Research in wealthy countries has shown that cooling down babies who are deprived of oxygen during birth can prevent brain damage and other health problems. But results from a study focused on births in low and middle income countries (LMICs) now suggest that in many cases the treatment fails to prevent brain damage, and instead is linked with increased mortality.
The findings published on 3 August in The Lancet Global Health1 are surprising, says Joanne Davidson, a fetal physiologist at the University of Auckland in New Zealand, because researchers conducted the trials at seven hospitals with "decent facilities". "We would have expected in those circumstances that the treatment would have been better and certainly wouldn't be deleterious," she says.
Sudhin Thayyil, a perinatal neuroscientist at Imperial College London who led the study, says that cooling should be suspended immediately in LMICs until further research can pinpoint exactly why the treatment is associated with more deaths.
The cooling technique, known as therapeutic hypothermia, is used to treat babies born at or near full term who are deprived of oxygen during birth a life threatening condition that causes nearly one million deaths worldwide every year, the majority of which occur in LMICs. Previous research2 suggests that lowering these babies' body temperature can prevent their health from deteriorating, by stopping the death of oxygen deprived cells.
International guidelines3 recommended in 2015 that oxygen deprived babies in LMICs should be cooled under strict conditions, but acknowledged that the underlying evidence was weak. Nevertheless, cooling is widespread in countries including India, where about half of advanced neonatal intensive care units have adopted the practice, says Thayyil.
To test the treatment, the researchers randomly assigned 202 full term babies with suspected brain damage from Sri Lanka, India and Bangladesh to receive the cooling treatment for 3 days; 206 babies received standard treatment without cooling. Thayyil and his colleagues found that 42% of the infants who were cooled died within 18 months, compared with 31% of babies who did not receive the treatment.
By contrast, a 2009 study4 of 325 babies born in the United Kingdom found that both cooled and non cooled infants had a mortality rate of about 26% after 18 months, but the treatment significantly reduced the risks of disability and motor impairments in survivors.
Uncovering why the approach is more successful in some countries than in others will not be easy, says David Edwards, a neonatologist at Kings College London who helped to conduct some of the first large scale trials of the treatment and was a co author of the 2009 study. "We're seeing the age old problem that a medical intervention is part of a system," he says. "And the system includes obstetrics, the nutrition of the patients, the drugs available all those things."
One plausible explanation, says Edwards, is that even mild cases of fetal infection during pregnancy which are more common in LMICs than in wealthy countries, and can go undetected can affect babies' brains at or before birth. Brain scans showed that all of the babies in this trial had significantly more damage to the white matter in their brains than did those in trials conducted in high income countries, suggesting that they might have been deprived of oxygen for longer. "These babies are not presenting with the same kind of patterns of injury as in a first world setting, and we need to understand why that is," says Davidson.
Cooling demonstrates the perils of what Davidson terms "therapeutic creep": when a clinical practice becomes widespread without evidence that its benefits are universal. "Once something makes it into clinical practice, it's hard to then get it out of clinical practice again or to even go back and then do the proper randomized control trials," she says.
But Marianne Thoresen, a neonatal neuroscientist at the University of Bristol, UK, isn't convinced that the practice should be stopped in LMICs. She says that the higher mortality might not be caused directly by cooling; it might, in part, be related to level of care given to each baby. Nurses in the study cared for two to four babies at once, whereas it is customary for nurses in higher income countries to care for a single baby undergoing cooling treatment. Dedicated care could prevent deaths while preserving the benefits of cooling seen in the study, such as improved motor performance and fewer cases of disabling cerebral palsy. "These babies need very high quality intensive care, because when you cool somebody, a lot of the physiology is affected," she says.
Edwards says that there will now be a greater impetus to find a treatment that does not involve cooling, and to collaborate with researchers in LMICs.
"These results should not discourage us," says Samanmali Sumanasena, a paediatric disabilities specialist at the University of Kelaniya Sri Lanka and a co author of the study. "We have to go forward and look at what factors contribute towards the mortality and how we can minimize it."