Objective: The aim of this study was to systematically review the applicability (attendance, achieved intensity, adverse events) and effects of physical exercise on physical functions, cognitive functions, and activities of daily living among people with dementia.Design: Randomized controlled trials were identified in PubMed, the Cumulative Index to Nursing and Allied Health, the Allied and Complementary Medicine Database, and the Cochrane Library on August 30 and September 1, 2010, according to predefined inclusion criteria. Two reviewers independently extracted predetermined data and assessed methodologic quality.Results: A qualitative analysis was performed, including ten studies. Most participants were people with Alzheimer disease in residential care facilities. Four studies reached "moderate" methodologic quality, and six reached "low." The studies of moderate quality evaluated the effects of combined functional weight-bearing exercise, combined functional and nonfunctional exercise, and walking exercise.Conclusions: Among older people with Alzheimer disease in residential care facilities, combined functional weight-bearing exercise seems applicable for use regarding attendance and adverse events, and there is some evidence that exercise improves walking performance and reduces the decline in activities of daily living. Furthermore, there is some evidence that walking exercise performed individually reduces decline in walking performance, but adverse events need to be evaluated. Among older people with various types of dementia disorders who are staying in a hospital, there is some evidence that combined functional and nonfunctional exercise over 2 wks has no effect on mobility. It seems important that the interventions last for at least a few months and that the exercises are task-specific and are intended to challenge the individual's physical capacity. Among older people with unspecified dementia disorders in residential care facilities, there is some evidence that walking exercise performed at a self-selected speed has no effect on cognitive functions. Whether physical exercise can improve cognitive functions among people with dementia remains unclear because studies evaluating this have either been of low methodologic quality or used an intervention of presumably insufficient intensity. There is a need for more studies of high methodologic quality, especially among people with dementia disorders other than Alzheimer disease.