Olanzapine versus Droperidol for the Treatment of Primary Headache in the Emergency Department
Presented at the SAEM Annual Meeting, May 2006, San Francisco, CA.
Address for correspondence and reprints: James R. Miner, MD; e-mail: Miner015@umn.edu.
Objectives: The objective was to determine if there is a difference in pain relief or frequency and severity of side effects in emergency department (ED) patients with primary headache treated with either intramuscular (IM) olanzapine or IM droperidol.
Methods: This was a prospective, randomized nonblinded clinical trial of adult ED patients undergoing treatment for suspected primary headache. Consenting patients were randomized to receive either droperidol 5 mg IM or olanzapine 10 mg IM. Prior to receiving treatment, patients were asked to complete a 100-mm visual analog scale (VAS) describing their pain and a 4-point verbal rating scale (VRS) describing their pain as none, mild, moderate, or severe. Patients also completed a 100-mm VAS describing their level of nausea. Pain and nausea measurements were repeated 30 and 60 minutes after medication administration. Patients also completed the Barnes Akathisia Scale (BAS) 30 and 60 minutes after medication administration. Descriptive statistics were used as appropriate. Pain relief was compared both in terms of the decrease in VAS scores and in the proportion of patients who reported moderate or severe pain whose report later changed to mild or no pain.
Results: One-hundred patients were enrolled; 13 were withdrawn before administration of the study medication, 8 in the droperidol group and 5 in the olanzapine group, leaving 87 patients for analysis. Forty-two patients received droperidol and 45 received olanzapine. In the droperidol group, 35/40 (87.5%) patients who had reported moderate or severe pain at baseline reported mild or no pain at 60 minutes. In the olanzapine group, 38/44 (86.4%) reported this change (p = 0.89). The mean percent change from baseline VAS pain score at 60 minutes was −37% (95% CI = −84% to 11%) for droperidol and −37% (95% CI = −64% to 10%) for olanzapine (p = 0.30). The mean percent change from baseline for the VAS nausea score was −59% (95% CI = −70% to −47%) for droperidol and −64% (95% CI = −77% to −51%) for olanzapine (p = 0.83). There was no difference in any report of akathisia by the BAS between the groups (p = 0.63).
Conclusions: Both olanzapine and droperidol are effective treatments for primary headaches in the ED. No significant differences were found between the medications in terms of pain relief, antiemetic effect, or akathisia. Olanzapine may be used to treat primary headache and it is an effective alternative to droperidol.
Headaches present a frequent diagnostic and therapeutic challenge in the emergency department (ED), accounting for more than 3 million ED visits a year. This makes headaches, representing 2.7% of all ED visits, a more common complaint than back pain.1 While a headache can be a sign of serious and potentially life-threatening pathology, 90% of headaches encountered in the ED are benign and are not secondary to occult pathology.2 These headaches are known as primary headaches and include the diagnoses of migraine, cluster, and tension-type headaches.3–7 While the subdivisions of primary headache are important for long-term therapeutic interventions, recent studies suggest that in the acute setting, it is reasonable to treat primary headache as a single entity. In this study, we evaluated all patients with primary headache as a single diagnosis.4,6,7
Several classes of medications have been used in the ED management of acute primary headaches, including antiemetics, opiates, and triptans. The use of antidopaminergic antiemetics has been well described for the treatment of undifferentiated benign headaches in the ED. These medications include droperidol, metoclopramide, and prochlorperazine.5,8–15 In one study of 168 patients, droperidol was shown to be more effective than prochlorperazine for the treatment of primary headache.5 Prochlorperazine, metoclopramide, and sumatripan have been shown in similarly sized studies to have equivalent effectiveness.9,10 Because of its efficacy, droperidol is the typical treatment for primary headache in our ED. Unfortunately, many of these medications, including droperidol, have the undesirable side effects of sedation and akasthesia.16 Droperidol has also been associated with the possible risk of adverse cardiac events, resulting in a “black box” warning from the Food and Drug Administration (FDA) in 2001 for a possible association with cardiac arrhythmia. Many recent studies have refuted the clinical significance of this risk,17,18 including an article reviewing 16,791 patients who received droperidol without torsades de pointes.19 While the efficacy and safety of droperidol is well supported, the FDA warning creates the need for a medication that is similarly effective in treating headaches without the perceived risks of a black box warning.
Olanzapine, a thenobenzodiazapine, is a neuroleptic medication with dopaminergic and serotonergic antagonism. Olanzapine has been associated with less QT prolongation and fewer extra pyramidal symptoms than traditional neuroleptics.20–22 Olanzapine has not been associated with adverse cardiac events. Recent research suggests that olanzapine may be effective at treating primary headache, likely through antidopinergic effects similar to droperidol, prochlorperazine, and metoclopramide.23,24 Olanzapine has also been shown effective in the treatment of nausea and vomiting, a frequent secondary complaint in patients with some types of primary headache.23 The objective of this study was to determine if there is a difference in pain relief or frequency and severity of side effects in ED patients with primary headache treated with either intramuscular (IM) olanzapine or IM droperidol.
This was a prospective randomized, single-blinded clinical trial of adult ED patients undergoing treatment for suspected primary headache. The Institutional Review Board of the Hennepin County Medical Center approved the study. Patients provided informed consent prior to enrollment.
Study Setting and Population
The study was performed at Hennepin County Medical Center, an urban Level 1 trauma center with approximately 102,000 visits per year. A convenience sample of adult (age >17 years) ED patients who were to receive droperidol for treatment of a suspected primary headache were eligible. Patients in whom head trauma, infection, vascular disorders, and disorders of facial or cranial structures were suspected were excluded. Patients who were pregnant, intoxicated, prisoners, over the age of 65, or allergic to either study medication were also excluded. The study took place between September 2006 and September 2007.
Eligible patients were approached to undergo informed consent. Consenting patients who were premenopausal females without a history of surgical sterilization underwent a urine pregnancy test prior to randomization. Consenting patients were then randomized to receive either our standard treatment of 5 mg IM droperidol or the study treatment of 10 mg of IM olanzapine. Randomization was done using a random-number generator to assign the order of the study medications. Sequentially number envelopes with the randomized study protocol were placed in the physician area of the ED and opened after patient selection for study enrollment. Prior to receiving treatment, patients were asked to complete a 100-mm visual analog scale (VAS) describing their pain and a 4-point verbal rating scale (VRS) describing their pain as none, mild, moderate, or severe. Patients also completed a 100-mm VAS describing their level of nausea. Patients were blinded to the medication they received. Pain and nausea measurements were repeated 30 and 60 minutes after medication administration. Patients also completed the Barnes Akathisia Scale (BAS) 30 and 60 minutes after medication administration. The BAS is a validated scale that scores patients based on observable akasthetic movements, subjective awareness of restlessness, and global assessment of akathisia.25 Patients then completed a headache diagnosis worksheet that differentiated headache subtype based on International Headache Society (IHS) criteria as migraine, tension type, or migrainous.5,7,26 The patient’s level of consciousness was determined by the treating physician using the 9-point altered mental status (AMS) scale27 at 0, 30, and 60 minutes. Patients who had inadequate pain control after 1 hour but before discharge were given rescue medications at the discretion of the attending physician and this rescue medication was recorded. Adverse events, including (but not limited to) the need for further therapeutic intervention, were recorded. As recommended by the FDA, it was our department policy that patients who received droperidol had a QT segment calculated before enrollment and had cardiac monitoring for 2 hours or the length of their ED visit.
Data were collected using trained research associates who were available from 7 am to 1 am daily and periodically between 1 am to 7 am as well. Data were entered into a database using Excel (Microsoft Corp., Redmond, WA) and analyzed using Stata 10.0 (StataCorp, College Station, TX). Descriptive statistics were used as appropriate. Proportions were compared with chi-square tests. Distributions were compared with Wilcoxon rank sum tests. Pain relief was compared both in terms of the decrease in VAS scores and in the proportion of patients who reported moderate or severe pain whose report later changed to mild or no pain. Power analysis showed that to detect a 20% difference between the changes in VAS scores from baseline at the subsequent time points between groups, assuming a 10% standard deviation, with a significance level of 0.05 and a power of 80%, 45 subjects were required in each group.
One-hundred patients were enrolled; 13 did not receive the study medication before discharge and were withdrawn, 8 in the droperidol group and 5 in the olanzapine group, leaving 87 patients for analysis. Forty-two patients received droperidol and 45 received olanzapine (Figure 1). Baseline data are described in Table 1. There were no differences in baseline data, including the headache subtypes differentiated using IHS criteria.3,4 In the droperidol group, 2 patients described their pain as mild, 13 moderate, and 27 severe. For the olanzapine group, 1 patient described their pain as mild, 11 moderate, and 33 severe (p = 0.61).
Table 1. Baseline Data
|Female (%)||31 (73.8%)||35 (77.8%)|
|Age, yr (±SD; range)||34.6 (9.3; 18 to 56)||32.5 (10.8; 18 to 54)|
|Median headache duration (IQR; range)|| 3 days (1 to 4; 0.1 to 10)|| 3 days (1 to 5; 0.1 to 7)|
|QTC 30 min postadministration (±SD; range)|| 0.405 (0.052; 0.186 to 0.470)|| 0.377 (0.029; 0.310 to 0.447)|
|History of migraines (%)||23 (54.8%)||33 (73.3%)|
| Migraine||24 (57.1%)||30 (66.7%)|
| Tension||10 (21.8%)|| 7 (15.6%)|
| Migrainous|| 8 (19.0%)|| 8 (17.8%)|
|Pain VAS (95% CI)||83.9 mm (78.2, 89.6)||84.2 mm (78.7, 89.6)|
|Nausea VAS (95% CI)||37.4 mm (27.2, 47.6)||36.6 mm (26.5, 46.7)|
|Median AMS score (IQR; range)|| 0 (0 to 0; 0 to 1)|| 0 (0 to 0; 0 to 1)|
All patients remained in the study 30 minutes after administration of the study drug; no patients received a rescue medicine in the first 30 minutes. Data for the 30-minute time point are summarized in Table 2. In the droperidol group, 27/40 (67.5%) patients who had reported moderate or severe pain at baseline reported mild or no pain at 30 minutes. In the olanzapine group, 27/44 (61.4%) reported this change (p = 0.68). There was no difference in any report of BAS scores between any of the groups (p = 0.54) or in the magnitude of the reported symptoms (see Table 2).
Table 2. Thirty-minute Summary Data
|Pain VAS (95% CI)|| 42.7 mm (31.6, 53.9)|| 44.0 mm (34.1, 54.0)||0.80|
|Mean % change in pain VAS from baseline (95% CI)||−52.2% (−63.5, −40.8)||−46.8% (−58.7, −35.0)||0.53|
|VAS nausea (95% CI)|| 16.0 (9.2, 22.9)|| 20.4 (12.5, 28.2)||0.47|
|Mean percent change in nausea VAS from baseline (95% CI)||−38.4% (−59.2, −17.6) ||−34.5% (−52.6, −16.4)||0.72|
|Median AMS score (IQR; range)|| 0 (−1 to 0; −2 to 2)|| −1 (−1 to 0; −3 to 2)||0.21|
|Median BAS awareness (IQR; range)|| 0 (0 to 1; 0 to 3)|| 0 (0 to 0; 0 to 3)||0.51|
|Median BAS distress (IQR; range)|| 0 (0 to 0; 0 to 3)|| 0 (0 to 0; 0 to 3)||0.63|
|Median BAS global (IQR; range)|| 0 (0 to 1; 0 to 4)|| 0 (0 to 0; 0 to 3)||0.28|
All patients remained in the study until 60 minutes after administration of the study medication. No patients received a rescue medication during the 60 minutes after study drug administration. Data for the 60-minute time point are reported in Table 3. Sixty-minute patient pain ratings for the droperidol group included 11 with no pain, 21 mild, 3 moderate, and 7 severe. For the olanzapine group, 17 reported no pain, 18 mild, 4 moderate, and 6 severe (p = 0.65). In the droperidol group, 35/40 (87.5%) patients who had reported moderate or severe pain at baseline reported mild or no pain at 60 minutes. In the olanzapine group, 38/44 (86.4%) reported this change (p = 0.89). For the patient-reported BAS scores in the droperidol group, 10/42 reported scores > 0 for the awareness portion, 5/42 for the distress portion, and 7/42 for the global portion. For olanzapine, 13/45 reported scores > 0 for the awareness portion, 9/45 for the distress portion, and 10/45 for the global portion. There was no difference in any report of akathisia between the groups (p = 0.63) or in the magnitude of the reported symptoms (see Table 3).
Table 3. Sixty-minute Summary Data
|Pain VAS (95% CI)|| 35.9 mm (26.1, 45.7)|| 29.7 mm (19.8, 39.5)||0.37|
|Mean percent change in pain VAS from baseline (95% CI)||−58.7% (−70.0, −47.4)||−63.9% (−76.6, −51.2)||0.30|
|VAS nausea (95% CI)|| 10.8 mm (5.7, 16.0)|| 13.6 mm (6.1, 21.1)||0.77|
|Mean percent change in nausea VAS from baseline (95% CI)||−37.0% (−84.9, 11)||−37.3% (−64.4, −10.2)||0.83|
|Median AMS score (IQR; range)|| 0 (−1 to 0; −3 to 1)|| 0 (−1 to 0; −3 to 1)||0.86|
|Median BAS awareness (IQR; range)|| 0 (0 to 0; 0 to 2)|| 0 (0 to 0; 0 to 3)||0.82|
|Median BAS distress (IQR; range)|| 0 (0 to 0; 0 to 2)|| 0 (0 to 0; 0 to 3||0.44|
|Median BAS global (IQR; range)|| 0 (0 to 0; 0 to 4)|| 0 (0 to 0; 0 to 2)||0.43|
Six of 42 patients in the droperidol group and 4/45 patients in the olanzapine group received rescue medications after 60 minutes (p = 0.43). For the droperidol group, 1 patient received more droperidol, 2 received sumatriptan, 2 received intravenous morphine, and 1 received olanzapine. For the olanzapine group, 3 received droperidol and 1 received a narcotic.
The treatment of primary headache is a frequent and challenging situation for the emergency physician. While many medications are available to treat headache, each has significant clinical side effects. Ergot alkaloids and triptan medications are contraindicated in hypertension, coronary artery disease, and peripheral vascular disease. Triptans can also cause the significant clinical side effect of chest pain and flushing.6 Prochlorperazine does not relieve pain as effectively as droperidol.5 Droperidol, prochlorperazine, and metoclopramide cause akathisia and sedation. Finally, while evidence supports the safety of droperidol, the FDA “black box” warning regarding adverse cardiac events limits its availability and acceptance in many practice environments. It was our intent to show that olanzapine, a drug with similar pharmacologic properties to droperidol, was similarly effective at treating primary headache.
We found that both IM olanzapine and IM droperidol can be effective for the treatment of pain and nausea from primary headache. Seventy-nine percent of patients received pain relief with olanzapine and 81.6% with droperidol. Sedation appeared to be similar between the two groups. No serious adverse events were noted. We also found no difference in the effectiveness of olanzapine and droperidol at controlling the nausea associated with headache. At 60 minutes, the mean changes in VAS scores for nausea were −37.0% for droperidol and −37.3% for olanzapine. This suggests that the medications are similar at controlling this frequently associated symptom of benign headache.
It should be emphasized that our study was conducted on patients thought to have a primary headache based on history, exam, and physician clinical judgment. The treatment of secondary headache will always be specific to the pathophysiology of the causative disease process, and we are not advocating the use of olanzapine or droperidol in the treatment or diagnosis of secondary headache.
Our results did not show a difference in the level of akathisia between droperidol and olanzapine. We did not find a difference in the number of patients with akathisia or in the severity of their symptoms either by objective observation, subjective reporting, or global assessment. Olanzapine has been found to have fewer extrapyramidal side effects compared to haloperidol,28 while droperidol and haloperidol have been shown to have equivalent akathisia and extrapyramidal side effects.16 Given this, it is reasonable to assume that olanzapine may have some advantage over droperidol with regard to akathisia. If this advantage exists, it was too small for our study to detect. Due to the safety of both medications, our study was also too small to detect the occurrence of any possible rare adverse phenomena.
It was our objective to determine if there was a difference between IM olanzapine and IM droperidol for the treatment of pain from primary headaches, and we were unable to find one. Droperidol is currently less expensive than olanzapine and has been used successfully for many years. While we did not find a reason to advocate for the use of olanzapine over the use of droperidol, it does appear that olanzapine may be a useful treatment option for primary headache when droperidol is not available or is considered contraindicated.
The primary limitation of the study is that the physicians were not blinded to the administered study medications due to the significant cost associated with blinding. To avoid bias associated with this, all data were obtained by research associates rather than the treating physician. Research associates were not aware of the study hypothesis and did not have familiarity or preference for either drug, making bias in the data recording less likely. However, lack of blinding may have affected the choice of rescue medications by the treating physician. As a result, we chose to limit the conclusions drawn from these data.
A further limitation is the lack of 24-hour follow-up. While we initially intended to follow-up with patients at 24 hours to evaluate for recurrence of symptoms or delayed side effects, this proved to be difficult, because we were unable to contact a large number of patients. With our poor rate of contacting patients with the phone numbers given, and the lack of staffing to ensure 24-hour contact, our attempt at 24-hour follow-up was abandoned.
Finally, the patients in this study represent a convenience sample, because patients were enrolled depending on availability of research associates to collect data and at the discretion of the treating physician. One of the enrollment criteria was the presumed diagnosis of primary headache. This diagnosis is a complex clinical decision process and is not easily standardized. It is likely that there is practice variability in the decision to classify a patient as having a primary headache or to perform further evaluation for causes of secondary headache.
Both olanzapine and droperidol are effective treatments for primary headaches in the ED. We did not find any significant difference between the medications in terms of pain relief, antiemetic effect, or akathisia. Olanzapine may be used to treat primary headache, and it is an effective alternative to droperidol.