LAUSR

BACKGROUND The number of persons living with dementia will increase. So far, pharmacological management is limited because of small effect sizes and side effects of the drugs. Therefore, it is important to assess non-pharmacological treatment options such as massage, which have nearly no side effects and are easy for caregivers to apply. OBJECTIVES To conduct a systematic review with meta-analysis, aiming to pool the evidence for the efficacy of manual massage for persons living with dementia. DESIGN A systematic review and meta-analysis. DATA SOURCES EMBASE, Medline, PubMed, PSYinfo, BIOSIS, EBM, PSYCINDEX, Osteopathic Research Web, and OSTMED.DR were searched, regardless of publication year, through August 2017. REVIEW METHODS Randomized controlled trials (RCTs) evaluating manual massage in patients with dementia with or without using (aromatic) oil were included. The intervention of the control group had to guarantee no physical contact between caregiver and patient. Only studies assessing behavioural and psychological symptoms of dementia, cognitive abilities or depressive symptoms with validated instruments were included. Two reviewers independently extracted data and assessed risk of bias using The Cochrane Collaboration's 'Risk of bias' tool. Continuous outcomes are given as standardized mean difference (SMD), with 95% confidence intervals (CI) if different scaling of outcome measurement was used, and as mean difference (MD), with 95% confidence intervals (CI) for identical scaling. Data were pooled using the random-effects model. Sensitivity analysis considered type of massage, oil and outcome. Funnel plots were performed. RESULTS Eleven RCTs, with a total of 825 persons living with dementia, were eligible for qualitative synthesis and nine for quantitative synthesis. Two studies, with a total of 95 participants, had a high risk of bias. A pooled analysis of the mean change showed a benefit of manual massage compared to the control group using the Cohen Mansfield Agitation Inventory (SMD = -0.56, 95% CI [-0.95, -0.17], P = 0.005), which included six studies with 395 participants, and using the Cornell Scale of Depression in Dementia (MD = -6.14 [-8.66, -3.61], P < 0.00001), which included three studies with 193 participants. No significant effect could be demonstrated using the Neuropsychiatric Inventory and Mini Mental State Examination. Subgroup analysis of 'acupressure' did not show significant group differences. CONCLUSIONS Manual massage may serve as a non-pharmacological strategy to improve behavioural and psychological symptoms in persons living with dementia. Thus, healthcare professionals and family caregivers should be encouraged to apply massage to their patients and relatives. More research is needed, however, to provide clearer recommendations with respect to frequency and types of massage.