Decades of data
During the Whitehall II study, researchers assessed average sleep duration (“How many hours of sleep do you have on an average weeknight?”) six times over 30 years of follow-up. Dr. Sabia’s group extracted self-reported sleep duration data at ages 50, 60, and 70. Short sleep duration was defined as fewer than 5 hours, or 6 hours. Normal sleep duration was defined as 7 hours. Long duration was defined as 8 hours or more.
A questioner during the Q&A period noted that this grouping is a little unusual. Many studies define 7-8 hours as normal. Dr. Sabia answered that they were unable to examine periods of 9 hours or more due to the nature of the data, and the lowest associated risk was found at 7 hours.
The researchers analyzed data from 7,959 participants (33.0% women). At age 50, compared with 7 hours of sleep, 6 or few hours of sleep was associated with a higher risk of dementia over the ensuing 25 years of follow-up (hazard ratio [HR], 1.22; 95% confidence interval [CI], 1.01-1.48). The same was true at age 60 (15 years of follow-up HR, 1.37; 95% CI, 1.10-1.72). There was a trend at age 70 (8 years follow-up; HR, 1.24; 95% CI, 0.98-1.57). For 8 or more hours of sleep, there were trends toward increased risk at age 50 (HR, 1.25; 95% CI, 0.98-1.60). Long sleep at age 60 and 70 was associated with heightened risk, but the confidence intervals were well outside statistical significance.
Twenty percent of participants had persistent short sleep over the course of follow-up, 37% had persistent normal sleep, and 7% had persistent long sleep. Seven percent of participants experienced a change from normal sleep to short sleep, 16% had a change from short sleep to normal sleep, and 13% had a change from normal sleep to long sleep.
Persistent short sleep between age 50 and 70 was associated with a 30% increased risk of dementia (HR, 1.30; 95% CI, 1.00-1.69). There were no statistically significant associations between dementia risk and any of the changing sleep pattern groups.
Dr. Sabia and Dr. Sexton have no relevant financial disclosures.