Sleep-wake cycle measures may flag dementia risk

Accelerometer-derived sleep-wake cycle measures are associated with dementia risk, modestly improving the accuracy of a dementia prediction model comprising established risk factors, a study has shown.

The analysis included data from two prospective UK population-based cohort studies, namely the UK Biobank (derivation study) and Whitehall II (validation study). UK Biobank consisted of 53,448 participants (mean age 67.5 years, 54.2 percent female, mean follow-up 7.8 years), while Whitehall II comprised 4,267 participants (mean age 69.4 years, 25.9 percent female, mean follow-up 10.6 years). The included participants were ≥60 years of age, were dementia-free, and had valid accelerometer and covariate data.

A total of 36 accelerometer-derived sleep-wake cycle metrics were extracted and analysed. In UK Biobank, nine sleep-wake cycle metrics were combined in two components. In component 1, higher values corresponded to shorter durations and less frequent bouts of moderate-to-vigorous physical activity, more time in low-intensity activity, lower diversity of activity intensities, and higher probabilities to transition from activity to rest during daytime. In component 2, higher values represented more extreme sleep durations, longer wake bouts during sleep, lower probabilities to transition from wake to sleep, and earlier waking time.

Both components of sleep-wake cycle were associated with increased risk of dementia (component 1: hazard ratio [HR], 1.43, 95 percent confidence interval [CI], 1.33–1.54; component 2: HR, 1.10, 95 percent CI, 1.04–1.17).  When added to a risk prediction model that included sociodemographic, behavioural, and health-related factors, sleep-wake cycle components improved the prediction of dementia (increase in C-index, 0.018, 95 percent CI, 0.011–0.025).

The findings were confirmed in the Whitehall II cohort.

Compared with an age-only prediction model, a model integrating the sleep-wake cycle components resulted in an increase in C-index equivalent to that for APOE genotype.