Questions about the HELIX trial – Authors' reply

Abstract

We appreciate the clinician's dilemma when existing practice is challenged by evidence, and have provided a combined response.1Kozlov M Doubts raised about cooling treatment for oxygen-deprived newborns.Nature. 2021; (published online Aug 11.)DOI:10.1038/d41586-021-02201-6Crossref Google Scholar 95% of patients enrolled in the HELIX trial had multiple antenatal visits. Data on Apgar scores beyond 5 min and cord pH are rarely available in outborn infants from low-income and middle-income countries (LMICs). However, admission blood pH was available in 91% of infants in our study. Only two infants in the hypothermia group and one infant in the control group had temperatures lower than 33°C at the point of admission (data not shown). Amplitude-integrated electroencephalography (aEEG) within 6 h has poor prognostic accuracy,2Chandrasekaran M Chaban B Montaldo P Thayyil S Predictive value of amplitude-integrated EEG (aEEG) after rescue hypothermic neuroprotection for hypoxic ischemic encephalopathy: a meta-analysis.J Perinatol. 2017; 37: 684-689Crossref PubMed Scopus (54) Google Scholar hence we used a structured neurological examination by a certified examiner to grade encephalopathy. Increased leukocyte count is common in encephalopathy and is not a marker of sepsis. Levels of coexistent sepsis (affecting 5% of infants in our study) and funisitis (affecting 16% of infants in our study) were similar to those in high-income countries (HICs).3Bingham A Gundogan F Rand K Laptook AR Placental findings among newborns with hypoxic ischemic encephalopathy.J Perinatol. 2019; 39: 563-570Crossref PubMed Scopus (11) Google Scholar All infants in our study had continuous vital signs monitoring. In our study, hypothermia was ineffective at all sites, including the site in Dhaka that had a 1:1 nurse to infant ratio; both in outborn and inborn infants. Assessing outcomes according to the encephalopathy stage or time to randomisation was not the purpose of the HELIX study; to do so would be underpowered and inappropriate, as would be adjusting for multiple baseline variables. None of these characteristics significantly differed between groups. The trial data were analysed five times by the Independent Data Monitoring Committee, who were masked to the allocation. Moreover, at each review, the committee unanimously voted for trial continuation. The conclusion of our study, that LMICs should immediately suspend use of cooling in neonates, was based on the primary outcome—the trial had 81% power to conclude that hypothermia does not reduce death or disability (50% in hypothermia vs 47% in the control group; p=0·55). Of the three predefined secondary outcomes, mortality at 18 months was higher in the hypothermic group, while severe disability and microcephaly were not significantly different. All infants who were discharged against medical advice subsequently died and were included in the 18-month mortality data. Only assessors not involved in the recruitment did the infant's Bayley assessment. Nevertheless, any outcome bias would have benefitted the hypothermia group. Mortality at 18 months was available in 399 (98%) infants; data for nine missing infants would not have affected the primary outcome even if they had all died, and the conclusion of the paper would remain unchanged. Cerebral palsy was an exploratory outcome and was different between the groups only for gross motor function classification systems (GMFCS) greater than or equal to 2, not when all infants with cerebral palsy were included, and the median GMFCS and Bayley scores were not different. Death and disability are competing outcomes; HELIX sites with the highest mortality had the lowest disability. Interventions (or lack of) that increased mortality in infants who were most sick, or who otherwise survived with disability, could prevent cerebral palsy, as shown by the low rates of prematurity-related cerebral palsy in Africa.4Kakooza-Mwesige A Andrews C Peterson S Wabwire Mangen F Eliasson AC Forssberg H Prevalence of cerebral palsy in Uganda: a population-based study.Lancet Glob Health. 2017; 5: e1275-e1282Summary Full Text Full Text PDF PubMed Scopus (63) Google Scholar To compare with other studies, the therapeutic hypothermia for neonatal hypoxic-ischaemic encephalopathy in India (THIN) trial recruited 50 infants, of whom only two had severe encephalopathy. Moreover, the primary outcome (measured using a magnetic resonance [MR] biomarker) was only available for 22 infants. No inference about hypothermic neuroprotection can be made from such a small study. The higher proportion of infants who survived to MRI merely reflects milder levels of encephalopathy. In HELIX, of 408 recruited infants, 114 died, eight left against medical advice before MR scan, and the caregivers of two denied consent. Of the remaining 284 infants, 267 (94%) had available MR data. Hypothermia did not reduce severity of brain injury, which was measured by validated qualitative and quantitative 3 Tesla MR biomarkers. The South African cooling study had no control arm. In addition, 32% of participants had mild encephalopathy, which might have led to the better outcomes than those reported in participants in the HELIX trial.5Kali GT Martinez-Biarge M Van Zyl J Smith J Rutherford M Therapeutic hypothermia for neonatal hypoxic-ischaemic encephalopathy had favourable outcomes at a referral hospital in a middle-income country.Acta Paediatr. 2016; 105: 806-815Crossref PubMed Scopus (15) Google Scholar Pooled data from all randomised controlled trials reported from LMICs suggest that hypothermia does not reduce death or disability at 18 months (risk ratio 0·92 [95% CI 0·77–1·10; appendix p 1). Although the clinical characteristics and early seizure onset of previous trials were similar to HELIX, all except one were pilot trials; many included mild encephalopathy, and six were from the same hospital with very high risk of bias (appendix pp 2–3). Meta-analyses and Cochrane reviews are only as good as the individual trials, hence a traffic light system to exclude poor quality (denoted by the colour red) trials has been developed recently to improve their credibility.6Moore A Red for danger in systematic reviews.Eur J Hosp Pharm. 2021; 28: 299-300Crossref PubMed Scopus (2) Google Scholar HELIX only involved hospitals from the academic public sector, as most private hospitals are the setting of a low proportion of deliveries (a high proportion of which are caesarean sections), are driven by commercial interests, offer cooling primarily to outborn infants and infants with mild encephalopathy using indigenous methods, and rely on observational data and personal experience, which departs from cooling protocols used in HICs.7Chandrasekaran M Swamy R Ramji S Shankaran S Thayyil S Therapeutic hypothermia for neonatal encephalopathy in Indian neonatal units: a survey of national practices.Indian Pediatr. 2017; 54: 969-970Crossref PubMed Scopus (17) Google Scholar Ten (5%) infants in the control group of our study had hyperthermia (>38°C); all died or had disability. The control groups of trials done in HICs had levels of hyperthermia ranging from 14% to 39%, which might have resulted in an overestimation of the therapeutic effect of cooling. Although original paediatric and adult post-cardiac arrest trials had reported beneficial effects of hypothermia, subsequent larger trials showed that targeted normothermia was as effective as hypothermia;8Moler FW Silverstein FS Holubkov R et al.Therapeutic hypothermia after in-hospital cardiac arrest in children.N Engl J Med. 2017; 376: 318-329Crossref PubMed Scopus (160) Google Scholar hence clinical practice changed to therapeutic normothermia from hypothermia.9Morrison LJ Thoma B Translating targeted temperature management trials into postarrest care.N Engl J Med. 2021; 384: 2344-2345Crossref PubMed Scopus (7) Google Scholar Seizures that occurred soon after birth indicated an earlier onset of fetal brain injury in the HELIX trial than trials done in other HICs. These infants do not respond to cooling, even in HICs,10Gluckman PD Wyatt JS Azzopardi D et al.Selective head cooling with mild systemic hypothermia after neonatal encephalopathy: multicentre randomised trial.Lancet. 2005; 365: 663-670Summary Full Text Full Text PDF PubMed Scopus (1739) Google Scholar as the therapeutic window period possibly closes before birth. The incidence of encephalopathy in most LMIC hospitals is several times higher than that in HICs, indicating sub-optimal antenatal or intrapartum care; HELIX sites were no different. Continued cooling in LMICs without informing parents about HELIX results is injudicious. In our view, global efforts should focus on preventing encephalopathy and avoiding hyperthermia in low-resource settings rather than on an expensive and possibly ineffective therapy that arguably requires 1:1 tertiary neonatal intensive nursing care1Kozlov M Doubts raised about cooling treatment for oxygen-deprived newborns.Nature. 2021; (published online Aug 11.)DOI:10.1038/d41586-021-02201-6Crossref Google Scholar and could harm infants from underprivileged settings. We declare no competing interests. Download .pdf (1.32 MB) Help with pdf files Supplementary appendix Questions about the HELIX trialWe congratulate Sudhin Thayyil and colleagues1 on publishing the HELIX trial in The Lancet Global Health. Based on their results, which showed increased mortality in the hypothermia group compared with the standard care group, they recommended that cooling for neonatal encephalopathy should not be offered in low-income and middle-income countries (LMICs).However, we feel that their results cannot be generalised. Full-Text PDF Open AccessQuestions about the HELIX trialDespite more than half a decade since the launch of the Every Newborn Action Plan, perinatal asphyxia remains one of the most important causes of perinatal mortality and morbidity in neonates in developing countries.1 Hence, the HELIX trial holds paramount importance. The study reports that therapeutic hypothermia was associated with increased mortality and therefore should not be offered in low-income and middle-income countries. 2 However, we have several critical observations. Full-Text PDF Open AccessHypothermia for moderate or severe neonatal encephalopathy in low-income and middle-income countries (HELIX): a randomised controlled trial in India, Sri Lanka, and BangladeshTherapeutic hypothermia did not reduce the combined outcome of death or disability at 18 months after neonatal encephalopathy in low-income and middle-income countries, but significantly increased death alone. Therapeutic hypothermia should not be offered as treatment for neonatal encephalopathy in low-income and middle-income countries, even when tertiary neonatal intensive care facilities are available. Full-Text PDF Open AccessQuestions about the HELIX trialIn the HELIX trial, Sudhin Thayyil and colleagues1 found no benefit in risk of mortality or severe disability following therapeutic hypothermia in neonates with asphyxia in low-income and middle-income countries (LMICs). The authors made firm conclusions on the use of therapeutic hypothermia in LMICs that warrant careful consideration. Full-Text PDF Open AccessQuestions about the HELIX trialIn the HELIX study,1 Sudhin Thayyil and colleagues suggested that all low-income and middle-income countries (LMICs) should immediately suspend use of cooling in neonates.1 The reported absence of benefit and increased mortality in neonates who were cooled are in contrast to our experience and the findings in meta-analyses in LMICs.2,3 Full-Text PDF Open Access