• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Harris B, Andrews PJD, Murray GD, et al. Systematic Review of Head Cooling in Adults After Traumatic Brain Injury and Stroke. Southampton (UK): NIHR Evaluation, Trials and Studies Coordinating Centre (UK); 2012 Nov. (Health Technology Assessment, No. 16.45.)

Cover of Systematic Review of Head Cooling in Adults After Traumatic Brain Injury and Stroke

Systematic Review of Head Cooling in Adults After Traumatic Brain Injury and Stroke.

Show details


The conditions and incidence: traumatic brain injury and stroke

Brain injuries resulting from stroke and trauma are common and costly in human and resource terms. In England, approximately 130,000 people have a stroke each year, of whom about one-quarter die and half of the survivors are left dependent on others.1 The incidence of head injury is similar to that for stroke,2 although the incidence of death is lower, at 6–10 per 100,000 population per year.3 However, head injury is more common in younger people, and it has been estimated that 4700 of those admitted to hospital each year would be unable to return to work at 6 weeks.2 A Scottish study found that 78% of patients with a severe injury had moderate or severe disability 1 year later.4

Aside from the often devastating consequences for patients and their families, these brain insults are expensive. Morbidity from head injury ‘far exceeds the capacity of UK neurorehabilitation services’3 and the costs of stroke to the NHS are estimated at £2.8B per year, with the cost to the wider economy about £1.8B more in disability and lost productivity.1

Although the primary mechanisms of brain injury are different in trauma, haemorrhage and ischaemia [whether focal, as in ischaemic stroke, or global, as in cardiac arrest and neonatal hypoxic–ischaemic encephalopathy (HIE)], the result is a cascade of excitotoxity, apoptosis and inflammation.5,6 Inflammation, cell death and infection, if present, mean that increased temperature is common after both stroke and brain injury.7,8 There is no universally agreed definition of the threshold for pyrexia or where and how temperature should be measured in these patients but, in one study, nearly 68% of patients had a rectal temperature ≥ 37 °C within 48 hours after severe traumatic brain injury (TBI)9 and 54% had an axillary temperature of > 37.5 °C within 48 hours after stroke.10

Increased temperature is associated with worse outcome after both stroke and TBI.9,11 The exact nature of the relationship in humans is hard to determine, as the time of onset of raised temperature has an influence and temperature elevation can be a marker of more severe injury and of infection, both of which are also associated with worse outcome,12 although one systematic review11 suggests that infection may not play a significant part in the relationship in stroke. There is considerable evidence from animal research that reducing temperature, and, more especially, inducing hypothermia, reduces the extent of injury and that the sooner cooling is instigated the more effective it is.6 However, there is insufficient high-quality prospective evidence to show that normothermic or hypothermic temperature interventions improve functional outcome in humans after TBI and stroke.1315 This may be because it is difficult to cool patients early and quickly enough and/or because the side effects of hypothermia, such as increased infection, may outweigh the benefits in some circumstances.

Nevertheless, the usual clinical goal in TBI and stroke is to reduce raised temperature to normothermia, although consistently achieving this can be difficult.16,17 In stroke it is recommended that temperature is treated if > 37.5 °C.18 In brain injury, body temperature control is recommended in the context of treating raised intracranial pressure (ICP).19 There are no standard recommendations on the site of temperature measurement or methods of temperature reduction. In practice, choice of site of measurement is variable20,21 and cooling interventions are usually systemic. Pharmacological intervention, generally with paracetamol, is the most common first-line treatment, followed by a variety of physical systemic cooling interventions, which include cooling blankets, ice packs and fanning.21,22

The intervention: non-invasive head cooling

Physical cooling methods can be classified into those targeted systemically and those targeted at the head to cool the brain directly, and include invasive and non-invasive methods. Non-invasive head cooling is the subject of this review and therefore invasive methods, such as antegrade and retrograde cerebral perfusion and devices applied to brain tissue, which are mainly used during surgery,23 are not included.

Methods of non-invasive head cooling are categorised into:

  • Heat loss from the upper airways This takes place by convection with gas or fluid flow or by conduction with nasal or pharyngeal balloons – whether or not these devices are truly non-invasive is a moot point, but they have been included in this review.
  • Heat loss through the skull This takes place by convection (fanning, hoods delivering cold air or water) or by conduction (passive, e.g. ice, gel caps or active, e.g. liquid cooling); some of the devices also have a neck band that theoretically may help cool the brain by reducing the temperature of the carotid blood supply.24,25

Heat loss occurs as flow down temperature gradients from warm to cool. Convective cooling methods use air/gas flow to remove heat; molecules are removed in bulk and transfer heat in the process. Convective methods also allow heat loss by evaporation, a form of convection in which bulk movement of molecules is achieved by water loss (changing water into water vapour requires large amounts of heat). With conductive methods energy (heat) moves but the molecules do not. Heat from the head is conducted through the wall of the device and either actively removed by the circulating liquid coolant or passively absorbed by the frozen material (ice/gel). Devices containing frozen material will warm up in this process and must be replaced regularly to maintain cooling efficiency.

Non-invasive head-cooling methods are generally quick and easy to apply and may be suitable for pre-hospital use, which are important considerations in reducing time to cooling if neuroprotection is the aim. They also have potentially wide application because they can be used in patients with a range of severity of illness, not just the most severely ill.

How the intervention might work

Although cooling interventions are more commonly delivered systemically, the logic behind head cooling is that it targets cooling where it is needed because it is brain rather than trunk temperature that is important in cerebral protection. It is also thought that head cooling may reduce the complications of hypothermia because less body temperature reduction is required, although the evidence for this is not robust.23

The great advantage of cooling, by comparison with most other neuroprotective interventions, is that it has many potentially beneficial effects with regard to secondary injury mechanisms and therefore cerebral protection. Hypothermia has even been described as ‘the ultimate neuroprotective cocktail’.7 The effects of cooling are not fully understood but include reduction in metabolic rate, modulation of cerebral blood flow, and the inflammatory response and reduction of excitotoxic damage and cerebral oedema.6,26 Because cooling can be very effective in reducing refractory ICP this is the most usual reason for instigating therapeutic hypothermia in severe traumatic and haemorrhagic brain injury.27,28 In ischaemic stroke it is considered possible that therapeutic hypothermia could extend the time window within which restoration of blood supply, for example with thrombolysis, might be effective.29

Measurement of temperature reduction

If cooling, however delivered, is to have a neuroprotective effect, brain temperature must be reduced. The primary measure of the effectiveness of head cooling with regard to temperature reduction is a decrease in intracranial temperature. For the purposes of this review, intracranial temperature is defined as temperature inside the skull and within the dura. In the absence of intracranial temperature data, the secondary measure for this review is reduction in core trunk temperature with head cooling, measured in an artery (usually pulmonary), the oesophagus, bladder or rectum, on the assumption that for core trunk temperature to be reduced there must have been some reduction in intracranial temperature. (For further explanation see Appendix 1.)

Cardiac arrest and neonatal hypoxic–ischaemic encephalopathy

The principal focus of this review is head cooling in TBI and stroke, in which the primary problem is in the brain. However, in global (whole body) ischaemia, following cardiac arrest, therapeutic hypothermia is considered to improve outcome, specifically with return of circulation after ventricular fibrillation,3032 although doubts have been raised over the quality of the evidence.33 Therefore, during the protocol review process, we were asked to include the cardiac arrest literature on head cooling in our searches because this could contribute information about how effective these interventions are in reducing temperature, and on their ease of use and side effects. Studies in cardiac arrest were not relevant for assessment of functional outcome in this review. However, in our opinion, it is not yet clear to what extent whole-body cooling, which includes myocardial cooling, contributes to improved outcome with hypothermia after cardiac arrest, and whether or not head cooling alone is as effective as systemic cooling in this systemic ischaemic injury. There is no comparative randomised controlled trial (RCT) but there is some evidence, for example, that myocardial reperfusion injury, which can be ameliorated by hypothermia, may contribute to post-arrest morbidity and mortality.34,35

Neonatal HIE is the other global ischaemic condition in which therapeutic hypothermia has been shown to be of benefit.36,37 Head cooling has been commonly used as the means of achieving hypothermia in neonatal HIE but whether or not it has advantages over systemic cooling has not yet been assessed in a comparative RCT.36,38 However, a recent systematic review and meta-analysis in neonatal HIE includes a subgroup analysis of systemic hypothermia (seven studies) and head cooling (six studies) compared with normothermia, which shows that more adverse functional outcomes were reduced with systemic cooling than with head cooling.38 It is relatively easy to cool infants with head cooling as they have a smaller body–head ratio than adults and therefore have less counterwarming from the trunk; also their skulls are not closed because their fontanelles have not fused. Intracranial temperature is not measured clinically in infants with neonatal HIE, but head cooling has a considerable ‘knock-on’ effect on body temperature and body warming is required to control systemic hypothermia.39 The effect of head cooling on temperature in neonates does not extrapolate to adults, but neonatal head-cooling research could contribute information on adverse effects of methods and devices therefore it was included in the review for this purpose.

The reason for undertaking this review

Systematic reviews of cooling interventions after brain injury and stroke have not differentiated between cooling methods. The only Cochrane review of a specific cooling intervention, for example, is that of paracetamol for fever in children.40 In the reviews of cooling for acute stroke14 and of hypothermia for head injury15 the effect of temperature reduction on outcome has been the focus rather than the method(s) of achieving this, although a distinction was made between pharmacological and physical methods in stroke. Yet physical cooling methods differ in their effectiveness and complications. The reason for using head cooling is that, theoretically, it may have advantages over systemic cooling. Cooling is targeted to the site of injury where it is most needed, therefore requiring less body temperature reduction relative to brain temperature, which means that it may have fewer side effects than systemic physical methods. In order to determine whether or not head cooling has an effect and whether or not there are advantages it was necessary to review head cooling as an intervention.

© 2012, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Bookshelf ID: NBK127510
PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.9M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...