Introduction Both short and long sleep duration and sleep fragmentation have immediate and long-term adverse impacts on physical and mental health. However, most population studies are based on self-reported sleep and health status rather than objective assessments. There are few longitudinal datasets that measure the impact of poor sleep over time. This study investigated the impact of objectively measured poor sleep on the long term metabolic and mental health status amongst the UK Biobank population.

Methods Sociodemographic, accelerometry and primary care records data were obtained from the UK Biobank (n=84,404). Sleep duration and fragmentation was objectively assessed with accelerometry (mean age= 62.4 years) and divided into five sleep groups: <5 hours, 5-6 hours, 6-7 hours, 7-8 hours and >8 hours. Sleep fragmentation related measurements including wake after sleep onset, activity level during the least active five hours and episodes of movement during sleep were also analysed. Binary regression models were adjusted for age, gender and Townsend deprivation score. There was then detailed assessment of the primary care records after a 6-year interval.

Results A ‘U-shaped’ relationship was found between sleep duration and incidence of many metabolic diseases, as well as mental illnesses such as depression. Fragmented sleep and both short and long sleep duration were associated with increased odds ratio (OR) of diseases including diabetes (OR= 1.42, p<0.001), hypertensive disease (OR= 1.22, p<0.001), ischaemic heart disease (OR= 1.12, p<0.001) and organic mental disorder (OR= 1.77, p<0.001).

Discussion Using the largest accelerometry cohort to date, 6-8 hours of sleep along with non-fragmented sleep was associated with better, future metabolic and mental health. However, sleep duration and quality alone were not sufficiently sensitive or specific to be a stand-alone biomarker for future health status. Future studies should consider additional lifestyle and health factors.