Effects of dance on gait and dual-task gait in Parkinson's disease

Gait impairments in Parkinson's disease (PD) limit independence and quality of life. While dance-based interventions could improve gait, further studies are needed to determine if the benefits generalise to different terrains and when dual-tasking. The aim was to assess the effects of a dance intervention, based on the Dance for PD® (DfPD®) program, on gait under different dual-tasks (verbal fluency, serial subtraction) and surfaces (even, uneven), and to determine if a larger scale follow-up RCT is warranted.

A dance group (DG; n = 17; age = 65.8 ± 11.7 years) and a control group (CG: n = 16; age = 67.0 ± 7.7 years) comprised of non-cognitively impaired (Addenbrooke's score: DG = 93.2 ± 3.6, CG = 92.6 ± 4.3) independently locomoting people with PD (Hoehn & Yahr I-III). The DG undertook a one-hour DfPD®-based class, twice weekly for 12 weeks. The CG had treatment as usual. The spatiotemporal variables of gait were assessed at baseline and post-intervention while walking on two surfaces (even, uneven) under three conditions: regular walking; dual-task: verbal-fluency (DTVERB), and serial-subtraction (DTSUBT). The data were analysed by means of a linear mixed model.

At baseline, there was no significant group difference for any spatiotemporal gait variable. The DG improved significantly compared to the CG with and without a dual task when walking on even surface. During regular walking, DG improved in gait velocity (p = 0.017), cadence (p = 0.039), step length (p = 0.040) and stride length (p = 0.041). During DTVERB significant improvements were noted in gait velocity (p = 0.035), cadence (p = 0.034) and step length (p = 0.039). The DG also exhibited significant improvement compared to the CG during DTSUBT in the measures of gait velocity (p = 0.012), cadence (p = 0.021), step length (p = 0.018), and stride length (p = 0.151). On the uneven surface, improvements were noted when walking while performing serial subtractions only. During regular walking, improvements were noted for the CG but not for the DG. CG has spent less time in double support following the intervention than DG. While DTVERB condition had no significant group differences for any gait parameter (p's >0.05), in the DTSUBT condition, the DG improved significantly compared to the controls on gait velocity (p = 0.048), cadence (p = 0.026), and step length (p = 0.051).

DfPD®-based classes produced clinically significant improvement in spatiotemporal gait parameters under dual-task conditions and on uneven surfaces. This could arise from improved movement confidence and coordination; emotional expression; cognitive skills (planning, multitasking), and; utilisation of external movement cues. A large-scale RCT of this program is warranted.

A protocol for this study has been registered retrospectively at the Australian New Zealand Clinical Trials Registry. Identifier: ACTRN12618001834246.