Long term follow-up of extremely preterm neonates

BMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e8252 (Published 04 December 2012) Cite this as: BMJ 2012;345:e8252
  1. Floris Groenendaal, associate professor of neonatology1,
  2. Cuno Uiterwaal, associate professor of clinical epidemiology2
  1. 1Wilhelmina Children’s Hospital, University Medical Centre Utrecht, Utrecht, Netherlands
  2. 2Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, Netherlands
  1. C.S.P.M.Uiterwaal{at}umcutrecht.nl

Good outcomes data are needed now more than ever

Cohort studies have been used to investigate morbidity and mortality in preterm neonates since the early days of neonatal intensive care and have shown adverse sequelae in many of those who survive.1 Such studies have identified many perinatal risk factors that affect outcome. This has improved quality of care and led to a decrease in the incidence and severity of cerebral palsy in preterm neonates.2 Studies of children followed up into puberty identified risks of abnormalities in psychological development and behaviour associated with preterm birth that were not detectable at younger ages.3 Advances in neonatal intensive care in the 1990s mean that neonates who are born as early as 22-23 weeks’ gestational age are now offered active treatment. Two linked research papers from the EPICure 2 study (doi:10.1136/bmj.e7976 and doi:10.1136/bmj.e7961) look at trends in mortality and long term morbidity in this highly vulnerable group.4 5

Costeloe and colleagues report on survival and neonatal morbidity for babies born at 22-26 weeks’ gestation in England during 2006 and assess changes in survival and major morbidity since 1995 for babies born at 22-25 weeks of gestation.4 They explored whether more emphasis on evidence based interventions in preterm neonates and increased centralisation of delivery and intensive care treatment for the most premature babies had reduced mortality in neonates born before 26 weeks of gestation. They found that survival of babies born at 22-25 weeks of gestation had improved, specifically during the first week of life.

However, patterns of major morbidity and proportions of survivors with morbidity had not changed, and a greater absolute number of neonates had major morbidity. Fifty one per cent of survivors born at 26 weeks had major morbidity, and this increased to 77% for babies born at 23 weeks. Moreover, predictors of short term outcome had not changed substantially. Striking findings were the low reported proportion of neonates transferred antenatally to a tertiary centre, changing trends in ethnicity for extremely preterm neonates, the high proportion of patients with chronic lung disease and surgery for retinopathy of prematurity, and the high proportion of surviving patients with abnormalities on cerebral ultrasonography.

Moore and colleagues, using the same cohorts, compared rates of neurological and developmental impairment at the age of 3 years for extremely preterm babies.5 Because of the high rates of neurological sequelae reported in the earlier cohort (EPICure), the authors hypothesised that increasing survival of extremely low gestational age babies would lead to an absolute increase in neurological and developmental problems if the burden of major morbidity remained the same. They found that the prevalence of important adverse neurological and developmental outcomes had indeed not improved over the period between the two studies.

These authors have undertaken a huge research project. Causal explanations are obviously challenging. The authors considered the recent changes that have occurred in the organisation of neonatal care of premature babies, and the care itself, but other changes may have influenced the studies’ findings. Comparison of longer term neurological and developmental outcomes was hampered by changes that had occurred in measurements, but an even more important factor was the limited access to follow-up data because of legislation. A substantial number of the 2006 cohort were lost to follow-up, which might have made the two cohorts less comparable. However, the investigators used all available means to deal with these difficulties, and their findings seem to be the best available estimates of outcome.

The findings of these studies have substantial clinical implications. The most powerful predictor of adverse outcome remains birth at a lower gestational age. Although mortality for extremely preterm neonates declined between 1995 and 2006, the rate of serious sequelae did not. Therefore, a higher absolute number of affected babies need lifelong special care. Out of every 100 neonates born at 24 weeks, 60 will die despite intensive care, and of the 40 survivors 12 will have serious impairments.5 Only 11 out of 100 neonates born at 23 weeks will survive without impairments. Impairments were still present in 20% of neonates born at a gestational age of 26 weeks, 5 who were used as the comparator group in Costeloe and colleagues’ study.4 Cognitive development seems to be most affected. Similar outcome data are seen in the Netherlands, and neonatal intensive care is therefore not offered routinely to neonates born before 24 completed weeks’ gestation. Cranial magnetic resonance imaging does not help predict which patients will have long term intellectual or behavioural impairment, so children born extremely prematurely must be carefully followed up long term in dedicated clinics.

It is a matter of concern that the authors of the current studies had limited access to follow-up data, because studies that can accurately estimate outcome are of utmost importance for counselling families and improving the planning of care offered by healthcare providers and society. Excellent follow-up is an essential part of neonatal intensive care, as the EPICure studies have elegantly shown. Access to outcomes data is needed now more than ever.


Cite this as: BMJ 2012;345:e8252


  • Research, doi:10.1136/bmj.e7976
  • Research, doi:10.1136/bmj.e7961
  • Competing interests: Both authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; not externally peer reviewed.