Longitudinal medial temporal lobe gray matter volume and cognitive changes in late middle‐aged and older adults with treated and non‐treated obstructive sleep apnea

Abstract Background Neuroimaging studies in obstructive sleep apnea (OSA) have found both gray matter atrophy and hypertrophy in medial temporal lobe subregions, the latter probably reflecting edema. Whether and how these changes progress over time when OSA is treated or untreated remains unclear. Here, we investigated gray matter volume changes in medial temporal lobe subregions in treated OSA, untreated OSA and control participants aged > 55 years. We also explored whether the medial temporal lobe volume at baseline could predict cognitive decline. Method Twenty‐three non‐treated OSA participants (65.8 ±5.4 years) and twenty controls (64.0 ±6.4 years) were evaluated with overnight polysomnography, MRI session and neuropsychological assessment. Twelve OSA participants then started continuous positive airway pressure (CPAP) therapy and used it at least 4 hours per night, 4 nights per week. Eleven participants refused treatment. Participants were evaluated again with MRI and neuropsychology after 18 months. We extracted the entorhinal, hippocampal and parahippocampal volumes using FreeSurfer 7.1. We performed repeated‐measure ANOVAs with Group (treated OSA, untreated OSA and controls) and Time (baseline and follow‐up) on bilateral volumes. We also performed partial correlations between bilateral medial temporal volumes at baseline and changes in cognition (Montreal Cognitive Assessment, score at follow‐up – baseline) in each group separately. Analyses were corrected for age, sex, and total intracranial volume. Result We found a significant Group X Time interaction for bilateral parahippocampal volumes (F = 7.2; p = 0.002) where treated OSA participants showed a decreased volume over time (p tukey<0.001), while no changes occurred in the two other groups. Regarding the predictive value of medial temporal lobe volume, in untreated participants, higher bilateral hippocampal volume at baseline was associated with more severe cognitive decline over time (r = ‐0.77; p = 0.01). In treated participants, higher bilateral entorhinal cortex volume at baseline was associated with more cognitive decline (r = ‐0.74; p = 0.035). No associations were found in controls. Conclusion These preliminary results showed that presenting hypertrophy of the medial temporal lobe subregions at baseline is associated with worse cognitive outcome at 18‐month follow‐up in OSA participants. In the CPAP group, the decrease in parahippocampal volume might represent a reduction in hypertrophy, possibly by limiting edema.