Sleep and Longevity: Why Sleep Is the Foundation of Health

What Sleep Is For

Sleep is not the absence of activity. It is a distinct physiological state with its own architecture, cycling roughly every ninety minutes through stages with different functions.

Deep sleep, or slow wave sleep, is concentrated in the first half of the night. Growth hormone is released predominantly in pulses during it. Glymphatic clearance, the movement of cerebrospinal fluid through brain tissue that removes metabolic byproducts including amyloid beta and tau, is most active during this stage. Sympathetic nervous system activity falls to its daily minimum, and blood pressure dips.

REM sleep is concentrated in the second half. It supports emotional memory processing, procedural learning consolidation and, on the current evidence, the integration of new information with existing knowledge. Muscle tone is suppressed and brain activity resembles waking.

Light sleep occupies the largest share and is not merely transitional. Sleep spindles during this stage participate in memory consolidation and in protecting sleep from disturbance.

Because deep sleep loads the first half of the night and REM the second, truncating sleep does not remove a proportional slice of each. Going to bed late preferentially removes deep sleep; waking early preferentially removes REM. Two people sleeping six hours can lose entirely different functions depending on which end they cut.

The Outcome Data

Duration

Large meta-analyses find a U-shaped association between habitual sleep duration and all-cause mortality, with the nadir around seven hours. Short sleep, generally under six hours, associates with higher mortality, cardiovascular events, metabolic dysfunction and cognitive decline. Long sleep, generally over nine hours, also associates with higher mortality, though causality there is more doubtful, since illness causes long sleep as much as the reverse.

Regularity

A more recent and important finding is that irregularity of sleep timing predicts mortality independently of duration, and in some analyses more strongly. Going to bed and waking at wildly variable times appears to carry a cost of its own, independent of total hours obtained. This has practical consequences: for someone who cannot extend sleep duration, stabilising timing may be the more available intervention.

Specific mechanisms with outcome links

Short sleep reduces insulin sensitivity measurably within days in controlled studies. It raises evening cortisol and shifts the appetite hormones leptin and ghrelin toward increased intake. It reduces antibody response to vaccination. It increases inflammatory markers. Slow wave sleep disruption specifically has been linked to reduced glymphatic clearance, which is the mechanistic thread connecting sleep to neurodegenerative processes.

Sleep apnoea

Obstructive sleep apnoea deserves separate mention because it is common, frequently unrecognised, and associates strongly with cardiovascular and metabolic outcomes. It is also the case where optimisation efforts are most often wasted: no supplement or routine addresses it, and it requires clinical assessment. Loud snoring, witnessed breathing pauses, unrefreshing sleep despite adequate duration and morning headache are reasons to seek that assessment rather than to buy a magnesium product.

The Circadian Layer

Sleep quality is largely determined by circadian alignment, and circadian alignment is largely determined by light.

The central clock in the suprachiasmatic nucleus is entrained principally by light reaching intrinsically photosensitive retinal ganglion cells, which are most responsive to short-wavelength light. Morning light exposure advances the clock and strengthens its amplitude; evening light delays it and weakens the melatonin rise. The intensity difference between outdoor daylight and indoor lighting is enormous, often a hundredfold, which is why fifteen minutes outdoors in the morning does more than any indoor arrangement.

Peripheral clocks in liver, muscle and adipose tissue are entrained partly by feeding time rather than by light. Eating late shifts peripheral clocks out of phase with the central clock, which is the mechanistic basis for the metabolic effects of late eating independent of total intake.

Core body temperature is the other lever. Sleep onset requires a fall in core temperature, which is why a warm bath ninety minutes before bed helps: peripheral vasodilation after leaving the water accelerates heat loss. A cool bedroom works through the same mechanism.

What Is Worth Measuring

Consumer sleep tracking has become good enough to be useful and is still poor enough to be misleading, and knowing which is which matters.

Reasonably reliable from wearables: total sleep time, sleep timing and regularity, resting heart rate, heart rate variability trends, respiratory rate, and movement-based awakenings. These are the metrics worth watching.

Less reliable: stage classification. Wrist-worn devices infer deep and REM sleep from heart rate, movement and temperature, and agreement with polysomnography for specific stages is moderate at best. Reacting to a single night's "deep sleep" figure is reacting to an estimate with substantial error.

Most useful practice: trends over weeks rather than individual nights, and a small number of subjective measures recorded consistently. Morning alertness, need for caffeine, and afternoon energy are crude and genuinely informative. Sleep tracking anxiety is a documented phenomenon, and a person whose sleep worsens because they are monitoring it has made a net loss.

The Interventions That Actually Move Sleep

Ranked roughly by effect size in the human literature.

• Consistent timing, including weekends. The single highest-leverage change for most people, and free.
• Morning outdoor light within an hour of waking. Strengthens circadian amplitude, advances sleep onset.
• Caffeine timing. Half-life of roughly five hours means afternoon caffeine measurably affects deep sleep even when subjective sleep feels unaffected. A cutoff eight to ten hours before bed is defensible.
• Alcohol reduction. Alcohol accelerates sleep onset and then fragments the second half of the night and suppresses REM. It is the most common cause of unexplained poor sleep quality in people who believe they sleep well.
• Cognitive behavioural therapy for insomnia. For persistent insomnia, this outperforms medication in durability and is the recommended first-line approach in clinical guidance. It is underused relative to its evidence.
• Temperature. Cool room, warm bath ninety minutes before bed.
• Light restriction in the evening. Dim, warm, low, and off screens where possible. The effect is smaller than morning light exposure but real.
• Exercise. Improves sleep quality overall; intense training within a few hours of bed can delay onset in some people.
• Supplements. Genuinely last in this ranking. Magnesium glycinate, glycine and low-dose melatonin for circadian shifting have modest evidence. None substitutes for the items above.

Sleep as the Rate-Limiting Pillar

Sleep has a distinctive position among the ten Pillars: its insufficiency degrades all of the others.

Short sleep reduces insulin sensitivity, which affects the Metabolic Pillar. It raises inflammatory markers, affecting Inflammation and Immune Defense. It suppresses testosterone in men and disrupts appetite and stress hormone signalling, affecting Hormonal Optimization. It impairs consolidation and glymphatic clearance, affecting Cognition and Neuroprotection. It reduces training adaptation and increases injury likelihood, affecting Structural Support. It raises NAD+ consumption through inflammatory signalling, affecting Cellular Energy.

The practical implication is a priority order rather than a parallel effort. Where sleep is genuinely insufficient, work on other Pillars is being done at a discount, and the resources spent on supplementation would return more if spent on sleep timing and duration. This is not a rhetorical point about sleep hygiene. It is a claim about where the marginal hour and the marginal pound are best deployed.