Cortisol levels in former preterm children at school age are predicted by neonatal procedural pain-related stress

Abstract

Early life stress can alter hypothalamic pituitary adrenal (HPA) axis function. Differences in cortisol levels have been found in preterm infants exposed to substantial procedural stress during neonatal intensive care, compared to infants born full-term, but only a few studies investigated whether altered programming of the HPA axis persists past toddler age. Further, there is a dearth of knowledge of what may contribute to these changes in cortisol. This prospective cohort study examined the cortisol profiles in response to the stress of cognitive assessment, as well as the diurnal rhythm of cortisol, in children (n=129) born at varying levels of prematurity (24-32 weeks gestation) and at full-term (38-41 weeks gestation), at age 7 years. Further, we investigated the relationships among cortisol levels and neonatal procedural pain-related stress (controlling for multiple medical confounders), concurrent maternal factors (parenting stress, depressive and anxiety symptoms) and children's behavioral problems. For each aim we investigate acute cortisol response profiles to a cognitive challenge as well as diurnal cortisol patterns at home. We hypothesized that children born very preterm will differ in their pattern of cortisol secretion from children born full-term, possibly depended on concurrent child and maternal factors, and that exposure to neonatal pain-related stress would be associated with altered cortisol secretion in children born very preterm, possibly in a sex-dependent way. Saliva samples were collected from 7-year old children three times during a laboratory visit for assessment of cognitive and executive functions (pretest, mid-test, end-study day acute stress profile) and at four times over two consecutive non-school days at home (i.e. morning, mid-morning, afternoon and bedtime-diurnal rhythm profile). We found that cortisol profiles were similar in preterm and full-term children, albeit preterms had slightly higher cortisol at bedtime compared to full-term children. Importantly, in the preterm group, greater neonatal procedural pain-related stress (adjusted for morphine) was associated with lower cortisol levels on the study day (p=.044) and lower diurnal cortisol at home (p=.023), with effects found primarily in boys. In addition, child attention problems were negatively, and thought problems were positively, associated with the cortisol response during cognitive assessment on the study day in preterm children. Our findings suggest that neonatal pain/stress contributes to altered HPA axis function up to school-age in children born very preterm, and that sex may be an important factor.

Keywords: Child; Cortisol; HPA axis; Internalizing behavior; Low birth weight; Maternal interaction; Pain; Preterm; Sex; Stress.

Figure 1

Figure 1A: Cortisol pattern during cognitive assessment (study day in the laboratory). Salivary cortisol (μg/dl, adjusted for time of testing, mean + S.E.M.) during cognitive assessment (pre-test, mid-test, end) for children born at extremely low gestational age (ELGA, 24–28 weeks), very low gestational age (VLGA, 29–32 weeks) and full-term at 7 years of age. All groups showed an initial drop in cortisol followed by a subsequent increase. Figure 1B: Study day cortisol by number of skin breaking procedures. Figure 1B displays salivary cortisol levels (μg/dl, mean + S.E.M.) from the laboratory visit for cognitive assessment, for children born preterm with either a low (lowest 4th Percentile) or high (highest 4th percentile) number of skin breaking procedures to illustrate the negative association between neonatal pain (corrected for morphine) and diurnal cortisol (p = 0.044) that was found with the neontal GEE model.

Figure 1

Figure 1A: Cortisol pattern during cognitive assessment (study day in the laboratory). Salivary cortisol (μg/dl, adjusted for time of testing, mean + S.E.M.) during cognitive assessment (pre-test, mid-test, end) for children born at extremely low gestational age (ELGA, 24–28 weeks), very low gestational age (VLGA, 29–32 weeks) and full-term at 7 years of age. All groups showed an initial drop in cortisol followed by a subsequent increase. Figure 1B: Study day cortisol by number of skin breaking procedures. Figure 1B displays salivary cortisol levels (μg/dl, mean + S.E.M.) from the laboratory visit for cognitive assessment, for children born preterm with either a low (lowest 4th Percentile) or high (highest 4th percentile) number of skin breaking procedures to illustrate the negative association between neonatal pain (corrected for morphine) and diurnal cortisol (p = 0.044) that was found with the neontal GEE model.

Figure 2

Figure 2A: Diurnal cortisol pattern (at home) Diurnal salivary cortisol (μg/dl, mean + S.E.M.), collected at home and averaged across two non-school days, for children born at extremely low gestational age (ELGA, 24–28 weeks), very low gestational age (VLGA, 29–32 weeks) and full-term, at 7 years of age. All groups showed the expected drop in cortisol levels throughout the day. Figure 2B: Diurnal cortisol by number of skin breaking procedures. Figure 2B displays diurnal salivary cortisol levels (μg/dl, mean+S.E.M.) for children born preterm with either a low (lowest 4^th percentile) or high (higest 4^th percentile) number of skin breaking procedures to illustrate the negative association between neonatal pain (corrected for morphine) and diurnal cortisol (p = 0.023) that was found with the neonatal GEE model.

Figure 2

Figure 2A: Diurnal cortisol pattern (at home) Diurnal salivary cortisol (μg/dl, mean + S.E.M.), collected at home and averaged across two non-school days, for children born at extremely low gestational age (ELGA, 24–28 weeks), very low gestational age (VLGA, 29–32 weeks) and full-term, at 7 years of age. All groups showed the expected drop in cortisol levels throughout the day. Figure 2B: Diurnal cortisol by number of skin breaking procedures. Figure 2B displays diurnal salivary cortisol levels (μg/dl, mean+S.E.M.) for children born preterm with either a low (lowest 4^th percentile) or high (higest 4^th percentile) number of skin breaking procedures to illustrate the negative association between neonatal pain (corrected for morphine) and diurnal cortisol (p = 0.023) that was found with the neonatal GEE model.