Therapeutic Hypothermia Effects on Brain Development

Neonatal encephalopathy (NE) is a clinical syndrome of disturbed neurologic function in the earliest days of life in an infant born at or beyond 35 weeks of gestation. It is manifested by a subnormal level of consciousness or seizures and often accompanied by difficulty initiating and maintaining respiration and depression of tone and reflexes. The incidence of NE is 3 in 1,000 live births.^1,2 Although there are many potential causes of NE, the most common is a hypoxic ischemic event, resulting in hypoxic ischemic encephalopathy (HIE), which has an incidence of 1.5 in 1,000 live births. The presence of NE with accompanying clinical and/or laboratory evidence of acute or subacute asphyxia leading to decreased cerebral blood flow and subsequent injury defines HIE.² The clinical or laboratory evidence for HIE includes:

1. hypoxic or ischemic event (ie, uterine rupture, placental abruption, umbilical cord prolapse, or chronic placental insufficiency);

2. low Apgar scores;

3. acidemia (decreased arterial cord pH <7 or increased base deficit ≥12); and

4. multisystem organ failure.²

Based on a clinical encephalopathy score from a seminal paper³ categorizing NE by features of the neurologic examination and early EEG findings (Table 1), HIE is classified as mild (stage 1), moderate (stage 2), or severe (stage 3). Neonatal HIE has historically resulted in significant morbidity and mortality, particularly in those with moderate and severe encephalopathy. The 21st century, however, has seen a revolution in the management of HIE with the advent of therapeutic hypothermia (TH)—the only proven neuroprotective therapy to improve outcomes for affected neonates.

Pathophysiology

The pathophysiology of HIE evolves over hours to days and is marked by 3 phases of injury: 1) acute or primary, 2) latent, and 3) secondary (Figure). The acute phase is defined by primary energy failure leading to an excitotoxic-oxidative cascade occurring over minutes.⁴

Decreased cerebral blood flow and resultant decreased oxygenation leads to cerebral anaerobic metabolism. This metabolic shift induces Na⁺/K⁺ pump failure causing reduced ATP production, increased lactic acid, and an influx of intracellular calcium and neurotransmitters. Energy failure triggers neuronal depolarization that releases glutamate and reactive oxygen species. These act as excitotoxins disrupting cellular integrity and mitochondrial respiration to cause neuronal cell death.

The latent phase is marked by reperfusion and partial recovery of oxidative metabolism 1 to 6 hours after injury, although the length of the latent period varies by the severity of the initial hypoxic-ischemic insult and is shorter in more severe injuries. There is also ongoing apoptotic activity from proinflammatory signals released during the initial excitotoxic-oxidative cascade from the acute injury.

The latent phase is followed by the secondary injury phase, which starts 6 to 15 hours after the initial injury and is marked by delayed energy failure from the recurrent failure of mitochondrial metabolism, cytotoxic edema, and ongoing inflammation.⁵ The latent phase is considered the optimal window for intervention to mitigate additional neuronal injury and prevent progression to secondary injury.

Pathophysiologic Effects of Hypothermia

Systematic reduction of cerebral temperature with TH affects multiple phases of the hypoxic-ischemic injury. For every 1 °C decrease in body temperature, cerebral metabolic rate decreases by 6% to 7%, reducing cerebral energy demands.⁶ Additionally, TH modulates the primary phase of injury by decreasing the release of excitatory amino acids and the synthesis of nitrous oxide to reduce excitotoxicity.⁶ In the latent and secondary phases, TH blunts the release of reactive oxygen species triggered by reperfusion, suppresses inflammation, decreases expression of proapoptotic mediators, and decreases cytotoxic edema by preserving the blood-brain barrier.⁶

History of Hypothermia in Clinical Use

Although there are reports of therapeutic cooling dating back to ancient times, modern clinical use was not established until the early 21st century. Improved survival and neurologic outcomes with TH were seen in adults who had an out-of-hospital cardiac arrest in 2 clinical trials in 2002.^7,8 Between 2005 and 2011, there were 6 major clinical trials evaluating TH for NE (Table 2).^9-14 Despite some methodologic differences, all studies included term (≥37 weeks gestation) and late preterm (≥35 weeks gestation) infants with moderate-to-severe HIE who had selective-head or whole-body cooling within 6 hours of birth to target temperatures of 33.5 °C to 34.5 °C for 72 hours. These trials demonstrated efficacy (defined as reduced mortality and neurodevelopmental morbidity) of TH for moderate and severe neonatal HIE.

Although TH is generally well-tolerated, it is not without side effects. Medical complications include sinus bradycardia, increased risk of sepsis and pneumonia, thrombocytopenia or other coagulopathies, need for additional interventions (eg, central venous access), and need for increased sedation owing to shivering.¹⁵ There are additional considerations such as delay to independent feeding and a decrease in early parental bonding. Therefore, it is important to appropriately assess and select infants who will most benefit from TH, making any potential risks worth the presumed benefit.

Outcomes Without Therapeutic Hypothermia

Before the routine use of TH, mortality or major neurologic disability (defined as cerebral palsy [CP] or motor or cognitive impairments >2 standard deviations [SD] below normative values) in neonates with HIE was 47% (95% CI, 36%-57%).¹⁶ Outcomes varied with HIE severity, with 94% to 100% of neonates with severe HIE and 24% to 67% of those with moderate HIE experiencing poor outcomes. Mild HIE outcomes had more variability. Early studies showed that neonates with mild HIE had similar rates of mortality and major neurologic disability compared with neonates without HIE when assessed at toddler age.^16-18 However, a more recent meta-analysis found mild-to-moderate neurologic disability in 25% of children with mild HIE at age 2 years that increased to 35% at age 5 years.¹⁹

Cognition

Although cognitive assessment measures in HIE survivors have varied across studies, there is a general trend towards worse cognitive outcomes when TH is not used. Recent outcomes data for uncooled infants primarily focus on those with mild HIE who do not meet current criteria for TH. In a prospective multicenter trial, neonates with mild HIE were assessed with the Bayley Scales of Infant and Toddler Development (BSID), a standardized neurodevelopment assessment tool with specific composite scores in cognitive, language, and motor domains at 18 to 24 months age.²⁰ There was a 16% rate of disability overall, with 40% of infants having any Bayley subset score >1 SD below the norm, most often in the language domain. A persistently abnormal exam at discharge, abnormal brain MRI, and prolonged hospitalization were associated with the presence of cognitive deficits. A pooled prospective cohort study of 55 infants with mild HIE who did not receive TH demonstrated that affected neonates had significantly lower composite cognitive scores on BSID at age 2 years compared with healthy controls. Their scores were more similar to those of infants with moderate HIE who were cooled, furthering evidence that untreated mild HIE affects development.²¹ Later cognitive outcomes show a widening gap in outcomes for neonates who had HIE but were not treated with TH, and school-age assessments may better predict difficulties and disabilities.²² A study of children age 8 years with a history of untreated HIE had mean IQ scores 10 points lower than children who had not had HIE, although both were within normal range.²³ Overall, studies have consistently demonstrated that school-age children with a history of untreated neonatal HIE are more likely to be at least 1 grade level behind in reading, spelling, and math, receive special education services, require extra academic support, and have significantly lower full scale, verbal, and performance IQs compared with age-matched children who did not have HIE.^23-25 These differences persist in the absence of CP and motor deficits.²⁶ Cognitive deficits persist with age. A Swedish population-based study of children age 15 to 19 years with a history of moderate HIE found that 81% had cognitive deficits, two-thirds of whom did not have concurrent CP.²⁷

Neurobehavior

Executive function deficits such as inattention and impulsivity are generally reported to be worse in patients with HIE compared with healthy controls, though data is mixed and studies are flawed by methodologic limitations and poor patient follow-up.^24,27-29 Poor behavioral assessments at school-age have been reported in children with a history of mild and moderate HIE, though this difference was only significant in those with moderate HIE.³⁰ This finding has not been replicated in other studies.³¹ Another study demonstrated that despite normal overall mean scores on scales of attention and executive function, children with HIE have a greater-than-expected proportion of scores that fall into the impaired range when compared with standardized data from controls.²⁸ This was more marked in those who had moderate HIE than in those who had mild HIE. Additionally, there is some suggestion that rates of autism spectrum disorder (ASD) are higher in HIE survivors, but data is limited. A population-based study showed increased rates of autism at age 5 years of age for children who had NE compared with children born at full term without NE. Those who had HIE were 5.9 times more likely to be diagnosed with ASD (relative risk [RR] 95% CI, 2.0-16.9), with a prevalence rate of 5% in HIE survivors compared with 0.8% in those without HIE.³²

Outcomes With Therapeutic Hypothermia

In the 6 major clinical trials described earlier in this article and in Table 2, selective head cooling (cooling targeted to the head while the body is kept normothermic) was used in 2 trials and whole-body cooling (temperature lowered for brain and body) was used in 4 trials.^9-14 The primary composite outcome was death or major disability at 18 to 24 months, and 4 studies demonstrated a neuroprotective benefit of TH, with an overall relative risk of 0.76 (95% CI, 0.69-0.84) in the hypothermia-treated group. The number needed to treat for the primary outcome was 7,³³ leading to TH becoming the standard of care for moderate-to-severe HIE. For 3 of the studies, follow-up data at age 6 to 8 years is available.^34-36

Motor

Motor outcomes improved in infants with HIE treated with TH compared with infants in the control group. In all 6 of the major neonatal TH trials, CP rates varied from 28% to 42% in the control group and 13% to 32% in the TH group at the time of primary outcome assessment (18-24 months).^9-14 Across all studies, there was a reduction in CP in infants treated with TH, with an RR of 0.66 (95% CI, 0.54-0.82) for CP in the hypothermia group at 18-24 months, seen both in selective head cooling and whole-body cooling trials.¹⁵

At later time points, lower rates of CP were maintained in the TH group. In 1 trial, follow-up at age 6 to 7 years demonstrated statistically significant differences in secondary outcomes of death or CP; CP was present in 17% vs 29% of infants treated with TH vs controls.³⁴ In another of the studies, follow-up at age 6 to 7 years showed that 21% vs 36% of those treated with TH vs control had CP.³⁵ More children in the TH group vs the control group had normal gross motor (78% vs 59%) and fine motor (77% vs 61%) function scores.

Cognition

Cognitive outcomes at initial follow up at 18 to 24 months using the BSID were reported for 3 of the trials.^34-36 No significant difference in cognitive outcomes was seen between treated and control groups; however, it should be noted that the initial TH trials were underpowered for secondary outcome assessment (eg, specific cognitive subdomains) because of deaths, children lost to follow-up, and the inclusion of survivors with neurologic impairments severe enough to preclude standardized cognitive testing. Despite these limitations, testing of school-age children with a history of HIE does show improved cognitive outcomes for those who received TH, further strengthening the argument for the long-term benefits of TH.

At age 6 to 7 years, follow-up assessment from 1 of the trials reported death or IQ<70 in 47% of children treated with TH vs 62% of the control group, although a significant difference was seen only between those with moderate vs severe HIE.³⁴ Full-scale IQ scores were <85 in 52% and <70 in 25% of patients in the TH group, compared with <85 in 53% and <70 in 32% of those in the control group.³⁷ Although the difference was not statistically significant, children who had HIE and were not treated with TH consistently performed lower on formal cognitive testing and had higher rates of special education service use.

In another of the trials, follow-up assessment at age 6 to 7 years also found significant cognitive differences between those who had TH vs those who did not. Of those who had TH, 52% had an IQ ≥85 compared with 39% of those who did not have TH (P=.05).³⁵ Additionally, 8% vs 27% of the trial participants who had TH vs control required special education (RR 0.3, 95% CI, 0.12-0.79, P=0.01). Although there were no significant differences in specific neuropsychologic domains (eg, working memory, long-term memory, and processing speed), children who had TH had significantly higher performance on attention and executive function tests (RR 1.02, 95% CI, 0.12-1.92, P=0.03).

For 1 of the studies, functional status at age 7 to 8 years was assessed with the survey-based WeeFIM instrument that measures self-care, mobility, and cognition.³⁶ There was no significant difference between children treated with TH vs control (P=.83), but positive or negative outcomes seen at age 18 months were still present. Severe disability persisted between toddler age and early school age, regardless of the treatment group.

Neurobehavioral

There are limited data published for behavior outcome measures after TH for HIE. A study of parental assessment of behavior demonstrated a mildly higher rate of behavioral problems in children with a history of HIE who did not have TH (9% normothermia vs 7% TH). Behavior problems, however, were found to be more closely related to maternal educational attainment, and no difference was found in parental ratings of attention or executive function.³⁷ Similarly, data from another of the studies at age 6 to 7 years found no difference in parentally assessed self-esteem for children who had HIE and were treated or not treated with TH.³⁶

Special Considerations

Although TH is approved for use in moderate-to-severe HIE, use in mild HIE is increasing such that approximately one-fifth of infants treated with TH had mild HIE only.³⁸ Few studies have assessed outcomes after TH in this group. A study comparing outcomes in infants with mild HIE showed no difference in overall BSID or cognitive, language, or motor domains at age 12 to 24 months for children with mild HIE who received TH compared with children without a history of HIE, suggesting normalization of developmental scores in the setting of TH. However, this has not been consistently replicated. A small number of children with mild HIE were unintentionally included in 1 of the 6 major TH clinical trials; no increase in mortality or major neurologic disability was observed, regardless of TH use.¹³

Prognostication and Follow-Up

Accurate prediction of outcomes for infants with HIE would lead to improved prognostication for clinicians and families, however, early prognostication remains difficult. Clinical variables from early in the infant’s course (eg, respiratory status, Apgar score, or Sarnat score) have limited prognostic usefulness. The addition of laboratory features (eg, cord pH or serum lactate) have yielded mixed results and are highly dependent on timing and persistence of the laboratory abnormality.

Early clinical data should guide clinical care and prognostic discussions with clear understanding and explanation of the limitations of these data for prognostication. Ancillary data such as MRI and EEG add prognostic value, but are not available acutely when many life-altering decisions are made. An MRI performed 3 to 10 days after insult predicts neurodevelopmental outcomes, but cooling apparatus and patient stability limit the ability to perform MRI earlier. The EEG background patterns have predictive value for mortality and neurodevelopment--normalization of the EEG background is associated with favorable outcomes, whereas the persistence of severely abnormal EEG backgrounds (eg, burst suppression, low voltage, or flat trace) is associated with unfavorable outcomes. However, EEG obtained later in the course (48-72 hours) is more predictive than early EEG. Use of TH can delay normalization of EEG background patterns for 24 to 48 hours, further delaying the prognostic utility of EEG. The search for an ideal biomarker for neonatal HIE is a research priority.

Neonates with HIE, particularly those with moderate and severe HIE, should be enrolled in follow-up programs after hospital discharge given the high risk of neurodevelopmental delays, epilepsy, visual impairment and hearing loss. Many hospitals have developed multidisciplinary follow up programs that provide access to pediatricians, developmental pediatricians, neurologists, psychologists, and physical, occupational, and speech therapists in a comprehensive visit for children and families. These visits are often scheduled at specific ages (such as 3, 6, 9, 12, 18, 24, 36, 48, and 60 months) and allow for standardized physical, neurologic, and cognitive assessments, which can help inform prognosis of various outcomes. These visits also provide the opportunity for ongoing communication, education, and support between families and medical providers who have often developed relationships during the acute hospitalization.

Conclusion

There are considerable neuroprotective effects of TH for neonates with hypoxic-ischemic injury, with improved outcomes demonstrated not just in the perinatal period but throughout early and middle childhood in motor, cognitive, behavioral, and neurologic domains. Despite impressive neuroprotective effects, questions remain on how to optimize TH for neonatal HIE. The incidences of death or major neurologic disability—although improved from the preTH era—remain unacceptably high at 40% to 55%.^39,40 Questions remain about the ability to maximize the neuroprotective effects of TH. For example, there is preliminary evidence demonstrating a reduction in mortality with initiation of cooling between 6 and 24 hours of life, potentially extending the time window for initiating TH.⁴¹ There is also preliminary data demonstrating benefits of TH in infants with a 34 to 35 week gestational age at a single center,⁴² and many centers are initiating hypothermia for mild HIE, although there are only limited and mixed outcome data. Although questions remain, TH remains a powerful tool for neuroprotection in neonatal HIE and has demonstrated long-reaching effects on all areas of neurologic development throughout childhood.