Abstract
Background
The association between Obstructive Sleep Apnea (OSA) and cognition in older‐adults is conflicting. We determined to characterize older OSA‐patients using the NIA‐AA Research Framework and test whether OSA accelerates cognitive decline in preclinical Alzheimer’s disease (AD)
Method
Community‐dwelling cognitively normal elderly participating on a study on memory, sleep and aging, with baseline AD biomarker data and both baseline and follow‐up neuropsychological data were included. OSA was defined using AHI4%. Data‐driven, clinically relevant thresholds for CSF‐Aβ42 (≤375pg/ml), and CSF P‐tau (≥53.7pg/ml) indicated OSA participants AT(N) status using the NIA‐AA Research Framework. Twenty‐four participants with non‐AD pathologic change defined as A‐T+ were excluded leaving 127 for the analysis. Main outcome was the annual rate‐of‐change in global cognition (calculated as an average composite Z‐score of episodic memory, language and executive function cognitive tests domains). Linear mixed‐effects models with random intercept and slope were used to assess associations between AT(N) characterized OSA subjects, and longitudinal changes in global cognition, controlling for age‐at‐baseline, sex, APOE4‐status, years‐of‐education, and their interactions with time.
Result
Of the 127 participants, 81 (63.8%) were women. The mean (SD) age was 69.6 (7.3) years and follow‐up time was 2.46 (0.64) years. Eight (6.3%) met biomarker criteria for AD (OSA+/A+/T+ [n=4] and OSA‐/A+/T+ [n=4]). Forty‐four participants (34.6%) were amyloid positive (OSA+/A+ [n=20] and OSA‐/A+ [n=24]). Sixteen (12.6%) were OSA+/A+/T‐, and 20 (15.7%) were OSA‐/A+/T‐. Eighty‐three (63.4%) had normal AD biomarkers (OSA+/A‐/T‐ [n=40] and OSA‐/A‐/T‐ [N=43]). Regardless of OSA status, relative to normal AD biomarkers, amyloid positive participants i.e. AD pathologic change, showed significant faster rate‐of‐decline in global cognition (β = −0.066, 95%CI, −0.088, −0.046; P < .001). However, OSA+/A+/T‐ participants showed even more significantly faster rate‐of‐decline in global cognition compared to OSA‐/A+/T‐ participants (β = −0.042, 95%CI, −0.063, −0.019; P < .001), suggesting an OSA/Aβ42 synergism independent of PTau. OSA+/A‐/T‐ or OSA‐/A‐/T‐ participants (normal AD biomarkers) did not show any significant cognitive change over time.
Conclusion
Among amyloid positive healthy‐elderly, OSA and Aβ demonstrate synergism related to cognitive decline that might be independent of tau deposition. Clinical trials in a population of elderly OSA cognitive‐normal individuals should target at minimum persons with Alzheimer’s pathologic change.