Patients were eligible for study if they exhibited psychosis or behavioral abnormalities making them out of control. Most were agitated, aggressive, destructive, or restless. Patients were enrolled if they were capable of harming themselves or others, and individuals with either drug-induced psychosis or psychiatric diagnoses were eligible. Patients with pure alcohol intoxication were excluded. Those with head trauma or CNS infections were included unless they exhibited instability. Ninety-eight patients prospectively received either lorazepam 2 mg IM or haloperidol 5 mg IM, or a combination of both. Additional doses were allowed one hour apart for a total of six doses. Following attempted tranquilization, a variety of parameters were measured.
More than half of the patients had a final diagnosis of schizophrenia, nondescript psychosis, or substance abuse. Although all patients exhibited a decrease in the rating scale used to measure agitation, patients receiving the combination treatment demonstrated significantly greater reductions in agitation than the patients who received either single drug. Most patients were adequately sedated with a maximum of three doses. Approximately 20 percent of the patients developed extrapyramidal symptoms secondary to haloperidol, including dystonia, hypertonia, rigidity, and tremor.
The authors concluded that the combination of lorazepam and haloperidol was more effective in establishing rapid tranquilization than the use of either drug alone. The combination therapy was more effective for all measured parameters, and it achieved tranquilization statistically sooner than either single drug treatment. When given alone, lorazepam and haloperidol were generally equivalent. There were no serious side effects from any regimen, although the potential for respiratory depression, hypotension, and stupor were noted. The authors conclude that rapid tranquilization is best achieved by a combination of lorazepam plus haloperidol.
Despite the almost daily use of rapid tranquilization, also known as chemical restraint, there is surprisingly little emergency medicine literature addressing a specific protocol or advantages or disadvantages of particular drugs. This is one of the few articles to evaluate different modes of tranquilization scientifically, but the study had some significant limitations, most notably that most patients were psychotic and not experiencing potentially fatal delirium secondary to cocaine, PCP, or amphetamines. Most psychotic patients are not in imminent danger of cardiac arrest because of hyperthermia, rhabdomyolysis, or metabolic acidosis, although they can be quite problematic and difficult to control.
For some reason, psychotic patients are almost always bigger and stronger than the staff, and they rarely understand reason. In my experience, however, they respond better to tranquilization than cocaine addicts do. Although the combination of lorazepam and haloperidol was more effective than either drug given alone, the results were not particularly impressive. In addition, the doses were miniscule from a modern day emergency medicine standpoint. Because these patients were controlled with relatively small doses of the study medications, I would intuit that none were at great risk for cardiac arrest or severe metabolic derangements. In my experience, 2 mg of lorazepam or 5 mg of haloperidol doesn't even begin to touch the acutely delirious cocaine or PCP user.
My problem with this study is that the drugs were given IM. I strongly believe that any acutely agitated or delirious patient requires IV tranquilization, and the time required for an IM drug to be absorbed may mean the difference between rhabdomyolysis and renal failure and a benign mild elevation in the CPK. Although vascular access is often problematic in the delirious patient, all attempts should be made to deliver sedatives by IV. The main goal is to achieve control over delirious patients in the most rapid manner possible because even a few minutes of struggling against restraints can significantly increase morbidity. The IV route is the most logical, reasonable, and rational choice.
Regardless of one's choice of medication, the key to success is to use enough drug. Internist- and pediatrician-type doses usually don't cut it. The PDR is clueless about effective doses for our purposes, and safety caveats in that publication are merely fodder for lawyers, not information for clinicians. The best guideline is to give enough to achieve the intended result.
Our major clinical conundrum is that we often don't know what's wrong, and we have to act and gain control with only a modicum of data, history, and time, so safe and familiar drugs are the best first choice. Of course, always, always, check the glucose, even if EMS tells you it is OK. Because benzodiazepines have such a wide margin of safety, their use is familiar to all emergency physicians, and because the potential side effects of major tranquilizers are not present, benzodiazepines have remained the drugs of choice under most circumstances. (See a good discussion by Bodkin, J Clin Psychiatry 1990;51:5 [Suppl].) They are generally very safe and effective in the undifferentiated patient, even when you don't have a clue about the correct diagnosis. There seems to be no great difference between diazepam, lorazepam, or midazolam, and all three are available for IV use.
Midazolam's advantage is that it is most rapidly absorbed by an IM injection, but it has a short half-life, maybe too short for some. All benzodiazepines can be reversed with flumazenil, although giving this benzodiazepine antagonist is tricky in the undifferentiated, recently sedated patient. It is difficult to overdose a patient with the judicious use of IV benzodiazepines, and I personally would not attempt to reverse inadvertent overly aggressive use of benzodiazepines in any patient who was delirious or widely agitated. It's probably better to institute mechanical ventilation for that unusual circumstance where respiratory depression occurs. Fortunately, that's the rare case, and concerns about excessive respiratory depression, even from aggressive benzodiazepine use, are overstated.
There is no maximum dose that has been established for intravenous benzodiazepines or for haloperidol, for that matter. They can be mixed in the same syringe. I am always amazed by the seemingly gargantuan doses of benzodiazepines that can be tolerated in alcohol withdrawal, cocaine, amphetamine, or PCP delirium, and there are some patients that don't seem to ever respond. Benzodiazepine resistance is well documented, and no amount may be enough for some, especially in alcohol withdrawal. Adding haloperidol probably decreases the benzodiazepine requirement and vice versa, but I would personally push the benzodiazepine dose first. If haloperidol is used, it's best given after a hefty dose of a benzodiazepine has been tried. Most of the downsides of haloperidol are merely theoretical, and high doses of the drug given as a single agent have been used successfully for years. Most clinical experience with haloperidol, however, has been with psychiatric patients.
As a general benzodiazepine guideline, I use the 10/100/10 rule: If the patient is still agitated after 10 mg of lorazepam or 100 mg of diazepam and 10 mg haloperidol, I will consider another regimen. It has been my clinical experience that once this upper level of benzodiazepine dose is ineffective, additional benzodiazepines are not all that helpful. The literature is replete with references for safely using more than 2000 mg diazepam (200 times the “safe” PDR dose) for DTs over only 24 hours, and the patient is still thrashing about (Acad Emerg Med 2000;7:308). Some individuals are benzodiazepine-resistant, and some become even more agitated after a few doses. In adults, I would start with 4 mg lorazepam and 5 mg haloperidol unless there are some reasons to choose the very modest 2 mg initial dose used in the previous study. This can be repeated two or three times about five to 10 minutes apart.
The major advantage of using the IV route is that you know within a few minutes whether your first dose was high enough. One simply cannot wait three to four hours for these drugs to kick in. Some authors believe that diazepam reaches its maximum effect (and maximum respirator depression) within five minutes, so this drug can be used every five minutes. It may take longer for lorazepam to reach its peak effect, but this is not certain. I actually have come to like diazepam better than lorazepam; it just seems to work faster and better, and it's so safe.
Likewise, there is no maximum dose of haloperidol. In the ICU, even patients with serious medical problems have been maintained on daily haloperidol doses in the range of 600 mg without serious side effects. (J Clin Psychiatry 1988;49:239; S Carolina Med Assoc 1992;88:386.) In my opinion, using less than 5 mg IV for the first haloperidol dose is counterproductive, but high doses generally keep the patient out for 12 to 24 hours. Haloperidol has minimal anticholinergic effects, and it does not precipitate hypotension or respiratory depression. Haloperidol is a good choice to put to sleep the obnoxious drunk who insists on leaving the hospital in an intoxicated state.
The combination of benzodiazepines and haloperidol has been studied in the past, and the combination has been generally hailed as safe and effective. Manza et al (Heart Lung 1988;17:238) administered haloperidol/benzodiazepine to medically ill patients. Side effects were negligible, although some patients did develop extrapyramidal effects. Some authors prefer to use more haloperidol because this drug has minimal respiratory and cardiovascular effects. There seems to be no overall support for the use of the older potent neuroleptics, such as thorazine, stelazine, prolixin, or mellaril. These drugs predictably produce hypotension in higher doses, and they can theoretically lower the seizure threshold.
Salzman et al (J Clin Psych 1991;52:177) also compared IV lorazepam with IV haloperidol for controlling acutely psychotic behavior. They concluded that parenteral lorazepam is an effective agent for controlling acute aggressive and assaultive behavior, and overall it is quite safe. It was preferred over haloperidol, but the arguments were not impressive.
The EP should know that high doses of many benzodiazepines (especially lorazepam used in continuous infusion in the ICU) can cause a hyperosmolar anion gap metabolic acidosis because of its preservative, propylene glycol (PG). (Crit Care Med 2004;32(8):1709.) As a toxicologist, I have tried to track down many an osmol gap in sedated patients only to find that it was due not to a toxic alcohol commonly considered but to the PG in the sedative. High doses of propylene glycol also can cause CNS depression, renal failure, hemolysis, seizures, and arrythmias. PG also is used with diazepam, pentobarbital, nitroglycerine, and etomidate.
Finally, I have found but have not seen studied whether patients will respond better to a different benzodiazepine if the first one is ineffective. I have been impressed that diazepam or midazolam just might work if lorazepam is not stopping a seizure or agitation. Any comments from readers?
Droperidol v. Haloperidol for Chemical Restraint of Agitated and Combative Patients
Thomas H, et al Ann Emerg Med 1992;21:407
The authors of this study compared the effects of droperidol with haloperidol for chemical restraint of agitated ED patients. The proposed benefit of droperidol is that it has a shorter half-life than haloperidol, a more rapid onset of action, and is generally effective for controlling aggressive and disruptive behavior in psychiatric patients. Both drugs were given either IM or IV in a 5 mg dose. Effectiveness was assessed for up to 60 minutes, and droperidol was found to reduce agitation better than haloperidol, but there was no great difference between the two. It was concluded that droperidol in a dose of 5 mg IM was preferred over haloperidol. With the current bizarrely conceived warnings about QT prolongation with droperidol, most would not use this drug anymore, although it has an excellent clinical track record. Curiously, haloperidol also prolongs the QT interval, maybe as much as droperidol.
Many have championed the use of droperidol. (J Clin Psych 1984;45:298.) I used it for years with great success (also for migraines and vomiting), but now it is eschewed by most EPs for medico-legal reasons, perhaps bogus at best. Why take the chance of a sudden death due to acidosis that is blamed on droperidol? Recently, Martel et al (Acad Emerg Med 2005, 12:1167) used droperidol safely to treat acute undifferentiated agitation in the ED.
There are some potential downsides to any neuroleptic use that have been highlighted in the literature. Most are theoretical and of no great medical consequence. While lack of hypotension or respiratory depression are strong points for haloperidol, extrapyramidal symptoms appear to be present in up to about half of the patients administered haloperidol in large doses. Some authors have advocated the routine and empiric use of Cogentin or Benadryl when large doses of haloperidol have been used, although I think that's a bad idea. It's theoretically detrimental (not proven) in patients who have the potential for hyperthermia because Cogentin impairs sweating. Occasionally it's difficult to differentiate psychosis from anticholinergic delirium.
Wilt (Ann Intern Med 1993;119:391) reported four cases of Torsades de pointes in patients whose delirium was treated with hefty doses of haloperidol. In three of the four patients, antiarrhythmic therapy was required. Haloperidol has been known to prolong the QT interval, and this is the proposed arrhythmogenic mechanism. Haloperidol use has been linked to the neuroleptic malignant syndrome, a potentially fatal condition. Although you will have to give haloperidol to a lot of patients and in impressively large doses before you see a case of the neuroleptic malignant syndrome or Torsades, these are theoretical reasons to eschew haloperidol as your first and only drug for chemical restraint.
I personally have seen neither complication in patients given high doses for short periods of time. Other reasons to avoid high doses of haloperidol alone without a benzodiazepine and other neuroleptic agents include impaired heat dissipation, anticholinergic effects, and a prolonged duration of action. Theoretically, neuroleptics can lower the seizure threshold, a potentially bad side effect in cocaine intoxication, although I do not believe that this has ever been proven with certainty, and it remains controversial. This specific concern should not limit haloperidol's use on a regular basis.
Although there is consensus that acutely agitated and violently delirious patients should be rapidly tranquilized; unfortunately, there is a paucity of data on the most appropriate or effective drug regimen. Likely the biggest mistake is to use too little of any drug and to give it too late. A number of safe and effective options are available to the practicing clinician, and as is the case with most situations where a variety of choices are available, many options probably work equally well. Obviously it's best to choose drugs with which one has most experience. For most emergency physicians, this would be the benzodiazepines. Whether you choose to add haloperidol is a matter of personal choice, but it seems reasonable to consider combination therapy if benzodiazepines alone are not effective.
Administer all drugs intravenously, and titrate to effect. With intravenous diazepam, lorazepam, or midazolam, one knows in a few minutes whether the dose was the right one and whether an additional drug is required to get the job done correctly. It's a common scenario that seemingly huge doses may be necessary, and one should not be concerned with doses that would appear gargantuan to the psychiatry department or the package insert. In reality, there is probably no maximum dose limit for any benzodiazepines or haloperidol, but if one is draining the pharmacy's supply of the benzodiazepines, it's time to add something else.
Literature would support haloperidol as the next logical addition, and while extremely safe under most scenarios, the physician must be cognizant of the potential downsides of these drugs. Some patients just can't be restrained with these chemicals and will require other interventions, including muscular paralysis and mechanical ventilation. Often this coincides with a continuous infusion of something. Continuous infusions of benzodiazepines or propofol are all right for the intensive care unit, but for the patient who just can't be controlled in the ED, the threshold for succinylcholine/Pavulon or one's favorite intubating/paralyzing combination should be quite low. Some authors raise the possibility of hyperkalemia, rhabdomyolysis, or prolonged duration of action in the presence of cocaine-induced delirium as reasons to avoid succinylcholine. To my reading, these are purely theoretical arguments that do not overshadow the other benefits of this superb paralyzing agent.
It is important to point out that narcotics or barbiturates are not particularly good choices for acute sedation, although various barbiturates have been used safely for decades. Fentanyl is great for conscious sedation, but it's not a major tranquilizer. Etomidate is super for intubation, but it does not last long enough. Ketamine is interesting because it can be given IM and will produce a profound effect in only a few minutes. Next month, I will focus on some newer agents for rapid tranquilization, such as ketamine, ziprasidone, and olanzapine.
Readers are invited to ask specific questions and offer personal experiences, comments, or observations on InFocus topics. Literature references are appreciated. Pertinent responses will be published in a future issue. Please send comments to firstname.lastname@example.org. Dr. Roberts requests feedback on this month's column, especially personal experiences with successes, failures, and technique.© 2007 Lippincott Williams & Wilkins, Inc.