BACKGROUND: Sleep disturbances, including reduced nocturnal sleep time, sleep fragmentation, nocturnal wandering, and daytime sleepiness are common clinical problems in dementia, and are associated with significant carer distress, increased healthcare costs, and institutionalisation. Although non-drug interventions are recommended as the first-line approach to managing these problems, drug treatment is often sought and used. However, there is significant uncertainty about the efficacy and adverse effects of the various hypnotic drugs in this clinically vulnerable population. OBJECTIVES: To assess the effects, including common adverse effects, of any drug treatment versus placebo for sleep disorders in people with dementia. SEARCH METHODS: We searched ALOIS (www.medicine.ox.ac.uk/alois), the Cochrane Dementia and Cognitive Improvement Group's Specialized Register, on 19 February 2020, using the terms: sleep, insomnia, circadian, hypersomnia, parasomnia, somnolence, rest-activity, and sundowning. SELECTION CRITERIA: We included randomised controlled trials (RCTs) that compared a drug with placebo, and that had the primary aim of improving sleep in people with dementia who had an identified sleep disturbance at baseline. DATA COLLECTION AND ANALYSIS: Two review authors independently extracted data on study design, risk of bias, and results. We used the mean difference (MD) or risk ratio (RR) with 95% confidence intervals (CI) as the measures of treatment effect, and where possible, synthesised results using a fixed-effect model. Key outcomes to be included in our summary tables were chosen with the help of a panel of carers. We used GRADE methods to rate the certainty of the evidence. MAIN RESULTS: We found nine eligible RCTs investigating: melatonin (5 studies, n = 222, five studies, but only two yielded data on our primary sleep outcomes suitable for meta-analysis), the sedative antidepressant trazodone (1 study, n = 30), the melatonin-receptor agonist ramelteon (1 study, n = 74, no peer-reviewed publication), and the orexin antagonists suvorexant and lemborexant (2 studies, n = 323). Participants in the trazodone study and most participants in the melatonin studies had moderate-to-severe dementia due to Alzheimer's disease (AD); those in the ramelteon study and the orexin antagonist studies had mild-to-moderate AD. Participants had a variety of common sleep problems at baseline. Primary sleep outcomes were measured using actigraphy or polysomnography. In one study, melatonin treatment was combined with light therapy. Only four studies systematically assessed adverse effects. Overall, we considered the studies to be at low or unclear risk of bias. We found low-certainty evidence that melatonin doses up to 10 mg may have little or no effect on any major sleep outcome over eight to 10 weeks in people with AD and sleep disturbances. We could synthesise data for two of our primary sleep outcomes: total nocturnal sleep time (TNST) (MD 10.68 minutes, 95% CI -16.22 to 37.59; 2 studies, n = 184), and the ratio of day-time to night-time sleep (MD -0.13, 95% CI -0.29 to 0.03; 2 studies; n = 184). From single studies, we found no evidence of an effect of melatonin on sleep efficiency, time awake after sleep onset, number of night-time awakenings, or mean duration of sleep bouts. There were no serious adverse effects of melatonin reported. We found low-certainty evidence that trazodone 50 mg for two weeks may improve TNST (MD 42.46 minutes, 95% CI 0.9 to 84.0; 1 study, n = 30), and sleep efficiency (MD 8.53%, 95% CI 1.9 to 15.1; 1 study, n = 30) in people with moderate-to-severe AD. The effect on time awake after sleep onset was uncertain due to very serious imprecision (MD -20.41 minutes, 95% CI -60.4 to 19.6; 1 study, n = 30). There may be little or no effect on number of night-time awakenings (MD -3.71, 95% CI -8.2 to 0.8; 1 study, n = 30) or time asleep in the day (MD 5.12 minutes, 95% CI -28.2 to 38.4). There were no serious adverse effects of trazodone reported. The small (n = 74), phase 2 trial investigating ramelteon 8 mg was reported only in summary form on the sponsor's website. We considered the certainty of the evidence to be low. There was no evidence of any important effect of ramelteon on any nocturnal sleep outcomes. There were no serious adverse effects. We found moderate-certainty evidence that an orexin antagonist taken for four weeks by people with mild-to-moderate AD probably increases TNST (MD 28.2 minutes, 95% CI 11.1 to 45.3; 1 study, n = 274) and decreases time awake after sleep onset (MD -15.7 minutes, 95% CI -28.1 to -3.3: 1 study, n = 274) but has little or no effect on number of awakenings (MD 0.0, 95% CI -0.5 to 0.5; 1 study, n = 274). It may be associated with a small increase in sleep efficiency (MD 4.26%, 95% CI 1.26 to 7.26; 2 studies, n = 312), has no clear effect on sleep latency (MD -12.1 minutes, 95% CI -25.9 to 1.7; 1 study, n = 274), and may have little or no effect on the mean duration of sleep bouts (MD -2.42 minutes, 95% CI -5.53 to 0.7; 1 study, n = 38). Adverse events were probably no more common among participants taking orexin antagonists than those taking placebo (RR 1.29, 95% CI 0.83 to 1.99; 2 studies, n = 323). AUTHORS' CONCLUSIONS: We discovered a distinct lack of evidence to guide decisions about drug treatment of sleep problems in dementia. In particular, we found no RCTs of many widely prescribed drugs, including the benzodiazepine and non-benzodiazepine hypnotics, although there is considerable uncertainty about the balance of benefits and risks for these common treatments. We found no evidence for beneficial effects of melatonin (up to 10 mg) or a melatonin receptor agonist. There was evidence of some beneficial effects on sleep outcomes from trazodone and orexin antagonists and no evidence of harmful effects in these small trials, although larger trials in a broader range of participants are needed to allow more definitive conclusions to be reached. Systematic assessment of adverse effects in future trials is essential.