Late effects of caffeine use on sleep of infants born prematurely

Abstract

Objective: This study aimed to evaluate whether the therapeutic use of caffeine for premature newborns is associated with changes in sleep habits and the presence of obstructive sleep apnea in childhood. Methods: This is a cross-sectional single-center study in which the caretakers of 87 children aged 5–10 years, born full-term or preterm, treated or not with caffeine in the neonatal period, answered questionnaires to screen for obstructive sleep apnea (Pediatric Obstructive Sleep Apnea Screening Tool [PosaST]) and to characterize the sleep habits (Children's Sleep Habits Questionnaire [CSHQ]) of their children. ANOVA and linear regression tests were performed to verify possible differences between the groups. Results: Children born prematurely who were treated with caffeine woke up significantly later on weekdays than those born at term (09h±00h58 and 07h43±1h15, respectively, p=0.022) and had longer total daily sleep time also compared to those born at term (10h24±1h08 and 09h29±1h08, respectively, p<0.001). There was no significant difference between the three groups in overall PosaST and CSHQ scores. Conclusions: Caffeine use in the neonatal period did not impair sleep habits later in life and did not lead to increased obstructive sleep apnea scores in prematurely born children compared to those born at term.

Caffeine therapy in preterm infants: effects in 2nd and 3rd childhood

“…Oliveira et al 3 also do not present an effect measure for findings with significant differences, which limits understanding of their clinical usefulness. It is necessary to develop a measure of effect in order to expose the relevance of the findings.…”

“…While reading the manuscript "Late effects of caffeine use on sleep of infants born prematurely" by Oliveira et al, 3 volume 42, 2024, methodological gaps were found in the analysis of variables regarding wake-up time and total daily sleep time in premature babies undergoing caffeine therapy in the neonatal period.…”

“…Therefore, we identified in the study by Oliveira et al 3 that the comparison of groups regarding the variables mentioned above is done through variance analysis with Tukey post-hoc, which is considered inadequate due to the imbalance between the groups. This may inflate the type I error due to the post-hoc selection, which should be Hochberg's GT2.…”