Journal of Cerebral Blood Flow and Metabolism 12:817-822 © 1992 The International Society of Cerebral Blood Flow and Metabolism Published by Raven Press, Ltd., New York

Methodological Requirements for Accurate Measurements of Brain and Body Temperature During Global Forebrain Ischemia of Rat T. Miyazawa and K.-A. Hossmann Department of Experimental Neurology, Max-Planck-Institute for Neurological Research, Cologne, Germany

Summary: The methodological requirements for accurate measurements of brain and body temperature during brain ischemia have been validated in Wistar rats submit­ ted to 30 min of four-vessel occlusion. During ischemia, brains were exposed to three different temperature pro­ files: spontaneous cooling from 36 to 31°C (n = 10), con­ stant hypothermia at 30°C (n = 19), and constant normo­ thermia at 36°C (n = 21). Direct and indirect brain tem­ perature recordings were carried out by placing fine thermocouples (200 !-Lm diameter) into the striate nucleus, the temporal muscle, and the epidural space. Body tem­ perature was measured with a flexible thermocouple in­ serted at various depths into the rectum. Accurate mea­ surements of body temperature required insertion of the

rectal probe to a depth of at least 6 cm; lesser insertion resulted in an underestimation of up to 6°C, Accurate estimates of brain temperature were obtained in all three experimental conditions by recording of the epidural tem­ perature. The temperature in the temporal muscle, by contrast, differed from the brain temperature by up to 2°C, depending upon the experimental condition and the duration of ischemia. We therefore suggest that indirect measurements of brain temperature during ischemia are carried out in the epidural space in order to avoid misin­ terpretations of temperature-sensitive pathological changes. Key Words: Brain temperature-Core tempera­ ture-Cerebral ischemia-Normothermia-Hypother­ mia-Rat.

The severity of brain damage after global cere­ brocirculatory arrest is markedly affected by slight variations of brain temperature (Busto et a!., 1987, 1989a). Investigations of ischemic injury, therefore, require precise measurements of brain temperature, particularly for the evaluation of pharmacological intervention (Buchan and Pulsinelli, 1990; Corbett et a!., 1990). A methodological problem associated with such measurements is the placement of the temperature probes. Measurements on the surface of the brain require exposure of the cortex and are greatly affected by changes in ambient temperature. Insertion of the probe into the brain produces a pen­ etrating lesion that may affect the ischemic injury (Takahata and Shimoji, 1986; Mushiroi et aI., 1989). Noninvasive temperature measurements by infra-

red thermography (Anderson et aI., 1970) or mag­ netic resonance imaging (Dickinson et aI., 1986) re­ quire expensive instrumentation and are not suit­ able for routine recordings. Therefore, the brain temperature is frequently estimated from extracere­ bral recordings, in particular in the temporal muscle ( Busto et aI., 1987, 1989a,b; Dietrich et aI. , 1990a,b). A validation of this approach has been carried out by Busto et al. ( 1987), who compared the spontaneous temperature decline in the striate nucleus and temporal muscle after bilateral carotid and vertebral artery occlusion in the rat. They ob­ served that brain and muscle temperature were sim­ ilar before and after ischemia but that during isch­ emia the muscle temperature exceeded the brain temperature by I-2°C. This error could be cor­ rected because the muscle and brain temperatures correlated linearly within the range of 30-38°C. Busto et al. ( 1987) therefore concluded that the tem­ perature of muscle provides a reliable estimate of the brain temperature. However, this conclusion may become invalid if

Received December 2, 1991; final revision received February 3, 1992; accepted February 7. 1992. Address correspondence and reprint requests to Prof. Dr. K.-A. Hossmann at Max-Planck-Institut fUr neurologische Forschung, Gleueler Str. 50, D-5000 K61n 41 (Lindenthal), Ger­ many.

817

818

T. MIYAZAWA AND K.-A. HOSSMANN

the brain temperature is manipulated during isch­ emia. In fact, external heat or cold sources are fre­ quently used to stabilize the brain temperature at a constant level during the ischemic impact. Tests to find the appropriate location of extracerebral tem­ perature probes, therefore, require recording of the corresponding temperature profiles not only under conditions of spontaneous brain cooling but also during induced temperature changes. Another source of temperature-associated error is the incorrect placement of temperature probes for the recording of body temperature. As previously reported by Lomax ( 1966) and Donhoffer ( 1980), the core temperature may be substantially underes­ timated by rectal probes that are not inserted deeply enough. In the present investigation, the method­ ological requirements for precise measurements of core and brain temperature have been validated in rats submitted to 30 min of four-vessel occlusion with or without additional manipulation of the brain temperature. The results obtained indicate that placement of the rectal probe is critical for accurate measurement of the body temperature and that tem­ perature recordings in the temporal muscle deviate substantially from the brain temperature. Indirect measurements of the brain temperature, therefore, should be carried out in the epidural space. MATERIALS AND METHODS Male Wistar rats weighing 250-350 g were used. Ani­ mals were anesthetized with 1.5-2% halothane and 70% N20. Thirty minutes of four-vessel occlusion (transient clipping of the carotid and permanent coagulation of ver­ tebral arteries) was carried out by a modification (Schmidt-Kastner et aI., 1989) of the method described by Pulsinelli and Brierley (1979). The body temperature was measured with a flexible thermocouple inserted at differ­ ent depths into the rectum. The brain temperature was recorded with a thin wire-type thermocouple (copper/ constantan wire, diameter of 200 flom, California Fine­ wire, Grover City, CA, U.S.A.) inserted stereotactically through a thin glass tube into the striate nucleus. Mea­ surements of the extracerebral temperature were carried out in the right temporal muscle or the right epidural space. For recording of the muscle temperature, the tip of the thermocouple was placed between muscle and bone. For placement in the epidural space, a small burr hole was made 3 mm lateral and 2 mm posterior to the bregma. The dura mater was separated gently from the calvarium, and the tip of the thermocouple was inserted into the epidural space and fixed with dental cement. Three groups of animals were compared. In the spon­ taneously hypothermic group (n = 10), the brain temper­ ature was allowed to decline spontaneously during isch­ emia at an ambient room temperature of about 21°C. In the induced hypothermic group (n = 19), the brain tem­ perature was lowered to 31°C immediately before isch­ emia by exposing the head to the vapors of liquid nitro­ gen, and then kept at this level throughout ischemia. In

J Cereb Blood Flow Metab. Vol. 12, No.5. 1992

the normothermia group (n = 21), the brain temperature was kept constant at 36°C by exposing the head to the heat of a 60-W electrical bulb placed at a distance of 25-30 cm over the head. Correlations between brain and extracerebral temper­ ature were established in each group by linear regression analysis.

RESULTS Measurements of body temperature At the beginning of the experiment, a flexible thermocouple was inserted 7 cm into the rectum, and the temperature recorded at this site was stabi­ lized for 15 min at 37°C, using a feedback-controlled heating system. Thereafter, the probe was with­ drawn in stages, and the recorded temperature was plotted against the distance of the thermocouple from the anus. At a distance of less than 6 cm, the temperature declined at first slowly and then steeply (Fig. l). Precise measurements of core tem­ perature, therefore, require insertion of the thermo­ couple to at least 6.0 cm. In the following experi­ ments, the thermocouple was placed at 6.5 cm from the anus, and the corresponding core temperature was kept constant at 37°C, by means of a feedback­ controlled heating system. Measurements of brain temperature The ambient room temperature during measure­ ments was 2 1. 2 ± 0.9°C (n = 50). The temperature 38

36 u

!... ev ... j

C; L-

34

ev Q.

E

,!. 32

i

2

i

3

i

4

s

6

7cm

Distance from anus FIG. 1. Measurement of body temperature with a flexible thermoprobe inserted at various depths into the rectum (n = 11). Recordings were taken after stabilization of the core temperature at 7 cm. Subsequently, the probe was withdrawn in stages without changing the heating condition of animals. Note the sharp decline of the recorded temperature as the probe is withdrawn. Values are means ± SO.

ACCURA TE TEMPERATURE MEASUREMEN TS DURING ISCHEMIA

profiles recorded during and after ischemia in brain and extracerebral tissue are shown in Fig. 2. Before ischemia, the striate temperature was about 1°C be­ low the core temperature; the temperatures of the temporal muscle and epidural space differed from the striatal temperature by less than 0.3°C. In the spontaneously hypothermic group, i.e., in animals whose head temperature had not been ex­ ternally manipulated, four-vessel occlusion resulted in an exponential fall in striatal temperature (Fig. 2A): 5 min after the onset of ischemia, the temper­ ature decreased by 2. 1 ± 0.75°C, and after 30 min by 4.0 ± 1.3°e below the preischemic value. In the

temporal muscle, the temperature decline was much slower, and after 30 min of ischemia it amounted to only 2.6 ± 0.94°e. Epidural recording, in contrast, closely followed the striatal tempera­ ture profile. During postischemic recirculation, the striatal temperature transiently increased 0.8°e above the preischemic value. The temperature in the temporal muscle-but not in the epidural space-underestimated this overshoot by about 0.6°e. During induced hypothermia, the striatal temper­ ature was lowered to 3 1.3 ± 1.4°e before ischemia (Fig. 2B). In this group, a similar dissociation be-

Epidural space

Temporal muscle

'0

1.0

38

38

� � E e

.. 0-

�

35

..

�

36

2..

3'

:>

3'

0-

E �

32

15

30 ReCirculatIOn

'5

.. 0-

E �

38

� 35

31.

I

'�"-0�J

'

Temporal muscle In,lO)

I-�._. I I i I

15

45

30

60 min

38

�

36

2..

31.

0-

E

..

�

32

32 30

15

30

45

15

Ischemia

60 min

30 ReCirculatIOn

40

40

./

38

Epidural space (no17) _

� �

36

:>

38 36

:>

o

I,

3'

E � 32

c

'_'

1

:>

ReCirculation

�

i

1

40

Ischemia

�

I '.

1

•

Ischemia

30

B

1'.

I

�_L��I

l(Ic--1-'I\r-F�==�==.-. \ '

•

ReCirculatIOn

:>

2

"

'\�'LU.

60 min

'0

�

i

=,,1

30 Ischemia

�

Striate nucleus (n,10)

E � 32

30

A

819

"

Strlate nucleus (n, 21

o :;; 3' 0E

)

..

�

_�

�

45

� o�","iI"".I =�,�_� -L._� _�-I I �-.=.�"""...-" . I 1

"

I

o-n_O-c -', _r i ! I I i

I

·

1

Te

1 1

I

1

poral muscle In,5)

'-0

1

1

I

I

I

,I 1

Striate nucleus I n, 21)

32 30

30 Ischemia

15

30 Recirculation

'5

50 m in

Ischemia

15

30

45

&0

min

ReCirculation

FIG. 2. Simultaneous recording of striatal and epidural temperature (left) and of striatal and temporal muscle temperature (right) in rats submitted to 30 min of four-vessel occlusion. A: Spontaneous fall in temperature at an ambient room temperature of 21°C (spontaneous hypothermia, n = 10). B: Cooling of striate nucleus temperature to 31°C by controlled exposure of the head to liquid nitrogen vapors (induced hypothermia, n = 19). C: Prevention of temperature decline in striate nucleus by controlled exposure of the head to an electric bulb (constant normothermia, n = 21). Note the marked difference between the temperature in the striate and temporal muscle, in contrast to the similar temperature profile in the striatum and epidural space. Values are means ± SO.

J Cereb Blood Flow Metab, Vol. 12, No.5, 1992

820

T. MIYAZAWA AND K.-A. HOSSMANN

tween striatal and temporal muscle temperatures was observed as in the spontaneously hypothermic group. Throughout ischemia, the temporal muscle temperature-but not epidural-underestimated the striatal temperature by about 2°C. The postischemic overshoot of the striatal temperature was underes­ timated in the temporal muscle by about l.O°C. In the induced normothermic group, the striatal temperature was stabilized close to 36°C throughout the ischemic period ( Fig. 2C). During this time, the temperature in the epidural space was about O.4°C higher, and in the temporal muscle it was about O.5°C lower than in the striate nucleus. The correlation of the mean temperature record­ ings obtained at different intervals after the onset of ischemia demonstrated that epidural recordings were virtually identical with striatal values in all Epidural space 1.0 Co

38

'" u

&.

'"

A

.

E '"

38

'"

36

&.

E

'" ...

30 36

31.

o

" E 31. ::

32

y=0.599x .11. 698 R=0.967

'"

...

30

32 30

.

•

30

1.0

38

Striatal temperature 1°C ) 1.0

l;-' Co

E

38

. ..

� 31.

E �

Ci.

lJJ

B

l;-' 1.0 .. '

o

E

38

'"

36

�

u

..

I

,

1.0

38

. ..

.

::

&. 32

E

� 30

30 32

31.

36

38

1.0

I

30

/1.0

. .'

38

E

�

.! 36 '" v

...

.. '

o

�

'

E'"

. .'

::

Ci. LLI

E ..

&. ...

30 30

32

•

31.

I

36

i

36

i

38

i

1.0

38

� 36 u '" " E 31.

o ...

i

i

31.

1.0

.'

� 32

i

i

32

Striatal temperature ,OC)

� 31.

c

i

36

y=0.60I.x .13.705 R=0.996

Striatal temperature 1°C)

ci

•

31.

'" " E 31.

32

30

�

I

32

Striatal temperature 1°C)

y=0.91.3x .1.757 R=0.998

� 36 �

The present validation of core and brain temper­ ature measurements during four-vessel occlusion in the rat revealed that measurements may be subject to substantial errors when temperature probes are placed inappropriately. The core temperature may be underestimated by as much as 6°C when the probe is not inserted deeply enough into the rectum, and the brain temperature may be overestimated by up to 2°C when the temperature is recorded in the temporal muscle. These errors may not only lead to

u

.

31.

-0 ... � 32

Ci. LLI

1.0

.. ..

DISCUSSION

Temporal muscle

Y= 0 899x. 3.711 R=0.995

E � 36

three groups (Fig. 3). Temporal muscle recordings, in contrast, considerably deviated from striatal measurements, particularly in the low temperature range.

I

38

,

1.0

Striatal temperature 1°C)

J Cereb Blood Flow Metab, Vol. 12, No.5, 1992

32 30 •

30

i

32

•

31.

i

36

,

38

Striatal temperature 1°C)

•

1.0

FIG. 3. Correlation between striatal and epidural temperature (left) and be­ tween striatal and temporal muscle temperature (right) in rats submitted to 30 min of four-vessel occlusion. A-C: Same experimental conditions as in Fig. 2. Measurements were taken before ischemia, at 10, 20, and 30 min after the onset of ischemia, and at 10. 30, and 60 min after the onset of recir­ culation. Solid line: calculated regres­ sion between variables; dotted line: line of identity. Note the close similar­ ity between striatal and epidural tem­ perature recordings under all experi­ mental conditions.

ACCURATE TEMPERATURE MEASUREMENTS DURING ISCHEMIA

erroneous interpretations of temperature effects but also to severe misadjustments of physiological vari­ ables if feedback-controlled heating or cooling sys­ tems are used for stabilization of the temperature at a constant level. For instance, the underestimation of the core temperature by only 3-4°C results in severe hyperthermia if the set point of the temper­ ature-controlling system is adjusted to 38°C. Con­ versely, overestimation of the brain temperature by placing the temperature probe into the temporal muscle may result in pathophysiologically signifi­ cant hypothermia if the temporal muscle tempera­ ture is adjusted to the normothermic range. Both hypo- and hyperthermia have substantial effects on the severity of ischemic brain injury, particularly in the selectively vulnerable regions of the brain after brief periods of cerebrocirculatory arrest (Dietrich et aI., 1990b; Kuroiwa et aI., 1990; Minamisawa et aI., 1990). Precise measurements of temperature, therefore, are of eminent importance for minimizing the variability of ischemic injury and for avoiding misinterpretations of temperature-sensitive patho­ logical processes. Our finding of a substantial difference between striate nucleus and temporal muscle temperatures in animals with spontaneous brain cooling during ischemia confirms similar observations by Busto et al. (1987), who also observed a discrepancy of up to 2°C after 30 min of four-vessel occlusion. This dif­ ference may be caused by the faster arrest of cere­ bral metabolic activity caused by selective brain cooling through fluid evaporation from the nasal mucosa (Hayward and Baker, 1969; Bamford and Eccles, 1983), or some residual blood flow in the temporal muscle after vascular occlusion. Busto et al. (1987) proposed to correct this tem­ perature difference by introducing an empirical cor­ rection factor derived from the correlation of brain and muscle temperatures during spontaneous brain cooling. Our data confirm that under constant ex­ perimental conditions, the two temperatures are tightly correlated. However, the regression coeffi­ cient changes when external heating or cooling sys­ tems are used. It also neglects variations in ambient room temperature or collateral blood supply to the temporal muscle and therefore has to be redeter­ mined for each experimental situation. The present study suggests a more reliable indi­ rect way to assess the brain temperature, i.e., plac­ ing an extracerebral temperature probe into the epi­ dural space. The temperature recorded in this posi­ tion is independent of extracerebral variations of metabolic activity or blood flow and therefore equil­ ibrates more easily with the temperature of the brain. The close agreement with striatal tempera-

821

ture measurements under all three experimental conditions demonstrates that correction factors are not required for the accurate estimation of brain temperature. Epidural measurements, in conse­ quence, are preferable to temporal muscle record­ ings if the appropriate correction factor is not known and if penetrating brain temperature mea­ surements have to be avoided. A surprising result of the present study was the large error obtained by inappropriate placement of the rectal temperature probe. Of particular concern is a possible underestimation of the body tempera­ ture if the temperature probe is not inserted deeply enough into the rectum. This error may lead to in­ advertent overheating of the animal that, in turn, may substantially aggravate the ischemic injury. In fact, ischemic or postischemic hyperthermia of as little as 1°C has been found to increase significantly the number of injured neurons in the CA l sector of the gerbil following 5 min of bilateral carotid artery occlusion (Dietrich et aI., 1990b; Kuroiwa et aI., 1990; Minamisawa et aI., 1990). Another complicating factor associated with er­ roneous underestimation of the body temperature is the exposure of the animal to an external heat source. In a parallel investigation, we have demon­ strated that the heat energy required to prevent the fall of the brain temperature during 30 min of four­ vessel occlusion impairs the general state of the an­ imal (Miyazawa and Hossmann, 1991). This impair­ ment is reflected by a significant deterioration of EEG activity and a loss of body weight not only in ischemic but also in nonischemic control animals submitted to the same heating procedure. Correct adjustments of body temperature are therefore just as important as the correct measurement of the brain temperature for avoiding complicating side ef­ fects. In conclusion, severe errors in the estimation of the brain and body temperatures may evolve from inappropriate placements of temperature probes. The present study provides methodological recom­ mendations for minimizing this pathophysiologi­ cally important problem. REFERENCES Anderson RE, Waltz AG, Yamaguchi T, Ostrom RD (1970) As­ sessment of cerebral circulation (cortical blood flow) with an infared microscope. Stroke I: 100-103 Bamford OS, Eccles R (1983) The role of sympathetic efferent activity in the regulation of brain temperature. Pflugers Arch 396:138-143 Buchan A, Pulsinelli WA (1990) Hypothermia but not the N-methyl-D-aspartate antagonist, MK-801, attenuates neu­ ronal damage in gerbils subjected to transient global isch­ emia. J Neurosci 10:311-316 Busto R, Dietrich WD, Globus MY-T, Valdes I, Scheinberg P,

J Cereb Blood Flow Me/ab, Vol. 12, No.5, 1992

822

T. MIYAZAWA AND K.-A. HOSSMANN

Ginsberg MK (1987) Small differences in intra-ischemic brain temperature critically determine the extent of ischemic neuronal injury. J Cereb Blood Flow Metab 7:729-738 Busto R, Dietrich WD, Globus MY-T, Ginsberg MD (l989a) The importance of brain temperature in cerebral ischemic injury. Stroke 20:1113-1114 Busto R, Globus MT, Dietrich WD, Martinez E, Valdes I, Gins­ berg MK (1989b) Effect of mild hypothermia on ischemia­ induced release of neurotransmitters and free fatty acids in rat brain. Stroke 20:904-910 Corbett D, Evans S. Thomas C, Wang D, Jonas RA (1990) MK801 reduced cerebral ischemic injury by inducing hypother­ mia. Brain Res 514:300-304 Dickinson RJ, Hall AS, Hind AJ, Young IR (1986) Measurement of changes in tissue temperature using MR imaging. J Com­ put Assist Tomogr 3:468-472 Dietrich WD, Busto R, Halley M, Valdes I ( I990a) The impor­ tance of brain temperature in alterations of the blood-brain barrier following cerebral ischemia. J Neuropathol Exp Neu­ rol 49:486-497 Dietrich WD, Busto R, Valdes I, Loor Y ( I990b) Effects of nor­ mothermic versus mild hyperthermic forebrain ischemia in rats. Stroke 21:1318-1325 Donhoffer SZ (1980) Homeothermia of the Brain. Cerebral Blood Flow, Metabolic Rate, and Brain Temperature in the Cold. The Possible Role of Neuroglia. Budapest, Akademiai

Kiado, pp 17-133

J

Cereb Blood Flow Metab, Vol. 12, No.5, 1992

Hayward IN, Baker MN (1969) A comparative study of the role of the cerebral arterial blood in the regulation of brain tem­ perature in five mammals. Brain Res 16:417-440 Kuroiwa T, Bonnekoh P, Hossmann K-A (1990) Prevention of postischemic hyperthermia prevents ischemic injury of CA-l neurons in gerbils. J Cereb Blood Flow Metab 10:550-556 Lomax P (1966) Measurement of "core" temperature in the rat. Nature (Lond) 210:854-855 Minamisawa H, Nordstrom CH, Smith ML, Siesj6 BK (1990) The influence of mild body and brain hypothermia on isch­ emic brain damage. J Cereb Blood Flow Metab 10:365-374 Miyazawa T, Hossmann K-A (1991) Changes of brain tempera­ ture during global forebrain ischemia in rats are underesti­ mated by measurement in temporal muscle. J Cereb Blood Flow Metab l1(suppl 2):SI25 Mushiroi T, Yoshimine T, Hayakawa T, Mogami H (1989) Stab wound prevents delayed neuronal death in the gerbil hippo­ campus. J Cereb Blood Flow Metab 9(suppl I):S188 Pulsinelli WA. Brierley JB (1979) A new model of bilateral hemi­ spheric ischemia in the unanesthetized rat. Stroke 10:267272 Schmidt-Kastner R, Paschen W, GroBe Ophoff B, Hossmann K-A (1989) A modified four-vessel occlusion model for in­ ducing incomplete forebrain ischemia in rats. Stroke 20:938-946 Takahata W, Shimoji K (1986) Brain injury improves survival of mice following brain ischemia. Brain Res 381:368-371