BACKGROUND: Obstructive sleep apnea (OSA) is a highly prevalent disorder associated with increased risk of Alzheimer's disease (AD) and cognitive decline. The mechanisms underlying the relationship between OSA and disease progression remain undefined, but may involve impairment in the glymphatic system, a perivascular network responsible for cerebrospinal fluid and interstitial fluid exchange and waste clearance. This study evaluated MRI-visible perivascular spaces (PVS) and parenchymal CSF mapping (pCSF) as non-invasive proxies of glymphatic function in untreated and CPAP-treated OSA relative to healthy controls. METHODS: Forty-two adults (n = 16 healthy controls, n = 14 untreated OSA, n = 12 CPAP-treated OSA) were retrospectively evaluated. Participants underwent MRI and cognitive testing, and a sub-cohort (n = 25) received 11C-PIB amyloid PET. Enhanced PVS contrast (EPC) mapping segmentation quantified MRI-visible PVS, while multi-echo FAST-T2 MR derived pCSF mapping provided voxelwise quantification of glymphatic fluid distribution. Between-group differences were assessed using nonparametric statistics and multivariable regression analyses controlled for age and sex. RESULTS: Segmented MR-visible PVS did not differ between groups. Comparatively, untreated OSA subjects had higher ADmask pCSF versus than controls (p = 0.0155), with exploratory analyses displaying a similar relationship across regions including cerebral grey and white matter. Intriguingly, CPAP-treated individuals exhibited pCSF levels statistically comparable to controls. In the sub-cohort, higher ADmask pCSF independently predicted greater amyloid burden on PET (β = 0.616, p = 0.007). CONCLUSIONS: Untreated OSA is associated with higher parenchymal glymphatic fluid burden consistent with impaired perivascular fluid dynamics. CPAP-treated participants exhibited pCSF levels comparable to controls. pCSF mapping represents a sensitive biomarker for evaluating perivascular fluid distribution, a positive association with amyloid deposition, and monitoring potential treatment-related differences in OSA.
