Hypothermia therapy for neonatal encephalopathy

Brain hypothermia, induced by cooling a baby to around 33°C for three days after birth, is a treatment for birth asphyxia. It has recently been proven to be the only medical intervention which reduces brain damage, and improves an infant's chance of normal survival. Birth asphyxia, caused by a reduction in the supply of blood or oxygen to a baby's brain during birth, is a major cause of death and brain damage, occurring in approximately 1 per 1000 births and causing around 20% of all cases of cerebral palsy.

Hypothermia for resuscitation
Many physicians over the centuries have tried to resuscitate babies after birth by altering their body temperatures, essentially aiming to animate the infant by inducing the onset of breathing. Little thought was given to brain protection, because cerebral hypoxia during birth was not linked with later neurological problems until William John Little in 1861, and even then this was controversial; Sigmund Freud, for example, famously disagreed, and when scientific studies of neonatal therapeutic hypothermia were begun in the 1950s researchers like Bjorn Westin still reported their work in terms of re-animation rather than neuroprotection. Investigators such as James Miller and Clement Smith carried out clinical observations and careful physiological experiments,   but although some babies were conscientiously followed up, they were not mainly concerned with long term neurological outcome.

However, by the 1960s physicians saw hypothermia after delivery was something to be avoided. The problem of infants who failed to breathe at birth had been solved by the invention of mechanical ventilation, so any benefit cooling might have for re-animation was no longer needed, and an influential trial showed that keeping small and preterm infants warm increased survival. These results, together with observational and experimental data made it an article of medical faith for decades that babies should not be allowed to get cold.

Consequently during the next two decades studies of neonatal hypothermia in Europe and the USA were sporadic and often unsuccessful. An interest in cooling for brain protection was beginning to emerge, but contemporary neuroscience provided few useful concepts to guide this research and little progress was made. Although across the Iron Curtain in the Soviet Union cooling was being applied empirically following birth asphyxia, the language barrier, cold war politics and the Russians’ failure to carry out randomised controlled trials contributed to an almost total ignorance of this work in the West. Indeed a group of Russian neonatologists who described hypothermic neural rescue during a visit to the Neonatal Unit in Bristol, UK, met with little interest

Neural rescue
In the late 1980s the development of a new set of concepts and problems led to a re-examination. A new generation of neonatal researchers were influenced by the growing evidence that protecting the brain against the effects of oxygen deprivation during labour might be possible. These researchers were aware that cooling produced powerful intra-ischaemic neuroprotection during cardiac surgery but a new concept of hypothermic post-insult neural rescue developed. This shift in thinking was possible because of at least three major new ideas that were developing at the same time: delayed post-ischaemic cell death; excitotoxicity; and apoptosis.

Delayed cell death
The first paradigm shift that affected neonatal researchers in particular was the idea that if a baby was resuscitated after cerebral hypoxia-ischaemia there was a period of time before brain cells started to die. Osmund Reynolds at University College London used the newly developed technique of Magnetic Resonance Spectroscopy (MRS) to show that the infant brain metabolism is normal in the hours after birth asphyxia and deteriorated only after a distinct delay. Robert Vannucci confirmed the effect with painstaking biochemistry, and delayed injury was also reported in neuropathological studies. Delayed brain injury (called ‘secondary energy failure’ by Reynolds) was a critical new idea. If brain cells remained normal for a time and the mechanism of the delayed death could be unravelled, it opened the possibility of therapeutic intervention in what had previously seemed an impossible situation.

Excitotoxicity
The new and transforming concept of excitotoxicity developed from the seminal experiments of John Olney and Brian Meldrum. They showed that at least some of the neural cell death caused by hypoxia-ischaemia is mediated by excess production of the excitatory neurotransmitter glutamate, and that pharmacological blockade of the N-methyl-D-aspartate receptor could provide good protection against hypoxic damage. Olney and Meldrum had shifted the paradigm, allowing researchers to think of hypoxic-ischaemic damage as a treatable disease.

Apoptosis
However, it was still a mystery how and why cells triggered by hypoxia-ischaemia should die hours or days later, particularly when it became clear that glutamate levels were not particularly high during secondary energy failure. The next critical idea came with the discovery of programmed cell death, a novel form of cell suicide. Originally observed as a pathological appearance and named apoptosis ("falling off", as of leaves) in the 1970s, Horvitz, Raff and Evan provided a molecular understanding and showed that apoptosis could be triggered by cellular insults. The radical idea that hypoxia-ischaemia triggered a cell suicide programme which could explain the perplexing phenomenon of delayed cell death was soon supported by experimental and human data, and many researchers believe this helps explain why neural rescue works in the newborn. However the picture is complex: both apoptosis and necrosis are present in variable proportions; and there seems to be prolonged neurodegeneration after an insult. Research into this problem continues.

Neonatal neural rescue
These ideas flowed through the perinatal research community, producing a new belief that neural rescue after birth asphyxia should be possible. Amongst the first to have attempt neonatal neural rescue in animals were Ingmar Kjellmer and Henrik Hagberg in Gothenburg, and Michael Johnston in Baltimore. The potential began to draw in other neonatal researchers from diverse fields to begin neuroprotection research, including those who came to form the informal neonatal hypothermia research group:

Peter Gluckman and Tania Gunn were endocrinologists in the University of Auckland New Zealand and interested in cooling for its effect on thyroid function; they had first cooled a sheep fetus for endocrine studies in 1983. Denis Azzopardi, John Wyatt and David Edwards, then young researchers working for Reynolds, were using Reynolds’s sophisticated MRS approach to replicate secondary energy failure in newborn piglets and immature rats; in Gluckman’s laboratory Alistair Gunn and Chris Williams developed a simple and elegant biophysical method using cerebral impedance to do essentially the same thing in fetal sheep. Marianne Thoresen, who was working on cerebral perfusion, was prompted to think about neuroprotection by stories of children who fell through the Norwegian ice and suffering prolonged drowning in iced water but emerged with preserved cerebral function.

There were many potential therapies around which might achieve neural rescue, and most of these workers did not immediately move to hypothermia. Magnesium was an appealingly simple excitoxin receptor antagonist that protected cells in culture: the Reynolds group tested it in their piglet model without success. Gluckman and Gunn started by looking unsuccessfully at flunarizine, a calcium entry inhibitor. Edwards picked on nitric oxide synthase inhibition which was also a failure. Gluckman had success with his innovative studies of IGF-1, but could not immediately translate this to clinical practice.

Experimental neonatal hypothermia
Most neonatal researchers recognise work published in 1989 from Myron Ginsberg’s group as starting point of their interest in post-insult hypothermia. Ginsberg showed that a short period of hypothermia after hypoxia-ischaemia in adult rats produced significant protection in the hippocampus. Soon Pusanelli’s group suggested that at least some of the strong neuroprotective effect of the canonical glutamate antagonist MK-801 was by reducing body temperature. Before long there was a significant body of experimental work on post-ischaemic neuroprotection by hypothermia in mature animals. These results also stimulated clinical scientists working with adults, who had apparent early success in clinical trials.

The informal neonatal hypothermia interest group was developing: Gluckman visited London; Reynolds and Wyatt went to Oslo; Edwards went to Auckland; and Thoresen contacted Westin, then went to work in the Reynolds laboratory. Over the next year or so together and separately they produced a series of reports in piglets,   immature rats   and fetal sheep which showed repeatedly that post-insult hypothermia significantly reduced hypoxic-ischaemic brain damage in the developing brain. Trying to understand how and why cooling might work, they showed that it specifically reduced apoptosis, and interrupted the excitiotoxic cascade.

Thoresen recalls the immediacy of the first hypothermia experiment in the Reynolds laboratory. The room was full and she watched with Reynolds, Wyatt, Edwards and others as the biomarkers remained stubbornly normal after a very severe hypoxic insult which would normally have caused catastrophic secondary energy failure. In New Zealand there were similar experiences: Gunn says that he realised that hypothermia was going to work during his third experiment, as he watched the expected delayed injury fail to materialise.

Early clinical studies
As the experimental data continued to accumulate, clinical pilot studies were already being organised, although with some trepidation because of the prevailing view that cold was very dangerous for infants. There was some controversy over the relative benefits of selective head against whole body cooling. Tania and Alistair Gunn, first out of the clinical blocks, set out in Auckland to study local head cooling which they argued would have fewer side effects; Edwards and Azzopardi, now based at Hammersmith Hospital, delayed clinical studies until they had used a computer model to decide between the two then started whole body cooling; Thoresen and Andrew Whitelaw, relocated from Oslo to Bristol tried both methods.

Cooling was now a topic of wider discussion in the neonatal community and other groups started to organise further experimental and further preliminary clinical studies,  (reviewed in Jacobs et al. ).

However, not everyone was convinced. The highly respected Vannucci laboratory had failed to find any protective effect of post-insult cooling, and worse, Ginsberg’s group reported that hypothermic protection in mature rats was only temporary. Many clinicians thought that the experiments models were over simplistic and unrepresentitive of the complex clinical situation. Others thought that brain damage might have occurred weeks or months earlier; although an MRI study from Hammersmith dispelled this myth

The CoolCap study
Nevertheless, cooling captured the imagination of one of the most influential figures in Neonatal Medicine. Jerry Lucey, the editor of the top-rated pediatric journal Pediatrics, had an extraordinary ability to spot new ideas, and he became a strong champion of cooling. He promoted hypothermia tirelessly, and in early 1997 made a critical introduction between Olympic Medical, a medium sized equipment company, and the cooling fraternity. In 1997, in his car on the way to Dulles International Airport after a series of meetings largely organised by Lucey, Jay Jones the owner of Olympic Medical decided to fund a randomised controlled trial of hypothermic neural rescue therapy in newborn infants. Olympic would construct a head cooling device, the Olympic Cool-Cap System, and provide practical and financial support for the trial. Gluckman and Wyatt would be principal investigators and the scientific committee which developed and ran the trial joined hypothermia researchers like Edwards and Whitelaw, with new experts Donna Ferriero, Richard Polin, Roberta Ballard and Charlene Robertson. The newcomers were not all convinced baby-coolers; Ferriero, also a distinguished neuroprotection experimentalist, was particularly sceptical; but they brought a range of new skills essential to taking hypothermia into a major randomised trial. Arguably the most important roles in delivering the trial were taken by Ted Weiler, an indefatigable engineer from Olympic Medical, and Gunn who became the Scientific Officer.

The CoolCap study studied cooling for 72 hours started within 6 hours of delivery. The protocol was based on the Auckland pilot studies, but was somewhat arbitrary; the relative merits of different temperatures or lengths of cooling were unclear, and the researchers used best estimates based on their animal experiments, cautiously choosing the upper band of the expected therapeutic range. No-one knew with certainty what effect size to predicit or even what the best primary outcome was. On the divisive question of selective head versus whole body cooling, in the almost total absence of useful data they compromised and used rectal temperature to control a head cooling device. CoolCap thus studied the effect of cooling the whole baby to 34.5oC with the expectation (some said hope) that the brain might be a degree or so cooler.

The first infant was enrolled at Columbia Presbyterian Medical Center, New York, on July 1, 1999 and all data were collected, audited, and prepared for analysis by fall 2003. The result was reported at Lucey's annual Washington, D.C. 'Hot Topics in Neonatology' meeting in December 2003 with full publication in The Lancet in 2005. CoolCap showed a non-significant trend to improvement with cooling in the primary outcome of death or disability at 18 months overall, but a clear and significant benefit when infants with very severe or long established brain injury were excluded. This was not quite the definitive result they hoped for, but the researchers remained resolute; Gluckman confided to Edwards as they relaxed together after the data presentation in Washington, that he thought they would never do more important work.

The National Institute of Child Health and Human Development Study
After the CoolCap trial another major study was published by the National Institute for Child Health and Human Development (NICHD) Neonatal Research Network. The CoolCap researchers had previously had lengthy discussions with the network about collaboration because although the network had no experience of hypothermia they had a strong track record of patient recruitment. However after detailed discussions around the final CoolCap protocol broke down, the Network began a separate trial led by Seetha Shankaran. Some network members were setting up animal models of hypothermic neural rescue, providing valuable experience for preliminary experiments. The NICHD trial found a significant effect of cooling, although the study was criticised in some quarters for temperature instability in the control group.

The TOBY Trial
The question which now faced the community was: should cooling become standard of care? Opinions were divided. Some observers thought the evidence was sufficient, particularly if the growing number of pilot or smaller studies were considered and in December 2006 the U.S. Food and Drug Administration (FDA) approved the Olympic Cool-Cap System for clinical use. However others pointed out that the results were at best statistically marginal and urged scepticism. Worryingly, hypothermia trials for head trauma in adults and older children were failing to show benefit. A review meeting organised by NICHD advised that cooling was an emerging therapy but not standard of care; the community awaited more data.

After a couple of years during which even the original researchers found it difficult always to agree, the last major trial that had grown out of the original experimentalist group was completed. The TOtal BodY hypothermia for pernatal asphyxia trial (TOBY), led by Azzopardi backed by distinguished trials specialist Peter Brocklehurst and the UK National Perinatal Epidemiology Unit, developed from the Hammersmith pilot study of whole body cooling, but collected the data needed to allow meta-analysis with both CoolCap and the NICHD studies. TOBY had the 'advantage' of being delayed in starting by the funding timetable of the UK Medical Research Council, so that it was still in progress when CoolCap and the NICHD trial reported marginal benefits, making it clear that the trial size should be increased. When TOBY reported it was considerably larger than either previous study and yet showed remarkable consistency with both those trials; the point estimate for effect was similar and meta-analysis (detailed below) showed unequivocally that cooling increases an infants chance of surviving without neurological deficits at 18 months and reduces neurodevelopmental impairment in survival.

The ICE Trial
A new generation of trial data on hypothermia is now becoming available, adding to the information collected in the first three major studies. The ICE trial was planned to determine the effectiveness and safety of moderate whole-body hypothermia in newborns with hypoxic-ischemic encephalopathy born in hospitals with and without newborn intensive care facilities or complicated hypothermia equipment, and used cooled gel packs as part of the cooling regime. Therapeutic hypothermia reduced the risk of death or major sensorineural disability at 2 years of age: 55 of 107 infants (51.4%) in the hypothermia group and 67 of 101 infants (66.3%) in the control group died or had a major sensorineural disability at 2 years (risk ratio, 0.77 [95% confidence interval, 0.62-0.98]; P = .03).

The memories of some of the researchers involved in bringing hypothermic neural rescue into clinical practice have recently been collated.

Mechanisms of hypothermic neuroprotection after birth asphyxia
Much of what is known about the mechanisms of hypothermic neuroprotection is gathered from studies in mature and adult models. What follows uses some of these data while trying to focus on the immature brain. More information, particularly on adults, is given in the article Therapeutic hypothermia.

Hypoxia-ischaemia
Cerebral hypoxia-ischaemia results in reduced cerebral oxidative metabolism, cerebral lactic acidosis and cell membrane ionic transport failure; if prolonged there is necrotic cell death. Although rapid recovery of cerebral energy metabolism occurs following successful resuscitation this is followed some hours later by a secondary fall in cerebral high energy phosphates accompanied by a rise in intracellular pH, and the characteristic cerebral biochemical disturbance at this stage is a lactic alkalosis. In neonates, the severity of this secondary impairment in cerebral metabolism are associated with abnormal subsequent neurodevelopmental outcome and reduced head growth.

Several adverse biological events contribute to this secondary deterioration, including: release of excitatory amino acids which activate N-methyl-D-aspartate (NMDA) and amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptors on neurons (30,37) and oligodendroglial precursors, accumulation of excitatory neurotransmitters, generation of reactive oxygen radicals, intracellular calcium accumulation and mitochondrial dysfunction. Whilst necrotic cell death is prominent in the immediate and acute phases of severe cerebral insults, the predominant mode of death during the delayed phase of injury appears to be apoptosis. Neuroprotective mechanisms need to interact with these mechanisms to have beneficial effect.

Newborn hypoxic-ischaemic brain injury differs from injury in the adult brain in several ways: NMDA receptor toxicity is much higher in the immature brain. Apoptotic mechanisms including activation of caspases, translocation of apoptosis-inducing factor and cytochrome-c release are much greater in the immature than the adult. The inflammatory activation is different with less contribution from polymorphonuclear cells and a more prominent role of IL-18 whereas IL-1, which is critical in the adult brain, is less important. The anti-oxidant system is underdeveloped with reduced capacity to inactivate hydrogen peroxide.

Actions of hypothermia
Mild hypothermi helps prevent disruptions to cerebral metabolism both during and following cerebral insults. Hypothermia decreases the cerebral metabolic rate for glucose and oxygen and reduces the loss of high energy phosphates during hypoxia-ischaemia and during secondary cerebral energy failure, and reduces delayed cerebral lactic alkalosis The simultaneous increase in cytotoxic oedema and loss of cerebral cortical activity that accompanies secondary energy failure is also prevented.

Hypothermia appears to have multiple effects at a cellular level following cerebral injury. Hypothermia reduces vasogenic oedema, haemorrhage and neutrophil infiltration after trauma. The release of excitatory neurotransmitters is reduced, limiting intracellular calcium accumulation. Free radical production is lessened, which protects cells and cellular organelles from oxidative damage during reperfusion. In addition mild hypothermia may reduce the activation of the cytokine and coagulation cascades through increased activation of suppressor signalling pathways, and by inhibiting release of platelet activating factor.

Many of the effects induced by mild hypothermia may help to reduce the number of cells undergoing apoptosis. Experimental and clinical studies indicate that the number of apoptotic neurons is reduced caspase activity is lessened and cytochrome c translocation is diminished by mild hypothermia,  and there may be an increase in expression of the anti-apoptotic protein BCl-2.

Effect of therapeutic hypothermia in clinical trials
Approximately 1000 patients have now been studied in randomised controlled trials of hypothermia, There is ongoing surveillance of hypothermia-treated patients through the TOBY registry in the United Kingdom, and also some data collection of neonatal encephalopathy data by the Vermont-Oxford database in the United States. The results of an extended follow-up of the CoolCap and NICHD trials until 6–7 years of age have been published and similar data for the TOBY trial is being accrued. It is possible therefore to reach some firm conclusions on the clinical value of hypothermia:

Meta-analysis of early childhood outcomes after hypothermia
A number of randomised controlled trials have now been carried out. Early synthesis of the available data by the Cochrane Collaboration suggested that there was a beneficial effect but did not provide unequivocal support for general application. However a meta-analysis of all studies published in late 2009 included new data, in particular the TOBY trial, and this provided conclusive evidence that cooling reduces the adverse effects of birth asphyxia.

This analysis found three trials, the CoolCap, NICHD and TOBY trials which measured neurological outcome at least 18 months after birth. These trials report a total of 767 infants. Taken together they showed a significant benefit of cooling with a significant reduction in death or disability at 18 months after birth: typical Risk Ratio, 0.81 (95% Confidence Intervals, 0.71, 0.93), P=0.002. Treatment with hypothermia was consistently associated with an increased rate of normal survival survival: typical Risk Ratio, 1.53 (95% Confidence Intervals, 1.22, 1.93), P <0.001. There was a significant reduction in cerebral palsy amongst infants treated with hypothermia compared with controls. The relative effects of selective head and whole body cooling seem indistinguishable.

11 trials reported mortality rates after cooling. Meta-analysis of these data showed that fewer infants treated with prolonged moderate hypothermia died: typical Risk Ratio 0.78 (95% Confidence Interval, 0.66, 0.93) P=0.004.

Extended follow-up of trial participants
Studies have been undertaken to determine the effects of hypothermia beyond early childhood. Participants in the CoolCap, NICHD and TOBY trials were entered into extended follow-up programmes. None of these programmes have sufficient power to make confident assessements of the long-term effect of hypothermia, however even these underpowered studies give important information on whether the therapeutic effects of cooling are sustained beyond the first 2 years after birth.

The most significant follow-up study published so far is the assessment of the NICHD trial participants at 6–7 years Of the 208 trial participants, primary outcome data were available for 190. Of the 97 children in the hypothermia group and the 93 children in the control group, death or an IQ score below 70 occurred in 46 (47%) and 58 (62%), respectively (P=0.06); death occurred in 27 (28%) and 41 (44%) (P=0.04); and death or severe disability occurred in 38 (41%) and 53 (60%) (P=0.03). The CoolCap study gathered data using the WeeFim questionnaire at 7–8 years of age, but only collected information on 62 (32 cooled; 30 standard care) of 135 surviving children who had had neurodevelopmental assessment at 18 months. Disability status at 18 months was strongly associated with WeeFIM ratings (P < 0.001) suggesting that the therapeutic effect persisted, but there was no significant effect of treatment (P = 0.83).

These results were not quite conclusive, as the effect in the NICHD trial appears to be on mortality rather than neurological function, but they gave considerable confidence that the therapeutic effects of hypothermia following birth asphyxia are sustained into later childhood, and the preliminary childhood outcomes of the Toby trial presented at the Hot Topics in Neonatology meeting in Washington DC in December 2013 appear to confirm persisting benefit of treatment on neurological function; these data await peer review and formal publication and if confirmed would provide the definitive data needed for evidence-based health policy decisions.

Current state of the evidence
Hypothermic neural rescue therapy is an evidence-based clinical treatment which increases a severely injured full term infant's chance of surviving without brain damage detectable at 18 months by about 50%, an effect which seems to be sustained into later childhood.

At present data relate only to full term infants, and all human studies of hypothermia treatment have so far been restricted to infants >36 weeks out of an expected 40 weeks gestation. There are both more potential side effects on the developing premature with lung disease, and there is more evident protection by hypothermia when a greater volume of complex brain is actively developing. During mid gestation to late term the fetal brain is undergoing increasingly complex progressive growth of first the mid-brain and then development of the cortex and "higher" centers. The effects of fetal asphyxia on the developing brain in sheep are dependent on gestational age with near term fetuses showing both less tolerance of asphyxia and maximal damage in the rapidly expanding cortex; while fetuses prior to the last third of development experience more extended tolerance of asphyxia with maximal effects on the growing mid-brain. The fetal sheep asphyxia model also suggests a six-hour window post asphyxia in which hypothermia will have greatest benefit.

There remains much that is unknown. Recognition of infants with marginal external signs of asphyctic damage at birth, who still develop moderate Hypoxic Ischemic Encephalopathy would be enhanced by finding more reliable bio-markers or physiologic tests accurately predicting the risk for progressive damage. These tests could also prevent unwarranted, expensive treatment of many infants. Long-term follow-up has yet to demonstrate show persisting benefit, but available data together with an imaging study nested in TOBY also found reduced brain tissue damage in cooled infants are encouraging.

The simplicity that attracted empyricists to cooling centuries ago now makes hypothermic neural rescue with accurate patient selection a potentially transforming therapy for low-resource environments where birth asphyxia remains a major cause of death and disability. Ironically this brings back the problem of cooling infants in an environment where modern resuscitation and intensive care are not available.