EPICure: facts and figures: why preterm labour should be treated


  • K Costeloe,

    Corresponding author
    1. Academic Unit for Child Health, Queen Mary’s School of Medicine and Dentistry, University of London, Homerton University Hospital, London, UK
    Search for more papers by this author
  • EPICure Study Group

    1. Academic Unit for Child Health, Queen Mary’s School of Medicine and Dentistry, University of London, Homerton University Hospital, London, UK
    Search for more papers by this author

Prof K Costeloe, Academic Unit for Child Health, Queen Mary’s School of Medicine and Dentistry, University of London, Homerton University Hospital, London E9 6SR, UK. Email Kate.Costeloe@homerton.nhs.uk


The principal objective of the EPICure studies was to determine both short- and long-term outcomes of extremely preterm birth. Data were collected for all births before 26 completed weeks of gestation in the UK and Republic of Ireland for 10 months in 1995. Of 811 infants admitted to neonatal units, 314 (39%) survived. Of these, 283 (92%) were assessed at 2.5 years and 241 (78%) at 6 years, together with a comparator group selected from classmates in normal schools. At 6 years, 32 (13%) of those assessed had disabling cerebral palsy and 31 (13%) had severe sensory impairment. Cognitive impairment was a more frequent adverse outcome: the mean intelligence quotient (IQ) was 82 (SD 19). Fifty children (21%) had IQs > 3 SD less than that of the comparator group, and in 48 children (20%), IQ was 2–3 SD less than that of the comparator group. Infants whose mothers had received antenatal steroids had fewer severely abnormal head scan findings. In this population, abnormal head scan findings are independent predictors of reduced, severe motor disability at 2.5 years. Using step-wise logistic regression analysis, postnatal transfer was associated with severe motor disability; prolonged membrane rupture with reduced Mental Development Index (MDI) and antenatal steroid with increased MDI. It is clear that factors around the time of birth are critical in determining outcome, irrespective of later complications during neonatal intensive care. Since 1995, there is good evidence of improved survival, but it is not clear whether or not the number of survivors with severe adverse outcomes has changed. A new cohort of births <27 completed weeks is currently being collected in English maternity units. Data collection around the time of birth is more detailed than in 1995 in order to better explore the relationship between perinatal factors and later outcomes.


The aim of the EPICure study1–4 was to measure health outcomes for children born in the UK and Republic of Ireland before 26 completed weeks of gestation. The purpose of this article was to review the clinical progress of the surviving children and associations between perinatal and neonatal clinical events and later outcomes and the relevance of these results to current practice.


A simple data set (‘working’ gestational age, birthweight, sex, presence of congenital anomaly, recording of signs of life and admission to a neonatal unit) was collected for all births (n = 4004) in the gestational age range of 20–25 completed weeks in all maternity units in the UK and Republic of Ireland for a period of 10 months in 1995.

Details were collected for all babies admitted to neonatal units (n = 811) with simple obstetric information, data to validate the gestational age, limited information about neonatal interventions and details of the following major complications of prematurity: hydrocephalus or parenchymal cysts on cerebral ultrasound, continued oxygen dependency at 36 and 40 weeks postmenstrual age and treated retinopathy of prematurity.

All surviving children have been offered standardised medical and neurodevelopmental assessment at 2.5 and 6 years. The children are currently being seen again at 10 years.


Around 50% of those born at 23 weeks with reported signs of life, 82% of those at 24 weeks and 92% at 25 weeks were admitted for neonatal intensive care. Overall survival to discharge from hospital of those who were admitted was 39%, rising from 20% at 23 completed weeks to 52% at 25 completed weeks.1

Of the 314 survivors, 6 are known to have died during the first year; the investigators are not aware of any other deaths. At 6 years, 241 children (78% of survivors), together with a comparator group of 160 children who were classmates of children from the main EPICure cohort who are in mainstream schools, were assessed by a paediatrician and psychologist.4 Of the 47 children who were not assessed at 6 years but who were seen at 2.5 years,2 36% had been classified as having a severe disability and 26% as having other disabilities; corresponding rates in the whole population assessed at 2.5 years were 23% with severe disabilities and 25% with other disabilities.

Of the children assessed at 6 years, 32 (13%) had neuromotor problems with associated disability, 17 (7%) had moderate or severe visual impairment and 14 (6%) had moderate or severe hearing loss.

Using the test norms for the Kaufman Assessment Battery for Children that was used to assess cognitive function, 50 (21%) of the EPICure children were found to have intelligence quotient (IQ) > 2 SD below the mean. If instead of simply using the test norm, which was derived from children born in the 1970s, the EPICure children are compared with their 160 classmates, whose mean IQ was 106 ± 12, then the number falling more than 2 SD below the mean rises to 98 (41%). The mean IQ of the 241 EPICure children who were assessed was 82 ± 19. While there was no difference in the IQs of boys and girls in the comparator group, the mean IQ of boys in the EPICure group was 10 (95% CI 5–15) points lower than that of the girls.

Linear regression analysis was used to determine independent factors associated with motor and developmental disability in the 283 (92%) of the 308 surviving EPICure children who were assessed at the age of 2.5 years. This confirmed the prognostic importance within this cohort of severe abnormality of the cerebral ultrasound scan appearance during the initial hospitalisation and of bronchopulmonary dysplasia (BPD) among survivors, defined as a continued requirement for supplemental oxygen at 36 weeks postmenstrual age.3

The independent effects of clinical factors around the time of birth and in the first 24 hours of life on death, serious abnormality of the cerebral ultrasound scan and continued oxygen dependency were also analysed using logistic regression analysis. The results showed independent associations between exposure to antenatal steroids and both decreased death (OR 0.57, 95% CI 0.38–0.85; P = 0.006) and reduced occurrence of severe cerebral scan findings in survivors (OR 0.39, 95% CI 0.20–0.77; P < 0.001) and between the use of tocolysis and both seriously abnormal scan findings (OR 2.02, 95% CI 1.04–3.94; P < 0.05) and long-term oxygen dependency (OR 2.53, 95% CI 1.42–4.51; P = 0.001). The importance of items describing the clinical condition of the baby in the first few hours of life were evidenced by the association of having an admission temperature more than 35°C with less death (OR 0.58, 95% CI 0.39–0.85; P = 0.006) and with less BPD (OR 0.50, 95% CI 0.27–0.96; P < 0.05), by the association of bradycardia at 5 minutes with increased risk of death (OR 1.99, 95% CI 1.07–3.67; P = 0.03) and by the association of high clinical risk index for babies score5 with both increased risk of death (OR 1.31, 95% CI 1.25–1.38; P < 0.0005) and of BPD (OR 1.12, 95% CI 1.03–1.22; P < 0.05).


While the babies born at extremely low gestation who survive labour to be admitted to our neonatal units represent only a very small proportion (around 0.2%) of total births in the UK and Republic of Ireland, there are nonetheless more than 1000 such babies each year, and the investment, both emotional and in terms of resource, in their care by their families, the NHS and later the educational services is considerable. It is therefore critically important that we study their outcomes and understand how these are affected and may be modified by clinical interventions. It is equally important that we understand the significance of the outcomes that we measure in studies such as EPICure, for the longer term health and welfare of our patients.

While for the individual child the occurrence of a severe motor or sensory impairment is clearly devastating, the rates of EPICure children with severe impairments in these domains are not higher than those in other published series of children slightly less preterm.6,7 It is the number of children in the EPICure cohort with cognitive problems, which are also those that are most difficult to interpret, that causes the most concern. The observation that the proportion of children with a moderate or severe cognitive problem rises from 21 to 41% depending on whether one uses the test norm or refers to the comparator group highlights the difficulty in understanding the significance of these data for these children. A further difficulty is the anxiety that there was an excess of children whose assessment had suggested disability at 2.5 years among those not seen at 6 years. It is also apparent that there is a high prevalence of behavioural problems among these children (D. Wolke, unpubl. obs.).

At the time of their last assessment, the EPICure children were only 6.5 years old, and considerable caution is needed in terms of predicting what these observations mean in terms of their journey through adolescence and out into the adult world.

The messages emerging from those long-term studies that have been reported would suggest that many children overcome their early problems.8 Unlike the EPICure cohort, the published series do not comprise solely extremely preterm children, and the observation that the EPICure children are more likely to have impairments in multiple domains emphasises the caution that must be exercised in extrapolating the findings from other cohorts to these children.

It is also important to consider whether things may have changed significantly in the decade that has passed since 1995 when these children were born. Existing population-based data suggest that survival is increasing at 24 and 25 completed weeks of gestation. Whether or not the proportion of surviving children with disabilities is changing is unknown.

It is likely that the number of women threatened with extremely preterm birth receiving antenatal steroids (65% in EPICure) has increased. It is also likely that early neonatal hypothermia (40% of babies in EPICure) has reduced and that in addition, increased attention is paid to the quality of respiratory and circulatory care in the early neonatal period. Another change, due to the evidence of their association with cerebral palsy, has been the reduction in the use of postnatal steroids given to reduce BPD, both in overall frequency of use and of the dose and duration. Dexamethasone was given to 72% of the surviving children from the EPICure study for a median of 24 days.

It is possible that the condition of babies at birth may improve further with increased sophistication of tocolytic agents and in more refined selection of those women for whom they may confer benefit. It is also believed that the clinical service in the UK will develop with increased capacity within perinatal centres, so as to facilitate in utero as opposed to ex utero transfer for specialist care.


The data from the EPICure study provide the most comprehensive overview that we have of outcomes of extreme prematurity. Many of these children are doing well. Depending on the interpretation (particularly of cognitive measures), between 54 and 75% of the population assessed at the age of 6 years were functioning within the normal range or had a mild disability that should not prevent them from being independent. The limited analyses that we can do examining for associations between interventions and outcomes suggest that significant associations do exist. This supports the view that it should be possible to further improve care of women and of their preterm babies around the time of birth to achieve a reduction in the number of children with impairments.