Sphemerg.ca

A Prospective, Double-blind, Randomized Trial ofMidazolam versus Haloperidol versus Lorazepam inthe Chemical Restraint of Violent and SeverelyAgitated Patients Flavia Nobay, MD, Barry C. Simon, MD, M. Andrew Levitt, DO, AbstractObjectives: To determine if midazolam is superior to CI = 0.5 to 19.3 minutes). Time to arousal was 81.9 minutes lorazepam or haloperidol in the management of violent for patients receiving midazolam, 126.5 minutes for halo- and severely agitated patients in the emergency department.
peridol, and 217.2 minutes for lorazepam. Time to arousal Superiority would be determined if midazolam resulted in for midazolam was significantly shorter than for both a significantly shorter time to sedation and shorter time to haloperidol and lorazepam (p < 0.05). The mean difference arousal. Methods: This was a randomized, prospective, in time to awakening between midazolam and lorazepam double-blind study of a convenience sample of patients was 135.3 minutes (95% CI = 89 to 182 minutes) and that from an urban, county teaching emergency department.
between midazolam and haloperidol was 44.6 minutes (95% Participants included 111 violent and severely agitated CI = 9 to 80 minutes). There was no significant difference patients. Patients were randomized to receive intramuscular over time by repeated-measures analysis of variance be- midazolam (5 mg), lorazepam (2 mg), or haloperidol (5 mg).
tween groups in regard to changes in systolic and diastolic Results: The mean (6SD) age was 40.7 (613) years. The blood pressure (p = 0.8965, p = 0.9581), heart rate mean (6SD) time to sedation was 18.3 (614) minutes for (p = 0.5517), respiratory rate (p = 0.8191), and oxygen sat- patients receiving midazolam, 28.3 (625) minutes for uration (p = 0.8991). Conclusions: Midazolam has a signif- haloperidol, and 32.2 (620) minutes for lorazepam. Mid- icantly shorter time to onset of sedation and a more rapid azolam had a significantly shorter time to sedation than time to arousal than lorazepam or haloperidol. The efficacies lorazepam and haloperidol (p < 0.05). The mean difference of all three drugs appear to be similar. Key words: agitation; between midazolam and lorazepam was 13.0 minutes (95% emergency department; restraint; midazolam; haloperidol; confidence interval [95% CI] = 5.1 to 22.8 minutes) and that lorazepam. ACADEMIC EMERGENCY MEDICINE 2004; between midazolam and haloperidol was 9.9 minutes (95% Violent and severely agitated patients (VSAPs) in the more staff members injured due to violence during emergency department (ED) are a common and the same period.1 The urban, county ED where this serious problem. In a survey of EDs, 32% reported study was based provides care for a sizable substance receiving at least one verbal threat daily, 18% reported and alcohol abusing population and also offers med- at least one threat with a weapon daily, and 25% ical clearance services for county emergency psychi- reported restraining at least one patient each day.
atric patients. Consequently, our ED encounters Seven percent of surveyed institutions had experi- a significant number of VSAPs for a variety of medical enced a death secondary to violent behavior in the ED and other reasons. Most VSAPs in our ED require within the previous five years, and 80% had one or physical and chemical restraint for the protection ofthemselves, other patients, and the ED staff.
While physical restraints are often necessary for From the Division of Emergency Medicine, Department of Internal VSAPs, some may also require augmentation of their Medicine, University of California, San Francisco (FN), San Fran- restraints by chemical sedation. A number of studies cisco, CA; and Department of Emergency Medicine, Alameda have documented the safety, efficacy, and expediency County Medical Center, Highland Campus (BCS, MAL, GMD), of chemical restraint.2–21 From that literature, buty- Oakland, CA.
Received November 15, 2002; revisions received March 20, 2003, and June 19, 2003; accepted June 23, 2003.
emerged as leaders in the pharmacologic arena for Presented at the SAEM annual meeting, Boston, MA, May 1999.
chemical restraint.2–8 Benzodiazepines (e.g., loraze- Address for correspondence and reprints: Barry C. Simon, MD, pam) have more recently been added to the list of safe Department of Emergency Medicine, Highland General Hospital, and effective options for chemical sedation.8–10 These 1411 East 31st Street, Oakland, CA 94602. Fax: 510-437-8322; e-mail:[email protected].
drugs used alone or in combination are effective in subduing patients who are a danger to themselves or ACAD EMERG MED d July 2004, Vol. 11, No. 7 d www.aemj.org others.12–14 Research has shown that combination vide consent. All patients who met inclusion criteria therapy (haloperidol and lorazepam) is better than during study hours were entered in the study.
single-drug therapy for the rapid sedation of agitatedpatients but may result in prolonged sedation.13 Study Setting and Population. This study was Additionally, when compared with benzodiazepines, performed at a university-affiliated, county teaching haloperidol has an increased frequency of untoward ED with an annual census of 64,000 patient visits. The effects such as extrapyramidal reactions and tachy- hospital supports a four-year emergency medicine cardia.13 The relatively longer half-lives of haloperidol residency program. The study lasted from December and lorazepam and the occasional need for additional 1997 to December 1999. The study population in- dosing can lead to excess sedation that delays medical cluded those patients who required emergency seda- evaluation, treatment, and disposition of the VSAP.
tion for the control of violent behavior or severe We sought to find a single sedating agent that would agitation (patients deemed to be physically threaten- result in a shorter time to arousal yet be equivalent or ing to themselves, deemed to be physically threaten- better than commonly used agents with respect to ing to those around them, with severely disruptive time to sedation. Rapid sedation is needed for obvious behavior, and with extreme psychomotor agitation).
reasons such as patient/staff safety, control of the ED All patients in our study were initially physically milieu, and prevention of increased agitation of other restrained. If the restraints were not sufficient to patients. Rapid time to arousal is desired to prevent mollify behavior and the patients continued to be delays in diagnosis and treatment and to limit staffing threatening, verbally abusive, and agitated, chemical restraints were administered. By our department’s Midazolam is a water-soluble benzodiazepine with protocols and care standards, all patients who met a pharmacologic profile that may be ideal for sedation inclusion criteria would have had chemical sedation of VSAPs. It has excellent intramuscular (IM) absorp- regardless of whether or not they participated in the tion, rapid time to onset, and a functional time to study. Patients were excluded from the study if they recovery of 30–120 minutes. When administered IM, it had known allergies to any of the drugs, were has very little effect on cardiorespiratory function, hypotensive (systolic blood pressure <90 mm Hg), thus eliminating the need for cardiac monitoring.15,16 were unable to maintain their own airway, were However, in elder patients and when combined with tachycardic ([140 beats/min) or bradycardic (<60 opioids, midazolam causes excess sedation; in these beats/min), were in respiratory distress (respiratory circumstances, some authorities recommend de- rate [40 breaths/min), were younger than 18 years of creasing the IM dose by as much as 50%.17–20 A sur- vey of emergency medicine residencies found thathaloperidol and lorazepam are often used as single Study Protocol. To determine the most commonly sedating agents (unpublished data, 1997). Therefore, used single sedating agents in settings such as ours, we chose to compare haloperidol and lorazepam with a mailed/faxed survey was sent to every academic ED across the country. The response rate was 59 of 108 Our hypothesis was that midazolam is superior in programs (54.6%). The medications used most fre- time to sedation and arousal than other commonly quently for sedation of VSAPs were haloperidol and used agents (haloperidol and lorazepam) when used lorazepam. Therefore, these were selected to compare as a single agent for sedation of VSAPs.
A convenience sample of patients was enrolled when a trained research coordinator was present inthe ED (Monday through Friday from 8 AM to 11 PM).
Study Design. This was a randomized, prospective, The research assistants were trained to observe and double-blind study that sought to compare midazolam record the time to sedation, time of patient recovery to with lorazepam and haloperidol in the sedation of an interviewable state, and to apply the sedation and VSAPs. Written consent was attempted for all patients.
In cases in which the patients were judged unable to The VSAPs who were to receive chemical sedation give consent, the need for emergency sedation, based were given one of the three drugs. The administered on the requirement of safety for the patient and the drugs were prepared in the pharmacy using a com- employees of the ED, was clearly documented by the puter-generated randomization code. The research physician. Consent was then obtained from family assistant, the administering physician, and the patient members when available. The institutional review were all blinded to the drug delivered. The doses of board approved the study with the understanding medication delivered were 5 mg midazolam, 5 mg that many patients would be judged unable to consent haloperidol, or 2 mg lorazepam. The dose of midazo- and no surrogate would be present. However, all three lam was chosen based on the Physicians’ Desk Reference drugs under study are routinely used to sedate agi- dosing guide for a 70-kg patient, and the doses of tated patients. When able, no patient refused to pro- haloperidol and lorazepam were based on published guidelines. All were delivered IM in a volume of 1 spiratory rate, and oxygen saturation, were analyzed mL. Considering the difficulty of establishing intra- with repeated-measures analysis of variance.
venous access in an agitated patient, an IM injection Due to limited available information on time to was considered a more practical and safe approach sedation and awakening with the study medications, to a VSAP. The efficacy of IM sedation has been an a priori sample size was not calculated. An interim analysis was conducted after 95 patients were en- After study medications were given, research assis- rolled. Based on these findings, a true sample size was tants trained to monitor patient levels of sedation and calculated using a b = 0.2 and #0.05 and a standard- arousal continuously observed the patients. Baseline ized effect size of 0.7. This effect size was used in vital signs were determined at time zero (time at sample size calculations. Pocock tables were used to initial IM dose), and every hour thereafter for the adjust sample size determination from this interim duration of the study period and at the time of analysis. A Bonferroni correction would be used for the multiple comparisons of data when p < 0.05 was Sedation was defined as when the patient exhibited no agitation and no longer required continuous super-vision by nursing or security to avoid self-injury. Athree-point scale modified from the study by Thomas et al. was used to determine when the patientachieved adequate sedation (Table 1).5 From the time Interim analysis with 95 patients showed that loraze- the study drug was administered until the time of pam demonstrated a statistically significantly longer arousal, patients were continuously observed, with time to sedation and time to awakening than mid- data collected every 15 minutes. Arousal was deter- azolam. Therefore, lorazepam was dropped from the mined to be complete if the patient awakened to randomization schedule, and the study continued verbal commands, was able to count backward from with midazolam and haloperidol. Based on the find- 10 to 1, and was able to follow simple commands.
ings of the interim analysis, it was determined that an Time to sedation and time to arousal were calculated additional seven patients would be needed in each from the time of the administration of the first dose of group to determine if a statistically significant differ- the study drug. If the patient continued to be disrup- ence existed between haloperidol and midazolam. A tive 20 minutes after the study drug was adminis- final total of 111 patients were entered into the study tered, a ‘‘rescue drug’’ could be given at the discretion protocol. There were no statistical differences demon- of the treating attending physician. Patient enrollment strated in the baseline demographics of the groups.
in the study was terminated if a rescue medication Recreational drug and alcohol use was self-reported was given. These patients were considered sedation failures, and their data were not included in the Twenty-seven patients received lorazepam, 42 pa- analysis. Patients were also monitored for any adverse tients received haloperidol, and 42 patients received midazolam. The mean (6SD) time to sedation was32.2 (620) minutes for patients receiving lorazepam,28.3 (625) minutes for haloperidol, and 18.3 (614) Data Analysis. Data were entered into Microsoft minutes for midazolam. The differences in time to Excel for Windows 98 and imported into Stat View sedation between the drugs are presented in Table 3.
5.0 (SAS Institute, Cary, NC) for statistical analysis.
The mean (6SD) time to arousal was 217.2 (6107) Categorical data are reported as absolute number and minutes for patients receiving lorazepam, 126.5 (685) percentage. Continuous data are reported as mean and minutes for haloperidol, and 81.9 (6 66) minutes for standard deviation. The differences in times to seda- midazolam. Times to awakening were significantly tion and times to awakening are reported as mean different between the medications (Table 3). Seven differences with 95% confidence intervals (95% CIs).
patients receiving lorazepam (26%), eight receiving Differences between groups were analyzed with the haloperidol (19%), and seven receiving midazolam chi-square test for categorical data and an analysis of (17%) needed rescue drugs (p = 0.6584).
variance for continuous data. Repeated-measures There were no statistically significant differences data, which included blood pressure, heart rate, re- over time in regard to change in systolic and diastolicblood pressure (p = 0.8965, p = 0.9581), heart rate(p = 0.5517), respiratory rate (p = 0.8191), and oxygen TABLE 1. Modified Thomas Combativeness Scale5 saturation (p = 0.8991) among patients receiving each of the medications. Two adverse events occurred (reviewed by the institutional review board), both in patients receiving haloperidol. One patient became hypotensive and another apneic after receiving halo- peridol, but both subsequently recovered fully.
ACAD EMERG MED d July 2004, Vol. 11, No. 7 d www.aemj.org TABLE 2. Demographics for the Three Study Groups All p-value comparisons were <0.05 and considered nonsignificant.
mens, combination therapy has also been evaluated.
Several studies have shown the superiority of a com- Violence in the ED can cause injuries to patients and bination of lorazepam and haloperidol.12–14 In a dou- staff and can delay the medical evaluation and treat- ble-blind, prospective trial of 98 VSAPs requiring ment of patients. Therefore, the ability to rapidly rapid tranquilization, Battaglia et al. showed that sedate VSAPs is essential to maintaining control in a combination of lorazepam and haloperidol was the ED. The use of chemical restraints in conjunction superior to either drug alone.13 Patients receiving with physical restraints has been studied widely. To the combination therapy scored better on all study out- best of our knowledge, no prospective studies have come measurements, including the Agitated Behavior analyzed the use of midazolam for sedation of VSAPs Scale, a modified Brief Psychiatric Rating Scale, a Clin- in the ED. Our study was designed to evaluate the ical Global Impressions Scale, and an Alertness Scale.
efficacy of midazolam versus haloperidol and loraze- However, this study did not take into consideration pam for VSAPs in the ED. The nature of these encoun- the time to patient return to an interviewable level of ters usually requires early intervention that precedes a final diagnosis or confirmation of the etiology of the Midazolam, a water-soluble short-acting benzodi- agitation; indeed, chemical restraint is partly aimed at azepine, is an excellent medication for conscious enabling a complete medical evaluation.
sedation and rapid sequence intubation. It has several In the mid-1980s, several studies examined the theoretical advantages over other drugs for use in efficacy of haloperidol to sedate VSAPs. Clinton et al.
sedating violent patients. It has rapid IM absorption studied 136 disruptive patients in the ED and found due to its lipophilic properties, a rapid onset of action, that haloperidol was an effective medication for predictable sedation levels, and a short half-life. A sedation.7 Disruptive behavior was alleviated within number of case reports and retrospective studies have 30 minutes in 83% of the patients, which is consistent found midazolam to be both safe and efficacious for with our findings. Another study confirmed these the sedation of patients in the ED.22–24 One study in results in 81 consecutive patients in an ED.6 Continu- psychiatric patients showed that midazolam was ing the search for faster and better medication regi- significantly more effective in controlling motor agi-tation than haloperidol.24 The time to arousal is an important property of any sedating drug used in the ED. Our previous experi- ence suggested that prolonged sedation led to anextended length of stay in the ED and a delay in the time to diagnosis. The short half-life characteristic of midazolam was therefore an attractive property for its use in VSAPs. Our study suggests that the pharma- cokinetics of midazolam are favorable in this regard; it is quickly and thoroughly absorbed IM, and its clinical duration of action is shorter than the other that droperidol is not used as frequently as our study drugs we evaluated. We found that 5 mg of midazo- drugs for control of agitated patients in the ED.
lam (IM) was more effective than haloperidol or We recognize that our sedation/arousal scale was lorazepam with respect to time to patient arousal.
not validated previously. However, there are no pre- The choice of a single predetermined dose as opposed viously validated scales that can be easily applied to to a weight-based dose was intended to simplify the the conditions and patients we are studying. The scale study and to facilitate clinical practice. Weighing or advanced by Thomas et al. seemed most clinically asking agitated patients their weight is usually not relevant to our purposes. We chose to modify the Thomas scale because it included data points that The time to sedation was not a major focus of this looked for subtle differences between groups; we were study. However, we did determine that the time to more interested in the end point of sedation and sedation was as favorable with midazolam as with the arousal and not subtle changes in level of conscious- other agents evaluated. Indeed, the data suggest that ness. We believe that the data points used in our scale time to sedation with midazolam was faster than with were practical, easy to apply, and objective.
Another limitation of our study is that there was no The attempt to find a clinically relevant and user- placebo control group. Because our study patients friendly scale to measure arousal and sedation is were extremely agitated, we believed that chemical challenging to ED clinical investigators. We looked at restraint was necessary for the safety of patients and a number of published and validated scales and staff. We believed it was ethically and practically found them to be more involved and detailed than impossible to use a placebo, given that these patients necessary for the purposes of our study. In the study would have been chemically sedated regardless of by Thomas et al.,5 a five-point sedation scale was used (see Table 1). Because we believed that points two and An additional limitation of the study is that patient four represented subtle differences that were difficult weight was not recorded and dosing was not adjusted to measure and did not offer additional clinical value, for weight. Our single-dose regimen was selected we modified the Thomas scale to create a three-point because of precedence in the literature and to facilitate scoring system. The arousal scale was designed to the protocol, especially as it related to the blinding of determine if patients could answer simple questions, the medications. In addition, the clinical situation thus enabling a focused medical history and physical eliminated the possibility of obtaining accurate weights in the majority of patients.
Although adverse events were rare in our study, It would have been interesting to obtain blood two patients had potentially serious but transient alcohol levels and toxicology screens on all patients reactions after receiving haloperidol (one became in both groups. However, it is not our practice to hypotensive, and another became apneic). Caution routinely obtain these levels in this patient popula- must be exercised when administering medication to tion. In our experience, many of these patients refuse patients before a thorough assessment can be com- to provide urine samples or to allow blood draws. In pleted. Subsequent to studies recommending haloper- addition, we find it difficult and sometimes danger- idol for the control of VSAPs, lorazepam was found ous to obtain blood and urine samples from agitated to be a practical option for sedating patients because it is associated with fewer side effects than haloper- In the future, research may be directed at validating idol.8–11 Studies comparing lorazepam with haloper- a scale and then applying it to a similar study.
idol suggest that lorazepam is equally efficacious but Including droperidol as a treatment arm would be does not have the same propensity for extrapyramidal advantageous if Food and Drug Administration con- cerns can be addressed. Weight-based dosing mightalso be studied to determine whether better efficacy This was a convenience sample of patients deter- mined and limited by the availability of our researchassistants. We did not collect demographics on pa- We found that midazolam is superior to haloperidol tients who were not entered into the study during off- and lorazepam in the sedation of VSAPs with respect study hours. A selection bias could therefore be to time to sedation and time to arousal. The use of present. The exclusion of droperidol as a treatment midazolam in the control of VSAPs can facilitate arm of the present study may be a theoretical limita- patient care, rapidly ease the disruption to the ED, tion. Some studies have found droperidol to be more and hasten disposition. These concerns have become effective in the control of agitated patients than both much greater in recent years given ED overcrowding haloperidol and lorazepam.5,8 However, our limited and the documentation and staffing demands re- preliminary survey (54.6% response rate) suggested quired on all restrained patients. With the goal of ACAD EMERG MED d July 2004, Vol. 11, No. 7 d www.aemj.org being safe, the effective, rapid, and short duration of 12. Garza-Trevino ES, Hollister LE, Overall JE, Alexander WF.
action of midazolam seems to be superior to haloper- Efficacy of combinations of intramuscular anti-psychotics and sedative-hypnotics for control of psychotic agitation. Am JPsychiatry. 1989; 146:1598–601.
13. Battaglia J, Moss S, Rush J, et al. Haloperidol, lorazepam, or The authors thank Matthew Waxman, Thomas Quigley, and Justin both for psychotic agitation? A multicenter, prospective, Young for their assistance with this study.
double-blind, emergency department study. Am J Emerg Med.
1997; 15:335–40.
14. Bieniek SA, Ownby RL, Penalver A, Dominguez RA. A double-blind study of lorazepam versus the combination of 1. Moore GP, Jackimczyk KC. The violent patient. In: Rosen P, haloperidol and lorazepam in managing agitation.
Barkin R (eds). Emergency Medicine Concepts and Clinical Practice. Volume 3. St. Louis, MO: CV Mosby, 1998.
15. Gerecke M. Chemical structure and properties of midazolam 2. Dubin WR, Feld JA. Rapid tranquilization of the violent compared with other benzodiazepines. Br J Clin Pharmacol.
patient. Am J Emerg Med. 1989; 7:313–20.
3. Resnick M, Burton BT. Droperidol versus haloperidol in the 16. Versed Product Manual. Nutley, NJ: Hoffman-La Roche, 1986.
initial management of acutely agitated patients. J Clin 17. Nishiyama T, Matsukawa T, Hanaoka K. The effects of age and gender on the optimal premedication dose of intramuscular 4. Lenehan GP, Gastfriend DR, Stetler C. Use of haloperidol in the midazolam. Anesth Analg. 1998; 86:1103–8.
management of agitated or violent, alcohol-intoxicated patients 18. Nordt SP, Clark RF. Midazolam: a review of therapeutic uses in the emergency department: a pilot study. J Emerg Nurs.
and toxicity. J Emerg Med. 1997; 15:357–65.
19. Taylor J, Kenneth S. Possible intramuscular 5. Thomas H, Schwartz E, Petrilli R. Droperidol vs. haloperidol midazolam-associated cardiorespiratory arrest and death. Ann for chemical restraint of agitated and combative patients. Ann 20. Wong HY, Fragen RJ, Dunn K. Dose-finding study of 6. Siverstein S, Frommer D, Marx J, et al. Parenteral haloperidol in intramuscular midazolam preanesthetic medication in the combative patients; a prospective study [abstract]. Ann Emerg elderly. Anesthesiology. 1991; 74:675–9.
21. Deschamp C. Calming the demon within: chemical vs. physical 7. Clinton JE, Sterner S, Stelmachers Z, Ruiz E. Haloperidol for restraint. J Emerg Med. 1994; 19:49–53.
sedation of disruptive emergency patients. Ann Emerg Med.
22. Wright SW, Chudnofsky CR, Dronen SC, Wright MB, Borron SW. Midazolam use in the emergency department. Am J Emerg 8. Richards JR, Derlet RW, Duncan DR. Chemical restraint for the agitated patient in the emergency department: lorazepam 23. Mendoza R, Djenderedjian AH, Adams J, Ananth J. Midazolam versus droperidol. J Emerg Med. 1998; 16:567–73.
in acute psychotic patient with hyperarousal. J Clin Psychiatry.
9. Ward ME, Saklad SR, Ereshefsky L. Lorazepam for the treatment of psychotic agitation. Am J Psychiatry. 1986; 143: 24. Wyant M, Diamond BI, O’Neal E, Sloan A, Borison RL. The use of midazolam in acutely agitated psychiatric patients.
10. Medell JG. Further experience and observations with Psychopharmacol Bull. 1990; 26:126–9.
lorazepam in the management of behavioral agitation. J Clin 25. Foster S, Kessel J, Berman ME, Simpson GM. Efficacy of lorazepam and haloperidol for rapid tranquilization in 11. Salzaman C. Use of benzodiazepines to control disruptive a psychiatric emergency room setting. Int Clin behavior in inpatients. J Clin Psychiatry. 1988; 49:13–5.

Source: http://sphemerg.ca/files/MidazolamvsHaloperidolvsLorazepamintheChemicalRestraintofViolentandSeverelyAgitatedPatients.pdf