The effect of an Iyengar yoga-based exercise programme versus a seated yoga relaxation programme on falls in people aged 60 years and older (SAGE): a pragmatic, two-arm, parallel randomised controlled trial.

Background: Exercises targeting balance and strength are proven to prevent falls. Yoga is growing in popularity and can improve balance and mobility in older adults, but its effects on falls have not been rigorously tested. In this study, we aimed to compare the effects of Iyengar yoga-based exercise and seated relaxation yoga on the rate of falls among older adults.

Methods: This pragmatic, two-arm, parallel randomised controlled trial recruited Australian community-dwelling people aged 60 years and older who were not currently practising yoga and who lived independently. Participants were randomly assigned (1:1) to the intervention (Iyengar yoga-based exercise) or control (seated relaxation yoga) group using a computer-generated sequence. Participants and yoga instructors were unmasked, but research staff verifying falls data and assessing goal attainment were masked to group allocation. Intervention participants received 80 supervised, 1-h, twice-weekly yoga classes over 12 months, and were encouraged to undertake unsupervised practice on 2 additional days per week. Control participants attended two 1-h supervised workshops focused on seated breathing and stretching. Most classes were held online due to the COVID-19 pandemic. The primary outcome was fall rate per year. Secondary outcomes were mental wellbeing, physical activity, quality of life, balance self-confidence, physical function, sleep quality, pain, and goal attainment, all assessed in the intention-to-treat population. The study protocol was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12619001183178).

Findings: Between Oct 3, 2019, and Oct 28, 2021, 2182 older adults expressed their interest in participating, 810 were assessed for eligibility, 110 were excluded, and 700 were randomly assigned to either the Iyengar yoga exercise programme or the control group (seated yoga relaxation programme; 350 participants per group). The mean age of participants was 67 years (SD 5·2), and 570 (81%) were female and 130 (19%) were male. Six intervention participants reported musculoskeletal-related adverse events associated with the yoga programme, and no serious adverse events occurred. Contrary to expectations, there was a higher fall rate in the intervention group than in the control group (0·87 vs 0·64 falls per person-year; incidence rate ratio 1·33 [95% CI 1·01-1·75; p=0·044]). The intervention improved the number of hours per week of planned physical activity (mean difference 0·96 h per week [95% CI 0·43-1·49]; p<0·0001), self-reported balance confidence (mean difference 2·94 [0·60-5·28]; p=0·014), and goal attainment (mean difference 0·60 [0·26-0·94]; p=0·0006). No significant between-group differences were identified for other secondary outcomes.

Interpretation: This Iyengar yoga-based programme should not be recommended for fall prevention in its current form. Modified programmes with dynamic balance and fall prevention strategies warrant investigation.

Funding: National Health and Medical Research Council.