Sleep regularity predicts all-cause mortality more reliably than sleep duration itself. In a prospective cohort of 60,977 UK Biobank adults followed for nearly eight years, individuals whose bedtimes and wake times varied the least from day to day showed a 20 to 48% lower risk of death compared to the most irregular individuals, independently of the number of hours slept (PubMed). This result, published in Sleep in 2024, reorders the usual hierarchy of sleep hygiene advice: when you go to bed matters more than how many hours you accumulate.
The study uses the Sleep Regularity Index (SRI), a score derived from more than ten million hours of accelerometer data. The mortality reduction is dose-dependent and covers three distinct causes: all-cause mortality, cardiometabolic mortality, and cancer mortality. The cross-cause consistency suggests a systemic mechanism rather than an organ-specific effect.
What the Sleep Regularity Index measures
The SRI quantifies the probability that an individual is in the same state (asleep or awake) at two identical moments separated by 24 hours, over a seven-day observation window. A score of 100 corresponds to a perfectly stable cycle; a score of 0 to a completely random cycle. Unlike duration averages, the SRI captures the stability of the circadian signal, not the total volume of sleep.
This distinction is biologically fundamental. The central clock, located in the suprachiasmatic nucleus of the hypothalamus, runs on an endogenous cycle of about 24.2 hours. It must be reset daily by external signals called zeitgebers ("time givers"), the most powerful of which is morning light. An unstable sleep-wake cycle degrades the suprachiasmatic nucleus's ability to synchronize peripheral clocks (liver, pancreas, adipose tissue, vascular endothelium), each of which possesses its own circadian machinery.
Difference in all-cause mortality risk between the most regular sleep quintile and the most irregular quintile, after adjustment for age, sex, sleep duration, and metabolic covariates.
The data: 60,977 participants, 7.8 years of follow-up
The Windred cohort includes 60,977 adults (mean age 62.8 years, 55% female) equipped with a wrist accelerometer for seven days between 2013 and 2015. After adjustment for classical risk factors, the most regular quintile shows a 20 to 48% reduction in all-cause mortality depending on the model, a 22 to 57% reduction in cardiometabolic mortality, and a 16 to 39% reduction in cancer mortality.
One methodological point deserves attention. When the authors directly compare regularity and sleep duration in nested models, regularity retains its predictive power even after controlling for duration. The reverse is not true: duration loses some of its predictive power when adjusted for regularity. This does not mean duration is unimportant, but that part of the effect of duration is mediated by regularity itself.
An independent cohort, the Hispanic Community Health Study / Study of Latinos, partially confirms the direction: a low SRI is associated with increased prevalence of type 2 diabetes in cross-sectional analysis (odds ratio 1.64 in the most irregular quartile) (PubMed). The prospective association does not reach significance in this 2,107-subject subsample, which calls for caution: the effect is robust in large cohorts such as the UK Biobank, but its magnitude varies across populations.
Why regularity matters more than duration: the circadian mechanisms
Sleep irregularity creates a misalignment between the central clock and peripheral clocks. Three biological axes document the consequences of this misalignment.
The cortisol-melatonin axis
Cortisol follows a precise circadian rhythm: morning peak (30 to 45 minutes after waking), gradual decline during the day, nadir in early night. Melatonin operates as a mirror: nocturnal secretion initiated 2 to 3 hours before habitual sleep onset, daytime suppression by light. When bedtime varies by two hours from day to day, these two curves desynchronize relative to tissue metabolic demand. The immediate consequence is a reduction in insulin sensitivity upon waking, a phenomenon measurable from the first night of irregularity (PubMed).
Low-grade inflammation
The Irwin et al. meta-analysis shows that chronic sleep disturbances are accompanied by sustained elevation of high-sensitivity C-reactive protein (hs-CRP) and interleukin-6, two markers of systemic inflammation (PubMed). The magnitude of the effect remains modest (effect size 0.12 for hs-CRP), but its persistence over time makes it a biological lever consistent with the increased cardiometabolic mortality observed in the Windred study. Low-grade inflammation is now considered an independent risk factor for cardiovascular events, on a par with LDL.
Lipid profile and insulin resistance
Circadian desynchronization directly affects lipid metabolism. The Wong et al. study of 447 adults shows that a phase shift greater than one hour between weekdays and weekends (social jetlag) is associated with decreased HDL, elevated triglycerides, increased fasting insulin, and higher HOMA-IR, independently of sleep duration and subjective quality (PubMed). The New Hoorn cohort (1,585 adults) confirms that people with more than two hours of social jetlag have a 1.64-fold increased risk of metabolic syndrome (PubMed).
The blood biomarkers that respond to irregularity
The clinical data converge on a set of biomarkers that deteriorate in response to chronic circadian irregularity, well before symptoms appear. These markers are accessible through a standard blood panel and allow the metabolic impact of unstable sleep hygiene to be quantified.
Glucose metabolism. Fasting glucose, insulin, HOMA-IR, and HbA1c reflect insulin sensitivity and long-term glycemic regulation. HbA1c, which integrates glycemic exposure over the last 90 days, is particularly sensitive to repeated misalignments. HOMA-IR, calculated from fasting glucose and insulin, captures hepatic insulin resistance quickly.
Systemic inflammation. hs-CRP is the marker of choice for the low-grade inflammation associated with circadian disruption. Values above 2 mg/L in a subject without acute infection may indicate persistent inflammation, partially reversible through the restoration of a stable sleep-wake cycle.
Lipid profile. Triglycerides, HDL, and apolipoprotein B (ApoB) are all modulated by the hepatic and intestinal clocks. ApoB, in particular, counts one atherogenic particle per molecule and reflects the burden of potentially harmful particles with precision. An elevated triglyceride/HDL ratio is an early marker of insulin resistance.
Thyroid balance. TSH shows a pulsatile circadian rhythm with a late-night peak. Measuring it at a fixed time allows detection of subtle drifts associated with chronically fragmented sleep.
How to stabilize your SRI
Three levers emerge from intervention studies with a favorable effort/benefit ratio.
The first is anchoring wake time within ± 30 minutes, seven days a week. Counterintuitively, it is wake time, not bedtime, that stabilizes the central clock. Bedtime tends to follow naturally when wake time is fixed, through the homeostatic sleep pressure accumulated during the day.
The second is retinal exposure to natural light within 30 minutes of waking, ideally outdoors, for 10 to 15 minutes. Outdoor light intensity, even on overcast days, far exceeds that of the most powerful indoor lighting. It is this morning light signal that resets the central clock and paces hormonal secretion for the next 24 hours.
The third is the reduction of blue light and heavy meals in the two hours before bed. Blue light suppresses melatonin secretion; late meals shift the phase of hepatic and pancreatic clocks. Both contribute to inter-clock misalignment.
The convergence of the data is clear. Circadian regularity is not a peripheral refinement of sleep hygiene, but a first-order biological factor whose effects are measurable on long-term mortality and on a coherent set of blood biomarkers in the short term. For an individual attentive to their longevity trajectory, stabilizing the SRI represents probably one of the simplest, least costly, and best-documented interventions in the currently available toolkit.
Frequently asked questions
References
- Windred DP, Burns AC, Lane JM, Saxena R, Rutter MK, Cain SW, Phillips AJK. Sleep regularity is a stronger predictor of mortality risk than sleep duration: A prospective cohort study. Sleep. 2024;47(1):zsad253 (PubMed).
- Wong PM, Hasler BP, Kamarck TW, Muldoon MF, Manuck SB. Social Jetlag, Chronotype, and Cardiometabolic Risk. J Clin Endocrinol Metab. 2015;100(12):4612-4620 (PubMed).
- Koopman ADM, Rauh SP, van 't Riet E, et al. The Association between Social Jetlag, the Metabolic Syndrome, and Type 2 Diabetes Mellitus in the General Population: The New Hoorn Study. J Biol Rhythms. 2017;32(4):359-368 (PubMed).
- Fritz J, Phillips AJK, Hunt LC, et al. Cross-sectional and prospective associations between sleep regularity and metabolic health in the Hispanic Community Health Study/Study of Latinos. Sleep. 2021;44(4):zsaa218 (PubMed).
- Irwin MR, Olmstead R, Carroll JE. Sleep Disturbance, Sleep Duration, and Inflammation: A Systematic Review and Meta-Analysis of Cohort Studies and Experimental Sleep Deprivation. Biol Psychiatry. 2016;80(1):40-52 (PubMed).
