Effectiveness of exercise programmes in improving physical function and reducing behavioural symptoms of community living older adults with dementia living in Asia, and impact on their informal carers: A systematic review and meta-analysis.

Only nine studies were found, most with 26 to 90 participants, and only four pooled analyses were possible, drawing on three of the nine studies. Several outcomes were likely too small to detect a real difference. Study quality was mixed. Four of the six randomised trials were rated at unclear risk of bias and two at high risk; of the three non-randomised studies, one was rated poor, one fair and one good. Only English-language published studies were searched and grey literature was excluded, so relevant work may have been missed. Programmes varied widely in exercise type, length (4 to 52 weeks), dose and intensity, and outcomes were measured with different tools, so the review cannot say which exercise, how much, or for how long. Only one study reassessed participants after the programme ended (12 weeks later), so it is unknown whether any benefits are sustained. Most studies came from more economically developed Asian countries and most participants had mild dementia, so results may not transfer to other Asian settings or to people with more advanced dementia. The authors state the small number of studies and small sample sizes may have over-estimated the pooled results.

The authors declare no conflicts of interest relevant to this paper. One author was funded by an Australia Awards Scholarship and another was supported by an NHMRC Investigator Grant; neither is a commercial sponsor of the interventions studied.

Do not take the improved reach-test balance result as proof that exercise improves balance or physical function in dementia. It comes from one pooled analysis of two small studies, the same review found no balance gain on a multi-item scale and no improvement in walking mobility or leg strength, and the authors concluded the overall findings were inconsistent.