Infants born extremely preterm (<28 weeks' gestation) are at high risk of neurodevelopmental impairment (NDI) with 50% of survivors showing moderate or severe NDI when at 2 years of age. We sought to develop novel models by which to predict neurodevelopmental outcomes, hypothesizing that combining baseline characteristics at birth with medical care and environmental exposures would produce the most accurate model. Using a prospective database of 692 infants from the Preterm Epo Neuroprotection (PENUT) Trial, which was carried out between December 2013 and September 2016, we developed three predictive algorithms of increasing complexity using a Bayesian Additive Regression Trees (BART) machine learning approach to predict both NDI and continuous Bayley Scales of Infant and Toddler Development 3rd ed subscales at 2 year follow-up using: 1) the 5 variables used in the National Institute of Child Health and Human Development (NICHD) Extremely Preterm Birth Outcomes Tool, 2) 21 variables associated with outcomes in extremely preterm (EP) infants, and 3) a hypothesis-free approach using 133 potential variables available for infants in the PENUT database. The NICHD 5-variable model predicted 3-4% of the variance in the Bayley subscale scores, and predicted NDI with an area under the receiver operator curve (AUROC, 95% CI) of 0.62 (0.56-0.69). Accuracy increased to 12-20% of variance explained and an AUROC of 0.77 (0.72-0.83) when using the 21 pre-selected clinical variables. Hypothesis-free variable selection using BART resulted in models that explained 20-31% of Bayley subscale scores and AUROC of 0.87 (0.83-0.91) for severe NDI, with good calibration across the range of outcome predictions. However, even with the most accurate models, the average prediction error for the Bayley subscale predictions was around 14-15 points, leading to wide prediction intervals. Higher total transfusion volume was the most important predictor of severe NDI and lower Bayley scores across all subscales. While the machine learning BART approach meaningfully improved predictive accuracy above a widely used prediction tool (NICHD) as well as a model utilizing NDI-associated clinical characteristics, the average error remained approximately 1 standard deviation on either side of the true value. Although dichotomous NDI prediction using BART was more accurate than has been previously reported, and certain clinical variables such as transfusion exposure were meaningfully predictive of outcomes, our results emphasize the fact that the field is still not able to accurately predict the results of complex long-term assessments such as Bayley subscales in infants born EP even when using rich datasets and advanced analytic methods. This highlights the ongoing need for long-term follow-up of all EP infants. Supported by the National Institute of Neurological Disorders and StrokeU01NS077953 and U01NS077955.

In the Preterm Erythropoietin (Epo) NeUroproTection (PENUT) Trial, potential biomarkers of neurological injury were measured to determine their association with outcomes at two years of age and whether Epo treatment decreased markers of inflammation in extremely preterm (<28 weeks’ gestation) infants. Plasma Epo was measured (n=391 Epo, n=384 placebo) within 24h after birth (baseline), 30min after study drug administration (day 7), 30min before study drug (day 9), and on day 14. A subset of infants (n=113 Epo, n=107 placebo) had interferon-gamma (IFN-γ), Interleukin (IL)-6, IL-8, IL-10, Tau, and tumour necrosis factor-α (TNF-α) levels evaluated at baseline, day 7 and 14. Infants were then evaluated at 2 years using the Bayley Scales of Infant and Toddler Development, 3rd Edition (BSID-III). Elevated baseline Epo was associated with increased risk of death or severe disability (BSID-III Motor and Cognitive subscales <70 or severe cerebral palsy). No difference in other biomarkers were seen between treatment groups at any time, though Epo appeared to mitigate the association between elevated baseline IL-6 and lower BSID-III scores in survivors. Elevated baseline, day 7 and 14 Tau concentrations were associated with worse BSID-III Cognitive, Motor, and Language skills at two years. Elevated Epo at baseline and elevated Tau in the first two weeks after birth predict poor outcomes in infants born extremely preterm. However, no clear prognostic cut-off values are apparent, and further work is required before these biomarkers can be widely implemented in clinical practice. PENUT was funded by the National Institute of Neurological Disorders and Stroke (U01NS077955 and U01NS077953).

Perinatal hypoxic-ischemic (HI) brain injury, often in conjunction with an inflammatory insult, is the most common cause of death or disability in neonates. Therapeutic hypothermia (TH) is the standard of care for HI encephalopathy in term and near-term infants. However, TH may not always be available or efficacious, creating a need for novel or adjunctive neurotherapeutics. Using a near-term model of inflammation-sensitized HI brain injury in postnatal day (P) 17 ferrets, animals were randomized to either the control group (n = 43) or the HI-exposed groups: saline vehicle (Veh; n = 42), Ur (uridine monophosphate, n = 23), Epo (erythropoietin, n = 26), or TH (n = 24) to test their respective therapeutic effects. Motor development was assessed from P21 to P42 followed by analysis of cortical anatomy, ex vivo MRI, and neuropathology. HI animals took longer to complete the motor assessments compared to controls, which was exacerbated in the Ur group. Injury resulted in thinned white matter tracts and narrowed cortical sulci and gyri, which was mitigated in Epo-treated animals in addition to normalization of cortical neuropathology scores to control levels. TH and Epo treatment also resulted in region-specific improvements in diffusion parameters on ex vivo MRI; however, TH was not robustly neuroprotective in any behavioral or neuropathological outcome measures. Overall, Ur and TH did not provide meaningful neuroprotection after inflammation-sensitized HI brain injury in the ferret, and Ur appeared to worsen outcomes. By comparison, Epo appears to provide significant, though not complete, neuroprotection in this model.

We aimed to evaluate diffusion tensor imaging (DTI) in infants born extremely preterm, to determine the effect of erythropoietin (Epo) on DTI, and to correlate DTI with neurodevelopmental outcomes at 2 years of age for infants in the Preterm Erythropoietin Neuroprotection (PENUT) Trial. Infants who underwent MRI with DTI at 36 weeks postmenstrual age were included. Neurodevelopmental outcomes were evaluated by Bayley Scales of Infant and Toddler Development (BSID-III). Generalized linear models were used to assess the association between DTI parameters and treatment group, and then with neurodevelopmental outcomes. A total of 101 placebo- and 93 Epo-treated infants underwent MRI. DTI white matter mean diffusivity (MD) was lower in placebo- compared to Epo-treated infants in the cingulate and occipital regions, and occipital white matter fractional isotropy (FA) was lower in infants born at 24–25 weeks vs. 26–27 weeks. These values were not associated with lower BSID-III scores. Certain decreases in clustering coefficients tended to have lower BSID-III scores. Consistent with the PENUT Trial findings, there was no effect on long-term neurodevelopment in Epo-treated infants even in the presence of microstructural changes identified by DTI.