Does Obstructive sleep apnea mediate the risk of cognitive impairment by expanding the perivascular space?

Obstructive sleep apnea (OSA) is a neglected global health issue and when left untreated could lead to cognitive impairment (CI), one of the most burdensome outcomes of OSA. Enlarged perivascular spaces (EPVS), an imaging feature as well as a subtype of cerebral small vessel disease and integral part of CSVD, are associated with cognitive function, but the relationship between EPVS and CI is not well understood and by extension the correlation between OSA and EPVS, how CI develops under the joint impact of OSA and EPVS remains unclear. It is the goal of This study to explore the associations among OSA, EPVS, and CI.

This cross-sectional study included 175 older adults with imaging features of EPVS with or without other CSVD subtype features by cranial magnetic resonance imaging between January 2021 and June 2023 at the Shanghai Fifth People's Hospital. We assessed OSA using polysomnography. Blood samples were collected to determine vascular risk factor indices. Cognitive scoring modalities included the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MOCA). To explore the relationship among OSA, EPVS, and CI, we used single-factor analysis, multifactorial analysis, and receiver operating characteristic (ROC) curves.

A total of 136 participants were analyzed. In our statistical process, MMSE showed a more distinguished performance than MoCA. Participants with OSA had greater EPVS burdens in the midbrain (p < 0.001) and hippocampus (p < 0.001) and more serious CI (p = 0.001). OSA positively influenced EPVS in the midbrain (β = 0.052; 95% confidence interval [CI]: 0.006, 0.097; p = 0.026) and hippocampus (β = 0.190, 95% CI: 0.104, 0.275, p < 0.001). Moreover, the apnea-hypopnea index (AHI; β = -0.514; 95% CI: -0.077, -0.031; p < 0.001) negatively affected cognitive e function. With each increase in the AHI by 1 unit, the risk of CI increased by 12.0% (odds ratio = 1.120; 95%CI: 1.062, 1.181; p < 0.001). The AHI (sensitivity, 67.20%; specificity, 92.20%; area under the ROC curve, 0.828; p < 0.001) had a certain degree of accuracy in ruling out CI in the EPVS population, as calculated using the ROC curve.

We identified significant relations among OSA, EPVS, and CI. The AHI is a potential marker for estimating cognitive function in patients with EPVS.