Per the 2018 Society of Critical Care Medicine’s Clinical Practice Guidelines, the panel recommends against using haloperidol, an atypical antipsychotic, dexmedetomidine, a β-Hydroxy β-methylglutaryl-Coenzyme A (HMG-CoA) reductase inhibitor, or ketamine to prevent delirium in all critically ill adults (conditional recommendation, very low to low quality of evidence). Additionally, routine use of haloperidol, an atypical antipsychotic, or an HMG-CoA reductase inhibitor should be avoided for the treatment of delirium. On the other hand, dexmedetomidine may be considered in mechanically ventilated adults with delirium where agitation precludes weaning/extubation (conditional recommendation, low quality of evidence). Of note, the recommendation for dexmedetomidine is based on a single placebo-controlled randomized trial; thus, there is no direct comparison with antipsychotics in this setting. 
A 2020 meta-analysis compared the efficacy and safety of dexmedetomidine and other agents for the management of delirium, with sub-analyses available specifically for the intensive care unit (ICU) setting. A total of 108 randomized controlled studies were included for analysis, with 24 (n= 1,846) focusing on treatment and 84 (n= 28,089) focusing on prevention. For treatment in ICU patients, the response rate for dexmedetomidine had an odds ratio (OR) of 2.66 (95% confidence interval [CI] 1.05 to 6.77). When compared to antipsychotics, quetiapine surpassed dexmedetomidine in OR, but the confidence interval was widely distributed, indicating possible heterogeneity within the studies (OR 8.00; 95% CI 1.41 to 45.41). Dexmedetomidine demonstrated a higher OR compared to lorazepam (OR 2.03), ziprasidone (OR 1.23) ondansetron (OR 1.25), haloperidol (OR 1.01), rivastigmine (OR 0.87), and midazolam (OR 0.28). 
For prevention of delirium in ICU surgical patients, dexmedetomidine reported an OR of 0.46 (95% CI 0.32 to 0.66), which favors the active drug. Only risperidone, combination diazepam/flunitrazepam/pethidine, and combination dexmedetomidine/acetaminophen report lower ORs compared to dexmedetomidine. For prevention of delirium in ICU medical patients, dexmedetomidine (OR 0.50; 95% CI 0.04 to 6.30) was only surpassed by trazodone, ramelteon, and suvorexant, although there were noticeably fewer antipsychotics for comparison. The quality of studies reporting on dexmedetomidine is generally rated moderate to high. Despite the perceived benefit of reducing delirium in the ICU setting, dexmedetomidine also observed a higher dropout rate which questions the tolerability of adverse events. Yet the authors also do not suggest routine use of antipsychotics and have established a preference for dexmedetomidine. 
A 2019 review discussing the use of dexmedetomidine in the management of delirium in intensive care units noted a favorable safety profile of dexmedetomidine compared to that of antipsychotics (QTc prolongation, extrapyramidal effects, and sedation). Other outstanding differentiating features included sedative effects without respiratory depression and reduced need for benzodiazepines. Despite the lack of supporting clinical data, the authors proposed these superior qualities of dexmedetomidine over antipsychotic medications in the management of delirium. 
Several reviews evaluated studies that compared the role of dexmedetomidine with other sedatives, analgesics, or antipsychotics for the prevention and treatment of delirium in ICU patients. A randomized, open-label, parallel-group pilot study that compared the effect of dexmedetomidine with haloperidol in the treatment of ICU delirium is commonly referred to when evaluating the effect of dexmedetomidine versus other antipsychotics. Dexmedetomidine, in comparison with haloperidol, significantly shortened the median time to extubation (19.9 vs. 42.5 hours; p= 0.016) and decreased median ICU length of stay (1.5 vs. 6.5 days; p= 0.004) in mechanically ventilated patients (Table 1). Based on the findings from this study, it was suggested that dexmedetomidine might be an effective agent in the treatment of ICU-associated delirium agitation. [4-6]
A 2020 review evaluated data from 10 systematic reviews on the pharmacological treatment of delirium in both intensive and non-intensive care settings to provide treatment alternatives to haloperidol or other standards of care in patients with COVID-19. Only quetiapine and dexmedetomidine demonstrated possible benefits in ICU settings among all antipsychotics, benzodiazepines, and other pharmacological agents being evaluated. However, a direct comparison between quetiapine and dexmedetomidine was not provided, and the overall quality of evidence remained low. [7,8]
A 2019 meta-analysis attempted to discern the best pharmacologic intervention for the management and prevention of delirium, non-specific to the ICU. A total of 58 studies were included for analysis (20 were treatment for delirium while 38 were preventative for delirium). A subanalysis of ICU patients was not performed. The OR for a treatment response favoring dexmedetomidine versus placebo/control was 2.06 (95% CI 0.51 to 8.34), which was surpassed by other antipsychotic agents, including haloperidol, olanzapine, ziprasidone, and quetiapine but not surpassed by risperidone or haloperidol+rivastigmine. For prevention, dexmedetomidine reported a lower delirium occurrence rate compared to placebo/control (OR 0.50; 95% CI 0.31 to 0.80), which was surpassed by risperidone and olanzapine but not surpassed by haloperidol and midazolam. Using surface under the cumulative ranking curve (SUCRA) to compare the outcomes with an optimally imaginative intervention, dexmedetomidine and midazolam are considered the most effective preventative regimens while associated with the least increase in overall mortality. Ramelteon is still considered the most effective prophylactic agent (OR 0.07; 95% CI 0.01 to 0.66), while haloperidol+lorazepam may be the most effective treatment regimen (OR 28.13; 2.38 to 333.08). The included trials were characterized by underpowered statistics and heterogeneity among baseline characteristics. While the results favor ramelteon, the majority of evidence was published by a single network of investigators.