Neurocognitive Outcomes at 10 Years of Age in Extremely Preterm Newborns with Late-Onset Bacteremia

Abstract

Objective: To evaluate the difference in 10-year neurocognitive outcomes between extremely low gestational age newborns without bacteremia and those with suspected or confirmed late-onset bacteremia.

Study design: Neurocognitive function was evaluated at 10 years of age in 889 children born at <28 weeks of gestation and followed from birth. Definite (culture-positive) late-onset bacteremia during postnatal weeks 2-4 was identified in 223 children, and 129 children had suspected bacteremia.

Results: Infants with the lowest gestational age and birth weight z-score had the highest prevalence of definite and suspected late-onset bacteremia. Compared with peers with no or suspected bacteremia, infants with definite bacteremia performed worse on tests of general cognitive ability, language, academic achievement, and executive function, even after adjustment for potential confounders. Adjustment for low IQ attenuated the associations between bacteremia and all dysfunctions at age 10 years. Children with suspected bacteremia did not differ appreciably from those with no evidence of bacteremia. The motor domain was unaffected.

Conclusions: Extremely low gestational age newborns who had definite late bacteremia during postnatal weeks 2-4 are at heightened risk of neurocognitive limitations at age 10 years.

Keywords: academic function; cognition; executive function; extremely low gestational age newborn; neurodevelopment; school age; sensorimotor impairment; sepsis.

Figure 1

Forest plots of odds ratios (ORs) and 95% confidence intervals of a Z-score ≤ −1 on each DAS-II and NEPSY-II neurocognitive assessment at age 10 associated with definite bacteremia and suspected bacteremia. Odds ratios in the top panel are adjusted for Black race, maternal education (≤ 12 years), public insurance, gestational age (23–24 and 25–26 weeks) and birth weight Z-score (<−2 and ≥ −2, < −1). Odds ratios in the bottom panel are also adjusted for DAS IQ ≤ −1. DAS Verbal and Nonverbal Reasoning do not appear in the bottom panel since DAS IQ is the average of the two. Models in both panels take into account the correlation between children from the same pregnancy.