Introduction with a case

An elderly woman was admitted to the ICU for COPD exacerbation requiring BiPAP.  Her COPD improved but she developed delirium.  She was increasingly agitated, attempting to remove her oxygen and intravenous lines.  She received 4 mg IV haloperidol, but continued to be agitated.  Eventually after titration to a total of 12 mg IV haloperidol she became somnolent.  At that point it was noted that despite her sleeping, she manifested myoclonic jerking of her extremities (a new finding).

What happened?  With hindsight, it is likely that she began experiencing akathisia as a side effect of the IV haloperidol used to sedate her (which caused worsening restlessness, agitation, and myoclonic jerking).  This is a potential pitfall of using haloperidol for treatment of agitation, because akathisia may actually exacerbate the patient’s agitation, leading to a vicious cycle involving further haloperidol administration.  Diagnosing akathisia in an agitated patient can be difficult.  How could this complication be avoided in the future? (1)

The evolution of intravenous olanzapine

Technically, haloperidol is not FDA-approved for intravenous use.  However, shortly after intramuscular haloperidol was released, physicians realized that it could be given intravenously.  Over time, IV haloperidol became widespread and accepted.

Olanzapine (ZYPREXA) seems to be undergoing a similar process.  Although olanzapine isn’t FDA-approved for intravenous administration, it is increasingly recognized that this is effective and safe.  For example, the “intramuscular” formulation of olanzapine has a neutral pH and contains substances which can be given intravenously (olanzapine, lactose, and tartrate – as in metoprolol tartrate).  IV olanzapine was initially utilized in Australia (Chan 2009, 2011).

The rationale for using IV olanzapine isn’t superior efficacy compared to haloperidol, but rather equivalent efficacy with fewer side effects (2).  What does the evidence show?

Efficacy of olanzapine for acute agitation

Efficacy of IM olanzapine

The efficacy of IM olanzapine has been investigated more thoroughly than IV olanzapine.  Five RCTs involving 613 patients have compared IM olanzapine versus IM haloperidol (Kishi 2015).  These studies either found that the two drugs were equally effective or that olanzapine was slightly more effective (depending on the relative doses utilized)(3).

Efficacy of IV olanzapine

Chan 2013 performed a prospective RCT of agitated emergency department patients who received titrated open-label midazolam in addition to being randomized to receive one of three treatments: 5 mg IV olanzapine, 5 mg IV droperidol, or placebo.  Droperidol and olanzapine were equally effective.

Safety of olanzapine: Torsade de pointes?

It can be hard to determine whether a drug causes torsade de pointes (TdP), because this event is rare.  Several sources of evidence bear consideration:

HERG channel affinity

The primary driver of drug-induced TdP is blockade of the HERG potassium channel involved in myocardial repolarization (Crumb 2006).  Drugs with a greater affinity for the HERG channel tend to cause more TdP.  As shown above, olanzapine's affinity for HERG channels is 6,000 times below haloperidol's affinity.  Additionally, a guinea pig model suggests that olanzapine tends to be less concentrated within myocardial tissue.

Typical therapeutic serum concentrations of olanzapine are 32-64 nM, which might correlate with a concentration of perhaps ~86-170 nM within the myocardium.  These concentrations are well below the half-maximal inhibitory concentration (IC-50) of 6,000 nM, predicting an absence of proarrhythmia (Titier 2004).

Animal models

Ando 2007 studied the arrhythmic potential of intravenous risperidone or olanzapine in a canine model.  Risperidone caused prolongation of ventricular repolarization only at supra-therapeutic doses.  In contrast, intravenous olanzapine had minimal effects on repolarization even at very high doses (3 mg/kg; figure below).

Effect on QT interval in humans

Although QT interval isn't perfect, it remains a clinically relevant measurement of how drugs affect cardiac repolarization.  Harrigan 2004 performed a prospective RCT comparing the effect of several antipsychotic agents on the QTc interval.  As shown, olanzapine had no effect on QTc interval (even at a dose of 20 mg/day).  Lack of effect on QTc has been validated by other studies (Lindborg 2003, Shafti 2014, Takeuchi 2014).

Case reports of TdP

Currently there is only one case report in the literature of a patient with TdP related to olanzapine (Huang 2014).  However, this woman also had cardiomyopathy status post placement of an implantable defibrillator, hypomagnesemia, and hypokalemia, making it impossible to establish causality between olanzapine and TdP.  Overall there is a notable absence of case reports of TdP.

Epidemiology

Population-wide correlations between antipsychotic use and sudden death yield conflicting results, without any consistent hierarchy of risk among different antipsychotics (Ray 2009, Danielsson 2016, Leonard 2013).  Due to confounding variables and contradictory results, no conclusions can be drawn from this data.

Bottom line on TdP

The best available evidence doesn’t reveal a risk of TdP due to olanzapine.  There is no indication for monitoring the QT interval in patients receiving olanzapine.

Safety of olanzapine: Extrapyramidal symptoms

Olanzapine causes fewer extrapyramidal symptoms than haloperidol.  This has been shown in RCTs comparing haloperidol vs. olanzapine administered via either oral (Tollefson 1997) or IM routes (Wright 2001, Breier 2002).

With haloperidol, intravenous use is associated with fewer extrapyramidal side effects compared to oral administration.  If the same relative reduction in extrapyramidal symptoms were true of IV olanzapine compared to oral olanzapine, it would suggest that IV olanzapine should have a very low rate of extrapyramidal symptoms.  One large study of IV olanzapine reported a 0.6% rate of akathisia and no dystonic reactions, consistent with this possibility (more on this study below; Martel 2015).

Other safety concerns?

Some concerns have been raised about over-sedation due from the combination of IM olanzapine with benzodiazepines.  However, the RCT by Chan 2013 discussed above reported no increase in adverse effects among 109 patients receiving IV olanzapine plus midazolam, compared to patients receiving placebo plus midazolam.  There was actually a trend towards increased desaturation in the placebo group, likely due to higher doses of midazolam required in the placebo group.

IV olanzapine may be safer than IM olanzapine, because IV olanzapine works faster, allowing it to be titrated accurately.  Alternatively, agitated patients who receive IM olanzapine may continue to receive several doses of benzodiazepine, before the olanzapine has taken effect.  When the IM olanzapine finally reaches peak effect, the combination of olanzapine plus several doses of benzodiazepine could cause over-sedation (“dose-stacking”).  Notably, the recommended starting dose of IM olanzapine is 10 mg, whereas most patients receiving IV olanzapine require only ~5 mg.  Thus, titrated doses of IV olanzapine may reduce the total dose administered.

Martel 2015 reported a retrospective cohort study examining side effects experienced by 713 patients who received IV olanzapine in the emergency department.  The only safety concern raised in this article were three intubations potentially related to olanzapine (adjacent box).  These intubations might have been required regardless of the choice of antipsychotic.  Overall this study is reassuring, since it represents a large pragmatic description of olanzapine use in acutely ill patients that failed to reveal any unexpected problems.

Pharmacoeconomics

Olanzapine is currently off-patent, but it may remain somewhat more expensive than haloperidol (4).  However, haloperidol is associated with costs of obtaining EKGs and more frequent treatment of extrapyramidal reactions (e.g. intravenous diphenhydramine).  Additionally, although TdP is rare, the cost of treating a single episode can be enormous.

Formal pharmacoeconomic analyses show that the primary determinant of the cost of managing an agitated patient is time: anything which prolongs the ED or ICU length of stay is expensive (e.g. one day in the ICU costs ~$3,000; Chan 2012).  This makes it unwise to choose a cheaper drug that may occasionally prolong length of stay.

Choosing: IV Olanzapine vs. IV Haloperidol?

Is IV olanzapine ready for prime time?  This boils down to a choice of the devil that we know (haloperidol) versus the devil that we don’t entirely know (olanzapine).  IV olanzapine is newer, so it is possible that additional side effects may emerge over time.  However, we already know that there are significant problems with IV haloperidol.

IV haloperidol can cause TdP.  Patients receiving this should be monitored with serial EKGs to minimize their risk.  However, in practice it may be impossible to obtain an EKG in an agitated patient.  Thus, compliance with recommended QT monitoring for IV haloperidol is extremely poor (Warnier 2015).  This represents a systemic source of risk to our patients.

The switch to olanzapine might be justifiable solely on the basis that we could spend less time checking EKGs and debating the damn QTc.  Humans have a limited capacity to make decisions and pay attention to details.  As discussed by Scott Weingart in the John Hinds Memorial Lecture, making too many decisions (even minor decisions) leads to decision fatigue.  Eliminating unnecessary decisions and interruptions could free up physicians and nurses to pay attention to more important problems.

Ultimately, weighing the known versus the unknown is a matter of judgment.  Some centers (e.g. Hennepin County Hospital) have already broadly adopted IV olanzapine.  Another approach might be to selectively use IV olanzapine for patients at higher risk of harm from haloperidol (e.g. patients with prolonged QT or patients who are too agitated to obtain an EKG).

Nuts and bolts: How to use IV olanzapine

For agitation, a reasonable dose appears to be 2.5-5 mg IV, repeated every 5-10 minutes to a maximal dose of 20mg (similar to IV droperidol).  Note that IV olanzapine appears to be roughly twice as powerful as haloperidol (e.g. 2.5-5 mg olanzapine may be equivalent to ~5-10 mg haloperidol).  As with any sedative, using lower doses and increasing the time interval between doses will reduce the risk of over-sedation.

Returning to the case

Akathisia is 3-5 times less common with olanzapine compared to haloperidol (Tollefson 1997, Kishi 2015).  Therefore, using IV olanzapine in the case described above might have avoided this complication entirely.  Alternatively, it is possible that even if did occur with IV olanzapine, it might have been less severe.

• Intravenous use of haloperidol is not FDA-approved, but is widely utilized and accepted in the literature.
• Recent studies suggest that olanzapine can also be given intravenously, even though it is labeled only for intramuscular use.
• One RCT showed that IV olanzapine was as effective as IV droperidol.
• Olanzapine does not cause torsade de pointes, nor does it require monitoring with serial EKGs.
• The rate of extrapyramidal side effects with olanzapine is lower than with haloperidol.
• Intravenous olanzapine may be an attractive alternative to intravenous haloperidol, particularly among patients at risk for arrhythmia.

To get the perspective of a clinician experienced with using IV olanzapine, I reached out to Dr. Marc Martel.  He is an Associate Professor of Emergency Medicine at Hennepin County Medical Center, sedation guru, and author of the recent paper discussed above in Academic Emergency Medicine. These are his comments:

The “right” answer is never hoping or waiting for an acutely agitated patient to calm down.  Working night shifts exclusively in a busy, inner-city ED has taught me that anything that can go wrong will go wrong – injuries, assaults, self-injurious behavior, guns, knives, and even attempted escapes (into the ceiling, no kidding, I’ve seen it twice).

The agitated and violent patient clearly represents a threat to patient and provider safety.  I teach (preach) the underappreciated risk of the “squeaky wheel effect.”  Attention is focused on the behavior, rather than the patient.  What is the underlying pathology in the agitated patient?  Who else is sick in your department?  And finally, who else is coming in sick?  Acute undifferentiated agitation prevents adequate evaluation of the patient at hand, distracts from others already in your care, and inhibits the assessment of new patients.  Agitation management is a priority.

Like airway management, a backup plan is crucial when approaching agitation.  Things to consider when selecting a sedative – degree of agitation, feasible routes of administration, time to medication onset of action, risk of respiratory suppression, risk of hypotension, incidence of arrhythmias (real, and perceived risk of TdP).  EKG’s before sedation has never quite made sense to me.  I’m not sure the patient climbing into the ceiling would have held still long enough.

Olanzapine has filled the void in my practice resulting from the droperidol drug shortage.  Our data suggests that a single 5 mg IV dose of olanzapine will control agitation in approximately two-thirds of patients (where IV access was obtainable).  5 mg IV is a good place to start, but repeat dosing for agitation is common.  In our ED, we are now approved to administer up to two 5-mg doses of olanzapine to patients in a 24-hour period.  I use IV olanzapine in moderately agitated patients, re-dosing if necessary.  In severely agitated patients, I prefer dosing between 10 and 20 mg IM.  I frequently start at 20 mg, the maximum daily dose.  In patients with persistent or extreme agitation, I will follow olanzapine with either haloperidol and/or lorazepam.  Although some have reservations about the combination of benzodiazepines and olanzapine, the reality is that the only significant risk is over sedation.  The benefit of agitation management, the ability to completely assess the patient, and ensuring patient and provider safety far outweigh this risk.  Ultimately, this “complication” can be anticipated, should be monitored for (SaO2, EtCO2 and close observation) and is completely manageable by the intensivist or emergency physician who deals with acute undifferentiated agitation on a regular basis.  In our study, 47 patients received benzodiazepines within 2 hours of IV olanzapine.  Personally, the combination does not make me nervous, but I realize I do not practice in a vacuum.

Truthfully, I’m a believer.  If you think of using olanzapine, you probably should.  The biggest downside is well known, sedation.  The only “poor” candidate for IV (or IM) olanzapine is one that needs to be discharged immediately.  The benefit of using IV olanzapine in this population is that lower doses, 1.25 – 2.5 mg, are very effective with less sedation.  I find it particularly valuable for nausea, vomiting, diarrhea, vertigo and opioid withdrawal.  The onset of action is relatively short for these indications, 5 – 10 minutes.  I can’t say that I have seen significant delays in the elderly, CHF patients or other states where circulation is impaired.  Anecdotally, although very effective, I haven’t felt that 5 mg IV olanzapine is as effective as the 10mg IM dosing for headache we’ve reported previously (5).

Stay tuned for an EMCrit podcast with Scott Weingart and Marc Martel, to discuss sedation, olanzapine, and patients escaping into the ceiling.

Related posts

• Allergies to haloperidol don't exist (PulmCrit)
• On human bondage and the art of the chemical takedown (EMCrit)

Conflicts of interest: Neither myself nor Dr. Martel have any.

Notes:

1. Eventually this patient woke up, her akithisia resolved, and she did fine. However, this event may be regarded as a “near-miss” because she did end up getting slightly oversedated and could potentially have been harmed.
2. This post focuses on comparing IV haloperidol vs. IV olanzapine, because IV droperidol isn't available to me.
3. One study involving multiple different doses of olanzapine found that 7.5 mg of IM haloperidol was equivalent to between 2.5-5 mg of IM olanzapine (Breier 2002). Other studies found 7.5 mg haloperidol IM to be less effective than 10 mg olanzapine IM, which may not be very surprising.
4. At Genius General Hospital, our goal is to get patients better regardless of cost. However, this issue invariably comes up, so this is an answer to it.
5. Hill CH, Miner JR, Martel ML. Olanzapine versus droperidol for the treatment of primary headache in the emergency department. Academic Emergency Medicine. 2008 Sep 1;15(9):806-11

Picture credits: opening image is from Sharma 1998 showing a telemetry strip from a patient who received IV haloperidol and subsequently developed TdP, with additional image from here.

• About
• Latest Posts

Social Me

Josh Farkas

Josh is the creator of PulmCrit.org. He is an assistant professor of Pulmonary and Critical Care Medicine at the University of Vermont (Burlington Vermont, USA).

Social Me

Latest posts by Josh Farkas (see all)

• IBCC:Guide to supportive care in critical illness - November 19, 2018
• IBCC chapter & cast:Hyperkalemia - November 14, 2018
• IBCC chapter & cast:Delirium - November 7, 2018