A global picture of outcomes after preterm birth: Is there a discrepancy?

Abstract

Understanding the long-term impact of preterm birth on childhood survivors is difficult. First, the number of births decreases substantially with decreasing gestational age: although approximately 10% of births are preterm, only around 0.5% of all births occur below 28 weeks' gestation (defined as extremely preterm by the World Health Organization); adverse consequences also disproportionately affect those born the earliest.¹ Second, due to increasing rarity with decreasing gestational age, collecting the sufficiently large sample sizes required for meaningful results becomes more difficult as data must be collated across large geographical regions and/or over long time periods. Third, study comparisons are difficult: harmonization of exposure and outcome definitions is complicated, with other methodological considerations including population differences, temporal changes, and variable follow-up rates and assessment ages.²

The seemingly contradictory results for very preterm births from Denmark (with an increasing prevalence of cerebral palsy)³ and Taiwan (with decreases in severe neurodevelopmental impairment)⁴ therefore needs close examination. Are these results actually conflicting, or is something else happening? In Denmark, Fogh et al. used data about all preterm births – divided into gestational age categories of extremely (23–27 weeks), very (28–31 weeks) and moderately (32–36 weeks) preterm – obtained from national registers to examine survival and cerebral palsy diagnoses. In Taiwan, Wang et al. studied a prospective, geographical cohort including the 90% of survivors at 2 years corrected age born less than 31 weeks – divided into categories of 22 to 25 weeks, 26 to 28 weeks, and 29 to 30 weeks – who underwent detailed developmental assessment. From such brief descriptions, it is already clear that direct comparison is challenging, and also that there is much more to learn from each study than the headline result.

Indeed, facilitating comparison is not the main objective of either study, but it is useful. Comparisons prompt thinking about differences and stimulate ideas about how to improve management. But comparisons can be misused: for example, to say that one country or region is better (or worse) than another. This is clearly not true, as local influences play important roles and require differing approaches to tackle them. A better way to consider similar studies (like these two articles^{3, 4}) might be to consider them as different but complementary colours, both required to complete different parts of an overall picture. We can then learn from the differences – and will also be led to remember the gaps that are as yet unfilled.

Taking this approach, it is clear that knowledge gaps relate to milder disease – about which neither study comments – and to a comprehensive assessment of outcomes (not just cerebral palsy) in children born at more than 30 weeks' gestation. Both of these issues were highlighted by the French national EPIPAGE-2 cohort of children born preterm at 24 to 34 weeks gestational age. Developmental coordination disorder and behavioural difficulties were included alongside more routine components (cerebral palsy with Gross Motor Function Classification System level I, and mild cognitive, visual, and hearing deficits) in the ‘mild’ category of neurodevelopmental impairment; at 5 years 6 months of age, around 30% of all children were classified in this category, with little variation according to gestational age.⁵

As ever, the devil is in the detail. The authors of these two disparate, robust studies require congratulations for providing detailed information about their respective regions and adding to global knowledge about preterm birth. We must, however, remain vigilant for what we do not know, particularly with respect to the potentially high burdens of milder impairment.

The author has stated that he has no interests which might be perceived as posing a conflict or bias.