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 5 of my response.

Some of these thoughts will be familiar to frequent readers of this blog.

The statement promotes inappropriate definitions of severe disability. Many of the articles quoted refer to infants who have a 2 year Bayley MDI of less than 70 as being severely impaired. Such a definition is entirely inappropriate. Most preterm infants with an age corrected Bayley 2 of less than 70 do not have a cognitive deficit (as measured by full scale IQ in later childhood).

Back in 2005 Maureen Hack published an article comparing scores on the Bayley MDI with IQ scores at 8 years of age. She showed that only 1/3 of babies who had a Bayley MDI below 70 at 20 months of age had an IQ at age 8 years below 70. 1/3 had an IQ between 70 and 85 and 1/3 had an IQ above 85.

The 5 year outcome of the CAP trial compared the 18-20 month Bayley MDI to 5 year IQ testing, and found a substantial overall improvement in cognitive scoring, only 18% of babies who had a Bayley MDI less than 70 actually had evidence of cognitive impairment, that is a WPPSI IQ test less than 70 at 5 years of age. 

Thus to withhold active care from a preterm infant because they are predicted to have an increased chance of having a 2 year Bayley MDI less than 70 is entirely inappropriate. Early childhood developmental delay among very preterm infants is not the same as impairment!

The WHO definition of impairment is as follows ‘An impairment is any loss or abnormality of psychological, physiological or anatomical structure or function’. A delay in reaching milestones or a score below the normal range on a Bayley is not an impairment.

From an editorial in Pediatrics last year by Colombo and Carlson

The BSID is a global test designed to identify developmental delay. Its role and place within the field of developmental science is relatively well established. The BSID is, to be charitable, only modestly related to school-age cognitive development (ie, the outcome that is most meaningful to investigators in this field). The BSID is a global measure of developmental status in infancy that assesses and aggregates the timely attainment of relatively crude milestones in infancy and early childhood.

Simply, the BSID is not an adequate indicant of specific cognitive skills that may be differentially affected by interventions or exposures, nutritional or otherwise, and so its use to evaluate the construct of infant cognition is seriously deficient in the context of recent advances in developmental science.

It is this ‘seriously deficient’ test that we have been using for years as an indicator of developmental ‘impairment’ that often leads us to consider re-orienting care. I seriously doubt that is an appropriate thing to do.

In a similar vein, most of the quoted articles state that any diagnosis of cerebral palsy is a serious impairment, whereas, in fact, most preterm infants with cerebral palsy are ambulant. Most preterm infants with CP do not have major abnormalities in function.

More recent articles use the GMFCS system, which accords a score reflecting functional impact to infants with cerebral palsy. A score of 1 refers to CP without functional limitation, 2 means some limitation but able to walk without assistance, 3 means ambulation with aids, 4 and 5 are more severe with little useful motor function at a GMFCS of 5. A score of 3, 4 or 5 could be considered a major impairment, but in fact a minority of preterm infants diagnosed with CP have such a major impairment. In the Epicure2 study 7% of survivors had severe motor impairment, equivalent to a GMFCS of 3 or more.

The Epicure1 study has shown that a Bayley score of more than 3 SD below the mean at 2.5 to 3 years of age is more predictive of having cognitive impairment at long term. Although GMFCS scores tend to stay stable, even when classified at 2 years, you really have to test later than 2 years of age to be able to predict cognitive outcomes.

The Epicure1 study also notes no significant difference in cognitive scores or the proportion of cognitively impaired children by gestational age, when measured at 5 years of age. Gregory Moore and colleagues from Ottawa have recently published a systematic review of the literature of cognitive scores, measured between 4 and 8 years of age, in large cohorts of infants born after 1995. They show no significant gradient by weeks of gestational age. In other words, any position statement which uses cognitive impairment as a justification for non-intervention at lower gestational ages is in error. The rates of cognitive impairment among survivors do not differ substantially by gestational age.

One of the things that puzzles me about very preterm infants is the way they continue to catch up to their term peers over an extremely long time, I posted recently about a study from Germany showing catch up of executive function after 8 years of age, and my colleague Mai Luu published a fascinating article in 2011 (now open access) showing continued catch up in cognition and language skills after 8 years of age, even up to 16 years for former very preterm infants. Why 3 months in the NICU should delay development which then slowly catches up in this extremely prolonged way is a conundrum.

There is also an important philosophical question, which is not addressed at all by the position statement; what severity of impairment is enough to appropriately make a difference in decision making around the provision of life-sustaining interventions? If we don’t address that issue, why discuss impairment at all?

If we can decide what impairments are profound enough to appropriately affect decision making, then we need to ask whether we can predict them, and in particular whether we can predict them before birth. I think that the answer to these questions is that we can’t predict impairment with any degree of certainty.

In that case we should really question whether rates of impairment are appropriately used as justification for not initiating intensive care.