Towards the end of last year the Canadian Pediatric Society published a new ‘position statement’. These are official proclamations of the society, supposedly based on the best available evidence to guide practice, and which become de facto standards of care. This particular one ‘Counselling and management for anticipated extremely preterm birth’ presented an opportunity to update a 20 year old statement. This is part 3 of my response.

Among many of the failings of this statement, the literature review misquotes many of the articles reviewed. Some of the misquotations are very serious, for example it is stated that “A large UK-wide study (EPICURE) showed that  … At 11 years of age, this cohort had serious cognitive impairment (score < 2 SD below mean) in 40% and functional disability in 45%. Approximately 50% of the children were free of serious disability. Similar results were reported at eight years of age for an Australian cohort.”

In fact the article quoted shows the following,  in table 2, the upper part of which is reproduced below, of 219 infants examined there were 34 with severe disability, or 15.5%, in other words 86.4% who were free of serious disability.

The Australian cohort did indeed have somewhat similar results, but similar to those actually in the Epicure publication not those misquoted in the position statement. The Australian publication which was quoted reports that 87% of their survivors of 23, 24 and 25 weeks gestation were free of serious disability. Such misquotation is an egregious failing. (also, Epicure was a study in the UK and Ireland, not just the UK).

The other serious example is the misquotation of all of the data in Table 2 of the statement. For example, the study by Synnes et al: the table shows “Survival free of severe adverse neurodevelopmental outcome” as being 35% for infants born at 23 weeks.  But the original article (which actually is quite hard to get hold of, being a non-peer-reviewed publication in a provincial medical association rag not included in PubMed (Synnes A, et a;. Management of the newborn delivered at the threshold of viability. BC Medical Journal. 2008;50(9):498-508.)) actually notes that survival of infants designated as being born at 23 weeks was 22%, and that 10% of survivors had severe neurodevelopmental impairment; they showed no difference in the frequency of severe impairment among survivors between 23 and 25 weeks gestation. Nowhere in that article does the figure 35% appear. If survival is 22% then it isn’t possible to have survival free of adverse outcome of 35%!

The data of Jacobs et al are quoted as showing 36, 57, and 63% probability of survival without severe adverse neurodevelopmental outcome after birth at 23, 24 and 25 weeks respectively. In fact survival was 21%, 58%, and 65% at each of those GA weeks, and the percentage of survivors with severe impairments was 18%, 21% and 12%. It isn’t possible to come up with the percentages quoted in the table from these figures.

Even more misleading are the figures in table 2 which are from an article which Charlene Robertson wrote to review the progress in outcomes of a 30 year period. The only place you can find those data are in bar charts at the end of the article, and you have to estimate what the actual numbers are, for example in 2001-2003 of the 24 babies born at 24 weeks it looks like about 20% died in the NICU, about 4% died after discharge (probably 1 baby), and about 10% were impaired, this is represented in table 2 as being 54% survival without severe impairment. Also Dr Robertson uses a different definition of impairment, which cannot be compared to the other 2 articles, including any cerebral palsy but only Bayley scores more than 3 SD below the mean, whereas Jacobs uses a Bayley 2 SD below the mean but only included severe cerebral palsy.

These numbers are those used to support the recommendations in this position statement. The fact that there are so many serious errors in these figures is very disturbing. This position statement will affect the way babies are treated, the evidence in the statement ought to at least be accurately quoted.