|

 

A randomized controlled trial of intravenous haloperidol vs. intravenous metoclopramide for acute migraine therapy in the emergency department

The Journal of Emergency Medicine, Vol. -, No. -, pp. 1–9, 2015 Copyright Ó 2015 Elsevier Inc.
Printed in the USA. All rights reserved 0736-4679/$ - see front matter Pharmacology inEmergency Medicine A RANDOMIZED CONTROLLED TRIAL OF INTRAVENOUS HALOPERIDOL VS.
INTRAVENOUS METOCLOPRAMIDE FOR ACUTE MIGRAINE THERAPY IN THE EMERGENCY DEPARTMENT Matthew E. Gaffigan, MDDavid I. Bruner, MD,† Courtney Wason, BSAmy Pritchard, DO,‡ and Kenneth Frumkin, MD, PHD *Naval Medical Center Portsmouth, Portsmouth, Virginia, †Naval Medical Center San Diego, San Diego, California, and ‡Naval Hospital Camp Pendleton, Camp Pendleton, California Reprint Address: David I. Bruner, MD, Naval Medical Center San Diego, 8022 Paseo del Ocaso, La Jolla, CA 92037 , Abstract—Background: Emergency Department (ED) groups were similar on all VAS measurements, side effects, headache patients are commonly treated with neuroleptic an- and in their satisfaction with therapy. Pain relief averaged tiemetics like metoclopramide. Haloperidol has been shown 53 mm VAS over both groups, with equal times to maximum to be effective for migraine treatment. Study Objective: improvement. Subjects receiving haloperidol required rescue Our study compared the use of metoclopramide vs. haloper- medication significantly less often (3% vs. 24%, p < 0.02).
idol to treat ED migraine patients. Methods: A prospective, Mean QTcs were equal and normal in the two groups and double-blinded, randomized control trial of 64 adults aged did not change after treatment. In telephone follow-up, 18–50 years with migraine headache and no recognized risks 90% of subjects contacted were ‘‘happy with the medication'' for QT-prolongation. Haloperidol 5 mg or metoclopramide they had received, with haloperidol-treated subjects experi- 10 mg was given intravenously after 25 mg diphenhydramine.
encing more restlessness (43% vs. 10%). Conclusions: Intra- Pain, nausea, restlessness (akathisia), and sedation were as- venous haloperidol is as safe and effective as metoclopramide sessed with 100-mm visual analog scales (VAS) at baseline for the ED treatment of migraine headaches, with less and every 20 min, to a maximum of 80 min. The need for frequent need for rescue medications.
Ó 2015 Elsevier Inc.
rescue medications, side effects, and subject satisfactionwere recorded. QTc intervals were measured prior to and , Keywords—migraine; haloperidol; pain management after treatment. Follow-up calls after 48 h assessed satisfac-tion and recurrent or persistent symptoms. Results: Thirty-one subjects received haloperidol, 33 metoclopramide. The This study was presented at the American Academy of Emer- gency Medicine Scientific Assembly Resident Research Forum Headache accounts for 2–5% of emergency department in February 2014, and it won first place. It was also presented at (ED) visits and is the fifth most common ED chief the Naval Medical Center Portsmouth local research competi- complaint Current first-line ED therapies typically tion for 2014 and won first place. It was then presented at a include a dopamine receptor antagonist like prochlorper- Navy-wide research competition and placed third.
azine or metoclopramide, often combined with diphenhy- The views expressed in this article are those of the authors and dramine. Studies have shown these medications to be safe do not necessarily reflect the official policy or position of theDepartment of the Navy, Department of Defense, or the United and more effective than opiates, nonsteroidal anti- States Government.
inflammatory medications, and sumatriptan .
RECEIVED: 30 October 2014; FINAL SUBMISSION RECEIVED: 9 March 2015;ACCEPTED: 24 March 2015 M. E. Gaffigan et al.
Haloperidol is another venerable dopamine antago- presenting to the ED from June 2013 to February 2014 nist, of the butyrophenone class. Like other neuroleptic with a chief complaint of migraine headache. The medi- antiemetics, haloperidol has been reported to be effective cal center's Institutional Review Board approved this in the treatment of nausea and migraine headaches .
study. The design and analysis closely resembles that of When neuroleptics are used alone, side effects (most Kostic et al., which compared prochlorperazine with su- commonly akathisia) can limit their usefulness. The matriptan for migraine relief in the ED .
addition of i.v. diphenhydramine can help reduce this The setting was the ED of a 360-bed U.S. Department of Defense teaching hospital with an emergency medi-cine residency and an annual census of 75,000 patients.
Selection of Participants Over the last several years, nationwide shortages of anti-emetics have narrowed the range of therapies available to Adult patients ages 18–50 years, presenting with their emergency physicians for treating headaches . The typical migraine headache, were identified by the triage frequency of medication allergies, plus strong recom- nurse or their assigned provider as potential subjects.
mendations to forego opiate use for headaches further Those meeting the Modified International Headache limit our options A new, readily available, safe Society's criteria for migraine () were offered and effective option for ED migraine treatment would participation by their treating physician or a research potentially benefit large numbers of patients.
coordinator when present . Exclusion criteria arelisted in .
Goals of the Investigation Our goal was to compare the efficacy of intravenoushaloperidol with intravenous metoclopramide (both in All subjects provided written informed consent. Patients combination with diphenhydramine) for the treatment declining enrollment received standard migraine therapy of acute migraine headache in the ED. The primary at the discretion of the treating physician. After informed outcome measure was pain relief, measured using a visual consent was obtained, each subject had a peripheral intra- analog scale (VAS). A difference in pain relief of 13 mm venous (i.v.) catheter placed and a bolus infusion of between the two groups was considered a priori to be clin- 1000 mL of normal saline begun. A point-of-care whole ically significant.
blood electrolyte panel (Istat9; Abbott Point-of-Care, Secondary outcome measures included time to East Windsor, NJ) was drawn, a cardiac monitor was maximal pain relief; the use of rescue medication; VAS placed, and an electrocardiogram (ECG) completed. Fe- measurements of nausea, sedation, and anxiety/restless- male patients provided a urine sample for a point-of- ness (akathisia); electrocardiographic Q-T intervals prior care pregnancy test. Diphenhydramine 25 mg i.v. was to and after treatment; and responses to a follow-up tele- administered, followed by the study medication. All parties were blinded as to the study medication adminis-tered: metoclopramide 10 mg i.v., or haloperidol 5 mg i.v.
MATERIALS AND METHODS Both were given over 2 min. Subjects were assigned toone of the two arms by means of a random-numbers table Study Design and Setting generated and maintained by the pharmacy. Study medi-cations were provided in identical coded syringes. Vital We conducted a prospective, double-blinded, random- signs (blood pressure, pulse, respiratory rate, oral temper- ized, controlled trial on a convenience sample of patients ature, and oxygen saturation) were assessed at triage, dur-ing the course of care (per ED protocol), and prior to Table 1. Inclusion Criteria discharge. Per protocol, subjects were to remain on car- Inclusion Criteria diac monitors throughout their ED stay, and a repeatECG was to be completed prior to discharge.
Ages 18 to 50 yearsAt least two of the following: Unilateral location Methods of Measurement Throbbing characterWorsening pain with routine activity The time of study-medication delivery was considered Moderate to severe intensity At least one of the following features: Time 0. Pain, nausea, restlessness (akathisia), and seda- Nausea or vomiting tion were each assessed via separate 100-mm nonhatched Photophobia or phonophobia VAS presented to the subject at 0, 20, 40, 60, and 80 min.
Haloperidol vs. Metoclopramide for Migraine Therapy in the ED Table 2. Exclusion Criteria Known hypersensitivity to haloperidol, metoclopramide, or diphenhydramineHistory of heart disease: congestive heart failure (CHF), myocardial infarction (MI)/coronary artery disease (CAD), atrial fibrillation (AF), Electrolyte abnormalities on IStat testing (defined as outside of the reference range)Use of prodysrhythmic drugs: quinidine, disopyramide, flecainide, propafenone, moricizine, amiodarone, dofetilide, sotalol, dronedaroneProlonged QT interval on screening ECG (>450 ms)History of uncontrolled hypertension with presenting diastolic blood pressure > 100 mm HgUse of an ergotamine derivative, triptan, or dopamine blocking antiemetic within the past 24 hHypothyroidismConcurrent administration or within 2 weeks of discontinuing an MAOI inhibitorConcurrent management of hemiplegic or basilar migraine or known neurologic disorderSevere hepatic impairmentPregnantBreastfeedingHistory of cancer (except nonmelanoma skin cancer)Previous involvement in the studyFebrile to 38.06C (100.5F) or greater on presentationAny indication for further diagnostic evaluation of their headache such as lumbar puncture or CT scanPresenting headache differs from the subject's normal migraine headache ECG = electrocardiogram; MAOI = monoamine oxidase inhibitor; CT = computed tomography.
The left side of the line represented no symptoms, with the right end representing the worst symptoms ever.
The four scales for each time interval were on a single The primary outcome measure was improvement in pain page along with four yes/no questions: as reported on the VAS within 80 min of therapy. An ab-solute difference of 13 mm or more was considered clin-ically significant . Secondary outcome measures Do you have nausea? Do you feel jittery, restless or anxious? were nausea, restlessness, sedation, the need for rescue Do you feel sleepy? medications, side effects and adverse reactions, and Do you have chest tightness, heaviness QTc interval change. Secondary outcomes assessed at callback were subject satisfaction with the study drug,and persistence of symptoms.
At the completion of their stay, subjects were asked if they would ‘‘want the same medication again to relieve Primary Data Analysis their migraine.'' If the subject was satisfied with pain re-lief and requested discharge prior to 80 min, this was al- Sixty-two subjects were required (31 per group) to detect lowed. After 80 min, the subject was either discharged a 13-mm VAS difference between the groups with a po- home if their pain control was adequate, or offered rescue wer of 81%. Differences between means were compared therapies at the discretion of the treating physician. The with the t-test. Differences in proportions were assessed use of rescue medicines was recorded along with any using chi-squared analysis.
adverse reactions. Prior to discharge, the subject had asecond ECG ordered to assess any changes in QT interval.
Subjects were to be contacted by the primary investi- gators at 48 h after their discharge from the ED and Characteristics of Study Subjects queried using a standard script. They were asked if theywere happy with the medication received, and if they Subject enrollment is detailed in . A total of 138 had any recurrent or persistent symptoms: specifically, patients presented during the enrollment time frame with headache, sleepiness, restlessness, agitation, nausea, a chief complaint of migraine headache. Potential candi- vomiting, or chest pain. If we were unable to contact dates were identified by the triage nurse, primary nurse, the subject at 48 h, further follow-up phone calls were or provider, and screened. A total of 64 patients were made again at 72 h, 96 h, 7 days, and finally at 2 weeks, enrolled in the study, with 31 randomized to receive halo- for a total of four follow-up attempts. If we were unable to peridol and 33 receiving metoclopramide. The study pro- follow up the subject with a phone call, we checked all tocol maintained an intention-to-treat analysis.
available Department of Defense electronic databases at The demographic characteristics of the subjects are the 2-week mark to see if they had made any return visits found in Subjects in the two groups did not differ to the ED or their primary care provider.
in age or sex.

M. E. Gaffigan et al.
Figure 1. Flow chart of subjects enrolled in the study.
Secondary Outcomes Nausea. Overall, both agents were effective at treating VAS scales for pain, nausea, restlessness, and sedation nausea, when present, with mean VAS reductions of were presented to the subjects at Time 0 (‘‘Pre'') and 27 mm for metoclopramide and 31 mm for haloperidol every 20 min until the last recorded measurement or the (p < 0.01 for both). There were no differences between 80-min period (‘‘Post''). When an individual subject's the haloperidol and metoclopramide groups in reported data collection was suspended prior to 80 min (usually nausea either prior to or after treatment (VAS - for requesting discharge), the last recorded measurementof that variable was used for ‘‘Pre-Post'' comparisons.
Restlessness. Within groups, neither agent produced a sta- Results are presented in tistically significant change in restlessness. There were nodifferences in restlessness between the groups prior to or after treatment (all p > 0.05).
The mean reduction in pain from baseline to the last re- Sedation. Subjects in both groups recorded a baseline corded measure of pain on the 100-mm VAS scale was level of sedation that did not change statistically within statistically and clinically significant for both haloper- each group after treatment, and that did not differ at base- idol- and metoclopramide-treated groups: 57 mm for line or study completion between groups (all p > 0.05).
the haloperidol group and 49 mm for those treated withmetoclopramide (p < 0.01 for each comparison). When Rescue Medications compared to each other, the VAS pain scores for the halo-peridol and metoclopramide groups did not differ at base- Eight of the 33 subjects in the metoclopramide group line, at the last recorded measurement, or in the (24%) were given rescue medications, compared with magnitude of the pre-post treatment change (p > 0.05).
only 1 of the 31 subjects (3%) receiving haloperidol The two regimens were also equivalent in the time to (p < 0.02). In the metoclopramide group, 7 patients achieve maximum pain relief (defined as the average received intravenous ketorolac, and one patient received measurement interval in which the subjects lowest VAS intravenous morphine. One patient in the haloperidol score was first recorded): 55 min for metoclopramide, group received intravenous ketorolac and methylprednis- 56 min with haloperidol (p > 0.05; VAS - olone as rescue medications. Only one patient requiredmore diphenhydramine, and that patient was in the meto-clopramide arm.
Table 3. Demographic Information There was no difference in the number of subjects report- ing adequate pain relief and requesting discharge prior to the 80 min point (35% in the haloperidol group and 21% Mean age in years (SD) in the metoclopramide group, p > 0.05).

Haloperidol vs. Metoclopramide for Migraine Therapy in the ED Figure 2. Mean visual analog scale (VAS) in mm with 95% confidence intervals (CI).
Patient Satisfaction jects. After the study drug was administered, an additional22% of patients reported new sleepiness. Only one subject When asked at the completion of their ED stay, 92% of 24 developed nausea after treatment. Sleepiness was statisti- responding subjects in the haloperidol group (and 75% of cally more common in the group that was to receive halo- 28 responding subjects in the metoclopramide group) peridol. There were no other differences between the would want the same medication again (p > 0.05, chi- groups in any of the other side-effect questions asked. In addition to the four specific questions, caregivers wereencouraged to report separately any observed adverse Side Effects/Adverse Events events. There were recorded events for four subjects inthe haloperidol group (‘‘restless''/‘‘loopy''; ‘‘restless Subjects were specifically queried regarding the presence legs''; ‘‘nausea – given ondansetron''; ‘‘difficult to of nausea, restlessness/anxiety, sleepiness, and chest wake''). Two metoclopramide-treated subjects had re- discomfort at baseline and at the same 20-min intervals corded observations: ‘‘jittery, momentarily short of breath as the VAS scoring. Results are presented in .
during i.v. push''; ‘‘symptoms relieved at 40 min but re- The Table lists the proportions of subjects in the two groups who either responded affirmatively at baseline(prior to study-drug administration), or who reported ECG and QTc Intervals the index symptom at least once beginning after interven-tion. The denominator was the number of subjects in each Subjects were to remain on cardiac monitors throughout treatment group.
their ED stay. QTc intervals were obtained from the Sleepiness and nausea were most commonly reported.
printed ECGs. All subjects had unremarkable pretreat- Even prior to study drug administration, both nausea and ment ECGs. Posttreatment cardiograms were performed sleepiness were acknowledged by two-thirds of all sub- on 45% of the subjects. Their pre- and posttreatment Table 4. Side Effects Haloperidol (n = 31) Metoclopramide (n = 33) Baseline (Time 0) (%) Baseline (Time 0) (%) Metoclopramide (%) *p < 0.02, chi-squared, baseline measurement: Haloperidol vs. Metoclopramide. All other paired comparisons (Haloperidol vs. Metoclo-pramide, baseline, or poststudy drug) were not significant (p > 0.05, chi-squared).
M. E. Gaffigan et al.
Table 5. Mean Electrocardiographic QTc Intervals (ms) infrequently reported and no subject contacted had vom-iting or chest pain (See ).
Haloperidol (H) Metoclopramide (M) p (H vs M) p (Pre vs. Post) In this, the first randomized, double-blinded comparison, NS = p > 0.05; t-test.
intravenous haloperidol (5 mg) was equivalent to intrave-nous metoclopramide (10 mg) in the successful ED treat-ment of pain and nausea from migraine headaches QTc intervals are presented in Mean QTcs were (The degree of pain relief averaged 53 mm equal and normal in the two groups and did not change on the VAS over both groups, and the time to maximum after treatment for either group.
pain relief was the same. Ninety percent of all subjects No dysrhythmias were reported. The four subjects re- contacted were ‘‘happy with the medication'' they had porting chest pain were re-evaluated and received repeat received. These outcomes compare favorably with other ECGs, and subsequently, they were not felt to require studies of neuroleptic antiemetics in migraine further cardiac evaluation. No subject complained of pal- pitations, had clinically relevant dysrhythmias, or had any obviating the need for rescue medications. Only one cardiac adverse event during treatment.
subject in the haloperidol group required rescue (3%),while 8 (24%) of the metoclopramide group required further pain control. These findings are mirrored in thevisit satisfaction data: over 90% in the haloperidol Of the 64 enrolled subjects, 43 were reached for a follow- group would take the same treatment, compared to 75% up phone interview. Overall, 90% answered affirmatively of the metoclopramide group (p = 0.11).
when asked if they were happy with the medication they There were no differences in side effects while in the received in the ED. Sixty percent of metoclopramide- ED, with a tendency for more restlessness with haloper- treated subjects and 48% of haloperidol-treated subjects idol (NS, p < 0.051). While there are few reports quanti- reported return of headache symptoms after discharge.
fying haloperidol's side effects in migraine treatment, it Three subjects, all in the metoclopramide group, returned may be more likely to promote akathisia than other neuro- to the ED. In response to questions about other symptoms leptic antiemetics. When haloperidol (without diphenhy- since ED discharge, sleepiness was the most commonly dramine) was successfully given for migraines, half of the reported, affecting 40% and 52% of the metoclopramide subjects reported akathisia, with 16% considering this and haloperidol groups, respectively. The only statisti- side effect ‘‘intolerable'' In contrast, when metoclo- cally significant difference in side effects between treat- pramide was used alone (for nausea), akathisia was found ment groups was in reported restlessness, reported in in 12% . Our own unpublished experience with 43% of the haloperidol group, but in only 10% of those higher doses of diphenhydramine (50 mg) and lower receiving metoclopramide. Agitation and nausea were doses of haloperidol (2.5 mg) support formal study ofsuch dosage adjustments as a means of further limitingthis symptom.
Table 6. Telephone Callback: Positive (Yes) Responses Haloperidol was more likely to be associated with ‘‘any restlessness'' in the contacted callback subjects (43% of the haloperidol group vs. 10% of those receiving Number contacted (%) Happy with medication* Sixty percent of metoclopramide-treated subjects and Headache recurred (%)* 48% of the haloperidol-treated subjects contacted re- Mean recurrent headache intensity/disability (out of 10)** ported return of headache symptoms after discharge.
Returned to ED for headache* This degree of headache recurrence supports recommen- dations for routinely adding corticosteroids to standard Any restlessness*** ED migraine abortive therapy To our knowledge, this is the first study of acute migraine therapy to directly compare haloperidol to me- toclopramide. Intravenous metoclopramide was one of ED = emergency department.
the first medications in the neuroleptic/antiemetic class *p > 0.05, chi-squared.
to be shown safe and effective for migraine It has **p > 0.05, t-test.
***p < 0.015 chi-squared.
been used as the comparison standard in studies of other Haloperidol vs. Metoclopramide for Migraine Therapy in the ED agents This study adds haloperidol to the list of safe ments (Time 0, coinciding with administration of the and effective ED headache therapies, which may be study drug). It is not clear why baseline sleepiness was particularly helpful in the current setting of frequent higher in the group to receive haloperidol.
medication shortages Pain and nausea reduction in Continuous ECG monitoring was inconsistently per- both study groups was substantial, and there was no sta- formed. Posttreatment ECGs were obtained in 45% of tistically significant difference in the timing or degree the subjects with no changes in QTc. No clinically detec- of those treatment effects between groups.
tible cardiac events occurred.
There were no reported incidences of palpitations or This was a single-center study, and as noted in dysrhythmias. QTc intervals were unaffected by either this was a relatively young study population with a large of the study drugs. Intervals were measured, subjects percentage of female subjects. Our exclusion criteria were monitored, and those with risk factors for QT pro- were designed to limit the likelihood of any adverse QT longation (QTP) excluded ). In their review of prolongation as well, and future studies or clinical appli- the evidence behind the 2007 U.S. Food and Drug cation should note the exclusion criteria.
Administration concerns for QTP and torsades de pointes The percentage of patients contacted for telephone with i.v. haloperidol, Meyer-Massetti et al. noted the rar- follow-up (67%) was adequate, but not large. This follow- ity of those events (70 cases in over a decade), and their up rate is better than some studies, but less than others per- association, in 97% of cases, with additional risk factors formed at our institution None of the subjects we for QTP or torsades de pointes (preexisting heart disease, failed to contact returned to our ED or another military medical facility during the 2 weeks after their enrollment.
drugs, and mechanical ventilation) . Their studyand others, in combination with our QTc data, suggest that ECGs and QTc measurements, electrolytes, and car-diac monitoring are unnecessary when using haloperidol In summary, haloperidol is at least as safe and effective as for headache in similar populations metoclopramide for the ED treatment of migraine head-aches, with less need for rescue medications. Despite an increase in restlessness with haloperidol, patient satisfac-tion was high for both therapies.
The study was adequately powered. It is possible that alarger sample might have allowed some of the trendsobserved to reach statistical significance: for example, 92% (haloperidol) vs. 75% (metoclopramide) subjectsatisfaction discharge more common with haloperidol (32% vs 12%, p = 0.0515); 3 subjects in the metoclopra- mide group returned to the ED (vs. 0, p < 0.054). Statis- tical significance in any of these trends would not alter our conclusions. The fact that 28% of subjects chose to leave when their symptoms were controlled precluded acquisition of a few data points. While the use of a con- venience sample is never optimal, the demographic char- acteristics of our subject groups were similar, and the fact that multiple providers could enroll patients they were seeing enabled more night and weekend subject acquisi- tions than studies relying on the availability of a research coordinator alone.
While the observed pretreatment nausea is an expected feature of migraine headache, the sedation noted at base- line in two-thirds of all subjects was likely a function of the antecedent administration of diphenhydramine by protocol. The precise timing of the diphenhydramine- study drug interval was not monitored. While the study drugs were sent from the pharmacy, the diphenhydramine was available in the ED. By protocol, all subjects received their diphenhydramine prior to the baseline measure- M. E. Gaffigan et al.
Haloperidol vs. Metoclopramide for Migraine Therapy in the ED 1. Why is this topic important? Migraines are common complaints to emergency de- partments (EDs). These patients may have multiple al-lergies, and medication shortages have limited ourchoices. As a result, we often need alternative treatmentregimens to help mitigate their pain without using opioids.
2. What does this study attempt to show? This randomized control trial compares intravenous haloperidol to intravenous metoclopramide in adult pa-tients presenting to the ED with acute migraine head-aches.
3. What are the key findings? Haloperidol is an equally effective migraine therapy when compared to metoclopramide. Haloperidol hadfewer recurrent or persistent headaches on follow-up,less need for rescue medications, superior patient satisfac-tion, and no clinically apparent effect on the QTc interval.
There was an increase in restlessness noted with haloper-idol.
4. How is patient care impacted? This study demonstrated that haloperidol is an effec- tive, alternative, nonopioid treatment for migraine pa-tients. Patients were satisfied with their pain control,and would want the same treatment on future ED visits.

Source: http://emupdates.com/perm/Gaffigan%202015%20Haldol%20vs%20Reglan%20For%20Headache%20-%20JEM.pdf