Sleep health may depend on more than simply counting the hours spent in bed. New research involving nearly 96,000 UK Biobank participants found that REM sleep, deep sleep, sleep consistency and nighttime wakefulness were associated with different patterns of future disease risk.
The findings suggest that both sleep duration and the internal structure of sleep may provide important clues about long-term health.
A Large Study Examines Real-World Sleep
Researchers analyzed data from 95,559 UK Biobank participants who wore wrist-based accelerometers continuously for seven days and nights. Instead of relying entirely on questionnaires, the research team used a deep-learning system called SleepNet to estimate several characteristics of sleep, including REM sleep, deep sleep, light sleep, total sleep duration, nighttime wakefulness and variation in sleep duration from one night to another.
Participants were followed for a median of 8.9 years, with health records used to examine 1,049 disease outcomes. After statistical adjustments, the researchers identified 156 significant associations between sleep characteristics and later disease diagnoses.
The approach offers an important advantage over studies based solely on self-reported sleep because wearable measurements can capture patterns that people may not accurately remember or estimate. However, accelerometers cannot measure sleep stages as precisely as laboratory polysomnography, so the estimated stages should be interpreted as movement-based measures rather than direct physiological recordings.
REM Sleep Had the Widest Associations
Among the different sleep stages examined, REM sleep produced the broadest pattern of associations. An increase equivalent to the study's interquartile range of 47.6 minutes was associated with lower risks for 83 diseases spanning 12 categories.
Several notable relationships involved cardiovascular and neurological conditions. Greater REM sleep was associated with lower risks of heart failure, atrial fibrillation, ischemic heart disease, dementia, Alzheimer's disease, multiple sclerosis and parkinsonism. For example, the reported hazard ratio was 0.54 for dementia and 0.20 for parkinsonism when comparing the relevant sleep exposure levels.
These results do not demonstrate that increasing REM sleep directly prevents those conditions. The study was observational, so the associations could also reflect other factors influencing both sleep and health.
Deep Sleep Added Another Dimension
Deep sleep showed a smaller but still meaningful group of associations. An additional 47.5 minutes, corresponding to the study's interquartile range, was associated with lower risks of seven diseases. Among the reported relationships were lower risks of type 2 diabetes, major depressive disorder, sleep apnea and Parkinson's disease. The researchers found hazard ratios of 0.89 for type 2 diabetes, 0.86 for major depressive disorder, 0.78 for sleep apnea and 0.70 for Parkinson's disease.
Sleep Fragmentation May Also Matter
The research examined what happened during the time normally classified as nighttime wakefulness. Greater wakefulness after sleep onset was associated with higher risks for six conditions, while greater night-to-night irregularity was associated with higher risks for three conditions.
The researchers reported links between increased nighttime wakefulness and several outcomes, including psychoactive substance dependence and osteoarthritis. Greater sleep irregularity was associated with higher risks of abdominal pain, anxiety disorders and major depressive disorder.
These findings suggest that consistency and continuity may complement total sleep duration when assessing sleep health. A schedule that changes substantially between nights or repeated periods of wakefulness may represent a different sleep pattern from a relatively stable and uninterrupted night.
The Six-to-Eight-Hour Range
Total sleep duration showed a nonlinear relationship with many health outcomes. Researchers identified 86 disease phenotypes with statistically significant nonlinear associations, and the lowest-risk duration for 69 of those outcomes generally fell between six and eight hours.
The strongest pattern appeared among people with very short sleep. Compared with the six-to-eight-hour reference group, participants sleeping fewer than five hours accounted for 37 of 41 significant adverse associations identified in the detailed analysis. Participants sleeping fewer than six hours also showed higher risks for several conditions, including heart failure, type 2 diabetes and chronic obstructive pulmonary disease.
Expert Insight
Dr. Nina Rzechorzek, MRC Clinician Scientist at the University of Cambridge, offered an important caution about interpreting the findings. Rzechorzek said the results “support sleep patterns as potentially informative markers of future disease risk”. She also emphasized the limits of the observational research, stating that the findings “do not show that increasing REM or deep sleep, or simply forcing sleep into a 6–8-hour window, will prevent disease”.
The large UK Biobank study indicates that sleep duration is only one part of a much broader sleep-health picture.Importantly, the research does not establish that sleep patterns directly cause or prevent these conditions. Instead, it provides large-scale evidence that sleep architecture, duration, regularity and nighttime continuity are all relevant factors for understanding long-term health.
