The association between sleep apnea and neurodegenerative disorders: A systematic review and meta‐analysis with an emphasis on precision‐medicine

Abstract Background Previous meta‐analyses that assessed the possible link between sleep apnea (SA) and dementia included cohorts with patients having self‐reported diagnoses of SA as exposures and nonspecific cognitive diagnoses as outcomes, undermining clinical implications. We aimed to assess the association between SA and specific neurodegenerative causes of dementia and to review the use of precision‐medicine tools supporting clinical diagnoses in studies. Method We performed a systematic review and meta‐analysis in conformity with the PRISMA guidelines. Two investigators searched the Web of Science Core Collection databases from inception to March 1 st , 2021. Cohort studies were included if they: 1) used either polysomnography (PSG) or International Classification of Diseases (ICD) codes for SA diagnosis, and 2) measured the risk of all‐cause dementia, mild cognitive impairment, Alzheimer’s disease (AD), vascular dementia (VaD), Parkinson’s disease (PD), Lewy body dementia (LBD), frontotemporal dementia (FTD), and/or mixed dementia. The use of biomarkers to support clinical diagnoses in eligible studies was collected. Studies of cross‐sectional design, that used self‐administered questionnaires for the diagnosis of SA, or that measured solely cognitive scores as outcomes were excluded. Pooled analyses of hazard ratios (HR) were obtained for every type of dementia using a random effects model. Result Among the 1,191 records identified, 8 studies were included, representing a total of 1,249,975 patients. Patients with SA had a 43% increased risk of developing any of the aforementioned types of dementia (HR: 1.43 [1.26‐1.62], 95% CI). They were 34% more likely to develop all‐cause dementia (HR: 1.34 [1.15‐1.57], 95% CI), 28% more likely to develop AD (HR: 1.28 [1.16‐1.41], 95% CI) and 54% more likely to develop PD (HR: 1.54 [1.30‐1.84], 95% CI). No statistically significant association was found with VaD. One study reported a two‐fold increased risk of LBD with SA. No study used biomarkers such as regional cortical atrophy on brain imaging or genetic testing. Results remained consistent after sensitivity analyses. Conclusion While SA appears to be associated with an increased risk of all‐cause dementia, notably for AD, PD and LBD, future studies will have to be conducted using more rigorous precision‐medicine tools in order to better characterize the specific associations between these conditions.