← Back
NeuroTrials.ai
Neurology Clinical Trial Database

PED-DEX

Randomized controlled trial of intravenous dexamethasone to prevent relapse in the treatment of migraine in a pediatric emergency department

Year of Publication: 2026

Authors: Tourigny-Ruel G, Bailey B, Jean-Charles S, Gravel J

Journal: Headache: The Journal of Head and Face Pain

Citation: Headache. 2026;66(6):1235-1243. doi:10.1111/head.70087

Link: https://doi.org/10.1111/head.70087

Bottom Line

In this small single-center RCT, adjunctive IV dexamethasone did not reduce 48-h migraine relapse or improve pain, functional recovery, or health-care use in children and adolescents.

Major Points

  • Primary outcome (48-h relapse) not different: 39% dexamethasone vs 44% placebo; risk difference −4% (95% CI −32% to 24%).
  • All secondary outcomes (VAS/VNS pain at 48 h and 7 days, return to school, return to normal activity, physician revisit) similar between arms.
  • Best-case/worst-case sensitivity analyses spanned from a 28% absolute benefit to a 22% absolute harm, reflecting the substantial 25% loss to follow-up.
  • Adverse events were infrequent and mild in both arms; no serious adverse events.
  • Trial was substantially underpowered — target 58/arm, but only 87 of 116 randomized provided 48-h outcome data after 12 years of recruitment.

Design

Study Type: Randomized Controlled Trial

Randomization: 1

Blinding: Double-blind (patients, treating physicians, research nurses, outcome assessors)

Enrollment Period: July 8, 2013 – February 24, 2025

Follow-up Duration: 7 days (primary outcome at 48 h)

Centers: 1

Countries: Canada

Sample Size: 116

Analysis: Modified intention-to-treat (complete-case) with pre-specified best/worst-case sensitivity analyses


Inclusion Criteria

  • Age 8–17 years
  • Presentation to the pediatric ED (CHU Sainte-Justine, Montréal) during research-nurse hours
  • Acute migraine attack meeting Irma ED criteria (headache 1–72 h with ≥4 of: moderate-severe activity impairment, focal localization, pulsatile quality, associated GI symptoms, photophobia or phonophobia, worsening with activity or relief with rest)
  • Required IV rescue therapy (metoclopramide + diphenhydramine) per treating physician
  • Improvement after IV abortive medication
  • Written informed consent from a parent and assent from the patient

Exclusion Criteria

  • Known allergy to any study drug or component
  • Absolute contraindication to oral corticosteroids (active untreated infections, systemic fungal infections, cerebral malaria, respiratory tuberculosis, hypertension, heart failure, renal or hepatic impairment, GI disease, myasthenia gravis, diabetes, cataracts, glaucoma, seizure disorder, thyroid dysfunction, thromboembolic tendencies)
  • Currently taking oral corticosteroids
  • Judged unsuitable for research by treating physician (e.g., language barrier, complex medical history, parental stress)
  • No improvement after initial IV abortive therapy

Baseline Characteristics

CharacteristicControl (Placebo, n = 46)Active (Dexamethasone, n = 41)
Median Age (years, IQR)13 (13, 15)14 (11, 15)
Sex - Female72% (33/46)85% (35/41)
Median Duration of Pain Before ED (h, IQR)96 (36, 168)108 (36, 168)
Nausea22% (10/46)26% (11/41)
Pallor26% (12/46)29% (12/41)
Fatigue61% (28/46)81% (33/41)
Visual Aura (seeing flashpoint)33% (15/46)17% (7/41)
Extremity Numbness11% (5/46)24% (10/41)
Prior Diagnosis of Migraine54% (25/46)51% (21/41)
History of Motion Sickness41% (19/46)56% (23/41)
Family History of Migraine59% (27/46)59% (27/41)
Previously Seen by Neurologist33% (15/46)29% (12/41)
Prior Head CT Scan41% (19/46)46% (19/41)
Median Pain Pre-Treatment - VAS (IQR)72 (64, 83)71 (59, 81)
Median Pain Pre-Treatment - VNS (IQR)8 (7, 8)7 (6, 9)
Median Pain at Discharge - VAS (IQR)4 (0, 16)2 (0, 13)
Median Pain at Discharge - VNS (IQR)1 (0, 2)1 (0, 2)
No Pain at Discharge35% (16/46)49% (20/41)

Arms

FieldDexamethasoneControl
InterventionStandard therapy (IV metoclopramide 0.5 mg/kg [max 10 mg] + IV diphenhydramine 0.5–1 mg/kg [max 50 mg]) followed by a single IV dexamethasone 0.6 mg/kg (max 15 mg) before discharge, plus oral naproxen 5 mg/kg twice daily for 48 hSame standard therapy plus a single IV normal saline (NaCl 0.9%) of equal volume before discharge, plus oral naproxen 5 mg/kg twice daily for 48 h
DurationSingle dose in ED; 48 h of oral naproxen at homeSingle dose in ED; 48 h of oral naproxen at home

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Relapse within 48 h of ED discharge (recurrence of any headache after pain-free status, or >1-point VNS increase in those discharged with residual pain)Primary44% (20/46)39% (16/41)Not applicable (no significant benefit)Not significant (95% CI crosses 0)
Median pain VAS at 48 h (0–100)Secondary10 (IQR 0–49)10 (IQR 0–49)
Median pain VNS at 48 h (0–10)Secondary2 (IQR 0–5)2 (IQR 0–5)
Median pain VAS at day 7Secondary10 (IQR 0–41)1 (IQR 0–46)
Median pain VNS at day 7Secondary2 (IQR 0–4)0 (IQR 0–5)
Returned to school at 48 hSecondary37% (17/46)39% (16/41)
Returned to normal activity at 48 hSecondary57% (26/46)65% (27/41)
Physician visit within 7 daysSecondary22% (10/46)17% (7/41)
Took additional medication within 48 hSecondary91% (42/46)85% (35/41)
New symptoms at 48 h (any)Adverse13% placebo (6/46) vs 5% dexamethasone (2/41); risk difference -8% (95% CI -24% to 10%), p = 0.272
Dexamethasone group new symptomsAdverseAbdominal pain (n=1), throat pain (n=1)
Placebo group new symptomsAdverseAbdominal pain (n=2), panic attack (n=1), dizziness (n=2), buccal aphthous lesions (n=1)
Serious adverse eventsAdverseNone reported in either arm

Subgroup Analysis

Sensitivity analysis of primary outcome using imputation for the 29 patients (25%) lost to follow-up: best case (missing placebo = relapse, missing dex = no relapse) yielded 28% dex vs 55% placebo (absolute difference 28%, 95% CI 2%–49%); worst case yielded a 22% higher relapse with dexamethasone (95% CI -3% to 45%). No prespecified subgroup analyses were reported.


Criticisms

  • Substantially underpowered: only 87 of the pre-specified 116 patients (target 58/arm) provided 48-h outcome data; sensitivity analyses ranged from large benefit to large harm.
  • Single-center trial at CHU Sainte-Justine over 12 years (2013–2025), limiting generalizability and raising the possibility of secular changes in practice.
  • 25% loss to follow-up via mailed diaries; children lost to follow-up had higher baseline VNS pain, introducing potential attrition bias.
  • Recruitment restricted to weekdays 8 a.m.–6 p.m. when a research nurse was on site, excluding most nighttime ED presentations.
  • Powered to detect a 25% absolute reduction in relapse — much larger than the ~10% benefit seen in adult meta-analyses — so a clinically meaningful smaller effect could have been missed.
  • Outcomes relied on self-reported diaries (VAS/VNS, functional recovery), susceptible to reporting bias.
  • Blinding assumed but not formally assessed.
  • Baseline headache frequency was not captured, so some 'relapses' could represent new attacks rather than true recurrences.

Funding

Not specified in the article; authors declare no conflicts of interest.

Based on: PED-DEX (Headache: The Journal of Head and Face Pain, 2026)

Authors: Tourigny-Ruel G, Bailey B, Jean-Charles S, Gravel J

Citation: Headache. 2026;66(6):1235-1243. doi:10.1111/head.70087

Content summarized and formatted by NeuroTrials.ai.