Middle-aged and older adults at high risk for obstructive sleep apnea demonstrate approximately 40% higher odds of concurrent mental health conditions and 44% increased odds of developing new psychiatric disorders over time, according to findings from a secondary analysis of the Canadian Longitudinal Study on Aging (CLSA) published in JAMA Network Open.

The study led by Tetyana Kendzerska, MD, PhD, of the Ottawa Hospital Research Institute and University of Ottawa, examined approximately 30,000 patients at baseline (median age, 62 years; 51% women) and nearly 28,000 at follow-up over a median of 3 years. Using the validated STOP questionnaire (assessing snoring, daytime somnolence, witnessed apnea, and hypertension), researchers classified 24% of participants at baseline and 27% at follow-up as high risk for obstructive sleep apnea (OSA).

Principal Findings

In adjusted multivariable analyses, high OSA risk was associated with significantly elevated odds of a composite poor mental health outcome—defined as Center for Epidemiologic Studies Short Depression Scale score of 10 or greater, Kessler Psychological Distress Scale score of 20 or greater, self-reported physician-diagnosed mental health condition, or antidepressant use.

Cross-sectional analyses demonstrated nearly identical associations at both time points, with high OSA risk corresponding to approximately 1.4 times the odds of the composite mental health outcome at both baseline and follow-up. The repeated-measures mixed-effects analysis found high OSA risk corresponded to 1.44 times the odds of the composite mental health outcome.

Among patients without mental health conditions at baseline (n = 19,990), high OSA risk predicted 1.2 times the odds of developing new mental health conditions at follow-up. Among this cohort, 7% developed new mental health conditions.

Disorder-Specific Associations

Secondary analyses examining individual psychiatric outcomes revealed the strongest associations with mood disorders and clinical depression. In the repeated-measures analysis, high OSA risk corresponded to:

- 1.48 times the odds of clinical depression
- 1.46 times the odds of mood disorder
- 1.48 times the odds of anxiety disorder.

Among those free of each specific condition at baseline, longitudinal associations strengthened considerably, with high OSA risk corresponding to 1.71 times the odds of clinical depression and 1.65 times the odds of mood disorder. The anxiety disorder association did not reach statistical significance in the longitudinal analysis.

Witnessed apnea during sleep—a secondary exposure with higher specificity but lower sensitivity compared with the STOP questionnaire—produced similar findings, with 1.47 times the odds of the composite outcome, 1.51 times the odds of clinical depression, 1.50 times the odds of mood disorder, and 1.36 times the odds of anxiety disorder in repeated-measures analyses.

Risk Factor Identification

Exploratory analyses identified characteristics associated with new mental health conditions among patients at high OSA risk without baseline psychiatric disorders (n = 3,213). Among this subgroup, 11% developed new composite mental health conditions at follow-up.

Female sex, low household income (less than $50,000), dissatisfaction with life, fair self-rated general health (vs excellent), and comorbid sleep disorders including restless legs syndrome, acting out dreams while asleep, and insomnia symptoms were associated with higher odds of new psychiatric outcomes.

In the repeated-measures analysis of all high-risk OSA patients regardless of baseline mental health status, additional risk factors emerged: younger age, apartment dwelling (vs house), abstinence from alcohol, lower body mass index, respiratory problems, traumatic brain injury, pain (with dose-response relationship for severity), and polypharmacy.

Mechanistic Considerations

The researchers posited several biological pathways linking untreated OSA with psychiatric morbidity. OSA-related intermittent hypoxemia may disrupt neural systems involved in mood regulation, while sleep fragmentation could alter neuroendocrine pathways. The condition is also associated with elevated inflammatory markers, which may contribute to depression. Additionally, OSA-related cardiometabolic comorbidities may elevate mental distress.

The findings align with prior research demonstrating lower odds of association among older adults compared with younger populations, suggesting that while OSA may contribute to poor mental health in this demographic, other factors also play a role.

Clinical Implications

The researchers emphasized the importance of systematic mental health screening for older adults at risk for OSA. They recommended incorporating mental health assessment tools into sleep evaluations to identify patients at greatest risk and support early intervention.

Modifiable factors identified—including respiratory problems, pain, comorbid sleep disorders, and polypharmacy—may represent potential intervention targets for improving mental health outcomes in this population. Sensitivity analyses suggested potentially stronger associations between OSA risk and composite mental health outcomes among women compared with men, though interaction terms did not reach statistical significance.

Methodological Considerations

The study employed the STOP questionnaire, which demonstrates sensitivity of 87% to 90% and specificity of 29% to 42% for any OSA (apnea-hypopnea index of five or greater), with negative predictive value typically exceeding 80% for moderate to severe OSA. Dr. Kendzerska and colleagues acknowledged this balance may have reduced misclassification and likely biased results toward the null, producing conservative effect estimates.

Models adjusted for 19 covariates at baseline and 17 at follow-up, including sociodemographic factors, lifestyle measures, medical conditions, and sleep-related symptoms beyond OSA. Variables were selected through a step-down procedure to obtain the lowest stable Akaike information criterion value.

Study Limitations

The researchers acknowledged several limitations inherent to observational design, including potential residual confounding and the possibility of reverse causality—for example, depression contributing to OSA risk via weight gain or medication effects. OSA risk and mental health diagnoses were self-reported without polysomnographic confirmation.

The CLSA cohort was not fully representative of the Canadian population; participants were predominantly White (94%), healthier, more educated, and from urban areas (83%), limiting generalizability. Treatment data were unavailable, precluding assessment of whether OSA therapy modifies psychiatric risk.

 Disclosures can be found in the study.

Source: JAMA Network Open