Exercise effects on backward walking speed in people with dementia: A randomized controlled trial.

Background: Multidirectional walking, including backward walking, is integral to daily activities, and seems particularly challenging in older age, and in people with pathology affecting postural control such as dementia.

Research Question: Does exercise influence backward walking speed in people with dementia, when tested using habitual walking aids and without, and do effects differ according to walking aid use?

Methods: This study included 141 women and 45 men (mean age 85 years) with dementia from the Umeå Dementia and Exercise (UMDEX), a cluster-randomized controlled trial study set in 16 nursing homes in Umeå, Sweden. Participants were randomized to a High-Intensity Functional Exercise (HIFE) program targeting lower limb strength-, balance and mobility exercise or to a seated attention control activity. Blinded assessors measured 2.4-meter usual backward walking speed, at baseline, 4 - (intervention completion) and 7-month follow-up; tested 1) with habitual walking aids allowed, and 2) without walking aids.

Results: Linear mixed models showed no between-group effect in either backward walking speed test at 4 or 7 months; test 1) 0.005 m/s, P = .788 and -0.006 m/s, P = .754 and test 2) 0.030 m/s, P = .231 and 0.015 m/s, P = .569, respectively. In interaction analyses, exercise effects differed significantly between participants who habitually walked unaided compared with those that used a walking aid at 7 months (0.094 m/s, P = .027).

Significance: In this study of older people with dementia living in nursing homes, the effects of exercise had no overall effects on backwards walking speed. Nevertheless, some benefit was indicated in participants who habitually walked unaided, which is promising and merits further investigation in future studies.