A Cochrane review included 5 studies with a total of 334 patients with dementia. Approximately 85% of the patients had a diagnosis of Alzheimer disease. The primary goal of 4 studies was to measure the effects of melatonin on sleep disorders as compared with placebo. Secondary outcomes included changes in cognitive and non-cognitive function, depressive and neuropsychiatric symptoms, and functions in activities of daily living. There were non-significant effects revealed from the pooled estimates from 3 trials (n=268, treatment time 4-7 weeks) of MMSE cognitive, and ADAS-cognitive change scores (WMD 0.29, 95% CI -0.63 to 1.22). Individual study estimates demonstrated a significant improvement for melatonin in behavioural and affective symptoms as measured by the ADAS non-cognitive scale in a study of 20 patients and in the Neuropsychiatric Inventory (NPI) (WMD -3.48, 95% CI -4.89 to -2.07), but not with 10 mg/day (IR) melatonin in a larger study of 157 patients (WMD 0.63, 95% CI -4.58 to 5.84). The remainder of the treatment effects for affect, behaviour and activities of daily living were non-significant.
Comment: The quality of evidence is downgraded by indirectness (differences in studied patients and outcomes) and by imprecise results (limited study size for each comparison).