Midazolam vs Haloperidol vs Lorazepam in the Chemical Restraint of Agitated Patients

Abstract & Commentary

Source: Nobay F et al. A prospective, double-blind, randomized trial of midazolam versus haloperidol versus lorazepam in the chemical restraint of violent and severely agitated patients. Acad Emerg Med 2004;11:744-749.

The pharmacologic profile of midazolam, a short-acting benzodiazepine, led the authors to hypothesize that this drug would compare favorably with lorazepam and haloperidol in the sedation of violent and severely agitated patients in the ED. All patients in the study were violent and/or agitated, and were eligible if they initially did not respond to physical restraints. Exclusion criteria included appropriate hemodynamic and cardiorespiratory parameters, age younger than 18 years, pregnancy, or study drug allergy. A convenience sample of 111 patients was randomized to receive a single intramuscular injection of lorazepam 2 mg (27 patients), haloperidol 5 mg (42 patients), or midazolam 5 mg (42 patients); interim analysis demonstrated that lorazepam was associated with significantly longer times to sedation and awakening; therefore, it was removed as a study arm. Rescue drugs could be given at the discretion of the treating physician after 20 minutes (leading to termination of that patient’s enrollment).

The mean time to sedation was faster in the midazolam group (18 ±14 min) than in the lorazepam (32 ± 20 min) or haloperidol (28 ± 25 min) groups. Correspondingly, the mean time to arousal was shorter with midazolam (82 ± 66 min) than with lorazepam (217 ± 107 min) or haloperidol (126 ± 85 min). As mentioned, lorazepam was dropped from the study due to poorer performance, but analysis of the difference between times to sedation and recovery with midazolam vs. haloperidol favored the former: 9.9 min (95%, CI = 0.5-19.3, P = 0.04) for time to sedation, and 44.6 min (95%, CI 9-80, P = 0.025) for time to arousal. A rescue drug was administered to seven patients in the lorazepam and midazolam groups, and eight patients in the haloperidol group, suggesting equivalent efficacy among the agents. Cardiorespiratory monitoring of patients during the study revealed no difference in key parameters. The authors conclude that intramuscular midazolam is superior to haloperidol and lorazepam in the sedation of violent and/or agitated patients with regard to time to sedation and arousal.

Commentary by Richard Harrigan, MD

Most ED physicians either have given or received the patient at shift turnover who has yet to wake up from chemical sedation in the ED necessitated by violent, agitated behavior. As the incoming physician, it is always disconcerting to receive such patients at sign-out. Pessimistically, the incoming team ruminates about the potential for developing occult pathology masked by this type of sedation, while the outgoing team reassures them how wild the patient was, and that "s/he will be fine after s/he wakes up." Yet, it is well known how important and valuable chemical sedation is to the expedited management of the violent/agitated patient, so the quest for a better mousetrap continues.

And why not use midazolam? The authors provide literature that supports its safety and efficacy in this domain, and indeed it seems to have performed well in this study. One-size-fits-all dosing could be criticized, but as the authors point out, accurate weight-based dosing is not feasible when confronted with the violent/agitated patient. Prior survey work by the authors had revealed that haloperidol and lorazepam were "often used as single sedating agents." They stopped short of saying they were the most often used agents—although they are in my experience. Prior research has shown that a combination of haloperidol and lorazepam works better than either one alone in achieving adequate sedation of violent, agitated ED patients, but time to arousal was not specifically addressed in this study.1

It is worth noting that the relatively large standard deviations from the mean times to sedation and arousal suggest a lot of overlap here. Similarly, the data suggest that midazolam brings sedation roughly 10 minutes faster than haloperidol and achieves return to the arousal state approximately 45 minutes faster than haloperidol. However, the confidence intervals are wide and again suggest there is much overlap. A larger study that will evaluate times to sedation and arousal in three groups (midazolam, haloperidol, and midazolam plus haloperidol) would be interesting.

Dr. Harrigan, Associate Professor of Emergency Medicine, Temple University Hospital and School of Medicine, Philadelphia, PA, is Editor of Emergency Medicine Alert.


1. Battaglia J et al. Haloperidol, lorazepam, or both for psychotic agitation? A multicenter, prospective, double-blind, emergency department study. Am J Emerg Med 1997;15:335-340.