The place of corticosteroids in migraine attack management: A 65-year systematic review with pooled analysis and critical appraisal

Abstract

Background and objectives

Headaches recur in up to 87% of migraine patients visiting the emergency department (ED), making ED recidivism a management challenge. We aimed herein to determine the role of corticosteroids in the acute management of migraine in the ED and outpatient care.

Methods

Advanced search strategies employing PubMed/MEDLINE, Web of Science, and Cochrane Library databases inclusive of a relevant gray literature search was employed for Clinical Studies and Systematic Reviews by combining the terms “migraine” and “corticosteroids” spanning all previous years since the production of synthetic corticosteroids ca. 1950 until August 30, 2014. Methods were in accordance with MOOSE guidelines.

Results

Twenty-five studies (n = 3989, median age 37.5 years, interquartile range or IQR 35–41 years; median male:female ratio 1:4.23, IQR 1:2.1–6.14; 52% ED-based, 56% randomized-controlled) and four systematic reviews were included. International Classification of Headache Disorders criteria were applied in 64%. Nineteen studies (76%) indicated observed outcome differences favoring benefits of corticosteroids, while six (24%) studies indicated non-inferior outcomes for corticosteroids. Median absolute risk reduction was 30% (range 6%–48.2%), and 11% (6%–48.6%) for 24-, and 72-hour headache recurrence, respectively. Parenteral dexamethasone was the most commonly (56%) administered steroid, at a median single dose of 10 mg (range 4–24 mg). All meta-analyses revealed efficacy of adjuvant corticosteroids to various abortive medications—indicating generalizability. Adverse effects were tolerable. Higher disability, status migrainosus, incomplete pain relief, and previous history of headache recurrence predicted outcome favorability.

Conclusions

Our literature review suggests that with corticosteroid treatment, recurrent headaches become milder than pretreated headaches and later respond to nonsteroidal therapy. Single-dose intravenous dexamethasone is a reasonable option for managing resistant, severe, or prolonged migraine attacks.

Background

Burden of recurrent migraine

Primary headache disorders are increasingly important causes of disabling public health and socioeconomic burden worldwide (1). In the United States (US) alone, more than 30 million people suffer from migraine (2–4), and an estimated 3 million headache patients visit emergency departments (EDs) annually (4,5). Migraines are often underestimated and misdiagnosed—and thus are under- or mistreated (2,6). Migraine costs the US economy nearly $17 billion annually; direct costs are increased by incomplete pain relief, and repeated ED visits. Lost productivity due to headache accounts for the bulk of the economic burden (7,8).

Studies indicate that post-ED recurrent headaches occur within 24 hours in up to 87% of migraine patients visiting the ED (5,9–15), and more than half of these recurrent headaches are functionally disabling (14,16). Previous history of headache recurrence, longer headache duration, severe baseline pain, and persistent pain at discharge all contribute to post-ED headache recurrence (5,15). Repeated ED visits are a continuing problem in the ED setting (17), making ED recidivism a significant management challenge (17,18). Undertreated or poorly treated headaches (18), “medication-seeking” behavior (18,19), and medication overuse (17) are features associated with the ED-repeater phenotype. Lower headache recurrence rates are found to be significant measures according to patients’ self-reported satisfaction for optimum outcome (20).

Algogenic properties of neurovascular inflammation in migraine

Current evidence indicates that migraine is a neurovascular inflammatory disorder (21–23). It is a dysfunction of sensory processing of painful stimuli (24). Pain is generated either peripherally from trigeminocervical neurons through abnormal activation at primary afferents (25) or centrally due to anomalous modulation at second-order neurons (26–28). Once nociceptive primary afferent trigeminocervical systems are peripherally activated, signals travel orthodromically and neuropeptides are released onto receptors on second-order neurons of the trigeminal nucleus caudalis (23,29,30). The same peptidergic nociceptive primary afferents subserve sensory-efferent functions whereby peptides are likewise released at their peripheral perivascular terminals (30). Neuropeptides modulate neuroregulation of arteriolar vasodilation, inflammation, and smooth muscle changes (23,31). Neurally induced inflammatory discharge of algogenic pro-inflammatory mediators, plasma protein extravasation, meningeal irritation, perivascular edema, and dural platelet aggregation are inherent in animal models of trigeminal ganglion stimulation or pouring pro-inflammatory mediators over animal dura mater (23,32). Vascular and inflammatory changes follow neuronal changes (23,32,33). This complex inflammation-pain interplay is broadly termed “sterile neurogenic inflammation” (SNI) (23,32). SNI contributes to development of hyperalgesia, pain prolongation, and peripheral sensitization of polymodal receptors in animal models (23,26,34). Here, it is important to note that the SNI theory has not been shown to be significant in the human condition of migraine; differences and translational challenges in neuroethological approaches need to be considered. In addition, multiple SNI inhibitors (e.g. neurokinin receptor antagonists (35), neurosteroids (36,37)) have failed to be efficacious for both acute and preventive treatment of migraine. Among neuropeptides, a recent systematic review (38) has shown four clinical studies reporting increased calcitonin gene-related peptide (CGRP) during migraine attacks (39–42), while another three clinical studies found no difference (43–45).

Corticosteroids are the main humoral mediators of stress, and their increased secretion in response to adverse stimuli normally results in a cascade of physiological and behavioral homeostatic mechanisms (46). The term “corticosteroids” embraces two prototypic steroids, i.e. glucocorticoids and mineralocorticoids. By virtue of their physical features of lipophilicity (47–49) and small mass size (ca. 300 Da) (50,51), corticosteroids can readily pass the blood-brain barrier (47–49,52) and participate in a variety of behavioral and physiological processes, as well as taking part in their own secretion (46). In this respect, they serve as messengers between the periphery and brain, and also between the external and internal environments and the brain. Since corticosteroids act against the background of increased monoamine secretion, it is thought that they act to fine-tune the individual’s response to stress (53). Unlike the transient monoamine response, corticosteroids exert sustained actions on cellular activity and behavior, and therefore are essential for ensuring the orchestration of a coordinated adaptive response as well as “preparedness” of the individual to cope with future challenges (53).

Corticosteroids have been used as pain relief adjuvants in different neurological conditions; an adjuvant pain medication should be considered at all stages of the World Health Organization’s (WHO) pain ladder for mild to severe painful conditions (54). Corticosteroids reduce pain by inhibiting prostaglandin synthesis, which leads to inflammation, and by reducing vascular permeability that results in tissue edema (55). Exogenous administration of corticosteroids suppresses excessive stress and inflammatory response from the hypothalamic-pituitary axis (46). Steroid receptors are found in the central and peripheral nervous systems and are responsible for growth, differentiation, development, and plasticity of neurons (55). In particular, corticosteroids reduce spontaneous discharge in an injured nerve, which reduces neuropathic pain (56). Corticosteroids offer extended suppression of SNI (57) and modulate neuroplasticity, thereby reducing pain (56). Corticosteroids have long been used to manage migraine (58–62).

Aim

In this systematic review, we make a pooled analysis and critical appraisal of the published literature on clinical studies of corticosteroid treatment for migraine. Our objectives are to assess the efficacy of corticosteroid administration in the ED and outpatient clinical settings on the primary outcome of reducing headache frequency. We also compared corticosteroid administration to other standard drug regimens. Identification, screening, selection, and inclusion of clinical studies followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) procedures as described within the Methods and materials part of our study.

Methods and materials

A combination of the following search strategies using different search databases was employed to capture our topic of interest, i.e. studies on corticosteroids for migraine management.

A PubMed/MEDLINE search was employed for Clinical Studies Categories and Systematic Reviews on the PubMed Clinical Queries tool combining the terms “migraine AND corticosteroids”; the Boolean logic operator “AND” was applied to connect the two search terms. The Clinical Studies Category was selected for “Therapy,” “Etiology,” “Diagnosis,” “Prognosis,” “Clinical Prediction Guides,” and the scope of the search was made specific to “Broad” to enable sensitivity and specificity search values of 99% and 70% (63), respectively. The PubMed equivalent for this search method was “((clinical[Title/Abstract] AND trial[Title/Abstract]) OR clinical trials as topic[medical subject headings (MeSH) Terms] OR clinical trial[Publication Type] OR random*[Title/Abstract] OR random allocation[MeSH Terms] OR therapeutic use[MeSH Subheading]).”

A second PubMed/MEDLINE search was employed for Clinical Studies Categories and Systematic Reviews on the PubMed Clinical Queries Tool combining the terms “migraine AND corticosteroids.” The Clinical Studies Category was selected for “Therapy,” and the scope of the search was made specific to “Narrow” to enable sensitivity and specificity search values of 93% and 97% (63), respectively. The PubMed equivalent for this search method was “(randomized controlled trial [Publication Type] OR (randomized [Title/Abstract] AND controlled [Title/Abstract] AND trial [Title/Abstract])).”

A third PubMed/MEDLINE search was employed without using the Clinical Queries Tool. Search terms used were “migraine AND corticosteroids.”

Advanced PubMed/MEDLINE search was used by implementing auto-suggested MeSH Terms and Boolean logic operator “AND” as “(migraine[MeSH Terms]) AND corticosteroids[MeSH Terms].”

The Cochrane Library Advanced Search strategy was used by adding the search line “Search All Text” and employing the search terms “migraine AND corticosteroids.”

A Web of Science Advanced Search was employed by using the field tag “TS” for topic, the Boolean operator “AND,” and parentheses to create our query as “TS = (migraine AND corticosteroids)” on Indexes = CPCI (Conference Proceedings Citation Index), Science Citation Index Expanded (SCI-EXPANDED), Timespan = All years. Results included all languages and all document types.

Unpublished studies and a relevant reference search were exhaustively conducted. This enabled us to locate and capture unpublished studies from the gray literature (e.g. conference abstracts or research letters) on our topic of interest, with the goal of elucidating the impact of unpublished studies and ultimately avoiding publication bias of negative or positive studies.

A PRISMA (64) flow diagram describing the selection method is displayed in Figure 1. An additional relevant reference search was made to extract complete previous data on corticosteroids and migraine. Scientific abstracts and several prepublication trial registries for investigations related to the research question were searched. All published studies up to August 30, 2014, were covered; this meant that all clinical studies since the production of synthetic corticosteroids were reviewed. Two authors (YWW and RPC) reviewed each abstract, and where disagreement occurred, discussion and consensus was achieved with input from the third author (AMR). Clinically relevant data, useful evidence-based recommendations, strengths, limitations, and unaddressed areas of selected studies were carefully examined in this review.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram displaying the method of how selection of studies was made.

Where applicable, study quality was evaluated using the Jadad scale (65) and the Consolidated Standards of Reporting Trials (CONSORT) checklist (66) (Table 1); the former measures adequacy of reported randomization, blinding, and management of withdrawals and dropouts, while the latter similarly assesses enrollment, allocation, loss to follow-up, and analysis. Each study was awarded a score out of a maximum of 5 points for the Jadad scale. Methods were in accordance with Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines (67).

Table 1.

Comprehensive summary of included clinical studies reviewed.

First author, title of article, name of journal, year of publication, and study setting	Study design	Outcomes among intervention arms				Appraisals and clinical bottom lines (CONSORT) and Jadad scale compiled, where possible)
1. Gallagher (60). Emergency treatment of intractable migraine. Headache, 1986. ED setting, suburban hospital. NJ, USA.	Three-arm retrospective chart review of six-month period; all parenteral administrations Overall average headache duration = 54 hours. n = 162 Mean age = 41 years. M:F ratio = 1:2.45. Minimum headache duration = 36 hours.		M, P (Meperidine (75-100 mg) and promethazine (50 mg); (n = 73); male-to-female (M:F) ratio = 1:2.47)	D, M (Dihydroergotamine (1 cc) and meperidine (75–100 mg); (n = 32); M:F ratio = 1:2.20)	M, P, D (Meperidine (75–100 mg), promethazine (50 mg) and dexamethasone (8 mg); (n = 57); M:F ratio = 1:2.56)	Favorable outcomes for dexamethasone ED convenience sampling may create unrecognized source population bias ARR of 35% between M,P,D and D,M; 33% between M,P,D and M.P. Retrospective chart review may introduce recall bias of written record on charts; difficulty to control confounders due to retrospective nature of study
No/mild headache (no interference with usual activity) 24 hours post-treatment	29%	37%	72%
2. Foroughipour et al. (68) Randomized clinical trial of intravenous valproate (orifil) and dexamethasone in patients with migraine disorder. Iran J Med Sci 2013. ED and Headache Clinic, Iran.	Randomized, two-arm, double-blind. All status migranosus patients. n = 31 M:F = 1:9.3 Mean age = 33.4 years 900 mg IV valproate vs 16 mg IV Dexamethasone IV dexamethasone diluted in 150 cc normal saline and infused for 10 minutes IHS criteria applied Clinical trial registration: IRCT13891146234N2		Dexamethasone (16 mg IV) n = 12		Valproate (900 mg IV) n = 19	Non-inferior outcomes for dexamethasone ED and Headache Clinic, convenience sampling used Six lost to follow-up from dexamethasone reduced final number analyzed – this may affect result CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Headache recurrence within 72 hours	66.67%		68.42%
		Severe recurrent migraine attacks provoking another physician visit	33.33%		26.32%
		Mean age (years)	32.50 ± 11.19		33.89 ± 13.34
		Mean migraine history (per month)	6.00 ± 8.07		6.26 ± 8.99
		Presence of aura	8.34%		38.84%
		Recovery from nausea	91.67%		68.42%
		Recovery from photophobia	91.67%		78.95%
Pain-free response after treatment	33.33%		26.32%
3. Soleimanpour et al. (69) Effectiveness of intravenous dexamethasone versus propofol for pain relief in the migraine headache: A prospective double blind randomized clinical trial. BMC Neurol 2012. ED, Hospital. Iran.	Double-blind, randomized, two-arm n = 90 Mean age = 35.9 years M:F ratio = 1:2.1 IV propofol (n = 45) (10 mg every 5–10 minutes to a maximum of 80 mg, slowly) versus IV dexamethasone (n = 45) (0.15 mg/kg to a maximum of 16 mg, slowly) IHS criteria applied Clinical trial registration: IRCT201008122496N4		Dexamethasone (0.15 mg/kg IV to a maximum of 16 mg, slowly) (n = 45)		Propofol (10 mg IV every 5–10 minutes to a maximum of 80 mg, slowly) (n = 45)	Non-inferior outcomes for dexamethasone Study could have benefited from follow-up post-discharge Propofol administered in titration as compared to dexamethasone bolus form—this may create pharmacokinetic difference ED setting: convenience sample used CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
1.		Mean age (years)	36.27		35.65
1.		M:F ratio	1:1.65		1:2
1.		Mean VAS score – VAS baseline	8.11 ± 1.31		8 ± 1.52
1.		VAS 10 minutes	5.13 ± 1.47		3.08 ± 1.7
1.		VAS 20 minutes	3.73 ± 1.81		1.87 ± 1.28
1.	VAS 30 minutes	3.06 ± 2		1.44 ± 1.63
4. Saadah (61). Abortive migraine therapy in office with dexamethasone and prochlorperazine. Headache 1994. Health Center, OK, USA	Inclusion: headache severity, protracted duration, or resistance to abortive home therapy n = 108 mean age = 39 years M:F = 1:7.6 IHS criteria applied for diagnosis Premedication with 40 mg of oral famotidine for those receiving dexamethasone		Dexamethasone 10 mg IV over five minutes (n = 22)	Dexamethasone 20 mg IV over 10 minutes (n = 39)	Prochlorperazine 3.5 mg IV over five minutes followed by 20 mg of dexamethasone over 10 minutes (n = 47)	Favorable outcomes for dexamethasone Increasing dexamethasone dose from 10 to 20 slightly improved outcome, while side effects were reduced. Definitions of outcome terms of relapse, resolution, remission not clearly stated. Non-randomized design lacking blinding, and being based on patient preference may introduce bias. All included patients had failed abortive therapy with NMEC (naproxen sodium 550 mg + metoclopramide 10 mg + ergotamine tartrate 1 mg + caffeine 100 mg), analgesics, intramuscular DHE or subcutaneous sumatriptan. Comparison of post-dexamethasone abortive treatment indicated oral or parenteral abortive therapies that were ineffective prior to dexamethasone became effective afterward. Study conducted before CONSORT and Jadad scale became applicable.
		Mean age (years)	35.5	43	38.5
		M:F ratio	1:8.3	1:4.6	1:5.7
		Complete resolution of headache, relapses, remissions (over one week)	80%, 29%, 57%	89%, 35%, 71%	89%, 28%, 83%
Side effects (GI discomfort, flushing, insomnia, sedation, akathasia)	19%	15%	22%
5. Monzillo et al. (70) Acute treatment of migraine in emergency room: Comparative study between dexametasone and haloperidol. Preliminary results [article in Portuguese]. Arq Neuropsiquiatr 2004. ED, Hospital, São Paolo, Brazil.	Dexamethasone (4 mg) IV versus haloperidol (5 mg) IV in ED. n = 29 Mean age = 31.5 years M:F ratio = 1:6.25 Patients were asked about pain intensity at 30, 60, 90 and 120 minutes after use of either drug. Recurrence defined as returning of headache after two hours and before 24 hours, once headache intensity has decreased from severe/moderate to mild/none. IHS criteria applied for diagnosis	Outcome: post-treatment percentage of patients with average VAS score		Dexamethasone 4 mg IV (n = 15)	Haloperidol 5 mg IV (n = 14)	Non-inferior outcomes for dexamethasone ED setting convenience sampling used Non-randomized design lacking blinding may introduce bias CONSORT checklist not mentioned; Jadad scale indicates poor score.
		Baseline		VAS 10 for all (100%)	VAS 10 for 100%
		30 minutes		86.7% (VAS 8.4)	64.3% (3.8)
		60 minutes		80% (6.6)	7% (4)
		90 minutes		65% (6.2)	0 (0)
Two hours		53.4% (6.6)	0 (0)
6. Krymchantowski et al. (71) Dexamethasone decreases migraine recurrence observed after treatment with a triptan combined with a nonsteroidal anti-inflammatory drug. Arq Neuropsiquiatr 2001. Private Headache Clinic. Rio de Janeiro, Brazil.	n = 400 mean age = 38 years M:F = 1:3.87 Sumatriptan/zolmitriptan/rizatriptan plus 4 mg dexamethasone oral (maximum of twice a week) to their usual triptans for six consecutive moderate/severe migraine attacks Recurrence defined as returning of headache after two hours and before 24 hours, once headache intensity has decreased from severe/moderate to mild/none. IHS criteria applied for diagnosis	Recurrence rate 23.4% (mean) in 18 patients 50% in two patients				Favorable outcomes for dexamethasone Study setting: private headache center: convenience sampling Side effects: heartburn in 30%, asthenia in 25%, somnolence in 15% For generalizability purpose, it was useful that study described including different abortive combinations of triptans and NSAIDs used by patients: rizatriptan 10 mg plus rofecoxib (n = 11); rizatriptan 10 mg plus tolfenamic acid (n = 4); zolmitriptan 2.5 mg plus rofecoxib (n = 3); zolmitriptan plus tolfenamic acid (n = 1); sumatriptan 100 mg plus tolfenamic acid (n = 1) Non-randomized design lacking blinding may introduce bias. CONSORT checklist not mentioned; Jadad scale indicates poor score.
7. Kelly et al. (72). Impact of oral dexamethasone versus placebo after ED treatment of migraine with phenothiazines on the rate of recurrent headache: A randomized controlled trial. BMJ 2007. ED, Hospital, Australia.	Double-blind, randomized, placebo-controlled, multi-centered Adult patients (n = 63), physician-diagnosed migraine; treated with IV phenothiazine M:F ratio = 1:1.29 Median age = 39.75 At discharge, patients were randomized to receive either 8 mg oral dexamethasone (n = 31) or placebo (n = 32) as a single dose. Telephone follow-up at 48–72 hours and proportion of patients with recurrent headache overall was recorded		Dexamethasone n = 31 (8 mg po)	Control n = 32		Favorable outcomes for dexamethasone Convenience ED sampling – followed by outpatient follow-up. ARR of 12% and 30% among those with headache lasting <24 hours is clinically significant. RR use provides valid measure as this was a prospective trial. Study could have included NNT computations, as this makes results of randomized trials more meaningful. NNT was (100/12=8.3) among all, and (100/30=3.33) among 40 patients with headache lasting <24 hours. Phenothiazines pre-treatment impact was not analyzed. Sample size not adequate. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Median age (years)	39	40.5
		Rate of recurrent headache at 48–72 hours	27% (8/31)	39% (12/32)
		Rate of recurrent headache for 40 patients with headache lasting <24 hours at 48–72 hours	15% (3/20)	45% (9/20)
Adverse effects	In dexamethasone group there was one report of facial flushing, one of nausea, two of new-onset transient tingling in hands or feet, one of blurred vision, one of a hot sensation in legs and one of diarrhea.	No adverse effects were reported in placebo group.
8. Baden E et al. (73) Intravenous dexamethasone to prevent the recurrence of benign headache after discharge from the emergency department: A randomized, double blind, placebo-controlled clinical trial. CJEM 2006. ED, Hospital, TX, USA.	Randomized, double-blind, placebo-controlled, two-arm n = 55 Mean age = 33.5 years M:F ratio = 1:1.75 Headache evaluation and therapy were determined by the treating physician, and, before discharge, patients were administered either 10 mg of IV dexamethasone or placebo Participating patients were contacted 48-72 hours following discharge and asked whether headache was - “better,” - “worse” or - “remained unchanged” when compared with symptoms at discharge. Headache recurrence defined as those with treatment failures whose headaches were “worse” or “unchanged,” and those who reported return of headache after being pain free at discharge. Recurrent headaches were further subclassified as - severe (i.e. provoking another physician visit or interfering with daily activity) - mild (i.e. requiring self-medication or no treatment).		Dexamethasone (10 mg IV) (n = 31)	Placebo (n = 24)		Favorable outcomes for dexamethasone ARR of 48.6% on headache recurrence 48–72 hours post-administration and 20.4% on severe headache. Convenience ED sampling used. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Mean age (years)	34.5	32.6
		M:F ratio	1:1.38	1:2.43
		Headache recurrence 48–72 hours post-administration	9.7% (3/31)	58.3% (14/24)
Severe headaches seeking repeated ED visits	12.9% (4/31)	33.3% (8/24)

9. Rowe et al. (74). Randomized controlled trial of intravenous dexamethasone to prevent relapse in acute migraine headache. Headache 2007. ED, Hospital, Canada.	All received standard IV abortive therapy at discretion of treating physician. n = 126 Mean age (years) = 35 M:F ratio = 1:4.26 Concealed allocation of patients was employed and patients were randomized to receive IV Dexamethasone (15 mg) or placebo in a double-blind design. Relapse was defined as return to ED, urgent clinic visit, or headache that precluded normal activity reported during follow-up telephone interviews 48–72 hours and seven days after ED discharge. Intention-to-treat analysis was conducted		Dexamethasone (IV, 15 mg) n = 64	Placebo n = 62		Favorable outcomes for dexamethasone Controlling for treatment assignment, relapse was more common when headache pain was incompletely relieved (VAS >2) at ED discharge (OR = 2.2; 95% CI: 1.1–5.4). ARR of 10 and 12 (in 72 hours and seven days) is clinically significant. OR is less valid to analyze effects in this study, as this was not a case-control study. Relapse was closely associated with incomplete pain relief at discharge. Convenience ED sampling; IHS criteria applied. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Age, mean (SD), years	35 (11)	34.6 (10)
		M:F ratio	1:4	1:4.55
		Relapses at 48–72 hours	14/64 (22%)	20/62 (32%)
Relapses by day 7	18/64 (28%)	25/62 (40%) (OR = 0.6; 95% CI: 0.3–1.3).
10. Jivad et al. (75) Is dexamethasone a suitable alternative for dihydroergotamine on migraine attacks? International Journal of Pharmacology 2005. ED, Hospital, Iran.	DHE 1 mg IV or dexamethasone 8 mg IV n = 72 M:F ratio = 1:2.6 Mean age = N/A Age range = 15-55 No differences in duration and frequency of headaches between two arms at baseline IHS criteria applied Outcome: Headache severity using McGill method on 1–15 scale method at baseline, 10 minutes, 30 minutes.	Mean headache score		DHE (n = 36) 1 mg IV	Dexamethasone (n = 36) 8 mg IV	Non-inferior outcomes for dexamethasone Convenience sampling ED setting 20 patients on DHE had side effects, i.e. chest tightness, hypertension, claudication, and vomiting. No patients on dexamethasone showed side effect. Lowering of headaches as early as 10 minutes among dexamethasone arm suggests another mechanism apart from anti-inflammation. Non-randomized design lacking blinding may introduce bias; CONSORT checklist not mentioned; Jadad scale indicates poor score.
		Baseline		12.9 ± 1.6	12.7 ± 1.9
		At 10 minutes		7.6 ± 2.6	8.6 ± 2.2
At 30 minutes		3.3 ± 1.9	4.3 ± 2.6
11. Taheraghdam et al. (76) Intravenous dexamethasone versus morphine in relieving of acute migraine headache. Pak J Biol Sci 2011. ED, Hospital, Iran.	Double-blinded, randomized, two-arm n = 190 mean age = 44.2 M:F ratio = 1:1.6 IHS criteria applied Dexamethasone 8 mg IV versus morphine 0.1 mg/kg IV Outcome: Headache severity using VAS at baseline, 10 minutes, 60 minutes, and 24 hours			Dexamethasone (8 mg IV) (n = 93)	Morphine (0.1 mg/kg IV) (n = 97)	Favorable outcomes for dexamethasone Convenience sampling ED setting In both groups headaches started to lower within 10 minutes. Difference between means of VAS baseline and VAS 10 minutes between two arms was not statistically significant (p = 0.236, 0.481). Means of VAS 60 minutes and VAS 24 hours were significantly lower in dexamethasone group than in morphine (p = 0.017, 0.010). CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
1.		Mean age		45.93	42.34
1.		M:F ratio		1:1.27	1:2.013
1.		VAS baseline		8.49 ± 1.5	8.75 ± 1.43
1.		VAS 10 minutes		5.6 ± 1.55	5.39 ± 2.1
1.		VAS 60 minutes		2.33 ± 1.73	2.89 ± 1.45
1.	VAS 24 hours		0.64 ± 0.71	1.03 ± 1.27
12. Fiesseler et al. (77) Steroids for migraine headaches: A randomized double-blind, two-armed, placebo-controlled trial. J Emerg Med 2011. ED, Hospital, NJ, USA.	Randomized double-blind, two-armed, placebo-controlled n = 181 mean age = 37 M:F ratio = 1:6.14 IHS criteria applied in 86% Dexamethasone (10 mg IV) if IV access was used or prednisone (40 mg by mouth× 2 days) if no IV access was obtained. Each medication was matched with identical-appearing placebo. Patients were contacted 24–72 hours after ED visit to assess headache recurrence.		Steroids (n = 94) Dexamethasone (n = 46), prednisolone (n = 48)	Placebo (n = 87)		Favorable outcomes for dexamethasone and prednisolone for recurrent headache Non-inferior results for acute attack management ED setting convenience sample used. Discharge VAS 0–100 while arrival 0–10. Study could have benefited from separately analyzing prednisolone and dexamethasone outcomes ARR of 10% on headache recurrence at 24–72 hours. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Age (years)	37 (SD = 10)	38 (SD = 10)
		Male	12 (7%)	13 (7%)
		Female	82 (45%)	74 (41%)
		M:F ratio	1:6.83	1:5.7
		Arrival VAS (mean 0–10)	8.9	8.9
		Discharge VAS (mean 0–109)	2.5	2.7
		Discharge VAS = 0	13	10
		Returned to ED	15	11
		Thought needed added treatment	33	36
		Mean patient satisfaction	4.9	4.7
		Patient felt needed further work-up	43	48
Recurrent headache at 24–72 hours	20/91 (22%) (95% CI 13.5–30.5)	26/82 (32%) (95% CI 21.9–42.1)
13. Donaldson et al. (78) IV dexamethasone vs placebo as adjunctive therapy to reduce the recurrence rate of acute migraine headaches: A multicenter, double-blinded, placebo-controlled randomized clinical trial. Am J Emerg Med 2008. ED, Hospital, MI, USA.	Multicenter, double-blinded, placebo-controlled randomized, two-arm n = 95 M:F ratio = 1:4.23 Mean age (years) = 36.33 IHS criteria applied placebo or 24 mg dexamethasone IV To ensure generalizability, all other aspects of patient care left to discretion of emergency physician. Baseline characteristics and adverse event profiles equivalent in both study groups.	Recurrence of migraine headache	Dexamethasone (n = 40) 24 mg IV	Placebo (n = 55)		Favorable outcomes for dexamethasone ED setting convenience sample used. ARR of 10% at three-day and 5% at 30-day follow-up. Small sample size. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Mean age	37.48	35.17
		M:F ratio	1:6.69	1: 2.7
		At three-day follow-up	35% (95% CI 24%-48%)	45% (95% CI 31%-60%)
At 30-day follow-up	19 (47.5%)	24 (42.9%)
14. Bigal et al. (16) A randomized double-blind study comparing rizatriptan, dexamethasone, and the combination of both in the acute treatment of menstrually related migraine. Headache 2008. Hospital, NJ, USA	Randomized, double-blind, six-attack crossover study n = 35 M:F ratio = Mean age = rizatriptan (RI, 10 mg po) vs dexamethasone (DE, 4 mg po); RI + DE vs RI; RI + DE vs DE 35 patients and 190 attacks IHS criteria applied		RI vs DE po	RI + DE vs RI po	RI + DE vs DE po	Favorable outcomes for dexamethasone ARR 19% for RI + DE vs RI ARR 48.2% for RI + DE vs DE Tolerability: More attacks treated with DE+RI (33.8%) were associated with side effects, compared to RI (18.6%) and DE (15.2%). AEs were mild, nonspecific, and typical of triptan sensations experienced in previous clinical trials of triptans (drowsiness, aches, tenderness in the neck, shoulders, and chest) and of steroids with DE (dyspepsia, etc.). No serious AEs reported. No special treatment required for AEs. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Primary endpoint: 24-hour sustained relief	62.7% vs 33.3%	81.5% vs 62.7%	81.5% vs 33.3%
Secondary endpoint: 24-hour sustained pain free	32.2% vs 12.1%	50.7% vs 32.2%	50.7% vs 32.2%
15. Rabe et al. (79) Prednisone for the treatment of withdrawal headache in patients with medication overuse headache: A randomized, double-blind, placebo-controlled study. Cephalalgia 2012. Headache Centers and Neurology Department, Germany.	Randomized, double-blind, placebo-controlled, parallel designed multicenter trial n = 96 M:F ratio = 1:5.5 Mean age = 41 years IHS criteria applied Prednisolone 100 mg or placebo for five days Data on 18 patients not complete because of patients not returning or not completing their headache diaries		Prednisolone/5 days 100 mg po (n = 48)	Placebo/5 days (n = 48)		Favorable outcomes for prednisolone. All patients reported decline in narcotic use. Benefit from each infusion lasted more than two weeks. All patients reported improved quality of life. Three patients did not respond. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts)
		Mean age	43.2	40
		M:F ratio	1:11.3	1:3.56
		Primary endpoint: Hours with severe or moderate headache within first three days	20.9	18.2
		Secondary endpoint: Hours with severe or moderate headache within five days	29.6	27.7
		Hours with severe or moderate headache within 14 days	74.9	69.8
		Rescue medication (intake frequency) within 14 days	3.6	6.4
		Rescue medication (intake frequency) within first five days	1.1	2.3
Further withdrawal symptoms (nausea, vomiting, photo-, phono-, osmophobia) within five days	39.2%	27.8%
16. Askari et al. (80) Pulse IV steroid and IV valproate combination therapy for sustained relief of chronic, treatment refractory headaches in patients with psychiatric disorders. Neurology 2010. Hospital, MO, USA	Patients who failed migraine therapy with or without narcotic use for >12 months were selected for treatment with IV methylprednisolone 500 mg and IV valproate 1 g. All patients also had bipolar disorder or depression. Medications were kept constant up to three months after infusions. Stat bolus of IV valproic acid, immediately followed by IV infusion of methylprednisolone 500 mg was administered over 60 minutes and repeated every three weeks.	Primary efficacy endpoints: 13 patients were treated. 10 patients showed >50% decline in severity and frequency (pain scale declined from 8–9/10 to 3–4/10). More than 60% decline was reported in the frequency of related ER visit. All patients reported decline in narcotic use. Benefit from each infusion lasted more than two weeks. All patients reported improved quality of life. Three patients did not respond.				Favorable outcomes for IV methylprednisolone Patients with chronic headaches having depression and/or bipolar disorder can show significant benefit from pulse therapy by IV methylprednisolone and IV valproic acid administered every three weeks. Improves patient’s quality of life, decreases ER visits and saves cost. Non-randomized design lacking blinding may introduce bias. CONSORT checklist not mentioned; Jadad scale indicates poor score.
17. Tfelt-Hansen et al. (81) Prednisolone reduces nitric oxide-induced migraine. Eur J Neurol 2009. Hospital, Denmark	Randomized, double-blind, and placebo-controlled, two-arm, crossover study N = 15 Mean age = 39.1 years M:F ratio = 1:14 15 migraineurs with migraine without aura were pre-treated with 150 mg of prednisolone or placebo followed by a 20-minute infusion of glyceryl trinitrate (GTN) (0.5 ug/kg/min). One hour after GTN infusion, participants were sent home, but continued to rate headache and possible associated symptoms by filling out a headache diary every hour for 12 hours. IHS criteria applied	Frequency of post-GTN migraine in first 90 minutes IHS criteria	Prednisolone (n = 15) 150 mg	Placebo (n = 15)		Favorable outcomes for prednisolone. All participants had been headache free for three days, and had to refrain from caffeine containing and alcoholic beverages, smoking as well as OTCs for 12 hours prior to study start. The patients were examined with a minimum of three-week intervals and at a fixed time of examination, 08.00 hours. One patient felt a slight, passing headache when bending forward after each intake of prednisolone, one patient felt slightly nauseated after intake of placebo. GTN infusion caused moderate hypotension in one and severe hypotension in another patient, reversed by raising the foot end of the beds. One patient had polypnoea during GTN infusion, and another felt tiredness after GTN infusion, both on placebo day. Prednisolone could exert its effect via inhibition of COX-2, CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained. Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts).
		Intensity (peak) of post-GTN headache in first 90 minutes	4/15	9/15
		Intensity (summed) of post-GTN headache in first 90 minutes	2 (0–5)	3 (0–7)
		Intensity (peak) of post-GTN headache in 12 hours	6 (0–17)	15 (0–55)
		Time to first migraine attack	2 (0–8)	4 (0–8)
		Intensity (summed) of post-GTN headache in 12 hours	6 hours (70 minutes–13 hours) mean	3.2 hours (10 miutesn–12 hours)
		Triptans as rescue medication	0 to 40 (mean of 7)	0 to 53 (18)
OTC as rescue medication	3	7
18. Innes et al. (82). Dexamethasone prevents relapse after emergency department treatment of acute migraine: a randomized clinical trial. CJEM 1999. ED, Canada.	Randomized, double-blind, two-centered; IHS criteria applied n = 98 Mean age = 35 years M:F ratio = 1:4.44 Dexamethasone (24 mg IV) or placebo. Patients recorded headache severity on VAS at time T = 0, T = 30 minutes and T = 60 minutes and at discharge. They were contacted 48–72 hours later and asked whether they had suffered a recurrence of their headache, categorized as class A (severe, provoking another physician visit), class B (severe, interfering with daily activity but not provoking a physician visit), class C (mild, requiring self-medication but not limiting activity) or class D (mild, requiring no treatment).		Dexamethasone (24 mg IV) n = 49	Placebo n = 49		Favorable outcomes for dexamethasone: ARR of 27% for severe headache. Lack of standardized abortive therapy could conceivably introduce a bias if one group received different treatment from the other. But randomization seemed to balance this. Abortive therapy included: antiemetics, IV nonsteroidal agents, DHE or opioids. Placebo recipients more likely to be female; other baseline characteristics similar between groups. OR use validity may not be appreciable as this was not a case-control study. Convenience sampling from ED setting may create unrecognized bias. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained. Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts).
		Mean age (years)	34	36
		M:F ratio	1:2.77	1:6
		Severe (classes A and B) recurrent headache 48 to 72 hours later	9 of 49 (18%)	22 of 49 (45%) (OR = 0.28; 95% CI: 0.11 to 0.69)
		class A	0	11
		class B	9	11
		class C	11	7
		class D	12	4
19. Bøe et al. (83) Prednisolone does not reduce withdrawal headache. A randomized, double-blind study. Neurology 2007. Hospital, Norway.	Randomized, double-blind, placebo controlled. study n = 97 M:F ratio = 1:2.85 Mean age (years) = 42.25 IHS criteria applied Placebo vs prednisolone 60 mg oral on days 1 and 2, 40 mg on days 3 and 4, and 20 mg on days 5 and 6 Sixty-five had migraine, 13 had tension-type headache, and 22 had both migraine and tension-type headache.		Prednisolone (n = 49)	Placebo (n = 48)		Non-inferior favorable outcomes for prednisolone. Mixture of tension headaches to migraine assessment Follow-up for six days. CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained. Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts).
		Age, years	43.2 (39.8–46.8)	41.3 (38.1–45.5)
		Total days of headache during last 30 days before withdrawal	25.0 (23.4–26.7)	25.7 (24.2–27.1)
		Headaches during first six days after withdrawal	Mean headache score 1.6 (1.4–1.8)	1.5 (1.3–1.7)
		Days with zero or mild headache	3.1 (2.5–3.6)	2.7 (2.1–3.2)
		Days with moderate or severe headache	2.9 (2.4–3.5)	3.4 (2.8–3.9)
Days without headache	1.1 (0.7–1.6)	1.1 (0.6–1.6)
20. Friedman et al. (84) Randomized trial of IV dexamethasone for acute migraine in the emergency department. Neurology 2007. ED, Hospital, NY, USA.	Randomized, double-blind, placebo-controlled multicenter, two-arm, dexamethasone 10 mg IV or placebo n = 205 Mean age = 36.5 years M:F ratio = 1:5.61 IHS criteria applied All patients (n = 205) received 20 mg IV metoclopramide and 2.5 mg diphenhydramine (both mixed in 50 ml normal saline, run-in 20 minutes) following dexamethasone/placebo. Patients were asked to characterize migraine pain as “none, mild, moderate or severe.” Scale was administered at baseline and every 30 minutes until two hours had elapsed. This scale was then administered by telephone 24 hours after medication administration. Functional impairment scale: patients described their disability as 1) none; 2) mildly impaired (having a little bit of difficulty doing what I usually do); 3) moderately impaired (having a great deal of difficulty doing what I usually do and can only do very minor activities); or 4) severely impaired (requiring bed rest). Two patients from dexamethasone group did not complete study. ‘Persistent pain free’ required participants to attain a pain level of ‘none’ on this categorical scale before two hours had elapsed in the ED and to maintain level of ‘none’ throughout post-ED discharge period.		Dexamethasone 10 mg IV (n = 106)	Placebo (n = 99)		Favorable outcomes for dexamethasone Side effect profiles were similar, with exception of acute medication reactions, which occurred more commonly in dexamethasone group. ARR 6% in headache recurrence reduction 24 hours ARR 25% in headache recurrence reduction 72 hours Study results generalizability is limited because all received metoclopramide and diphenhydramine. Convenience ED sampling CONSORT: enrollment, randomization, blinding, allocation, loss to follow-up, and analysis methods clearly explained. Jadad scale: 5 – rigorous (2 for randomization, 2 for blinding, 1 for description of withdrawals and dropouts).
1.		Mean age (years)	36	37
1.		M:F ratio	1:4.58	1:7.25
1.		Persistent pain-free outcome of 24 hours (primary outcome)	25%	19%
1.		No functional impairment after ED discharge (24 hours)	67%	59%
1.	Persistent pain free among migraine >72 hours	38%	13%
21. Krymchantowski et al. (85) Out-patient detoxification in chronic migraine: Comparison of strategies. Cephalalgia 2003. Outpatient, Hospital, Rio de Janeiro, Brazil	Prospective randomized open-label trial; IHS criteria applied 150 chronic migraine patients (125 women, 25 men; ages 18–80 years, mean 40.3 ± 13.8) with overuse of symptomatic medication divided into three study arms n = 50 each Mean age = 40.3 years M:F = 1:5 In each group, 50 patients received education and orientation and were then abruptly withdrawn from all symptomatic medication. All patients had similar profiles of headache characteristics and consumption (quality and quantity) of symptomatic medication before initiation of treatment, but most were not severe headache sufferers, heavy medication overusers or were overusing opioids. Day 7 onward: For all patients: Six initial days: (atenolol 10 mg + nortriptyline 4 mg + flunarizine 1 mg) capsules. Six following days: (atenolol 20 mg + nortriptyline 8 mg + flunarizine 2 mg) capsules. Four following weeks: (atenolol 30 mg + nortriptyline 16 mg + flunarizine 2.5 mg) capsules. After five weeks headache frequency and intensity, prevalence and frequency of withdrawal symptoms and consumption of rescue medications during first six days were compared between groups. Adherence to treatment (who returned or not and for which reasons, between groups) and headache frequency, week by week, among the groups of patients were also compared.		Prednisone PO (60 mg/ day two days, 40 mg/day two days and 20 mg/day two days) for first six days (n = 50)	No regular medications to take for first six days (n = 50)	Naratriptan (2.5 mg twice a day) for first six days (n = 50)	Non-inferior outcome for prednisolone Comparable results of headache frequency reduction among three arms during five weeks’ post-intervention Open-label nature of study may introduce bias Showed efficacy of steroids in a non-ED setting i.e. outpatient setting; this helps save costs and time associated with ED-setting therapy or inpatient hospitalization. No major side effects CONSORT: enrollment, randomization, allocation, loss to follow-up, and analysis methods clearly explained; blinding not performed. Jadad scale: 3 (2 for randomization, 0 for blinding, 1 for description of withdrawals and dropouts).
1.		Baseline percentage of seven-day headache	27 (54%)	23 (46%)	25 (50%)
1.		Adherence to treatment after five weeks	44 (88%)	41 (82%)	35 (70%)
1.		Baseline headache frequency seven days per week for last four weeks	27 (54%)	23 (46%)	25 (50%)
1.		Baseline headache intensity: severe headache	13 (26%)	12 (24%)	21 (42%)
1.		Total number of headache days during the first 6 days	264	246	210
1.		Absence of headaches during the first 6 days (refers to the total number of days considering the total number of six-day periods among all patients who returned in all three groups.	69 (26.2%)	62 (25.2%)	62 (29.5%)
1.		Number of patients with withdrawal symptoms during first six days	36 (81.8%)	40 (97.5%)	24 (68.6%)
1.		Headache frequency reduction of >50%	From week 3 onward	From week 2 onward	From week 3 onward
1.	Headache frequency reduction of >50% after five weeks	66% 63.4%		57.1%
22.Krymchantowski et al. (86). Prednisolone as initial treatment of analgesic-induced daily headache. Cephalalgia 2000. Tertiary headache center, Rio de Janeiro, Brazil	Prospective, single-arm study using six-day course of oral prednisone taper, comparing mean headache frequency at baseline with mean frequency after 30-day period n = 400; 318 women and 82 men; M:F ratio = 1:3.88; mean age 38 years (19–64); sequential selection all presented with headache for >28 days per month for longer than six months mean baseline headache frequency = 0.96 (calculated considering number of headache days divided per 30-day period of each patient and taking into account mean number for all patients) during previous two months mean headache intensity = 1.9 (using analog scale of 0=none, 1=mild, 2=moderate, 3=severe) average daily headache duration was four to eight hours during previous two months Exclusion criteria: barbiturate, benzodiazepine, opioid habituation; cardiovascular, cerebrovascular, hematopoietic diseases, active peptic disease, basilar migraine, medical and psychiatric comorbidities According to patient’s own judgment: 319 (79.7%) patients had frequent presence of anxiety and/or depression; 281 (70.2%) reported sleep disturbances Method of intervention: Instructed to suddenly stop overused medications. Stop gradually over period of seven days preventive medications that were being taken. Take only prednisone and ranitidine for first six days and preventive medications afterward. Dosage of prednisone, oral: 60 mg/day for two days, 40 mg/day for two days, 20 mg/day for two days plus ranitidine 300 mg /day		Baseline (two months previous to study)		30 days after treatment	Favorable outcome for prednisolone Showed efficacy of steroids in a non-ED setting i.e. outpatient setting, helps save costs and time associated to ED-setting therapy or inpatient hospitalization. Not randomized, not blinded No major side effects Inadequate description to assess using CONSORT checklist and Jadad scale.
		24 headache days/30 days	–		176 (44%)
		25 headache days/30 days	–		121 (30.2%)
		26 headache days/30 days	–		56 (14%)
		27 headache days/30 days	–		26 (6.5%)
		28 headache days/30 days	169 (42.2%)		21 (5.2%)
		29 headache days/30 days	132 (33%)		–
30 headache days / 30 days	99 (24.8%)		–
23. Kaniecki (87). Management of status migrainosus with dexamethasone—A prospective comparison versus methylergonovine and symptomatic care. Headache 2006. University Hospital, Headache Center, Pittsburgh, PA, USA	Sample size: total n = 1654, 2758 instances of status migrainosus IHS criteria employed for all patients with migraine attacks extending beyond 72 hours. Only patients using triptans as abortive were included. Patients with headaches lasting more than seven days were excluded. All patients instructed to use symptomatic agents etodolac for pain (400 mg tid prn) and metoclopramide for nausea (10 mg tid prn). Random and prospective assignment of patients to one of three treatment arms: - three days of dexamethasone (4 mg tid, bid, qd each for one day) - three days of methylergonovine (2 mg tid, bid, qd each for one day) - no additional care. Response defined as complete resolution of migraine pain within three-day treatment phase (primary endpoint).		Response on intent-to-treat analysis		Patients with AEs necessitating withdrawal from study	Favorable outcome for dexamethasone Superior results showing efficacy of three-day dexamethasone course to be superior than ergots or symptomatic nonsteroidal care (ARR = 9.3, 19.3). Adequate randomization; blinding information not available; attrition rate not fully explained. CONSORT: enrollment, randomization, allocation, and analysis methods clearly explained. Jadad scale: 2 (2 for randomization, 0 for blinding, 1 for description of withdrawals and dropouts).
		Three days of oral dexamethasone (n = 942, instances)	90.3% (n = 851)		37
		Three days of oral methylergonovine (n = 900)	81.6% (n = 734)		58
No additional care (n = 916)	70.6% (n = 696)		14
24. Kopjas et al. (88). Migraine treatment with dexamethasone. Headache 1967. Outpatient, IL, USA	112 selected patients (81 females and 31 males, M:F ratio of 1:2.61, average age of 37 years). Contraindication to steroid administration ruled out. Each patient received IV 4 mg of dexamethasone twice a week for first two weeks, then once a week for two weeks and then once a month for two months. All patients had acute headaches with history of an average of five years migraine. Study conducted prior to development of IHS criteria.	In around 90% of patients, severity of migraine headache subsided in an average of two hours after slow IV dexamethasone injection. Average headache-free periods after first injection lasted 60 hours. Second injection increased pain-free period to an average of three and one-half days. Second series of IV injections once a week showed very satisfactory results in more than 60% of patients. Usual daily attacks not noticed for several days, and the experienced headaches were mild and much shorter in duration. Monthly injections for the two-month period were mostly beneficial in female patients who were still noticing some headaches around their menstrual periods. Around 35% of patients returned approximately once every three months for an intravenous injection, stating that after the therapy headaches are relieved from six to eight weeks and the intensity is lessened.			Favorable outcome for IV dexamethasone. Study showed efficacy of IV dexamethasone administered with decreasing frequency of twice a week for first two weeks, then once a week for two weeks and then once a month for two months. Selection method unclear. Jadad and CONSORT not applicable.
25. Frohner (58). Cortisone in migraine and histamine headaches. Rocky Mt Med J 1953. MT, USA	20 patients with migraine included in study over period of two years. 50 mg of oral cortisone given when headache was clearly established Study design not clear. Objective outcome measurands not clearly described. Study conducted prior to development of IHS criteria.	Number of patients (%)	Outcome		Favorable outcome for oral cortisone. Selection criteria not clear. Jadad and CONSORT checklist not applicable.
		9 (45%)	Complete dramatic relief two hours post-administration All patients in this cohort had migraine by most rigid study criteria
		3 (15%)	Partial relief with headache recurrence
8 (40%)	No benefit This cohort was suspected to have probable tension headache on closer, more rigid re-examination

First author, title of article, name of journal, year of publication, and study setting are described in the first column. The second column provides important methods employed for study design. Outcomes are fully elaborated in the third column. Critical appraisals, clinical bottom lines, Consolidated Standards of Reporting Trials (CONSORT) and Jadad scale are presented in the last (fourth) column. M: male; F: female; IHS: International Headache Society; tid: three times per day; prn: as needed; IV: intravenously; ED: emergency department; VAS: visual analog scale; CI: confidence interval; OTC: over-the-counter; po: orally; ER: emergency room; GI: gastrointestinal; ARR: absolute risk reductions; NSAIDs: nonsteroidal anti-inflammatory drugs; RR: relative risk; OR: odds ratio; DHE: dihydroergotamine; AEs: adverse events.

To help interpret the results of the different studies included, the following methods were applied. For the primary outcome, i.e. frequency of headache, relative risk (RR) ratios were calculated from the numbers of headache events in the control and intervention groups of each study. Where headache recurrence was reported as a dichotomous variable (present/absent), the presence of headache was included as an event for the primary outcome. For adverse events, the number of events in each study group was scored. Absolute risk reductions (ARRs) were calculated by subtracting event rate (headache frequency) in the intervention arm from event rate in the control (or study arm not of primary interest). Numbers needed to treat (NNT) were calculated by dividing 100% by the ARR values. All published studies relevant to our question stem were included and their outcomes were completely assessed so as to avoid publication bias and selective reporting.

Results

A total of 25 studies with an aggregate number of 3989 (median age of 37.5 years, interquartile range or IQR 35–41 years; median male:female ratio of 1:4.23, IQR 1:2.1–6.14) (Figure 7(a) and (b)) patients were found under the combined search methodologies described within the Method and Materials section (Table 1). Four systematic reviews and meta-analyses were found under the Systematic Reviews search (Table 2). Two clinical studies were found from a gray literature search on conference abstracts. Fourteen (56%) studies were designed as randomized, double-blinded, controlled trials. Thirteen (52%) studies were from ED-based settings, while the remaining 12 (48%) were from non-ED-based outpatient settings. Sixteen (64%) studies applied International Headache Society (IHS) headache diagnostic criteria.

Figure 2.

Methodological quality of included studies as assessed using the Jadad score. The Jadad score measures adequacy of reported randomization, blinding, and management of withdrawals and dropouts. Fourteen studies (56%) had a Jadad score (63) of 5, and fulfilled the complete CONSORT (64) checklist—indicating rigorous methodological quality; two studies (8%) had a Jadad score 3, and missed blinding from the CONSORT checklist—indicating adequate methodological quality; four studies (16%) had a Jadad score of 0 and did not fulfill any of the CONSORT checklist.

Table 2.

Comprehensive summary of included systematic reviews and meta-analysis.

First author, title of article, name of journal, year of publication	Systematic review, meta-analysis results	Clinical bottom lines
Neill A et al. (89) Towards evidence-based emergency medicine: Best BETs from the Manchester Royal Infirmary. BET 2: Dexamethasone for reduction of migraine recurrence. Emerg Med J 2013.	Two systematic reviews with meta-analysis of seven randomized controlled trials each; total of 738 and 742 patients each; outcome: recurrence of headache at 24 hours; key results: recurrent headache reduction RR of 0.74 (95% CI 0.60–0.90) and 0.87 (95% CI 0.80–0.95).	Patients who have received successful abortive treatment for migraine in the ED should be considered for a single dose of IV dexamethasone before discharge to reduce risk of recurrence, in absence of any of usual relative contraindications to steroid therapy.
Colman et al. (5). Parenteral dexamethasone for acute severe migraine headache: Meta-analysis of randomized controlled trials for preventing recurrence. BMJ 2008.	Seven randomized controlled double-blinded studies. All trials used standard abortive therapy and subsequently compared single-dose parenteral dexamethasone with placebo, examining pain relief and recurrence of headache within 72 hours. Dexamethasone and placebo provided similar acute pain reduction (weighted mean difference 0.37, 95% CI 0.20–0.94). Dexamethasone was, however, more effective than placebo in reducing recurrence rates (fixed-effects weighted RR of 0.74, 95% confidence interval 0.60 to 0.90). Similar side effect profiles between dexamethasone and placebo groups.	When added to standard abortive therapy for migraine headache, single-dose parenteral dexamethasone is associated with 26% relative reduction in headache recurrence (number needed to treat=9) within 72 hours.
Huang et al. (90). Steroids for preventing recurrence of acute severe migraine headaches: A meta-analysis. Eur J Neurol 2013	Eight studies with 905 patients were included. Pooled analysis showed that when steroids were added to standard abortive therapy they reduced rate of moderate or severe headache recurrence after 24–72 hours of follow-up evaluation (RR = 0.71; 95% CI = 0.59–0.86). No significant benefit of steroids compared with placebo in proportion of totally resolved migraines (RR = 1.11; 95% CI = 0.94–1.32). Side effects of steroids are mild and not significant except for dizziness. Subgroup meta-analysis showed parenteral dexamethasone tends to be more effective in reducing moderate or severe recurrent headaches (RR = 0.68; 95% CI = 0.55–0.84). However, no significant differences found between oral administration and parenteral administration of steroids (p = 0.37).	When steroids are added to standard abortive therapy for migraine headaches, they are effective and safe for preventing moderate or severe headache recurrence.
Singh et al. (91). Does the addition of dexamethasone to standard therapy for acute migraine headache decrease the incidence of recurrent headache for patients treated in the emergency department? A meta-analysis and systematic review of the literature. Acad Emerg Med 2008.	Seven ED-based randomized, controlled trials involving 742 patients. Pooled results suggest a modest but significant benefit when dexamethasone is added to standard antimigraine therapy to reduce rate of patients with moderate or severe headache on 24- to 72-hour follow-up evaluation (RR = 0.87, 95% CI = 0.80–0.95; absolute risk reduction = 9.7%). The treatment of 1000 patients with acute migraine headache using dexamethasone in addition to standard antimigraine therapy would be expected to prevent 97 patients from experiencing outcome of moderate or severe headache at 24–72 hours after ED evaluation. Sensitivity analysis yielded similar results with sequential trial elimination, indicating that no single trial was responsible for overall result. Adverse effects related to administration of a single dose of dexamethasone were infrequent, mild and transient.	Dexamethasone is efficacious in preventing headache recurrence and safe when added to standard treatment for management of acute migraine headache in the ED.

Table 2 displays included systematic reviews and meta-analysis (first author, title of article, name of journal, and year of publication in the first column), important results of clinical relevance combined from the statistical analysis (second column), and clinical bottom lines (third column). The systematic reviews by Colman et al. and Singh et al. included the same seven randomized-controlled trials (RCTs); the former further included a meta-analysis of adverse events, while the latter provided symmetric inverted funnel plots to describe the lack of systematic difference and publication bias between the larger and the smaller study effect sizes. The systematic review by Huang et al. included one more RCT than the seven; this study added a meta-analysis elucidating the superiority of adjunctive parenteral dexamethasone compared to placebo with standard abortive therapy. The review by Neill and Brannigan further discussed the weaknesses of three selected RCTs. RR: relative risk; CI: confidence interval; ED: emergency department; IV: intravenously.

Fourteen studies (56%) had a Jadad score (65) of 5, and fulfilled the complete CONSORT (66) checklist—indicating rigorous methodological quality; two studies (8%) had a Jadad score 3, and missed blinding from the CONSORT checklist—indicating adequate methodological quality; four studies (16%) had a Jadad score of 0 and did not fulfill any of the CONSORT checklist (Figure 2).

Figure 3.

(a) Median headache reduction (shown by horizontal dotted lines) was found to range from 30% in 24-hour to 11% in 72-hour respectively following administration of corticosteroids. Error bars indicate interquartile range (IQR) values; two studies each from both 24-hour and 72-hour comparison were found to be outliers to the IQR. (b) Similar to the results displayed on Figure 3a, the Numbers Needed to Treat (NNT) was found to be 3 for 24-hour and 10 for 72-hour headache recurrence.

Nineteen studies (76%) indicated observed outcome differences favoring benefits of corticosteroids, while six (24%) studies indicated non-inferior outcomes for corticosteroids. Stratifying the studies to those that addressed headache recurrence (56%) and acute reduction of migraine attacks (68%), respectively, revealed that 78.6% and 61.2% of the studies observed favorable outcome differences indicating benefits of corticosteroids; 21.4% and 38.8% indicated non-inferior outcomes for corticosteroids (Figure 4). Median absolute risk reduction was 30% (range 6%–48.2%), and 11% (6%–48.6%) for 24-, and 72-hour headache recurrence (Figure 3(a)), making the NNT 3.3, and 9.1 for 24-, and 72-hour headache recurrence (Figure 3(b)), respectively. Parenteral dexamethasone was the most commonly used (14 studies, 56%) corticosteroid. Dexamethasone was administered with an average single dose of 12.8 mg (range 4–24 mg; interquartile range 8–16 mg; median 10; mode 8, 10 mg) (Figure 5). All the ED-based studies used parenteral dexamethasone, and showed favorable observed outcome efficacy of dexamethasone. One study used 500 mg intravenous methylprednisolone. Ten of the 13 non-ED-based studies administered oral corticosteroids, nine of which showed favorable observed outcome efficacy. Two studies administered 4 mg oral dexamethasone; one study used 8 mg oral dexamethasone; another study used a five-day course of 100 mg oral prednisolone; three studies used a six-day tapering course of oral prednisolone 60 mg on days 1 and 2, 40 mg on days 3 and 4, 20 mg on days 5 and 6; one study administered pre-treatment with 150 mg oral prednisolone; one study administered a three-day course of oral dexamethasone (4 mg three times per day (tid), twice per day (bid), once per day (qd) each for 1 day); and one study administered a single oral dose of 50 mg cortisone (Figure 6). Parenteral dexamethasone (56%), oral prednisolone (12%), and oral dexamethasone (8%) were the top three common steroid forms administered, in descending order (Figure 6).

Figure 4.

Nineteen studies (76%) indicated observed outcome differences favoring benefits of corticosteroids, while six (24%) studies indicated non-inferior outcomes for corticosteroids. Stratifying the studies to those that addressed headache recurrence (56%) and acute reduction of migraine attacks (68%), respectively, revealed that 78.6% and 61.2% of the studies observed favorable outcome differences indicating benefits of corticosteroids; 21.4% and 38.8% indicated non-inferior outcomes for corticosteroids.

Figure 5.

Dexamethasone was administered with an average single dose of 12.8 mg (range 4–24 mg; interquartile range 8–16 mg, shown by error bars; median 10, shown by the dotted horizontal line; mode 8, 10 mg).

Figure 6.

Parenteral dexamethasone was the most commonly used (14 studies, 56%) corticosteroid. Oral prednisolone (3 studies, 12%) and oral dexamethasone (2 studies, 8%) were the other two common corticosteroid forms administered.

Figure 7.

(a) Median (37.5 years) and interquartile range (35–41 years) of the mean age of all participants (both sexes) within the included studies. (b) Median (1:4.23) and interquartile range (1:2.1–6.14) of Male:Female ratio among all participants within the studies included.

In the systematic reviews and meta-analysis appraised in this study, the objectives and inclusion criteria were clear among all. Relevant sources were searched without restriction by language or publication status. Steps were taken to minimize bias and error by having more than one reviewer independently conduct study selection, validity assessment and data extraction. Appropriate statistical methods were used to combine studies and to assess for heterogeneity and publication bias. Potential biases and sources of confounding were well addressed in the text. Although the trials were small, they were of acceptable quality and their results were consistent. All systematic reviews were well conducted and the authors’ conclusions appear reliable. All systematic reviews and meta-analyses revealed the importance of adjuvant corticosteroids to various abortive medications—indicating generalizability of results. Three meta-analyses showed that a single dose of adjuvant parenteral dexamethasone was associated with 11.1% ARR (absolute risk reduction), NNT of 9, and 26% RRR for 72-hour headache recurrence. The fourth meta-analysis indicated that parenteral dexamethasone delivered the highest efficacy. Adverse effects were tolerable.

Cumulative evidence from our systematic review shows that higher disability, prolonged migraine duration, status migrainosus, incomplete pain relief, and headache recurrence were the settings in which corticosteroids were beneficial. One study indicated that abortive medications that were ineffective prior to dexamethasone became effective subsequently (61). Among the ED-based studies, there was an increased chance of headache recurrence among patients with higher levels of pain at ED discharge than among patients with lower levels of pain at ED discharge; this may indicate that migraine attacks treated when the pain is mild are less likely to recur. One randomized double-blind study showed that adjunctive dexamethasone resulted in higher sustained relief periods as compared to triptan monotherapy (16).

Discussion

Migraine recurrence is commonly observed following “successful” ED management. Various clinical features have been associated with poor pain outcomes after treatment for headache. In the outpatient setting, a patient’s history of headache-related functional disability can be used to predict long-term medication needs prospectively and to decrease the economic burden of illness. Some ED patients cannot access neurology or headache specialty care. Timely primary care appointments can also be difficult to obtain, and ED patients frequently fail to appear for scheduled appointments. Therefore, it is important for the emergency physician to understand the magnitude of post-discharge headache and to identify which patients are most at risk. Having this information can be used to guide choice of medication, discharge instructions, and urgency of specialty referral.

During this study, we did not perform a sub-analysis of outcomes depending on the type of abortive treatment used because, in the real world, different drug combinations are often required and because all patients do not respond to a standard regimen. Lack of standardized abortive therapy could conceivably introduce a bias if one group received different treatment from another. Moreover, patients often know what agents work best for them, and physicians often have treatment preferences. We felt that mandating specific abortive therapies would make the study more complex, and would not emulate clinical heterogeneity in practice. Given the inter-study differences in design, in methodology, and in outcome and efficacy measurands, we were not able to perform statistical heterogeneity and weighted meta-analysis tests.

The fact that dexamethasone recipients suffered fewer severe recurrences and had fewer repeat physician visits (73,74,80,82) suggests that there are quality-of-life and productivity benefits to this therapy; performing formal economic analysis and demonstrating these benefits will further support this notion. While being efficacious in both situations of reducing headache recurrence and acute headache attacks, corticosteroid administration showed higher potency in the former, i.e. controlling headache recurrence (Figure 4).

Parenteral administration within the ED setting using single-dose intravenous dexamethasone at 10 mg shows adequate efficacy in controlling attacks, and transferring care to the outpatient. Parenteral administration can be useful to avoid poor drug absorption secondary to gastric stasis commonly experienced by chronic migraineurs. In the outpatient setting, six to eight short, oral, tapering courses of steroids per year with close follow-up of adverse effects shows efficacy; it is essential to note that corticosteroid administration is intended to control acute, severe attacks in an effort to smoothe transfer of care to better acute care and/or preventive management targets (92).

Being a long-acting corticosteroid with a biologic half-life of 36–72 hours (93,94), dexamethasone should effectively suppress inflammation during the period when patients are most likely to experience a headache recurrence—potentially making it an ideal agent for a one-time administration before ED discharge. At equipotent anti-inflammatory doses, dexamethasone almost completely lacks the sodium-retaining property of hydrocortisone, minimizing unwanted side effects. In our review the incidence of minor effects was rare and was similar to placebo, with no reported severe adverse reactions. That outcomes compare favorably both in the ED- and non-ED settings indicates that convenience sampling is not a likely source of confounding bias; this, in fact, further promotes the generalizability of benefits of steroids for relief of recurrent headache in both outpatient and ED settings.

It is worthwhile to differentiate migraine sufferers presenting to the ED with “the last straw headache syndrome” of consecutive bouts of headaches from those presenting with “the first or worst headache syndrome” alarming the patient and causing the visit to the ED. The former comprise the ED-frequenter migraine sufferers who have been described as sharing certain salient features of emotional decompensation, compulsive preoccupation with symptoms, and occasional manipulative behavior, besides their head pain (19). These “refractory” patients might be victims of misdiagnosis, mistreatment, undertreatment, or overmedication, while the latter present as a more challenging subset of new headache or genuine refractoriness that is difficult to manage with readily available medications. Addressing limitations of current therapy is important while monitoring refractory headaches. Reviewing treatment plans at regular visits and balancing short-term pain relief with longer-term goals of disease management needs to be emphasized. Management of expectations is an increasingly important tool in handling patients with chronic pain in both the ED and outpatient settings.

In patients presenting with frequent migraine attacks, careful selection needs to be made as to which attack to treat aggressively; this may involve less convenient but more effective non-oral medications at the cost of more side effect burden. Always observe for inadequate trial duration, inappropriate dosage and consideration of medication overuse when patients are not doing well. Using a headache calendar or diary to log headache types and frequency, treatment effectiveness, menstrual cycle, ED visits, and other health care visits are essential to the care of migraineurs, especially difficult-to-manage migraineurs.

Dexamethasone causes less fluid retention than other steroids due to the fact that it has less mineralocorticoid effect. Because these headaches often recur over the days and months following ED discharge, the responsibility of the emergency physician includes identifying as yet unmet treatment needs and ensuring successful transition of care of these patients to an appropriate outpatient health care provider. Prior to administering dexamethasone, it is important to remember the various contraindications, i.e. pregnancy, known peptic ulcer disease, diabetes, or systemic fungal infections. Clinicians deciding which medication to prescribe for recurrence can therefore be guided by considerations such as cost, contraindications, adverse effects, and a patient’s previous overall experience with the medication. Careful evaluation and management for opioid administration, barbiturate use, and stressful life events are salient features characteristic of multidisciplinary approaches toward a long-term control of prolonged and refractory migraine. Where resources are available, it is important to use existing opportunities to optimize emergency care of migraine attacks by real-time consultations between the emergency and headache specialty departments.

Steroids have been efficacious for detoxification of overused substances in patients with medication-overuse headache (MOH) (79,95). Concurrent administration of nonsteroidal anti-inflammatory drugs (NSAIDs) and steroids must be avoided to reduce potential risk of gastrointestinal bleeding. Comparison between titration versus bolus dosage of dexamethasone administration can affect headache recurrence and other outcome measures. Some of the studies reviewed herein defined headache recurrence as 24 hours after headache relief from the study drug, while others used a 48-hour or longer timepoint. Panel A (available online) provides commonly applied definitions of recurrence and efficacy measures; these terms are inter-related and their consistent use can help enhance daily practice and research methods.

Conducting mega-trials and mega-studies has its own challenges of operational limitations in organizing resources for longer durations, recruiting larger number of study patients, and securing funding. For these reasons, performing systematic reviews and pooled analysis provide perceivable options to assess accumulating evidence and aggregate data from different studies addressing similar question. Locating studies from literature database searches was the most important component while preparing this systematic review. One of the limitations we faced was the inconsistent availability of variables, and outcome and efficacy measurands across the different included studies. Future studies can benefit by providing consistently available data so that weighted meta-analysis, heterogeneity statistics, and meta-regressions can accurately be computed, and trends and interactions can easily be detected. Forthcoming studies can elucidate further on the efficacy differences between different dosages of oral and parenteral corticosteroid administration, and mixture parenteral infusions. Our recommendations include conducting full clinical studies expanding to other migraine patient populations apart from prolonged migraine attack cohorts and emergency settings.

Broader global utility

In lower-resource emergency settings or on an outpatient basis, where the burden of migraine is increasing and where triptans and parenteral infusion forms of emergency abortive headache medications (e.g. intravenous magnesium, dihydroergotamine, levetiracetam, and valproic acid) (1,96,97) are unavailable, corticosteroids provide efficacious and low-cost therapeutic options to control prolonged and highly disabling migraine attacks. Corticosteroids are readily and inexpensively accessible both in oral and parenteral forms in lower-resource clinical settings; hence, our study can be help optimize their application in a clinically relevant manner.

Conclusions

Our literature review suggests that with corticosteroid treatment for acute migraine attacks, recurrent headaches decrease, become milder than pretreated headaches and then may respond to nonsteroidal therapy. Intravenous dexamethasone provides a reasonable option for managing resistant, severe, recurrent or prolonged migraine attacks in the ED. Our recommendations are consistent with the literature review and include up to six to eight administrations of corticosteroids per year with follow-up of adverse effects. Although documented by fewer studies, we consider it is reasonable to use oral dexamethasone on an outpatient basis in similar settings. It can be beneficial to provide a one-time dexamethasone administration before ED discharge to a headache specialist for long-term management. Reasonable and achievable goals of pain reduction should be made clear to the presenting patient in the ED setting or in the office during the initial visit.

Clinical implications

Studies indicate that post-emergency department (ED) recurrent headaches occur within 24 hours in up to 87% of migraine patients visiting the ED.

With corticosteroid treatment, recurrent headaches become milder than pretreated headaches and later respond to nonsteroidal therapy.

Corticosteroids provide a reasonable option for managing resistant, severe, prolonged or recurring migraine attacks.