The Sleep Sweet Spot: 6 to 8 Hours Slows Biological Ageing Across 17 Organs

For roughly half a million middle-aged and older adults in the UK, the relationship between how long they slept and how fast their bodies were ageing traced a shape researchers have seen before — but never at this scale or with this much granularity. In a study published this week in Nature, a team led by computational neuroscientist Junhao Wen at Columbia University analysed self-reported sleep-duration data from the UK Biobank against 23 distinct biological ageing clocks spread across 17 organ systems. The result was a consistent U-shaped curve: both people who slept less than six hours and those who slept more than eight hours showed signs of accelerated biological ageing relative to their chronological age, while the sweet spot — where the gap was narrowest — fell between roughly 6.4 and 7.8 hours, depending on the organ and the person’s sex.

“Sleep affects every organ of the body,” Abigail Dove, a postdoctoral neuroepidemiologist at the Karolinska Institute who was not involved in the study, told Nature News. “And sleep is somewhat modifiable. This is a tool that could help.”

The MULTI Consortium’s approach — which the authors call the Sleep Chart — goes further than previous efforts in two important respects. First, the sheer sample size: roughly 500,000 participants aged 37 to 84, drawn from one of the most intensively phenotyped population cohorts in existence. Second, and more novel, is the breadth of the biological readouts. The team did not rely on a single ageing clock — the kind of composite score that collapses organ-level variation into one number. Instead, they deployed 23 separate clocks that span three distinct omics layers: structural and functional brain and body imaging, blood proteomics, and circulating metabolites. That meant they could ask not just whether sleep duration correlates with biological age in aggregate, but whether the same U-shaped pattern holds up when you examine the cardiovascular system, the liver, the kidneys, the lungs, the immune system, and the brain, each on its own terms.

It mostly did. The U-shaped relationship was not confined to one or two organ systems; it appeared across cardiovascular, metabolic, hepatic, renal, immune, pulmonary, and musculoskeletal domains. The brain-ageing clocks, derived from MRI-based measures of grey- and white-matter integrity, showed the same pattern. The precise nadir of the curve shifted slightly by organ — the lowest biological-age gap for cardiovascular markers sat near 7.2 hours, while for some metabolic clocks it dipped closer to 6.8 — but the overall shape was consistent. The hazard ratios for all-cause mortality reinforced what the clocks were showing: short sleepers, defined as under six hours, carried a 50 percent higher risk of death over the follow-up period (HR 1.50, 95% CI 1.44–1.55, P < 1 × 10⁻²⁰). Those sleeping more than eight hours faced a 40 percent elevated risk (HR 1.40, 95% CI 1.36–1.44).

Wen summarised the takeaway bluntly in an interview with Down to Earth: “Our study goes further and shows that too little and too much sleep are associated with faster aging in nearly every organ.”

The U-shaped sleep-mortality curve is not new. Epidemiologists have been reporting it for more than two decades, and meta-analyses have consistently pegged the lowest-risk window at somewhere between seven and eight hours. What the Sleep Chart adds is a plausible mechanism — or at least a mechanistic correlate. If both short and long sleep are linked to mortality, and those same sleep durations are linked to accelerated biological ageing across organ systems, then biological ageing may be part of the causal pathway. It is not proof, but it narrows the explanatory gap in a way that a mortality-only association cannot.

Still, there are solid reasons to hold the conclusions at arm’s length, at least for now. The UK Biobank is not a random sample of the population. Its participants skew healthier, wealthier, and whiter than the UK average, a well-documented healthy-volunteer bias that can distort associational estimates — sometimes inflating them, sometimes attenuating them, depending on the exposure-outcome pair. More immediately relevant to this study, sleep duration was self-reported via a single questionnaire item. People are notoriously poor at estimating how long they actually sleep; in older adults, self-report agrees with wrist-worn actigraphy only modestly, with kappa statistics as low as 0.24. That means some fraction of the “long sleep” group may include people who spent nine hours in bed but only slept for six, and some of the “short sleep” group may be under-reporting. The true U-shape could be narrower, flatter, or shifted.

The direction of causation is another open question. Observational data can tell you that short and long sleep accompany faster biological ageing. It cannot tell you whether poor sleep accelerates ageing, whether an already-ageing body disrupts sleep architecture, or whether both are downstream of a third factor — chronic low-grade inflammation, socioeconomic deprivation, undiagnosed disease — that the statistical models did not fully capture. Wen and his co-authors acknowledged this in the paper, noting that longitudinal tracking with repeated sleep measurements and interventional designs are the logical next steps.

Those caveats do not neutralise the findings. They just make them provisional, which is what most large-scale epidemiology is. And the signal here is strong enough — and consistent enough across organs, clocks, and mortality endpoints — that it probably warrants attention from clinicians who talk to patients about sleep. Dove, for her part, sees the modifiability of sleep as the most actionable thread. Unlike many of the biological clocks themselves, which reflect accumulated damage from genetics, environment, and decades of lifestyle, sleep duration is something people can often change, whether through cognitive behavioural therapy for insomnia, stimulus control, or, for some, simply deciding to protect it.

Wen echoed that framing. “Sleep might be more environmental,” he said. “It’s a strong message for the public that this can be modifiable.”

What the study cannot answer — and what no single observational study could — is whether moving a short sleeper from five and a half hours to seven would measurably slow their biological ageing over a period of years. That question requires a randomised trial, and at least one is already under way: the ALPS trial is testing whether a structured sleep-extension intervention changes ageing-related biomarkers in middle-aged adults. Results are expected within the next two to three years. If they show that extending sleep moves the needle on biological ageing clocks, the public-health case for treating sleep as a first-line healthy-ageing intervention becomes considerably stronger.

In the meantime, the Sleep Chart provides something genuinely useful: a high-resolution, multi-organ map of where the epidemiological evidence currently stands. It suggests that for most people, somewhere around seven hours — give or take 40 minutes — is associated with the least biological wear and tear across most of the body’s systems. It also suggests that consistently sleeping substantially more or less than that range is worth investigating, not because either extreme is necessarily pathological, but because both are associated with outcomes that deserve a closer look.

The researchers are careful not to prescribe a single number. Individual variation is real, and the optimal range shifted slightly by organ system and by sex in their data — a reminder that “one size fits all” recommendations are almost always oversimplifications. But if the U-shaped curve is replicated in more representative cohorts and with objective rather than self-reported sleep measurement, the implications would be substantial. Sleep — freely available, non-pharmacological, and already recommended for reasons spanning cognitive performance, metabolic health, and immune function — would gain yet another credential: a plausible and potentially modifiable role in slowing how fast we age.

References

1. Wen J, et al. Sleep chart of biological ageing clocks in middle and late life. Nature. 2026. https://www.nature.com/articles/s41586-026-10524-5
2. Ledford H. Sleep linked to slower ageing: huge study pinpoints the right amount. Nature News. 2026. https://www.nature.com/articles/d41586-026-01506-8
3. Keyes KM, Westreich D. UK Biobank, big data, and the consequences of non-representativeness. Lancet 393(10178):1297. 2019. https://ncbi.nlm.nih.gov/pmc/articles/PMC7825643/
4. Van Den Berg JF, et al. Disagreement between subjective and actigraphic measures of sleep duration in a population-based study of elderly persons. J Sleep Res 17(3):295-302. 2008. https://pubmed.ncbi.nlm.nih.gov/34338629.