Topical review

Hyperthermia exacerbates ischaemic brain injury

C. X. Wang1,2, A. Stroink1, J. M. Casto2, and K. Kattner1,2

Hyperthermia frequently occurs in stroke patients. Hyperther- Hyperthermia is a common complication of ischaemic
mia negatively correlates with clinical outcome and adversely stroke; studies show it is detrimental to the ischaemic injured
effects treatment regiments otherwise successful under nor- brain and correlates with poor outcome (6). Furthermore,
mothermic conditions. Preclinical studies also demonstrate hyperthermia also adversely affects treatment regimens that
that hyperthermia converts salvageable penumbra to ischae- work under normothermic conditions. For example, previous
mic infarct. The present article reviews the knowledge accu- animal studies have shown that mild hyperthermia abolishes
mulated from both clinical and preclinical studies about the effects of an otherwise neuroprotective N-methyl-d-aspar-
hyperthermia and ischaemic brain injury, examines current tate (NMDA) agonist, MK-801 (7). Similarly, in a recent study,
treatment strategies and discusses future research directions. we demonstrated that hyperthermia masks the beneficial
effects of tPA in a focal embolic model of cerebral ischaemia,
Key words: cerebral, hyperthermia, infarction, stroke, tem-
despite the enhanced fibrinolytic activities of tPA due to
perature
increased temperature (8).
While there is extensive research on the effects of hypother-
mia on neuronal damage (beyond the scope of this discussion)
and putative pharmacotherapy, less intention has focused
Introduction
on molecular mechanisms of hyperthermia in ischaemic
Stroke is a disorder affecting approximately 15 million people brain injury. The present article reviews the current knowledge
worldwide. It is the third leading cause of death in the majority about hyperthermia and ischaemic brain injury, accumulated
of industrialised countries and also in many developing from both clinical and preclinical studies, with intentions of
countries (13). More than 85% of strokes are due to ischaemic stimulating more research in both clinical and preclinical
brain injury; haemorrhagic brain injury accounts for the levels.
remaining cases (1). Although thrombolysis with tissue plas-
minogen activator (tPA) is the only effective pharmacological
treatment proven in randomised-controlled multicentre clin- Body temperature and thermoregulation
ical trials, more than 95% of patients are ineligible to receive
Body temperature is regulated by the hypothalamus, which
this treatment. Unfortunately, neuroprotective therapies have
requires integrated autonomic, endocrine and skeletomotor
not proven to be effective treatment strategies. To date, 15 000
responses. Although both anterior and posterior hypothalamic
patients have participated in more than 65 clinical trials of
areas are involved in temperature regulation, the detectors of
neuroprotective agents (4, 5). Despite enormous efforts, all
body temperature, both low and high temperatures, are
compounds reaching clinical trials have failed due to a lack of
located in the anterior hypothalamus. The anterior hypotha-
demonstrable efficacy or problems with intolerable adverse
lamus contains cold-sensitive and warm-sensitive neurons
effects.
which serve as the thermoregulatory centre. The activity of
Correspondence: Chen Xu Wang, 232A FSA, Illinois State University, these neurons is highly influenced by the temperature of local
Normal, IL 61790-4120, USA. Tel: 11 309 438 4494; Fax: 11 309 438 3722; blood flow. Warm-sensitive neurons increase firing when local
e-mail: [email protected] tissue is warmed; likewise, cold-sensitive neurons actively
1
The Central Illinois Neuroscience Foundation, Bloomington, IL, USA respond to local cooling. The warm-sensitive neurons, in
2
Department of Biological Science, Illinois State University, Normal, addition to responding to local warming, are generally excited
IL, USA
by warming of the skin or spinal cord and are inhibited by
Conflict of interest: None. cooling of skin or spinal cord. The cold-sensitive neurons
Funding: This work is supported by CINF career initiative funding and exhibit opposite behaviors. Thus, these neurons serve to
Katter funding. integrate thermal information from the periphery with that

& 2009 The Authors.

274 & 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284
 C. X. Wang et al. Topical review
from the brain and act to maintain the normal range of body stroke insult, body temperature independently relates to initial
temperature (9). stroke severity, infarct size, mortality and poor outcome.
Hyperthermia refers to a pathological increase in body For each 11C increase in body temperature the odds of
temperature when the thermostatic centre in the hypothala- poor outcome rise by 22. To examine whether this relationship
mus is unable to compensate for temperature alteration (10 was also valid in patients admitted at later time points, the
12). Hyperthermia occurs when the body produces more heat researchers extended the analyses to patients admitted between
than it can dissipate, and it has been observed in both stroke 6 and 24 h after stroke onset (18). Body temperature was
patients and animals with experimentally induced ischaemia. found to independently and significantly related to stroke
Several mechanisms may contribute to increasing body tem- severity and mortality in patients admitted between 6 and
perature. Hyperthermia can result from damage to the ther- 12 h, but not in patients admitted between 12 and 24 h from
mostatic centre when infarction is concentrated in the stroke onset. Mortality is higher at 60-month poststroke in
hypothalamus. Support for this hypothesis is evidenced in patients presenting with hyperthermia in the initial phase of
preclinical studies demonstrating that spontaneous increases onset (o6 h), suggesting that hyperthermia may have long-
in body temperature occur in experimental animals following term effects (19). That hyperthermia correlates with poor
cerebral ischaemia (13). Peripheral infection can also stroke outcome is also observed by other groups (13, 20, 21).
cause hyperthermia, especially when the patient is comatose Findings from several retrospective studies also support that
(1416). hyperthermia correlates with higher mortality, more severe
neurological deficits and less favourable final outcomes (15,
Clinical studies 16, 22, 23).
In a study of 725 patients, Boysen and Christensen (12)
examined the relationship between body temperature and
Correlation between hyperthermia and stroke
outcome, at various time points, in patients with both
outcome
ischaemic and haemorrhagic stroke. No correlation was found
The relationship between body temperature on admission and in patients admitted within 2 h of stroke onset and outcome at
stroke severity was investigated in a prospective study of 3 months. For patients admitted within 6 h of stroke onset,
390 patients with acute stroke (17) (Table 1). Within 6 h of lower body temperature was associated with less favourable

Table 1 Hyperthermia and ischaemic stroke in clinical studies

Patient BT recorded (h Hyperthermia

Type (number) post ictus) (%) Major findings References

(A) Prospective
390 6 25 Hyperthermia worsens mortality and outcome. BT correlates with initial (17, 19)
stroke severity, lesion size, mortality and outcome in survivors. For 11C
increase in BT, the relative risk of poor outcome rises by 22
398 12 and 24 N/G BTcorrelates to stroke severity and mortality in patients admitted between (18)
612 h from stroke onset, but not admitted 1224 h
260 72 608 Hyperthermia initiated in 24 h from insult associated with poor outcome (14)
and large infarcts. The earlier the hyperthermia occurs, the higher relation
between the BT increase and brain damage
725 24 N/G At 1012 h after stroke onset, increased BT relates to stroke severity and (12)w
poor outcome. Initial increased BT (o8 h) does not relate to stroke severity
and outcome
183 168 43 Higher BT correlates with poor stroke outcome (20)w
229 24 376 Hyperthermia correlates with higher infarct volume, poor outcome (21)
(B) Retrospective
346 72 N/G Higher BT correlates with larger infarct volume and more severe neuro- (22)
logical deficits
100 24 53 Hyperthermia associated with unfavourable outcome following throm- (24)
bolysis with tPA
150 72 31 Hyperthermia correlated to mortality (15)
3790 N/G N/G Hyperthermia correlated to morbidity and mortality (25)
509 N/G N/G Hyperthermia associated with both short- and long-term mortality. 11C (23)w
increase in BT increases relative risk of 1-year mortality by 34
119 48 32 Increased BT associated with more severe neurological deficits (16)
Hyperthermia occurred in the recording period.wIncudes ischaemic and haemorrhagic stroke. N/G, data not given; BT, body temperature.

& 2009 The Authors.

& 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284 275
 Topical review C. X. Wang et al.

outcome. However, starting at 8 h and for every 2-h interval Preclinical studies
over the next 10 h, higher body temperatures correlated with
poor outcomes.
Focal ischaemia
Because clinical studies with small numbers of
patients often provide conflicting results, Hajat et al. (25),
conducted a retrospective meta-analysis using published data. Elevation of body temperature aggravates ischaemic brain
Analyses from the data of 3790 patients revealed that injury
hyperthermia is highly associated with both morbidity The deleterious effects of hyperthermia in ischaemic brain
(Po00001) and mortality (Po000000001). These data injury are more clearly demonstrated in animal models (Table
clearly demonstrate that hyperthermia is harmful to the 2). In a permanent model of focal cerebral ischaemia the effects
injured brain. of hyperthermia on ischaemic injury were examined. Chen
et al. (26) reported that when the brain temperature was
increased to 401C either 1 h before or immediately after
ischaemic insult and maintained for a period of 1 h after the
Hyperthermia and thrombolysis
injury, infarct volume, measured 4 days after the injury, was
In 100 consecutive patients treated with tPA for acute significantly larger in the animals receiving hyperthermia
stroke, patients with unfavourable outcomes had elevated treatment than in normothermic rats.
body temperatures when compared to patients with favourable The effects of increased brain temperature were also com-
responses to thrombolytic therapy (24). After adjustment for pared between transient and permanent models of ischaemic
baseline characteristics, the presence of hyperthermia was injury (28). In the transient model, inducing 391C hyperther-
significantly associated with a reduced probability of better mia for 2 h, starting immediately after middle cerebral
outcome following thrombolytic treatment. artery (MCA) occlusion, significantly increased infarct volume
in the cortex after a survival period of 3 days. This treatment
does not result in significant changes to infarct volume
in the permanent model. Further analyses show that in the
Hyperthermia and infection
transient model, the infarct volume positively correlates to
Hyperthermia with or without an infectious cause was cerebral blood flow during the first 30 min of recirculation.
also studied in stroke patients (1416). In a study of 158 Brain temperature also positively correlates to cerebral
stroke patients with hyperthermia within first 72 h, an blood flow during MCA occlusion and the early recirculation
infectious cause was found in 91 patients (576%). Broncho- period. These results suggest the net effect of hyperthermia is
pulmonary infection was identified in 47 patients, urinary harmful to the injured brain even with improved local blood
infection in 40 patients and thrombophlebitis in nine flow.
patients. Body temperature and mortality rate are higher To investigate whether delayed hyperthermia has detrimen-
in patients with hyperthermia of infectious aetiology tal effects on the pathological outcome of focal ischaemia,
than those with hyperthermia of noninfectious origin in brain temperature was increased to 39401C at 24 h after MCA
all time periods studied (up to 72 h after stroke onset). occlusion. Infarct volumes were measured on day 4 after MCA
Hyperthermia commencing within the first 24 h from occlusion. Delayed hyperthermia extended ischaemic injury
stroke onset independently relates to larger infarct volumes, from the caudoputamen to the overlying cortex in most
increased neurological deficits and functional dependency at animals with brain temperatures of 401C. The cortical infarct
3 months. volume increased by 66-fold and total infarct volume by
In summary, convincing evidence supports that hyperther- threefold following hyperthermia treatment. In addition, this
mia correlates with increased mortality, severe neurological treatment also worsened neurological deficits. However, treat-
deficits and poor outcome. While randomised-controlled ment with 391C did not exacerbate the brain damage, indicat-
clinical studies are needed to verify whether hypothermia can ing both the timing and degree of temperature modulation are
reduce ischaemic damage, currently, the main efforts should be important factors in influencing the thermal effects on ischae-
directed towards an immediate and effective reduction of body mic outcome (27).
temperature when it is higher than 3751C, particularly in the Recently, we have also studied hyperthermia and ischaemic
early phase of stroke as high temperature in this period brain injury using a focal embolic model of stroke in rats
independently contributes to poor prognosis. Further, studies (8, 27). In these studies, body temperature was elevated before
demonstrate that infection, particularly pulmonary and ur- and continued for 3 h after the ischaemic induction.
inary, is a frequent complication and a major reason for Hyperthermia (391C) worsened neurological deficits, and
increased body temperature in acute stroke patients. Clinicians significantly increased infarct volume, measured at 48 h
should always search for an infectious origin in acute stroke after the MCA occlusion. In addition, the mortality rate was
patients and any infection should be controlled properly and also significantly higher in the hyperthermic rats than in
effectively. normothermic rats. Findings from a second study showed

& 2009 The Authors.

276 & 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284
 C. X. Wang et al. Topical review
Table 2 Hyperthermia and ischaemic stroke in preclinical studies

Hyperther- Initiate (Duration)

Model Species mia (1C) (h post ischaemia) Major findings References

(A) Focal
AI. Brain temperature
Rat 3940 24 (3) Hyperthermia 401C increases infarct volume by 3 fold at 4 days (27)
post-MCA occlusion, but not 391C
Rat 40 1 prior or 0 (12) Hyperthermia, starting at both 1 h before or 1 h after ischaemia, (26)
increases infarct volume
Rat 39 0 (2) Hyperthermia increases infarct volume in transient MCA occlusion (28)
model, but not in permanent MCA occlusion model
Rat 39 0 (25) Hyperthermia increases extracellular glutamate release (29)
AII. Rectal temperature
Rat 39 05 (2) Hyperthermia enhances tPA-induced recanalisation and increases (30)
infarct volume
Rat N/A N/A Spontaneous hyperthermia occurs with hypothalamic injury. (31)
Spontaneous hyperthermia (48 h after ischaemia) does not
increase infarct volume
Rat N/A N/A Spontaneous hyperthermia increases infarct volume, antagonises (32)
beneficial effects of reperfusion and accelerates penumbra to
infarct
Rat N/A N/A Spontaneous hyperthermia nullifies the neuroprotective actions (7)
of NMDA antagonist
Rat 3839 03 prior (3) Hyperthermia increases infarct volume, worsens neurological (8, 33)
deficits, antagonises actions of thrombolytic agent
Rat N/A N/A Selective hypothalamic infarction produces significant and sus- (13)
tained spontaneous hyperthermia
(B) Global
BI. Brain muscle temperature
Gerbil 39 0 prior (5 min) Hyperthermia exacerbates ischaemia-induced cell death, causes (34)
complete destruction of neurons in CA1
BII. Temporalis muscle temperature
Dog 3839 03 prior (1) Small increase of temperature worsens neurological function and (35)
severity of injury
Rat 39 0 (03) Hyperthermia enhances ischaemic brain damage and mortality, (36)
fosters transformation of ischaemic injury into infarction, accel-
erates morphological appearance of ischaemic brain injury
BIII. Skull temperature
Rat 39 Prior (5, 10, or Increase of temperature by 21C enhances brain damage (37)
15 min)
BIV. Rectal temperature
Rat N/A N/A Spontaneous hyperthermia (43851C) occurs from 21 to 63 h (38)
following ischaemia. Depressing body temperature to normother-
mic diminishes neuronal damage
Rat 396 24 (3) Delayed hyperthermia exacerbates ischaemic neuronal injury (39)
Before ischaemic induction.

that hyperthermia (391C) exacerbated brain injury which is 1520 min after MCA occlusion and reaches 394051C during
consistent with the first study. Furthermore, the second study the first hour. Sustained hyperthermia is observed during the
also showed that treatment with hyperthermia (381C) in- rest of the first 24 h. Spontaneous hyperthermia has also been
creased infarct volume, ischaemic brain oedema and seizure observed by several other groups (31, 40).
activities compared with normothermic rats. Spontaneous hyperthermia increases infarct volume in both
permanent and transient MCA occlusion (32), which is due to
Spontaneous hyperthermia accelerated transformation of salvageable tissues to infarct.
Spontaneous hyperthermia has been examined in several The peripheral zone of a focally ischaemic region, the so-called
studies using focal models of ischaemic brain injury. Reglodi ischaemic penumbra, is known to be a temperature-sensitive
et al. (32) reported that hyperthermia occurs in rats experien- area. In 2,3,5-triphenyltetrazolium chloride-stained brain sec-
cing transient ischaemia with durations of 90 min, 2 h tions, the border zone between normal (red) and infarct
and permanent ischaemia. Temperature starts to increase at (white) zones are separated by tissue with pink colour. These

& 2009 The Authors.

& 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284 277
 Topical review C. X. Wang et al.

pink tissues are an early indicator of ischaemia and are also been demonstrated (30). Despite better re-canalisation of the
believed to represent the penumbra region. Compared with the occluded arteries, treatment with tPA enlarged the size of
normothermic controls, the infarct size increases and penum- infarcted tissue in the hyperthermic animals. These results
bra area decreases significantly at 4, 12 and 48 h in the suggest that hyperthermia can convert an effective treatment to
hyperthermic animals, indicating the transformation of pe- harmful toxic agent in the injured brain. Based on these results,
numbra to infarct. authors of this study have suggested that hyperthermia should
To examine whether the hypothalamus, the thermoregulat- be avoided in clinical trials of thrombolytic therapy and in
ing centre, is responsible for spontaneous hyperthermia, patients with acute ischaemic stroke in general. This study is
temperature variations were recorded when ischaemic damage highly significant since thrombolytic therapy with tPA is
was induced only in hypothalamus or in both neocortex presently the only proven effective pharmacological treatment
(MCA territory) and nonneocortical structures (13). Rats for acute stroke patients.
with hypothalamic infarction exhibit persistent hyperthermia We also studied whether hyperthermia eliminates the
for 72 h, whereas rats with infarction at both neocortex beneficial effects of tPA treatment in an embolic model of
and nonneocortical structures display transient hyperthermia. cerebral ischaemia (8). We found that hyperthermia signifi-
In addition, a lesion involving the medial hypothalamus cantly increased infarct volume and brain oedema. Treatment
alone, by occlusion of tuberal artery, causes hyperthermia with tPA significantly reduced infarct volume in normother-
irrespective of the obstruction of the anterior cerebral artery mic and 381C rats. However, this treatment did not reduce the
or anterior choroidal artery. These results provide infarct volume when the body temperature was increased to
evidence that hypothalamus is involved in the occurrence of 391C. Treatment with tPA reduced brain oedema only in the
hyperthermia. 381C group but not in the 391C group. Neurological deficit
scores were significantly higher in the hyperthermic rats than
Hyperthermia affects the efficacy of neuroprotective and in the normothermic rats. Treatment with tPA improved
thrombolytic agents neurological deficits in the 381C group but not in the 391C
The noncompetitive NMDA antagonist dizocilpine maleate group. In addition, in vitro experiments showed that hy-
(MK-801) is an effective treatment to reduce infarction in perthermia increased the fibrinolytic activity of tPA. Thus,
animal models of focal ischaemia. Unexpectedly, it was in- these data clearly show that hyperthermia abolishes the
effective in a transient model, occluding the MCA with an therapeutic actions of thrombolytic treatment with tPA, even
intra-arterial suture (7). In exploring the possible reasons for though the fibrinolytic activity of tPA is increased with
this difference among models, researchers found that sponta- increasing temperature. These findings indicate that the
neous hyperthermia occurred following MCA occlusion. To increased fibrinolytic activity of tPA is offset by its deleterious
examine whether hyperthermia nullified the protective effects actions; and the net effect of hyperthermia in ischaemic brain
of MK-801, ischaemic injured rats were divided into groups in injury is neurodestructive.
which body temperature was allowed to increase sponta-
neously (393951C) and groups in which body temperature
was maintained within normal range during ischaemia. MK-
Global ischaemia
801 failed to reduce infarct size in animals with spontaneous
hyperthermia. In contrast, this drug treatment markedly The findings from global model of ischaemic brain
reduced infarct volume in animals when the body temperature injury are consistent with those from focal models. In a global
was controlled. These results indicate that amelioration of model of cerebral ischaemia induced by vessel occlusion
focal ischaemic damage cannot be expected if hyperthermia is combined with systemic hypotension in rats, increased
allowed to occur. brain temperature accelerated ischaemic injured cells
Hyperthermia also blunts the beneficial action of reperfu- to necrotic death (36, 37). Mortality rates were also increased
sion in ischaemic brain injury (32). In a suture model of in animals with increased brain temperature (36). Sub-
ischaemic brain injury, reperfusion is achieved by removing sequent study reveals that transient hyperthermia, even
the intra-arterial nylon suture. A comparison between animals imposed 1 day following a brief episode of global ischaemic
with different reperfusion times revealed that infarct sizes were insult, exacerbates the extent of ischaemic neuronal
smaller when reperfusion occurs at 90 min than 120 min, even injury (39). Conversely, prevention of transient postischaemic
in animals with postischaemic spontaneous hyperthermia. hyperthermia reduces the extent of neuronal damage (37).
However, infarct sizes of the hyperthermic rats with 120-min In gerbils, hyperthermia results in accentuated ischaemia-
transient ischaemia exceed those of the normothermic rats induced loss of CA1 neurons in hippocampus. Increasing
with permanent ischaemia, indicating hyperthermia abolishes cerebral temperature to 391C completely destroys the
the beneficial actions of reperfusion. neurons in CA1 regions and also produces an extension of
In a focal embolic model of ischaemic brain injury a positive ischaemic neuropathology to regions CA2, CA3 and CA4
relationship between body temperature during the initial where there is no significant loss of neurons in normothermic
phase of infarct development and final infarct volume has animals (34).

& 2009 The Authors.

278 & 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284
 C. X. Wang et al. Topical review
Occurrence of spontaneous hyperthermia is also observed neurotransmitters, damage of the microvessels and cytoskele-
in the global model of ischaemia (38). In rats subjected to tal structures and activation of death signals.
10 min of transient ischaemia induced by occlusion of com-
mon carotid arteries in combination with hypotension, body
Metabolic functions
temperature persistently increased to above 3851C starting at
1 day from reperfusion and continuing for 3 days. Depressing In a model of global cerebral ischaemia in cats, the effects of
body temperature to normothermic range by the antipyretic mild whole body hyperthermia were investigated on metabolic
agent, dipyrone, markedly diminished neuronal damage in the recovery during recirculation using in vivo phosphorus-31
neocortex and hippocampus when evaluated 7 days after nuclear magnetic resonance spectroscopy (45). Hyperthermia
ischaemia. Cooling the animals to normothermic levels with (4061C) was induced 1 h before ischaemia and maintained
physical therapy was also beneficial for the injured brain. These during 152 h of recirculation. In hyperthermic cats, b-
observations suggest that the processes leading to neuronal adenosine 50 -triphosphate (b-ATP) and phosphocreatine
death are very sensitive to temperature increase, not only (PCr) as well as the ratio of PCr to inorganic phosphate fail
during the ischaemic insults, but also in the late reperfusion to return to preischaemic levels during recirculation, in
phase, which is relevant to the design of clinical therapeutic contrast to normothermic cats. Hyperthermia also caused
treatment as stroke patients usually arrive at the hospital a few lower intracellular cerebral pH during recirculation than in
hours following the insult. the normothermic cats. These data suggest mild hyperthermia,
Hyperthermia affecting functional changes and histological concurrent with global cerebral ischaemia in cats, worsens
changes has also been studied in a global model of cerebral metabolic and physiological functions.
ischaemia using large animals (35). Complete cerebral ischae- Following reductions in cerebral blood flow and oxygen
mia, 125 min in duration, is produced by arterial hypotension supply, high-energy phosphorylated compounds, such as
plus intracranial hypertension. Compared with the controls ATP and PCr, are reduced and tissue acidosis occurs. On the
(371C), dogs with increased brain temperature of 1 or 21C other hand, hyperthermia, if present, increases energy utilisa-
exhibit significantly worse neurologic function during a tion. Studies using newborn swine have shown a linear
72-h postischaemic observation period. At the end of the relationship between brain energy utilisation and brain tem-
experiments, 72 h after ischaemia or at the time of ischaemia- perature over a range of 28411C. An elevation in brain
related death, the severity of injury was determined histologi- temperature increases energy utilisation by 53% per 11C
cally. Structures rich in grey matter were more susceptible (48). Thus, increased energy utilisation further depletes energy
to ischaemic injury and more likely to be influenced by substrates and cultivates accumulation of ions and toxic
temperature increases. When compared with the 371C con- metabolites, which can increase the brains vulnerability to
trols, significantly increased injury was observed in both injury under ischaemic conditions. Furthermore, ischaemia
381C and 391C dogs. The severity of injury was also denatures proteins and leads to protein aggregation within
higher in the 391C dogs than in the 381C dogs. Moreover, neurons (49). Renaturing of proteins by a cellular chaperone
mortality rates also increased in dogs with increased body system requires ATP. Therefore, protein denaturation may
temperature. cause damaging effects to the injured brain when energy stress
In summary, compelling evidence from both focal and is combined.
global models of ischaemic cerebral injury demonstrates that
the brain is very sensitive to temperature changes. Increased
Free radicals
brain temperature, either spontaneous or extrinsically in-
duced, worsens neuronal function, accelerates injury pro- Hyperthermia enhances ischaemia-induced increases of free
cesses, converts salvageable tissue to macro-infarct and radicals. In a model of global ischaemic brain injury, the levels
aggravates stroke outcome. If these observations are transfer- of hydroxyl radicals were examined using a method based on
able to stroke patients, these data strongly suggest hyperther- chemical trapping in the form of stable adducts 2,3- and 2,5-
mia is harmful to the injured brain and hyperthermia should dihydroxybenzoic acid following salicylate administration (42,
be corrected efficiently and effectively. 43). In normothermic rats, the levels of hydroxyl radicals
decreased during ischaemic injury and significantly increased
during reperfusion. Hyperthermia treatment results in even
Possible mechanisms of increased cerebral
higher increases of hydroxyl radicals. In addition, results also
damage by hyperthermia
show that these hydroxyl radicals arise within brain parench-
Currently, there is no comprehensive explanation for why yma itself.
hyperthermia exacerbates damage to neuronal cells following Free radicals are extremely active compounds that can
ischaemic brain injury, but it is likely due to multiple mechan- interact with lipids, enzymes and DNA to produce a variety
isms working in combination (Table 3, Fig. 1). These mechan- of harmful actions. The brain is particularly vulnerable to
isms include metabolic changes, free radical production, oxidative damage because it is rich in unsaturated fatty acids
increased inflammatory reactions and extracellular excitatory and relatively poor in antioxidant defenses. The injurious role

& 2009 The Authors.

& 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284 279
 Topical review C. X. Wang et al.

Table 3 Possible mechanisms of hyperthermia worsening stroke outcome

Hyperthermia Initiate (Duration)

Model Species (1C) (h post ischaemia) Major findings References

(I) Preclinical studies

A. Focal
Rat 39 0 (1) Hyperthermia increases degradation of cytoskeletal proteins by (41)
enhancing calpain activities
Rat 3839 03 prior (3) Hyperthermia compromises microcirculation, damages BBB (8)
B. Global
Rat 39 3 min prior (7 min) Hyperthermia increases production of free radicals (42, 43)
Rat 39 5 min (15 min) Hyperthermia enhances the reduction of PKC activities (44)
Cat 406 1 prior (152) Hyperthermia results in lower intracellular pH, and b-ATP phos- (45)
phocreatine
Gerbil 39 Prior (5 min) Raising intracerebral temperature increases the inhibition of CaM (34)
kinase II, and aggravates the loss of MAP 5 2
Rat 39 0 (03) Hyperthermia increases permeability of microvessels (46)
(II) Clinical studies
Hyperthermia correlates with higher plasmatic pro-inflammatory (21)
cytokines
Hyperthermia increases glutamate and glycine in the CSF in stroke (47)
patients

CSF, cerebrospinal fluid; MAP, microtubule-associated protein.

of these radicals in the setting of ischaemia is evident from pyrogenic. Thus, hyperthermia may be part of a stroke-
studies demonstrating that free radical scavengers confer induced inflammatory reaction (11).
protection from ischaemic damage (5).
Excitatory neurotransmitters
Hyperthermia exacerbates ischaemic brain injury by modify-
Inflammatory reactions
ing the release of excitatory neurotransmitters. Clinical study
A clinical study involving 214 stroke patients was carried out to reveals that the concentrations of glutamate and glycine in the
determine the relationship between pro-inflammatory cyto- cerebrospinal fluid increase significantly in patients with
kines and hyperthermia as it relates to larger infarct size (21). hyperthermia (47). Higher body temperature also relates to
Cytokines were detected at admission and also 48 h later. worsening neurological deficits and larger infarct volume. In
Results from this study show that the plasmatic levels of preclinical research, the hyperthermia-enhanced release of
interleukin (IL)-6 and tumour necrosis factor a (TNFa), glutamate has been investigated in the ischaemic injured brain
measured at admission, significantly correlate to infarct vo- by measuring extracellular concentration of glutamate using
lume. The plasmatic levels of IL-6, TNFa, intercellular adhe- intra-cortical microdialysis (29). In hyperthermic rats (391C),
sion molecule (ICAM)-1 and vascular cell adhesion molecule- peak glutamate concentration in the cortex during MCA
1, measured at 48 h after the admission, also correlate to infarct occlusion is 31-fold above baseline, compared with 65-fold
volume. Moreover, a significant association was found be- elevations in normothermic rats. Significantly increased glu-
tween the cytokines (IL-6, TNFa and ICAM-1) and poor stroke tamate release is also observed in the brain following global
outcome. ischaemia (6).
Preclinical studies have also shown that inflammatory Glutamate is the major endogenous excitatory neurotrans-
reactions occur in the ischaemic injured brain (50). These mitter in the central nervous system (51). Increased release of
reactions are characterised by a sequential series of processes, these neurotransmitters causes neurotoxicity through an ex-
including the release of pro-inflammatory cytokines, increased cessive activation of pro-synaptic glutamate receptors, which
expression of endothelial adhesion molecules and chemotactic leads to an increase in intracellular free calcium ion concen-
factors, activation of microglia and macrophages, and leuco- tration, triggering a cascade of events which eventuate in
cyte infiltration. Overwhelming evidence supports that in- neuronal death.
flammatory reactions contribute considerably to the
pathogenesis of ischaemic brain injury through the increase
Microcirculatory and bloodbrain barrier (BBB)
of secondary expansion of ischaemic infarction. In addition,
pro-inflammatory cytokines, TNFa, IL-6 and IL-1b, which are The permeability of the BBB has been examined in the global
pivotal for the inflammatory response, are known to be model of cerebral ischaemia, by evaluating the leakage of

& 2009 The Authors.

280 & 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284
 C. X. Wang et al. Topical review

Metabolic
Changes

Free
Radicals

Inflammation

Ischemia Glutamate Neuronal

Death
Microvessel

Cytoskelaton

Apoptotic
Signal

Fig. 1 Mechanisms of hyperthermia-enhanced neuronal death in the ischaemic injured brain.

horseradish peroxidase (46). Hyperthermia (391C), instituted occlusion, indicating that hyperthermia compromises the
during the 20-min ischaemic induction, significantly increased microcirculation in the ischaemic injured brain. To examine
the BBB breach in the CA1 area of the hippocampus and the effects of hyperthermia on BBB damage, concentrations of
ventrolateral thalamus. Importantly, the sites of increased extravagated Evan blue dye were measured in the striatum, the
vascular permeability were spatially correlated with dark cortex above the striatum and the cortex more peripheral to
shrunken (damaged) neurons, indicating hyperthermia may the striatum. The extravasations of Evan blue dye were
enhance neuronal death through BBB damage. significantly increased in both the peripheral cortex and in
We have examined whether hyperthermia affects the micro- the striatum in the rats with hyperthermia (391C).
circulation, and BBB permeability, in an embolic model of
focal ischaemia (8, 52). In these studies, perfusion deficits were
Cytoskeleton damage
revealed using Evan blue staining and BBB permeability was
evaluated by detecting extravagated Evan blue dye. Compared Spectrin is a cytoskeleton protein; it is degraded by activated
with normothermic rats, perfusion deficits in the hyperther- calpain, a calcium-sensitive cysteine protease. Immunohisto-
mic rats (391C) were significantly larger at 3 and 6 h after MCA chemical localisations of spectrin breakdown products have

& 2009 The Authors.

& 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284 281
 Topical review C. X. Wang et al.

been studied using an MCA occlusion model (41). In nor- Summary and future directions
mothermic rats spectrin immunoreactivity is present occa-
sionally in the cortical neurons at 1 h; sparse spectrin Findings of clinical studies support that hyperthermia relates
immunoreactivity appears at 4 h after reperfusion. No spectrin to poor stroke outcome and increased mortality. Preclinical
immunoreactivity could be detected at 24 h after reperfusion. studies manipulating brain temperature have provided more
In contrast, moderate to intensive spectrin immunostaining is direct and clear evidence that hyperthermia is detrimental to
present in the cortical neurons at 1 h after reperfusion in the ischaemic injured brain. These findings have prompted
hyperthermic animals. At 4 and 24 h, most brains exhibit dense investigators to plan clinical trials on cooling therapy with
immunoreactivity associated with morphologically shrunken physical or pharmacological treatments in acute stroke pa-
neurons. These data support that hyperthermia is involved in tients with the aim of improving stroke outcomes.
the damage of cytoskeleton during ischaemic brain injury. Acetaminophen is a safe, clinically proven antipyretic that
Microtubule-associated protein (MAP)-2, a cytoskeleton- acts at the hypothalamic thermal regulating centre. In a
related protein, selectively associates with neuronal soma and randomised-controlled trial, early administration of acetami-
dendrites. Microtubule-associated protein-2 regulates the sta- nophen, 39 g/day, to afebrile acute stroke patients resulted in
bility of microtubules and also mediates the interaction of very modest, but not significant, reduction in body tempera-
cytoskeletons and other cell components. Changes of MAP-2 ture (54). However, treatment with a daily dose of 6 g of
have been evaluated with immunohistochemistry using a acetaminophen significantly lowers body temperature by
global model of ischaemia in gerbils (53). Compared with 041C in acute stroke patients (55). Further studies are needed
sham-operated animals, MAP-2 immunostaining significantly to determine whether this small reduction in body tempera-
decreased at 48 h following ischaemic injury. However, com- ture leads to improved outcome.
pared with the normothermic animals, MAP-2 immunostain- Owing to its anti-platelet properties, aspirin is recom-
ing was significantly lower in the hyperthermic animals at 24 h, mended treatment in ischaemic stroke as soon as a CT or
but not at 48 h, after ischaemia, indicating that hyperthermia MRI scan has ruled out intracerebral haemorrhage (11).
accelerates enzyme activities responsible for MAP-2 degrada- Preclinical studies also show that aspirin prevents glutamate-
tion in the injured brain. These results suggest that hyperther- induced neurotoxicity, reduces inflammatory reactions, im-
mia may foster the disassembly of microtubules and proves microcirculation and protects the ischaemic injured
destruction of cytoskeleton structures, which in turn contri- brain (56, 57). These data support aspirin as an ideal anti-
butes to neuronal death. pyretic medication in every case of hyperthermia, if the cause
of ischaemic stroke is not from a cardio-embolic origin and no
other general contraindications exist. However, more research
Kinase activities and other signal transduction
is needed to examine whether treatment with aspirin is
proteins
effective in stroke patients with hyperthermia.
The effects of hyperthermia on the activities of two kinases, Physical cooling therapy has also been evaluated to reduce
CaM kinase II and protein kinase C (PKC), have been studied body temperature. Normal range of body temperature (36
in the ischaemic injured brain. CaM kinase II activity is 371C) can be attained in hyperthermic stroke patients within
significantly inhibited in the brain following ischaemia (34). 33 h by application of a water-perfused cooling mattress (58).
Hyperthermia intensifies ischaemia-induced inhibition of An even lower body temperature can be achieved with more
CaM kinase II activity. CaM kinase II plays an important intensive physical cooling therapy. However, physical cooling
role in the transient increase of intracellular Ca21. This kinase is limited by its potential adverse effects including cardiac
is predominantly expressed in neurons, and comprises 2% of arrhythmia, metabolic derangements and propensity for in-
total hippocampal proteins. The alteration in CaM kinase II fection. Additionally, physical cooling also requires sedatives
caused by ischaemia may trigger changes in calcium-regulated and muscle relaxation adjuvants.
processes, such as ion fluxes, transmitter release and cell Currently, no evidence exists from randomised trials to
transport mechanisms which eventually lead to neuronal support the routine use of pharmacological or physical cooling
death. in acute stroke. Given the high incidence of hyperthermia in
In the vulnerable brain regions, ischaemia induces signifi- stroke patients, and compelling evidence suggesting that
cant reductions in PKC activities, which are irreversible for up hyperthermia has deleterious effects in stroke outcome, multi-
to 24 h after reperfusion (44). Similar to the changes in CaM centre randomised-controlled clinical trials are worthwhile to
kinase II, hyperthermia also leads to further reduction of PKC examine whether the prevention of hyperthermia by prophy-
activities. Upon activation, PKC enzymes are translocated to lactic therapy can lead to an improvement in stroke outcome.
the plasma membrane. The activated PKC can phosphorylate Although the present article intends to review the research
many types of proteins, such as cytoskeleton protein, and thus progress during recent decade, two pioneer studies on hy-
alter their functions. Therefore, hyperthermia-induced alter- perthermia and stroke are worth mentioning. The first retro-
nation of PKC activities may contribute to the secondary spective study was reported by Hindfelt in 1976 (59). In this
injury processes in the brain following ischaemia. study, prognostic influence of hyperthermia was analysed in

& 2009 The Authors.

282 & 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284
 C. X. Wang et al. Topical review
110 patients with varying neurological disabilities. Results 8 Noor R, Wang CX, Shuaib A. Hyperthermia masks the neuroprotective
from this study support that hyperthermia correlates to an effects of tissue plasminogen activator. Stroke 2005; 36:6659.
9 Kupferman I. Hypothalamus and limbic system; in Kandel ER,
unfavourable prognosis with respect to residual symptoms. Schwartz JH, Jessell TM (eds): Principle of Neural Science, 3rd edn.
Based on this observation, the author suggests that hyperther- Norwalk, USA: Appleton & Lange, 1991: 75060.
mia, irrespective of its genesis, should be intensely combated 10 Di Rosa AE, Morgante L, Spina E, Meduri M. Epidemiology and
during early stages of ischaemic stroke. In 1994, Castillo et al. pathoetiology of neurological syndromes with hyperthermia. Funct
Neurol 1995; 10:1119.
(60) reported the first prospective clinical trial. In this study,
11 Zaremba J. Hyperthermia in ischemic stroke. Med Sci Monit 2004;
177 patients with ischaemic stroke are analysed. Results 10:RA14853.
demonstrated that hyperthermia is associated with poor out- 12 Boysen G, Christensen H. Stroke severity determines body temperature
comes in patients with ischaemic stroke. Both temperature and in acute stroke. Stroke 2001; 32:4137.
glucose levels were significantly higher in the patients that died 13 He Z, Yamawaki T, Yang S, Day AL, Simpkins JW, Naritomi H.
than in the surviving patients during a 6-month follow-up Experimental model of small deep infarcts involving the hypothalamus
in rats: changes in body temperature and postural reflex. Stroke 1999;
period; however, further analyses indicate that temperature is 30:274351.
the only variable that has a significant influence on morbidity. 14 Castillo J, Davalos A, Marrugat J, Noya M. Timing for fever-related
More than 30 years have passed since the first retrospective brain damage in acute ischemic stroke. Stroke 1998; 29:245560.
study on hyperthermia and ischaemic stroke was reported. 15 Seo WK, Yu SW, Kim JH, Park KW, Koh SB. The impact of
hyperthermia and infection on acute ischemic stroke patients in the
Only recently has attention been focused on the actions of
intensive care unit. Neurocrit Care 2008; 9:1838.
hyperthermia, in comparison to the large amount of research 16 Grau AJ, Buggle F, Schnitzler P, Spiel M, Lichy C, Hacke W. Fever
that has been conducted on hypothermia. Additionally, little and infection early after ischemic stroke. J Neurol Sci 1999; 171:
attention has been devoted to research towards understanding 11520.
the cellular and molecular mechanisms of hyperthermia- 17 Reith J, Jorgensen HS, Pedersen PM et al. Body temperature in acute
enhanced injury processes. Although research has shown stroke: relation to stroke severity, infarct size, mortality and outcome.
Lancet 1996; 347:4225.
that hyperthermia affects changes in some biological markers, 18 Jorgensen HS, Reith J, Pedersen PM, Nakayama H, Olsen TS. Body
it is not clear if these changes play significant roles in neuronal temperature and outcome in stroke patients. Lancet 1996; 348:193.
death or if they are simply the epiphenomena of cell injury. 19 Kammersgaard LP, J^rgensen HS, Rungby JA et al. Admission body
Further studies to identify these cellular and molecular me- temperature predicts long-term mortality after acute stroke: the
chanisms will provide important information not only for Copenhagen Stroke Study. Stroke 2002; 33:175962.
20 Azzimondi G, Bassein L, Nonino F et al. Fever in acute stroke worsens
mechanisms of hyperthermia-enhanced neuronal death but prognosis. A prospective study. Stroke 1995; 26:20403.
also for the fundamental mechanisms of ischaemic injury. This 21 Leira R, Rodrguez-Yanez M, Castellanos M et al. Hyperthermia is a
knowledge will certainly aid in the design of neuroprotective surrogate marker of inflammation-mediated cause of brain damage in
strategies for the treatment of stroke patients. acute ischaemic stroke. J Intern Med 2006; 260:3439.
22 Audebert HJ, Rott MM, Eck T, Haberl RL. Systemic inflammatory
response depends on initial stroke severity but is attenuated by
Acknowledgements successful thrombolysis. Stroke 2004; 35:212833.
23 Wang Y, Lim LL, Levi C, Heller RF, Fisher J. Influence of admission
Authors are grateful for the careful revision of the manuscript body temperature on stroke mortality. Stroke 2000; 31:4049.
by Jennifer Troyanovich and Mira Coleman, technical support 24 Ernon L, Schrooten M, Thijs V. Body temperature and outcome after
from Leslie Campbell, and graphic design by Judith Soliman stroke thrombolysis. Acta Neurol Scand 2006; 114:238.
and John Soliman. 25 Hajat C, Hajat S, Sharma P. Effects of poststroke pyrexia on stroke
outcome: a meta-analysis of studies in patients. Stroke 2000; 31:4104.
26 Chen H, Chopp M, Welch KM. Effect of mild hyperthermia on the
References ischemic infarct volume after middle cerebral artery occlusion in the
rat. Neurology 1991; 41:11335.
1 Green AR, Shuaib A. Therapeutic strategies for the treatment of stroke. 27 Kim Y, Busto R, Dietrich WD, Kraydieh S, Ginsberg MD. Delayed
Drug Discov Today 2006; 11:68193. postischemic hyperthermia in awake rats worsens the histopathological
2 He J, Klag MJ, Wu Z, Whelton PK. Stroke in the Peoples Republic of outcome of transient focal cerebral ischemia. Stroke 1996; 27:227480.
China. I. Geographic variations in incidence and risk factors. Stroke 28 Morikawa E, Ginsberg MD, Dietrich WD et al. The significance of brain
1995; 26:22227. temperature in focal cerebral ischemia: histopathological conse-
3 Editorial: tackling the global burden of stroke. Lancet Neurol 2005; quences of middle cerebral artery occlusion in the rat. J Cereb Blood
4:689. Flow Metab 1992; 12:3809.
4 Hankey GJ. Neuroprotection for acute ischaemic stroke: hope reig- 29 Takagi K, Ginsberg MD, Globus MY, Martinez E, Busto R. Effect of
nited. Lancet Neurol 2006; 5:2878. hyperthermia on glutamate release in ischemic penumbra after middle
5 Wang C, Shuaib A. Neuroprotective effects of free radical scavengers in cerebral artery occlusion in rats. Am J Physiol 1994; 267:H17706.
stroke. Drug Aging 2007; 24:53746. 30 Meden P, Overgaard K, Pedersen H, Boysen G. The influence of body
6 Ginsberg MD, Busto RS. Combating hyperthermia in acute stroke: a temperature on infarct volume and thrombolytic therapy in a rat
significant clinical concern. Stroke 1998; 29:52934. embolic stroke model. Brain Res 1994; 647:1318.
7 Memezawa H, Zhao Q, Smith ML, Siesjo BK. Hyperthermia nullifies 31 Li F, Omae T, Fisher M. Spontaneous hyperthermia and its mechanism
the ameliorating effect of dizocilpine maleate (MK-801) in focal in the intraluminal suture middle cerebral artery occlusion model of
cerebral ischemia. Brain Res 1995; 670:4852. rats. Stroke 1999; 30:246470.

& 2009 The Authors.

& 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284 283
 Topical review C. X. Wang et al.

32 Reglodi D, Somogyvari-Vigh A, Maderdrut JL, Vigh S, Arimura A. 46 Dietrich WD, Halley M, Valdes I, Busto R. Interrelationships between
Postischemic spontaneous hyperthermia and its effects in middle increased vascular permeability and acute neuronal damage following
cerebral artery occlusion in the rat. Exp Neurol 2000; 163:399407. temperature-controlled brain ischemia in rats. Acta Neuropathol 1991;
33 Noor R, Wang CX, Shuaib A. Effects of hyperthermia on infarct volume 81:61525.
in focal embolic model of cerebral ischemia in rats. Neurosci Lett 2003; 47 Castillo J, Davalos A, Noya M. Aggravation of acute ischemic stroke by
349:1302. hyperthermia is related to an excitotoxic mechanism. Cerebrovasc Dis
34 Churn SB, Taft WC, Billingsley MS, Blair RE, DeLorenzo RJ. Tempera- 1999; 9:227.
ture modulation of ischemic neuronal death and inhibition of calcium/ 48 Laptook AR, Corbett RJ. The effects of temperature on hypoxic-
calmodulin-dependent protein kinase II in gerbils. Stroke 1990; ischemic brain injury. Clin Perinatol 2002; 29:62349.
21:171521. 49 White MG, Luca LE, Nonner D et al. Cellular mechanisms of
35 Wass CT, Lanier WL, Hofer RE, Scheithauer BW, Andrews AG. neuronal damage from hyperthermia. Prog Brain Res 2007; 162:
Temperature changes of 4or 5 11C alter functional neurologic out- 34771.
come and histopathology in a canine model of complete cerebral 50 Wang CX, Shuaib A. Involvement of inflammatory cytokines in central
ischemia. Anesthesiology 1995; 83:32535. nervous system injury. Prog Neurobiol 2002; 67:16172.
36 Dietrich WD, Busto R, Valdes I, Loor Y. Effects of normothermic versus 51 Wang C, Shuaib A. NMDA/NR2B selective antagonists in the treatment
mild hyperthermic forebrain ischemia in rats. Stroke 1990; 21:131825. of ischemic brain injury. Curr Drug Target 2005; 4:14351.
37 Minamisawa H, Smith ML, Siesjo BK. The effect of mild hyperthermia 52 Wang CX, Shuaib A. Critical role of microvasculature basal lamina in
and hypothermia on brain damage following 5, 10, and 15 minutes of ischemic brain injury. Prog Neurobiol 2007; 83:1408.
forebrain ischemia. Ann Neurol 1990; 28:2633. 53 Eguchi Y, Yamashita K, Iwamoto T, Ito H. Effects of brain temperature
38 Coimbra C, Boris-Moller F, Drake M, Wieloch T. Diminished neuronal on calmodulin and microtubule-associated protein 2 immunoreactiv-
damage in the rat brain by late treatment with the antipyretic drug ity in the gerbil hippocampus following transient forebrain ischemia.
dipyrone or cooling following cerebral ischemia. Acta Neuropathol J Neurotrauma 1997; 14:10918.
1996; 92:44753. 54 Kasner SE, Wein T, Piriyawat P et al. Acetaminophen for altering body
39 Baena RC, Busto R, Dietrich WD, Globus MY, Ginsberg MD. Hy- temperature in acute stroke: a randomized clinical trial. Stroke 2002;
perthermia delayed by 24 hours aggravates neuronal damage in rat 33:1345.
hippocampus following global ischemia. Neurology 1997; 48:76873. 55 Dippel DW, van Breda EJ, van Gemert HM et al. Effect of paracetamol
40 Zhao Q, Memezawa H, Smith ML, Siesjo BK. Hyperthermia compli- (acetaminophen) on body temperature in acute ischemic stroke:
cates middle cerebral artery occlusion induced by an intraluminal a double-blind, randomized phase II clinical trial. Stroke 2001;
filament. Brain Res 1994; 649:2539. 32:160712.
41 Morimoto T, Ginsberg MD, Dietrich WD, Zhao W. Hyperthermia 56 Grilli M, Pizzi M, Memo M, Spano P. Neuroprotection by aspirin and
enhances spectrin breakdown in transient focal cerebral ischemia. sodium salicylate through blockade of NF-kappaB activation. Science
Brain Res 1997; 746:4351. 1996; 274:13835.
42 Globus MY, Busto R, Lin B, Schnippering H, Ginsberg MD. Detection 57 Uddin G, Hussain M, Wang C, Todd K, Shuaib A. Acetylsalicylic acid
of free radical activity during transient global ischemia and recircula- reduces perfusion deficit in ischemic injured brain in rats. NeuroReport
tion: effects of intraischemic brain temperature modulation. J Neu- 2003; 14:17535.
rochem 1995; 65:12506. 58 Knoll T, Wimmer ML, Gumpinger F, Haberl RL. The low normother-
43 Kil HY, Zhang J, Piantadosi CA. Brain temperature alters hydroxyl mia concept maintaining a core body temperature between 36 and 37
radical production during cerebral ischemia/reperfusion in rats. degrees C in acute stroke unit patients. J Neurosurg Anesthesiol 2002;
J Cereb Blood Flow Metab 1996; 16:1006. 14:2778.
44 Busto R, Globus MY, Neary JT, Ginsberg MD. Regional alterations of 59 Hindfelt B. The prognostic significance of subfebrility and
protein kinase C activity following transient cerebral ischemia: effects fever in ischaemic cerebral infarction. Acta Neurol Scand 1976; 53:
of intraischemic brain temperature modulation. J Neurochem 1994; 729.
63:1095103. 60 Castillo J, Martinez F, Leira R, Prieto JMM, Lema M, Noya M. Mortality
45 Chopp M, Welch KM, Tidwell CD, Knight R, Helpern JA. Effect of mild and morbidity of acute cerebral infarction related to temperature and
hyperthermia on recovery of metabolic function after global cerebral basal analytic parameters. Cerebrovasc Dis 1994; 4:6671.
ischemia in cats. Stroke 1988; 19:15215.

& 2009 The Authors.

284 & 2009 World Stroke Organization International Journal of Stroke Vol 4, August 2009, 274284