Mortality in US veterans with insomnia, sleep apnea, and comorbid insomnia and sleep apnea (COMISA); an exploratory,...

This study examined the association of insomnia, sleep apnea, and comorbid insomnia and sleep apnea (COMISA) with all-cause mortality in a large Veteran sleep clinic cohort.

We retrospectively identified Veterans with insomnia and/or SA using ICD-9/10 codes, requiring two diagnoses within 13 months (≥ 30 days apart). Insomnia cases also required relevant prescriptions and/or cognitive behavioral therapy for insomnia (CBT-I); SA cases required positive airway pressure prescriptions within three years. Patients were classified as insomnia only (INS), SA only, or COMISA. Adjusted odds ratios (aOR) for mortality were estimated using SA as the reference. Model 1 adjusted for age, sex, race, ethnicity, and Charlson Comorbidity Index (CCI); Model 2 additionally adjusted for trazodone and quetiapine; Model 3 additionally adjusted for benzodiazepines and benzodiazepine receptor agonists. Model 4 included all Model 3 covariates plus CBT-I.

Among the 1,720,090 patients, 47.6% had SA (mean age 56.3 ± 13.7; 93.3% male; 69% White; 26.2% with CCI > 2), 41.1% had INS (mean age 57.0 ± 16.8; 87.1% male; 70.3% White; 25.4% with CCI > 2), and 11.2% had COMISA (mean age 51.9 ± 13.2; 89.4% male; 66.7% White; 19.6% with CCI > 2). Compared with SA, unadjusted ORs for mortality were 1.75 (95% CI: 1.73-1.76) for INS and 0.68 (95% CI: 0.67-0.69) for COMISA. Fully adjusted aORs were 1.51 (95% CI: 1.49-1.52) for INS and 0.79 (95% CI: 0.78-0.80) for COMISA.

Among the Veterans referred to sleep clinics, insomnia alone was associated with the highest mortality, followed by SA, and the lowest in COMISA. These findings contrast with published studies and need to be seen considering the limitations of our approach, including the possibility that lack of a pre-specified hypothesis increases the risk of spurious findings.

Insomnia and sleep apnea commonly coexist as comorbid insomnia and sleep apnea (COMISA), which has been linked to increased mortality in population-based studies. Evidence from clinical cohorts, particularly among Veterans, remains limited.

In this exploratory, hypothesis-generating large cohort of Veterans referred to sleep clinics, insomnia alone was associated with the highest mortality, followed by sleep apnea, and the lowest in COMISA. These results contrast with prior population studies and should be interpreted cautiously given study limitations (including possible chance findings without a pre-specified hypothesis), warranting confirmation.