Association of Pre-existing Insomnia with Short- and Long-Term Mortality in Patients with Sepsis Admitted to the Intensive Care Unit: A Retrospective Cohort Study Using MIMIC-IV
PURPOSE The association between pre-existing insomnia and mortality in ICU patients with sepsis is uncertain. We assessed mortality at 28, 90, and 365 days. METHODS Adults meeting an operational Sepsis-3 definition in MIMIC-IV version 3.1 were included. Pre-existing insomnia was identified from diagnoses recorded before the index admission. Associations were estimated using adjusted Cox models, propensity score matching, overlap weighting, and doubly robust models. Sleep-disorder subtypes were examined in exploratory 1:1 matched analyses. RESULTS The cohort included 19,609 patients; 403 had pre-existing insomnia and 18,457 had no prior sleep disorder. In the unmatched cohort, adjusted hazard ratios (HRs) for 28-, 90-, and 365-day mortality were 1.24 (95% confidence interval [CI], 1.05–1.48), 1.28 (1.10–1.50), and 1.31 (1.14–1.50). After 1:3 matching, 393 patients with insomnia and 1,179 controls remained. Associations remained for 90-day (HR, 1.19; 95% CI, 1.04–1.36) and 365-day mortality (HR, 1.24; 1.09–1.40), but not for 28-day mortality (HR, 1.13; 0.98–1.32). In exploratory 1:1 matched analyses, 365-day mortality HRs were 1.16 (95% CI, 0.99–1.36; P = 0.073) for insomnia and 1.05 (0.93–1.20; P = 0.420) for sleep apnea. CONCLUSION Pre-existing insomnia was associated with higher 90- and 365-day mortality, but the matched 28-day estimate was not statistically significant. These observational findings do not establish causality.
Authors
- Haixia Wang (ORCID: https://orcid.org/0000-0002-0403-7797)
- Zhenguang Li¹
- Jinbiao Zhang
- Mengyuan Wang
Institutions
- Binzhou Medical University (CN)
- Shandong University (CN)
- Weihai Municipal Hospital (CN)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-19
- DOI
- https://doi.org/10.21203/rs.3.rs-11584831/v1
- Primary Topic
- Sleep and related disorders
- Type
- preprint