Article Text


Treatment of hypoxic-ischaemic brain damage by moderate hypothermia
  1. A D Edwardsa,
  2. J S Wyattb,
  3. M Thoresenc
  1. aSection of Paediatrics Imperial College School of Medicine Hammersmith Hospital London, bDepartment of Paediatrics University College London School of Medicine London, cDepartment of Child Health University of Bristol St Michael’s Hospital Bristol
  1. Professor A D Edwards Section of Paediatrics Imperial College School of Medicine Hammersmith Hospital London W12 ONN. email dedwards{at}

Statistics from

To many people, especially writers of science fiction interested in preserving brains for narrative purposes, it seems self evident that cooling the brain protects it against hypoxic–ischaemic damage. Indeed, every day, cardiac surgeons and anaesthetists cool the brains of children during surgery to protect them against the effects of cardiac arrest or cardiopulmonary bypass.

However there has long been a hope that cooling the brainafter hypoxia–ischaemia might lessen cerebral injury. Observational data in support of this were collected by Westin and colleagues 40 years ago,1 but experimental studies in animal models at that time failed to support the hypothesis and it fell from favour.2 Now the belief is gaining ground again among basic researchers that moderate brain cooling to around 32oC is one of several interventions which can be applied after hypoxia–ischaemia to modify the process of brain cell death and so lessen cerebral damage.

Delayed cerebral damage after hypoxia–ischaemia

A cornerstone of this growing consensus was the realisation that not only do some cells die during hypoxia–ischaemia, but many more may die hours or days later.3 In the 1980s this delayed cerebral injury was shown in infants with birth asphyxia using31P magnetic resonance spectroscopy (MRS): asphyxiated infants were usually found to have normal cerebral energy metabolism soon after resuscitation, but oxidative phosphorylation became impaired 9 to 24 hours later and remained low for many days.4 The delayed impairment of energy metabolism was not contingent on continued hypoxia–ischaemia, nor was it associated with intracellular acidosis,5 but its magnitude was linearly related to the severity of later neurodevelopmental impairment and reduced brain growth.6

This was clear evidence that the effects of birth asphyxia might become manifest only some days after resuscitation, and indeed more recent data suggest that abnormal cerebral energy metabolism and cell death can persist for weeks and months.7 8 The mechanisms of this delayed cell death are complex, but the recognition that apoptosis (genetically controlled cell suicide which requires a biochemical pathway to be completed for death to occur) was an important aspect of hypoxic–ischaemic cerebral injury made it easier to understand why cells should die late after injury.9 10

These results suggested that it might be possible to intervene before delayed cell death occurred and so lessen cerebral damage. The hypothesis was supported by experiments in animals which showed that a large number of treatments, ranging from administration of glutamate receptor antagonists to injection of growth factors, would reduce brain injury if given soon after the insult.11 Thus rescue treatment of hypoxia–ischaemia after resuscitation was confirmed as a real possibility, and in some cases treatment could still be effective even if delayed for a considerable time after the hypoxic–ischaemic insult.12

Neural rescue treatment using post-insult hypothermia

Many recent studies, mainly in adult animals, have shown that cooling of the brain by 3–4oC after experimental hypoxia–ischaemia reduces the severity of brain injury.13The more drastic temperature reductions used in older experiments do not seem to provide any additional benefit and indeed may be less effective as they have been associated with systemic toxicity.14 Consequently most current research focuses on moderate rather than deep hypothermia.

Moderate hypothermia is one of the most robust and effective techniques available for reducing hypoxic–ischaemic cerebral damage. However, many experiments have used short-term histopathological examination to assess neural cell loss, which raised concern that hypothermia may not prevent injury, merely delay it. But a major study of adult gerbils cooled after hypoxia–ischaemia, found that they performed better than controls in neuropsychological tests six months later, showing that a persistent beneficial effect of cerebral function could be produced.15 Some researchers now believe that in adult animals at least, a short period of hypothermia lasting a few hours will delay cell death while more prolonged cooling for a day or two permanently prevents cerebral injury.16

Neural rescue using moderate hypothermia in the developing brain

Several groups have shown that post-insult hypothermia can reduce cerebral injury in the developing brain. In 7 day old rat pups hypothermia for a period as short as 3 hours after hypoxia–ischaemia had some neuroprotective effect,17 and histological differences between treatment and control brains could still be detected 6 weeks later.18 In 21 day old rat pups 72 hours of hypothermia was highly protective, although a delay of 6 hours before starting treatment significantly reduced the benefit.19

One study of 7 day old rats failed to show any benefit, but this may be explained by the severe nature of the hypoxic–ischaemic insult combined with a relatively brief period of 3 hours cooling after the insult.20 In a study of newborn piglets 3 hours of cooling exerted only a modest neuroprotective effect and there was no protection at all in more severely injured animals.21 By contrast, 12 hours of cooling by 4oC in newborn piglets produced a major reduction of both the delayed impairment in cerebral energy metabolism, and histological injury.22 23

All these studies have examined the effect of whole body cooling, but investigations of fetal sheep subjected to total cerebral ischaemia found that a substantial protective effect could be produced by a cooling device positioned around the fetal head.24 In this model cooling was maintained for 72 hours, but clinically significant protection was achieved even if treatment was delayed for 5 hours after ischaemia.25 Focal brain cooling was also effective in 7 day old rat pups.26 Taken together these results suggest that optimal benefit is obtained by extending the duration of moderate hypothermia for 12 hours or longer, although they leave unanswered the question of whether whole body or selective head cooling is better.

How cooling works

Hypothermia applied during hypoxia–ischaemia is thought to be protective by preventing the decline in high energy phosphates that seem to initiate both apoptotic and necrotic cell death.27 28 The mechanisms by which coolingafter hypoxia–ischaemia prevent cell death are less clear. Hypothermia prevents the delayed decline in phosphocreatine and adenosine triphosphate, as well as the simultaneous increase in cerebral lactate concentration, seen 8–12 hours after hypoxia–ischemia in newborn piglets.22 29 However, it is not clear if preservation of energy metabolism in the delayed phase of injury is the primary mechanism by which hypothermia operates, or whether it represents an indicator of cellular protection mediated by other pathways.

Additional effects are probably involved. Hypoxia–ischaemia leads to high concentrations of glutamate in the synaptic cleft which induce excitotoxic neuronal death.30 In both newborn piglets31 and adult rats32 hypothermia reduces the delayed increased in extracellular glutamate seen after hypoxia–ischaemia, and in a cell culture model of hypoxia–ischaemia mild cooling decreased the impairment in glutamate re-uptake that is an important factor in acute injury.33Hypothermia also reduced production of a downstream mediator of the excitotoxic process: nitric oxide.31 34 However, these mechanisms may not be a sufficient explanation, as in at least one model where hypothermia is effective there is no increase in nitric oxide production after hypoxia–ischaemia (Edwards et al, unpublished data), and the role of impaired glutamate re-uptake in delayed cell death is unclear.30

Hypothermia reduces the number of ischaemic depolarisations in injured but viable tissue,35 and prevents the increase in cerebral impedance (which reflects impaired membrane function) during delayed cerebral injury in fetal lambs.24 Reduced body temperature may also increase catecholamine secretion,36 and as stress also improves neuropathological outcome in developing rats,37 sympathetic stimulation may be another mechanism of protection.

Hypothermia reduces the number of apoptotic cells seen in the newborn piglet brain after hypoxia–ischaemia without affecting the number of necrotic cells, suggesting that it may specifically inhibit the apoptotic pathway.23 However, comparable hypothermia delays rather than prevents apoptosis in cell culture systems,38 and an attractive hypothesis is therefore that hypothermia delays commitment to apoptosis for long enough to enable endogenous protective mechanisms, including the production of growth factors,39 40 to be induced. If this is correct it suggests that hypothermia may have a role in prolonging the therapeutic window during which other treatments might be applied.

Studies of adult animals have suggested other mechanisms through which hypothermia may act including: suppression of free radical action; prevention of ischaemia induced protein kinase C inhibition; or activation of transcription factors.41 It remains to be seen whether these mechanisms are important in the developing brain. However, the success of hypothermic treatment probably depends on it, affecting several of the many mechanisms of damage which are activated by hypoxia–ischaemia.

Hyperthermia after hypoxia–ischaemia

There is a developing body of experimental evidence to suggest that increased brain temperature after hypoxia–ischaemia may worsen cerebral damage. Observational data from adults who have had a stroke42 are supplemented by experiments which have shown increased glutamate release in hyperthermic animals, and worse histological outcome even if the period of hyperthermia is delayed for 24 hours after the period of hypoxia–ischaemia.43 44 Of particular interest are epidemiological data showing a ninefold increase in the incidence of cerebral palsy in infants weighing more than 2500 g who were born to mothers with a fever exceeding 38oC.45

Adverse effects of hypothermia

Profound hypothermia to less than 30oC affects many physiological functions. It decreases perfusion and oxygenation by: impairing myocardial contractility; reducing cardiac output; and making myocardial muscle more prone to dysrhythmia46 47; as well as causing peripheral vasoconstriction; increasing blood viscosity48; and shifting the oxygen dissociation curve of blood to the left.49 This deranged circulation can lead to renal failure, pulmonary oedema, metabolic acidosis and inadequate cerebral blood flow.46 Cooling also impairs clotting,50 depresses the immune system,51disrupts serum potassium homeostasis,52 alters acid base balance,53 and is associated with hypoglycaemia.54 Hypothermia may cause gastrointestinal lesions in the developing animal, although the data are somewhat contradictory.55 56These and other changes can significantly worsen the outcome of experimental subjects,14 and early observational studies of newborn infants showed a number of these adverse events, of which pulmonary haemorrhage (probably due to raised left atrial pressure) was the most severe.57

However, although moderate hypothermia to around 32oC has been less well studied, adverse effects seem to be less severe. Indeed, newborn animals physiologically induce moderate hypothermia in response to hypoxia.58 Recent studies of newborn piglets subjected to moderate whole body cooling for 12 hours as a treatment for either cerebral hypoxia–ischaemia or 3 hours for whole body hypoxia, showed little evidence of circulatory or metabolic disruption during cooling22 or systemic pathology at necropsy.59 60

Nevertheless, systematic studies of the potentially toxic effects of moderate hypothermia maintained for longer periods are required, as the risk of adverse effects from moderate hypothermia seems likely to be increased if hypothermic treatment is prolonged to maximise its neuroprotective effect. There is likely to be a complex trade-off between optimising protection and minimising side effects which will involve decisions not only about the length and depth of hypothermia but also about the appropriate balance between selective head and whole body cooling.


Cerebral hypoxia–ischaemia leads to a delayed period of cell death that begins some hours after adequate oxygenation is restored to the brain. Cooling the brain to around 32oC for between 12 and 72 hours, beginning after resuscitation, significantly reduces cerebral damage and long term sequelae in experimental models. This moderate hypothermia may be associated with relatively few adverse systemic effects, and although the mechanisms of cerebral protection are incompletely understood, interest is growing in the possibility of clinical neuroprotection by reduction in brain temperature.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.