Development of early predictors of long-term neurodevelopmental outcomes for extremely preterm and/or extremely low birth weight infants

Abstract

IntroductionAccurate prediction of long-term neurodevelopmental outcome is currently not possible for extremely preterm (EP) and/or extremely low birth weight (ELBW) infants before their discharge home from the hospital. While most EP and/or ELBW infants have normal outcomes, some will develop a significant neurodevelopmental disability and/or delay (NDD), including cerebral palsy (CP) and/or functional impairment (FI). These outcomes may not become apparent for months or even years after discharge. Significant time and resources are spent on surveillance of EP and/or ELBW infants assessing for possible complications of their preterm birth, which is a substantial burden for the child, family, and health system.Outcome prediction before hospital discharge relies upon a combination of birth history, neonatal complications, results of neuroimaging (cranial ultrasound [CUS], and magnetic resonance imaging [MRI]), and more recently, the general movements (GMs) assessment. Neuroimaging and GMs are qualitative assessments requiring specialist training and are subject to potential bias.The addition of quantifiable measures of the quality of movement before discharge from the hospital of an EP or ELBW infant may increase the sensitivity and specificity of the prediction of long-term neurodevelopmental outcomes. Accelerometers provide a low-cost method of quantifying movement and are easily applied without moving the baby. AimsThe aims of this thesis were to;·         Review the literature surrounding outcome prediction for EP and/or ELBW infants focusing on investigations and assessments suitable for use before hospital discharge.·         Undertake a prospective observational study to determine in EP and/or ELBW infants if;Routine inpatient CUS performed on EP and/or ELBW infants at approximately day five of life, day 28 of life, and 36-weeks post menstrual age (PMA), predicts NDD (i.e., CP or FI).General movements (GMs) assessment in EP and/or ELBW infants at 28-, 32-, or 36- weeks PMA predicts NDD.Developmental trajectories (calculated from two or more GMs assessments before hospital discharge) in EP and/or ELBW infants predicts NDD.Quantitative measures of motor activity at 28-, 32-, or 36-weeks PMA, obtained from a movement detection system (MDS) (comprising of four wireless tri-axial accelerometers), correlate with the GMs assessments undertaken before hospital discharge.Quantitative measures of motor activity in EP and/or ELBW infants at 28-, 32-, or 36-weeks PMA, obtained from the MDS, predicts NDD.Quantitative measures of motor activity in EP and/or ELBW infants obtained from the MDS change according to PMA at the time of assessment (28-, 32-, and 36-weeks).The PMA at which quantitative measures of motor activity obtained from the MDS in EP and/or ELBW infants first correlate with GMs assessment and predicts NDD. Research design and methodologyA prospective observational cohort was used to determine the predictive value of movement parameters obtained from an MDS consisting of four wearable sensors (tri-axial accelerometers). Participants were EP and/or ELBW infants admitted to the Mater Mothers’ Hospital (MMH), Brisbane. Movement detection system and video recordings for the GMs assessments were undertaken on infants at 28-, 32-, and 36-weeks PMA depending on their initial age at enrolment, clinical condition, and age at discharge. Results of routine CUS and long-term neurodevelopmental follow-up at one and/or two-years CA were recorded.ResultsThe CUS had an overall sensitivity of 50% and a specificity of 98% in predicting CP. The GMs assessment at 36-weeks PMA had an overall sensitivity of 20% and specificity of 100% for predicting CP and a sensitivity of 14% and specificity of 100% for predicting FI. Assessing the developmental trajectory on GMs did not improve the prediction of CP or FI. The intra- and inter- scorer reliability of the GMs assessments was poor. The MDS parameters did not differ significantly when comparing measures between; infants with and without cramped synchronised (CS) movements, infants with and without a diagnosis of CP, and infants with and without a diagnosis of FI. Many MDS parameters decreased in value with increasing PMA demonstrating a significant maturation effect.                ConclusionCranial ultrasound findings continue to provide clinically useful information for long-term neurodevelopmental outcome prediction. General movements assessments performed before hospital discharge had poor sensitivity for predicting CP and FI. Isolated use of quantitative measures of movement, obtained from four tri-axial accelerometers before hospital discharge, did not appear to assist with outcome prediction for EP and/or ELBW infants. The study was limited by the large number of parameters used for comparison. Future research should use adequately powered studies and investigate the use of pattern recognition and artificial intelligence in movement analysis as well as the combination of new neuroimaging techniques and other biomarkers. Surveillance and assessments post-hospital discharge remain essential for the detection of neurodevelopmental problems in infants born EP and/or ELBW.