To assess any benefit or harm, we conducted a systematic review of randomised clinical trials (RCTs) allocating adults to dexmedetomidine versus placebo/no intervention for the prevention of delirium in intensive care or post-operative care units. We searched Medline, Embase, CENTRAL and other databases. The last search was 9 April 2022. Literature screening, data extraction and risk of bias volume 2 assessments were performed independently and in duplicate. Primary outcomes were occurrences of serious adverse events (SAEs), delirium and all-cause mortality. We used meta-analysis, Trial Sequential Analysis, and GRADE (Grading Recommendations Assessment, Development and Evaluation). Eighty-one RCTs (15,745 patients) provided data for our primary outcomes. Results from trials at low risk of bias showed that dexmedetomidine may reduce the occurrence of the most frequently reported SAEs (relative risk [RR] 0.69; 95% CI 0.43-1.09), cumulated SAEs (RR 0.70; 95% CI 0.52-0.95) and the occurrence of delirium (RR 0.62; 95% CI 0.43-0.89). The certainty of evidence was very low for delirium. Mortality was very low in trials at low risk of bias (0.4% in the dexmedetomidine groups and 1.0% in the control groups) and meta-analysis did not provide conclusive evidence that dexmedetomidine may result in lower or higher all-cause mortality (RR 0.47; 95% CI 0.18-1.21). There was a lack of information from trial results at low risk of bias for all primary outcomes. Trial results at low risk of bias showed that dexmedetomidine might reduce occurrences of SAEs and delirium, while no conclusive evidence was found for effects on all-cause mortality. The certainty of evidence ranged from very low for occurrence of delirium to low for the remaining outcomes.