ObjectivesTo investigate the effects of a high-intensity functional exercise program on independence in activities of daily living (ADLs) and balance in older people with dementia and whether exercise effects differed between dementia types.DesignCluster-randomized controlled trial: Umeå Dementia and Exercise (UMDEX) study.SettingResidential care facilities, Umeå, Sweden.ParticipantsIndividuals aged 65 and older with a dementia diagnosis, a Mini-Mental State Examination score of 10 or greater, and dependence in ADLs (N = 186).InterventionNinety-three participants each were allocated to the high-intensity functional exercise program, comprising lower limb strength and balance exercises, and 93 to a seated control activity.MeasurementsBlinded assessors measured ADL independence using the Functional Independence Measure (FIM) and Barthel Index (BI) and balance using the Berg Balance Scale (BBS) at baseline and 4 (directly after intervention completion) and 7 months.ResultsLinear mixed models showed no between-group effect on ADL independence at 4 (FIM=1.3, 95% confidence interval (CI)=−1.6–4.3; BI=0.6, 95% CI=−0.2–1.4) or 7 (FIM=0.8, 95% CI=−2.2–3.8; BI=0.6, 95% CI=−0.3–1.4) months. A significant between-group effect on balance favoring exercise was observed at 4 months (BBS=4.2, 95% CI=1.8–6.6). In interaction analyses, exercise effects differed significantly between dementia types. Positive between-group exercise effects were found in participants with non-Alzheimer's dementia according to the FIM at 7 months and BI and BBS at 4 and 7 months.ConclusionIn older people with mild to moderate dementia living in residential care facilities, a 4-month high-intensity functional exercise program appears to slow decline in ADL independence and improve balance, albeit only in participants with non-Alzheimer's dementia.