Is It Time for a Randomized Controlled Trial of Hypothermia for Mild Hypoxic-Ischemic Encephalopathy?

The time is right to perform a randomized controlled trial (RCT) of neuroprotection for neonates with mild hypoxic-ischemic encephalopathy (HIE). According to published reports, more than 600 infants with mild HIE have undergone hypothermia at 33.5°C for 72 hours.^1–6 Data on neuroimaging findings following cooling for mild HIE from these reports are now becoming available. Depending on the cohort evaluated, whether retrospective or prospective, and whether infants received cooling or not, 11%–40% of infants with mild HIE have brain injury and watershed injury is seen more frequently than basal ganglia or thalamus injury.^1,2,5,6 In addition, a recent survey of practices in the United Kingdom noted that physicians from 76% of the centers that offered hypothermia for moderate or severe HIE had cooled infants with mild HIE.⁷ In a small study of infants with mild HIE who were cooled for less than 72 hours owing to rapid clinical improvement, brain injury was still present on magnetic resonance imaging.⁸ The majority of these reports do not report monitoring of safety with cooling or provide any neurodevelopmental outcome data. Therefore, neither the safety nor efficacy of hypothermia for mild HIE is known.

A recent systematic review noted that up to 25% of infants with mild HIE are at risk for neurodevelopmental deficits in infancy or childhood.⁴ However, studies reporting outcomes of infants with mild encephalopathy are challenging to evaluate because they include an inconsistent definition of mild HIE, most are retrospective in nature, the inclusion of infants is not limited to HIE as a possible etiology, the diagnosis of mild encephalopathy was made over hours to days, and there were varying rates of follow-up and different psychometric measures used to assess outcome. Last, there is a lack of a control group in the majority of reports published in the hypothermia era.^9–15 A meta-analysis of infants with mild HIE who had received hypothermia compared with those who were not cooled in randomized and quasi-RCTs noted that current evidence was insufficient to recommend cooling for infants with mild HIE and significant benefits or harm could not be excluded.¹⁶ A recent prospective observational study of infants with severe birth acidosis or the need for resuscitation diagnosed with mild HIE (any abnormality in Sarnat staging at <6 hours of age) has noted that 16% of infants with mild HIE had a disability at 18–22 months of age.^17,18 A recent multicenter study performed in the United Kingdom and the US evaluating infants with neonatal encephalopathy who received hypothermia therapy reported the outcome of 37 infants with mild encephalopathy who were cooled. Eleven percent of the infants had moderate or severe injury on magnetic resonance imaging (all white matter injury). Magnetic resonance spectroscopy was within normal limits among all infants with mild encephalopathy and 30 of the 31 infants evaluated at 2 years had a normal outcome.¹⁹

Should hypothermia therapy be offered to infants with mild HIE? This strategy was discussed in a recent consensus paper.²⁰ Preclinical data justify cooling for mild hypoxiaischemia; hypothermia decreases white matter injury in the mouse model of mild hypoxia-ischemia.²¹ Mildly injured brains demonstrated the greatest neuroprotection with hypothermia. Regarding the duration of cooling, there is evidence from fetal sheep that cooling for 48 hours was less protective than cooling for 72 hours.²²

What is the incidence of mild HIE? Evaluation of the incidence from clinical observational studies, national registry, or survey data is challenging partly because the number of infants evaluated with a neurologic examination following acidosis or a need for resuscitation at birth is not always available.^{1–3,7,12–15} Lee et al have noted that, among 125 million births in 2010, 1.15 million neonates were diagnosed with encephalopathy after intrapartum related events or perinatal respiratory depression; of these, 38.2% were mild, 38.0% moderate, and 23.7% severe encephalopathy.²³ The type of perinatal events, status at birth, and diagnosis of encephalopathy varied in the reports included in the analyses.²³ In the California registry of neonates with recognized encephalopathy within 72 hours of birth in 2010–2012, 237 of 829 infants with HIE (28.5%) were characterized as having mild HIE.³ Saw et al performed a systematic review of observational studies from 2013 to 2017 on neonates with HIE receiving hypothermia; of 2783 who received hypothermia, 573 (20.5%) had mild encephalopathy.²⁴ The National Institute of Child Health and Human Development (NICHD) Neonatal Research Network published a RCT of longer and deeper cooling among infants with moderate or severe HIE in 2017.²⁵ Among infants with severe birth acidosis or need for resuscitation who were screened for trial eligibility with a standardized neurologic examination at less than 6 hours (n = 657), 293 (44.5%) had either mild HIE or a normal examination; unfortunately, the data collection did not distinguish between a normal or mild HIE exam. The Prospective Research for Infants with Mild Encephalopathy (PRIME) study noted that of 311 term infants with birth acidosis or need for resuscitation (NICHD screening criteria for HIE evaluation) evaluated at less than 6 hours of age, 160 had moderate or severe encephalopathy, 75 were normal, and 76 (24%) had mild encephalopathy.¹⁷ Thus, the incidence of mild HIE among infants diagnosed with encephalopathy ranges from 20% to 36%, depending on the cohort evaluated.

Any RCT evaluating the safety and efficacy of hypothermia for mild HIE should evaluate cognitive and behavioral outcomes, because the deficits that occur in some cases of mild HIE tend to be subtle as compared with those with moderate or severe HIE.^11,26 The mortality rate is anticipated to be low. The trial should have an adequate effect size between groups based on a robust measurable outcome (ie, the Bayley III composite cognitive, language, and motor score). The study screening criteria should be consistent with those for moderate or severe HIE: 36 weeks of gestation or greater, severe acidosis or in the absence of blood gases, evidence of an acute sentinel event, and need for birth resuscitation continued for at least 10 minutes.^25,27 These criteria will limit inclusion of infants with mild HIE to those who are at greatest risk for neurodevelopmental deficits.

Currently there are no early (<6 hours of age) neurophysiologic or imaging biomarkers of outcome after mild HIE, so the modified Sarnat neurologic examination needs to be performed. The qualifying neurologic examination for mild HIE should capture infants at greatest risk, such as including infants with more than only one abnormal category (out of six categories in the modified Sarnat examination); a recent scoring system that can be performed at less than 6 hours of age may discriminate higher risk of poor outcome after mild HIE.²⁸ The examination should be performed by trained examiners using a manual of procedures as has been used in trials of moderate or severe HIE.²⁹ The role of amplitude integrated electroencephalogram should be explored as part of the inclusion criteria. Many of the current publications on mild HIE document that selected infants had abnormal electroencephalograms before discharge.^11,24,30 The depth and duration of cooling should continue to be the current practice for moderate or severe HIE until further data are available regarding a shorter duration or less depth of cooling. The intervention of cooling (72 hours at 33.5°C), or normothermia, should be monitored for adverse events associated with cooling in moderate or severe HIE cases, including thrombocytopenia.³¹ Elevated temperatures should be avoided in the infants assigned to the normothermia group and before and after cooling in those assigned to the cooled group. Maternal sedation use should be documented. Investigators should use a current pain scale to assess neonatal distress during cooling and also evaluate other measures of stress such as salivary cortisol or heart rate variability as secondary studies. The clinical management strategy should be standardized to diminish variability in clinical practice in the 2 study intervention groups.

Neonatal magnetic resonance and diffusion-weighted imaging should be performed within the first week and read by more than 1 central reader because inter-rater and intrarater variability among neuroradiologists can be significant.^32–34 The magnetic resonance imaging scoring system used should be validated with neurodevelopmental follow-up.^6,35–39 The cost of neuroimaging should be included in the study budget so that all infants with mild HIE can be assessed by neuroimaging. The age of neurodevelopmental assessment should be 24 months (rather than 18 months) with continued assessment at 6–7 years of age.

An RCT of hypothermia for mild HIE will require a multicenter setting owing to the large sample size; 50% of infants with severe birth acidosis/resuscitation are either normal or have mild HIE at less than 6 hours of age.^17,25 The primary outcome should be the Bayley III composite cognitive score at 2 years of age, with motor, language, and behavioral scores as secondary outcomes. A clinically meaningful effect size would be a 0.33 difference (5 points) on the cognitive score between the 2 intervention groups (normative value is 100 ± 15). This would require a total of 430 infants to be enrolled, with 90% power, a survival rate of 98%, and a follow-up rate of 90%. Infants should be stratified by clinical center into hypothermia and normothermia groups.

The impact of a favorable outcome in terms of decreasing the number of children with neurodevelopmental impairments after mild HIE is difficult to assess because prospectively collected data on outcome of all infants with mild HIE diagnosed within 6 hours of age is limited. In the PRIME study of infants with mild HIE, the follow-up rate for a Bayley III examination performed at 18–22 months of age was only 68% among infants with mild HIE enrolled at less than 6 hours of age. Among the 43 infants evaluated, a Bayley III score of less than 85 in either the cognitive, motor, or language domain was noted among 40% of infants. Any disability defined as a cognitive score of less than 85 with or without a gross motor function level of 1 or greater was noted among 7 infants (16%), including 1 infant with cerebral palsy and 1 with autism.¹⁸ In the systematic review of mild HIE and outcome, among the infants with mild HIE who were inadvertently included in RCTs of hypothermia for moderate or severe HIE (n = 91; 45 cooled and 46 not cooled), an adverse outcome was noted among 29% of the cooled and 37% of the infants who were not cooled.⁴ In the second meta-analysis, evaluating 117 infants in five cooling trials, adverse outcome occured in 11 of 56 (19.6%) of cooled and 12 of 61 (19.7%) of infants who were not cooled.¹⁶ It is difficult to estimate what impact cooling may have on outcomes of infants with mild HIE; we suggest that performing cognitive, neurologic, and behavioral evaluations in infancy and childhood may provide an answer.

The challenges of performing an RCT for mild HIE are many. There is the possibility of an increase in length of hospital stay; however, it should be noted that the median (IQR) length of stay for infants with 1 or more abnormal neurologic category of mild HIE is 5 (IQR, 3–9) days.¹⁷ The costs of the entire duration of hospital stay are unlikely to be covered by a funding agency. However, no information is available regarding how hospital costs are currently covered for infants with mild HIE who have been cooled as part of clinical care. Another challenge will be that some infants will need to be separated from their mothers for transfer to a center participating in the RCT. Last, the neurologic examination should be standardized across all the participating sites with training and certification of all physician examiners. The Table provides a format for the neurologic examination, adapted from the Sarnat evaluation.^17,25 Video documentation of the neurologic examination confirming the presence of mild encephalopathy will be important. Mild HIE is heterogeneous, so stratifying the infants based on the examination and evaluating associations with outcomes will be an important aspect of the trial.

Table.

Modified Sarnat criteria for designation of HIE stage

Categories	Normal	Mild	Moderate	Severe
1: Level of consciousness	Alert, responsive to external stimuli	Hyperalert, stare, jitteriness, high pitched cry, exaggerated response to minimal stimuli, inconsolable	Lethargic	Stupor, coma
2: Spontaneous activity*	Normal	Decreased, with or without periods of excessive activity	Decreased	No activity
3: Posture	Predominantly flexed when quiet	Mild flexion of distal joints (fingers, wrist)	Strong distal flexion, complete extension	Intermittent decerebration
4: Tone	Strong flexor tone in all extremities	Slightly increased peripheral tone	Hypotonia or hypertonia	Flaccid, rigid
5: Primitive reflexes†
Suck*	Strong, easy to elicit	Weak, poor	Weak or has bite	Absent
Moro	Strong, easy to elicit	Low threshold to elicit	Incomplete	Absent
6: Autonomic nervous system†
Pupils	Normal size and reactive to light	Mydriasis	Miosis	Skew deviation/dilated and nonreactive/nonreactive
Heart rate	Normal heart rate	Tachycardia (>160)	Bradycardia(<100/min)	Variable heart rate
Respirations	Normal	Hyperventilation (>80/min)	Periodic breathing	Apnea or on ventilator with or without spontaneous respiration

^*If the category overlaps between mild and moderate as in spontaneous activity and suck, that category will be assigned the same severity of encephalopathy as the infant’s level of consciousness.

^†Primitive Reflexes has 2 signs (suck and Moro) and autonomic nervous system has 3 signs (pupils, heart rate, respiration); if more than 1 sign in either category is coded as abnormal, choose the most severe encephalopathy grade among the signs within that category. Suggest the entry criteria include 2 abnormal categories in the examination, but not meeting current criteria for cooling (3 abnormal categories in moderate or severe HIE stage).

An RCT will provide the needed answers; if the RCT demonstrates that cooling infants with mild HIE is neuroprotective, cooling can be offered to all infants with mild HIE. If an RCT does not show safety and efficacy, this practice of therapeutic creep of offering cooling for mild HIE can be discontinued. The ongoing challenge for researchers is the development and assessment of other neuroprotective therapies for HIE with early biomarkers of outcome.

Glossary

HIE

Hypoxic-ischemic encephalopathy

NICHD

National Institute of Child Health and Human Development

PRIME

Prospective Research for Infants with Mild Encephalopathy

RCT

Randomized controlled trial