No abstract

Background Large systematic reviews have pointed to the efficacy of cognitive training and physical exercise on cognitive performance in older adults, making them the most common interventions in multidomain dementia prevention trials. However, it remains unclear to which extent combinations of these interventions yield additive effects beyond their individual components and what combination strategies are most beneficial. Our aim therefore was to synthesize the evidence from randomized controlled trials of combined cognitive and physical training on cognitive, physical, psychosocial and functional outcomes in older adults with or without cognitive impairment, and to compare and rank the efficacy of the three main types of combined intervention delivery formats (simultaneous, sequential or exergaming) relative to either intervention alone or control conditions. Methods and Findings We systematically searched MEDLINE, Embase and PsycINFO from inception to 23 July 2019. Change from baseline to post-intervention were extracted for each outcome and results were analyzed using random-effects models. The overall efficacy of combined interventions was assessed using multivariate pairwise meta-analysis and the comparative efficacy across different intervention delivery formats and control conditions was investigated using network meta-analysis. A total of 47 trials encompassing 4052 participants were eligible, including cognitively healthy older adults (k=30), mild cognitive impairment (k=13), dementia (k=2) and Parkinson disease (k=2). Due to the small number of identified studies in dementia and Parkinson disease, these were excluded from the pooled analyses and instead summarized narratively. Relative to any control condition, combined interventions were associated with small and heterogeneous effects on overall cognitive (k=41, Hedges g = 0.22, 95% CI 0.14 to 0.30, prediction interval -0.34 to 0.78) and physical function (k=32, g = 0.25, 95% CI 0.13 to 0.37, prediction interval -0.46 to 0.96). Cognitive and physical effects were similar for cognitively healthy older adults and those with mild cognitive impairment. No robust evidence for benefit on psychosocial function (k=9) or functional abilities (k=2) was found. The efficacy of simultaneous and sequential training for cognition exceeded all control conditions apart from cognitive training alone, with small to moderate effect estimates and moderate certainty of the evidence. The efficacy of simultaneous and sequential training on physical outcomes was comparable but not significantly greater than physical exercise alone. Exergaming was ranked low for both outcomes and superior only to passive control. The certainty of the evidence was low for physical outcomes, and we were not able to compare interventions in Parkinson disease and dementia. Conclusion Combined cognitive and physical training, delivered either simultaneously or sequentially, is efficacious for cognitive and physical outcomes in cognitively healthy and mildly impaired older adults, but not superior to cognitive or physical exercise alone in each domain. There is little systematic evidence to support potential additive effects of combined interventions or the use of exergaming in older adults. Given simultaneous training is as efficacious but less burdensome than sequential design, the former may be better suited for practical implementation to promote cognitive alongside physical health in late life. More research is needed to establish the effects of combined interventions on everyday function and well-being.