ObjectiveBenzodiazepines and related drugs (BDZRs) have been associated with an increased risk of Alzheimer's disease (AD) in later life. Despite this, it remains unclear whether ongoing BDZR use may further accelerate cognitive decline in those diagnosed with mild to moderate AD. DesignThis study was embedded within NILVAD, a randomized controlled trial of nilvadipine in mild to moderate AD. Cognition was measured at baseline and 18 months using the Alzheimer Disease Assessment Scale, Cognitive Subsection (ADAS-Cog). We assessed predictors of long-term BDZR use and analyzed the effect of ongoing BDZR use on ADAS-Cog scores at 18 months. Additionally, the impact of BDZR use on adverse events, incident delirium, and falls over 18-month follow-up was assessed adjusting for relevant covariates. Setting and Participants448 participants with mild to moderate AD recruited from 23 academic centers in 9 European countries. ResultsOverall, 14% (62/448) were prescribed an ongoing BDZR for the study duration. Increasing total number of (non-BDZR) medications was associated with a greater likelihood of BDZR prescription (odds ratio 1.16, 95% confidence interval 1.05-1.29). At 18 months, BDZR use was not associated with greater cognitive decline on the ADAS-Cog controlling for baseline ADAS-Cog scores, age, gender, study arm, and other clinical covariates (β = 1.62, −1.34 to 4.56). However, ongoing BDZR use was associated with a greater likelihood of adverse events [incidence rate ratio (IRR) 1.19, 1.05-1.34], incident delirium (IRR 2.31, 1.45-3.68), and falls (IRR 1.66, 1.02-2.65) over 18 months that persisted after robust adjustment for covariates. Conclusions and ImplicationsThis study found no effect of ongoing BDZR use on ADAS-Cog scores in those with mild to moderate AD over 18 months. However, ongoing use of these medications was associated with an increased risk of adverse events, delirium, and falls. Thus, BDZR use should be avoided where possible and deprescribing interventions should be encouraged in older adults with AD.