Evidence describing the association between hypnotics use and dementia risk is conflicting. It is unknown if the controversy is related to the type or dose of hypnotics or if hypnotics affect different populations.

We sought to derive lessons learned and future projections based on evidence from longitudinal studies.

In the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort, 1,543 older adults without dementia (mean age = 73.3 years, female = 45%) were followed for four years. The association between hypnotics and the risk of Alzheimer’s disease (AD) was investigated using Cox proportional hazards regressions. Next, electronic databases were searched until March 2022 to conduct the evidence synthesis of the associations of hypnotics with incident risk of dementia.

In the ADNI cohort, ever use of hypnotics was associated with an increased risk of AD (hazard ratio = 1.96, 95% confidence intervals = 1.23–3.11, p < 0.01). This association was significant for benzodiazepines and Z-drugs but not for melatonin. The association was stronger in long-term (more than one year) users and those with high cumulative doses. A meta-analysis of 26 longitudinal studies with 3,942,018 participants revealed a correlation between the use of hypnotics and the risk of dementia (relative risk = 1.23, 95% confidence intervals = 1.13–1.33, p < 0.001, median risk difference = 4%). It is a linear dose-response relationship, if a person takes the daily recommended dose for 100 days, their risk of developing dementia increases by 5% relative to non-users. According to subgroup analyses, neither association was significant among patients with a history of insomnia.

Individuals who use hypnotics, especially high-dose or long-term users, are at a higher risk of dementia and AD. The main issue with conclusion credibility is heterogeneity.