466 - Neonatal sleep scoring based on visual cues in extremely premature infants

Poster Number: 466
Publication Number: 466.342

Andie Kwon, Division of Autonomic Medicine, Department of Pediatrics, Ann & Robert H. Lurie Children’s Hospital of Chicago and Stanley Manne Children’s Research Institute, Aurora, IL, United States; Erin Lonergan, Ann & Robert H. Lurie Children's Hospital of Chicago, Palos Heights, IL, United States; Brad S. Hopkins, Lurie Children’s Hospital, Douglas, MI, United States; Casey Rand, Ann & Robert H. Lurie Children's Hospital of Chicago, chicago, IL, United States; Narayanan Krishnamurthi, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States; Allaa Fadl-Alla, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States; Susan M. Slattery, Ann & Robert H. Lurie Children's Hospital, Chicago, IL, United States; Debra E. Weese-Mayer, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States

Background: Sleep is essential to early development, including maturation of the neuronal network. Recent research indicates that emergence of quiet sleep (QS) stages in preterm infants is correlated with postnatal neural maturation. One of the developmental changes seen during the first year of life is an increase in QS. Though changes in sleep duration/structure may serve as an early, non-invasive marker of brain development, literature is sparse on accurate neonatal sleep scoring methodology and lacks a standardized scoring system.

Objective: To develop sleep scoring methods in a cohort of preterm-born neonates and achieve high inter-rater reliability (IRR).

Design/Methods: We enrolled infants born between 24 and 28 6/7 weeks gestational age (wGA) for bedside recording at 36 weeks postmenstrual age (wPMA) in the NHLBI-funded Pre-Vent study. Infants with congenital anomalies, imminent death, or family circumstances that might preclude reliable follow-up were excluded. Respiratory inductance plethysmography (chest, abdomen), electroencephalogram, and electrocardiogram waveforms, and visual were continuously recorded for up to four hours. Thoman’s sleep behavioral sleep scale and criteria were applied (well-documented validity and versatility in broad infant age range). Each 30-second epoch was scored as quiet sleep (QS), active sleep (AS), wakeful state, or caregiving, with files independently scored by two trained individuals. IRR was calculated based on the agreement for 4046/4334 epochs (remaining 288 epochs resolved when individual scorers met to review). All QS:AS ratio were then calculated, along with % time in QS and AS.

Results: The cohort included 26 infants (Table 1). A median of 134 epochs were scored (IQR: 126-179) (range of 82 to 399 30-second epochs). Infants spent a median of 32.4% (IQR: 19.2-51.5) and 50.9% (IQR: 39.7-64.7) in QS and AS, respectively. The QS:AS ratio range was 0.03 to 2.94, with a median of 0.6 (IQR: 0.36-1.0). The IRR median was 96.2% (IQR: 91.7-98.3%). IRB #2017-1178 NIH grant support: 5U01HL133704 (Northwestern University) 5U01HL 133708 (LDCC; University of Virginia)Conclusion(s): In our cohort of preterm neonates, this methodology empowered non-invasive sleep scoring with strong IRR. The broad QS:AS range warrants further investigation in a larger cohort over an extended PMA range. While this cohort is not directly comparable to previously published studies, overall comparison of %QS calculations was similar. The difficulty in comparison demonstrates the essential need for a methodical system that establishes a transferrable, objective neonatal sleep state training.
Andie Kwon Medical Student CVAndie Kwon CV.pdf