To test the hypothesis that subjective and objective sleep disturbances are associated with risk of incident falls in older men.

The prospective observational MrOS Sleep Study.

Six academic clinical centers in the United States.

Community‐dwelling men aged 67 and older (mean 76) (n = 3,101).

Subjective sleep measurements included daytime sleepiness (Epworth Sleepiness Scale (ESS)), sleep quality (Pittsburgh Sleep Quality Index (PSQI)), and total sleep time (TST). Objective sleep measurements included actigraphic TST and sleep efficiency (an index of fragmentation) and sleep disordered breathing (measured using in‐home polysomnography). Fall frequency during the subsequent year was ascertained three times per year using questionnaires. Recurrent falling was defined as having two or more falls in the subsequent year.

In multivariable‐adjusted models, participants with excessive daytime sleepiness (ESS > 10) but not poor subjective sleep quality (PSQI > 5) had greater odds of experiencing two or more falls in the subsequent year (odds ratio (OR) = 1.52 95% confidence interval (CI) = 1.14–2.03). Based on actigraphic recordings, the odds of having recurrent falls was higher for men who slept 5 hours or less (OR = 1.79, 95% CI = 1.22–2.60) than for those who slept 7 to 8 hours. Actigraphically measured sleep efficiency was also associated with greater risk of falls, as was nocturnal hypoxemia (≥10% of sleep time with arterial oxygen saturation <90%; OR = 1.62, 95% CI = 1.17–2.24) but not apnea hypopnea index.

Subjective and objective sleep disturbances were associated with risk of falls in older men, independent of confounders.