Predictors of Insomnia Severity Index Profiles in United States Veterans With Obstructive Sleep Apnea

The Insomnia Severity Index (ISI) has been used to define insomnia symptoms in individuals with obstructive sleep apnea (OSA). However, whether distinct ISI profiles exist in individuals with OSA is unclear. The aims of this study were to determine (1) empirically-based ISI profiles in veterans with OSA and (2) predictors of these ISI profiles.

Participants were 630 veterans with a new diagnosis of OSA over a 12-month period. Individuals completed the ISI and other questionnaires on the polysomnography (PSG) night. Latent profile analysis was performed to detect ISI subgroups based on individual ISI items. Age, Charlson Comorbidity Index, apnea-hypopnea index (AHI), mood disorder, posttraumatic stress disorder, and chronic pain diagnoses were used to predict between ISI profiles.

Latent profile analysis identified five ISI subgroups in veterans with OSA. The "asymptomatic" group (12% prevalence) had low scores across all ISI items. The "moderate insomnia" (30% prevalence) and "severe insomnia" (44% prevalence) groups had elevated scores for all ISI items but differing in severity. Last, the "moderate" (6% prevalence) and "severe daytime symptoms" groups (8% prevalence) were characterized by absence of nocturnal complaints but high scores on daytime impairment items. Age, AHI, mood disorder, posttraumatic stress disorder and chronic pain diagnoses discriminated between ISI profiles.

We describe data-driven ISI profiles in veterans with OSA. Older age was associated with lower insomnia and daytime symptom complaints whereas psychological comorbidities were related to more severe insomnia. Caution should be used in interpreting the ISI score in individuals with OSA because a subset had elevated total scores without insomnia.