Abstract
Objective. Modest reduction in brain temperature is a promising therapy to reduce brain damage after neonatal encephalopathy as a result of acute perinatal asphyxia. The efficacy of modest hypothermia may in part be dependent on the stability of the desired brain temperature. The objective of this study was 1) to evaluate in newborn animals a commercially available cooling system (Blanketrol II Hyperthermia-Hypothermia system) to control brain temperature during whole-body hypothermia and 2) to use the results of the animal experiments to perform a pilot study evaluating the feasibility of whole-body hypothermia as a neuroprotective therapy for newborns with encephalopathy at birth.
Methods. In the animal investigation, 3 miniature swine were instrumented and ventilated, and temperature probes were placed in the esophagus and the brain (1 cm and 2 cm beneath the parietal cortical surface and the dura). Body cooling was achieved using the automatic control mode (servo) of the cooling system. In the human investigation, 19 term infants with moderate or severe encephalopathy were randomized to either normothermia (n = 10) or hypothermia (n = 9) within 6 hours of birth. Whole-body hypothermia was achieved using the hyperthermia-hypothermia cooling system with servo control of esophageal temperature to 34.5°C for 72 hours followed by slow rewarming.
Results. In the animal investigation, body cooling with the animal lying on a single blanket resulted in rapid cooling of the body within 90 minutes. Repetitive cyclical swings in esophageal temperature of 1.7 ± 0.2°C (mean ± standard deviation) around the set point of 33.5°C were reduced to 0.7 ± 0.2°C when a second, larger blanket was attached and suspended. Esophageal temperature was a good marker of deep brain temperature (esophageal to 2-cm brain difference: 0.1 ± 0.3°C). In the human investigation, the infants were randomized at 4.1 ± 1.3 hours (mean ± standard deviation) after birth. Age at randomization was similar in the 2 groups. Cooling was initiated at an average age of 5.3 hours. Target temperature of 34.5°C was achieved within 30 minutes and remained constant throughout the intervention period. Heart rate decreased to 108 ± 14 beats per minute (bpm) at 60 minutes and remained between 115 and 130 bpm for the duration of cooling compared with 130 to 145 bpm in the normothermia group. Blood pressure was similar in the 2 groups. No adverse events occurred during 72 hours of cooling. The mortality rate and frequency of persistent pulmonary hypertension, renal failure, hepatic dysfunction, and need for pressor support were similar in both groups.
Conclusions. Animal studies showed that a simple modification of a commercially available cooling system (2 blankets attached, subject lying on 1 and the second hanging freely) results in stable core body and brain temperature when used in the automatic control mode. The pilot study in term infants with encephalopathy using this cooling system demonstrates feasibility of initiating whole-body hypothermia at <6 hours of age to a constant esophageal temperature using servo control and provides no evidence that hypothermia involved greater hazard than benefit.
Hypoxic-ischemic encephalopathy (HIE) associated with acute perinatal asphyxia in the term or near-term newborn remains an important problem because of the potential for permanent neurodevelopmental deficits identified later in childhood.1 A number of specific neuroprotective therapies have emerged during the past 10 to 15 years. Modest reduction of brain temperature, in the order of 2°C to 4°C, seems to be the most promising intervention.2–7 In animal models, the extent of neuroprotection associated with brain cooling is related to the duration and the depth of cooling.3,8,9 Gunn et al10 demonstrated the presence of a therapeutic window in near-term fetal sheep, allowing modest hypothermia to be initiated at an interval of up to 5.5 hours after ischemia.
Three small pilot studies have examined modest hypothermia for newborns with encephalopathy after acute perinatal asphyxia.11–13 Methods of inducing modest hypothermia for the brain have included head cooling, head cooling combined with body cooling, and body cooling alone.11–14 It remains unclear whether the neuroprotective effects of modest hypothermia are appreciably different as a function of the method used to cool the brain. Irrespective of the method used, it would be desirable to be able to control brain temperature with as simple a method as possible.
None of the studies used an automatic control mechanism to regulate the temperature at a desired value but depended on investigators’ making frequent changes at the bedside. The efficacy of modest hypothermia may depend in part on the stability of the lowered brain temperature, and an automatic control mechanism may avoid large fluctuations in brain temperature. Furthermore, automatic temperature adjustments would reduce effort of the bedside staff caring for critically ill infants. On the basis of these considerations, the present report had 2 objectives: 1) to evaluate in newborn animals a commercially available cooling system (Blanketrol II Hyper-Hypothermia system, Model 222R, Cincinnati Sub-Zero Products, Inc, Cincinnati, OH) to control brain temperature during whole-body hypothermia and 2) to use the results of the animal experiments to design and perform a pilot study evaluating the feasibility and safety of whole-body hypothermia as a neuroprotective therapy for newborns with encephalopathy presumably from hypoxic-ischemic brain injury after acute perinatal asphyxia.
METHODS
Animal Investigation
Animal studies were performed at the University of Texas Southwestern Medical Center at Dallas after approval by the Institutional Animal Care and Research Advisory Committee. The studies were conducted between December 1998 and April 1999. Three miniature swine (age 3, 10, and 13 days; weight 1.3, 1.6, and 2.4 kg) were tracheotomized and ventilated with N2O and O2, followed by insertion of intravascular catheters, and placement of temperature probes into the brain tissue via a burr hole over the parietal cortex at a depth of 1 cm and 2 cm beneath the cortical surface corresponding to the cortical gray matter and basal ganglia, respectively, and the dura. The burr holes were filled with bone wax and the skin was sutured closed.15 Anesthesia was achieved with intravenous propofol, which was continued for the duration of the study. Pentothal was administered for creation of the burr holes. For monitoring core body temperature, a temperature probe was placed in the lower third of the esophagus. Body temperature was maintained at 38.5°C with a heating pad (physiologic temperature of miniature swine).
After 90 minutes of stabilization, animals were transferred to a 25“ × 33” (infant size) Maxi-Therm Lite Blanket and positioned on their side. The blanket was attached to a 9-ft hose from a Blanketrol II Hyper-Hypothermia system. The automatic control mode (servo) was used in all animals, and the controlling temperature site was the esophagus. Animals were stabilized at an esophageal temperature of 38°C to 39°C and underwent body cooling with a set point (target body temperature) of 33.5°C. Each animal was used to assess the automatic control mode under 1 of the following conditions: 1) with and without a second 25“ × 64” (adult size) blanket attached to the Blanketrol, 2) with a second 25“ × 64” blanket either suspended or folded and attached to the Blanketrol, or 3) with a second 25“ × 64” blanket attached to the Blanketrol with or without warming and humidification of inhaled gases. The second blanket was suspended vertically and positioned alongside the Blanketrol cooling unit to dampen the rate of change of the water temperature in the infant blanket. Temperatures were measured before and during cooling from 4 different sites: 1 cm and 2 cm beneath the parietal cortical surface, on the overlying dura, and in the esophagus. The temperature probes, manufactured by Cincinnati Sub-Zero Products, Inc, are accurate to within ± 0.2°C.
Human Investigation
The purpose of this pilot study was to determine the feasibility, in terms of enrollment, achievement of target body temperatures, and lack of serious adverse events, of conducting a whole-body cooling study with this equipment in neonates with HIE secondary to acute perinatal asphyxia. The study was performed in the participating centers in the National Institute of Child Health and Human Development Neonatal Research Network between November 1999 and June 2000. The institutional review board of each center approved the protocol. A subcommittee of the site principal investigators developed a manual of operations. The manual included the definitions of the neurologic examination for eligibility and details of the cooling and rewarming procedure.
Before initiation of the study, each site principal investigator reviewed the definitions of the categories of the neurologic examination. Certification of additional neonatologists to perform examinations entailed agreement between the center principal investigator (designated as the gold standard) and other designated faculty members on specific items of the neurologic examinations of 2 infants with abnormal examinations. A training session was held for research nurses from all of the clinical sites to review the screening of infants for eligibility, randomization process, and cooling/rewarming procedure.
Screening
All term infants who were ≥36 weeks’ gestation and admitted to the neonatal intensive care unit at ≤6 hours of age with a diagnosis of neonatal depression, acute perinatal asphyxia, or encephalopathy were screened for eligibility.
Inclusion Criteria
Infants were evaluated in 2 steps: evaluation by clinical and biochemical criteria (step A) followed by a neurologic examination (step B) (Fig 1).
Inclusion criteria for study entry.
Step A
When blood gas results were available, eligibility criteria included an umbilical cord pH or any postnatal pH within 1 hour of age ≤7.0 or a base deficit on the cord or postnatal blood gas within 1 hour of ≥16 mEq/L. When a blood gas was not available or pH ≤1 hour was between 7.01 and 7.15 or base deficit was 10 to 15.9 mEq/L, an infant was eligible when there was a history of an acute perinatal event and either an Apgar score ≤5 at 10 minutes or continued need for ventilation initiated at birth and continued for at least 10 minutes. An acute perinatal event was defined as late or variable decelerations resulting in an emergent cesarean section delivery, cord prolapse, placental abruption, uterine rupture, maternal trauma, or maternal respiratory arrest.
Step B
Once criteria in step A were met, all infants had a standardized neurologic examination performed by a certified examiner. Infants were candidates for the study when there was evidence of moderate or severe encephalopathy or the presence of seizures at the time of or before this neurologic examination. An infant was designated as having moderate or severe encephalopathy when 1 or more signs were present in 3 of the following 6 categories: 1) level of consciousness: lethargy (moderate), stupor, or coma (severe); 2) spontaneous activity: decreased (moderate), absent (severe); 3) posture: distal flexion (moderate), decerebrate (severe); 4) tone: hypotonia (moderate), flaccid (severe); 5) primitive reflexes: suck, weak (moderate), absent (severe) or Moro, incomplete (moderate), absent (severe); and 6) autonomic nervous system: pupils, constricted (moderate), skew deviation or nonreactive to light (severe), heart rate, bradycardia (moderate), variable heart rate (severe) or respiration, periodic breathing (moderate), apnea (severe). Variable heart rate was defined as rate fluctuating from ≤80 to ≥100 beats per minute (bpm), and ventilatory support was categorized as apnea.
Exclusion criteria were 1) inability to perform random assignment by 6 hours of age, 2) chromosomal abnormality, 3) major congenital anomaly, 4) severe growth restriction (≤1800 g birth weight), 5) infant unlikely to survive, and 6) parent or attending neonatologist refuses consent for study participation. The inclusion criteria were designed to include infants with HIE secondary to acute perinatal asphyxia, eg, acute perinatal events and fetal acidemia. Other causes of encephalopathy, including infection and metabolic disease, cannot be excluded within 6 hours of birth.
Study Intervention
After informed consent was obtained, the random assignment was performed by telephone by the Data Coordinating Center at the Research Triangle Institute (Research Triangle Park, NC). The randomization was stratified by center, and the assignments were generated by a random, permuted block algorithm with block sizes of 2 and 4. Infants were randomly assigned to either the hypothermia or the normothermia group. Assignments were not masked.
Hypothermia Group
Infants in the hypothermia group were placed supine on an infant-size (25“ × 33”) blanket that was precooled to 5°C. The rationale for precooling was to minimize the time to reach the desired temperature. The occiput of the head rested on the blanket. An esophageal probe (Cincinnati Sub-Zero Products) was inserted and the Blanketrol system was used in the automatic (servo) control mode with a target or set point temperature of the esophagus set at 34.5°C. The esophageal site was chosen as a representative site of core body temperature that would allow easy stabilization of a temperature probe for several days. A second blanket (25“ × 64”) was attached to the Blanketrol cooling system as described in our mini-swine experiments (Fig 2). Neither overhead radiant warmers nor other heat sources were used during the study intervention period. A probe to monitor skin temperature was placed on the lower abdominal wall. Esophageal and skin temperatures were monitored continuously and recorded every 15 minutes for the first 4 hours, every hour for the next 8 hours, and every 4 hours during the remaining period of cooling (total 72 hours). Infants received each center’s routine clinical care, including monitoring of vital signs and surveillance for organ dysfunction. Blood gases in the hypothermia group were corrected for body temperature. At the end of the 72 hours of induced hypothermia, the automatic control set point was increased 0.5°C per hour on the Blanketrol system until a set point of 36.5°C was reached. The esophageal probe was then removed, and servo control of the abdominal skin temperature from a radiant warmer was resumed with monitoring of temperature until normothermia was achieved.
The infant lies supine on the infant-size blanket. The adult-size blanket is suspended vertically alongside the cooling unit. Both blankets are attached to the cooling unit with water circulating through them simultaneously.
Normothermia Group
Infants in the normothermia group were cared for under an overhead radiant warmer. Abdominal skin and esophageal temperatures were monitored for 72 hours as in the hypothermia group. Temperature control was achieved with servo control of the abdominal skin to a radiant warmer using a control temperature of 36.5°C. Infants received routine clinical care with the same monitoring of vital signs and surveillance for organ dysfunction as described for the hypothermic group. In both groups, anticonvulsants, sedatives, and paralytic agents were administered according to each clinical center’s practice.
Adverse Events and Monitoring
During 72 hours of study intervention, the presence of the following predefined potential serious adverse events related to cooling were monitored in both groups: 1) cardiac arrhythmia, 2) persistent acidosis despite volume infusion and pressor support for >3 hours after initiation of study, 3) presence of major vessel thrombosis or bleeding, 4) evidence of loss of skin integrity, 5) equipment malfunction, and 6) death. The incidence of a composite adverse event based on the sum of 1), 2), and 3) was compared between the 2 groups using sequential analysis methods. All adverse events were reported within 72 hours of occurrence on MedWatch forms (FDA medical products reporting program) and on specially developed adverse event reporting forms.
All infants had serial assessments of metabolic status (serum electrolytes, blood urea nitrogen, and creatinine) measured at baseline and at 24, 48, and 72 hours of study intervention. Blood gas measurements were monitored at baseline and at 4, 8, 12, 24, 48, and 72 hours. Heart rate and blood pressure were recorded at baseline, hourly for 12 hours, and every 4 hours for 72 hours. Evaluation of metabolic status, blood gas measurements, and vital signs beyond 72 hours of study intervention was based on clinical indications. Urine output was recorded and monitored daily. A neurologic examination was performed daily by the certified examiner until 72 hours and at discharge. The coagulation status was evaluated by prothrombin time and partial thromboplastin time at baseline, at 24 hours, and as clinically indicated. A cranial sonogram was performed within 24 hours of study initiation, and magnetic resonance imaging (MRI) of the head was obtained at 44 weeks’ postmenstrual age or at discharge. MRI at all sites included multiplanar, multisequence T1 and T2 imaging of the brain. Additional sequences, based on the site’s capabilities, included fast-spin echo T2, fluid-attenuated inversion recovery, and diffusion-weighted imaging. The sonogram and MRI were read by radiologists at each center. All protocol deviations were monitored and reported to the data center.
Data Analysis
Data coordination and management was performed by the Data Center at Research Triangle Institute. Data were transmitted electronically from each participating center on a weekly basis. Because this was a feasibility study, it was not powered to detect differences in outcomes; hence, statistical comparisons between groups were not performed. Instead, comparisons were made on the basis of clinical relevance of the treatment differences within the constraints of the limited sample size. Data are presented as the mean and standard deviation (SD). An external Data Safety and Monitoring Board periodically monitored progress of the study, including review of all adverse events.
RESULTS
Animal Studies
Body cooling with a single blanket attached to the Blanketrol system resulted in rapid cooling of the body to the set point temperature within 90 minutes (Fig 3). Once the body was cooled to the set point, there were repetitive cyclic swings in esophageal temperature of 1.7 ± 0.2°C around the set point of 33.5°C. Addition of a second blanket (animal lying on a 25-in × 33-in blanket, a second 25-in × 64-in blanket suspended) reduced the magnitude of the esophageal temperature swings to 0.7 ± 0.2°C. Removal of the second blanket resulted in 1.6 ± 0.1°C temperature swings. The temperature of the blanket varied from 5°C to 42°C with 1 blanket in place. Addition of the second blanket led to smaller fluctuations in blanket temperature from 15°C to 35°C (data not shown).
Temperatures of the esophagus (top) and multiple brain sites (bottom) are plotted for 1 piglet during whole-body hypothermia. The esophageal temperature was servo controlled as indicated by the dashed line in both panels. Animals were initially cooled while lying on a 25“ × 33” blanket. Addition of a second blanket is indicated by the solid bar from 300 to 500 minutes (animal lying on a 25“ × 33” blanket, a second 25“ × 64” blanket suspended) and reduced the magnitude of the temperature swings.
Esophageal temperature seems to be a good marker of deep brain temperature because the esophageal-brain temperature difference was 0.1 ± 0.3, 0.4 ± 0.5, and 1.1 ± 0.9°C for the 2-cm and 1-cm depths and the dural surface, respectively. Brain temperature mirrored the pattern and magnitude of esophageal temperature fluctuation around the set point (Fig 3, bottom). The addition of the second blanket reduced the magnitude of the temperature swing of brain temperature at a 2-cm depth from 2.0 ± 0.1°C to 0.9 ± 0.3°C. Removal of the second blanket resulted in resumption of 2.0 ± 0.2°C temperature fluctuations. Similar results were found for other brain sites (1-cm depth and the dural brain surface). Two other animals were used to demonstrate that folding the second blanket on itself (reducing the length to approximately 12“) resulted in an increase in the magnitude of the esophageal temperature swings (1.0 ± 0.1°C to 1.5 ± 0.1°C) and that the magnitude of the esophageal temperature swings was not affected by the use of warmed and humidified inhaled gases (data not shown).
Human Studies
Seventy-eight infants were screened. Twenty infants were eligible in 10 of the 14 clinical sites. Informed consent was obtained for 19 infants; 9 were randomly assigned to the hypothermia group, and 10 were assigned to the normothermia group. The perinatal characteristics are shown in Table 1. Complications during the intrapartum period were similar between both groups. The location of birth, mode of delivery, and need for resuscitation at birth were also similar.
Perinatal Characteristics
The neonatal characteristics are shown in Table 2. The age at randomization was similar (range: 2.3–6.0 hours in the hypothermia and 2.2–5.9 hours in the normothermia group). Cooling was initiated in the hypothermia group within 55.5 minutes of randomization at an average age of 5.3 hours. The birth weight, proportion of infants with 10-minute Apgar score <5, and the presence of acidosis in the cord or postnatal blood gas were similar between the infants in the hypothermia and those in the normothermia group. The degree of encephalopathy at randomization (Table 3) was similar between the 2 groups.
Neonatal Characteristics at Randomization
Neurologic Examination of Study Infants at Entry
Esophageal and skin temperatures of infants in both groups are plotted in Fig 4. At initiation of cooling, esophageal temperature was the same in both groups. Body cooling resulted in an overshoot in esophageal temperature to 32.9 ± 0.7°C, 1.6°C lower than the control set point. The target temperature of 34.5°C was achieved within 30 minutes, stabilized after overshoot within 90 minutes, and remained constant throughout the intervention period. The temperature of the blanket (data not shown) initially increased from the precooled temperature of 5°C to 42°C in response to the rapid fall in body temperature and overshoot in esophageal temperature (Fig 4). Once the esophageal temperature stabilized, the blanket temperature ranged between 30°C and 40°C during the intervention period. The administration of anticonvulsants or sedatives/hypnotic agents was not accompanied by a decrease in esophageal temperature. In the normothermia group, esophageal temperature remained constant between 37.0°C and 37.5°C. The skin temperature was approximately 32.5°C in the hypothermia group and 36.5°C in the normothermia group during the intervention period. This resulted in a temperature gradient between the core and the skin of 1.7 ± 0.7°C for the hypothermia group and 0.4 ± 0.1°C for the normothermia group.
Temperatures of the esophagus (top) and abdominal skin (lower) are plotted for the normothermic group (▵) and the hypothermic group (•). Values are mean ± SD. Normothermic infants were servo controlled at 36.5°C using the abdominal skin as the control site, and hypothermic infants were servo controlled at 34.5°C using the esophageal temperature as the control site. The hypothermic group was rewarmed after 72 hours of cooling.
Heart rate (Fig 5) in the hypothermia group was 160 ± 14 bpm at the start of cooling and decreased to a minimum average value of 108 ± 14 bpm (range: 90–142 bpm) simultaneous with an esophageal temperature of 32.9 ± 0.7°C at 60 minutes after the start of cooling. Once the esophageal temperature stabilized, heart rate remained at approximately 120 bpm for the duration of cooling. For the normothermia group, heart rate was 148 ± 15 bpm at the start of the study interval and average values varied between 130 and 145 bpm during the next 72 hours. Despite the differences in esophageal temperature, blood pressure patterns (Fig 6) seemed similar between the 2 groups. Pressor support was used in 5 infants in the hypothermic group and in 3 infants in the normothermic group during the study intervention.
Heart rate is plotted for the normothermic (▵) and hypothermic (•) groups during the 72-hour study intervention. Values are mean ± SD.
Systolic blood pressure (top) and diastolic blood pressure (bottom) are plotted for the normothermic (▵) and hypothermic (•) groups during the 72-hour study intervention. Values are mean ± SD.
There were no adverse events during the 72-hour study interval in the hypothermic group. There were 2 adverse events in the normothermic group: 1 infant had persistent acidosis, and 1 infant died after life support was withdrawn. The number of infants who were receiving anticonvulsant and/or sedation in the hypothermia group at 48 (n = 5) and 72 hours (n = 3) was similar to the normothermia group (3 at 48 hours and 4 at 72 hours of the intervention). None of the infants in the hypothermia group was noted to be shivering.
Characteristics of the hospital course and status at discharge are shown in Table 4. The number of infants with pulmonary hypertension, oliguria, hepatic dysfunction, and hypotension requiring support was similar in the 2 groups. Cranial imaging with MRI was performed on all survivors. Abnormal MRI findings in 3 of 7 infants in the hypothermia group included diffuse cystic encephalomalacia (1 infant), cystic changes in the thalami (1 infant), and decrease in periventricular white matter (1 infant). Abnormal MRI findings in 3 of 7 survivors in the normothermia group included high-intensity signal in the anterior limb of the internal capsule (1 infant); small punctate hemorrhage in the posterior periventricular white matter (1 infant); and increased signal in the posterior limb of the internal capsule, posterior lentiform nucleus, and thalamus (1 infant).
Hospital Course and Status at Discharge
DISCUSSION
These studies demonstrate the feasibility of performing a clinical trial using whole-body hypothermia as an experimental therapy for term infants with moderate or severe encephalopathy after acute perinatal asphyxia. Within a 6-hour interval after birth, it was feasible to screen patients for eligibility criteria, perform a standardized neurologic assessment by a certified examiner, obtain informed consent, randomly assign patients to the appropriate study arm, and initiate treatment. The Blanketrol II Hyper-Hypothermia system provided rapid body cooling for the hypothermia group, and use in the automatic control mode provided excellent control of core temperature during the 72-hour cooling period. We identified no evidence in this pilot study that whole-body hypothermia has hazards that would preclude additional study. This study did not compare the Blanketrol system with any other mode of cooling; therefore, no conclusions can be drawn about 1 method of cooling versus another.
For experimental interventions that involve lowering body temperature, precise control of temperature is highly desirable given the widely recognized hazards of cold stress in newborns.16–18 The results of the animal experiments demonstrated that temperature regulation in the automatic control mode using a single blanket attached to the Blanketrol system readily achieved reductions in body temperature, but 2°C fluctuations in temperature around the set point occurred (Fig 3). An undulating pattern of temperature may not only increase the adverse effects of cold exposure when at the lowest point of the temperature fluctuation but also compromise the degree of neuroprotection at the peak temperature because even small changes (as little as 1–2°C) in brain temperature may modulate the extent of hypoxic-ischemic damage.19,20 The servo mechanism triggered cooling or heating as the temperature rose above or fell below the set point; a more sensitive servo mechanism that would adjust the blanket water temperature incrementally as the core temperature approached the control set point may decrease temperature fluctuations. The addition of a second blanket attached to the Blanketrol system reduced the magnitude of the swings in temperature. The magnitude of temperature fluctuations could be influenced by variables such as whether the second blanket was folded. These observations help characterize the impact of systemic thermal control on brain temperature in clinical trials that will use body cooling.
The addition of the second blanket did not prevent the overshoot of esophageal temperature relative to the control set point on initial cooling of human newborns (Fig 4). Although means to minimize or prevent the overshoot were considered, none was implemented because the overshoot was transient and seemed to be well tolerated and cooling the brain as rapidly as possible was desired. The Blanketrol II Hyper-Hypothermia system required no manual adjustments to the blanket temperature, unlike the adjustments needed to the water temperature of the cooling cap and output of the overhead heater in the head/body cooling strategy, or adjustments to air temperature in the body cooling strategy used in studies reported to date.11–14
The results of this feasibility trial agree with other pilot studies of brain cooling, suggesting that modest hypothermia compared with normothermia is not associated with adverse events such as arrhythmia, acidosis, bleeding, or thrombosis. Neuroimaging was performed on all 14 infants who survived, and the frequency of abnormal MRI findings was similar in the hypothermia and normothermia groups. The abnormalities noted on MRI are similar to those described by Azzopardi et al,13 who also found no preponderance of thrombosis or infarction in cooled infants. Until larger groups of infants are studied, these observations should be considered preliminary. Mean arterial blood pressure has been reported to decrease when head cooling is combined with body cooling and there is a large increase in output of the overhead heater.12 In our study, the hypothermic infants who were cooled with the Blanketrol system (without an overhead heater) had blood pressure measurements that were not different from those of the normothermic infants. The number of infants with evidence of pulmonary artery hypertension was similar in the 2 groups in our study, and additional interventions were not required to maintain adequate oxygenation during cooling, as noted by other investigators.12 Rapid falls in body temperature during combined body/head cooling have been reported with the use of anticonvulsants, sedatives, or intercurrent hypoxemia.12 In contrast, esophageal temperature was stable after the initiation of cooling with the Blanketrol system and was not affected by the use of anticonvulsants or sedatives/hypnotic agents. Reduction in esophageal temperature in the hypothermia group was associated with a fall in heart rate consistent with other reports12–14 and the physiologic effects of cold.21
There are a number of distinguishing features of this feasibility study compared with previous pilot studies. The entry criteria consisted of clinical and biochemical data and did not include use of an amplitude-integrated EEG (aEEG) as described by Thoresen and Whitelaw12 and Azzopardi et al.13 Abnormalities of the aEEG have been advocated as useful entry criteria; however, these data were not available at the time our study was designed. Hellstrom-Westas et al22 evaluated 47 term infants with asphyxia with an aEEG performed within 6 hours of birth and accurately predicted outcome at 12 to 18 months in 43 of the infants. In the study by al Naqeeb et al, 23 aEEG recorded at a median age of 5 hours (range: 3–12 hours) among 24 of 56 infants with encephalopathy and 14 normal infants demonstrated a close association with subsequent outcome at 18 to 24 months of age. Toet et al24 assessed the prognostic value of an aEEG at 3 and 6 hours after birth in 73 term asphyxiated infants and found the predictive value of the 6-hour reading superior to the 3-hour reading. The aEEG demonstrated a change in pattern in 31% of infants, with 3 infants who improved and 2 who deteriorated between the 2 readings. This raises important questions regarding entry criteria and timing of the protective therapy. Although the aEEG was not used in the present study, an extremely high-risk population was identified using clinical, biochemical, and neurologic criteria as confirmed by 5 deaths, 5 surviving infants with neurologic abnormalities at discharge, and 6 infants with an abnormal MRI. For ensuring compliance with the narrowly defined entry criteria and consistency among the examiners, all neurologic examinations were performed by the site principal investigator or designee after a standardized certification process. The study protocol was conducted only at the participating centers and was not initiated during uncontrolled intervals such as transport of infants to the centers, and patients were assigned to study groups by a true randomization process. Given the relationship between esophageal and brain temperatures, there is reasonable certainty that the brain was cooled and maintained at 34.5°C. Finally, all blood gases were corrected for temperature (pH-stat concept) as there continues to be controversy as to whether blood gas samples from hypothermic patients should be analyzed at actual body temperature or at 37°C.25
Modest hypothermia seems to be a promising therapy for acute hypoxic-ischemic brain injury. There are currently several ongoing randomized, controlled trials of hypothermia for HIE using different entry criteria and cooling strategies. We have demonstrated the feasibility and safety of whole-body hypothermia to 34.5°C in infants with moderate or severe encephalopathy. A simple modification of a commercially available cooling system results in stable body and brain temperatures when used in the automatic control mode. We are currently conducting a large efficacy study (n = 200) to evaluate whether induced hypothermia with whole-body cooling initiated before 6 hours and continued for 72 hours in term infants with encephalopathy will reduce the incidence of death or disability at 18 months of age.
Acknowledgments
This study was supported by the National Institute of Child Health and Human Development through Cooperative Agreements U10 HD21385, U10 HD40689, U10 HD27871, U10 HD27853, U10 HD21364, U10 HD34167, U10 HD21373, U01 HD36790, U10 HD34216, U10 HD27856, U10 HD27904, U10 HD27851, U10 HD27881, U10 HD21397, U10 HD27880, and U10 HD21415.
Footnotes
- Received October 17, 2001.
- Accepted March 13, 2002.
- Reprint requests to (S.S.) Children’s Hospital of Michigan, 3901 Beaubien, Detroit, MI 48201. E-mail: s_shankaran{at}wayne.edu
REFERENCES
- Copyright © 2002 by the American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- Outcomes in the Era of Therapeutic Hypothermia
- Risk factors for EEG seizures in neonates treated with hypothermia: A multicenter cohort study
- Neonatal seizures in the era of therapeutic hypothermia: Keeping it cool
- Induced Hypothermia for Neonatal Hypoxic-Ischemic Encephalopathy: Pathophysiology, Current Treatment, and Nursing Considerations
- Temperature control during therapeutic moderate whole-body hypothermia for neonatal encephalopathy
- A Comparison of Cooling Methods Used in Therapeutic Hypothermia for Perinatal Asphyxia
- Neurological outcomes at 18 months of age after moderate hypothermia for perinatal hypoxic ischaemic encephalopathy: synthesis and meta-analysis of trial data
- Treatment of asphyxiated newborns with moderate hypothermia in routine clinical practice: how cooling is managed in the UK outside a clinical trial
- Therapeutic hypothermia can be induced and maintained using either commercial water bottles or a "phase changing material" mattress in a newborn piglet model
- Pediatric Cardiopulmonary Resuscitation: Advances in Science, Techniques, and Outcomes
- Time to Adopt Cooling for Neonatal Hypoxic-Ischemic Encephalopathy: Response to a Previous Commentary
- The International Liaison Committee on Resuscitation (ILCOR) Consensus on Science With Treatment Recommendations for Pediatric and Neonatal Patients: Neonatal Resuscitation
- 2005 American Heart Association (AHA) Guidelines for Cardiopulmonary Resuscitation (CPR) and Emergency Cardiovascular Care (ECC) of Pediatric and Neonatal Patients: Neonatal Resuscitation Guidelines
- Hemodynamics Among Neonates With Hypoxic-Ischemic Encephalopathy During Whole-Body Hypothermia and Passive Rewarming
- Therapeutic hypothermia following perinatal asphyxia.
- Hypothermia: A Neuroprotective Therapy for Neonatal Hypoxic-Ischemic Encephalopathy
- Part 13: Neonatal Resuscitation Guidelines
- Part 7: Neonatal Resuscitation
- Mild Hypothermia and the Distribution of Cerebral Lesions in Neonates With Hypoxic-Ischemic Encephalopathy
- Cardiac output, pulmonary artery pressure, and patent ductus arteriosus during therapeutic cooling after global hypoxia-ischaemia