Abstract Introduction Substance use is a growing public health concern with important implications for sleep health. However, the extent to which substance use–related disorders are associated with the subsequent development of sleep disorders remains insufficiently characterized. This study aimed to evaluate that relationship using a large multicenter electronic health record (EHR) dataset. Methods This retrospective multicenter cohort study utilized a de-identified EHR database spanning 20 years from 64 U.S. healthcare organizations. Adults with and without documented substance use–related disorders who had outpatient encounters and no prior sleep disorder diagnoses were included. Propensity score matching was performed 1:1 for demographics, comorbidities, and commonly used substances such as tobacco. Relative risks (RRs) were calculated for the development of any sleep disorder and specific subtypes, including insomnia, hypersomnia, circadian rhythm sleep disorders, and narcolepsy or cataplexy. Analyses were stratified by age ( 60 vs ≥60 years). Results The substance use–related disorder cohort included 580,004 individuals, while the comparison group included 4,825,877 individuals. Following propensity score matching (n = 403,891 per group), individuals with substance use–related disorders demonstrated a significantly higher risk of developing any sleep disorder (RR 1.29; 95% CI, 1.27–1.32). Subtype analyses showed that these individuals were more likely to develop insomnia (RR 1.14; 95% CI, 1.68–1.80), hypersomnia (RR 1.14; 95% CI, 1.07–1.22), circadian rhythm sleep disorders (RR 1.23; 95% CI, 1.02–1.49), and narcolepsy or cataplexy (RR 1.46; 95% CI, 1.05–2.02). These associations remained consistent across both age strata. Conclusion In this large EHR-based cohort, substance use–related disorders were associated with a significantly increased risk of developing sleep disorders across multiple diagnostic categories. These findings underscore the importance of recognizing sleep health vulnerabilities in this population and highlight the need for prospective studies to explore underlying mechanisms and dose-response relationships. Support (if any)
Pande et al. (Fri,) studied this question.