Managing Agitation in the ED
Kudos to Dan for his thoughtful analysis of three papers highlighting aspects of agitation management in the emergency department.
Take-home points from Journal Club this week:
* 5mg of midazolam (versed) was significantly faster at causing sedation in violently agitated patients at a large academic ED, compared to 5mg of haldol and 2mg of lorazepam (ativan). Patients getting midazolam also had significantly faster times to recover from sedation, also the least incidence of sedation failure.
* Droperidol has an FDA black box warning for QT prolongation leading to torsades de pointes and death. While the association between droperidol use and QT prolongation is clear, the risk of sudden cardiac death as caused by droperidol is much less clear. Many centers still use droperidol to good effect; an ECG after administration and prior to discharge is strongly recommended (and, if possible, prior to use).
* In treating behavioral emergencies, the accepted and endorsed practice in this country is to confront the violent or uncooperative patient with a ‘show of force’ to obtain consent, and if that fails, forcible medication ensues. Covert administration of meds, while maybe expedient or supported by the patient’s family, threatens the therapeutic alliance and exposes the institution to liability and the practitioner to claims of criminal battery. There exists an opportunity for further study and policymaking, both within our institution and beyond.
An in-depth discuss of the three papers awaits you, below.
Nobay 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)7:744-9.
This was a prospective randomized trial of three IM agents for agitation control: 5mg versed (midazolam) vs. 2mg ativan (lorazepam) vs. 5mg haldol (haloperidol). The authors hypothesized that midazolam, because it’s water-soluble with great IM absorption, rapid onset, and lack of cardiovascular effects through that route, would be the fastest agent for sedation as well as for fastest for recovery.
This trial took place at an urban academic ED with a volume of 64,000 / year. Enrolling patients only when convenient (M-F, 8a-11p) they enrolled 111 patients over 2 years – just over one per week. Perhaps expanding enrollment to weekends and overnights, when the bars close, would capture more of the agitated population we deal with daily.
Also important to setting was demographics: the authors found between the three arms were largely similar, with about a third of patients in each group having known EtOH use, a quarter with known recreational drug use, and about half with prior psych histories. Other studies (reviewed in the ACEP clinical policy) differ wildly on this account – from a preponderance of meth users to 100% psych patients to studies in which alcohol intoxication was grounds for exclusion – so this paper was unusually applicable to our practice environment.
Their results supported their hypothesis, with midazolam’s time to sedation being 18.3 +/- 14 min , compared to 28.3 +/- 25 min for haldol and 32.2 +/- 20 min for lorazepam. Mean time to arousal was 82 +/- 66 min for versed, 126.5 min +/- 85 for haldol, and 217.2 +/- 107 min for lorazepam. There were no significant changes in BP, HR, RR, or O2 sat, though the haldol arm experienced 1 apneic patient and 1 hypotensive patient (both had full recoveries).
They note some significant differences (with reasonable confidence intervals) in time to sedation between midazolam and lorazepam (5.1-22.8 min at 95% CI) and midazolam to haldol (0.5 min to 19.3 min at 95% CI) – sure, it’s just minutes, but when you’re dealing with a violent patient every minute seems like an hour. If they had included more data rather than just reporting means, we might’ve been able to calculate the risk of not being sedated at 15, 30, 45 min, etc, which might’ve led to some other illuminating comparisons.
Also, it’s regrettable they didn’t include, in their calculation of sedation and arousal times, the 26% of patients receiving lorazepam, the 19% of haldol patients, and 17% of versed patients that needed “rescue” drugs because sedation was inadequate at 20 minutes. Since the rate of rescue mirrored the efficacy of the drugs, however, including this data would have only reinforced their findings. Though there was no significant difference between these groups, as an academic exercise you could say absolute risk reduction from haldol to versed was 0.02, so NNT = 1/0.02 = 50 — meaning, you’d have to treat 50 people with versed to prevent one inadequate sedation that would occur with haldol. Along similar lines, with an ARR of 0.09, you’d have to treat 11 patients with versed to prevent one rescue that would be needed for ativan.
All told, the groups regarding this as a well-designed and executed trial with conclusions very relevant to practice. Though further comparisons – particularly to the “5 and 2” haldol/ativan combo that is very popular, or to droperidol – will have to be abstracted from other studies (many of which are summarized in the ACEP clinical policy on psychiatric patients).
Kao et al. Droperidol, QT Prolongation, and Sudden Death: What Is the Evidence? Ann Emerg Med 2003;41(4):546-58. Paper.
This was an elegant academic assessment of the FDA black box warning on droperidol for QT prolongation and sudden cardiac death.
The authors begin by noting that we can’t easily measure torsades incidence and death, so we must ask: What is clinically significant QT prolongation? They survey the literature and note a bump of 60ms, or a baseline of 500ms, has some risk for torsades. They note other risks for torsades, beyond QTc, included being female, bradycardic below 50 bpm, having liver or kidney disease, CAD, and antifungal use.
The authors then proceed to review three trials, an abstract, 7 case reports, and the postmarketing surveillance data from MedWatch. What they uncover is:
An association exists between droperidol administration and QT prolongation. This association is dose-dependent and temporally related, but they cannot point to a dose- or temporal- relationship between droperidol and torsades. Such events are rare and there are no RCTs or observational trials to demonstrate the association between droperidol and life-threatening cardiac events (despite 30+ years of studies and clinical experience).
Where does that leave us? Well, we now leave the realm of science and enter that of hospital policy and risk management. The FDA strongly recommends ECG monitoring before and after droperidol administration, and it certainly should be done whenever possible. We could easily imagine scenarios where a straightforward sedation with droperidol turns into a protracted observation period or admission because of the discovery of a long QTc on ECG. Furthermore, if a violent patient meets an untimely end within a few days of discharge (which is never beyond the realm of possibility), your use of droperidol will come under scrutiny.
For the conspiracy theorists among us, it’s worth noting that many of the MedWatch case reports highlighting the toxicity of droperidol were delivered to the FDA on the same day in July 2001, well in violation of the 15 day period that manufacturers are supposed to report adverse outcomes. It’s also worth noting that the manufacturer of droperidol, Jannsen-Cilag, also made risperdol – which was a good deal more expensive than droperidol and also helps treat agitation.
Lewin et al. An Unusual Case of Subterfuge in the Emergency Department: Covert Administration of Antipsychotic and Anxiolytic Medications to Control an Agitated Patient. Ann Emerg Med. 2006;47:75-78.
Finally, the group discussed the Lewin paper , which detailed an interesting case involving covert administration of haldol and ativan in orange juice, to help coax a manic psychiatric patient into medical clearance prior to psych ED assessment. To summarize this article’s many complicated arguments for and against would do it a disservice; we can simply be thankful we work in EDs where medical clearance before psych assessment is rare. If you’re interested in helping establishing a policy in our institution, please ask – resources are available to you.
Posted
on Friday, December 18th, 2009 at 4:54 pm by Nick. Filed under
Post-Conference Letter, Psychiatry, Arrhythmias, Journal Club, Blog.
You may post a comment.
