Self-reported sleep difficulties and association with cognitive function in pediatric cancer survivors.

Abstract

10065 Background: Preventing and ameliorating cognitive impairment following treatment for pediatric cancer remains a persistent challenge, with recent literature examining lifestyle factors as possible targets for intervention. An association between sleep and cognitive impairment has been identified in adult cancer patients and survivors, yet information is limited for pediatrics. Thus, the primary goal of our retrospective study was to explore self-reported sleep difficulties in pediatric patients and survivors with mixed cancers. A secondary aim was to explore the impact sleep problems on specific cognitive functions. Methods: Children and adolescents who completed neuropsychological assessments at two large pediatric cancer centers (N = 298) were identified via medical record review. In addition to demographic data, clinically collected cognitive tests included Wechsler Working Memory (WMI), Processing Speed (PSI), and General Ability (GAI) indices. Objective attention measures included Conners Continuous Performance Test (versions II and 3) Omissions (O), Commissions (C), and Reaction Time (RT). Sleep Disturbance (SD) and Sleep-Related Impairment (SRI) were captured using pediatric self-report PROMIS scales. PROMIS cut points were used to categorize sleep issues within the study population and PROMIS scores were compared to reference population means. Multiple linear regression was used to identify relationships between cognitive domains and sleep factors. Common cancer-related clinical and demographic characteristics were controlled as covariates. Results: 21.5% of patients reported severe SD, while 12.8%, 15.9% and 49.7% reported moderate, mild, and within normal limits (WNL) SD, respectively. For SRI, 13.3% reported severe SRI, while 20.0%, 15.9% and 50.8% reported moderate, mild and WNL SRI, respectively. Mean SD and SRI were higher than in the reference populations (p = 0.000 for both measures). Pearson correlation results suggested a moderate relationship between SD and SRI (r = 0.489, p = 0.000). After controlling for covariates, multiple linear regressions indicated significant relationships between SD and GAI (p = 0.006), SD and O (p = 0.018), and SRI and WM (p = 0.006). Conclusions: Results suggested that ~50% of our pediatric cancer patients and survivors reported SD and SRI. SD was correlated with SRI in our sample, and regression analysis indicates a significant relationship between sleep problems and O, WM, and GAI in our sample. Findings from this retrospective study suggest a relationship between sleep problems and cognitive performance in pediatric cancer survivors. As a potentially modifiable factor to improve cognitive function in pediatric cancer survivors, sleep difficulties warrant more research to potentially guide interventions to better manage and improve cognitive function in this population.