Two new publications from multicenter trials of hypothermia as a treatment for post-asphyxial encephalopathy. The first from the NICHD network, and the whole body cooling study led by Seetha Shankaran. They examined the children and performed IQ testing at 6 to 7 years. (Shankaran S, Pappas A, McDonald SA, Vohr BR, Hintz SR, Yolton K, et al. Childhood Outcomes after Hypothermia for Neonatal Encephalopathy. New England Journal of Medicine. 2012;366(22):2085-92. http://www.nejm.org/doi/full/10.1056/NEJMoa1112066#t=abstract) There were 208 infants originally in the trial and only 18 of the surviving 140 infants were lost to follow-up.

The results confirm that therapeutic hypothermia improves survival without disability, diagnosing disability if the IQ was less than 70 (although it was not quite ‘significant’; p value was 0.06) and a reduction in severe disability (which was IQ less than 55, disabling cerebral palsy or blindness).

The second was a follow-up of the CoolCap trial, the parents were telephoned in order to complete the WeeFIM, a functional classification of the children’s abilities, at 7 to 8 years of age. (Guillet R, Edwards AD, Thoresen M, Ferriero DM, Gluckman PD, Whitelaw A, et al. Seven- to eight-year follow-up of the CoolCap trial of head cooling for neonatal encephalopathy. Pediatr Res. 2012;71(2):205-9. http://www.nature.com/pr/journal/v71/n2/full/pr201130a.html) Unfortunately only half of the babies were followed, which reduces the reliability of the results. (As an aside Pediatric Research now puts the methods at the end of the abstract, and at the end of the paper in smaller type. I think this is just weird. The methods are the most important part of any paper, how can you make anything out of the results if you read them before the methods? Can we start a campaign to make them change this back?)

What is really strange also is that the authors do not present the results from the cooled versus the control groups. Although they appropriately note that they had limited power because of the large numbers of drop-outs, I don’t understand why they didn’t at least present the data in that way, they focus instead on the predictability of the WeeFIM from the 18 months Bayley scores. I don’t see that as being any less affected by the low follow-up rate, but the authors claim that the data show that the Bayley is a relatively good prediction of functional outcome at 7 to 8 years. In fact though when I look at their figure 3 it seems that there are quite a few infants with an 18 month Bayley of 70 or less who have quite good functional scores.

If we examine the same relationship from Shankaran’s paper (looking at the supplementary material in the on-line appendix table 3) we can see that of 36 infants identified with a low Bayley at 18 months, 33 of them had an IQ less than 70 at 6 to 7 years. And of 86 infants without low Bayleys then 76 had an IQ over 70.

This is a much higher positive predictive value of an 18month Bayley in this study (even though the numbers are becoming quite small) than is the case for the preterm infant. It could make some sense that the cerebral injury in asphyxia is more diffuse, leading to a lower capacity for the plasticity of the brain to overcome the injury, whereas in preterm infants the injuries may be more patchy, especially periventricular hemorrhage, so they are more likely to slowly improve and catch up.

This confirms the value of hypothermia,  which has now been studied in a large number of similar trials, recently an updated systematic review was published, (Tagin MA, Woolcott CG, Vincer MJ, Whyte RK, Stinson DA. Hypothermia for Neonatal Hypoxic Ischemic EncephalopathyAn Updated Systematic Review and Meta-analysisHypothermia for Neonatal Encephalopathy. Arch Pediatr Adolesc Med. 2012;166(6):558-66. http://archpedi.jamanetwork.com/article.aspx?doi=10.1001/archpediatrics.2011.1772 ) Tagin and colleagues included RCTs that had follow-up to at least 18 months of age, which meant 7 trials with a total of over 1200 infants. This analysis confirmed that hypothermia decreases death, and decreases neurologic abnormalities and low Bayley scores.

I have only one beef with this well done meta-analysis: the very weak recommendation at the end of the abstract: ‘Clinicians should consider offering therapeutic hypothermia as part of routine clinical care to these newborns’.

Should ‘CONSIDER’ ?!

How about ‘Clinicians who do not offer therapeutic hypothermia are negligent’! We should sometimes go beyond the limits of the usual niceties of standard scientific writing when we are talking about saving babies lives and reducing handicap. Fortunately I don’t have an editor here on my blog!