Rapid eye movement
sleep reduction in pediatric epilepsy: A propensity score‐matched study
Alen Juginović, Laura Rodman Abstract
Objective
This study was undertaken to characterize sleep architecture assessed by polysomnography (PSG) in children with epilepsy versus matched nonepilepsy clinical comparators and disentangle disease from antiseizure medication (ASM) effects.
Methods
In this cross‐sectional analysis of the Nationwide Children's Hospital Sleep DataBank (3647 PSG studies from 3392 patients aged ≤18 years), a clinical cohort referred for PSG, we compared 560 PSG studies in children with epilepsy to 3087 studies in nonepilepsy clinical comparators. Propensity score matching (1:1) on age, sex, and body mass index percentile yielded 560 pairs. Fifteen PSG outcomes were compared using Wilcoxon signed‐rank tests with Hedges g and false discovery rate (FDR) correction. A secondary analysis compared 208 matched pairs of on‐ versus off‐ASM epilepsy patients. Seven sensitivity analyses addressed age subgroups, medication exclusions, and comorbidity matching.
Results
Children with epilepsy had reduced rapid eye movement (REM) sleep (median 18.2% vs. 20.9%, g = −.32 [95% confidence interval = −.41 to −.24], FDR p < .001), lower sleep efficiency ( g = −.14, p = .007), shorter total sleep time ( g = −.12, p = .015), prolonged REM latency (g = .14, p = .029), and lower arousal index ( g = −.12, p = .013). ASM use amplified the REM deficit (15.8% vs. 19.8%, g = −.48, p < .001). The effect followed a developmental gradient: strongest at age < 6 years ( g = −.38, p < .001), intermediate at 6–12 years ( g = −.31, p = .002), and not detected at >12 years ( g = −.07, p = .953). Excluding benzodiazepine users attenuated the signal ( g = −.24), whereas excluding melatonin users preserved it ( g = −.29).
Significance
In the largest pediatric epilepsy PSG study to date, epilepsy is associated with an REM‐predominant sleep disruption concentrated in children younger than 6 years and amplified by ASMs, although the independent contribution of epilepsy itself remains uncertain. These cross‐sectional findings identify early childhood as a period of heightened vulnerability that warrants prospective study.