Sleep Duration is Associated with Clinically Silent Brain Injury in Middle-Aged Persons without Stroke
Santiago Clocchiatti-Tuozzo1, Cyprien Rivier2, Daniela Renedo2, Victor Torres-Lopez2, Sam Payabvash4, Kevin Sheth5, Thomas Gill3, Guido Falcone6
1Yale University, Department of Neurology, 2Yale University, 3Internal Medicine, Geriatrics, Yale University, 4Yale University, Radiology, 5Yale UniversityDivision of Neuro and Critical Care, 6Yale School of Medicine
Objective:
To study whether suboptimal sleep duration worsens neuroimaging brain health profiles in middle-aged persons without stroke.
Background:
Evidence indicates that optimization of cardiovascular health during middle age leads to brain health benefits later in life. The AHA Life’s Simple 7, a research and public health construct for cardiovascular health, recently added sleep as a risk factor, becoming the Life’s Essential 8 (LE8).
Design/Methods:
We conducted a cross-sectional neuroimaging analysis within the UK Biobank. We included participants without stroke/dementia who underwent research brain MRI. We created a 6-category sleep score according to hours of sleep (best to worse): 7 to <9h; 9 to <10h; 6 to <7h; 5 to <6 or >=10h; 4 to <5h; and <4h. We evaluated 3 neuroimaging traits: white matter hyperintensity volume (natural log-transformed), fractional anisotropy (FA) and mean diffusivity (MD). For FA and MD, we evaluated the first principal component of measurements obtained across 48 neuroanatomical regions.
Results:
Of 502,408 participants enrolled in the UKB, 39,937 (7.9%) stroke/dementia-free enrollees participated in the brain MRI study (mean age 55, 53% female). The distribution of sleep categories was: 1 (n=28,958, 72.51%), 2 (n=2060, 5.16%), 3 (n=7165, 17.94%), 4 (n=1,562, 3.91%), 5 (n=163, 0.41%) and 6 (n=29, 0.07%). In multivariable linear regression analyses, a higher (worse) sleep score was associated with larger WMH volume (beta 0.026, SE=0.0046; p<0.001) and worse FA profile (beta 0.003, SE=0.0008; p<0.001), with no association observed for MD (p=0.77). For FA, when evaluating the 48 neuroanatomical regions separately, the most compromised areas were the right cerebral peduncle (p<0.001) and left and right cerebellar peduncles (p<0.001).
Conclusions:
Among middle-aged participants enrolled in the UKB, suboptimal sleep duration was significantly associated with adverse neuroimaging brain health profiles. Sleep duration may thus determine brain health in middle-aged persons who have not developed clinically evident manifestations of poor brain health (stroke or dementia).