656 - Active Sleep During Extrauterine Third Trimester is Positively Associated with Brain White Matter Volume in Preterm Infants

Publication Number: 2639.656

Sudeepta Basu, Children's National Hospital, Washington, DC, Washington, DC, United States; Kushal Kapse, Developing Brain Institute, Washington, DC, United States; Katherine Ottolini, George Washington University School of Medicine and Health Sciences, Washington, DC, United States; Julius Ngwa, Children's National Health System, Washington, DC, United States; Jonathan Murnick, Children's National Health System, Washington, DC, United States; Adre J. du Plessis, Children's National Health System, Bethesda, MD, United States; Catherine Limperopoulos, Children's National Health System, Washington DC, DC, United States; Rathinaswamy Govindan, Children's National Hospital, Washington, DC, United States

Background: Sleep, particularly active sleep (AS), is vital for early brain development. Frequent sleep disturbances in a stressful NICU environment after premature birth may underlie the increased risk of cognitive deficits in preterm infants. The neuroprotective role of AS is supported by recent findings of a positive association between AS duration% in a cross-sectional window of 29-32 weeks postmenstrual age (PMA) and white matter volume at term-age in preterm infants.

Objective: To investigate the temporal and dose-effect of AS, we tested our hypothesis that AS duration% (%of total sleep time) in weeks preceding the serial second- and third-trimester magnetic resonance imaging (MRI) will be associated with the brain 3-D white matter volume in preterm infants.

Design/Methods: In a prospective cohort of preterm infants born at < 36 weeks PMA, we acquired serial MRIs at 4-week intervals during the extrauterine second- and third-trimester and quantified 3-D brain volumes. We retrieved beat-to-beat heart rate (btb-HR) data from the bedside monitor from admission through NICU discharge. We characterized sleep states throughout the NICU stay using a baseline heart rate shift of >18 beats per minute (bpm) within a 30-second epoch to delineate AS, and epochs with a shift of ≤18 bpm to delineate quiet sleep (QS) (Fig. 1a, validated using video EEG). The AS duration% for each postmenstrual day and the average AS duration% across weeks preceding each MRI were calculated. The relationship between AS duration% within each averaging window of weeks preceding an MRI, and white matter volumes was characterized using a linear-mixed effects model.

Results: Baseline characteristics of 52 preterm infants (born at PMA 28.9 ± 3.2 weeks) included for analysis are summarized in Table 1. AS% increased with advancing PMA (r=0.5, p= 0.04) (Fig. 1b). The linear mixed-effects model demonstrated a positive association between AS duration% in weeks preceding the MRI (p values < 0.05 for preceding 1-5 weeks) and 3-D white matter volumes, adjusted for covariates (Fig. 2). The estimated effect-size of AS duration% on white matter volume increased sequentially and peaked at the preceding 3-week mark (estimate 12 cc, p value < 0.001), suggesting a strong temporal and dose-dependent relationship.

Conclusion(s): Our results showed a temporally linked positive dose-effect of the average % time spent in AS in weeks preceding each MRI on the preterm infant's white matter maturation; generating a hypothesis that optimizing sleep during early postnatal NICU stay may be a neuroprotective target to improve cognitive outcomes in preterm infants.

Table 1. Baseline characteristics of preterm infant cohort


Fig.1. Sleep-state analysis, (a) AS-QS delineation using btb-HR, and (b) Longitudinal Active Sleep Trajectory during Third Trimester


Fig.2. Linear mixed effects estimates for AS duration% in weeks preceding MRI for white matter brain volume
(Footnote): The linear mixed-effect model estimates (adjusted for PMA at AS% calculation and MRI acquisition, sex, respiratory support, cardiotonic and sedative medications, and infection) are shown as bars for each AS% average leading to MRI and 3D-white matter brain volumes. The P-values of the estimates are shown at the top of the bar (negative estimate obtained for week 7 average).