The cohort for this prospective observational study consisted of 1390 community-dwelling men and women aged 70 years, examined in a health survey between July 2012 and November 2014.
Gait patterns were measured using a computerized walkway system during normal-speed, fast-speed, and dual-task trials. Triaxial accelerometers were used to collect objective data on physical activity, and self-reported fall data were collected by telephone 6 and 12 months after examination. Incident low-energy falls were defined as unexpected events in which participants came to rest on the ground.
During the follow-up period, 148 study participants (88 women, 60 men; P = .01) reported falls. After adjusting for multiple confounders, including objective measures of physical activity, socioeconomic factors, cardiovascular disease, and cognitive function, the odds ratio for falling in women was 1.49 (95% confidence interval [CI] 1.02–2.19). Variations in gait pattern were significantly (20%–40%) increased in fallers compared with nonfallers during the dual-task trial for step width, step length, stride length, step time, stance time, stride velocity, and single support time (all P < .05). Furthermore, women showed 15% to 35% increased variability in all of these gait parameters during the dual-task trial compared with men (all P < .01).
In the present cohort, 70-year-old women were at greater risk of falls compared with their male counterparts. This increased risk was associated with increased variation in gait pattern during dual-task activities, and may contribute to women's greater fracture risk compared with men.