Little sleep all week, then a big sleep-in?

Cohort study. 38,015 people. 13 years of follow-up (1 Oct 1997 to 31 Dec 2010) via national death register.

People were watched for years. Nobody was told what to do. A clue, not proof.

Who: Swedish adults (18+) who took part in the Swedish National March Cohort, a 1997 fundraising walk for the Swedish Cancer Society, and returned a 36-page questionnaire (43,880 returned it; 38,015 analysed after exclusions). Main findings only in people under 65 at baseline

• Short every nightMoredied
• Short, then sleep inNo moredied

Under 65: short weekday sleep + longer weekend sleep → death rate no different from 6–7 h sleepers

Nobody was told how long to sleep.

More to know, from our read of the paper:

• Observational cohort: association, not causation. The authors call the compensation interpretation 'speculative' and say long weekend sleep may be 'appetitive behaviour' rather than true compensation
• Sleep duration was self-reported once, at baseline only (answer options <5, 5, 6, 7, 8, ≥9 hr); no repeat measures over 13 years
• 'Weekend' sleep is actually sleep on 'days off', which the authors simplify to weekend
• All findings hold only for people under 65; in people 65+ no association was found for any sleep pattern
• The short-weekday + medium/long-weekend group was small (1,438 people, 85 deaths overall), so the authors could not separate 'medium' from 'long' weekend catch-up

38,015 Swedish grown-ups said once how long they usually slept on work days and on days off, and scientists then checked deaths for 13 years (a cohort study: linked, not proven). Under 65, people who slept 5 hours or less every night had a 65% higher death rate than people who slept 6–7 hours. Under 65, people who slept 5 hours or less on work days but longer on days off did not differ from the 6–7-hour group. Over 65, no sleep pattern was linked to deaths. Nobody was told how long to sleep, so this is a clue, not proof.

We read the abstract and the full text of this paper and checked every number on this page against it (18 facts, checked on 2026-10-01). The study design is named above. Cohort and other observational studies are described as linked, not proven. Found an error? Email theis@onestudyaday.com.