Have school-age cognitive outcomes improved for preterm infants in the last decade of neonatal care?

Abstract

Neurodevelopmental follow-up studies of very preterm infants over the years have shown a high prevalence of neurological, sensory, cognitive, and behavioural deficits. While there has been good evidence to show the reduction in neonatal mortality in these infants following the introduction of neonatal intensive care, the evidence for an improvement in disability rates in more recent cohorts is less compelling. There may be many reasons for the inconsistencies in the results published, but variation in definitions of deficits, cohort selection, individual exclusions, assessments used, and the age at which they are applied may all operate. Many reports look for associations between perinatal variables such as illness severity, brain scans or post-natal growth, and outcomes. Whilst significant relationships are often found, they are rarely useful as precise predictors of future disability. The widescale use of cranial imaging by ultrasound and magnetic resonance imaging in preterm infants in the past couple of decades have helped to explain the connections between neonatal illness in this high-risk population and subsequent neurological and cognitive deficits. White matter damage, whether from periventricular haemorrhage, periventricular leukomalacia, or simply diffusely poor myelination is common, and significantly associated with later function. Poor white matter development in early life is associated with lower grey matter volumes later. Early white matter damage is particularly related to inflammation in the preterm and chorioamnionitis, later sepsis, or necrotizing enterocolitis markedly increase the risk of it occurring. Antenatal steroid use is associated with fewer white matter lesions, and its use has been shown to equate with better developmental outcomes. Its wider use may have contributed to any improvements in outcome seen in the most recent years, as may better management of neonatal sepsis. Lind et al. describe a Finnish cohort of very low birthweight infants born in 2001 to 2003 and followed up at 5 years together with term controls. They found, as expected, lower scores in the preterm children compared with the term controls on most of the sub-sets of the NEPSY-II, but the mean IQ score for the preterm children was 100. They conclude that as the cohort was born in this century, these results represent evidence for improved outcomes for neonatal care in recent years. In 2007, Wilson-Costello reviewed the published evidence for improved outcomes after intensive care in preterm infants born from 1959 to 2002. For neurological impairment there was sufficient evidence to suggest that it had shown a significant decrease in the most recent periods reported. All studies reported poorer cognitive abilities in preterm infants, and no real evidence that these results were improving over time. The most recent longitudinal study of outcomes in preterm infants (<1000g) compared results at 20 months from 1982 to 1989, 1990 to 1999, and 2000 to 2002. The latter period showed significantly better neurodevelopmental outcomes than the previous period, including a small but significant increase in Psychomotor Developmental Index score (but not Mental Developmental Index score) on the Bayley-II scale. These improved results were accompanied by fewer lesions noted on cranial ultrasound scans in the neonatal period. Can the results from this paper be used as evidence that cognitive outcome for Finnish preterm children, and especially those of lowest gestation have improved in the last decade? I think not. The relatively good performance of the very low gestation infants can be attributed to the exclusion of untestable children instead of including them as having a nominal score of say less than 50. When the NEPSY-II scores were examined the preterm infants performed significantly less well than term-born children. The IQ test scores of the preterm infants were also significantly poorer than the controls. The difference was due to the high mean IQ scores of the term children, which were nearly 1 SD above the standardized test mean. It emphasizes the importance of controls in such studies. As the authors suggest, this is almost certainly an example of the Flynn effect where mean IQ scores appear to creep upwards over time. We will need to await larger and longitudinal studies to confirm whether neonatal care improves cognitive outcome as well as survival and cerebral palsy rates.