PEDIATRICS Vol. 106 No. 1 July 2000, pp. 92-99

Cardiovascular Changes During Mild Therapeutic Hypothermia and Rewarming in Infants With Hypoxic-Ischemic Encephalopathy

Marianne Thoresen, MD, PhD^* and Andrew Whitelaw, MD

From the Division of Child Health, University of Bristol, ^* St Michael's Hospital, and  Southmead Hospital, Bristol, England.

	ABSTRACT

Top Abstract Methods Results Discussion References

Background.  Clinical trials of mild cooling to 35°C or below in infants with early hypoxic-ischemic encephalopathy are under way. The objective of this study was to systematically document cardiovascular changes associated with mild therapeutic hypothermia and rewarming in such infants.

Patients and Methods.  Nine infants with gestational ages of 36 to 42 weeks, with 10-minute Apgar scores of 5 or less, clinical encephalopathy, and an abnormal electroencephalogram before 6 hours were cooled by surface cooling the trunk (n = 3) or by applying a cap perfused with cooled water (n = 6) for a median of 72 hours. The target core temperature was 34.0°C to 35.0°C for head-cooled infants and 33.0°C to 34.0°C for surface-cooled infants. Maintenance heating and rewarming were provided by an overhead heater.

Results.  Mean arterial blood pressure increased by a median of 10 mm Hg during cooling and fell by a median of 8 mm Hg on rewarming. Heart rate decreased by a median of 34 beats/minute on cooling and increased by a median of 32 beats/minute on rewarming. A large increase in the output of the overhead heater decreased mean arterial blood pressure in 5 infants. Anticonvulsant drugs, sedatives, or intercurrent hypoxemia also produced falls in temperature. The inspired oxygen fraction had to be increased by a median of .14 to maintain oxygenation during cooling with 2 infants requiring 100% oxygen, an effect probably attributable to pulmonary hypertension, which was reversible with rewarming.

Conclusions.  Therapeutic cooling produces changes in heart rate and blood pressure that are not hazardous, but the combination of inadvertent overcooling and inappropriately rapid rewarming, together with sedative drugs that can impair normal thermoregulatory vasoconstriction, can cause hypotension in posthypoxic newborn infants. Infants who already require 50% oxygen should be cooled cautiously because pulmonary hypertension may develop. Knowledge of these cardiovascular changes, careful monitoring, anticipation, and correction should help to avoid potential adverse effects in the upcoming clinical trials.  Key words:  hypothermia, cardiovascular, newborn, encephalopathy, head cooling.

There is now considerable evidence that mild systemic cooling after cerebral hypoxia-ischemia can reduce brain damage in adult and neonatal animal models.^1,2 Reductions in body temperature of 2°C to 6°C have been shown to reduce neuropathologically measured brain injury^3-7 and also secondary energy failure⁸ and lactate accumulation⁹ on magnetic resonance spectroscopy. Only 1 neonatal study, however, has shown a long-lasting effect.¹⁰

These studies have reduced core temperature by reducing environmental temperature or applying cooling to the trunk. Experimentally, systemic cooling increases the release of catecholamines, peripheral vasoconstriction, and systemic arterial pressure.¹¹

Cooling the head after birth asphyxia by passing cold water through a cap containing rubber coils was pioneered in Russia by Kopchev¹² in the 1970s and 1980s, but no controlled evaluation has been published. In the 1990s, Gunn et al¹³ applied this principle to their fetal sheep model of cerebral ischemia. They showed that cooling the extradural temperature to 30°C to 33°C for 72 hours reduced brain damage even when cooling started as late as 5.5 hours after reperfusion. In this model, the core (esophageal) temperature fell by 1.5°C in response to selective head cooling. The core temperature of the fetal sheep was prevented from falling further by the placenta perfusing the body with warm blood. There was a small increase in systemic arterial pressure when cooling was started and a similar fall in systemic arterial pressure when cooling was stopped. Gunn et al¹⁴ then applied selective cooling to infants with perinatal asphyxia and early neurological signs. Head cooling was still applied by cold water in rubber coils applied to the head but core temperature was maintained by a different method. In place of perfusion with warm blood (from the placenta), surface warming via an overhead radiant heater was used. Their experience with a pilot study of 6 asphyxiated infants cooled to a rectal temperature of 36.3° ± .2°C and 6 asphyxiated infants cooled to 35.7° ± .2°C was published in 1998.¹⁴ The entry criteria for this study were: 1) pH at or below 7.09, 2) 5-minute Apgar score at or below 6, 3) clinical signs consistent with encephalopathy, and 4) gestation at or above 37 weeks. No infants developed arrhythmia, hypotension (mean arterial pressure: <40 mm Hg), bradycardia (heart rate [HR]: <80/minute) during cooling. One infant, whose rectal temperature inadvertently fell to 34.2°C, developed an increase in oxygen requirement of 40% and pH fell to 7.29, but both of these changes reversed when the rectal temperature was raised to 35.5°C. The article concluded, "Mild selective head cooling combined with mild systemic hypothermia in newborn infants after perinatal asphyxia is a safe and convenient method of quickly reducing cerebral temperature."¹⁴

In light of these encouraging results, 2 multicenter trials of mild hypothermia for infants with early hypoxic-ischemic encephalopathy (HIE) have started with a third being planned. All of these trials are planning a target rectal temperature lower than the temperature used in the safety study by Gunn et al¹⁴ and recruitment of infants with more severe entry criteria. Because avoidance of hypotension is 1 of the few principles generally agreed upon in the management of infants with HIE, we believed there was a need for data on possible cardiovascular changes in critically ill infants being managed by cooling to systemic temperatures <35.5°C.

We report the findings of an exploratory study of cardiovascular changes in 9 infants with perinatal asphyxia and electroencephalogram (EEG)-confirmed encephalopathy, 3 of whom were systemically cooled and 6 of whom were cooled by a cooling cap. The observations were part of a pilot study and were made, of necessity, before a complete treatment protocol had been developed. The larger randomized trial currently running has adopted different treatment strategies designed to avoid the problems described here. Nevertheless, we believe that our observations illustrate important pathophysiological processes and may help a wider readership of colleagues planning or taking part in hypothermia trials to avoid undesirable circulatory changes.

	METHODS

Top Abstract Methods Results Discussion References

Systemic hypothermia by surface cooling and mild hypothermia by head cooling have both been approved as pilot studies by the Research Ethics Committees of Southmead Health Services, National Health Service Trust and the United Bristol Health Care, National Health Service Trust. Informed parental consent was given for each infant before cooling was initiated.

Nine full-term infants were prospectively recruited based on the following criteria:

1. Apgar score of 5 or less at 10 minutes, or failure to establish spontaneous respiration at 10 minutes, or pH < 7.0 on cord blood or within 30 minutes of birth.
2. Signs of encephalopathy, eg, lethargy, stupor, hypotonia, absent suck, and clinical seizures.
3. Reduced amplitude on EEG or cerebral function monitor (CFM) within 6 hours of birth.¹⁵
4. Gestational age of 36 weeks or more.
5. Absence of congenital abnormalities.

HR, mean arterial blood pressure (MABP), and oxygen saturation were measured continuously and stored every minute in a Hewlett Packard 66S monitor (Hewlett Packard, Palo Alto, CA). For the purposes of comparison, 15-minute average values for HR and MABP before and after temperature manipulations were used. During the hypothermic period, values at 60-minute intervals were used for calculations of mean and standard deviation (SD). Rectal temperature was measured continuously with a YSI 4499E thermistor (Yellow Springs Industries, Yellow Springs, OH) with the tip 5 cm inside the rectum. The peripheral temperature on the plantar surface of 1 foot was measured continuously with a Yellow Springs thermistor. The accuracy of the thermistors was checked against a mercury thermometer with a .1°C resolution.

Transcutaneous oxygen saturation was measured using an Oxisensor 2 N25 probe (Nellcor Puritan Bennett, Coventry, England) on the sole of the contralateral foot and the accuracy of oxygen saturation was checked by confirming that the HR on the pulse oximeter was within 5 beats/minute of the HR on the electrocardiogram.¹⁶ Several studies have documented that transcutaneous oxygen saturation in infants is accurate if the skin temperature is >29°C or core temperature is >31.3°C.^16-18 In all the infants studied, rectal temperature was maintained above 31.3°C, and in the 1 systemically cooled infant, whose peripheral skin temperature fell temporarily below 29.0°C, the oximeter pulse rate concurred with the electrocardiographic HR. The principal respiratory management strategy was to maintain transcutaneous oxygen saturation at 95% or above. Oxygen saturation is not affected by a reduction in body temperature from 37°C to 32.8°C; however, partial pressure of arterial oxygen is affected.¹⁹ There is controversy as to whether blood gas samples from hypothermic patients should be analyzed at actual body temperature or at 37°C.²⁰ Because there is evidence from hypothermic cardiac surgery that neurological dysfunction is more likely if treatment is based on pH and partial pressure of carbon dioxide in arterial gas (PaCO₂) corrected for temperature (pH-stat strategy), we analyzed all blood gas samples at 37°C.^21,22 The ventilator strategy was to maintain PaCO₂ in the range of 40 to 50 mm Hg.

Systemic cooling was achieved by removing all heating, exposing the infant, and placing rubber gloves filled with water at ~10°C next to the groins, axillae, and neck. Occasionally an electric fan was used, and cooling was adjusted to maintain the rectal temperature in the range of 33.0°C to 34.0°C for 72 hours.

Selective head cooling was achieved by using a prototype plastic cooling cap based on the work of Gunn et al.¹⁴ The cap applied to the infant's head contained channels through which water was circulated from a cooling unit in which water temperature could be controlled. The water temperature was initially set between 10°C and 15°C. When the infant's rectal temperature fell below 35.5°C, a radiant heater over the infant's trunk was switched on. A plastic shield with reflecting foil covered the infant's head. A manually controlled radiant heater (Draeger Babytherm 8000, Draeger, Lübeck, Germany) provided temperature maintenance and was initially set on maximum output (14 mW/cm²) but was reduced if the infant's rectal temperature was inappropriately high or the blood pressure fell. The temperature of the water circulating in the cap was adjusted to maintain the infant's rectal temperature between 34.0°C and 35.0°C. Cooling to the target temperature was achieved within 30 to 90 minutes. Mild hypothermia was maintained for a median of 72 hours (range: 18-72 hours). Rewarming to a rectal temperature of 37.0°C was conducted slowly over a minimum of 5 hours by removing the cooling cap and adjusting the overhead heater to give a rise in rectal temperature of not more than .5°C every hour.

A single-channel, amplitude-integrated EEG was recorded continuously in 8 infants using the Cerebral Function Monitor (CFM 4640, Lectromed, Letchworth, UK).¹⁵ In the first infant to be systemically cooled, 2-channel EEG was recorded using the Medilog 9000 system (Oxford Medical Systems, Abingdon, UK).²³

The medical management of the infants otherwise was similar to that outlined by Levene²⁴ with maintenance of oxygenation, normoglycemia, normotension, appropriate respiratory assistance, and treatment of clinical seizures, but not subclinical electrical seizures. Phenobarbitone was not given prophylactically.

Table 1 shows the 9 infants with birth weights between 2100 g and 4200 g and gestational ages between 36 and 42 weeks. Ten-minute Apgar scores ranged from 1 to 5. 

Initial pH ranged from 6.8 to 7.1 and initial base deficit from 12 to 22 mmol/L. Initial EEG showed a burst-suppression pattern in 2 cases, severely depressed CFM pattern in 2, and moderately abnormal pattern in 5.^15,23 Seven infants had clinical seizures with onset between 2 and 24 hours (median: 9 hours). All were intubated after birth as a part of resuscitation, 7 were mechanically ventilated at the start of cooling, and 5 were still ventilated at rewarming. No infant had received inotropic support when cooling was started, but 3 had received volume expansion for hypotension (MABP: <40 mm Hg). Five infants received inotropic support (dopamine) because of hypotension during hypothermia and/or rewarming, and 6 infants received volume expansion because of hypotension during hypothermia/rewarming.

Anticonvulsant therapy of clinical seizures was initially 20 mg/kg of phenobarbitone intravenously over 20 minutes (7 infants), followed by a further 10 mg/kg intravenously if necessary. Further anticonvulsant therapy or sedation, if the infant was unsettled, was given in the form of 100 µg/kg of midazolam, followed by an infusion at 30 to 60 µg/kg/hour for a variable duration (6 infants). Infants who were mechanically ventilated were given morphine if their own respiratory efforts were compromising the effectiveness of the ventilator (3 infants). If morphine was inadequate for this purpose, pancuronium was added (1 infant).

	RESULTS

Top Abstract Methods Results Discussion References

Patient characteristics are given in Table 1.

Systemic Arterial Pressure and HR Changes on Cooling and Rewarming

When cooling was commenced, median MABP was 43 mm Hg (range: 30-50). When the target temperature was reached, median MABP was 54 mm Hg (range: 38-60; Fig 1). This median change in MABP was +26% of baseline value (range: 10% to +57%). HR fell by a median of 34 beats/minute (range: 14-60) on cooling. The increase in MABP is illustrated in Fig 1 and also in Fig 6. Calculating the average HR for each infant during the entire hypothermia period gives a median value of 100/minute (range: 88-127/minute). Five of the 9 infants had an HR <80/minute for at least 30 minutes during hypothermia (Table 2), but sinus rhythm was maintained throughout.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   MABP before and after cooling and before and after rewarming in all 9 infants.

View larger version (35K): [in this window] [in a new window]   Fig. 6.   A, Infant K was cooled using a head cap. A bolus of midazolam was administered at 34 hours.* Rectal temperature fell, more heating was applied, and MABP fell temporarily. B, Considerable instability of MABP and rectal temperature in infant M who was cooled by a head cap. Phenobarbitone and midazolam were given.# C, infant P was treated with whole body cooling with adequate mean arterial pressure and rectal temperature throughout. She was extubated after 24 hours.

When rewarming was commenced, median MABP was 53 mm Hg (range: 43-64). When a rectal temperature of 37°C was achieved, median MABP was 45 mm Hg (range: 3056; Fig 1). Median change in MABP on rewarming was 8 mm Hg or 13% of baseline (range: +2% to 31%). HR increased by a median of 32 beats/minute (range: 20-70) on rewarming.

Pulmonary Hypertension Associated With Cooling

At the start of cooling, 5 of the infants were breathing air and the other 4 infants required 29% to 55% oxygen. The fraction of inspired oxygen (FIO₂) required to maintain oxygen saturation of 95% during hypothermia increased by a median of +14% (range: 0%-45%). In 2 infants with meconium aspiration, 100% oxygen was required during hypothermia. One of these infants had echocardiographic evidence of right-to-left shunting through the foramen ovale and adequate ventilation with a PaCO₂ of 30 mm Hg. The oxygen requirement fell to ~50% after starting inhalation of nitric oxide and rewarming. The second infant had a partial pressure of arterial oxygen of 41 mm Hg and PaCO₂ of 30 mm Hg, while ventilated at a pressure of 18/4-cm water and FIO₂ of 1.0. There was no sustained improvement in oxygenation after surfactant treatment or paralysis, chest radiograph showed well-expanded lungs, and it was only after rewarming that FIO₂ could be reduced to below .3 (Fig 2).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Reversible increase in FIO2 required to achieve oxygen saturation of 90% in infant C as rectal temperature fell.

The most likely explanation for the worsening of oxygenation during hypothermia is that cooling induced pulmonary vasoconstriction and pulmonary hypertension. Such an effect of cooling is well-documented.²⁵ In these 2 cases, cooling was discontinued early because of the concerns over pulmonary hypertension. In all cases, the increase in oxygen requirement was reversible with rewarming.

Effect of the Overhead Heater on Circulation

Increasing the output of the overhead heater decreased MABP in 5 infants. This is illustrated in Fig 3, where increasing the abdominal temperature (by increasing the output of the overhead heater from 50% to 100%) decreased MABP from over 50 mm Hg to below 40 mm Hg, despite the use of dopamine at 10 µg/kg/minute. This necessitated a reduction in the output from the overhead heater, and this action was quickly followed by a drop in abdominal temperature and a rise in MABP.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   The reversible effect of increasing abdominal skin temperature (by overhead heating) on MABP in infant M.

Effect of Anticonvulsant or Sedative Therapy

Anticonvulsant doses of phenobarbitone or midazolam reduced spontaneous muscular activity. A marked reduction in muscular activity was followed by a fall in rectal temperature in 5 of 6 infants receiving midazolam. In Fig 4, administration of midazolam to an infant with adequate MABP and rectal temperature was followed by a drop in rectal temperature of 1°C and a fall in MABP of 15 mm Hg. Intercurrent hypoxemia reduce heat production and rectal temperature (Fig 5). Also, drops in temperature followed by increasing warming were associated with a drop in blood pressure in Fig 6, A and B. 

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Intravenous midazolam is followed by a drop in rectal temperature and MABP in infant S.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   A period of marked desaturation is accompanied by a drop in rectal temperature to <33°C in infant C.

Variability of Temperature and MABP

Table 2 shows the mean and SD for rectal temperature and MABP during the whole cooling period for each infant. The variability was similar in the infants cooled via the body and in the infants cooled via the head. Eight of 9 infants had rectal temperatures below the target temperature range (median overshoot: .8°C; range: .3°C-1.6°C]) for a variable duration (median: 5.5 hours; range: .5-20 hours).

3 Illustrative Cases

Figure 6A shows mean arterial pressure and rectal temperature throughout cooling and rewarming using a cooling cap in infant K. He was initially ventilated but extubated ~1 hour after cooling and did not receive any inotropic drugs. A bolus dose of midazolam (100 µg/kg) was administered at 34 hours because of tremor and irritability. Reduced movements reduced heat production and rectal temperature fell. Overhead heating was increased and the blood pressure fell. This effect was reversible.

Figure 6B shows mean arterial pressure and rectal temperature throughout the treatment of infant M. She was ventilated throughout and needed inotropic support and had frequent seizure episodes for which she received phenobarbitone 30 mg/kg and a continuous infusion of midazolam starting at 9 hours. It was difficult to maintain a stable temperature and blood pressure through periods of seizures and drug therapy. Rectal temperature fell below the target range for 19 hours. A cranial ultrasound examination before commencement of cooling was normal, but at 60 hours, ultrasound and computed tomography revealed a significant right-sided subdural hematoma with midline shift associated with coagulopathy and thrombocytopenia.

Figure 6C shows MABP and rectal temperature on infant P who was treated with whole body cooling. She was extubated at 24 hours of age. Her EEG showed frequent seizure activity (mostly subclinical) for the first 48 hours. She received phenobarbitone and midazolam by infusion but maintained her blood pressure throughout and never required any inotrope.

	DISCUSSION

Top Abstract Methods Results Discussion References

This pilot study was conducted at a time when there was little knowledge or experience of therapeutic cooling of critically ill, posthypoxic human infants. A number of potentially serious events that seem to be related to our therapeutic strategies occurred. The cardiovascular events encountered can be summarized by the following statements:

1. Inadvertent reduction in core temperature below the target range from active cooling and/or anticonvulsant/sedative drugs necessitating active rewarming.
2. Too powerful or too rapid rewarming, leading to peripheral vasodilatation lowering the blood pressure.
3. Increasing oxygen requirement during hypothermia probably attributable to pulmonary hypertension.
4. Sinus bradycardia during hypothermia, which did not seem to have clinical consequences, and no episodes of arrhythmia detected.

Inadvertent Reduction in Core Temperature Below Target Range

Infants differed greatly in their ability to resist cooling and it was difficult to predict how fast and how far the core temperature would fall. A fall in core temperature could be precipitated in a previously stable infant by the intravenous administration of drugs, such as phenobarbitone or midazolam, reducing the infant's spontaneous activity, and because muscular activity is an important source of heat production, it is not surprising that the body temperature fell. Midazolam could have accentuated this tendency still further by impairing thermoregulatory vasoconstriction.²⁶ Intercurrent hypoxemia (with oxygen saturation <60%, as happened to infant C in Fig 5) can be expected to reduce heat production.

Systemic Hypotension on Rewarming

The median 8-mm Hg (13%) drop in MABP associated with rewarming was not associated with any adverse clinical event in any of our infants, but it could be clinically detrimental in a critically ill infant whose arterial pressure is already borderline. Five of the infants received inotropic therapy and 6 of the infants received volume expansion during rewarming, and the full potential for hypotension has, thereby, been masked. The maintenance of systemic blood pressure is dependent on the ability of the sick neonate to vasoconstrict peripherally. This is an important compensatory mechanism for dealing with hypovolemia and maintaining cerebral and renal perfusion. Warming the trunk and limbs with an overhead heater may interfere with the ability of the infant to regulate the peripheral circulation in the interests of the vital organs. Whether the strategy of cooling the whole body results in different circulatory regulation from a strategy of cooling the head and warming the trunk and limbs is a question deserving further investigation. Benzodiazepine drugs, including midazolam, facilitate peripheral vasodilatation and reduce cardiac output^26,27 and may, therefore, have accentuated this hypotension on rewarming in our patients. A further consideration is that the hepatic metabolism of benzodiazepines and barbiturates may be reduced when core temperature is reduced and drugs given by continuous infusion, such as midazolam, therefore, may accumulate.

Pulmonary Hypertension

Although no formal measurements of pulmonary artery pressure were made, the evidence is highly suggestive of increased pulmonary vasoconstriction in the 2 infants who required 100% oxygen during cooling. Both of these infants had significant oxygen requirements, 50% and 55% before cooling started.

A theoretical possibility is impairment of pulmonary surfactant by hypothermia, but we considered that explanation to be unlikely because of the lack of carbon dioxide retention and absence of radiologic evidence of surfactant deficiency. This marked increase in oxygen requirement during hypothermia could have had serious clinical consequences if intensive monitoring and therapy had not been available.

Reduction in HR

The drop in HR associated with cooling was a sinus bradycardia not associated with any apparent underperfusion as indicated by change in EEG, oliguria, or increased metabolic acidosis. No arrhythmia but only a sinus bradycardia down to 75 to 90/minute was also reported by Westin et al,²⁸ who cooled asphyxiated infants immediately after birth. The low HR combined with increased stroke volume may be adequate for the reduced metabolic rate during mild hypothermia.

We report our experience of cardiovascular changes with therapeutic hypothermia and associated rewarming, because our findings differ from those in the safety study by Gunn et al.¹⁴ In our study, the infants had EEG confirmation of early HIE and were likely to have been sicker than those in the article by Gunn et al. In addition, the target core temperature in our study was >1 degree lower than that in the report by Gunn et al. Our treatment strategy differed from that of Gunn et al¹⁴ in our liberal use of midazolam as anticonvulsant and sedative and in our manual control of the overhead heater rather than servo-control.

Recommendations to Avoid Adverse Cardiovascular Events

We suggest the following measures to reduce the risk of adverse cardiovascular events during therapeutic hypothermia:

1. If the infant is in poor condition at birth, umbilical artery and venous catheters should be inserted for reliable continuous arterial pressure and blood gas measurements and infusion of volume and inotrope as needed. If the umbilical venous catheter is in the vena cava, continuous measurement of central venous pressure can help to distinguish hypovolemia from myocardial insufficiency and so guide volume expansion or inotropic support.
2. When the rectal temperature falls below 35.5°C, the rate of cooling should be reduced and carefully monitored by 1 member of the medical or nursing staff with no other patient responsibilities until the temperature is stable in the target range to avoid overshooting. If the rate of cooling is reduced by overhead heating, servo-control of the heater to the abdominal skin temperature allows an appropriate fraction of power to be given automatically without the nursing or medical staff having to decide.
3. When stable hypothermia has been achieved, the subsequent administration of sedatives, anticonvulsants, opiates, or muscle relaxants can be expected to reduce heat production by the infant and, thereby, reduce core temperature unless active cooling is reduced. The same may happen if the infant becomes hypoxic during hypothermia.
4. Infants who have an oxygen requirement of 50% or more before cooling should be carefully monitored to determine whether cooling further increases the oxygen requirement, in which case, the clinician should reconsider any further cooling.
5. Rewarming at the end of therapeutic hypothermia, or because the core temperature has inadvertently fallen below the target range, should be conducted slowly at not >.5°C per hour. Whenever rewarming at maximum radiant heater output is being applied to the skin of the sick hypothermic infant, arterial pressure should be closely monitored and volume expansion or inotrope support must be readily available. Systemic hypotension is particularly undesirable if it is combined with pulmonary hypertension because right-to-left shunting may be accentuated.

The evidence from animal models of cerebral hypoxia-ischemia strongly supports the protective effect of mild hypothermia after cerebral hypoxia-ischemia^1-10; therefore, it is essential that clinical trials of therapeutic hypothermia are conducted in newborn infants. It should be emphasized that none of the circulatory changes we observed caused actual harm to any of our patients. We hope that knowledge of cardiovascular changes and how to avoid potential complications will contribute toward the safe completion of these randomized trials.

	ACKNOWLEDGMENTS

We thank the parents of the infants studied for their courage and support, and the neonatal nurses and medical staff of Southmead Hospital, Bristol, and St Michael's Hospital, Bristol.

We thank Dr Saulius Satas for translating reference 12 from Russian to English, as well as for practical assistance.

We thank Professors John Wyatt and Peter Gluckman and Dr Alistair Gunn for helpful discussions on this manuscript.

	FOOTNOTES

Received for publication Sep 16, 1999; accepted Jan 10, 2000.

Reprint requests to (A.W.) Division of Child Health, University of Bristol, Medical School, Southmead Hospital, Bristol, BS10 5NB, England. E-mail: andrew.whitelaw{at}bristol.ac.uk

	ABBREVIATIONS

HR, heart rate; HIE, hypoxic-ischemic encephalopathy; EEG, electroencephalogram; CFM, cerebral function monitor; MABP, mean arterial blood pressure; SD, standard deviation; PaCO₂, partial pressure of carbon dioxide in arterial gas; FIO₂, fraction of inspired oxygen.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Wyatt JS, Thoresen M Hypothermia treatment and the newborn. Pediatrics 1997; 100:1028-1030 [Free Full Text]
2. Gunn AJ, Gunn TR The "pharmacology" of neuronal rescue with cerebral hypothermia. Early Human Dev 1998; 53:19-35 [CrossRef][Medline]
3. Thoresen M, Bågenholm R, Løberg EM, Apricena F, Kjellmer I Posthypoxic cooling of neonatal rats provides protection against brain injury. Arch Dis Child 1996; 74:F3-F9
4. Sirimanne ES, Blumberg RM, Bossano D, Gunnong M, Edwards AD, Gluckman PD The effect of prolonged modification of cerebral temperature on the outcome after hypoxic-ischemic brain injury in the infant rat. Pediatr Res 1996; 39:591-597 [Medline]
5. Gunn AJ, Gunn TR, deHaan 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]
6. Haaland K, Løberg EM, Steen PA, Thoresen M Posthypoxic hypothermia in newborn piglets. Pediatr Res 1997; 41:505-512 [Medline]
7. Trescher WH, Ishiwa S, Johnston MV Brief post-hypoxic-ischemic hypothermia markedly delays neonatal brain injury. Brain Dev 1997; 19:326-338 [CrossRef][Medline]
8. Thoresen M, Penrice J, Lorek A, Mild hypothermia following severe transient hypoxia-ischemia ameliorates delayed ("secondary') cerebral energy failure in the newborn piglet. Pediatr Res 1995; 5:667-670
9. Amess PN, Penrice J, Lorek A, Mild hypothermia after severe transient hypoxia-ischemia reduces the delayed rise in cerebral lactate in the piglet. Pediatr Res 1997; 41:803-808 [Medline]
10. Bona E, Hagberg H, Løberg EM, Bågenholm R, Thoresen M Protective effects of moderate hypothermia after neonatal hypoxia-ischemia: short- and long-term outcome. Pediatr Res 1998; 43:738-745 [Medline]
11. Schubert A Side effects of mild hypothermia. J Neurosurg Anesthesiol 1995; 7:139-147 [Medline]
12. Kopchev SN. Cranio-cerebral hypothermia in obstetrics. Moscow, Russia: Medicina; 1985:1-112
13. Gunn AJ, Gunn TR, Gunning MI, Williams CE, Gluckman PD Neuroprotection with prolonged head cooling started before postischemic seizures in fetal sheep. Pediatrics 1998; 102:1098-1106 [Abstract/Free Full Text]
14. Gunn AJ, Gluckman PD, Gunn TR Selective head cooling in newborn infants after perinatal asphyxia: a safety study. Pediatrics 1998; 102:885-892 [Abstract/Free Full Text]
15. Al Naqeeb N, Edwards AD, Cowan FM, Azzopardi D Assessment of neonatal encephalopathy by amplitude-integrated electroencephalography. Pediatrics 1999; 103:1263-1271 [Abstract/Free Full Text]
16. Macnab AJ, Baker-Brown G, Anderson EE Oximetry in children recovering from deep hypothermia for cardiac surgery. Crit Care Med 1990; 18:1066-1069 [Medline]
17. Picardo S, DiChiara L, Averardi M, Testa G, Giamberti A, Catena G Unreliability of pulse oximetry in hypothermic children after cardiovascular surgery with deep hypothermic circulation. Minerva Anestesiol 1998; 64:427-430 [Medline]
18. Iyer P, McDougall P, Loughnan P, Mee RB, Al-Tawil K, Carlin J Accuracy of pulse oximetry in hypothermic neonates and infants undergoing cardiac surgery. Crit Care Med 1996; 24:507-511 [CrossRef][Medline]
19. Dill DB, Forbes WH Respiratory and metabolic effects of hypothermia. Am J Physiol 1941; 132:685-697 [Free Full Text]
20. Ream AK, Reitz BA, Silverberg G Temperature correction of PCO₂ and pH in estimating acid-base status: an example of the emperor's new clothes? Anesthesiology 1982; 56:41-44 [Medline]
21. Stephan H, Weyland A, Kazmaier S, Henze T, Menck S, Sonntag H Acid-base management during hypothermic cardiopulmonary bypass does not affect cerebral metabolism but does affect blood flow and neurological outcome. Br J Anaesth 1992; 69:51-57 [Abstract/Free Full Text]
22. Patel RL, Turtle MR, Chambers DJ, Newman S, Venn GE Hyperperfusion and cerebral dysfunction: effect of differing acid-base management during cardiopulmonary bypass. Eur J Cardiothorac Surg 1993; 7:457-463 [Abstract]
23. Connell J, Oozeer R, de Vries L, Dubowitz LM, Dubowitz V Continuous EEG monitoring of neonatal seizures: diagnostic and prognostic considerations. Arch Dis Child 1989; 64:452-458 [Abstract]
24. Levene MI Management of the asphyxiated full-term infant. Arch Dis Child 1993; 68:612-616 [Medline]
25. Benumof JL, Wahrenbrock EA Dependency of hypoxic pulmonary vasoconstriction on temperature. J Appl Physiol 1977; 42:56-58 [Abstract/Free Full Text]
26. Matsukawa T, Hanagata K, Ozaki M, Iwasashita H, Koshimizu M, Kumazawa T Intramuscular midazolam as premedication produces a concentration-dependent decrease in core temperature in male volunteers. Br J Anaesth 1997; 78:396-399 [Abstract/Free Full Text]
27. Shekerdamian L, Bush A, Redington A Cardiovascular effects of intravenous midazolam after open-heart surgery. Arch Dis Child 1997; 76:57-61 [Abstract/Free Full Text]
28. Westin B, Nyberg R, Miller JA, Wedenberg E Hypothermia and transfusion with oxygenated blood in the treatment of asphyxia neonatorum. Acta Paediatr Scand 1962; 51:1-80