Original Contributions
Timing for Fever-Related Brain Damage in
Acute Ischemic Stroke
José Castillo, MD; Antoni Dávalos, MD; Jaume Marrugat, MD; Manuel Noya, MD
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Background and Purpose—The association between hyperthermia and early neurological deterioration, increased
morbidity, and mortality in acute ischemic stroke is well known. However, the timing at which the cerebral lesion may
be aggravated by high temperature has not been firmly established. The aim of this study was to determine the
prognostic value of body temperature measured at different times after onset of stroke.
Methods—Axillary temperature was recorded every 2 hours for 72 hours in 260 patients with a hemispheric cerebral
infarction of ,24 hours’ duration. A potential infectious focus was examined in all patients with hyperthermia
(temperature .37.5°C in any of the assessments). Stroke severity was quantified with the Canadian Stroke Scale on
admission. The relationship between the highest temperature recorded in each 6-hour interval from stroke onset and
stroke outcome (Canadian Stroke Scale and Barthel Index at 3 months) or infarct volume was evaluated by correlation
analyses. The importance of the time at which hyperthermia was first detected was assessed by logistic regression
analysis.
Results—During the first 72 hours, 158 patients (60.8%) had hyperthermia, and in 57.6% of them an infectious cause was
identified. Mortality rate at 3 months was 1% in normothermic patients and 15.8% in hyperthermic patients (P,0.001).
The correlation coefficients between the final infarct volume, Canadian Stroke Scale and Barthel Index scores at 3
months, and each temperature recording decreased progressively over time from symptom onset. Hyperthermia initiated
within the first 24 hours from stroke onset, but not afterward, was independently related to larger infarct volume (odds
ratio [OR]53.23, 95% CI51.63 to 6.43; P,0.001) and higher neurological deficit (OR53.06, 95% CI51.70 to 5.53;
P,0.001) and dependency (OR53.41, 95% CI51.69 to 6.88; P50.002) at 3 months. The infectious origin of
hyperthermia was not associated with poorer outcome or greater infarct volume.
Conclusions—The relationship between brain damage and high temperature is greater the earlier the increase in
temperature occurs. However, only body temperature within the first 24 hours from stroke onset is associated with poor
outcome and large cerebral infarcts. (Stroke. 1998;29:2455-2460.)
Key Words: fever n hyperthermia n stroke, ischemic n stroke outcome n temperature
H
yperthermia is a proven factor for cerebral injury in
experimental models of focal cerebral ischemia,1– 4 and
its harmful effect persists even if it appears days after the start
of ischemia.5,6 The association between hyperthermia, even if
slight, and early neurological deterioration, increased morbidity, and mortality has also been documented in patients
with acute ischemic stroke.7–11 However, the timing at which
high temperature may aggravate the cerebral lesion has not
been firmly established. The relationship between fever and a
poor prognosis may be easily explained in patients with
central hyperthermia due to extensive infarcts that affect the
anterior region of the hypothalamus,12 but it is unknown
whether infectious fever may contribute to greater cerebral
damage in patients with small cerebral infarctions located
distant to the thermoregulatory center. The objective of this
study was to determine the prognostic value of hyperthermia
of infectious and noninfectious origin recorded at different
times after onset of stroke.
Subjects and Methods
We prospectively studied 297 patients with a cerebral hemispheric
infarction confirmed by CT, admitted within the first 24 hours after
onset of symptoms (or of sleep, if the patient awoke with stroke).
Thirty-seven patients died during the first 3 days of data collection;
therefore, 260 patients (59% male; mean age, 69.8610.2 years) were
finally evaluated in the study.
Temperature was taken for 5 minutes in the dry axilla of patients
on admission and every 2 hours for 3 days. For the purpose of this
study, only the highest temperature recorded in each 6-hour period
from the onset of stroke (not from admission) was considered in the
Received July 10, 1998; final revision received September 14, 1998; accepted September 14, 1998.
From the Department of Neurology, Complexo Hospitalario Universitario de Santiago (Hospital Xeral de Galicia), Santiago de Compostela (J.C.,
M.N.); Section of Neurology, Hospital Universitari Doctor Josep Trueta, Girona (A.D.); and Lipid and Cardiovascular Epidemiology Unit, Institute
Municipal d’Investigació Médica, Barcelona (J.M.), Spain.
Presented at the 22nd International Joint Conference on Stroke and Cerebral Circulation, Anaheim, Calif, February 6 – 8, 1997.
Correspondence to Dr José Castillo, Department of Neurology, Hospital Xeral de Galicia, 15705 Santiago de Compostela, Spain. E-mail
[email protected]
© 1998 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
2455
2456
Timing for Fever-Related Brain Damage in Acute Stroke
Figure 1. Distribution of the number of
patients with and without hyperthermia
in each of the time periods studied.
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
analyses of data. Hyperthermia was defined as an axillary temperature .37.5°C in $1 determinations. In all patients with hyperthermia, blood analyses, chest x-ray, hemocultures for aerobic and
anaerobic germs, and sputum and urine cultures were performed to
exclude a potential infection. If these results were negative, additional examinations were performed, depending on the symptoms:
coproculture (5 patients), spinal tap and cerebrospinal fluid culture (6
patients), catheter culture (3 patients), and bronchial aspirate culture
(11 patients). Hyperthermia was considered to be of infectious origin
if one of the preceding tests was positive; otherwise, hyperthermia
was considered to be of noninfectious origin. Although patients with
infectious hyperthermia were treated with antibiotics and those with
body temperature $38°C received antipyretics, these drugs were not
registered in the database.
Stroke severity was quantified with the use of the Canadian Stroke
Scale (CSS) on admission, on day 7, and at 3 months. The functional
capacity of the patient was evaluated with the Barthel Index (BI) at
3 months. All CT examinations were performed on a CT Systec 3000
Plus (GEC) scanner with a 5123512 matrix display. A cerebral CT
was performed on all patients on admission; a second CT was
completed between days 4 and 7 after the patient’s inclusion in the
study. The final infarct volume was calculated with the CT scan
performed on days 4 to 7 according to the following formula:
0.53a3b3c, where a and b are the largest perpendicular diameters
measured on the CT and c is slice thickness. Infarct size was
measured by one radiologist who was unaware of each patient’s
clinical results.
Twenty-six patients (10%) died between day 3 and the last
evaluation at 3 months, and 7 patients were not present at the final
visit. The value of 0 was imputed to the CSS and BI scores at 3
months in those patients who died before the end of the study, and
the CSS score at day 7 was carried forward as the CSS score at 3
months in those who were lost to follow-up after the first week.
Statistical Analyses
Proportions were compared with the x2 test. Comparison between 2
groups of continuous variables was performed with the Student’s t
test or the Mann-Whitney test, depending on whether or not the
distribution was normal.
Correlations between the highest body temperature recorded at
each 6-hour period and stroke outcome measures at 3 months (CSS
and BI scores) or final infarct volume were performed with Spearman’s and Pearson’s analyses, respectively. A log transformation of
infarct volume was performed to achieve a normal distribution.
Correlation coefficients obtained at each interval from stroke onset
were plotted in charts showing the magnitude of the relationship
between temperature and the outcome variables.
The importance of the time at which hyperthermia was recorded
during the first 72 hours for stroke outcome was assessed by stepwise
logistic regression analysis. CSS and BI at 3 months and the ultimate
infarct volume were considered 2 categories because they were not
normally distributed. Cutoff values for CSS (0, good outcome: $7
points; 1, poor outcome: ,7 points), BI (0, good outcome: $60
points; 1, poor outcome: ,60 points) and infarct volume (0, small:
,30 cm3; 1, large: $30 cm3) were calculated as described by Robert
et al,13 with consideration of the mean values in normothermic and
hyperthermic patients. The time interval from stroke onset at which
hyperthermia was initially recorded (0 to 24 hours, 24 to 48 hours, 48
to 72 hours), age, the highest temperature recorded during the study
period, and coexistent infections within the first 3 days (15yes,
05no) were included as covariates. This procedure allowed us to
determine whether high temperature initiated in a particular interval
from symptom onset was related to stroke outcome and infarct
volume independently of the absence of fever in the subsequent
intervals as a result of treatment.
Results
During the first 72 hours after symptom onset, 158 patients
(60.8%) developed hyperthermia. Hyperthermia was first
recorded within the first 24 hours in 91 patients, from 24 to 48
hours in 49, and from 48 to 72 hours in 18 patients. The
number of patients with and without hyperthermia in each
time period is shown in Figure 1.
Body temperature on admission was significantly higher in
the 37 excluded patients who died within the first 3 days of
hospitalization than in the study population (37.860.5°C
versus 36.960.7°C; P50.028). Thirty-three of the 37 excluded patients had hyperthermia, in 28 of them initiated
within the first 24 hours from onset and in 5 afterward.
Among the 260 survivors at the third day, mortality rate at 3
months was 1% in normothermic patients and 15.8% in
patients who developed hyperthermia within the first 72
hours (P,0.001). Mortality was higher when hyperthermia
Castillo et al
December 1998
2457
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Figure 2. Lines showing the distribution
of correlation coefficients between each
temperature recording and stroke outcome variables (Canadian Stroke Scale
[CSS] and Barthel Index [BI] at 3 months
[3m]) (top, Spearman’s correlation) and
log-transformed final infarct volume (bottom, Pearson’s correlation). Note that
the coefficients decreased progressively
over time from symptom onset (coefficients .0.2, P,0.001).
was recorded in the first 48 hours than later (17.1% versus
5.6%; P5NS).
CSS score on admission was significantly lower in hyperthermic patients (4.962.5) than in normothermic patients
(6.961.9) (P,0.001). At 3 months, poor outcome in neurological deficit (49.4% versus 13.7%; P,0.001) and functional capacity (46.2% versus 11.8%; P,0.001) was more
frequent in patients with hyperthermia. CT showed a large
cerebral infarct in 9.8% of normothermic patients and in
39.2% of hyperthermic patients (P,0.001).
The relationship between brain damage and high temperature was greater the earlier the increase in temperature. The
correlation coefficients between the final infarct volume, CSS
and BI scores at 3 months, and each temperature recording
decreased progressively over time from symptom onset (Figure 2).
In 91 (57.6%) of 158 patients with hyperthermia within the
first 72 hours, an infectious cause was found (bronchopulmonary infection in 47 patients, urinary infection in 40 patients,
sepsis in 4 patients, thrombophlebitis in 9 patients, and
salmonellosis in 1 patient). The proportion of patients with
hyperthermia of infectious and noninfectious origin in each
time period is shown in Figure 3. Body temperature was
higher in patients with hyperthermia of infectious etiology
than in those with hyperthermia of noninfectious origin in all
the time periods studied, although this difference only
reached statistical significance from 36 hours onward (data
not shown). Mortality was greater in patients with infectious
hyperthermia than in those with noninfectious hyperthermia
(22% versus 7.5%; P,0.013), although patients with infections did not have a higher stroke severity on admission in
comparison with those with fever of noninfectious origin
2458
Timing for Fever-Related Brain Damage in Acute Stroke
Figure 3. Frequency of patients with hyperthermia of infectious and noninfectious origin at each time interval from symptom
onset.
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
(CSS: 4.962.4 versus 4.862.6; P50.76). The final infarct
volume (46662 versus 36638 cm3; P50.54) and functional
capacity at 3 months (BI: 48643 versus 61636; P50.14)
were not significantly different between the groups of patients with infectious and noninfectious hyperthermia, although the former group had a significantly higher stroke
severity at the end of the follow-up (CSS: 5.463.6 versus
6.862.8; P50.035).
The correlation coefficients between stroke severity at 3
months, final infarct volume, and each body temperature
recording were similar in patients with infectious and noninfectious hyperthermia during the first 30 hours from symptom
onset. From 30 to 72 hours, the coefficients decreased more
notably in cases of infectious hyperthermia (Figure 4).
Logistic regression analyses showed that hyperthermia
initiated within the first 24 hours from stroke onset, but not
afterward, was independently related to large infarct volumes
and poor outcome in neurological deficit and functional
ability at 3 months (Table). Coexistent infections within the
first 3 days were not independently associated with poor
prognosis.
Discussion
In experimental models of cerebral ischemia, even moderate
hyperthermia is a factor that increases cerebral lesion and
volume of infarcted tissue.1– 6,14 Hyperthermia facilitates the
transformation of ischemic penumbra into infarction and
accelerates the development of ischemic necrosis.1,2,5,14 In the
human acute stroke, it has also been demonstrated that
hyperthermia is an independent risk factor of poor outcome
and that it is associated with greater morbidity and mortality9 –11 and larger infarct volumes.11 The present results are in
accord with the previous findings. We have found an independent relationship between high body temperature within
the first 24 hours and larger infarct volume, higher neurological deficit, and dependency at 3 months.
Experimental and clinical studies have emphasized the
importance of the time of onset of hyperthermia on the
magnitude of the cerebral injury.1–3,14,15 Larger lesions are
provoked when hyperthermia coincides with the onset of
cerebral ischemia.1–3,14 Similarly, hypothermia yields beneficial effects when it is simultaneously induced with the
development of ischemia, but it is not effective if its onset is
Multiple Logistic Regression Analyses of Infarct Volume and Stroke Outcome
Infarct Volume
Covariates
Odds Ratio (95% CI)
Greater Neurological Deficit at 3 mo
P
Odds Ratio (95% CI)
P
Poor Functional Capacity at 3 mo
Odds Ratio (95% CI)
P
Age
1.02 (0.99–1.05)
0.173
1.04 (1.01–1.07)
0.010
1.08 (1.05–1.12)
,0.001
Infection
0.72 (0.32–1.62)
0.427
0.92 (0.43–2.01)
0.845
1.49 (0.65–3.39)
0.311
Highest temperature, °C
2.81 (1.34–5.89)
0.006
1.68 (0.84–3.40)
0.142
1.85 (0.88–3.88)
0.213
Time at which hyperthermia
was initiated
0–24 h
3.23 (1.63–6.43)
,0.001
3.06 (1.70–5.53)
,0.001
3.41 (1.69–6.88)
0.002
24–48 h
1.14 (0.52–2.51)
0.736
1.47 (0.78–2.80)
0.235
1.41 (0.66–3.05)
0.371
48–72 h
0.23 (0.05–1.09)
0.064
0.33 (0.10–1.03)
0.055
0.20 (0.04–0.96)
0.044
Castillo et al
December 1998
2459
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Figure 4. Distribution of correlation
coefficients between each temperature
recording and Canadian Stroke Scale
(CSS) score at 3 months (top, Spearman’s correlation) and log-transformed
final infarct volume (bottom, Pearson’s
correlation) in patients with hyperthermia
of infectious and noninfectious origin
within the first 72 hours. Note that the
coefficients were similar in patients with
infectious and noninfectious hyperthermia during the first 30 hours of evolution
but decreased more notably in cases of
infectious hyperthermia from 30 hours
onward (coefficients .0.2, P,0.001).
hours afterward.16,17 Models of focal and global cerebral
ischemia have been developed with delayed induction of
hyperthermia to reproduce the clinical situation in humans,
showing the persistence, even 24 hours afterward, of the
pernicious effects of an increase in temperature.5,6
The timing for hyperthermia-related brain damage in acute
ischemic stroke had not been investigated previously. In the
present study body temperature was monitored during 72
hours, ie, during the period in which most of progressing
strokes occur. According to our results, the relationship
between the intensity of hyperthermia and stroke outcome or
infarct volume is stronger the earlier the fever develops, and
body temperature within the first 24 hours is the key value
associated with greater cerebral damage; when hyperthermia
appears afterward it is not an independent factor of poor
outcome. The use of antipyretics in some patients, a fact that
was not recorded in this study, could modify body temperature profile after admission. Nevertheless, its influence in our
results is unlikely since we used a logistic model that allowed
us to determine whether high temperature initiated in a
particular interval from symptom onset was related to stroke
outcome and infarct volume independently of the absence of
fever in the subsequent intervals as a result of treatment.
The most frequent etiology of fever in acute stroke is
infection, but hyperthermia is occasionally an expression of
cell necrosis or of changes in thermoregulatory mechanisms
that occur when lesions are located in the anterior region of
the hypothalamus.12,18 –20 However, we cannot rule out the
occurrence of infectious and noninfectious fever in some
patients, particularly in those with large infarcts, which are
2460
Timing for Fever-Related Brain Damage in Acute Stroke
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
more prone to aspiration and urinary complications and
therefore to infection. In our study infection was demonstrated in 58% of the patients who developed hyperthermia,
while in 42% of patients fever could be due to tissue necrosis
itself or to changes in thermoregulation. The relationship
between hyperthermia within the first 24 hours from symptom onset and stroke outcome or infarct volume was independent of the infectious or noninfectious origin of fever.
However, after 30 to 36 hours from onset, the significant
correlation between infarct volume, neurological deficit at 3
months, and body temperature disappeared in patients with
infections. These facts support the hypothesis that in some
patients fever could be directly related to larger cerebral infarctions or to the acute phase response in more severe strokes.21
In conclusion, the relationship between the degree of hyperthermia and stroke outcome or infarct volume is highest when it begins
within 24 hours of onset of symptoms. While further studies are
needed to confirm these results, our main efforts should be directed
toward an immediate and effective reduction of body temperature
when it is .37.5°C, particularly within the first 24 hours of cerebral
ischemia, since most likely only high temperature in this period
independently contributes to poor prognosis.
Acknowledgment
This study was supported by a grant from the Xunta de Galicia
Research Project (XUGA20807B96).
References
1. Busto R, Dietrich WD, Globus MY-T, Valdés I, Scheinberg P, Ginsberg
MD. Small differences in intraischemic brain temperature critically
determine the extent of ischemic neuronal injury. J Cereb Blood Flow
Metab. 1987;7:729 –738.
2. Minamisawa H, Smith M-L, Siesjö BK. The effect of mild hyperthermia
and hypothermia on brain damage following 5, 10, and 15 minutes of
forebrain ischemia. Ann Neurol. 1990;28:26 –33.
3. Ginsberg MD, Sternau LL, Globus MY-T, Dietrich WD, Busto R. Therapeutic modulation of brain temperature: relevance to ischemic brain
injury. Cerebrovasc Brain Metab Rev. 1992;4:189 –225.
4. Meden P, Overgaard K, Pedersen H, Boysen G. The influence of body
temperature on infarct volume and thrombolytic therapy in a rat embolic
stroke model. Brain Res. 1994;647:131–138.
5. Kim Y, Busto R, Dietrich WD, Kraydieh S, Ginsberg MD. Delayed postischemic hyperthermia in awake rats worsens the histopathological outcome of
transient focal cerebral ischemia. Stroke. 1996;27:2274–2281.
6. Baena RC, Busto R, Dietrich WD, Globus MY-T, Ginsberg MD. Hyperthermia delayed by 24 hours aggravates neuronal damage in rat hippocampus following global ischemia. Neurology. 1997;48:768 –773.
7. Castillo J, Dávalos A, Noya M. Progression of ischaemic stroke and
excitotoxic amino acids. Lancet. 1997;349:79 – 83.
8. Dávalos A, Castillo J, Pumar JM, Noya M. Body temperature and
fibrinogen are related to early neurological deterioration in acute ischemic
stroke. Cerebrovasc Dis. 1997;7:64 – 69.
9. Hindfelt B. The prognostic significance of subfebrility and fever in
ischaemic cerebral infarction. Acta Neurol Scand. 1976;53:72–79.
10. Castillo J, Martı́nez F, Leira R, Prieto JM, Lema M, Noya M. Mortality
and morbidity of acute cerebral infarction related to temperature and basal
analytic parameters. Cerebrovasc Dis. 1994;4:66 –71.
11. Reith J, Jørgensen HS, Pedersen PM, Nakayama H, Raaschou HO,
Jeppesen LL, Olsen TS. Body temperature in acute stroke: relations to
stroke severity, infarct size, mortality, and outcome. Lancet. 1996;347:
422– 425.
12. Przelomski MM, Roth RM, Gleckman RA, Marcus EM. Fever in the
wake of a stroke. Neurology. 1986;36:427– 429.
13. Robert C, Vermont J, Bosson JL. Formulas for threshold computations.
Comput Biomed Res. 1991;24:514 –529.
14. Chen H, Chopp M, Welch KMA. Effect of mild hyperthermia on the
ischemic infarct volume after middle cerebral artery occlusion in the rat.
Neurology. 1991;41:1133–1135.
15. Jørgensen HS, Reith J, Pedersen PM, Nakayama H, Olsen TS. Body
temperature and outcome in stroke patients. Lancet. 1996;348:193.
Letter.
16. Ginsberg MD, Sternau LL, Globus MY-T, Dietrich WD, Busto R. Therapeutic modulation of brain temperature: relevance to ischemic brain
injury. Cerebrovasc Brain Metab Rev. 1992;4:189 –225.
17. Chopp M, Chen H, Dereski MO, Garcia JH. Mild hypothermia intervention after graded ischemic stress in rats. Stroke. 1991;22:37– 43.
18. Castillo J, Noya M. The prevention and treatment of medical complications in hospitalized stroke patients. In: Castillo J, Dávalos A, Toni D,
eds. Management of Acute Ischemic Stroke. Barcelona, Spain: SpringerVerlag Ibérica; 1997:67– 80.
19. Powers JH, Sheld WN. Fever in neurologic diseases. Infect Dis Clin
North Am. 1996;10:45– 66.
20. Lee-Chiong TL, Stitt JT. Disorders of temperature regulation. Compr
Ther. 1995;21:697–704.
21. Dinarello CA. The acute phase response. In: Bennett JC, Plum F, eds.
Celcil Textbook of Medicine. Philadelphia, Pa: WB Saunders Co; 1996:
1535–1537.
Timing for Fever-Related Brain Damage in Acute Ischemic Stroke
José Castillo, Antoni Dávalos, Jaume Marrugat and Manuel Noya
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Stroke. 1998;29:2455-2460
doi: 10.1161/01.STR.29.12.2455
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1998 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/29/12/2455
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/