Why sleep has a U-shaped mortality curve, not a "more is better" one

Of all seven Helixspan factors, sleep is probably the one most likely to surprise people — not because short sleep is risky (that's widely understood), but because long sleep shows up as a risk factor too. The relationship isn't "more sleep is better up to a point and then flat." It's a genuine U-shape, with elevated risk at both ends.

The evidence for the U-shape

A frequently cited 2010 meta-analysis in the journal Sleep, pooling data across sixteen studies and over a million participants, found that both short sleep (typically defined as under 6 hours) and long sleep (typically over 9 hours) were associated with elevated all-cause mortality compared to the reference range of around 7 hours. This pattern has been replicated across many subsequent cohort studies in different populations, with the low point of the risk curve consistently landing somewhere in the 7–8 hour range.

Why short sleep is harmful — the more intuitive half

Chronic short sleep is linked to impaired glucose metabolism, elevated blood pressure, increased inflammatory markers, and impaired immune function. It also has same-night effects on cognition, mood regulation, and accident risk that compound over time even before considering longer-term disease risk. The mechanisms here are relatively well understood: sleep is when much of the body's repair and metabolic regulation happens, and chronic restriction interferes with those processes.

Why long sleep is a risk marker too

The long-sleep side of the curve is less intuitive and the causal story is murkier. Some of the association is likely reverse causation — long sleep duration can be a symptom of underlying illness (depression, chronic inflammation, undiagnosed conditions) rather than a cause of poor health in its own right. Some studies also find that long sleep correlates with lower physical activity and other confounding lifestyle factors. That said, the association persists in a meaningful number of studies even after adjusting for these confounders, which is why most sleep researchers treat it as a real, if smaller and less mechanistically clear, part of the risk curve rather than dismissing it entirely.

What the model rewards

The Helixspan sleep factor runs from −4 to +2 years, centered on the 7–8 hour range. It's intentionally a smaller positive range than exercise, reflecting that sleep's upside is largely about avoiding the harms of insufficient or (to a lesser extent) excessive sleep, rather than an open-ended benefit from "optimizing" further within the healthy range.