A Randomized Trial of Ketorolac and Metoclopramide for Migraine in the Emergency Department

Lawrence P. Richer MD, MSc; Samina Ali MD; David W. Johnson MD; Rhonda J. Rosychuk PhD; Amanda S. Newton PhD; Brian H. Rowe MD, MSc


Headache. 2022;62(6):681-689. 

In This Article


Our randomized controlled trial failed to demonstrate any clinically meaningful or statistically significant difference in pain relief between combination iv treatment (metoclopramide + ketorolac) and metoclopramide monotherapy for children or adolescents presenting to the ED for acute treatment of migraine headache. There was also no significant difference observed between groups for secondary outcome measures after ED treatment, including complete pain relief, headache improvement, or other associated symptoms.

Unfortunately, there is a dearth of evidence to support some of the most common treatments used for pediatric patients in the ED, which leads to significant practice variation.[2] This study explored two commonly used treatments for migraine headache, metoclopramide and ketorolac, with the specific goal of answering the question: does the combination of these agents improve outcomes compared to monotherapy with metoclopramide. The scientific rationale for including iv ketorolac, alone or in combination with other agents, is compelling given the benefits that others have reported in reducing central sensitization in animal models of migraine[5] and adult patients.[6] Ketorolac, alone or in combination, remains one of the most commonly used drugs, being given in almost 40% of ED presentations for acute treatment of migraine headache.[2–4] To our knowledge, no studies examining the comparative effectiveness of using ketorolac in combination with metoclopramide have previously been published in the pediatric literature.

Overall, less than 20% of participants in our study reported pain freedom within 2 h of treatment—the most desirable outcome for patients. This is much lower than the 50% previously reported in observational studies of prochlorperazine,[16] for example, and demonstrates the need for controlled comparative effectiveness studies in this population. Metoclopramide, the active treatment in both arms of our study, is also one of the most used drugs in the treatment of pediatric migraine headache in the ED.[2] There is growing evidence, however, to suggest that other neuroleptics may be more effective. In one large observational study of migraine treatment in 35 academic children's hospital EDs, there was 31% increased odds of returning to the ED with metoclopramide when compared with prochlorperazine.[3] Sheridan et al. also observed that the use of prochlorperazine was associated with lower use of rescue opioids when compared with metoclopramide or promethazine in the pediatric population.[17] One important safety concern is the higher incidence of akathisia and dystonic reactions with prochlorperazine compared to other agents. In one large observational study of 4588 clinical encounters at a single tertiary pediatric hospital, the relative risk of a dystonic reaction was almost nine times higher with prochlorperazine.[18]

It was noteworthy that 41% of our study participants left the ED prior to the 120 min pain assessment, as it was deemed unreasonable to require that they remain in the ED if their treatment was effective. The number who left earlier than 120 min was equal in both groups though and the only significant predictor of the final VAS on multiple linear regression was baseline pain severity. There was no difference in pain severity at any of the time points from 30 to 90 min. Future ED studies need to account for variable time of last assessment and baseline pain severity in the statistical planning. The time to therapeutic benefit could then be estimated, which may have an impact on ED flow.

Almost one third of participants reported the presence of moderate to severe headache in follow-up and almost one third reported headache recurrence after achieving effective relief in the ED (i.e., my headache is "a lot better"). This is comparable to the recurrence rate observed in adult studies (19%–31%)[19] and the only other pediatric study of prochlorperazine versus ketorolac (30%).[8] Sustained pain freedom was only observed in six participants (6/27, 22%); however, five of these participants were in the ketorolac group. The presence of moderate or severe headache at 24 h post-discharge was also 20% lower in the combination therapy group. Although our study was not powered to detect outcomes like sustained pain-freedom or headache recurrence after leaving the ED, future studies may be needed to explore this potential effect.


This study has several limitations. First, administering additional hydration in addition to study medications may have lessened the likelihood of detecting a treatment effect. In a previous trial of ours on treatment expectation and saline infusions, the overall decrease in pain with iv fluids alone was small and clinically insignificant.[20] The expectation of treatment also did not significantly affect pain intensity when assessed at 30 min in the latter study.

Second, generalizability to the formally diagnosed pediatric and adolescent headache population may not be possible, because patients with "probable migraine" were included and PedMIDAS scores did not capture any patients with severe migraine-related disability (scores were 44 or less). It is notable, however, that these inclusion criteria reflect the real-world point-of-care decision making that emergency physicians would be required to make when treating children and adolescents with migraine.

Third, the sample size was powered to detect a minimum clinically important difference of 20 mm and smaller effects may have been missed. The study was also not powered to detect differences in follow-up. Although the baseline characteristics were mostly similar for both groups, the duration of the current migraine attack prior to presenting to the ED differed. Importantly, in logistic regression modeling, the differences in baseline characteristics failed to demonstrate an impact on the primary outcome and effect estimate for treatment. Fourth, the study protocol also did not assess or exclude participants with a history of medication overuse, and the use of abortive medications prior to ED presentation may have created a cumulative effect of medications on the outcome. Fifth, because migraine is an episodic disorder, an expected limitation and clinical requirement was the need to allow patients to leave the ED when their symptoms were resolved/improved. Twenty-two children left the ED before the 2-h data collection time point and the lack of difference between the ketorolac group and the monotherapy group may be attributed to this. Six, with only two Canadian pediatric EDs participating, generalizability to EDs serving mixed populations and non-Canadian settings may be limited. Finally, the standard treatment used in the trial as the control group (metoclopramide) may not be the current standard of care in other centers.