Combined exercise and cognitive interventions for adults with mild cognitive impairment and dementia: A systematic review and network meta-analysis.

Background: Exercise and cognitive interventions are beneficial for adults with preclinical and clinical dementia, but it is unclear whether the combination of these two components could generate synergistic benefits and what intervention designs would optimize this effect.

Objectives: This review aims to compare the effects of combined exercise and cognitive interventions on cognitive, psychological, functional outcomes, and health-related quality of life with the corresponding single approach and control groups in adults with mild cognitive impairment and dementia. It also aims to identify the optimal intervention design and factors affecting treatment effects.

Methods: A comprehensive search was conducted in ten databases from inception to 23rd November 2022. The methodological quality of studies was evaluated by the Cochrane risk of bias tool. Pairwise meta-analyses were performed to assess the effects of combined interventions relative to the single type of intervention and control groups, with further subgroup analysis to explore the factors affecting treatment effects. Network meta-analyses were used to identify the optimal intervention components.

Results: Twenty-nine randomized controlled trials involving 2910 participants were included. The results of pairwise meta-analyses indicated that combined interventions were superior to exercise in improving response inhibition, working memory, and delayed recall, but were not superior to cognitive interventions in all outcomes. Combined interventions were superior to active/passive controls in improving global cognition, response inhibition, immediate recall, delayed recall, category fluency, processing speed, and visuospatial ability. Influences of the clinical severity of dementia (mild cognitive impairment vs dementia), combination format (sequential vs simultaneous combination), mode of delivery (group-based vs individual-based vs mixed), training duration (short: ≤12 weeks vs medium: 13-24 weeks vs long: >24 weeks), and types of control (active vs passive control) were not detected. The network meta-analysis results indicated that the optimal intervention components varied across different outcomes, with multimodal exercise combining cognitive training demonstrated the greatest effects among all other combined or single component interventions in improving global cognition.

Conclusions: This review suggests the advantage of combined interventions over exercise with comparable effects when compared with cognitive interventions in the population with mild cognitive impairment and dementia. Full scale multi-arm randomized controlled trials to compare the effects of combined interventions with cognitive interventions are warranted.