QUESTIONS: What is the effect of physical exercise on cognitive decline and behavioural problems in people with mild cognitive impairment (MCI) or dementia? What is the effect of physical exercise on particular domains of cognitive function? How do training protocols and patients' characteristics influence the outcomes? DESIGN: Systematic review and meta-analysis of randomised trials. PARTICIPANTS: People with MCI or dementia as their primary diagnosis. INTERVENTION: Physical exercise. OUTCOME MEASURES: Cognitive function including global cognition, memory, executive function, reasoning, attention, language, and behavioural problems. RESULTS: Forty-six trials involving 5099 participants were included in this review. Meta-analysis of the data estimated that aerobic exercise reduced the decline in global cognition, with a standardised mean difference (SMD) of 0.44, 95% CI 0.27 to 0.61, I2 = 69%. For individual cognitive functions, meta-analysis estimated that exercise lessened working memory decline (SMD 0.28, 95% CI 0.04 to 0.52, I2 = 40%). The estimated mean effect on reducing the decline in language function was favourable (SMD 0.17), but this estimate had substantial uncertainty (95% CI -0.03 to 0.36, I2 = 67%). The effects of exercise on other cognitive functions were unclear. Exercise also reduced behavioural problems (SMD 0.36, 95% CI 0.07 to 0.64, I2 = 81%). CONCLUSION: Physical exercise can reduce global cognitive decline and lessen behavioural problems in people with MCI or dementia. Its benefits on cognitive function can be primarily attributed to its effects on working memory. Aerobic exercise at moderate intensity or above and a total training duration of > 24 hours can lead to a more pronounced effect on global cognition.

Full text (sometimes free) may be available at these link(s):      help

• DOI
• PubMed
• PDF locator
• publisher