PEDIATRICS Vol. 108 No. 5 November 2001, pp. 1103-1110

Differences in Brain Temperature and Cerebral Blood Flow During Selective Head Versus Whole-Body Cooling

Abbot R. Laptook, MD^*, Lina Shalak, MD^*, and Ron J. T. Corbett, PhD

From the Departments of ^* Pediatrics and  Radiology, University of Texas Southwestern Medical Center, Dallas, Texas.

	ABSTRACT

Top Abstract Methods Results Discussion References

Objective.  To compare brain temperature and cerebral blood flow (CBF) during head and body cooling, with and without systemic hypoxemia.

Methods.  Seventeen newborn swine were studied for either measurement of brain temperature alone (n = 9) or measurement of brain temperature and CBF (n = 8). All animals were ventilated and instrumented, and temperature probes were inserted into the rectum, into the brain at depths of 2 and 1 cm from the cortical surface, and on the dural surface. Blood flow was measured with microspheres. The protocol consisted of a control period, an interval of either head or body cooling, and cooling with 15 minutes of superimposed hypoxia. After a 1-hour recovery period, animals were exposed to the same sequence except that the alternate mode of cooling was evaluated.

Results.  Head cooling with a constant rectal temperature resulted in an increase in the temperature gradient across the brain from the warmer central structures to the cooler periphery (brain 2 cm  dura temperature: 1.3 ± 1.1°C at control to 7.5 ± 3.5°C during cooling). Hypoxia superimposed on head cooling decreased the temperature gradient by at least 50%. In contrast, body cooling was associated with an unchanged temperature gradient across the brain (brain 2 cm  dura temperature: 1.5 ± 1.2°C at control to 1.1 ± 0.9°C during cooling). Hypoxia superimposed on body cooling did not change brain temperature. Both modes of brain cooling resulted in similar reductions of global CBF (~40%) and O₂ uptake.

Conclusion.  Brain hypothermia achieved through head or body cooling results in different brain temperature gradients. Alterations in systemic variables (ie, hypoxemia) alters brain temperature differently in these 2 modes of brain cooling. The mode of brain cooling may affect the efficacy of modest hypothermia as a neuroprotective therapy.  Key words:  hypothermia, brain cooling, head cooling, body cooling, cerebral blood flow, hypoxia.

Perinatal hypoxia-ischemia remains an important cause of adverse neurodevelopmental outcomes diagnosed during early childhood.¹ In contrast to other causes of neurodevelopmental dysfunction, perinatal hypoxia-ischemia can potentially be diagnosed shortly after birth and may be amenable to therapy. As delineated by Vannucci and Perlman,² there has been encouraging progress during the past 15 years in demonstrating neuroprotection for hypoxia-ischemia by a number of different therapeutic strategies. Modest hypothermia seems to be the most promising of the current brain-oriented therapies being examined.³ Investigations of fetal and newborn animals have established that temperature reductions of 2°C to 6°C during or immediately after ischemia are effective in decreasing brain damage.^4-6 Although the number of studies are limited, modest hypothermia as a neuroprotective strategy can be effective even when initiated up to 6 hours after the acute insult.^7,8 The optimal method to reduce brain temperature and maintain the brain at a specified temperature is unknown.

Clinical trials are in progress to compare hypothermia and conventional temperature regulation in reducing mortality and neurodevelopmental deficits in term infants with hypoxia-ischemia. Current trials have been preceded by small pilot studies in newborns with hypoxic-ischemic encephalopathy to determine feasibility and safety of hypothermia.^9-11 Methods to achieve brain hypothermia in the current clinical trials include whole-body cooling and selective head cooling combined with body cooling. These different approaches to cooling the brain may lead to important differences in the extent of brain cooling and could influence the extent of neuroprotection. Furthermore, the mode of brain cooling may affect the ability to maintain the brain at a specified temperature when alterations in oxygenation or ventilation occur, as encountered in cardiopulmonary disorders associated with asphyxia (eg, meconium aspiration syndrome, pulmonary artery hypertension). Thus, the purpose of this investigation was to 1) compare the change in brain temperature during head and body cooling, 2) determine the change in brain temperature during the 2 modes of cooling in response to systemic hypoxemia, and 3) determine whether the mode of brain cooling is associated with differences in cerebral blood flow (CBF).

	METHODS

Top Abstract Methods Results Discussion References

Seventeen newborn Yucatan miniature swine (age: 5 ± 3 days; weight: 1.4 ± 0.4 kg) were studied after approval by the Institutional Animal Care and Research Advisory Committee. Nine animals were used to delineate the change in brain temperature during head and body cooling. These animals were premedicated with ketamine (15 mg/kg intramuscularly), followed by rapid tracheostomy and ventilation using a small animal respirator with nitrous oxide and oxygen (70%/30%). All surgical sites were infiltrated with 1% Xylocaine (Steris Laboratories Inc, Phoenix, AZ), and intravascular catheters were placed in the left common carotid artery and the left external jugular vein. Nalbuphine (0.15 mg/kg intravenously) and D-tubocurarine Cl (0.2 mg/kg intravenously) were administered for analgesia and muscle relaxation, respectively. Pentothal (20 mg/kg intravenously) was administered followed by retraction of the scalp and creation of bilateral burr holes located 0.75 cm lateral to the sagittal suture and 0.25 cm posterior to the coronal suture. Temperature probes (Physitemp Instruments, Clifton, NJ) were positioned in the brain (2 and 1 cm deep, corresponding to cortical gray matter and basal ganglia, respectively), on the dural surface, on the calvarium, on the skin overlying the calvarium, and in the ambient air 5 cm above the head. Burr holes were sealed with bone wax, probes were exteriorized through the skin, and the retracted scalp was reapproximated and sutured closed. Temperature probes were also placed in the rectum (5-cm deep). The body was wrapped in a thermostatically controlled warming blanket, which provided control of core body temperature. Animals were allowed to stabilize for 90 minutes, and nalbuphine was given every 4 hours for the duration of the study.

Eight additional animals were used to determine brain blood flow changes associated with different modes of brain cooling. Preparation was the same as above except that catheters were positioned in the left ventricle via the left common carotid artery, the mid- and lower abdominal aorta via the right and left femoral arteries, the superior sagittal sinus (after creation of a midline 4-mm diameter burr hole over the sagittal suture), and the superior vena cava via the right external jugular vein.

After 90 minutes of stabilization, baseline measurements were acquired during a control period (20 minutes). After control measurements, animals underwent either head or body cooling. Head cooling was achieved by positioning inflatable cuffs around the top half of the head and circulating ice water (temperature ~1°C) through the cuffs. This water temperature was used because it resulted in brain temperatures at 2- and 1-cm depths that were above and below, respectively, the brain temperature of the body cooling group. Thus, brain temperature of each group was reduced to values that would be considered modest hypothermia. Head cooling was continued until temperatures at each of the sites were stable over a 30-minute interval (typical total duration of cooling was 70-80 minutes). A thermal blanket wrapped around the body was used during head cooling to maintain normothermic body temperature (~38.5°C) by circulating warm water (40-45°C) through the blanket. Head cooling was followed by 15 minutes of head cooling with superimposed hypoxia (inhalation of 8% oxygen). Animals were then allowed to recover for 1 hour under normothermic and normoxic conditions. A second control period followed, and animals were exposed to the same sequence except with body cooling alone and subsequent superimposed hypoxia. Body cooling was achieved by circulating cool water (initially 1°C followed by an increase to 33°C) through the thermal blanket, which did not contact the head. The order of head and body cooling was alternated between successive animals studied.

Mean arterial pressure (MAP) and heart rate were monitored continuously. Arterial blood was sampled for pH, blood gases, and plasma lactate and glucose concentrations during each control period, at the completion of each interval of cooling (head or body), and at 12 minutes of hypoxia superimposed on head or body cooling. In the 9 piglets that did not undergo blood flow measurements, an additional arterial pH and blood gas measurement was obtained at 3 minutes of hypoxia to demonstrate a steady state of PO₂ reduction during hypoxemia. In the 8 piglets that underwent blood flow measurements, blood sampling was identical to the above except that samples obtained during cooling with superimposed hypoxia were taken at 10 minutes of hypoxia (because of the blood flow measurement), and cerebral arteriovenous differences of O₂ content were included at each blood sampling. Temperatures were recorded from the sites listed above during control, at 10-minute intervals during cooling, and at 3-minute (non-blood flow group) or 5-minute (blood flow group) intervals during hypoxia superimposed on cooling.

Blood flow was measured using fluorescent microspheres. The method is based on an adaptation of the radioactive-labeled microsphere technique, which has been described in detail¹² and was used previously in this laboratory.¹³ Fluorescent microspheres (Molecular Probes Inc, Eugene, OR) are 15 µm in diameter, and each microsphere contains a single fluorescent dye that is spectrally distinct. The microspheres were diluted with 0.9% NaCl, stirred by vortexing, and infused into the left ventricular catheter over 20 to 30 seconds. Starting 15 to 20 seconds before and continuing during and after each microsphere infusion, an arterial reference sample was withdrawn from the abdominal aorta into vials at a rate of 1.03 mL/min for 2 minutes using a calibrated pump. Six different microspheres were injected in each study during control, the last 5 minutes of cooling, and the last 5 minutes of cooling with superimposed hypoxia for both head and body cooling. At study completion, the animal was killed and the brain was removed and stored in 10% buffered formalin. The entire brain was cut into 1- to 2-g sections with the cerebral cortex grouped into regions located 1, 2, or 3 cm beneath the brain cortical surface, which corresponded to cortical gray, subcortical white matter, and basal ganglia/thalamus, respectively. The remainder of the brain was sectioned and grouped as brainstem and cerebellum. The fluorescence of both brain and blood samples was determined by separating the entrapped microspheres from the tissue and reference blood sample (chemical digestion with ethanolic KOH), extraction of the fluorescent dye by dissolving the microspheres, and quantitation of fluorescence (SLM Aminco-Bowman Series 2 Luminescence Spectrometer, SLM Aminco, Urbana, IL). Blood flow was computed as previously described,¹² and wet tissue weight was used to express blood flow as milliliters per minute per 100 g. CBF was derived from the fluorescence of all tissue samples of the cerebral hemispheres. CBF was used to calculate cerebral O₂ uptake (CMRO₂) as the product of CBF and the cerebral arteriovenous difference of O₂ content. We previously compared brain blood flow over a wide range of values in 2 piglets measured with radioactive microspheres (CBF_R) and fluorescent microspheres (CBF_F). There was an excellent correlation between the 2 methods (r² = 0.96), although CBF_F systemically underestimated CBF_R (regression line, CBF_F = 0.10 + 0.76 CBF_R). These data support the use of fluorescent microspheres to measure relative changes in CBF accurately.

All results were expressed as X ± SD. Body and head cooling were analyzed by examining results from 3 time points for each mode of cooling (control, cooling, and cooling with hypoxia). Values during cooling were those recorded after 30 minutes of constant temperature, and values during cooling with hypoxia were those recorded at the completion of the 15-minute interval of hypoxia. An analysis of variance with repetitive measurements (SPSS version 10.0, SPSS, Inc, Chicago, IL) was used to determine differences between the 2 modes of cooling and among the 3 values (control, cooling, cooling with hypoxia) for each mode of cooling. Variables analyzed included temperature, blood flow, and systemic hemodynamic and biochemical variables. Significant differences were localized with a multiple comparisons test using Student-Newman-Keuls and Duncan procedures. Regression analysis and analysis of covariance were used to examine the relationship between CBF and CMRO₂. Significance was designated at P < .05.

	RESULTS

Top Abstract Methods Results Discussion References

During the control period, the head of the animal was exposed to ambient laboratory temperature (~23°C), and there were small temperature gradients across the brain: 0.2 ± 0.4°C and 1.3 ± 1.1°C for the difference between the brain 2-cm and 1-cm sites and the brain 2-cm and dura sites, respectively. Brain temperatures were recorded for each mode of brain cooling under steady-state conditions with systemic and brain temperatures constant for at least 30 minutes. The time to reach steady-state temperatures during head and body cooling was variable from animal to animal; therefore, temperatures were plotted during the first 30 minutes of cooling and during the final 30 minutes of cooling when temperatures had become constant. Head cooling (Fig 1, top) resulted in a reduction in temperature at all sites, with larger falls in temperature for the brain 1-cm depth compared with the deeper brain depth. When hypoxia was superimposed on head cooling, brain temperature increased at all brain sites. The magnitude of the temperature gradient across the brain was quantified using the difference between the 2-cm depth and the dura (Fig 1, bottom). During head cooling, the gradient increased from 1.3 ± 1.1°C at control to 7.2 ± 3.5°C. When hypoxia was superimposed on head cooling, the temperature gradient decreased to 4.4 ± 2.4°C, indicating a reduction in the extent of brain cooling. All 3 temperature gradients (control, cooling, and cooling with hypoxia) were different from each other (P < .05). During body cooling (Fig 2, top), rectal temperature was reduced by 4°C and similar magnitude of changes in temperature were measured at all brain sites. In contrast to head cooling, hypoxia superimposed on body cooling was not associated with any increase in brain temperature. The temperature gradient across the brain during body cooling (Fig 2, bottom) was similar to control, 1.5 ± 1.2°C to 1.1 ± 0.9°C, for control and body cooling, respectively, indicating homogeneous cooling of the brain. During hypoxia superimposed on body cooling, the temperature gradient remained similar to control. The temperature gradient differed between head and body cooling during cooling alone and cooling with hypoxia (P < .05).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Top, Temperature of the rectum, brain, dura, and scalp during control, head cooling, and head cooling with superimposed hypoxia. Temperature was monitored from 5 sites in each animal (n = 9), and the symbol for each site is listed in the key. Temperatures before and after the break in the time axis represent values during the first 30 minutes of cooling and during the final 30 minutes of cooling. Bottom, Change in temperature gradient between brain (2-cm depth) and the dura during control, head cooling, and head cooling with hypoxia. Temperatures used to derive the gradients were recorded after 30 minutes of constant temperatures during head cooling and at the completion of hypoxia superimposed on cooling.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Top, Temperature of the rectum, brain, dura, and scalp during control, body cooling, and body cooling with superimposed hypoxia. Symbols for the different sites monitored are indicated in the key and are identical to Fig 1. Bottom, Change in temperature gradient between brain (2-cm depth) and the dura during control, body cooling, and body cooling with hypoxia. Temperatures used to derive gradients were identical to Fig 1.

To illustrate further the differences between head and body cooling, the temperature of multiple brain depths was plotted during cooling and during hypoxia superimposed on each mode of cooling (Fig 3). For head cooling, the temperature of the dura, brain at a 1-cm depth, and brain at a 2-cm depth all increased when hypoxia was superimposed on cooling compared with cooling alone (P < .05). In contrast, brain temperature during body cooling was constant when hypoxia was superimposed on cooling.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Temperature of the dura, brain 1-cm site, and brain 2-cm site during head and body cooling and hypoxia superimposed on each mode of cooling. , head-cooled animals;  body-cooled animals. Values plotted are from the completion of head and body cooling (ie, 130 minutes in Figs 1 and 2, top) and after 15 minutes of superimposed hypoxia. For head cooling, the temperatures of the dura, brain 1-cm site, and brain 2-cm site all increase when hypoxia is superimposed on cooling compared with cooling alone (*P < .05).

Systemic variables were examined to determine whether animals were in similar physiologic states while undergoing head versus body cooling. At control, both head and body cool groups had values of pH (7.40 ± 0.06 and 7.39 ± 0.07, respectively), PaO₂ (131 ± 37 and 129 ± 32 mm Hg), PaCO₂ (39.0 ± 4.5 and 37.7 ± 5.9 mm Hg), MAP (79 ± 15 and 77 ± 20 mm Hg), and heart rate (231 ± 34 and 224 ± 52 beats per minute) that were similar to previously reported control data for this model.⁴ These values were not changed by head or body cooling. Blood gases and arterial pH obtained at 3 minutes of hypoxia superimposed on cooling did not differ from values obtained at 12 minutes of hypoxia. Superimposed hypoxia resulted in a similar reduction in pH (7.28 ± 0.07 and 7.26 ± 0.07) and PaO₂ (24 ± 2 and 25 ± 3 mm Hg) for head and body cooling, respectively, whereas PaCO₂ and MAP were unchanged. Heart rate was lower when body cooling with hypoxia was compared with head cooling with hypoxia (206 ± 21 vs 252 ± 14 beats per minute; P < .05). Cooling was associated with reductions in arterial lactate concentrations (1.6 ± 0.6 to 1.4 ± 0.5 mM and 1.7 ± 0.8 to 1.2 ± 0.6 mM for control to head and body cooling respectively; P < .05 vs control). Superimposed hypoxia led to similar increases in lactate (5.8 ± 1.2 and 5.1 ± 1.5 mM for head and body cooling, respectively; P < .05 vs control). There were no differences between head and body cooling during each study interval for plasma glucose concentration (data not shown).

Cooling of the brain either by head or body cooling resulted in decreases in CBF that were of similar magnitude (Table 1). Furthermore, hypoxia superimposed on either mode of cooling resulted in similar increases in CBF. In each cooling mode, the 3 values of CBF differed (P < .05). Other regions of the brain (cerebellum and brainstem) followed the same pattern of change as CBF during head and body cooling and with superimposed hypoxia. Because the temperature profile of the cerebral cortex differed between head and body cooling, CBF was examined in each centimeter of tissue beneath the cortical surface (Table 2). There was no difference between CBF of head and body cooling at any tissue depth. There was a trend for CBF of the first centimeter of cortex in head cooling to decrease to a lower CBF (35% of control) compared with the second and third centimeters of tissue (47% and 56% of control, respectively; P = .07). In contrast, CBF during body cooling was uniformly reduced to ~60% of control for the first, second, and third centimeters of tissue from the cortical surface.

During cooling, the reduction in CBF was paralleled by similar reductions in CMRO₂. For both head and body cooling, there were direct linear correlations between CBF and CMRO₂, and the slopes and intercepts of these correlations were similar (Fig 4). The parallel decreases in CBF and CMRO₂ were also indicated by an unchanged ratio of CBF/CMRO₂ for both head and body cooling (18.2 ± 2.4 to 18.2 ± 3.5, 16.7 ± 4.1 to 15.6 ± 2.5 for control and either head or body cooling, respectively). Animals that were used for blood flow measurements showed the same temporal profile and magnitude of change in temperatures and systemic physiologic variables as the animals without microsphere injections.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   The effect of brain cooling on the relationship between CBF and CMRO2. Results during control and brain hypothermia are plotted for head cooling (open squares, solid line) and body cooling (solid circles, dashed line). During each mode of brain cooling, a direct linear correlation was found described by the following: for head cooling, r2 = 0.92 (P < .001) and CBF = 16.7 CMRO2 + 3.4; for body cooling, r2 = 0.89 (P < .001) and CBF = 13 CMRO2 + 10.

	DISCUSSION

Top Abstract Methods Results Discussion References

Results of in vivo laboratory work using animal models support modest reduction in brain temperature as a promising neuroprotective strategy for acute brain injury. Clinical trials have now been initiated in human newborns to determine the efficacy of modest hypothermia for asphyxial injury at birth. The mode of brain cooling may be an important determinant of the efficacy of modest hypothermia if thermal characteristics of the cooled brain differ substantially, depending on the method used to achieve modest brain hypothermia. Other considerations that may influence the mode of brain cooling include the feasibility of achieving the desired temperature, effect of systemic hemodynamic events on brain temperature, monitoring of the brain temperature, and associated adverse effects of hypothermia.

Temperature Gradients

A number of animal studies have examined selective head cooling to achieve brain hypothermia and avoid core body hypothermia with its well-recognized associated problems.^14-16 Most methods used to cool the head also result in cooling of the body irrespective of whether the animal is small (8-day old rat pups¹⁵ or large (cat and pigs,^14,16) and preclude assessment of brain hypothermia by head cooling alone. Gelman et al¹⁷ selectively cooled the brain without a change in core body temperature of 2- to 3-week-old swine after resuscitation from cardiac arrest and measured brain temperatures at 0.5 cm beneath the cortical surface and in the caudate nucleus. Selective brain cooling over 45 minutes resulted in progressively larger increases in the temperature gradient between the deep and superficial brain sites, but measurements were not obtained during steady-state conditions.¹⁷ Head cooling after asphyxia in term human newborns has been examined in a pilot study,⁹ and the cooling protocol was based on fetal sheep experiments using circulating water at 10°C passed through a coil of tubing wrapped around the head.^7,8 One experiment in fetal sheep was conducted to characterize this mode of brain cooling; the temperature gradient increased from deep to superficial brain (magnitude not reported), and core body temperature was mildly reduced.⁷ Although brain temperature gradients are expected on the basis of principles of thermal conduction, the magnitude of the gradient is a complex function dependent on multiple variables that determine brain temperature, including tissue heat production, local blood flow, perfusing blood temperature, insulating tissue, ambient air temperature, nasal mucosa temperature, and non-central nervous system cranial blood flow. The temperature gradient and associated physiologic effects with selective head cooling have not been well characterized or compared with an equivalent episode of whole-body cooling.

This investigation used newborn swine to determine the thermal characteristics of the brain during head and body cooling. We have used this model to demonstrate the neuroprotective effect of modest hypothermia during and immediately after brain ischemia.^4,6 These experiments were designed to compare body and selective head cooling as 2 distinct modes of brain hypothermia, using decreases in temperature consistent with modest hypothermia, and achieving constant temperatures. Head cooling performed with a constant rectal temperature increased the temperature gradient across the brain with the deep brain warmer than superficial brain. To achieve modest reductions in brain temperature (~2-3°C) of deep cortical and thalamic structures (brain 2-cm depth), the surface of the brain (approximated by the dura temperature) had to be cooled by almost 10°C. In contrast, body cooling to a rectal temperature of ~34°C did not change the temperature gradient between the brain 2-cm site and the dura compared with control. Brain hypothermia induced via body cooling resulted in homogeneous cooling of the cerebral cortex, whereas head cooling exaggerated temperature gradients between deeper brain and the brain surface. Given the extent of head surface hypothermia necessary to cool the deep brain structures, mild reduction in body temperature would facilitate some cooling of the deeper brain. These observations provide a firm rationale for clinical trials that combine less intense head cooling (eg, 10°C⁹) with some decrease in core body temperature. Consistent with the latter is the heat sink model applied to a 3-dimensional model of the infant head to examine temperature distributions in response to surface cooling at 10°C.¹⁸ The model demonstrated that surface cooling reduces deep brain temperature only when core body temperature is lowered.

Cerebral Blood Flow

This investigation also compared brain blood flow changes associated with head and body cooling. Whole-body cooling to produce brain hypothermia in animals and humans (including children and adults) consistently reduces CBF.^19-22 Furthermore, the reductions in CBF are coupled to decreases in metabolic rate, the latter correlated with the degree of temperature drop. The results from body cooling in the present investigation (Fig 4; Table 2) are consistent with these observations. In contrast, the effect of selective brain cooling on CBF is less consistent. Gelman et al¹⁷ used microspheres and demonstrated an unchanged CBF at 15 and 45 minutes of selective brain cooling after cardiac arrest and CPR. However, it is difficult to separate the effects of selective brain cooling and postischemic changes. In 6-week-old pigs, selective brain hypothermia (brain temperatures of 25°C and 30°C) via bicarotid perfusion with extracorporally cooled blood decreased CBF and CMRO₂ to a degree comparable to whole-body cooling.²³ Walter et al²³ speculated that the cerebrovascular response to selective brain cooling may depend on the method of cooling the head (surface cooling vs perfusion with cooled blood) because of the presence or absence of centripetal temperature gradients within the brain. Specifically, surface cooling and the resulting temperature gradients may lead to such profound drops in the peripheral brain temperature that vasoparalysis, loss of autoregulation, and an increase in cortical blood flow occur. Perfusion of the head alone with cooled blood using an extracorporeal circuit results in more homogeneous brain cooling and reductions in CBF.²³ The results from the present investigation show this not to be the case. The relationship between CBF and CMRO₂ during brain cooling was not affected by the mode of cooling and the presence or absence of temperature gradients (Fig 4). Furthermore, the direction of effect for head cooling on CBF was for larger reductions in the cooler first centimeter of tissue beneath the cortex compared with the warmer central brain (Table 2).

Hypoxia During Cooling

An important characteristic of brain cooling is maintenance of a constant temperature, because fluctuations of as little as 1°C to 2°C in brain temperature can modify neurologic outcome after hypoxia-ischemia.²⁴ This is pertinent for asphyxiated newborns with multiorgan dysfunction, whereby the brain and the brain temperature may be affected by direct injury and systemic events.²⁵ Therapeutic hypothermia may exacerbate pulmonary disorders because lower temperature increases pulmonary vascular resistance. However, clinicians may need to decide whether brain hypothermia is to be used (at present entered into trials) when hypoxic-ischemic encephalopathy is associated with meconium aspiration syndrome and pulmonary artery hypertension, depending on the specifics of each patient. In one pilot study of brain cooling, the inspired fraction of oxygen had to be increased by a median of 0.14 to maintain oxygenation.¹⁰ Given these considerations, the current investigation examined head and body cooling when hypoxia occurs. Hypoxia superimposed on head cooling resulted in elevations of brain temperature at all sites with the largest increases, ~4°C, noted for the superficial brain (Fig 3). In contrast, a constant brain temperature was achieved when hypoxia was superimposed on body cooling. Animals that underwent head and body cooling had similar increases in CBF in response to reductions in arterial O₂ content.²⁶ The difference in brain temperature between head and body cooling with superimposed hypoxia therefore reflected delivery of either warm or cool blood, respectively, from the body to the head via hypoxia-induced increase in CBF. Contrary to the speculation of Walter et al²³ that vasoparalysis occurs, CBF during superimposed hypoxia indicated vasodilation during head cooling, even in the peripheral brain. These observations are consistent with reports that hypercapnia, hypoxia, or hypoperfusion result in vasodilation under conditions of reduced metabolic rate (hypothermia, barbiturates^20,27,28). In each group, rectal temperature was well maintained during hypoxia. It is possible that the increase in brain temperature when hypoxia was superimposed on head cooling may be less in a nursery setting because mechanisms to maintain core temperature may be less efficient than used in this study.

Monitoring Brain Temperature

The ability to monitor brain temperature is a salient aspect of brain cooling as a therapeutic intervention. The temperature gradients across the brain during head cooling preclude a single non-central nervous system site as a valid index of brain temperature. Furthermore, there was greater variability in the temperatures during head cooling compared with body cooling, as indicated by the standard deviations (Figs 1 and 2, top). During body cooling, though, rectal temperature seems to be a valid index of brain temperature in view of the small temperature gradients. In adult humans undergoing profound whole-body hypothermia either with or without cardiopulmonary bypass, the esophageal temperature is the best index of brain temperature, and rectal temperature mirrored esophageal measurements.^29,30 Monitoring both brain and core temperature is important given the adverse effects associated with hypothermia. Multiorgan system dysfunction arising from systemic cold injury has been reviewed.³¹ Selective head cooling obviates these hazards and is attractive for asphyxiated infants because systemic manifestations of cold injury and hypoxia-ischemia can be similar. If the goal of therapeutic brain hypothermia is to reduce brain temperature by at least 2 to 3°C, then the temperature of the outermost layer of brain may need to be reduced to <30°C to cool deeper brain structures sufficiently. The latter represents a temperature at which the frequency of adverse effects increase, especially when used over a prolonged time interval (days compared with hours^32,33). However, it is unclear whether similar adverse effects would occur when a sustained temperature reduction to <30°C is localized to the surface of the head.

This investigation demonstrated that there are multiple distinguishing features between brain hypothermia induced by either head or body cooling. The differences include the extent of brain temperature gradients, the effect of systemic events such as hypoxia on the maintenance of a target temperature, and the ability to use a systemic site for temperature measurements as an indicator of brain temperature. Which mode of brain cooling should be used for clinical trials is a difficult question because there are advantages and disadvantages with each. These differences could affect the extent of neuroprotection provided by modest hypothermia.

	ACKNOWLEDGMENTS

We thank Damian Garcia for help in performing the experiments, Karen Kirby for secretarial expertise, and the support of the Department of Pediatrics.

	FOOTNOTES

Received for publication Feb 1, 2001; accepted May 14, 2001.

Reprint requests to (A.R.L.) Department of Pediatrics, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-9063. E-mail: abbot.laptook{at}utsouthwestern.edu

	ABBREVIATIONS

CBF, cerebral blood flow; MAP, mean arterial pressure; CMRO₂, cerebral O₂ uptake.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Paneth N. The causes of cerebral palsy. Recent evidence. Clin Invest Med. 1993;95-102
2. Vannucci RC, Perlman JM Interventions for perinatal hypoxic-ischemic encephalopathy. Pediatrics 1997; 100:1-11 [Abstract/Free Full Text]
3. Gunn AJ, Gunn TR The pharmacology of neuronal rescue with cerebral hypothermia. Early Hum Develop 1998; 53:19-35
4. Laptook AR, Corbett RJT, Sterett R, Burns DK, Tollefsbol G, Garcia D Modest hypothermia provides partial neuroprotection for ischemic neonatal brain. Pediatr Res 1994; 35:436-442 [Medline]
5. Thoresen M, Bagenholm R, Loberg EM, Apricena F, Kjellmer I Posthypoxic cooling of neonatal rats provides protection against brain injury. Arch Dis Child 1996; 74:F3-F9
6. Laptook AR, Corbett RJT, Sterett R, Burns DK, Garcia D, Tollefsbol G Modest hypothermia provides partial neuroprotection when used for immediate resuscitation after brain ischemia. Pediatr Res 1997; 42:17-23 [Medline]
7. Gunn AJ, Gunn TR, de Hann HH, Williams CE, Gluckman PD Dramatic neuronal rescue with prolonged selective head cooling after ischemia in fetal lambs. J Clin Invest 1997; 99:248-256 [Medline]
8. Gunn AJ, Gunn TR, Gunning MI, Williams CE, Gluckman PD Neuroprotection with prolonged head cooling started before postischemic seizures in fetal sheep. Pediatrics 1998; 102:1098-1106 [Abstract/Free Full Text]
9. Gunn AJ, Gluckman PD, Gunn TR Selective head cooling in newborn infants after perinatal asphyxia: a safety study. Pediatrics 1998; 102:885-892 [Abstract/Free Full Text]
10. Thoresen M, Whitelaw A Cardiovascular changes during mild therapeutic hypothermia and rewarming in infants with hypoxic-ischemic encephalopathy. Pediatrics 2000; 106:92-99 [Abstract/Free Full Text]
11. Azzopardi D, Robertson NJ, Cowan FA, Rutherford MA, Rampling M, Edwards AD Pilot study of treatment with whole body hypothermia for neonatal encephalopathy. Pediatrics 2000; 106:684-694 [Abstract/Free Full Text]
12. Heymann MA, Payne BD, Hoffman JIE, Rudolph AM Blood flow measurements with radionuclide-labelled particles. Prog Cardiovasc Dis 1977; 20:55-79 [CrossRef][Medline]
13. Laptook AR, Corbett RJT, Nguyen HT, Peterson J, Nunnally RL Alterations in cerebral blood flow and phosphorylated metabolites in piglets during and after partial ischemia. Pediatr Res 1988; 23:206-211 [Medline]
14. Horn M, Schlote W, Henrich HA Global cerebral ischemia and subsequent selective hypothermia. Acta Neuropathol 1991; 81:443-449 [CrossRef][Medline]
15. Towfighi J, Housman C, Heitjan DF, Vannucci RC, Yager JY The effect of focal cerebral cooling on perinatal hypoxic-ischemic brain damage. Acta Neuropathol 1994; 87:598-604 [Medline]
16. Tadler SC, Callaway CW, Menegazzi JJ Noninvasive cerebral cooling in a swine model of cardiac arrest. Acad Emerg Med 1998; 5:25-30 [Medline]
17. Gelman B, Schleien CL, Lohe A, Kuluz JW Selective brain cooling in infant piglets after cardiac arrest and resuscitation. Crit Care Med 1996; 24:1009-1017 [CrossRef][Medline]
18. Van Leeuwen GMJ, Hand JW, Lagenduk JJW, Azzopardi DV, Edwards AD Numerical modeling of temperature distributions within the neonatal head. Pediatr Res 2000; 48:351-356 [Medline]
19. Michenfelder JD, Milde JH The relationship among canine brain temperature, metabolism and function during hypothermia. Anesthesiology 1991; 75:130-136 [Medline]
20. Hagerdal M, Harp J, Nilsson L, Siesjo BK The effect of induced hypothermia upon oxygen consumption in the rat brain. J Neurochem 1975; 24:311-316 [Medline]
21. Busija DW, Leffler CW Hypothermia reduces cerebral metabolic rate and cerebral blood flow in newborn pigs. Am J Physiol 1987; 253:H869-H873 [Abstract/Free Full Text]
22. Greely WJ, Kern FH, Ungerleider RM, The effect of hypothermic cardiopulmonary bypass and total circulatory arrest on cerebral metabolism in neonates, infants, and children. J Thorac Cardiovasc Surg 1991; 101:783-794 [Abstract]
23. Walter B, Bauer R, Kuhnen G, Fritz H, Zwiener U Coupling of cerebral blood flow and oxygen metabolism in infant pigs during selected brain hypothermia. J Cereb Blood Flow Metab 2000; 20:1215-1224 [CrossRef][Medline]
24. Wass CT, Lanier WL, Hofer RE, Scheithaner BW, Andrews AG Temperature changes of 1°C alter functional neurologic outcome and histopathology in a canine model of complete cerebral ischemia. Anesthesiology 1995; 83:325-335 [CrossRef][Medline]
25. Perlman JM, Tack ED, Martin T, Shackelford G, Amon E Acute systemic injury in term infants after asphyxia. Am J Dis Child 1989; 143:617-620 [Abstract/Free Full Text]
26. Jones MD, Traystman RJ, Simmons MA, Molteni RA Effects of changes in arterial O₂ content on cerebral blood flow in the lamb. Am J Physiol 1981; 240:H209-H215 [Abstract/Free Full Text]
27. Prough DS, Stump DA, Roy RC, Response of cerebral blood flow to changes in carbon dioxide tension during hypothermic cardiopulmonary bypass. Anesthesiology 1986; 64:576-581 [CrossRef][Medline]
28. Donegan JH, Traystman RJ, Koehler RC, Jones MD, Rogers MC Cerebro-vascular hypoxic and autoregulatory responses during reduced brain metabolism. Am J Physiol 1985; 249:H421-H429 [Abstract/Free Full Text]
29. Whitby JD, Dunkin LJ Cerebral, oesophageal and nasopharyngeal temperatures. Br J Anaesth 1971; 43:673-676 [Abstract/Free Full Text]
30. Stone JG, Young WL, Smith CR, Do standard monitoring sites reflect true brain temperature when profound hypothermia is rapidly induced and reversed? Anesthesiology 1995; 82:344-351 [Medline]
31. Laptook AR, Corbett RJT Therapeutic hypothermia: a potential neuroprotective and resuscitative strategy for neonatal hypoxia-ischemia. Prenat Neonat Med 1996; 1:199-212
32. Steen PA, Soule EH, Michenfelder JD The detrimental effect of prolonged hypothermia in cats and monkeys with and without regional cerebral ischemia. Stroke 1979; 10:522-529 [Abstract/Free Full Text]
33. Weinrauch V, Safar P, Tisherman S, Kuboyama K, Radovsky A Beneficial effect of mild hypothermia and detrimental effect of deep hypothermia after cardiac arrest in dogs. Stroke 1992; 23:1454-1462 [Abstract/Free Full Text]