Wong · Sleep 2021 · prospective cohort study and meta-analysis · n=713,150 participants (15 prospective studies with 65,410 cases in meta-analysis)

Sleep duration and breast cancer incidence: results from the Million Women Study and meta-analysis of published prospective studies.

Cited 47 times in the scientific literature.

Level 3 - non-randomized controlled study

Meta-analysis of prospective cohort studies and prospective cohort analysis

PubMed 32886784 · doi:10.1093/sleep/zsaa166 · record verified 2026-08-31

What was done

In the UK Million Women Study, self-reported 24-hour sleep duration was assessed in 2001 among 713,150 women without prior cancer, cardiovascular disease, or diabetes (mean age 60 years). Participants were followed for an average of 14.3 years via national cancer registry linkage. Cox regression models calculated multivariable-adjusted relative risks (RRs) for breast cancer after excluding the first 5 years of follow-up to limit reverse causation. In addition, the authors conducted an inverse-variance weighted meta-analysis combining their findings with 14 other published prospective studies of sleep duration and breast cancer.

What was found

In the Million Women Study (24,476 incident breast cancer cases), compared to 7-8 hours of sleep per night, RRs for breast cancer were 1.01 (95% CI, 0.95-1.07) for <6 hours, 0.99 (95% CI, 0.96-1.03) for 6 hours, 1.01 (95% CI, 0.96-1.06) for 9 hours, and 1.03 (95% CI, 0.95-1.12) for >9 hours. In the pooled meta-analysis of 15 prospective cohorts encompassing 65,410 breast cancer cases, breast cancer risk was not associated with short sleep (<7 hours: RR 0.99, 95% CI 0.98-1.01) or long sleep (>8 hours: RR 1.01, 95% CI 0.98-1.04) relative to average sleep.

Why it matters

These findings provide definitive, highly powered prospective evidence that sleep duration is not an etiological risk factor for breast cancer incidence, clarifying ambiguity generated by smaller or retrospective studies.

Limits

Sleep duration was self-reported rather than objectively measured by actigraphy or polysomnography. The study focused on duration and did not assess sleep quality, circadian disruption, shift work, or sleep architecture. Meta-analysis results are subject to between-study differences in categorization of sleep hours.

Cited by