P0758 - Salivary Cortisol Correlates with Electrocorticol Power on High Density EEG of Hospitalized Preterm Infants

Background: The hypothalamic-pituitary-adrenal (HPA) axis mediates hormone production in response to stress beginning in the second trimester of gestation. Salivary cortisol levels reflect HPA activity and is a commonly used biomarker of stress response. Both prenatal and early childhood stress exposure are associated with altered neurodevelopment. Preterm infants spend a critical period of development, equivalent to the fetal third trimester, in the neonatal intensive care unit (NICU), a potentially stressful environment. High density electroencphalography (HD EEG) can serve as an early marker of neurodevelopment in preterm infants.

Objective: To quantify the relationship between salivary cortisol and HD EEG in hospitalized preterm infants participating in the NICU-HEALTH cohort study.

Design/

Methods: NICU-HEALTH enrolls infants born between 28 and 33 weeks gestation at birth. Among the study procedures, paired saliva specimens are collected weekly for cortisol measurement and a 30 minute sleep-coded 32-lead HD EEG is completed in the week prior to discharge. Of enrolled NICU-HEALTH participants, 73 had at least one analyzable saliva sample collected in the week prior to an acceptable EEG study. HD EEGs were screened for artifact and EEG power was computed per minute and averaged over sleep states. The log transformed EEG frequencies were considered both individually, and as averaged groups of three (e.g. 0-3 Hz, 4-6 Hz, etc.). Unadjusted Pearson correlations were constructed for each EEG frequency against the log of minimum salivary cortisol concentration. Generalized linear models of EEG frequency group against the log of minimum salivary cortisol were subsequently adjusted for the significant covariates sex, gestational age, race (white/non-white), insurance type (Medicaid or private) as a surrogate of socioeconomic status, postmenstrual age (PMA) at time of EEG acquisition, and first blood gas base excess after birth as a surrogate measure of degree of illness at birth. All analyses were then stratified by sex and repeated.

Results: In this preliminary analysis of 73 participants, univariable analysis was significant (P < 0.05) for females in frequency 2 and spanning frequencies 5 - 21 Hz with effect size ranging from 0.35 to 0.41. Pearson correlation analyses indicated that the log of salivary cortisol concentration (nmol/L) was significantly correlated with HD EEG power by frequency in male and female infants but not the same frequency ranges (Figure 1). In adjusted generalized linear models, salivary cortisol was significantly associated with EEG power in alpha and beta frequency ranges among female infants and in the beta frequency range in males (Table 1).

Conclusion: Salivary cortisol was associated with HD-EEG electrocortical power among girls at alpha and beta frequencies in this small population of preterm infants. This implies that NICU-based stress may impact preterm infant neurodevelopment in a sexually dimorphic manner.