Cerebral Hyperthermia in Children after Cardiopulmonary Bypass

After obtaining scientific and ethical approval of the Human Subjects Review Committee of the Hospital for Sick Children and written informed consent, 15 infants and children, American Society of Anesthesiologists physical status II or III, between 3 months and 6 yr of age, undergoing elective repair of CHD necessitating mild to moderate hypothermic CPB, were studied. All patients included in this observational, prospective study had been full-term. Preoperative exclusion criteria included children undergoing palliative procedures, patients with known neurologic disorders, or family history of malignant hyperthermia. Children requiring significant inotropic support for low cardiac output, defined by the administration of dopamine (more than 15 μg · kg⁻¹· h⁻¹) or the presence of poor myocardial contractility seen during transesophageal echocardiography (TEE), and those with major postoperative hemorrhage (> 50% of the circulating volume) would have been excluded.

Premedication and induction of anesthesia was left to the discretion of the anesthetist, with the understanding that the child’s trachea would remain intubated for 6 h postoperatively. At arrival in the operating room, routine monitoring devices were applied to the patient. Anesthesia was induced with use of either an inhalational agent or fentanyl (10–30 μg/kg) and thiopental (up to 5 mg/kg) and was maintained with either a volatile agent or midazolam and a supplemental narcotic as necessary. Muscular relaxation was achieved using pancuronium (0.2 mg/kg). A maximum of 1 minimum alveolar concentration (MAC) of volatile anesthetic (end-tidal) was administered. When necessary, patients received sodium nitroprusside (2–5 μg/kg) for peripheral vasodilation and 1% isoflurane during CPB according to level of mean arterial pressure. CPB was conducted in the routine fashion used at The Hospital for Sick Children and the University of Toronto, with flow rates of 100–150 ml/kg and alpha-stat p  H management with a arterial partial pressure of carbon dioxide (Paco²) at 40–45 for moderate hypothermic CPB. At weaning from CPB, inotropic agents were administered at the discretion of the anesthetist or the surgeon. Propofol was used postoperatively to provide sedation in those children suitable for early tracheal extubation. Invasive monitoring consisted of an indwelling arterial catheter for continuous blood pressure and arterial blood gas analysis, a central venous catheter, and a right retrograde jugular bulb catheter. The catheter positioned within the jugular bulb consisted of a 22-gauge (0.48 mm ID) catheter (Arrow, Reading, PA) and was inserted using a technique previously described. 14The use of a right jugular bulb catheter allowed continuous measurement of cerebral venous blood temperature during the study period.

The room temperature was set at 17°C, and the heating blanket was not used before the institution of CPB. The lowest temperature of the child during CPB was chosen according to the preference of the surgeon and monitored from the rectal temperature probe. The child was rewarmed using CPB until the rectal temperature was 35°C and the repair of the congenital heart lesion was completed. The temperature of the perfusate was not greater than 37°C at any time. The heating blanket was set to 42°C, and a convective warming system (Bair Hugger, Eden Prairie, MN) turned to medium temperature. The convective heat surface warmer was turned off when the rectal temperature reached 35°C. The room temperature was increased to 22°C. Modified ultrafiltration was performed on all children for a maximum of 20 min or until the circuit was empty. All patients were transferred to the critical care unit. Intermittent positive-pressure ventilation and muscle paralysis were maintained for a minimum period of 6 h. No external warming sources were used during the first 6 h postoperatively. All patients were monitored according to the critical care unit protocol with use of a rectal temperature probe and received paracetamol or the application of ice on their heads if the temperature reached 38.5°C. The nurse caring for the patient was not aware of the JVBT during the study period.

Temperature probes (Mallinckrodt, St. Louis, MO) were placed in the following sites: in the rectum, in the esophagus, on skin surfaces, in the tympanic membrane, and in the axilla for the duration of the procedure and for the first 6 h after admission to the critical care unit. Skin surface temperature sites included the toe, thigh, calf, finger, forearm, and forehead. All temperature probes were fast-response thermocouples. The temperature recording of the jugular venous bulb was obtained using a type T, copper-constantan thermocouple microprobe (Columbus Instruments, Columbus, OH) with a diameter of 0.41 mm, a time constant of 0.08 s, and a precision of 0.1°C (range, 0°C–50°C). The thermocouple was advanced within the jugular bulb catheter to avoid coiling of the tip during insertion or downward displacement by the venous flow within the jugular vein. The final position of the thermocouple was confirmed radiologically in all patients. The temperature of the blood returning to the patient was measured using a thermocouple inserted into the aortic cannula (Mallinckrodt). The time from induction of anesthesia to surgical preparation was noted. The minimum temperature achieved during CPB and the duration of time at that temperature was recorded. The duration of active rewarming, defined as the beginning of rewarming until the end of the CPB, was also monitored and recorded electronically. This period of time does not represent the time taken to rewarm to 35°C, but rather to complete the repair. However, it represents the total duration of heat energy transfer to the child. All temperature data were monitored continuously and recorded on a computer database for further analysis. The temperature data acquisition interface (Columbus Instruments) was calibrated every week. All data were acquired every minute for the duration of the study period, which was discontinued after 6 h in the critical care unit.

All data with parametric values are expressed as the mean ± SD. The number of patients needed to identify a clinically relevant increase in brain temperature postoperatively was based on the expectation to observe a difference between JVBT and tympanic temperature greater than 1.0°C. This determination was calculated based on a statistical power of 0.8, an α of 0.05, and a β of 0.2. A total of 12 patients was suggested. Fifteen patients were studied to account for methodologic difficulties that could have led to exclusion from the study. Mean jugular bulb and tympanic and esophageal temperatures were analyzed using analysis of variance and the Dunnett test for multiple comparisons, with the temperature recorded at the end of the rewarming period in the operating room and subsequently every 100 min until completion of the study. Data analysis between different temperature sites at similar times was analyzed using analysis of variance and the Student-Newman-Keuls test for multiple comparisons. The JVBT rewarming kinetics for all patients were analyzed using linear regression analysis, and the coefficient of determination (r²) was determined. A P < 0.05 was used to express statistical significance.

Fifteen patients were enrolled and studied. Demographic data for all patients are summarized in table 1. The median age was 15 months (range, 3 to 71 months) and the mean weight was 11.5 ± 8.4 kg. The duration of rewarming and the minimum temperature during CPB are also reported in table 1. No complications associated with the study were noted.

Figure 1shows the temperature profile recorded from patient 8 (see table 1). The JVBT increased continuously after rewarming and for the next 6 h to reach 39.2°C. The coefficient of correlation between the JVBT and the lower esophageal, rectal, and tympanic temperatures were 0.97, 0.85, and 0.98, respectively.

Figure 2shows the temperature:time relation obtained for each individual patient and the mean JVBT. The mean temperature recorded at the end of CPB was 36.9 ± 1.3°C but reached 39.6 ± 0.8°C at the end of the study period (P < 0.01). The JVBTs recorded after 200 and 300 min were also statistically different from the temperature obtained at the end of the rewarming period (38.6 ± 0.9 and 39.3 ± 0.8°C, respectively;P < 0.01) The JVBT variability decreased progressively as the brain temperature became hyperthermic from 1.4°C at the end of the rewarming period to 0.8°C at the end of the study period. The kinetics of rewarming of the brain in the postoperative period was determined by the slope of the mean JVBT and corresponded to y = 0.006×+ 37.21. The coefficient of determination of the temperature:time relation was highly significant (r²= 0.97). In one patient, JVBT was as high as 41.4°C at the end of the sixth hour, whereas the tympanic and rectal temperatures were 39.9 and 36.9°C, respectively.

Figure 3shows a comparison between the JVBT and the lower esophageal, rectal, and tympanic temperatures. Although there were no statistical differences between each site at the end of CPB and rewarming, JVBT continued to increase progressively to become statistically different from tympanic temperature after 200 min (P < 0.05) and from lower esophageal (P < 0.05) and rectal (P < 0.001) temperatures after 300 min of the recovery period. The temperature of the lower esophageal probe was 38.4 ± 0.7°C at the end of the study period, which was significantly different from the JVBT at 39.6 ± 0.8°C (P < 0.05). After 6 h of recording, the tympanic and rectal temperatures reached a maximum of 37.8 ± 0.9 and 37.7 ± 0.6°C, respectively, which was also different from the JVBT (P < 0.001). There were no differences among the esophageal, rectal, and tympanic temperatures during the 6 h of recovery in the intensive care unit.

In this study, significant jugular venous bulb hyperthermia occurred in all infants and children after hypothermic cardiopulmonary bypass. Mean JVBT was significantly higher than mean core temperature between 2 and 6 h. Although the correlation between the lower esophageal, rectal, and tympanic temperatures showed a close approximation with JVBT during the recovery period, the lower esophageal and rectal temperature sites did not accurately reflect the change in brain temperature after 2 or 3 h, whereas the tympanic temperature did not reflect the changes in JVBT.

The cerebral metabolic rate of the brain is one of the highest in the body and, as a result, heat is produced. 15,16The normal brain temperature is an average of 0.5°C warmer than the core temperature. 17–19However, the increase in JVBT observed in this study cannot be explained simply on this basis. It is possible that this increase in temperature also may be associated with surgery or the consequences of hypothermia alone. This observational study lacks the presence of a non-CPB control group. Alternatively, there is evidence that CPB could be responsible for the recorded increase in JVBT recorded. Cardiac surgery and CPB activate a systemic inflammatory response characterized clinically by systemic vasodilation, hyperthermia, postoperative bleeding, and reperfusion injury. 16,17It has been shown in a study of 20 children undergoing CPB that removing inflammatory mediators using zero-balanced hemofiltration resulted in lower maximum body temperature in the first 24 h postoperatively. 20Although they did not measure brain or jugular bulb temperature, these observations correlate with the current study. It may be that cerebral hyperthermia occurs as a result of increased endogenous pyrogens, such as cytokines.

The first hours after CPB in infants and children are critical. Previous investigators studied pediatric patients undergoing CPB and speculated that some patients could be at risk of cerebral ischemia postoperatively because of the presence of uncoupling of cerebral blood flow and cerebral metabolic rate for oxygen (CMr^o2). 22Burrows et al.  23studied 21 neonates and infants undergoing CPB and deep hypothermia circulatory arrest and reported significant increases in anterior fontanel pressure at the end of rewarming. These authors correlated this increase in intracranial pressure with the delayed return of cerebral electrical activity and showed that the duration of the increase correlated with the time taken for the visual evoked potentials to reappear. O’Hare et al.  24studied 10 infants undergoing deep hypothermia circulatory arrest and reported a significant decrease in cerebral blood flow velocity until 6 h postoperatively, despite optimal cerebral perfusion pressure. Although the current study did not investigate patients undergoing deep hypothermia circulatory arrest, neurologic injury is known to occur during CPB. 25Cerebral hyperthermia might important in the presence of an existing brain injury. It has been shown that brain temperature increases after cerebral ischemia, despite normal ambient and core body temperature. 26 

Experiments have shown that the central nervous system is sensitive to heat. 27Shum-Tim et al. , 28using a piglet model, reported that cerebral hyperthermia (40°C) was associated with persistent deterioration of neurobehavioral outcome after ischemia, and Kuroiwa et al.  29confirmed that such hyperthermic episodes are implicated in the exacerbation of neurologic injury. Studies of the pathologic changes in the human brain have shown that, within 6 h of whole-body hyperthermia (40°C), the cerebral cortex becomes congested and degenerative changes in the neurons are prominent. 27The effects of cerebral hyperthermia on the cerebellum are more serious and independent of the duration of the heat exposure. 27Previous studies of focal ischemia in animals showed a direct correlation between elevated intracerebral temperature and worsened outcome. 11,30Temperature increases of 1 or 2°C resulted in a significant deterioration of neurologic function, which correlated with histopathologic disease lesions. 11 

Cerebral hyperthermia at 40°C has been shown to increase extracellular glutamate concentration in cats after ischemia. 31Excitatory amino acid receptors are present at birth and their concentrations increase drastically during the first few weeks of life. 32–34The level of these excitatory amino acids in a 3-month-old infant approximates the adult concentration. The occurrence of seizures in infants after cardiac surgery has been suggested to caused by excitotoxic mechanisms via  neuronal injury and the release of excitatory amino acids. 3In a recent article, which reported use of astrocyte cell cultures during and after rewarming with sustained moderate hyperthermia (38.5°C), a continuous and significant increase in extracellular glutamate was seen. 35This astrocytic dysfunction and the resultant accumulation of extracellular glutamate may account for the observed neurologic dysfunction seen in infants after repair of CHD. Marked cerebral swelling, mainly cortical in origin, has been documented by use of magnetic resonance imaging immediately after hypothermic (28°C) and normothermic CPB. 36,37In an editorial about brain protection during anesthesia, Drummond 38raised concerns about the occasional use of hyperthermic perfusate during rewarming and suggested that iatrogenic hyperthermia contributes to the substantial incidence of neuropsychiatric dysfunction observed after CPB in adults. In the current study, cerebral hyperthermia occurred independently of the perfusate temperature.

Potential methodologic flaws in the current study include accuracy in positioning of the jugular bulb catheter and the reliability of this site to monitor global brain temperature. In the supine position, the right jugular bulb and vein usually are larger than the left bulb and vein and are known to drain the majority of the venous blood from both cerebral hemispheres, whereas the flow through the paravertebral plexus is minimal. 39,40Although it may be assumed that JVBT does not represent the temperature of the brain, the value of its monitoring in the current study is based on the second law of thermodynamics. The temperature recorded in the venous blood cannot be greater than the brain temperature itself. This was confirmed by a previous study that used direct measurement of cerebral temperature in head-injured patients. 41It is also known that JVBT does not reflect brain temperature because it measures core temperature as a result of the admixture of extracranial venous blood from the facial and mastoid veins. 41Finally, jugular venous bulb oxygen saturation during CPB has been reported to be of clinical value because it is a measure of cerebral blood flow. 42 

It has been suggested that “deep in the esophagus” could correlate with brain temperature. 15It has also been suggested that lower esophageal temperature readings provide an approximate indication of the cerebral temperature in the absence of an open thorax or a rapid transfusion of cold blood. 18In the current study, the lower esophageal temperature probe followed closely the temperature of the jugular bulb after rewarming but failed after 2 h to reflect it accurately at higher temperatures. Mechanical ventilation can significantly affect the temperature of the esophagus. 43 

The tympanic membrane temperature did not correlate with JVBT. Temperature probes used in this context do not always touch the tympanum because of the presence of cerumen or because of the small ear canal of children. This results in delayed response time and inaccuracy. Furthermore, it has been shown that convective air flow applied to the face of the patient decreases, 15or radiant heating of the head increases, 44tympanic temperature without affecting brain temperature. None of the patients in the current study were exposed to an external source of heating, either convective or radiating, during the initial 6 h in the critical care unit. They were treated for systemic hyperthermia when the rectal temperature probe recorded a temperature greater than 38.5°C, at which time the patient either received paracetamol or the application of ice packs on the head.

The importance of maintaining normothermia has been a fundamental premise in the neonatal and cardiac intensive care unit for many years. 28This has led to the application of aggressive rewarming techniques for even modest degrees of hypothermia, such as occur frequently after CPB. 28To our knowledge, this is the first study of infants and children that shows a significant increase in JVBT during the first 6 h of recovery after CHD repairs during CPB. Whether CPB, hypothermia, or surgery alone is responsible for this increase in temperature seen in this pediatric population is unclear. No matter what the cause of neurologic injury after repair of CHD, ample evidence in animal studies suggests that cerebral hyperthermia worsens neurologic outcome.

The authors thank the staff of the Cardiac Anaesthesia Division; the staff of the Cardiovascular Surgical Division; and Peter Cox, M.D., and Desmond Bohn, M.D., Critical Care Unit, for their collaboration with this study, and the Cardiovascular Operating Room nurses; the staff of the Perfusion Services; and the nurses of the Critical Care Unit for their patience and cooperation with this study.