Journal Article Annotations
2021, 3rd Quarter
Annotations by Natalie O. Fedotova, MD, PhD and Jordan H. Rosen, MD
The authors performed a post-hoc analysis on the 1495 critically-ill adults who were delirium-free at ICU admission and randomized to receive IV Haldol (1 mg vs. 2mg q8h) vs placebo prophylactically (REDUCE trial). Patients were monitored for delirium three times daily and, once a patient developed delirium, clinical teams were encouraged to begin open-label symptom-driven haloperidol treatment starting at 2mg TID (up to a max of 5mg). Mortality benefits were seen for the patient subgroup who developed delirium in the 28 days following ICU admission (“incident delirium”; n=542, 36%). For delirium-positive patients only, each milligram of treatment haloperidol was associated with a 7% decrease in mortality at 28 days. This association held when total haloperidol exposure (preventive and treatment) and treatment arm allocation were taken into account. The mortality benefit was stronger for haloperidol administered earlier in the ICU course: at 90 days, mortality benefits declined to 3% per mg of haloperidol.
Strength and weaknesses:
This study focused on data gathered via a randomized, placebo-controlled trial across multiple settings and prospectively included multiple potential confounders in their models. When patients developed delirium, although clinical ICU teams were encouraged start haloperidol treatment, more than 10% never received treatment. Further, as RASS scores are not available, it is difficult to comment on potential differences in dosing related to agitation, hyperactive vs. hypoactive subtypes, and overall patient frailty/illness severity that were not captured in the models. It bears noting that in a prior secondary cohort analysis of the REDUCE trial, Duprey et al. (2020) did not find that incident delirium or the number of days spent with delirium were associated with mortality at 28 or 90 days, which raises questions about how management with haloperidol reduces mortality.
This adds to the mixed literature on haloperidol’s effect on delirium in critical illness and calls to mind haloperidol’s immune-modulating effects for managing delirium in critically ill patients. Given that this is a post-hoc analysis, it would be important to replicate in a second trial as the primary endpoint. These results are not strong enough to justify a change in practice.