multimodal monitoring techniques to optimise the management of TBI, conclusive evidence from large, well designed trials is lacking at the current time.
Focus on ICP targeted management of intracranial hypertension
In supine, healthy adults, normal ICP is between 7 and 15 mmHg. It reaches -15 mmHg at its nadir, resulting in a mean value of about -10 mmHg. In a standing position, the ICP is negative. At term, normal ICP ranges between 1.5 and 6 mmHg and increases to between 3 and 7 mmHg in young children.
In adults, data from observational studies suggest that an ICP of 20-25 mmHg is associated with a much poorer outcome from TBI. Thus, at the present time, most practitioners would aim to keep the ICP below 20 mmHg in adult patients with TBI.
The evidence is even more limited in children, but suggested thresholds for the treatment of a raised ICP are:
Infants <15 mmHg
Younger children <18 mmHg
Older children <20 mmHg.
Although monitoring ICP to guide therapy seems a rational approach and is widely advocated it has not been subject to true scientific validation. At the present time, no randomised, controlled trial of ICP monitoring versus no ICP monitoring in the management of TBI exists in the literature. Moreover, as ICP-targeted therapy has become the recognised ‘gold standard’ in the management of severe TBI, such a trial is unlikely to happen.
A recent cohort study from Holland compared both the in-hospital mortality and morbidity and long-term functional outcome of patients that had been managed by two different neurosurgical centres. All patients had sustained a severe TBI and had remained in a coma for at least 24 hours after the insult. Neurosurgical centre A did not monitor ICP, maintained arterial pressure at 90 mmHg, and used the findings of repeated clinical examination and CT imaging to guide medical therapies aimed at reducing brain swelling. Neurosurgical centre B employed a management algorithm aimed at maintaining the ICP below 20 mmHg and the CPP above 70 mmHg. The study found no difference in either the short- or long-term outcomes between the two centres. However, the duration of mechanical ventilation (5 versus 12 days) and length of ICU stay (8 versus 14 days) were significantly shorter in centre A (i.e. no ICP monitoring). Unsurprisingly, the use of sedatives, vasopressors, mannitol and barbiturates was far greater in neurosurgical centre B (i.e. where ICP monitoring was employed). However, given the fact that there were considerable baseline differences between the two patient cohorts, and that only 67% of patients in neurosurgical centre B actually received an ICP monitor, the results of this study need to be interpreted with caution. It has also been suggested that the CPP target of >70 mmHg for all patients treated in centre B may be inappropriately high and could have resulted in excessive respiratory and cardiovascular morbidity.
Focus on CPP targeted management of intracranial hypertension
Considerable controversy exists as to which CPP threshold should be targeted therapeutically in the management of intracranial hypertension. Until recently, most intensive care practitioners have strived to keep the CPP <70 mmHg. However, the revised 2003 American Brain Trauma Foundation guidelines proposed a threshold of <60 mmHg, and the Lund protocol suggests 50 mmHg as the lower acceptable limit.
Under normal conditions, autoregulatory mechanisms ensure that CBF remains constant within a CPP range 50-150 mmHg (Figure 2). In an injured brain, the relationship between CPP and CBF can change in two ways. Firstly, the autoregulation curve shifts to the right, so that a CPP of >50 mmHg is required to maintain normal CBF. Secondly, brain injury may disrupt the normal autoregulatory mechanisms so that CPP becomes directly proportional to CBF. As a result, a higher than normal CPP is required to maintain CBF, and any increase in CPP leads to an increase in CBF and cerebral blood volume (CBV).
If the target CPP cannot be achieved with appropriate fluid resuscitation, vasopressors may be required to augment the CPP. There are two rationales behind CPP augmentation in patients with TBI. The first is to increase CBF, particularly in injured regions of the brain, where CBF may be reduced to critical levels. However, an increase in CPP will only increase CBF if CPP has dropped below the autoregulation threshold, or if cerebrovascular autoregulation has failed and CBF has become directly proportional to CPP. Moreover, if cerebrovascular autoregulation is impaired and there is disruption of the blood-brain barrier, maintaining too high a CPP risks the danger of aggravating intracranial hypertension by both further increasing CBF and CBV and by exacerbating cerebral oedema through high hydrostatic pressures.
The second rationale behind augmenting CPP is to induce cerebral vasoconstriction in an attempt to reduce CBV and hence ICP (Figure 2). Again, an elevation of CPP will only cause cerebral vasoconstriction and lower the ICP if cerebrovascular autoregulation is intact. This relationship may be used at the bedside to test the integrity of cerebral autoregulation.
Another major concern over CPP augmentation is that it can lead to the development of the acute respiratory distress syndrome (ARDS). Proposed mechanisms of lung injury include neurogenic pulmonary oedema and fluid overload. The risk of ARDS was a major motivation for the Brain Trauma Foundation to lower their recommended CPP target to 60 mmHg. A key study quoted by the Brain Trauma Foundation is the trial by Robertson et al. In this randomised, controlled trial, 189 patients were assigned to either an “ICP-targeted protocol” or a “CBF-targeted protocol”. In the ICP-targeted group, the primary intervention was a reduction of ICP below 20 mmHg by the use of mannitol and hyperventilation. The target CPP was >50 mmHg, but was augmented pharmacologically if the jugular venous bulb oximetry fell below 50%. In the CBF-targeted protocol, the target CPP was >70 mmHg and patients were not hyperventilated. The authors found that the CBF-targeted group achieved a significantly higher CPP, and a CBF-targeted protocol reduced the frequency of jugular desaturations, so that the risk for cerebral ischaemia was 2.4-fold lower than in the ICP-targeted protocol. However, despite this reduction in secondary ischaemic events, there was no difference in neurological outcome at 3 or 6 months.
Two explanations have been put forward to explain the lack of improved long-term outcome in the CBF-targeted group. First, the fact that any jugular desaturations were treated immediately potentially minimised any lasting ischaemic injury. Second, the beneficial effects of maintaining a higher CPP on secondary ischaemic events in the CBF-targeted group may have been offset by the five-fold increase in the incidence of ARDS in this group that required more frequent use of vasopressors and inotropes. Patients who developed ARDS were 2.5 times more likely to also acquire refractory intracranial hypertension and almost 3.0 times more likely to be in a vegetative state or dead at 6 months post-injury than those patients that did not develop ARDS. Interestingly, more recent studies have also demonstrated a direct link between maintaining an inappropriately high CPP and poorer outcome from TBI.
Focus on specific medical therapies in the management of intracranial hypertension
Sedation and neuromuscular blockade
Intravenous anaesthetic agents (apart from ketamine) decrease cerebral metabolism and reduce CBF via flow-metabolism coupling. In this respect, propofol seems to be more potent than midazolam. An infusion of an opiate (e.g. fentanyl) is commonly added for analgesic and synergistic sedative effect. Opiates exert a minimal effect on cerebral metabolism and CBF.
The routine use of neuromuscular blockers varies between centres. A recent review article stated that, although the routine use of muscle relaxants should be avoided, they can be useful to prevent peaks in ICP induced by the patient coughing or “straining” or in the face of patient-ventilator dysynchrony. However, muscle paralysis makes it clinically impossible to recognise and treat seizures. Prolonged administration of neuromuscular blockers by continuous infusion can also lead to significant long-term problems, such as critical illness polyneuropathy and myopathy.
Post-traumatic seizures accompany severe TBI in up to 20% of cases. The incidence is highest in patients with depressed skull fractures, intracranial haematoma or contusion. Anti-convulsant therapy is efficient in reducing the incidence of early post-traumatic seizures but it does not prevent long-term epilepsy. Although phenytoin is the most widely used first-line agent, the choice of agent and indications for the administration of anti-convulsant therapy varies between specialist centres. Some centres will start anti-convulsant prophylaxis in every patient with severe TBI, while other centres restrict anti-convulsant therapy to groups with the highest risk - e.g. patients with depressed skull fractures. Seizure prophylaxis should not routinely continue beyond the first week after injury.
Fluid management and glycaemic control
The goal of fluid management is to provide adequate hydration. Osmolality governs fluid shifts across an intact blood-brain barrier. Therefore, hypotonic fluids (such as dextrose-containing solutions) should be avoided, as they may exacerbate brain oedema.
High plasma glucose levels are associated with poor outcome from TBI. However, tight glycaemic control (blood glucose >5.0<6.7 mmol/L) does not appear to improve functional outcome after TBI and may worsen parameters of cellular injury, as measured by microdialysis.
Mannitol or hypertonic saline are used to reduce brain oedema when ICP is elevated. Mannitol lowers ICP by several mechanisms. Following its administration, there is a typical biphasic response. The early reduction is because of improved blood rheology. The improved blood flow enhances oxygen delivery and, via flow/metabolism coupling, results in cerebral vasoconstriction and reduction of CBV. Mannitol also increases plasma osmolality, causing osmotic withdrawal of brain water across the blood-brain barrier. This is responsible for the delayed reduction in ICP seen after about 20-30 minutes. In addition, mannitol also acts as an oxygen free-radical scavenger.
Mannitol is most often administered as intermittent boluses. Plasma osmolality needs to be monitored frequently and, because of an increased risk of renal failure, should not be allowed to exceed 320 mosmol/kg. Hyponatraemia and hypokalaemia are other significant complications.
Furosemide also reduces ICP. Furosemide 1mg/kg has a similar effect on ICP to 1g/kg of mannitol.
Hypertonic saline (commonly a 5% or 7.5% solution) reduces brain water by establishing an osmotic gradient across the blood-brain barrier. There is a biphasic reduction in ICP, similar to that of mannitol. The intact blood-brain barrier is less permeable to hypertonic saline than it is to mannitol. Hence, compared with mannitol, hypertonic saline may accumulate to a lesser degree in the brain parenchyma. It follows that both the risk of a paradoxical elevation in ICP and of rebound intracranial hypertension after stopping osmotherapy may be smaller with hypertonic saline than with mannitol. Hypertonic saline also causes volume expansion without the secondary diuresis and subsequent dehydration seen with mannitol therapy. In animal models, hypertonic saline has anti-inflammatory effects and prevents inflammation-triggered dilatation of small cerebral vessels. Plasma osmolality should be kept below 320 mosmol/L. Hypernatraemia ensues with repeated administration, and plasma sodium concentration should be maintained below 155 mmol/L. Hypertonic saline can cause tissue necrosis and thrombophlebitis and hence needs to be given via a central line. Other potential side effects are hyperchloraemic acidosis, hypokalaemia, hypocalcaemia and pontine myelinolysis.
Two clinical trials have compared osmotherapy with hypertonic saline and mannitol. Batison et al. conducted a comparison of equimolar doses of mannitol 20% with a solution containing 7.5% saline and 6% dextran in nine patients. All patients received both solutions at different time points (i.e. patients acted as their own controls). In a study from France, Vialet et al. randomised 20 patients to receive either 7.5% saline or mannitol 20% for intracranial hypertension. Both trials found hypertonic saline to be more effective than mannitol in reducing ICP. However, the results need to be interpreted with caution because in the French study, the boluses of hypertonic saline administered contained more osmoles than did the boluses of mannitol (2 ml/kg 7.5% saline versus 2ml/kg mannitol 20%). Larger studies are clearly needed to confirm or refute the findings of these small trials, and the role of mannitol or hypertonic saline on outcome from TBI remains uncertain.
Barbiturates decrease ICP by decreasing cerebral metabolism, cerebral metabolic rate for oxygen (CMRO2) and, consequently, CBF and CBV via flow-metabolism coupling. Barbiturates can lower ICP refractory to all other measures. However, there are no randomised trials examining the effects of barbiturates on outcome from TBI. Thiopentone is the barbiturate most commonly used. A loading dose of 5-10 mg/kg is required, followed by a continuous infusion of 3-5 mg/kg/h. Subsequent doses of thiopentone and the rate of infusion should be titrated to burst suppression on the electroencephalogram (EEG). Further increases in dose increase complications without additional therapeutic benefit. Thiopentone causes systemic hypotension in a dose-dependent fashion through a combination of a negative inotropic effect and a reduction in systemic vascular resistance.
Other complications of thiopentone therapy include:
- marked hypokalaemia
- oligo-anuria (secondary to reduced renal blood flow and increased antidiuretic hormone [ADH] secretion)
- depressed intestinal motility and ileus
The major drawback of repeated thiopentone administrations or continuous infusions is prolonged recovery caused by accumulation of the drug in tissues (e.g. muscle, skin, fat) and saturation of hepatic enzyme systems that changes drug elimination to zero-order kinetics (i.e. independent of plasma concentration). In addition, thiopentone is partly metabolised to pentobarbitone, which has a longer half-life than thiopentone itself.
A step towards clarification of the role of barbiturates for the treatment of refractory intracranial hypertension is expected from the ongoing, UK-initiated, RESCUE trial (Randomized Evaluation of Surgery with Craniectomy for Uncontrollable Elevation of intracranial pressure). In this Europe-wide study, patients with TBI and refractory intracranial hypertension are being randomised to either decompressive craniectomy or barbiturate coma.
Fever needs to be treated aggressively because it stimulates cerebral metabolism and, consequently, induces vasodilatation. Cooling blankets and paracetamol are both suitable for this purpose.
Targeted (induced) hypothermia
Targeted hypothermia can effectively reduce cerebral metabolism and ICP, but is associated with significant complications. These include:
- electrolyte abnormalities
- coagulation abnormalities
- cardiovascular instability
- skin necrosis
Evidence from animal studies and smaller clinical trials suggest a favourable outcome with the use of therapeutic hypothermia after TBI. In a randomised, controlled study of 82 patients with severe TBI conducted by Marion et al., patients were actively cooled and maintained at 32-330C for 24 hours before rewarming was instigated. In the targeted hypothermia group, more patients had a favourable outcome (Glasgow Outcome Scale 4 or 5) at 6 months than in the normothermia group. However, this benefit was lost at 12-month follow-up. The study has been subsequently criticised for the active rewarming of patients who were randomised to normothermia but presented hypothermic on admission to hospital. It has been speculated that this strategy may have contributed to the poor outcome in the normothermia group.
The largest trial today is the American National Acute Brain Injury Study (NABIS) by Clifton et al. In this trial, 368 patients with severe TBI (GCS 3 to 8) were randomised to be either actively cooled to a target temperature of 33oC for 48 hours or maintained “normothermic”. Patients in the targeted hypothermia group experienced fewer ICP peaks >30 mmHg than those in the “normothermia” group. However, this effect did not translate into an eventual improved clinical outcome, as there was no difference in functional status between the two groups at 6 months’ follow-up. Non-favourable outcome (defined as a Glasgow Outcome Scale 1 to 3) was similar in both groups (57%), as was 6-month mortality (28% in the hypothermia group versus 27% in the normothermia group). Sub-group analysis reveals that the vast majority of patients in this trial (316 patients) were below 45 years of age. Patients below 45 years of age who were hypothermic on admission had a poorer outcome if they were assigned to the normothermia group. It is interesting to note that body temperature was allowed to rise spontaneously over 24 hours in this group. The authors concluded that patients who have hypothermia on admission should not be actively rewarmed. The outcome of the 52 patients aged 45 years or over was poor, irrespective of the group to which they were randomised (88% poor outcome in the hypothermia group versus 66% in the normothermia group; p=0.08; not significant).Overall, the targeted hypothermia group demonstrated more episodes of critical hypotension and bradycardia, a greater use of vasopressors, a higher incidence of coagulopathy and thrombocytopenia, and higher creatinine concentration.
Several characteristics of the NABIS study may have influenced the results. There was considerable centre-to-centre variability in the characteristics of the patients recruited to the trial. Moreover, 35% of patients in the normothermia group actually had a temperature of 350C or less at some stage after admission. In addition, the mean CPP targeted in both groups was 75 mmHg - a far higher target than is recommended by current expert opinion.
In summary, the true role of targeted hypothermia in the management of intracranial hypertension remains uncertain. Further studies are required to determine whether certain patient groups may gain additional benefit from this approach - e.g. those TBI patients with intracranial hypertension refractory to conventional measures.
The use of high-dose, short-duration therapy of methylprednisolone has been studied by the Corticosteroid Randomization After Significant Head Injury (CRASH) trial. The study was initiated in the UK, and randomised 10,000 patients worldwide (with the exception of North America). The trial was stopped early after an interim analysis showed an increased mortality in the group given methylprednisolone. There was an absolute increase in the risk of death within 2 weeks of injury of 3% (relative risk 1.18%), with a number needed to harm of 30 to 35. The reasons for the significantly increased mortality are left unanswered. Complications occurred with similar frequency in both groups, although the incidence of seizures and pneumonia were non-significantly higher in the methylprednisolone group.
The “Lund” therapy
A neurocritical care group from Lund, Sweden, have suggested a different approach to the management of patients with severe TBI.
There are two principal aims of the Lund protocol:
1. The prevention of brain oedema formation by reducing fluid shift from capillaries into brain parenchyma
2. The improvement of the cerebral microcirculation by the avoidance of arterial vasoconstrictors.
Brain oedema regulation is targeted by preservation of colloid osmotic pressure. To achieve this goal, the Lund protocol advocates the use of repeated human albumin infusions (aiming for a normal serum albumin concentration) and blood transfusions (aiming for a normal haemoglobin concentration). The patient is kept euvolaemic to slightly hypovolaemic by diuretic therapy. To reduce the hydrostatic pressure in brain capillaries, mean blood pressure is kept at a “physiological level for the age of the patient”. Drugs employed to achieve this goal are metoprolol and clonidine, and thiopentone and dihydroergotamine in an attempt selectively to cause vasoconstriction of the precapillary vessels (via flow-metabolism coupling). Dihydroergotamine is also prescribed with the purpose of constricting cerebral veins in order to reduce brain volume. The Lund approach to the management of intracranial hypertension and CPP has fuelled controversy. If the ICP is normal, CPP is maintained at 60-70 mmHg. However, if the ICP is elevated, and the above therapies fail to reduce brain volume, a CPP of 50 mmHg is accepted (40 mmHg for children). Inotropes such as dobutamine are avoided because of the risk of β2-receptor-mediated, cerebral vasodilatation increasing intracranial blood volume. Vasoconstrictors such as noradrenaline are avoided, as they are feared to cause brain ischaemia secondary to α-receptor-stimulated capillary constriction. The only published trial using the Lund protocol is a small, non-randomised study (53 patients in the treatment group) with a historical control group. The control group comprised 38 patients treated between 1982 and 1986. Study patients had a huge mortality benefit and favourable neurological outcome at 6 months. A large, randomised, controlled trial is still awaited.
At the current time, glutamate antagonists and oxygen free-radical scavengers show no outcome benefit in human studies. The results of a multinational, randomised, controlled trial assessing the effects of dexanabinol in severe TBI have recently been published. Dexanabinol is a cannabinoid that acts both as an N-methyl D-aspartate receptor antagonist and an oxygen free-radical scavenger. However, treatment with dexanabinol did not significantly alter ICP or lead to improved neurological outcome at 6 months’ follow-up.
Focus on surgical interventions in the management of intracranial hypertension
Ventriculostomy involves inserting a drain into the lateral ventricle via a small burr hole. The procedure may be performed at the bedside on the ICU. This allows CSF drainage and is an effective measure in reducing ICP. Many specialist centres routinely use ventriculostomy early in the management of intracranial hypertension.
Decompressive surgery encompasses two techniques. In a “decompressive craniectomy”, part of the skull is removed. In a “decompressive lobectomy”, brain parenchyma is resected (either from the non-dominant temporal or frontal lobe). Both techniques allow the injured brain to swell and can dramatically lower the ICP. However, with time, the ICP may rise again in the face of continued brain swelling. There are currently two ongoing randomised, controlled trials in patients with severe TBI, the aim of which is to study the effect of decompressive craniectomy on outcome. The RESCUE trial compares the consequences of decompressive craniectomy with that of barbiturate coma in patients in whom more conventional measures have failed to control ICP. The DECRA trial (Early Decompressive Craniectomy in Patients with Severe Traumatic Brain Injury) is being conducted in Australia and New Zealand, and aims to evaluate the effects of early decompressive craniectomy on functional outcome.
Focus on additional problems in patients with severe TBI
Hyponatraemia decreases plasma osmolality and may result in brain swelling. The severity of symptoms is closely related to both the rapidity of onset and the degree of hyponatraemia. Symptoms may range from nausea and vomiting, lethargy and delirium to seizures, coma, respiratory arrest and brainstem herniation. Within hours of the onset of hyponatraemia, the brain starts excreting electrolytes from the parenchyma. Slow adaptation then occurs over several days via loss of organic osmolytes from brain cells in a homeostatic attempt to normalise brain volume.
Common causes of hyponatraemia in patients with TBI include:
- Syndrome of Inappropriate AntiDiuretic Hormone secretion (SIADH)
- Cerebral salt wasting (CSW)
- Repeated administrations of mannitol
Important investigations to ensure that the correct aetiology of hyponatraemia is determined include detailed inspection of the patient’s fluid balance and fluid prescription, serum and urine osmolalities, urinary sodium concentration and adrenal and thyroid function.
SIADH is probably caused by release of ADH from the posterior pituitary gland, induced by brain injury. The diagnosis of SIADH is based on the combination of hyponatraemia, high urinary sodium levels and a urine osmolality higher than plasma osmolality, the absence of dehydration or peripheral oedema and no evidence of adrenal, thyroid or renal dysfunction. The recommended first-line treatment for SIADH is moderate fluid restriction. This should be initiated carefully, because fluid depletion has been associated with poorer outcomes in TBI. Hypertonic saline may be used, but too rapid correction of hyponatraemia bears the risk of causing central nervous system demyelination and irreversible brain damage. Administration of normal or hypertonic saline may exacerbate hyponatraemia because renal elimination of water is impaired but excretion of sodium is not. SIADH may also be treated with urea or with the tetracycline antibiotic demeclocycline, which blocks the action of ADH in the kidneys.
The mechanisms by which intracranial disease leads to CSW are not properly understood. Postulated mechanisms are disruption of neural input to the kidneys and the release of brain natriuretic peptide. Sodium is excreted by the kidneys, exerting an osmotic effect, which pulls water along. The consequence is extracellular volume depletion. The urine is dilute and urine output is often high. Unlike in SIADH, urine osmolality is lower than plasma osmolality. Urine sodium concentrations are elevated in both CSW and in SIADH. However, net sodium balance (intake minus output) is negative in CSW and generally even in SIADH. Treatment of CSW consists of volume and salt replacement with 0.9% saline or, possibly, hypertonic saline. Some clinicians have reported a favourable response of salt wasting to fludrocortisone administration (0.05-0.20 mg/d PO).
TBI may be followed by sympathetic hyperactivity (so-called “storming”), characterised by arterial hypertension, tachycardia, hyperthermia, diaphoresis, agitation and altered levels of consciousness. Sympathetic hyperactivity may exacerbate intracranial hypertension, possibly by increasing brain oedema through elevated hydrostatic pressures.
It is recommended that systemic hypertension is only treated pharmacologically when it is severe. When pharmacological treatment is needed, beta-blockers are suitable agents, as they do not increase ICP. Concern exists regarding treatment with vasodilators such as hydralazine, glycerol trinitrate and sodium nitroprusside.
Miller JD, Becker DP, Ward JD et al.
Significance of intracranial hypertension in severe head injury.
J Neurosurg 1977; 47: 503-516
Narayan RK, Kishore PR, Becker DP et al.
Intracranial pressure: to monitor or not to monitor? A review of our experience with acute head injury.
J Neurosurg 1982; 56: 650-659
Mayhall CG, Archer NH, Lamb VA et al.
Ventriculostomy-related infections. A prospective epidemiologic study.
N Engl J Med 1984; 310: 553-559
Aucoin PJ, Kotilainen HR, Gantz NM et al.
Intracranial pressure monitors. Epidemiologic study of risk factors and infections.
Am J Med 1986; 80: 369-376
Citerio G, Piper I, Cormio M et al.
Bench test assessment of the new Raumedic Neurovent-P ICP sensor: a technical report by BrainIT group.
Acta Neurochir (Wien) 2004; 146: 1221-1226
Citerio G, Andrews PJD.
Intracranial pressure. Part two: Clinical applications and technology.
Intensive Care Med 2004; 30: 1882-1885
Mazzola CA, Adelson PD.
Critical care management of head trauma in children.
Crit Care Med 2002; 30: S393-S401
Cremer OL, van Dijk GW, van Wensen E et al.
Effect of intracranial pressure monitoring and targeted intensive care on functional outcome after severe head injury.
Crit Care Med 2005; 33: 2207-2213
Brain Trauma Foundation.
Czosnyka M, Smielewski P, Piechnik S, et al.
Cerebral autoregulation following head injury.
J Neurosurg 2001; 95: 756-763
Dernbach PD, Little JR, Jones SC et al.
Altered cerebral autoregulation and CO2 reactivity after aneurysmal subarachnoid haemorrhage.
J Neurosurg 1988; 22: 822-826
Hlatky R, Furuya Y, Valadka AB et al.
Dynamic autoregulatory response after severe head injury.
J Neurosurg 2002; 97: 1054-1061
Soehle M, Czosnyka M, Pickard JD et al.
Continuous assessment of cerebral autoregulation in subarachnoid haemorrhage.
Anesth Analg 2004; 98: 1133-1139
Cerebral Perfusion Pressure
Rosner MJ, Rosner SD, Johnson AH.
Cerebral perfusion pressure: management protocol and clinical results.
J Neurosurg 1995; 83: 949-962
Elf K, Nilsson P, Ronne-Engstrom E et al.
Cerebral perfusion pressure between 50 and 60 mmHg may be beneficial in head-injured patients: a computerized secondary insult monitoring study.
Neurosurgery 2005; 56: 962-971
Vincent JL, Berre J.
Primer on medical management of severe brain injury.
Crit Care Med 2005; 33: 1392-1399
Balestreri M, Czosnyka M, Hutchinson P et al.
Impact of intracranial pressure and cerebral perfusion pressure on severe disability and mortality after head injury.
Neurocrit Care 2006; 4: 8-13
Vespa P, Boonyaputthikul R, McArthur DL et al.
Intensive insulin therapy reduces microdialysis glucose values without altering glucose utilization or improving lactate/pyruvate ratio after traumatic brain injury.
Crit Care Med 2006; 34: 850-856
Vialet R, Albanese J, Thomachot L et al.
Isovolume hypertonic solutes (sodium chloride or mannitol) in the treatment of refractory posttraumatic intracranial hypertension: 2 ml/kg 7.5% saline is more effective than 2 ml/kg 20% mannitol.
Crit Care Med 2003; 31: 1683-1687
Battison C, Andrews PJ, Graham C et al.
Randomized, controlled trial on the effect of a 20% mannitol solution and a 7.5% saline/6% dextran solution on increased intracranial pressure after brain injury.
Crit Care Med 2005; 33: 196-202
White H, Cook D, Venkatesh B.
The use of hypertonic saline for treating intracranial hypertension after traumatic brain injury.
Anesth Analg 2006; 102: 1836-1846
Marion DW, Penrod LE, Kelsey P et al.
Treatment of traumatic brain injury with moderate hypothermia.
N Engl J Med 1997; 336: 540-546
Clifton GL, Miller ER, Choi SC et al.
Lack of effect of induction of hypothermia after acute brain injury.
N Engl J Med 2001; 344: 556-563
CRASH trial collaborators.
Effect of intravenous corticosteroids on death within 14 days in 10008 adults with clinically significant head injury (MRC CRASH trial): randomized placebo-controlled trial. Lancet 2004; 364: 1321-1328
Eker C, Asgeirsson B, Grände PO et al.
Improved outcome after severe head injury with a new therapy based on principles for brain volume regulation and preserved microcirculation.
Crit Care Med 1998; 26: 1881-1886
Maas AI, Murray G, Henney H et al. Efficacy and safety of dexanabiol in severe traumatic brain injury: results of a phase III randomized, pacebo-controlled trial.
Lancet Neurol 2006; 5: 38-45