Effects of Pharyngeal Cooling on Brain Temperature in Primates and Humans: A Study for Proof of Principle

• Pharyngeal cooling decreases brain temperature by cooling the carotid arteries. This study evaluated the principle of pharyngeal cooling in monkeys and humans.

• Pharyngeal cooling rapidly and selectively decreased brain temperature in primates and tympanic temperature in humans, and did not have adverse effects on return of spontaneous circulation even when initiated during cardiac arrest in primates.

MANY laboratory data and clinical data indicate that neuroprotective effects of hypothermia are greatly reduced by delaying the onset of hypothermia.1,2To compensate for the delay in onset of hypothermia, a long duration of hypothermia is needed.3,4Although use of ice packs or intravenous infusion of cold saline has been recommended by ILCOR CoSTR5for rapid induction of hypothermia, these techniques are generally used after the return of spontaneous circulation to avoid detrimental effects of hypothermia on systemic circulation. Since neuronal damage has already commenced during cardiopulmonary resuscitation because of excess release of glutamate6and increase in oxidative stress,7development of a technique that can decrease brain temperature before return of spontaneous circulation is needed.

Compared with conventional whole body cooling, the brain, which is only about 2% of body weight, can be cooled rapidly by small heat exchange. Since bilateral common carotid arteries run along with the pharynx and upper esophagus in humans (see figure, Supplemental Digital Content 1, https://links.lww.com/ALN/A849, which shows 3-dimensional images of carotid arteries and the pharynx), cooling the pharyngeal region hematogenously decreases brain temperature in a short period of time by cooling arterial blood without lowering systemic temperature.6,8 

We have developed a pharyngeal cooling cuff that is perfused with cold water or saline at the rate of 500 ml/min and at a pressure of 50 cm H²O. The present study was designed to evaluate the effects of pharyngeal cooling on brain temperature, return of spontaneous circulation, and pharyngeal epithelium when it is initiated simultaneously with the onset of resuscitation in monkeys. In a clinical study, the effects of pharyngeal cooling on tympanic temperature, systemic circulation, and pharyngeal epithelium were evaluated.

The animals were housed and manipulated in accordance with the guide for the care and use of laboratory animals.9All experiments were approved by the Animal Research Control Committee of Okayama University Medical School and were performed in an area that has passed the regulations of the regional health care center. National Bio-resource Project, which is funded by the Ministry of Education, Culture, Sports, Science and Technology of Japan, supplied the animals after examination of the facility and the research protocol. Eleven Japanese monkeys (National BioResource Project “Japanese Monkeys,” Okazaki, Japan), weighing 8.2 ± 2.1 kg, were used in this study. Since the use of animals that have a carotid rete may result in overestimation of the effects of pharyngeal cooling, we used Japanese monkeys, which do not have a carotid rete (see figure, Supplemental Digital Content 2, https://links.lww.com/ALN/A850, which is a figure showing a cerebral angiogram of a Japanese monkey) and have a pharynx that is similar in shape and length to the human pharynx. The animals had free access to water and were fasted overnight before the experiments. To avoid postoperative infection and pain, all animals received intravenous injection of antibiotics and subcutaneous injection of local anesthetics at the surgical sites. All surgical procedures were performed with an aseptic technique according to the guidelines for the prevention of surgical site infection.10 

Anesthesia was induced by injection of ketamine (20 mg/kg, intramuscular) and was maintained with 1.5% isoflurane in 50% O²balanced with nitrogen. After placement of an intravenous catheter (ID = 0.6 mm) and injection of vecuronium (0.4 mg/kg, IV), oral tracheal intubation and artificial ventilation (Newport 100; Newport Medical Instruments, Costa Mesa, CA) were performed. A polyurethane catheter (ID = 0.6 mm) was placed in the left radial artery for continuous mean arterial blood pressure monitoring and blood sampling. Blood gases were measured (i-STAT 300F; i-STAT Corporation, Windsor, NJ) and maintained within normal ranges before the onset of cardiac arrest. Small thermocouples (disc diameter = 1 mm) were placed at the epidural space and at 3 cm below the cortical surface through burr holes at 2 cm bilateral of the sagittal line and 1 cm posterior to the bregma for continuous monitoring of surface brain temperature and core brain temperature, respectively. A laser Doppler flow probe (ALF2100; Advance, Tokyo, Japan) was placed on the parietal cortex 2 cm right of the sagittal line and 1 cm posterior to the bregma for confirmation of cardiac arrest. Rectal temperature was monitored by a thermocouple and was maintained at 36.5 ± 0.5°C using a forced-air warming blanket (Bair Hugger, Arizant, MN). A pacing catheter with an inflation balloon was inserted from the right femoral vein into the right ventricle. After a 60-min equilibration period, 12 min of cardiac arrest was initiated by electrical stimulation (60 Hz, 10 mA, 5 min; SEN-3201; Nihon Kohden, Tokyo, Japan) through the pacing catheter. Resuscitation was performed according to the guidelines of Pediatric Advanced Life Support.11Briefly, at the end of 12 min of cardiac arrest, monophasic direct current shock (2 J/kg; Cardiopac 3M33; NEC, Tokyo, Japan) was applied for defibrillation. Epinephrine (0.01 mg/kg, IV) and subsequent direct current shock (4 J/kg) was repeatedly applied every 3 min. For prolonged cardiac arrest, sodium bicarbonate was administered according to the results of arterial blood gas analysis. Ten animals were used as the pharyngeal cooling group (n = 5) and the control group (n = 5). One animal underwent pharyngeal cooling without suffering from cardiac arrest.

As shown in figure 1, the pharyngeal cooling cuff (Daiken Medical Co., Osaka, Japan) is designed to fit to the upper esophagus and pharynx. Size 2 (for body weight of 10–20 kg) and size 4 (for body weight of 50–70 kg) were used for the animal study and the clinical study, respectively. In the animal study, cold water (5°C) was perfused simultaneously with the onset of resuscitation at the rate of 500 ml/min for 30 min. In the clinical study, cold saline (5°C) was perfused at the rate of 500 ml/min for 30 min. Intracuff pressure was monitored at the outlet tube and was controlled at 50 cm H²O by applying resistance.

After the onset of resuscitation, blood pressure and an electrocardiogram were continuously monitored for 2 h. Blood gas analysis was performed every hour. After subcutaneous injection of 1% lidocaine at the surgical sites (skin incisions for bilateral burr holes and for a femoral vein), all intravascular catheters, temperature sensors, and a laser Doppler flow probe were removed and surgical sites were closed. Then the animals were returned to their cages for neurologic evaluation and histologic observation 5 days later. As shown in table 1, neurologic deficit score was used for neurologic evaluation (0 = normal, 500 = full deficit).12 

Anesthesia was induced with ketamine (20 mg/kg, intramuscular) injection. Following tracheal intubation, anesthesia was maintained with 2% isoflurane in 100% O². After inserting a cannula into the abdominal aorta, animals were perfused first by heparinized physiologic saline (20 U/ml) for 60 s. The right atrium was cut to drain blood and perfusate. Perfusion was then continued with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 30 min. During perfusion fixation, perfusion pressure was continuously monitored and controlled at 60 mmHg. Following in situ  fixation for 30 min, the pharynx and esophagus were removed. After being paraffin-embedded, the pharynx and esophagus were coronally sectioned (5 μm in thickness). The sections were stained with hematoxylin-eosin. To search for the occurrence of cold injury or mechanical injury during the pharyngeal cooling, direct damage in the epithelium, infiltration of inflammatory cells, and appearance of edema in connective tissue were evaluated.

The clinical study was performed in accordance with the Declaration of Helsinki13and was conducted at Okayama University Hospital and Tsuyama Central Hospital, where the study protocol was approved by the each institutional review board. Three patients (see table, Supplemental Digital Content 3, https://links.lww.com/ALN/A851, which is a table showing clinical characteristics of 3 patients) who were being treated in intensive care units, had been orally intubated for artificial ventilation, and were going to undergo therapeutic hypothermia were enrolled in this study. Pharyngeal cooling was performed for 30 min before the initiation of conventional therapeutic hypothermia that decreases whole body temperature using a water blanket. During the procedures, patients were anesthetized with continuous infusion of propofol (100–200 mg/h), and arterial blood pressure, heart rate, bilateral tympanic temperatures, and bladder temperature were continuously observed. The cuff was inserted along the endotracheal tube without use of a laryngoscope until the feeling of resistance. The maneuver for inserting the cuff was almost identical to that for inserting a supraglottic airway device. The pharyngeal epithelium was macroscopically observed for 3 days using a fiberscope.

Values are expressed as means ± SD. Changes in temperature after the initiation of pharyngeal cooling were analyzed by repeated-measures ANOVA followed by Scheffé F test for multiple comparisons. Otherwise, statistical analyses were performed with the use of the Student t  test (two-tailed) or Mann–Whitney U test, if applicable. Calculations were analyzed using computer software (Stat Flex Ver. 6.0; Artech Co., Osaka, Japan) P < 0.05 was considered to be significant in all statistical tests.

The values of physiologic variables obtained before the initiation of cardiac arrest are shown in table 2. Although the values of all parameters were within normal limits, there was a difference in blood glucose level between the control group (88 ± 8 mg/dl) and pharyngeal cooling group (64 ± 7, P = 0.001) because of the small standard deviations.

Cardiac arrest initiated by electrical stimulation was confirmed by sudden drop of arterial blood pressure and cerebral blood flow. In the control group, arterial blood pressure decreased from 103 ± 8 mmHg to 17 ± 7 mmHg, and cerebral blood flow decreased to 8 ± 4% of the preischemia level during a period of 2 min. In the pharyngeal cooling group, arterial blood pressure decreased from 109 ± 22 mmHg to 17 ± 4 mmHg, and cerebral blood flow decreased to 10 ± 7% of the preischemia level. At the end of cardiac arrest, all animals showed ventricular fibrillation in an electrocardiogram. No animal showed asystole or pulseless electrical activity. With the initiation of resuscitation, spontaneous circulation recovered in all animals, but 1 of the 5 animals in each group died from arrhythmia and hypotension within 2 h. In the pharyngeal cooling group, direct current shock was applied 3.4 ± 1.5 times (P = 0.20, compared with the control group, 2.4 ± 0.5 times), 0.3 ± 0.3 mg of epinephrine was administered (P = 0.87, compared with the control group, 0.3 ± 0.2 mg), and mean arterial blood pressure had recovered to more than 50 mmHg at 3.2 ± 1.3 min (P = 1.0, compared with the control group, 3.2 ± 0.8 min) after the onset of resuscitation.

Figure 2A shows changes in core brain temperature in the parietal cortex. The differences in core brain temperature between the two groups were 0.1°C (SD = 1.0, P = 0.54), 1.9°C (SD = 0.8, P < 0.001), 3.0°C (SD = 0.9, P < 0.001), and 3.1°C (SD = 1.0, P < 0.001) at 0 min, 10 min, 20 min, and 30 min after the onset of pharyngeal cooling, respectively.

Figure 2B shows changes in surface brain temperature in the parietal region. The differences in surface brain temperature between the two groups were 0.1°C (SD = 0.9, P = 0.70), 2.0°C (SD = 1.4, P < 0.001), 2.5°C (SD = 1.4, P < 0.001), and 2.7°C (SD = 1.2, P < 0.001) at 0 min, 10 min, 20 min, and 30 min after the onset of pharyngeal cooling, respectively.

Figure 2C shows changes in rectal temperature. In the control group, rectal temperature was unchanged during the observation period (onset of ischemia: 36.3 ± 1.5°C, end of 12-min ischemia: 36.2 ± 1.0°C, 30 min after onset of resuscitation: 36.3 ± 1.1°C). In the pharyngeal cooling group, rectal temperature was also unchanged during the observation period (onset of ischemia: 36.9 ± 0.4°C, end of 12-min ischemia: 37.0 ± 0.6°C, 30 min after onset of resuscitation: 36.1 ± 0.8°C, P = 0.08 compared with the control group).

An animal in which mean arterial pressure was moderately low (40–50 mmHg) because of instability of the circulatory condition showed maximal decrease in core brain temperature (4.2°C) during 30 min of pharyngeal cooling. In contrast, an animal in which mean arterial pressure was high (100–120 mmHg) because of reactive hypertension showed minimal decrease in core brain temperature (0.3°C) during 30 min of pharyngeal cooling (fig. 3).

In the control group, 2 out of 4 animals showed full neurologic recovery (score = 0) on day 5. One animal did not clean itself and seemed to be delirious (score = 55). The remaining animal did not have light reflex, corneal reflex, or auditory response, and it had no response to toe pinch and needed full respiratory support (score = 489). In the pharyngeal cooling group, 3 out of 4 animals showed full neurologic recovery (score = 0) on day 5. The remaining animal had light reflex, corneal reflex, and sluggish response to auditory stimulation (clapping sound). It showed inappropriate response to toe pinch and survived with spontaneous respiration (score = 290). There was no statistically significant difference between the groups on day 5 (P = 0.65).

Microscopic images of the pharynx on day 5 are shown in figure 4. Histologic findings indicating progress of cold injury, such as appearance of microvacuolization, eosinophilic staining of the cytosol, pyknotic nuclei, and infiltration of inflammatory cells, were not observed in the control group (fig. 4A) or pharyngeal cooling group (fig. 4B). As an internal control, epithelium of the lower esophagus, where the cooling cuff did not make contact, was observed (fig. 4C) in the pharyngeal cooling group. Histologic findings were almost identical to those in the pharyngeal epithelium.

Surface brain temperature, core brain temperature, and tympanic temperature decreased by 0.8°C, 0.9°C and 1.1°C, respectively, during the 30 min of pharyngeal cooling (fig. 5). Tympanic temperature decreased to almost the same extent as those of brain surface temperature and core brain temperature during 30 min of pharyngeal cooling.

The cooling cuff was easily inserted into the pharynx and upper esophagus along the underside of the endotracheal tube. During cuff insertion, heart rate and mean arterial blood pressure were unchanged (fig. 6A).

With the initiation of pharyngeal cooling, tympanic temperature started to decrease without affecting bladder temperature, heart rate, or mean arterial blood pressure. Thirty minutes after the onset of cooling, tympanic temperature was significantly decreased by 0.6 ± 0.1°C (fig. 6B, P < 0.001, compared with the preperfusion level). The difference between decreases in left and right tympanic temperatures was 0.2 ± 0.1°C. Bladder temperature was unchanged (fig. 6C, 0.0 ± 0.0°C). After removal of the cuff, blood residue suggesting the occurrence of mechanical damage of the pharynx or upper esophagus epithelium was not observed on the cuff surface.

Macroscopic observation of the pharynx was performed using a flexible fiber scope. Mechanical damage, edema, and color change (pale and reddish) of the epithelium were not observed immediately after the pharyngeal cooling or on day 1, day 2, and day 3 (fig. 7).

In the present study, pharyngeal cooling was performed before the return of spontaneous circulation in monkeys. Compared with the control group, core brain temperature and surface brain temperature were significantly decreased by 1.9°C and 2.0°C, respectively, at 10 min after the onset of pharyngeal cooling without decrease in systemic temperature. There was no adverse effect of pharyngeal cooling on return of spontaneous circulation. Total dose of epinephrine and number of direct current shocks for resuscitation were not different between the two groups. There were no statistical differences in recovery of blood pressure after resuscitation. In the clinical study, insertion of the pharyngeal cooling cuff and start of perfusion did not affect heart rate or blood pressure. These results provided proof of the principle of pharyngeal cooling that can selectively decrease brain temperature in a short period of time without having adverse effects on systemic circulation.

During chest compression, the brain suffers from the first impact of reperfusion injury, including excess release of glutamate6and exposure to oxygen radicals,14,15leading to intracellular calcium accumulation16and mitochondrial dysfunction17that may initiate necrosis and may trigger the apoptotic process.18Since glutamate release and production of oxygen radicals are attenuated by lowering brain temperature,6,19,–,21hypothermia initiated before the return of spontaneous circulation would be beneficial for brain protection.

In the present study, pharyngeal cooling showed a maximum cooling effect in an animal in which blood pressure was moderately low because of instability of systemic circulation after resuscitation. It was thought that slower blood flow increased residence time at the carotid arteries and enhanced heat exchange leading to the rapid decrease in brain temperature. In clinics, we sometimes experience difficulty in achieving return of spontaneous circulation or sustained low blood pressure after return of spontaneous circulation. Although these patients need brain protection, current hypothermia is usually initiated after confirming stability of the circulatory condition. Pharyngeal cooling would be suitable for these patients by decreasing brain temperature during chest compression and during an unstable circulatory condition.

In the clinical study, tympanic temperature was decreased by 0.6°C during 30 min of pharyngeal cooling. Since pharyngeal cooling was performed in patients whose systemic circulation was stable, the cooling effect on tympanic temperature was attenuated compared with that in the animal study in which pharyngeal cooling was performed during resuscitation. As shown in figure 5, core brain temperature was decreased by only 0.9°C during 30 min of pharyngeal cooling in an anesthetized animal in which systemic circulation was stable. Therefore, it was thought that pharyngeal cooling can decrease tympanic temperature in humans similar to or slightly less than that in the animal study.

The factors that limit the decrease in brain temperature during pharyngeal cooling are thought to be the flow rate of carotid arteries, distance between the carotid arteries and pharynx, heat conductance of the material of the pharyngeal cooling cuff, and flow rate of the perfusate. To improve the efficacy of brain cooling, the last two factors are changeable. Vinyl chloride was used as the material of the pharyngeal cooling cuff to pass the biologic safety test. Since the heat conductance of vinyl chloride is relatively low (0.15–0.2°C/mK), use of other materials may improve heat conductance. The flow rate of perfusate was 500 ml/min in both the monkey and patients regardless of body weight (8.2 ± 2.1 kg vs.  47.3 ± 12.4 kg) and cuff volume (size 2, 40 ml vs.  size 4, 115 ml). Since core brain temperatures in the anesthetized monkey and tympanic temperature in patients were similarly decreased by 0.9°C and 0.6 ± 0.1°C, respectively, during 30 min of pharyngeal cooling, the flow rate of 500 ml/min may exceed the optimum flow rate for a size-2 cuff. The relationship between flow rate and decrease in brain temperature needs to be evaluated.

The pharyngeal epithelium was observed at 5 days after cooling in the animal study and was observed for 3 days in the clinical study. There was no epithelial injury associated with mechanical damage and cold damage of the pharynx. Supraglottic airway devices are similar in shape to the pharyngeal cooling cuff. Serious but reversible complications with the use of supraglottic airway devices have been reported to be recurrent laryngeal nerve palsy22and hypoglossal nerve palsy.23It has been postulated that they are mostly preventable by gentle insertion and maintaining cuff pressure lower than 60 cm H²O.24Since the pharyngeal cooling cuff was perfused at the pressure of 50 cm H²O, serious complications would be preventable as in the case of using supraglottic airway devices.

In general, two types of mechanisms have been postulated as a cause of cold injury: acute damage because of the formation of extracellular ice crystals and progressive damage because of the release of inflammatory mediators.25Since the temperature of perfusate was monitored and controlled at 5°C, it is unlikely that ice crystals were generated in the pharynx. However, the possibility of progressive damage cannot be ruled out. Although no epithelial damage was observed in the present study, pharyngeal cooling should be used for a limited duration until data on safety have accumulated.

An air-circulating blanket, water-circulating blanket, and water-circulating gel pad are commonly used to induce and maintain systemic hypothermia in clinics and have been reported to take 375 min,26210 min,27and 137 min28to reach target temperature of 32–34°C, respectively.

Intravenous rapid infusion of cold saline or Ringer's solution (2,000 ml or 30 ml/kg) is one of the fastest techniques for induction of systemic hypothermia. A randomized clinical trial performed in the field after return of spontaneous circulation showed that esophageal temperature is decreased by 1.2°C on arrival at the hospital without having detrimental effects on systemic circulation.29Ice packs and subzero cooling pads are also commonly used to initiate systemic hypothermia and take 2 h30and 70 min,31respectively, to reach target temperature of 32–34°C.

Nasal cooling is designed to exchange heat with the deep venous sinus of the brain and/or bone at the base of the skull.32In a randomized clinical trial,33nasal cooling was used before return of spontaneous circulation, and tympanic temperature was decreased to the target temperature of 34°C at 102 min after the collapse of circulation. The treated group tended to show better survival rate, suggesting the importance of initiation of hypothermia before return of spontaneous circulation.

In contrast to other devices, pharyngeal cooling is designed to exchange heat with bilateral carotid arteries at the pharynx and upper esophagus. In our previous study using rats, brain surface temperature and hippocampal temperature were equally decreased to 31°C in 6 min and 7 min, respectively,8and glutamate release was significantly reduced by initiating pharyngeal cooling simultaneously with the onset of resuscitation.6During pharyngeal cooling, oxygen extraction from arterial blood in the head region was reduced because of the suppression of metabolic rate.

In the present study, we used monkeys to observe the effects of pharyngeal cooling on brain temperature since the shape of the pharynx in the Japanese monkey is similar to that of the human pharynx and since monkeys do not have a carotid rete (see figure, Supplemental Digital Content 2, https://links.lww.com/ALN/A850, which is a figure showing a cerebral angiogram of a Japanese monkey). Animals such as cats, dogs, pigs, sheep, and goats34,–,36have a carotid rete or a rete-like structure in which the carotid artery branches like capillaries at the skull base to exchange heat with venous blood cooled by the nasal epithelium. These animals can decrease brain temperature by panting without sweating in a hot environment and have adapted to a dry environment during evolution.35The use of these animals for evaluation of pharyngeal cooling would overestimate its cooling effects.

There are several limitations in the present study. First, the use of pharyngeal cooling in the animal study was not blinded. Though resuscitation was performed according to the guidelines of Pediatric Advanced Life Support,11and total amount of epinephrine and number of direct current shocks were not different between the two groups, the nonblinded manner may have biased the survival rate in the treated group. Second, the animal study was not powered to detect differences in survival rate and neurologic outcome. Third, the concentration of blood sugar before initiation of ischemia was higher in the control group than in the pharyngeal cooling group for an unknown reason. Fourth, since brain temperature was measured through a burr hole, brain surface temperature was affected by ambient temperature.

In conclusion, the principle of pharyngeal cooling was proved. Pharyngeal cooling rapidly and selectively decreased brain temperature without having adverse effects on systemic circulation and the pharyngeal epithelium, even when initiated before return of spontaneous circulation. Since the cooling effect was evident in an animal with low blood pressure, it would be appropriate to initiate pharyngeal cooling before return of spontaneous circulation, i.e. , during chest compression.

The authors thank the following for their support with this study: Hisami Aoe, M.D., Ph.D., Postdoctoral; Motomu Kobayashi, M.D., Ph.D., Assistant Professor; Hideki Taninishi, M.D., Ph.D., Assistant Professor; Toshihiro Sasaki, M.D., Ph.D., Postdoctoral; Minako Arai, M.D., Ph.D., Postdoctoral; Kensuke Shiraishi, M.D., Ph.D., Postdoctoral; Nobuyuki Kobayashi, B.E., Technician, Department of Anesthesiology, Okayama University Medical School, Okayama, Japan; Ken Takata, M.D., Ph.D., Assistant Professor, Department of Anesthesiology, Kawasaki Medical School, Kurashiki, Japan; Itaru Yoshii, M.D., Ph.D., Lecturer, Department of Anatomy, Kawasaki Medical School; Noriaki Akagi, B.E., Radiological Technologist, Central Division of Radiology, Okayama University Hospital, Okayama, Japan; Hidekazu Tsuji, B.E., Engineer, and Masatomo Kokubu, B.E., Engineer, Daiken Medical Co., Osaka, Japan.