Abstract Introduction Obstructive sleep apnea (OSA) involves repeated airway obstruction during sleep that causes intermittent hypoxemia and respiratory-effort-related sleep fragmentation. Glymphatic transport, a brain waste clearance process driven by cerebrospinal fluid (CSF) movement, depends on stable vascular pulsations and their coordination with CSF flow, both of which may be disrupted by hypoxic vascular injury and fragmented sleep. However, the association between OSA and glymphatic transport, and whether hypoxic burden or sleep fragmentation drives this relationship, remains unknown. Methods This study examined the association of OSA diagnosis, hypoxic burden, and respiratory-effort-related sleep fragmentation with CSF mobility in 32 middle-aged and older adults. OSA was assessed via in-lab polysomnography or WatchPAT home testing. Hypoxic burden was quantified as T90 (i.e., time spent under 90% oxygen saturation). To examine distinct pathways, sleep fragmentation was defined as non-hypoxic respiratory effort related arousals (NH-RERAs), computed as the respiratory disturbance index minus the apnea hypopnea index using the 4% criterion (AHI4%). CSF mobility was defined as gBOLD-CSF coupling, which represent anti-coupling strength between gray matter blood oxygen dependent signals, a measure of cerebral blood flow, and ventricular CSF flow. gBOLD-CSF coupling was assessed using a 9 to 15 minute resting-state fMRI scan (TR=1.5 seconds, voxel size=1.5mm3) collected on a 7T scanner. A hierarchical approach was used to model gBOLD-CSF coupling as a linear function of T90, NH-RERAs, age, and sex. Results Mean age was 66 (SD=6.32), 59% were female, and 63% identified as White. Seventy percent had moderate to severe OSA (AHI4%≥15). T90 averaged 23.4 minutes (SD=48.4), and NH-RERAs averaged 6.9 events/hour (SD=6.7). Moderate to severe OSA was associated with lower gBOLD-CSF coupling (ΔgBOLD-CSF=0.25,p=0.002). In hierarchical models, longer T90 predicted reduced gBOLD-CSF coupling (β=0.002,p=0.03) and T90 explained unique variance beyond NH-RERAs and covariates (ΔR2=0.15,F=5.03,p=0.03). NH-RERAs were not associated with gBOLD-CSF coupling. Posthoc analyses found that lower mean oxygen saturation was likewise associated with reduced gBOLD-CSF coupling (β=-0.05,p=0.02). Conclusion OSA may be a risk factor for deficits in CSF mobility needed for brain waste clearance, with hypoxic burden rather than respiratory effort related sleep fragmentation as the primary driver. Support (if any) This study was supported by the NIH (R01AG066870) and the Alzheimer’s Association (AARFD-24-1306796).
Cho et al. (Fri,) studied this question.