Randomized Trial Comparing Low- vs High-Dose IV Dexamethasone for Patients With Moderate to Severe Migraine

Abstract

Background and Objectives

Dexamethasone decreases the frequency of migraine recurrence after emergency department (ED) discharge. However, the optimal dose of dexamethasone is unknown. We hypothesized that dexamethasone 16 mg IV would allow greater rates of sustained headache relief than 4 mg when coadministered with metoclopramide 10 mg IV.

Methods

This was a randomized double-blind study. Adults who presented with a headache meeting International Classification of Headache Disorders, 3rd edition, migraine criteria were eligible if they rated the headache as moderate or severe in intensity. Pain intensity was assessed for up to 2 hours in the ED and through telephone 48 hours and 7 days later. The primary outcome was sustained headache relief. Secondary outcomes included headache relief within 2 hours and the number of headache days during the subsequent week. Relying on a priori criteria, the data safety monitoring committee recommended halting the study early for futility.

Results

A total of 1,823 patients were screened, and 209 patients were randomized. The mean age was 38 years (SD 11). One hundred seventy-nine of 209 (86%) identified as women. One hundred fifty-one of 209 (72%) of the population reported severe intensity; the rest reported moderate. Thirty-five of 102 (34%) participants in the metoclopramide +4 mg arm achieved sustained headache relief as did 42/102 (41%) participants in the metoclopramide +16 mg arm (absolute difference 7%, 95% CI −6% to 20%). Headache relief within 2 hours occurred in 77/104 (74%) low-dose and 82/105 (78%) high-dose participants (absolute difference 4%, 95% CI −8% to 16%). During the week after ED discharge, low-dose participants reported a median of 2 headache days (25th, 75th percentile 1, 5); in the high-dose arm, this was also 2 (25th, 75th percentile 0, 4) (mean difference 0.4, 95% CI −0.3 to 1.2).

Discussion

When added to 10 mg IV metoclopramide, doses of dexamethasone greater than 4 mg are unlikely to benefit patients in the ED with migraine.

Trial Registration Information

This study was registered at ClinicalTrials.gov on October 2, 2019 (NCT04112823). The first patient was enrolled on December 22, 2019.

Classification of Evidence

This study provides Class I evidence that 16 mg of IV dexamethasone is unlikely to provide greater rates of sustained headache relief than 4 mg of IV dexamethasone among patients in the ED with migraine treated concurrently with IV metoclopramide.

Introduction

Migraine, a typically episodic headache disorder characterized by recurrent exacerbations, causes 1.2 million visits to US emergency departments (EDs) annually.¹ Nearly two-thirds of patients with migraine experience worsening or recurrence of headache within 48 hours of successful initial treatment in the ED.² About half of patients suffer more than 2 days of headache during the week after ED discharge.³

The corticosteroid medication dexamethasone, when administered during the ED visit, can minimize the recurrence of postdischarge headache. The American Headache Society recommends dexamethasone for use in ED patients with acute migraine.⁴ Meta-analytic data indicate that dexamethasone has a number needed to treat of approximately 9 to decrease the frequency of moderate or severe headache within 72 hours of ED discharge.⁵ The mechanism of action of corticosteroids for migraine has not been described. It may mitigate the neurogenic inflammation that is believed to underlie trigeminovascular activation, a key component of migraine pathophysiology.⁶ Clinical data from ED-based randomized clinical trials indicate that dexamethasone may begin to take effect within 1 hour of administration.^7-9

However, the optimal dose of dexamethasone is not known. Aggregated data do not reveal an association between dose of steroids and outcome.^5,10-18 Two randomized studies that used very large doses of dexamethasone (20 mg¹⁵ and 24 mg¹¹) vs placebo failed to show a statistically significant difference regarding headache recurrence and reported point estimates than were similar to studies that used just 10 mg of dexamethasone.^13,19 We therefore performed a randomized study to determine whether high-dose IV dexamethasone (16 mg) affords more relief for patients in the ED with migraine than low-dose IV dexamethasone (4 mg) when coadministered with metoclopramide 10 mg IV, a standard-of-care migraine abortive medication.⁴ We hypothesized that the dose of dexamethasone would be associated with the rate of sustained headache relief.

Methods

Overview

This was a randomized, double-blind, comparative efficacy study conducted among patients who presented to 1 of 2 EDs in the Bronx, NY, with migraine. All participants received metoclopramide 10 mg IV as the primary abortive medication. In addition, all participants received a dose of dexamethasone IV. There was no placebo arm. Outcomes were assessed in the ED and through telephone 48 hours and 7 days after the ED visit using a standardized instrument with a closed-question format.

Standard Protocol Approvals, Registrations, and Patient Consents

All participants provided written informed consent. The study was reviewed and approved by the Albert Einstein College of Medicine Institutional Review Board. It was registered online at ClinicalTrials.gov (NCT04112823). The study protocol and statistical analytic plan are presented as supplemental data (eSAP 1, links.lww.com/WNL/D60).

Population of Interest

Eligible patients were adults presenting to the ED for the treatment of headache fulfilling International Classification of Headache Disorders (ICHD) migraine without aura criteria, although we only required participants to have experienced one similar headache previously and we did not exclude patients if the headache duration was >72 hours.²⁰ Therefore, some participants had probable migraine without aura, and some patients had status migrainosus. We also did not exclude patients who presented with a headache that met all migraine without aura criteria except item E (not better accounted for by another ICHD-3 diagnosis) if that other diagnosis was migraine with aura. At the time of enrollment, the patient had to rate the headache as moderate or severe in intensity. Patients were excluded for suspicion for secondary headaches, including fever or focal neurologic findings on physical examination. Patients were also excluded if already using corticosteroids, for contraindications or allergies to the investigational medications, or if pregnant or breastfeeding.

Salaried, bilingual (Spanish and English) research associates, who staff our EDs 24 hours per day, 7 days per week, screened potentially eligible patients using a diagnostic checklist. Diagnosis of migraine was confirmed by the clinical attending physician.

Intervention

We randomized study participants to receive metoclopramide 10 mg as an IV drip over 15 minutes + dexamethasone 4 mg IV or metoclopramide 10 mg + dexamethasone 16 mg IV. We chose the 4 mg dose because it was the lowest dose of dexamethasone previously tested in migraine clinical trials.^21,22 The 16 mg dose is approximately the mean dose used in previous ED-based clinical trials,⁵ is a plausible high dose of dexamethasone, and, as a multiple of 4, was pragmatic for the pharmacists to prepare. Randomization occurred in blocks of 4 based on a random number generator. Study participants, clinicians, and research personnel were blinded to assignment using the following mechanism: the research pharmacist prepared the investigational medication in a secure location removed from the ED. Dexamethasone is a clear solution. The pharmacist diluted the 4 mg dose so that a volume of investigational medication comparable with the 16 mg dose was presented to the clinical nurse, who inserted both the metoclopramide and dexamethasone into a 50 mL bag of normal saline, which was then administered to the patient as an IV drip over 15 minutes.

Measures and Outcomes

The research associates obtained all data through oral interview. The 48-hour and 7-day follow-ups were obtained through telephone. We attempted to complete follow-up thrice daily until successful or deemed futile.

Baseline variables of interest included age, sex, duration of headache, presence of aura symptoms, and use of medication to treat the headache before ED presentation. We assessed headache intensity using the 4-item descriptive scale commonly used in headache clinical trials on which participants describe their headache intensity as “severe,” “moderate,” “mild,” or “none.”²³ We assessed adverse events by asking participants whether they developed any new symptoms that they attributed to the study medication and followed up affirmative answers using an open-ended format. The use of additional medication for the treatment of headache was assessed in the ED by asking the clinical team and at 48 hours by asking the patient.

The primary outcome for this study was sustained headache relief for 48 hours, defined as achieving a headache intensity of “none” or “mild” within 2 hours of medication administration and maintaining this level, without requiring additional headache medication, for the entire 48-hour follow-up period. Thus, participants who had a moderate or severe recurrence at 24 hours, which abated before 48 hours, would have been counted as an outcome failure.

Secondary efficacy end points included obtaining headache relief (achieving a headache intensity = mild or none) within 2 hours of medication administration and the number of headache days during the week after ED discharge. As an exploratory outcome, we asked participants during the 48-hour follow-up whether they would want to receive the same combination of medications during a hypothetical subsequent visit.

Sample Size Calculation and Analysis

The sample size was designed to detect a 15% absolute difference. Assuming a primary outcome rate of 30% in the high-dose group and 45% in the low-dose group, we calculated the need for 163 patients in each arm, using α = 0.05 and β = 0.20. To this, we added a 10% margin to account for protocol violations and patients lost to follow-up. Therefore, we intended to enroll 360 patients.

As per the study protocol, we performed an interim analysis after 200 patients had been enrolled, slightly past the halfway point. The intention of this interim analysis was to halt the study for futility, clear superiority, or a large discrepancy in adverse medication events. Futility was defined as <10% absolute difference, as continuing the study beyond this point would be unlikely to result in a statistically significant difference between the groups. Clear superiority was defined as >20% absolute difference. A large discrepancy in adverse events was defined as >20% absolute difference in the overall rate of participants who suffered any adverse event. After completing the interim analysis, the data monitoring committee recommended that enrollment be halted.

Baseline characteristics are reported as mean with SD, median with 25th and 75th percentile, or n with percent. Dichotomous outcomes are reported as n/N with percent. The absolute difference is reported with 95% CI. Headache days are reported as median (25th, 75th percentile). The between-group difference is reported as mean with 95% CI. We report the correlation between headache duration and primary outcome using Spearman ρ. We report the impact of discrepant baseline characteristics on the primary outcome using Spearman ρ for correlation and by looking at the impact of including or removing the discrepant baseline characteristic in multivariable logistic regression models.

Data Availability

Deidentified data will be made available by request to the principal investigator.

Results

Enrollment commenced in December 2019 and paused between March and June 2020 when all in-person clinical research halted at our institution, and after completion of the interim analysis, enrollment was permanently halted in August 2022. Altogether, 1,823 patients were screened, and 209 patients were randomized. Please see the Consolidated Standards of Reporting Trials flow diagram (Figure).

Figure. CONSORT Flow Diagram.

Participants lost to 48-hour follow-up were included in the primary analysis if they failed the primary outcome before ED discharge. CONSORT = Consolidated Standards of Reporting Trials; ED = emergency department.

Baseline characteristics are presented in Table 1. Medications used before the ED visit are presented in eTable 1 (links.lww.com/WNL/D61). Only about two-third of study participants used any medication for headache before ED presentation. Of the ones who used medication, most used nonsteroidal anti-inflammatory drug or acetaminophen.

Table 1.

Baseline Characteristics

Regarding the primary outcome, 35/102 (34%) patients in the metoclopramide +4 mg arm achieved sustained headache relief as did 42/102 (41%) patients in the metoclopramide +16 mg arm (absolute difference 7%, 95% CI −6% to 20%). Although sex was associated with the dose of dexamethasone, it was not correlated with the primary outcome (ρ = 0.11, p = 0.11). Including sex in a logistic regression model did not substantially affect the odds of sustained relief (without sex: odds ratio [OR] 1.0, 95% CI 1.0–1.1; with sex: OR 1.0, 95% CI 1.0–1.1) (eTables 2a and 2b, links.lww.com/WNL/D61). Secondary outcomes are presented in Table 2. Duration of headache was not correlated with the primary outcome (Spearman ρ = 0.00, p = 0.96) and did not affect the relationship between the dose of dexamethasone and the primary outcome (eTables 2a and 2c). Outcomes among the 52 participants with headache duration >72 hours are presented in Table 3; headache duration did not meaningfully affect the results. Similarly, the presence of nausea did not affect the relationship between the dose of dexamethasone and sustained headache relief (eTables 2a and 2d). Most participants and a comparable number between the study groups reported that they would want the same combination of medications during a subsequent visit (Table 4).

Table 2.

Secondary Outcomes

Table 3.

Outcomes Among Participants With Headache Duration >72 Hours

Table 4.

At 48 Hours, We Asked the Patients Whether They Would Want the Same Treatment the Next Time They Came to the ED for Migraine

Adverse events are presented in Table 5. There was 1 serious adverse event requiring hospital admission; a 23-year-old man was diagnosed with a cerebellar stroke within 1 week of study enrollment. A second participant reported markedly elevated blood sugar. Both of these participants received 4 mg of dexamethasone.

Table 5.

Adverse Medication Events

Discussion

In this randomized, double-blind ED study, a 16 mg dose of IV dexamethasone did not improve outcomes more than a 4 mg dose for patients with migraine when added to metoclopramide 10 mg IV. We identified no clinically important absolute differences between groups in the ED, at 48 hours, or up to 1 week later. These results are particularly disappointing given the overall low success rate observed in this clinical trial; despite using evidence-based, guideline recommended treatment, nearly two-thirds of our patients were not able to achieve and maintain headache relief without the use of additional headache medication.

The efficacy of corticosteroids for ED patients with migraine has been established for more than a decade,⁴ but optimal dosing strategies were not known. It is clear that when administered to a general population of patients with acute migraine, corticosteroids decrease the frequency of headache recurrence with a number needed to treat of 9 or 10.^18,24 In previous work, we showed that a longer-acting depot intramuscular corticosteroid did not decrease the frequency of “after-shock” headaches more than intramuscular dexamethasone during the week after ED discharge.³ Those results, combined with the results of this study, seem to indicate that more is not better when it comes to the treatment of migraine with corticosteroids. Some have suggested that a 1-week course of oral steroids may have increased efficacy. To us, the data cited above about the depot intramuscular corticosteroids indicated that this latter strategy is unlikely to be effective. Similarly, data do not indicate that any particular phenotype of migraine is more likely to benefit.

One subgroup of patients worth discussing is those with status migrainosus, a complication of migraine characterized by debilitating and unremitting headache for >72 hours. Some have suggested that corticosteroids are more likely to be of benefit among these patients.²⁵ Although the focus of this study was not status migrainosus, about one-fourth of our sample reported headache duration >72 hours. Yet even after controlling for headache duration, we did not identify any impact of the higher dose.

Other strategies exist for treating postdischarge headaches. In 1 randomized study, sumatriptan 100 mg or naproxen 500 mg were dispensed to patients with migraine at the time of ED discharge to be used as needed for the treatment of headache recurrence.²⁶ This strategy effectively treated headache recurrence for 75% of patients who used naproxen and 68% who used sumatriptan.

As has been shown elsewhere, many ED patients with migraine continue to suffer long after the ED visit.^2,3 In our sample, more than one-fourth reported moderate or severe headache over the 48 hours after ED discharge and nearly one-third reported experiencing headache often or always throughout the week after ED discharge. This may be a greater problem for patients in underserved communities, who cannot access specialty care in a timely manner. These data do make a powerful argument for emergency physicians who work in this setting to familiarize themselves with affordable outpatient abortive treatments.

Limitations of this study are mostly related to generalizability. We conducted this study in 2 academic EDs in the Bronx. It is unclear whether these results can be generalized to patients in other settings, those who receive abortive treatments other than metoclopramide, and those who receive oral rather than IV corticosteroids. Because the study was halted early, we do not know the true magnitude of the absolute differences; there remains a possible 20% absolute benefit for the higher dose of dexamethasone. Finally, we did not determine the impact of medical comorbidities on the relationship between dose and efficacy.

In conclusion, IV doses of dexamethasone greater than 4 mg are unlikely to benefit patients in the ED with migraine when added to metoclopramide 10 mg IV.

Glossary

ED

emergency department

ICHD-3

International Classification of Headache Disorders, 3rd edition

OR

odds ratio

Appendix. Authors

Study Funding

National Center for Advancing Translation Sciences UL1TR001073.

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/N for full disclosures.