PEDIATRICS Vol. 106 No. 4 October 2000, pp. 684-694

Pilot Study of Treatment With Whole Body Hypothermia for Neonatal Encephalopathy

Denis Azzopardi, MD, FRCP^*, Nicola J. Robertson, MD, MRCP^*, Frances M. Cowan, MD, PhD^*, Mary A. Rutherford, MD, MRCP^*, Michael Rampling, MD, PhD^*, , and A. David Edwards, FRCP^*

From the ^* Department of Paediatrics and the  Biomedical Sciences Division, Imperial College School of Medicine, London, United Kingdom.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Background.  There is extensive experimental evidence to support the investigation of treatment with mild hypothermia after birth asphyxia. However, clinical studies have been delayed by the difficulty in predicting long-term outcome very soon after birth and by concern about adverse effects of hypothermia.

Objectives.  The objectives of this study were to determine whether it is feasible to select infants with a bad neurological prognosis and to begin hypothermic therapy within 6 hours of birth, and to observe the effect of this therapy on relevant physiologic variables.

Methods.  Sixteen newborn infants with clinical features of birth asphyxia (median cord blood pH: 6.74; range: 6.58-7.08) were assessed by amplitude integrated electroencephalography (aEEG), and mild whole body hypothermia was instituted within 6 hours of birth in the 10 infants with an aEEG prognostic of a bad outcome. Rectal temperature was maintained at 33.2 ± (standard deviation) .6°C for 48 hours. Rectal and tympanic membrane temperature, blood pressure, heart rate, blood gases, blood lactate, full blood count, blood electrolytes, high and low shear rate viscosity, and coagulation studies were monitored during and after cooling. A preliminary assessment of neurological outcome was made by repeated magnetic resonance imaging (MRI) and neurological examination.

Results.  All infants selected to receive hypothermia developed convulsions and a severe encephalopathy. During 48 hours of hypothermia infants had prolonged metabolic acidosis (median pH: 7.30; base excess: 6.3 mmol · L¹, a high blood lactate (median lactate: 5.3 mmol · L¹) and low blood potassium levels (median value: 3.9 mmol · L¹). Hypothermia was associated with lower heart rate and higher mean blood pressure. However, these changes did not seem to be clinically relevant and no significant complication of hypothermia was encountered. Blood viscosity and coagulation studies were similar during and after cooling. Unusual MRI findings were noted in 3 infants: transverse sinus thrombosis with subsequent small cerebellar infarct; probable thrombosis in the straight sinus; and hemorrhagic cerebral infarction. Six of the 10 cooled infants had minor abnormalities only or normal follow-up neurological examination; 3 infants died and 1 had major abnormalities. None of the 6 infants with a normal aEEG developed severe neonatal encephalopathy or neurological sequel.

Conclusions.  After birth asphyxia infants can be objectively selected by aEEG and hypothermia started within 6 hours of birth in infants at high risk of developing severe neonatal encephalopathy. Prolonged mild hypothermia to 33°C to 34°C is associated with minor physiologic abnormalities. Further studies of both the safety and efficacy of mild hypothermia, including further neuroimaging studies, are warranted.  Key words:  hypothermia, neonatal encephalopathy, birth asphyxia, amplitude integrated electroencephalography.

Animal studies have shown that a reduction of body temperature by 3°C to 4°C after hypoxic-ischemic and other neuronal injuries preserves cerebral energy metabolism, reduces cytotoxic edema, and improves histologic and behavioral outcome.^1-4 This evidence suggests that mild hypothermia is a potent and potentially clinically useful neuronal rescue therapy, and preliminary studies have shown promising results in adults after head trauma and stroke.^5,6

However, clinical trials of hypothermia in infants suffering perinatal asphyxia have been delayed by several problems. First, low body temperature has been conventionally associated with adverse outcomes in newborn infants.^7,8 Second, because hypothermic therapy probably has to be started within ~6 hours of birth, and because it has been difficult to predict long-term outcome so soon after birth asphyxia, there has been doubt whether it is possible accurately to select and enroll infants in time.

Recent research suggesting that neonatal encephalopathy may often have causes other than hypoxia-ischemia has underlined the difficulty of adequate trial selection and enrolment.⁹ However, several studies have demonstrated that in infants with evidence of neonatal encephalopathy amplitude integrated electroencephalogram recording soon after birth allows precise and accurate prediction of later neurodevelopmental impairment, and offers a suitable tool for selection of infants for trials of neural rescue therapy.^10-12

We have, therefore, undertaken an observational study to assess the feasibility of attempting trials of hypothermic neural rescue therapy. We studied a series of infants born both within and outside of our referral center with clinical evidence of birth asphyxia. The aims of the study were to determine: whether infants could be selected for treatment using amplitude integrated electroencephalography (aEEG) criteria that predicted a high probability of severe cerebral palsy or death and treatment initiated within 6 hours of delivery; and whether whole body cooling to between 33°C and 34°C for 48 hours after birth asphyxia was associated with physiologic disturbances additional to those characteristic of severe birth asphyxia. We also observed the early neurological and neuroimaging outcome of treated infants. These data are necessary preliminary information for the planning of randomized clinical trials of whole body hypothermic therapy after suspected birth asphyxia.

	METHODS

Top Abstract Methods Results Discussion Conclusion References

The study was carried in the neonatal units of Hammersmith Hospital and Queen Charlotte and Chelsea Hospital London, with the approval of the hospital ethics committee. Infants were enrolled over a period of 18 months. Parents were provided with written information and an explanation of the study and parental consent was obtained before hypothermia was initiated in every infant.

Selection of Infants

Selection was in 2 stages. First, infants suspected of having suffered perinatal hypoxic-ischemic injury were defined using clinical criteria and were entered into the study. These infants then underwent objective assessment of their neurological prognosis using aEEG, and those infants with a poor prognosis were selected for treatment with hypothermic neural rescue therapy.

Clinical Evidence of Birth Asphyxia Infants were suspected of having suffered birth asphyxia if there was evidence of fetal distress from fetal heart rate monitoring and/or the infant had a metabolic acidosis from birth with a blood pH <7.0 and/or a base excess >14 mmol · L¹, and artificial ventilation was needed from birth. Infants who had not apparently recovered after resuscitation and who displayed excessive irritability or neurological depression were then assessed by aEEG.

aEEG The aEEG was recorded with a Lectromed CFM (model 5330, Lectromed, Herts, United Kingdom), as previously described.¹² Briefly, a single channel electroencephalography (EEG) signal is obtained from biparietal electrodes, filtered, and rectified, and the range of amplitude of the signal (in µV) is recorded on an integral printer at 6 cm/hour. The normal record consists of a band of activity with an upper margin >10 µV and a lower margin >5 µV.¹² The aEEG was considered abnormal if the upper and/or lower margins of the dense band of aEEG activity were outside the normal values or if seizures were identified. Seizures were defined as a sudden increase in voltage usually accompanied by a narrowing of the amplitude. The aEEG was performed before sedation was administered and a minimum of 30 minutes of recording was obtained for the initial assessment. The aEEG correlates well with the standard EEG and has been shown to predict neurological outcome accurately 3 to 6 hours after birth asphyxia.^10,11,13 An abnormal aEEG is associated with a positive predictive value (PPV) >70% for severe cerebral palsy or death.^10-12 The aEEG was continuously recorded throughout the cooling period in the treated infants and for at least 24 hours in the infants who did not receive hypothermic treatment.

Hypothermic Treatment

Hypothermia was initiated within 6 hours of birth and maintained using a commercial air cooling system (Polar Air, Augustine Medical, Eden Prairie, MD) that induces hypothermia by blowing cool air through a translucent perforated paper blanket placed over the infant. The system is not servo-controlled but the air temperature can be regulated. Air temperature was adjusted to maintain the rectal temperature between 33°C and 34°C for 48 hours. The infant was then rewarmed at .5°C per hour.

Hypothermic treatment was started in inborn infants as soon as assessment was completed and parental consent to treatment was obtained. Infants born outside of the study centers were assessed, including an aEEG examination, by retrieval teams that were familiar with assessment of the aEEG. Parental consent was obtained, and hypothermic treatment was initiated at the referring hospital if admission to the study center was likely to be delayed beyond 6 hours after birth. These infants were nursed in standard open or closed incubators, and hypothermia was achieved by not actively warming the infant. The rectal and abdominal skin temperatures, oxygen saturation, heart rate, and blood pressure were monitored during transfer and excessive hypothermia prevented.

Clinical Care

Intensive care was conducted according to our usual practice. After resuscitation, ventilatory requirement was judged by assessing the infant's spontaneous breathing efforts and blood gas analysis. Sedation was with morphine infusion (10-40 µg · kg¹/hour¹) for ventilated infants and chloral hydrate (25-50 mg · kg¹/dose¹) when mechanical ventilation was discontinued. Seizures (whether noted on aEEG or clinically) were treated initially with phenobarbitone (20 mg/kg¹ followed by 5-10 mg · kg¹ · 24¹ hours) and with lidocaine infusion (2-4 mg · kg¹/hour¹) and/or midazolam infusion (30-60 µg/kg¹ · hour¹) if persistent. Blood electrolyte analysis, urine volume and analysis, and infant weight guided fluid management. Oliguria was defined as a urinary flow rate <1 mL · kg¹ · hour¹ averaged over 6 hours.

Monitoring

Routine investigations on admission to the treatment centers included: arterial blood gas (measured at 37°C); coagulation studies; full blood count; and blood electrolytes, urea, and creatinine. Rectal and abdominal skin temperatures were monitored continuously, and the tympanic membrane temperature was recorded by infrared thermography (Thermoscan, San Diego, CA) every 30 minutes during hypothermic treatment until the infant was rewarmed. Coagulation studies, blood platelet count, and blood electrolytes were measured every 12 hours or more frequently during treatment. High and low shear rate blood viscosity was measured repeatedly (Contraves LS30 Viscometer, Contraves AG, Zurich, Switzerland) with samples collected during cooling measured at 34°C ± .5°C. Intraarterial blood pressure and heart rate were monitored continuously according to our routine practice. Investigations for sepsis were conducted and repeated if clinically indicated.

Neuroimaging and Neurological Assessment

The aEEG was monitored continuously for at least 72 hours and a standard EEG was performed within 3 weeks of birth. Serial magnetic resonance imaging (MRI) was performed starting at 3 to 4 days of age and survivors were assessed regularly by a neurologist using a standard neurological assessment.^14,15

Data Analysis

A series of statistical analyses were performed. Data distributions were inspected and log-transformed to normality where appropriate.

Multiple linear regression with dummy variables to account for intersubject variation was used to determine the relation between: 1) heart rate, mean arterial blood pressure, and rectal temperature; and 2) blood high and low shear viscosity, packed cell volume, and rectal temperature.

Analysis of variance for repeated-measures was used to compare group means for coagulation studies and platelet counts during and after hypothermic treatment.

The relationship between rectal and tympanic membrane temperature was defined using the Bland Altman method to calculate mean differences and limits of agreement. To define the trend of agreement with temperature, linear regression with dummy variables was used.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Sixteen infants suspected of having suffered birth asphyxia (median cord blood pH: 6.75; range: 6.58-7.08; median base excess: 25.5; range: 11 to 31.8 mmol · L¹) were studied. Ten infants had an abnormal aEEG, and after parental consent was obtained, hypothermic treatment was started within 6 hours of birth.

The 10 infants, including 6 born outside of the study centers, who received hypothermic treatment were born at, median (range), 38.5 (34-40) weeks' gestation and weighed 3.41 (1.86-3.8) kg at birth. The Apgar scores were 1.5 (0-4) at 5 minutes and 3.5 (0-5) at 10 minutes. The initial blood pH and base excess were 6.73 (6.58-7.08) mmol · L¹ and 25.5 (11 to 31.8) mmol · L¹, respectively. In 3 infants no cause for the asphyxia could be determined. The obstetric and clinical details of the infants who were cooled are given in Table 1.

All infants predicted to have a poor neurological prognosis and treated with hypothermia developed the multisystem disorder characteristic of severe birth asphyxia. All required mechanical ventilation at birth, although 3 infants were extubated during cooling. All infants developed oliguric renal failure but none required dialysis and in 5 a diuresis began before the end of the cooling period. Seizures were noted in all of the cooled infants and were treated with phenobarbitone only in 7 infants and also with lidocaine and/or midazolam in the other 3 infants.

The interval between birth and the initiation of hypothermia was 4 (1-6) hours. The admission rectal temperatures were 36.3°C to 36.5°C for the 4 inborn infants; 33.0°C to 34.5°C for the 5 outborn infants who were cooled during transfer; and 36.3°C in the 1 outborn infant who was kept warm during transfer. The mean (± standard deviation) rectal temperature during cooling was 33.2°C ± .6°C, and the rectal temperatures of the 10 infants are shown in Fig 1. The mean rectal temperature of the first 3 cooled infants was slightly lower than that of the last 7 cooled infants (32.8°C ± .6°C and 33.5°C ± .58°C, respectively). The rectal temperature was consistently higher than the tympanic membrane temperature particularly at the lower temperature range: the mean difference was .49°C and the trend of agreement was:

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Rectal temperatures in the 10 infants treated with hypothermia. Different symbols represent data from individual infants.

Clinical and Laboratory Findings in Infants Treated With Hypothermia

The blood pH, base excess, and partial pressure of carbon dioxide in arterial gas (PaCO₂) values are shown in Fig 2. The values were similar in the cooled and noncooled mildly asphyxiated infants during the first hours after birth. However, they seemed lower in the cooled infants between 4 and 60 hours of life.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Blood pH, PaCO2 and base excess measurements in infants treated () and in infants not treated () with hypothermia.

Blood pressure was negatively related (mean blood pressure = .72 [rectal temperature] + 73; P < .00001) and heart rate positively related (heart rate = 6.5 [rectal temperature]  105.5; P < .00001) but this was of little clinical significance, and no arrhythmia or significant hypotension was seen. The relationship between heart rate, mean blood pressure, and rectal temperature is shown in Fig 3. A slightly low blood potassium level (to below 3.8 mmol · L¹) was noted in 8 infants during hypothermia, and the median and lowest potassium levels were 3.9 and 2.8 mmol · L¹, respectively.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Relationships among mean blood pressure, heart rate, and rectal temperature in cooled infants. Different symbols represent data from individual infants.

Four infants had abnormal coagulation on admission that improved during the cooling period after treatment with fresh frozen plasma. Clotting studies became abnormal again in 1 infant (infant 3 in Table 1) after rewarming, in association with large hemorrhagic cerebral infarction. The median and maximum values of the prothrombin time, thrombin time, and partial thromboplastin time are shown in Table 2 and analysis of variance showed that the group means during and after cooling were not statistically different.

The blood viscosity was measured repeatedly in 5 infants. Both high and low shear viscosity were linearly related with packed cell volume (P < .001), as is shown in Fig 4, but not to body temperature. The platelet count was low on admission in 3 infants and fell further during cooling necessitating platelet transfusions in 2 of the infants. Thrombocytopenia was not observed in the other infants. However, mean platelet count was not statistically different during and after hypothermic treatment.

View larger version (9K): [in this window] [in a new window]   Fig. 4.   Relationship between high and low shear rate viscosity and packed cell volume in infants during () and after () treatment with hypothermia.

aEEG

A burst suppression pattern or suppressed pattern¹¹ was noted on the initial aEEG examination in all 10 infants. In 5 infants the aEEG remained abnormal throughout the cooling period, but in the other 5 infants aEEG activity returned to a more continuous pattern by 13 hours, although brief periods of abnormal activity continued to be observed. Seizures occurred even when the background pattern had recovered. The aEEG activity was more suppressed in the 3 of the 4 infants with a poor outcome, compared with the 6 infants who recovered. The standard EEG was first performed within 1 week of birth in 7 infants: EEG background activity was similar to that observed on the aEEG (Table 1).

Neurological and Neuroimaging Assessment in Infants Selected for Hypothermia

Infants With Normal Neurological Outcome Six infants had a normal neurological outcome on follow-up examination. Five of these 6 infants had minor abnormalities on MRI commonly seen in mild hypoxic-ischemic injury consistent with a good neurological prognosis. In the other infant (infant 2), MRI appearances suggestive of a transverse sinus thrombosis and abnormal low signal intensity consistent with a small right inferior cerebellar infarct were noted. However, neurological examination at 12 months of age was normal.

Infants With an Abnormal Outcome Three of the 10 infants selected for cooling died. One infant developed extensive bilateral cerebral hemorrhagic infarction at the end of the cooling period, although the exact time of onset of this complication could not be determined. The infant developed disseminated intravascular coagulation and worsening neurological state and died after intensive care was withdrawn according to parental wishes. Intensive care was also withdrawn in the 2 other infants when the clinical, EEG, and MRI findings confirmed extensive cerebral injury. In one of these infants, intensive care was withdrawn after 24 hours of cooling because the infant remained unresponsive from birth, the EEG showed little activity and abnormal signal intensity in basal ganglia and thalami and probable thrombosis in the straight sinus were observed on MRI.

Infants Not Selected for Treatment With Hypothermia

The 6 infants with similar clinical features of birth asphyxia but normal aEEG who were not treated with hypothermia were born at 40.5 (38-41) weeks' gestation, weighing 3.35 (2.9-4.02) kg at birth. Their 5- and 10-minute Apgar scores were 5 (3-7) and 6.5 (3-8), blood pH 6.74 (6.7-6.89), and base excess 23.5 (14.5 to 28.4) mmol · L¹. Four of the 6 infants were born by emergency cesarean section because of abnormal fetal heart rate; 1 infant was born at home, was severely acidotic on arrival to hospital, and suffered from meconium aspiration syndrome. One infant suffered severe meconium aspiration syndrome and polycthemia that necessitated exchange transfusion. None of these 6 infants developed a severe encephalopathy. Minor abnormalities only were seen on MRI and the neurological examination at 12 months of age was normal.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

These results suggest that after suspected birth asphyxia, infants with a high probability of severe cerebral palsy or death can be selected and hypothermic therapy instituted within 6 hours of birth even if the infant is born outside of the primary treatment centers. In this small preliminary study, 48 hours of whole body cooling to a rectal temperature of between 33°C and 34°C was undertaken with apparent safety, although minor physiologic abnormalities were noted during hypothermic treatment. MRI abnormalities were observed in both treated and untreated infants, but major abnormalities were seen only in infants predicted to have a poor prognosis from the aEEG. No problems related to hypothermia during transport were noted.

Study Design

Our study was planned to investigate a protocol that might be used to enroll infants in future randomized trials of hypothermic therapy. We, therefore, applied objective selection criteria and observed the clinical course of infants treated with hypothermia. Our pilot study was not randomized because the group size would have been too small to be meaningful. Randomized clinical trials will be required to assess the efficacy and safety of prolonged systemic hypothermia after perinatal asphyxia. Because clinical criteria alone are poorly predictive of subsequent outcome, randomized, controlled studies of neuroprotective therapies that rely solely on clinical entry criteria will need to enroll a large number of infants. Our results indicate that by using staged entry criteria infants with a high risk of developing severe encephalopathy can be selected, and the study size may be reduced; ~125 infants would need to be enrolled to observe a 30% reduction in a poor outcome (: .5; power: .9).

Because severe asphyxia is rare, future trials will need to enroll infants from several participating centers. We, therefore, investigated the practicality of selecting and treating infants before admission to our hospital. The infants who were not selected for cooling because aEEG was normal had clinical criteria that would have qualified them for selection into previous studies of neuroprotective therapies, including hypothermia. Therefore, we described the clinical course of these infants to confirm that neuroprotective therapy was correctly withheld. However, because these infants must have suffered a less severe insult than did the infants treated with hypothermia, only limited comparisons can be made between the 2 groups and we avoided statistical analysis of this comparative data.

Rationale of Hypothermic Treatment

Hypothermia seems to have profound effects on the cerebral response to hypoxic-ischemic and other neuronal injuries: the increases in extracellular glutamate and free radical and nitric oxide synthesis are suppressed,¹⁶ cerebral energy phosphates are preserved, and cerebral alkalosis and lactate reduced^1,17; and the activation of transcription factors, heat shock protein production, and microglial response are altered¹⁸ and the number of apoptotic cells reduced.¹⁹ Studies in adult and newborn animals have shown that a reduction of body temperature of 3°C to 4°C after cerebral insults is associated with improved histologic and behavioral outcome.⁴ Although there may be differences between the experimental models and neuronal injuries in infants,^20,21 the promising results reported in clinical studies of hypothermia after cranial trauma and stroke^5,6 suggest that hypothermia may be a clinically useful neuronal rescue therapy also in asphyxiated infants.

Selection of Infants for Treatment

Clinical trials of neuronal rescue therapies will need to select those infants who are most likely to benefit from treatment and to avoid exposing infants who have a good prognosis to potentially toxic therapies. Unfortunately clinical assessment is not sufficiently accurate.^22-24 Electrophysiological tests seem currently to be the best early predictors of subsequent neurological outcome after suspected asphyxia,^10,11 and the aEEG is currently the most precisely defined and clinically useful of these tests because it can be applied quickly by neonatal unit staff after little training.¹² Previous studies have shown that if aEEG demonstrates burst suppression, low voltage, and/or seizure patterns at 3 to 6 hours after birth, the infants have >80% PPV for death or severe cerebral palsy.^10,13 We have previously found that if the maximum aEEG activity is <10 µV and/or the minimum activity is <5 µV, then the PPV for adverse outcome is >70%.¹²

In the present study we were able to use aEEG to make decisions about hypothermic treatment within 6 hours of birth. Consistent with previous data, aEEG correctly predicted and selected for treatment those infants who developed a severe encephalopathy and excluded infants who had similar clinical criteria at birth but did not develop severe encephalopathy over the days following or evidence of neurodevelopmental impairment in the ensuing months. These results suggest that the aEEG is a practical clinical tool for selecting infants into trials of treatment with hypothermia.

Initiation of Hypothermia

Animal studies indicate that greater neuroprotection is obtained if hypothermia is started soon after the insult.³ We were able to enroll infants and initiate hypothermia within 2 hours of birth in infants born at the treatment centers, but treatment was started later in the outborn infants. To minimize delay we enrolled these infants and started hypothermia at the hospital of birth, transferring them in that state without complication. Because we experienced delays in arranging transfer, some infants would not have been treated within 6 hours of birth if treatment had been delayed until admission to the treatment centers. Although ~7000 infants were born at the treatment centers during the study, only 4 infants achieved the selection criteria for hypothermic treatment. Thus, it is likely that in future clinical trials most candidates for hypothermia treatment will be born outside the treatment centers and arrangements to transfer infants promptly are required. Alternatively, hypothermia will need to be started before admission to the treatment centers.

Monitoring Brain Temperature

A major difficulty facing clinical trials of hypothermic treatment in infants is the assessment of brain temperature during treatment. Direct brain temperature measurement is too invasive for routine clinical use and noninvasive methods have not yet been developed.²⁵ The rectal temperature may not reflect brain temperature reliably,²⁶ but studies in adults found close correlation between core temperature and direct measurements of brain temperature during mild hypothermia.^6,27 We used rectal temperature as a surrogate for brain temperature in preference to the nasopharyngeal temperature because this may be inaccurate during treatment with continuous positive airway pressure or oxygen administered by nasal prongs (G. Simbrunner, personal communication, 1999). We also measured the tympanic membrane temperature but found that the rectal temperature was consistently higher than the tympanic membrane temperature particularly at the lower range of temperature. This could have been attributable to poor measurement technique, or it is possible that the tympanic membrane temperature was affected by cool air blown over the infant. Our observations call into question the validity of the tympanic membrane temperature during hypothermia and emphasize the need for a noninvasive technique to measure brain temperature directly.

Previous Studies of Hypothermia in Newborns

Several studies have reported brief hypothermic therapy for resuscitation during asphyxia.^28-31 These early studies seemed to demonstrate improved outcomes without side effects but few controls were included and the relevance of their findings to prolonged postasphyxia hypothermic treatment is doubtful. One controlled pilot study of prolonged hypothermia after birth asphyxia has recently been reported.³² In this study mild hypothermia was achieved over 72 hours by focal head cooling. The study was too small to assess benefit and because the minimum temperature was 35.5°C, there are no data on the effects of cooling to 33°C to 34°C. Although some experimental evidence suggests that selective head cooling may be effective,^32,33 we used whole body cooling because commercial systems were available and because we were uncertain whether head cooling alone effectively lowers deep brain temperature. A steep temperature gradient has been observed between the surface of the head and deep brain structures, and the deep brain temperature remains very close to core temperature even when extreme cold is applied to the surface of the head.²⁶ The possibility of temperature gradients within the brain and the importance of damage to deep structures in causing severe neurodevelopmental impairment³⁴ mean that future studies of hypothermic therapy will benefit from precise, noninvasive methods for measuring regional brain temperature.

Toxicity

Clinicians will be concerned about the risk of side effects of hypothermia given the poor outcome conventionally associated with hypothermia.^7,8 Impaired cardiac function, disordered coagulation, and increased sepsis have been reported with profound hypothermia. Thrombocytopenia, hypokalemia, and increased sepsis have been noted in studies of hypothermia (to a rectal temperature of ~33°C) in adults, but these were not associated with an adverse outcome.⁶ There are no data on the safety of whole body cooling to 33°C to 34°C in newborn infants.

The risk of adverse effects may be increased with prolonged hypothermia. Some experimental studies have demonstrated benefit with brief periods of hypothermia,¹ while in others, benefit was only observed when hypothermia was prolonged for 72 hours.² We chose to cool infants for 48 hours followed by a gradual rewarming to limit the risk of toxicity and because most experimental and clinical studies have used hypothermia for a similar duration.

Our results suggest that minor physiologic effects occur during treatment with mild hypothermia for up to 48 hours. Cooled infants had more prolonged acidosis than did noncooled infants. However, it is not possible to determine whether this was attributable to hypothermia or whether it was a consequence of the increased severity of the hypoxic/ischemic insult. Blood gases were not temperature corrected because the changes caused by temperature are very small.

Mild hypokalemia and a sinus bradycardia occurred during hypothermia. Hypothermia-induced hypokalemia has been commonly observed in experimental and clinical studies^35,36 and is thought to be caused by intracellular shift of potassium. Excessive potassium supplementation must be avoided because of the risk of hyperkalemia during rewarming.³⁷

We aimed to keep the rectal temperature above 33°C. However, because of initial unfamiliarity with the equipment, the rectal temperature during cooling in the first 3 cooled infants was slightly lower than in the other infants. Because the Polar air cooling equipment is not servo-controlled and does not have a continuously variable temperature control, the rectal temperature sometimes fell briefly below the target temperature, but without apparent consequences other than bradycardia. The PR and QT intervals may be prolonged during hypothermia but significant arrhythmia has only been reported when the rectal temperature is <32°C.³⁵ We have no data on QT intervals in our patients, but no arrhythmia occurred during cooling. Although the mean blood pressure was statistically higher during cooling, there was no clinically relevant effect of hypothermia on blood pressure during cooling or on rewarming. We rewarmed infants gradually, in part, to avoid sudden vasodilatation and the risk of hypotension. Inotropes were given primarily because of the bradycardia.

Viscosity and Coagulation

Hypothermia to <30°C in adults causes a significant increase in blood viscosity and packed cell volume that is reversible with normothermia.^38,39 These changes may cause reduced cardiac output, increased systemic vascular resistance, and reduced cerebral blood flow.^40,41 Because of technical difficulties, we obtained data on the blood viscosity in 5 of the study infants. As expected, we found a linear relationship between blood viscosity and packed cell volume, but treatment with cooling did not significantly alter the relation. The recent observations that coagulation factors are often abnormal in infants who develop cerebral palsy emphasize the need for careful investigation of the clotting and thrombophilic profile in hypothermic studies.⁴² We did not observe a change in coagulation with hypothermia but cannot exclude this possibility because laboratory measurements were performed at 37°C. There was no clinical evidence of prolonged bleeding time, although 1 infant did suffer hemorrhagic cerebral infarction, which was detected after cooling had been completed.

Neuroimaging

We performed repeated MRI to identify unusual patterns of cerebral injury. Unexpected findings were observed in 3 infants: 1 infant with extensive hemorrhagic infarction at ~48 hours of age that was not noted in previous cranial ultrasound examinations; another infant with MRI appearance suggestive of transverse sinus thrombosis and who subsequently had a small right cerebellar infarct; and a third infant with evidence of probable thrombosis in the straight sinus. We do not know whether hypothermia contributed to these lesions, but the cases highlight the importance of performing detailed neuroimaging in studies of neuronal rescue therapies.

	CONCLUSION

Top Abstract Methods Results Discussion Conclusion References

After birth asphyxia infants can be objectively selected by aEEG and hypothermia can be started within 6 hours of birth in infants at high risk of developing a severe neonatal encephalopathy. In our patients, physiologic changes attributable to hypothermia were generally mild and intensive care was uninterrupted. Neuroimaging with MRI should be performed in future studies of hypothermia to identify unusual patterns of cerebral injury that may be associated with hypothermia. Randomized, controlled studies of treatment with hypothermia after perinatal asphyxia are warranted.

	FOOTNOTES

Received for publication May 24, 1994; accepted Jan 10, 2000.

Reprint requests to (D.A.) Department of Paediatrics, Imperial College School of Medicine, London, United Kingdom. E-mail: d.azzopardi{at}ic.ac.uk

	ABBREVIATIONS

aEEG, amplitude integrated electroencephalography; EEG, electroencephalography; PPV, positive predictive value; MRI, magnetic resonance imaging; PaCO₂, partial pressure of carbon dioxide in arterial gas.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. Thoresen M, Penrice J, Lorek A, Mild Hypothermia following severe transient hypoxia-ischaemia ameliorates delayed cerebral energy failure in the newborn piglet. Pediatr Res 1995; 37:667-670 [Medline]
2. Gunn AJ, Gunn TR, De Haan 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]
3. Sirimanne E, Blumberg RM, Bossano D, The effect of prolonged modification of cerebral temperature on outcome following hypoxic-ischaemic brain injury in the infant rat. Pediatr Res 1996; 39:591-597 [Medline]
4. Colbourne F, Corbett D Delayed postischemic hypothermia: a six month survival study using behavioral and histological assessments of neuroprotection. J Neurosci 1995; 15:7250-7260 [Abstract]
5. Marion DW, Penrod LE, Kelsey SF, Treatment of traumatic brain injury with moderate hypothermia. N Engl J Med 1997; 336:540-546 [Abstract/Free Full Text]
6. Schwab S, Schwarz S, Spranger M, Keller E, Bertram M, Hacke W Moderate hypothermia in the treatment of patients with severe middle cerebral artery infarction. Stroke 1998; 29:2461-2466 [Abstract/Free Full Text]
7. Silverman WS, Fertig JW, Berger AP. The influence of the thermal environment on the survival of newly born permature infants. Pediatrics. 1958;876-885
8. Mann TP, Elliott RIK. Neonatal cold injury due to accidental exposure to cold. Lancet. 1957;229-234
9. Badawi N, Kurinczuk JJ, Keogh JM, Intrapartum risk factors for newborn encephalopathy: the Western Australian case-control study. Br Med J 1998; 317:1554-1558 [Abstract/Free Full Text]
10. Eken P, Toet MC, Groenendaal F, de Vries LS Predictive value of early neuroimaging, pulsed Doppler and neurophysiology in full term infants with hypoxic-ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed 1995; 73:F75-F80 [Abstract]
11. Hellstrom Westas L, Rosen I, Svenningsen NW Predictive value of early continuous amplitude integrated EEG recordings on outcome after severe birth asphyxia in full term infants. Arch Dis Child Fetal Neonatal Ed 1995; 72:F34-F38 [Abstract]
12. Al Naqueeb N, Edwards AD, Cowan F, Azzopardi D Assessment of neonatal encepalopathy by amplitude integrated electroencephalography. Pediatrics 1999; 103:1263-1271 [Abstract/Free Full Text]
13. Toet MC, Hellstrom Westas L, Groenendaal F, Eken P, de Vries LS Amplitude integrated EEG 3 and 6 hours after birth in full term neonates with hypoxic-ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed 1999; 81:F19-F23 [Abstract/Free Full Text]
14. Dubowitz LM, Dubowitz V, Palmer P, Verghote M A new approach to the neurological assessment of the preterm and full-term newborn infant. Brain Dev 1980; 2:3-14 [Medline]
15. Griffiths R. The Abilities of BabiesA Study in Mental Measurement. London, United Kingdom: University of London Press; 1954
16. Thoresen M, Satas S, Puka-Sundvall M, Post-hypoxic hypothermia reduces cerebrocortical release of NO and excitotoxins. Neuroreport 1997; 8:3359-3362 [Medline]
17. Amess PN, Penrice J, Lorek A, Mild hypothermia after transient hypoxia-ischemia reduces the delayed rise in cerebral lactate in the newborn piglet. Pediatr Res 1997; 41:803-808 [Medline]
18. Globus MY, Busto R, Lin B, Schnippering H, Ginsberg MD Detection of free radical activity during transient global ischemia and recirculation: effects of intraischemic brain temperature modulation. J Neurochem 1995; 65:1250-1256 [Medline]
19. Edwards AD, Yue X, Squier MV, Specific inhibition of apoptosis after cerebral hypoxia-ischaemia by moderate post-insult hypothermia. Biochem Biophys Res Commun 1995; 217:1193-1199 [CrossRef][Medline]
20. Gluckman PD, Williams CE Is the cure worse than the disease? Caveats in the move from laboratory to clinic. Dev Med Child Neurol 1992; 34:1015-1018 [Medline]
21. Gluckman PD, Williams CE When and why do brain cells die? Dev Med Child Neurol 1992; 34:1010-1014 [Medline]
22. Ruth VJ, Raivio KO Perinatal brain damage: predictive value of metabolic acidosis and the Apgar score. Br Med J 1988; 297:24-27
23. Perlman JM, Risser R Can asphyxiated infants at risk for neonatal seizures be rapidly identified by current high-risk markers? Pediatrics 1996; 97:456-462 [Abstract/Free Full Text]
24. Ekert P, Perlman M, Steinlin M, Hao Y Predicting the outcome of postasphyxial hypoxic-ischemic encephalopathy within 4 hours of birth. J Pediatr 1997; 131:613-617 [CrossRef][Medline]
25. Roberts I, Azzopardi D, Hand J, Edwards AD Discrimination of deep from surface temperature with emitted microwave radiation in a brain-specific phantom. Early Hum Dev 1997; 47:210 [CrossRef]
26. Mellergard P Monitoring of rectal, epidural, and intraventricular temperature in neurosurgical patients. Acta Neurochir Suppl Wien 1994; 60:485-487 [Medline]
27. Mellergard P Changes in human intracerebral temperature in response to different methods of brain cooling. Neurosurgery 1992; 31:671-677 [Medline]
28. Cordey R, Chiolero R, Miller JA Jr Resuscitation of neonates by hypothermia: report on 20 cases with acid-base determination on 10 cases and the long-term development of 33 cases. Resuscitation 1973; 2:169-181 [CrossRef][Medline]
29. Westin B Infant resuscitation and prevention of mental retardation. Am J Obstet Gynecol 1971; 110:1134-1138 [Medline]
30. Dunn JM, Miller JA Jr Hypothermia combined with positive pressure ventilation in resuscitation of the asphyxiated neonate: clinical observations in 28 infants. Am J Obstet Gynecol 1969; 104:58-67 [Medline]
31. Ehrstrom J, Hirvensalo M, Donner M, Hietalahti J Hypothermia in the resuscitation of severely asphyctic newborn infants: a follow-up study. Ann Clin Res 1969; 1:40-49 [Medline]
32. Gunn AJ, Gluckman PD, Gunn TR Selective head cooling in newborn infants following perinatal asphyxia; a safety study. Pediatrics 1998; 102:885-992 [Abstract/Free Full Text]
33. Natale JA, D'Alecy LG Protection from cerebral ischemia by brain cooling without reduced lactate accumulation in dogs. Stroke 1989; 20:770-777 [Abstract/Free Full Text]
34. Rutherford MA, Pennock JM, Schweiso J, Cowan F, Dubowitz L Hypoxic-ischaemic encephalopathy: early and late magnetic resonance imaging findings in relation to outcome. Arch Dis Child 1996; 75:F145-F151
35. Clifton GL, Allen S, Berry J, Koch SM Systemic hypothermia in treatment of brain injury. J Neurotrauma 1992; 9:S487-S495
36. Sprung J, Cheng EY, Gamulin S, Kampine JP, Bosnjak ZJ Effects of acute hypothermia and beta-adrenergic receptor blockade on serum potassium concentration in rats. Crit Care Med 1991; 19:1545-1551 [Medline]
37. Sprung J, Gamulin S, Bosnjak ZJ, Kampine JP Potassium correction of hypothermic hypokalemia induces hyperkalemia after rewarming. Can J Anaesth 1990; 37:S69
38. Poulos ND, Mollitt DL The nature and reversibility of hypothermia-induced alterations of blood viscosity. J Trauma 1991; 31:996-998 [Medline]
39. Chen RY, Chien S Hemodynamic functions and blood viscosity in surface hypothermia. Am J Physiol 1978; 235:136-143
40. Lehot JJ, Villard J, Piriz H, Hemodynamic and hormonal response to hypothermic and normothermic cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1992; 6:132-139 [CrossRef][Medline]
41. Su JY, Amory DW, Sands MP, Mohri H Effects of ether anesthesia and surface induced hypothermia on regional blood flow. Am Heart J 1979; 97:53-60 [CrossRef][Medline]
42. Nelson KB, Dambrosia JM, Grether JK, Phillips TM Neonatal cytokines and coagulation factors in children with cerebral palsy. Ann Neurol 1998; 44:665-675 [CrossRef][Medline]