Both short and long sleepers show older-looking organs

Sleep is routinely described as central to healthy ageing, but the evidence usually rests on one outcome at a time: a memory test, a blood marker, a disease diagnosis. What has been missing is a systemic view, one that asks whether sleep duration tracks with ageing across many organs and many types of biological measurement at once. Biological ageing clocks, which estimate how old a tissue or a blood profile looks relative to chronological age, offer a way to do that. The open question is which sleep durations line up with the youngest-looking biology, and whether the answer differs by organ or by sex.

The researchers built what they call a Sleep Chart, relating self-reported sleep duration to 23 biological ageing clocks. Those clocks were derived from three different technologies: in vivo imaging, plasma proteomics and metabolomics, covering nine brain and body systems. The analysis drew on UK Biobank participants aged 37 to 84. Beyond the cross-sectional pattern, the team looked at genetic correlations, time-to-incident survival predictions for disease and death, mediation pathways towards late-life depression, and Mendelian randomisation to probe the direction of the relationship.

A systemic U-shaped pattern emerged between sleep duration and biological age gaps across all nine systems and all three measurement technologies. The lowest biological age gaps, meaning the youngest-looking biology, fell between 6.4 and 7.8 hours, with the exact figure varying by organ and by sex. Short sleep under 6 hours and long sleep over 8 hours, compared with 6 to 8 hours, were both associated with increased risk of systemic diseases beyond the brain and with all-cause mortality, including depression and diabetes. The routes differed: ageing clocks may partly mediate the link between long sleep and late-life depression, while short sleep showed a more direct link. Mendelian randomisation did not give strong evidence that disease causally affects sleep, but the authors say reverse causality cannot be fully excluded.

U-shaped associations between sleep duration and mortality have been reported before, and this work extends the pattern from disease endpoints down to measurable biology across multiple organ systems and multiple omics platforms at once. Finding the same shape in imaging, proteins and metabolites makes it harder to dismiss as an artefact of one measurement method. What remains unresolved is direction: long sleep in particular may be a symptom of underlying illness rather than a cause of accelerated ageing, and the authors are explicit that they cannot completely rule this out. Self-reported sleep is also a rough proxy for what actually happens overnight.

This is a reason to think of sleep duration as something with an apparent middle range rather than a more-is-better dial. It does not show that lengthening or shortening your sleep would change how old your organs look, because nobody in this study was assigned a sleep schedule. If your sleep sits far outside the 6 to 8 hour band, the finding is a reason to be curious about the cause, since illness can drive both very short and very long sleep. Treat it as a map of where the biology sits, not as a target to hit.