INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

WHOLE BODY HYPERTHERMIA ACTIVATES thermoregulatory responses involving both systemic and central responses, including vasodialation, nausea, and altered cognitive function. Animal studies have identified several regions within the brain that may control or modulate these responses. For example, the preoptic area/anterior hypothalamus (POAH) is regarded as the center for integration of thermal signals from the periphery and is sensitive to local temperature changes (1, 10, 32). In rodents, raising the temperature of the POAH activates cooling mechanisms such as saliva grooming, body extension (22), and panting (19, 28). However, rather than being the sole site of integration, the POAH may coordinate the activity of other integrating mechanisms at lower levels of the neuroaxis (3, 24).

Although there have been studies of human cerebral response to localized, noxious thermal stimuli (4), a search of the literature revealed no published studies of human cerebral metabolism during whole body systemic heating. Knowing which brain regions have altered metabolism during whole body hyperthermia is important for understanding the neuronal mechanisms responsible for altered cognitive and systemic processes. It would seem plausible that in humans and animals there would be similar central responses to hyperthermia. However, differences exist in the cognitive complexities of humans and animals and in the ability of humans and animals to sweat. Only a few animals (e.g., padus monkeys, horses) possess the ability to sweat. With the recent advancements in cerebral imaging techniques, the capability now exists to image changes in brain activity in the human during a whole body hyperthermic period. In the present experiment, positron emission tomography (PET) was used to examine regional changes in cerebral metabolic rate in humans during steady-state hyperthermia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The experiments were conducted in the Research Imaging Center at the University of Texas Health Sciences Center at San Antonio (UTHSCSA). The protocol was approved by the Institutional Review Boards at UTHSCSA and Brooks Air Force Base, as well as by the Radioisotope Safety Committee at UTHSCSA, and each subject gave written, informed consent before participating.

Subjects were three women and seven men, 20-39 yr of age, screened to ensure that each was in good health with no history of heat intolerance. All were lean in appearance; their height averaged 176 cm (range 168-189 cm) and weight 72 kg (range 53-81 kg). All engaged in some level of regular physical activity, they were heat acclimated to the extent of living in south Texas, but they were not highly trained athletes. Women stated that they were not pregnant, and this was confirmed on the morning of the experiment by means of a negative urine pregnancy test. Subjects were familiarized with equipment on an initial visit to the laboratory, at which time magnetic resonance imaging of the head was obtained to establish individual anatomic landmarks for later image processing.

Heating and respiratory measurements. Whole body systemic hyperthermia was induced by using a liquid-conditioned garment (LCG) consisting of a suit of long underwear lined with six runs of small-diameter plastic tubing (Carleton Technology, Tampa, FL). Water from a temperature-controlled bath was pumped through the suit at 1 l/min. Heat loss was minimized by using a rain suit to prevent evaporation of sweat and adding external insulation (blankets) to limit convective cooling. Each subject's temperature was monitored by using a rectal probe. Although this site lags esophageal and oral temperatures under dynamic conditions, it has been used in a variety of heating studies and was deemed suitable for this profile, which involved slow heating to a predetermined level followed by maintenance at that level. Furthermore, it was not practical to request that the subjects tolerate an esophageal thermistor while wearing the PET head restraint and possibly feeling nauseous.

Cerebral imaging and time line. The subject reported to the laboratory at 0730 in a fasting condition, ate a snack of bread and crackers, inserted the rectal thermistor to a depth of 10 cm, and then dressed in shorts and the LCG. The subject then lay supine on the PET bed for insertion of venous and arterial cannulas that remained in place throughout the day. The subject was kept thermally comfortable during the 2-h morning (control) session, which included collection of expired gas, fitting of a PET head restraint, a transmission scan used to correct for photon attenuation, and imaging. Each PET image required an intravenous injection of a bolus of [¹⁸F]fluorodeoxyglucose followed by a 20 min uptake period and a 10-min period of data collection during which the subject lay still in a quiet, dimly lit room.

Image processing. Individual PET images were spatially normalized into bicommissural coordinate space by employing a nine-parameter affine transformation derived from 14 structural landmarks (13, 29). Within-subject changes in cerebral metabolic rate were measured by using pixel-by-pixel subtraction (cool from hot). Data for all subjects were then merged, and an average data set was computed (6-8, 17). Statistical parametric mapping was used to identify voxels that had statistically significant changes (5). Local maxima and minima with a Z score exceeding 2.0 (P < 0.05, 2 tailed) were identified by the location of their peak voxel (search cube volume of 125 mm³). The database was cleared of significant voxels resulting from edge artifacts located along the inside edge of the ventricle and at the outside edge of the brain.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Heating and respiratory measurements. Figure 1 displays thermal data for a typical experiment. Heating produced an immediate rise in skin temperature accompanied by a small, paradoxical decline in rectal temperature. After rectal temperature began to rise, it followed a linear course with a slope that averaged 0.03°C/min. The time required to reach the target level averaged ~90 min. As rectal temperature approached 38.0°C, subjects experienced profuse sweating and awareness of their rapid heart rate. Water bath temperature was then lowered to reduce discomfort, control overshoot, and stabilize rectal temperature during data collection. Nine subjects completed the protocol, whereas a tenth was unable to maintain immobility for the final scan. Heart rate was counted by hand and recorded at 10-min intervals, in part to monitor the health and safety of the subjects. Mean heart rate immediately before the cool scans for nine subjects was 62 beats/min (range 60-66 beats/min), whereas the value during steady-state heating was 124 beats/min (range 116-120 beats/min) with one subject at 160 beats/min.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Plot of suit (Suit Temp) and rectal (Tre) temperatures during heating for a single subject. M and Gluc, times of measurement for whole body metabolism (spirometry) and brain metabolism (tagged glucose), respectively.

Cerebral metabolism. Analysis of cerebral metabolic rate showed a number of significant changes with heating (see Table 3 and Fig. 2). Significant increases were located in the hypothalamus, thalamus, corpus callosum, cingulate gyrus, and cerebellum. In contrast, significant decreases were noted in the caudate, putamen, insula, and posterior cingulum. Technical difficulties in the majority of our subjects interfered with the sequential arterial sampling. Thus the limited data set did not permit any meaningful statistical analysis related to changes in global brain metabolism.

View this table: [in this window] [in a new window]   Table 3.   Activation or deactivation of regional [18F]fluorodeoxyglucose labeling having a highly significant Z score magnitude of >2.8 or less than 2.8 during systemic hyperthermia

View larger version (83K): [in this window] [in a new window]   Fig. 2.   Significant decreases (A) and increases (B-D) in [18F]fluorodeoxyglucose activity during whole body hyperthermia. Coordinates in Z plane according to the atlas of Talairach and Tournoux (29) are the following: A and B, 18; C, 4; D, 26.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Elevation of rectal temperature by 1.5°C in our subjects produced a 23% rise in resting metabolic rate, which corresponds well with Saxton's (25) finding in subjects heated to produce a 2.0°C rise measured at the tympanic membrane. Consistent with the animal literature (12, 14, 16, 18, 26), the PET data from our human subjects showed that the POAH and associated thermoregulatory centers exhibit altered levels of metabolism during systemic hyperthermia.

However, not expected from the animal literature were the results showing that there was an increase in the cerebral metabolic rate in the lateral cerebellum with heating. The medial cerebellum, consisting of the vermis and intermediate hemisphere, contains somatotropic maps of the entire body and receives sensory information from the periphery, including input from labyrinthine, visual, and auditory receptors. Increased activity in the vermis has been reported in response to forearm thermal stimulation (4). In contrast to the vermis and intermediate hemisphere, the lateral hemispheres of the cerebellum receives input exclusively from the pontine nuclei and have traditionally been viewed as being involved in the planning and initiation of movement. However, recent experiments indicate that the lateral cerebellum is also involved in acquisition and discrimination of sensory information (9).

Some of the regions showing altered cerebral metabolic rate may be due to somatosensory input. Casey et al. (4) used PET to study brain function during application of thermal stimuli to small areas of the human forearm or hand. They found that discrimination between two levels of warmth was associated with increased cerebral blood flow in the thalamus, whereas progression to a painfully hot stimulus activated both the thalamus and the cingulate gyrus. Our subjects showed increased cerebral metabolic rate in both areas. Although they did not experience pain from heating during the scans, the combination of systemic hyperthermia, warm skin, profuse sweating, and nausea together with the requirement to lie entirely still may have been sufficiently noxious to activate the cingulate area. This concept is supported by the findings of Rainville et al. (21), who found that hyponotic reduction of the affective component of painfully hot stimuli reduced activation of the anterior cingulate cortex.

Although our subjects sweated profusely and were thirsty at the end of the day, the increased cerebral metabolic rate in the hypothalamus in the present study does not appear to be due primarily to the influence of osmotic thirst. It is thought that osmoreceptors comprise portions of the hypothalamus and it is well known that the the hypothalamus releases arginine vasopressin into the posterior pituitary in response to osmotic stimulation (11, 23, 27). Using PET imaging, Denton et al. (5) showed deactivation in the anterior hypothalamus, parahippocampal, frontal gyri, caudate, and thalamus, whereas increased activation was present in the cingulate region, medial and temporal gyri, and periaqueductal gray after infusion of hypertonic saline into human subjects. The activation observed in the hypothalamus during the present study does not appear to fit that pattern.

Some brain regions had decreased cerebral metabolic rates during the hyperthermic period. Some of these regions were the caudate, putamen, and insula. Altered cerebral metabolic rates in response to sensory stimuli (e.g., hyperthermia) are consistent with known neuronal connections. The caudate putamen receives efferents from the somatosensory cortical regions and thalamus and then projects back to the cortex via the thalamus.

Our analysis also showed that the insula had a depressed cerebral metabolic rate during hyperthermia. The insula has numerous connections, including the cingulate cortex, caudate putamen, and thalamus (2). Through its connections with the amygdala, the insula provides a pathway for somatosensory information to reach the limbic system (15).

In summary, the results showed that the metabolic activity in the hypothalamus, thalamus, corpus callosum, cingulated gyrus, cerebellum in humans is increased during whole body systemic hyperthermia. In contrast, the metabolic activity of the caudate, putamen, insula, and posterior cingulum is decreased. In light of recent PET research implicating the lateral cerebellum in sensory acquisition and discrimination (9), it was interesting to find altered activity of the lateral cerebellum during a hyperthermic period. Whether the altered cerebral metabolic rate in this area was due to its role in thermoregulation or in response to somatosensory input remains unknown and warrants further investigation. Furthermore, the relationship between these changes in cerebral metabolism and cognitive processes also requires further research.