Pharmacologic interventions for prevention of delirium in hospitalized older people: A meta-analysis

To comprehensively assess the effects of pharmacologic interventions for prevention of delirium in hospitalized older people.

A systematic review with meta-analysis following Preferred Reporting Items for Systematic reviews and Meta-Analyses methodology was performed. Hospitalized people aged 65 and older, recruited to randomized controlled clinical trials. The electronic databases MEDLINE, EMBASE, WOS and Cochrane Central Register of Controlled Trials were consulted (March 2019). Predefined criteria were used to determine inclusion of studies and to assess their methodologic quality.

1855 records were identified in the database, and after removing the duplicates, the titles and abstracts evaluated were 1250 records. Finally, 25 randomized controlled trials contributed to meta-analysis (n = 5820): 1 anti-epileptics (n = 697), 2 anti-inflammatories (n = 615), 4 antipsychotics (n = 1193), 2 cholinesterase inhibitors (n = 87), 13 hypnotics/sedatives (n = 2909), 1 opioids (n = 52), 1 psychostimulants/nootropics (n = 81), 1 yokukansan (n = 186). Olanzapine (RR = 0.36; 95 %CI: 0.24, 0.52; k = 1; n = 400), rivastigmine (RR = 0.36; 95 %CI: 0.15, 0.87; k = 1; n = 62), dexmedetomidine (RR = 0.52; 95 %CI: 0.38, 0.71; I² = 55 %; k = 6; n = 2084), and ramelteon (RR = 0.09; 95 %CI: 0.01, 0.64; k = 1; n = 65) reduced the incidence of delirium compared to placebo/usual care. Only dexmedetomidine was also associated with a shorter duration of delirium (0.70 days reduction) and a lower consumption of psychotropic drugs (48 %). No effect was found in mortality, adverse events, urinary tract infections or post-operative complications.

This meta-analysis suggests that dexmedetomidine is effective in reducing the incidence and duration of delirium in hospitalized older patients. Individual studies reveal effects of ramelteon, olanzapine and rivastigmine on the incidence of delirium but the evidence is insufficient to draw a robust conclusion.