Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Nearly half of all the global under-five deaths take place during the neonatal period and 99% of them are in low-income and middle-income countries (LMICs). Preterm birth complications and neonatal encephalopathy account for majority of these deaths. Two recent multi-country randomised clinical trials to evaluate interventions to reduce these deaths have major implications on neonatal care in LMICs.
First, a WHO study investigated the impact of immediate kangaroo mother care (iKMC) on survival of low birth weight neonates weighing 1000–1799 g.1 In the intervention arm, neonatal intensive care units (NICUs) were redesigned for a family-centred care allowing treatment of neonates in zero separation from mother or a surrogate during the NICU stay. In the control arm, neonates received care as usual in an incubator with mothers giving brief KMC sessions at earliest 24 hours from birth. Neonates from both groups were moved to a usual KMC ward once deemed sufficiently stable for at least 24 hours. Median duration of the NICU stay was 6.4 days in both groups and during this time median daily skin-to-skin contact in the iKMC group was 16.9 hours and in the usual care group 1.5 hours. Neonates in the iKMC group had 25% reduction in the risk of 28-day (neonatal death) compared with the control arm (12.0% vs 15.7%, risk ratio (RR) 0.75; 95% CI 0.64 to 0.89).
Cochrane’s systematic review has shown that low birth weight neonates cared with KMC after stabilisation of some days to weeks have better survival and lower rates of sepsis and hypothermia compared with usual care in incubators.2 WHO iKMC trial found that even lower rate of suspected sepsis (RR 0.82; 95% CI 0.73 to 0.93) and episodes of hypothermia (RR 0.65; 95% CI 0.51 to 0.83) can be achieved with iKMC compared with the usual care.
The second trial, HELIX,3 investigated the effect of therapeutic hypothermia in term or near-term post-asphyxiated neonates with neonatal encephalopathy comparing it with usual NICU care with normothermia. The study showed that cooling was potentially dangerous with 1.35 times higher risk of death during the first 18 months of life in the intervention group compared with usual care group (RR 1.35; 95% CI 1.04 to 1.76) and found that cooling was inefficient for neuroprotection (RR 1.06; 95% CI 0.87 to 1.30; p=0.55 for the main outcome of death or moderate to severe disability at 18–22 months).
The HELIX study aimed to provide definitive evidence on a technology-based intervention in LMIC setting with highest burden of disease. Therapeutic hypothermia was provided using state of the art servo-controlled whole body cooling device with excellent safety record in high-income trials. All participating centres had consultants with previous experience in cooling and protocol violations were exceptionally few. Only major difference from high-income practice was not requiring a blood sample to confirm acidosis. However, cord blood samples in those 10% of neonates that did undergo testing indicated severe acidosis and all neonates were moderately or severely encephalopathic in a clinical examination performed by certified evaluators. Most neonates underwent MRI during hospitalisation and the follow-up using Bayley III test achieved 97% coverage. Cooled neonates fared worse across all sites and in several secondary outcomes. Importantly, they also had no better disability-free survival than the usual care group.
The HELIX results are likely to be contested as cooling has become the standard of care in many LMIC NICUs with majority of previous randomised controlled trials and observational studies showing good safety record.4 Even the International Liaison Committee on Resuscitation guidelines 2020 had cautiously recommend cooling in LMICs.5 However, few of the previous trials were powered to detect differences in death and neurodevelopmental disability.4 Nor were they conducted with the same rigour as the HELIX trial.
The reported results of the HELIX study give little information as to why cooling did not work as expected. The two groups were similar at baseline except somewhat higher rate of sentinel events and lower rate of caesarean sections in the control group. The proportion of infants with culture-confirmed sepsis was no higher than in the high-income trials, whereby the mechanism of inflammation is unlikely to explain the observed lack of neuroprotection.
Sentinel events were quite rare overall, most neonates had seizures within 6 hours of birth and white matter injury was much more prevalent than basal ganglia injury on MRI. All these findings suggest a prolonged hypoxia that could have been compounded by a poor intrapartum care during difficult deliveries. As therapeutic hypothermia is based on slowing down brain metabolism during latent phase 6 to 72 hours after a hypoxic-ischaemic event, it will be less effective for subacute injuries.
More than two-thirds of the neonates were born outside the cooling centres. Cooling was started at median age of 4.3 hours and while time to achieving target temperature was not reported, in figure 2B it appears that this took place at around 5 to 6 hours from randomisation and thus beyond the optimal window of opportunity. Lastly, it is also possible that supportive care provided in these public hospitals was not up to the standard required for safe management of hypothermia.
Further research is needed to clarify any population-level differences in cooling outcomes and centres already practicing therapeutic hypothermia should evaluate their practices. Prevention of intrapartum-related neonatal encephalopathy by improving delivery care and optimising neonatal resuscitation are currently the best ways of helping term and near-term neonates at risk of brain injury in LMICs. Rigorous large-scale trials in the vein of HELIX are also needed to assess up and coming neuroprotective interventions. Identifying suitable biomarkers in LMIC to assess the appropriateness of treatment is also needed, especially for neonates born outside the cooling centres.
The promising results of the iKMC trial now suggest a three-stepped approach for its implementation in LMIC settings. First, understanding and benchmarking the required changes for implementation of iKMC from birth until discharge within the hospital. The assessment includes the current healthcare providers’ practice on zero separation at birth, current NICU designs and human resources to support the NICU and continuity of care following discharge. Second, reviewing the current recommendation for small and sick newborn care and developing new standards for care. Third, gradual implementation of the standards in hospitals and documenting good implementation practice for change. Most of the hospitals in LMIC settings require additional resource allocation to redesign the current NICU set up for accommodating mothers or parents. Therefore, a gradual benchmarking of implementation of iKMC will be required. This will further guide country and hospital managers on the required set up aimed at. A continuous quality improvement process will be key to seamlessly implement the iKMC intervention.
Almost two-thirds of all preventable neonatal deaths in LMIC are accounted by preterm birth complications and neonatal encephalopathy, so scale-up of effective interventions is needed to bend the curve of neonatal mortality. Both iKMC and therapeutic hypothermia are established interventions in high-income settings, but the results of these two recent trials suggest that only iKMC can be endorsed in LMIC. The HELIX results might be specific to the settings in which it was undertaken and may therefore not be generalisable to all LMICs but do call for further debate on management of neonatal encephalopathy.
Patient consent for publication
Contributors AK and AKC conceptualised the write up. AK and AKC wrote the first draft together and revised it. Both of them agree to the final version.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Commissioned; externally peer reviewed.