Which aspects of sleep architecture and respiratory-related sleep disturbances are associated with cognitive function in middle-aged to older adults?

In a recent study published in JAMA Network Open, researchers created the Sleep and Dementia Consortium, comprising five population-based cohort studies.

These studies were the Atherosclerosis Risk in Communities (ARIC) study, Osteoporotic Fractures in Men Study (MrOS), Framingham Heart Study (FHS), Cardiovascular Health Study (CHS), and Study of Osteoporotic Fractures (SOF) to evaluate the association between sleep architecture, obstructive sleep apnea (OSA) and cognitive function.

Background

There is an overall shortage of data on associations between poor sleep and dementia. The Lancet Commission on Dementia Prevention did not include poor sleep in 12 modifiable risk factors for dementia.  

There are many possible mechanisms through which sleep could decrease the dementia risk. For instance, adequate and good quality sleep augments the glymphatic clearance of Alzheimer’s disease proteins, facilitating memory consolidation and synaptic remodeling. 

It reduces the risk of cardiometabolic diseases and vascular brain injury, both well-recognized factors associated with dementia risk. Thus, advancements in sleep research could inform novel dementia prevention strategies.

About the study

The Sleep and Dementia Consortium performed overnight, home-based polysomnography (PSG) and neurocognitive assessments to investigate associations between baseline sleep metrics and cognitive function measured in the subsequent five years.

All study participants were at least 45 years old and had never had dementia or stroke. The standardized protocol for sleep assessments recorded readings from an electroencephalogram, electrooculogram, finger pulse oximeter, a single bipolar electrocardiogram, and ambient light levels, to name a few.

Trained PSG technicians scored sleep in 30-second epochs per established guidelines but remained blinded to data. In addition, the team measured sleep macroarchitecture, including total sleep time and sleep maintenance efficiency. Likewise, they used an eight-item questionnaire, 'The Epworth Sleepiness Scale' to measure daytime sleepiness, with higher scores indicating higher daytime sleepiness.

Assessments of obstructive sleep apneas (OSAs), and hypopneas numbers accompanied by >30% reduction in airflow and ≥4% oxygen desaturation helped the researchers compute the apnea-hypopnea index (AHI). 

While global cognition was the primary study outcome, the team explored cognitive domain scores as secondary outcomes, and methods used for each varied across cohorts. So, the researchers performed a principal component analysis (PCA) of all cognitive test scores to derive a general cognitive score.

The team used a neuropsychological framework in the secondary analysis to group cognitive tasks into broader cognitive domains. These were attention, executive function, processing speed, verbal memory, learning, language, and visuospatial function. The study had multiple covariates, e.g., age, gender, education status, etc., which the team assessed at PSG time or the clinical examination closest to the PSG.

Finally, the researchers explored the associations between each sleep variable and cognitive outcome using a linear regression model that adjusted for covariates. They also performed cohort-specific analyses.

Results

Overall, there were 5946 participants, of which 1875 (31.5%) were females with a median age between 58 and 89 years at PSG acquisition. Cohort-wise, there were 1791, 701, 640, 2619, and 195 participants from ARIC, CHS, FHS, MrOS, and SOF studies. 

The proportion of people with mild OSA (AHI ≥5) ranged between 45.2% and 63.9% in the FHS and MrOS, showing age and sex variations in these cohorts. The percentage of people with moderate OSA was lowest in the FHS and highest in MrOS (16.9% vs. 28.9%). Similarly, median sleep efficiency was the highest and lowest among participants in the FHS and MrOS, respectively.

With superior global cognition, there was an independent association between higher sleep maintenance efficiency and lower wake after sleep onset. However, the authors noted no association of global cognition level with sleep stage percentages, N1%, N2%, N3%, REM% sleep, and sleep time with oxygen saturation <90% across all cohorts. 

Note that N3 sleep plays a crucial role in memory consolidation as it is the period for optimizing glymphatic clearance. Yet, in this study, the researchers found no association between differences in N3 percentage and cognitive function. It is possible that other sleep stages compensated for age-related N3 declines or that single-night PSG were inadequate to quantify N3 sleep.

Further, pooled estimates revealed an association between short total sleep time and poorer attention and processing speed. Furthermore, the authors noted that the association between rapid eye movement (REM) sleep percentage and global cognition was positive only in female participants of the CHS cohort. Likewise, they observed that OSA was associated with poorer global cognition only in people with more daytime sleepiness.

Conclusions

Taken together, these results demonstrated an association between poorer sleep consolidation and OSA with poorer global cognition within five years. Results showed no associations between sleep stage percentages and cognition. Moreover, of all cognitive domains, there was an association between short sleep duration and poorer attention and processing speed.

Further work is needed to ascertain whether there are sensitive periods in adult life when good sleep is more critical for preventing late-life dementia or exposure to suboptimal sleep is enough for improved cognitive outcomes. 

Written by

Neha Mathur

Neha is a digital marketing professional based in Gurugram, India. She has a Master’s degree from the University of Rajasthan with a specialization in Biotechnology in 2008. She has experience in pre-clinical research as part of her research project in The Department of Toxicology at the prestigious Central Drug Research Institute (CDRI), Lucknow, India. She also holds a certification in C++ programming.