Overview

We conducted a systematic review to assess the effectiveness of pharmacologic and nonpharmacologic therapies for acute treatment of episodic migraine in adults. Recognizing the opioid epidemic in the United States of America, therapies were divided into opioids, nonopioid drugs (e.g. acetaminophen, nonsteroidal anti-inflammatory drugs [NSAIDs], triptans, ergot alkaloids, gepants, ditans, combination analgesics, muscle relaxants, antiemetic medications), and nonpharmacologic therapy (e.g. acupuncture, eye movement desensitization reprocessing, noninvasive neuromodulation devices). Further, we assessed the adverse events.

High and moderate strengths of evidence (SOE) support the effectiveness of triptans and NSAIDs; respectively. Two systematic reviews evaluated the combination of sumatriptan and naproxen and suggested that the combination of triptans and NSAIDs is also effective, well tolerated and can be used for patients with partial response to either agent.^47,48 These along with dihydroergotamine, antiemetics, and acetaminophen are considered established acute treatments for migraine. In general, the adverse events of these drugs are mild and transient. Newer therapies for acute treatment of migraine such as the gepants and the 5-HT1F receptor agonist, lasmiditan, were more effective than placebo in improving pain relief at 2 hours, 1 day, and at 1 week. Adverse events of newer medications require further study. Noninvasive neuromodulation devices are an area of new innovation for the acute treatment of migraine and there are several devices that have been given Food and Drug Administration (FDA) clearance; however, our search revealed that these therapies lack a strong evidence base with very few randomized controlled trials (RCTs). Other nonpharmacologic therapies had low to insufficient evidence.

Although opioids are commonly prescribed for acute treatment of migraine across all clinical settings and age groups,^13–18,181 this systematic review has shown that very few studies (15 studies with 2,208 patients) evaluated the use of opioids for acute treatment of migraine. The strength of evidence supporting the use of the various opioids for acute treatment of migraine was low or insufficient, and increased adverse events were noted. No included studies in the systematic review evaluated instruments to help in predicting risk of opioid misuse, developing opioid use disorders, or overdose in patients with migraine. Moreover, none of the included studies evaluated risk mitigation strategies to be used when prescribing opioids for acute treatment of episodic migraine.

The findings of this systematic review can inform shared decision making and choice of therapy, recognizing the variety in treatment types, clinical settings and routes of administration. Considering efficacy and harm outcomes as well as individual factors, such as characteristics of the migraine attack (including frequency, duration, severity, accompanying symptoms) and patient characteristics and comorbidities will help when selecting acute treatments.

Findings in Relation to What Is Known

This review provides a comprehensive overview of interventions in acute treatment of episodic migraine to address an urgent need to provide an updated summary of the current state of evidence. In addition to summarizing the evidence on established therapies, this review includes a summary of the newer agents for acute treatment of migraine, such as the gepants and lasmiditan, and nonpharmacologic interventions. With the discovery of novel therapeutic targets in migraine, the literature on acute interventions has proliferated substantially in recent years. Additionally, there are numerous published systematic reviews on guideline-recommended acute interventions, such as triptans and NSAIDs. It has become a difficult task for providers, health policymakers and other end users of the evidence to keep up with the constantly increasing body of evidence on the acute treatment of migraine and appraise this information relative to historical practice patterns. Guidelines for outpatient acute treatment recommend the use of triptans and NSAIDs as first line interventions, as well as acetaminophen for non-incapacitating attacks.²⁴ The findings of our systematic review align with these guidelines. Additionally, we have summarized the evidence showing moderate to high strength of evidence for a number of newer acute treatment options, such as gepants and ditans. These newer treatments have been FDA approved and are awaiting guideline updates to determine their place among the established therapies like triptans and NSAIDs.

Opioids were examined separately, and this report highlights the low or insufficient strength of evidence for their use despite them frequently being prescribed. Studies in several acute care settings have shown the use of opioids for migraine ranging from 16 to 71 percent.¹⁴ Both opioids and non-pharmacologic treatments have low strength of evidence. However, despite poor direct evidence on harms from these studies, there are known risks for opioids. The adverse effects captured in this review are those seen during the immediate exposure. Other adverse effects may only become apparent with frequent or long-term use of some of these treatments. Harms with frequent or long-term use of medications may relate to end-organ damage (e.g. nephrotoxicity and cardiotoxicity with NSAIDs, hepatotoxicity with acetaminophen, ergotism or peritoneal fibrosis with ergot alkaloids) and well as secondary conditions that may develop in the setting of consuming medications (e.g., medication overuse headache, misuse, opioid use disorder, and overdose). Medication overuse and the potential to develop medication overuse headache must be considered with the use of pharmacologic interventions for the acute treatment of migraine. Medication overuse headache (MOH) is operationally defined based on headache frequency (15 or more days per month for greater than 3 months) and days of use per month of specific medications.¹⁹⁸ The use of triptans, ergot alkaloids, combination analgesics, or opioids on 10 or more days per month meets criteria for medication overuse. Conversely, simple analgesics including NSAIDs and acetaminophen can be used on 15 or more days per month before this criteria for medication overuse is met.¹⁹⁸ In addition, use of more than one class of medications, for example a triptan and an NSAID, on 10 or more days per month also meets criteria for MOH.¹⁹⁸ Acute treatment options do not have an equal risk of MOH development.¹⁹⁹ Opioids and butalbital-containing medications have a two-fold higher risk of MOH development compared with simple analgesics and triptans.²⁰⁰ Analgesics and opioids have been associated with a higher risk of developing MOH compared with other treatments.¹⁹¹ Past studies from the 1980s and 1990s have raised concerns about opioid addiction secondary to treatment of migraine; noting conversion rates of opioid addiction secondary to treatment of migraines at 13 individuals per one million people and drug abuse in 19 percent of patients with three or more emergency department (ED) visits for migraine in a 42-month period and in 2.5 percent of patients with one or two visits.^201,202 More recent studies have found that opioid use is common in migraine patients and the risk of gastrointestinal-related adverse events and opioid abuse increased with long-term use of opioids.¹² No included studies in the systematic review evaluated instruments to help in predicting risk of opioid misuse, developing opioid use disorders or overdose in patients with migraine. No included studies evaluated risk mitigation strategies to be used when prescribing opioids for acute treatment of episodic migraine. The lack of risk assessment tools and mitigation strategies has major implications for practical implementation of treatment algorithms that include opioids. Noninvasive neuromodulation is a cutting-edge area of research for migraine treatment. While there are 4 devices currently FDA-approved for the acute treatment of migraine, our analysis of the literature revealed that there are actually few randomized trials to evaluate the effectiveness of their use.

Clinical Implications and Applicability of Findings

Considering patient characteristics, including comorbidities, when selecting acute treatments for migraine is important. For example, triptans and ergot alkaloids are considered vasoactive medications. Studies have shown that numerous patients with migraine have cardiovascular risk factors that may preclude the use of vasoactive medications.^203,204 Lasmiditan, as well as the gepants showed high SOE for acute treatment of migraine and, given their mechanisms of action, are believed to be nonvasoconstrictive.²⁰⁵ Treatment guidelines will need to be updated to reflect the evidence supporting newer therapies for acute treatment of migraine, such as the gepants and ditans, especially when considering certain patient populations like those with vascular risk factors. Our analysis uncovered that current acute treatments lack data in specific subpopulations, including the elderly, individuals with specific forms of migraine such as hemiplegic, as well as others with certain medical comorbidities. It will be important for future research to include these groups so that there can be a more robust evidence base to help guide treatment recommendations.

This review captured the acute treatment of migraine in different settings, including ambulatory as well as ED and urgent care. When considering the results of this systematic review, one should consider that the scenarios that would prompt an individual to seek care at an ED or urgent care versus self-treat at home may be different. The implication with presenting to the ED or urgent care with a migraine is that often the attack is refractory to treatments already tried at home and more likely to be severe or incapacitating, and accompanied with significant nausea, vomiting, and potentially dehydration. Although sub-group analyses were conducted to evaluate how settings might have affected outcomes (Appendix Tables K.1 to K.5), these subgroups were inherently limited by small numbers of studies and typically underpowered to detect true differences.

The evidence for opioids in acute treatment of migraine is low or insufficient based on this review. Although not captured in this review, risk for adverse outcomes pertaining to frequent or long-term intermittent use, such as misuse, opioid use disorder, and overdose, must be considered. These findings should prompt a review of the common prescribing practices of opioids for the acute treatment of migraine, as the evidence supports the current guidelines that opioids not be used as first line therapy for migraine. The lack of tools to select patients for opioids use or stratify their risk for abuse and misuse, can greatly impact the applicability of the evidence. Patients would benefit from improved implementation efforts to ensure clinical practice is consistent with guideline recommendations.

Access to interventions can be a barrier to obtaining acute treatment for migraine. Certain medications used for the acute treatment of migraine may not be reliably stocked by pharmacies due to potentially serious adverse effects that require close monitoring (e.g. ergot alkaloids) or due to insurance restrictions and cost (e.g. gepants and ditans). Some of the newer drugs may not be accessible or afforded by all patients. The noninvasive neuromodulation devices, despite being given FDA clearance for the acute treatment of migraine, are not routinely covered by insurance and can be cost prohibitive. Conversely, even though many of the nonpharmacological treatments have only few trials and lack robust evidence per our analysis, guidelines may consider their favorable safety profile in recommending these as treatment alternatives and they are becoming more commonplace in clinical practice. As they are used more routinely, additional studies will help clarify their role in the treatment algorithms. Furthermore, with the development of gepants and ditans, guidelines may consider how these new drugs may fit among the established options of first line therapies, as well as among other alternatives such as nonpharmacologic options with low SOE but favorable safety profiles, and opioid options with low or insufficient SOE but known risks. Currently, the lack of comparative effectiveness studies amongst all of these different treatment choices is an important research gap.

Although only studied in one or a few small trials, several other therapies may improve migraine pain compared with placebo, including dexamethasone, dipyrone, lidocaine, magnesium sulfate, octreotide, and secobarbital (low SOE). Evidence was insufficient to draw conclusions about serious adverse events of these interventions. Although the strength of evidence is low, clinically these interventions are considered if patients do not respond, encounter side effects, or have contraindication to the more established treatments.

Patients are often advised to use combinations of acute therapy to treat migraine attacks. This combination can include an antiemetic as well as migraine specific therapy such as a triptan and a nonspecific analgesic such as an NSAID. The trials we analyzed did not sufficiently evaluate these potential combination therapies for the acute treatment of migraine and it remains unclear which combination of treatment may have the best evidence, which combinations may have a synergistic benefit, and which may simply be additive.

Limitations and Suggestions for Future Research

For many interventions, very few RCTs, in some cases only one, were available (Key Questions 1-3) and some were small, which limits inferences from the quantitative analysis. Consequently, failure to detect statistical significance for many of the outcomes could have resulted from type II error and lack of power. Although we used very broad search terms and sought Supplemental Evidence and Data from the public to identify all relevant interventions, it is still difficult to be certain that all the appropriate literature and relevant interventions (e.g., nonpharmacologic therapy) has been included.

Most of the studies compared the interventions with placebo. Future trials should focus on comparative trials between different acute medication choices, particularly those that have the highest levels of evidence, to help clinicians decide amongst all of the available options and trials that look at combinations of therapies. This is important to also help clarify the place for the newer therapeutic options, such as gepants and ditans, and nonpharmacologic options compared with the more established therapies, such as NSAIDs and triptans.

The clinical trials included in our analysis generally excluded many important populations, including those with cardiovascular problems, cerebrovascular problems, hemiplegic migraine, and frequently individuals over the age of 65. Further studies evaluating the efficacy of acute treatments in these specific populations will be important, particularly now that we have options that are believed to be safe in some of these groups.

Pain as a main outcome of migraine research is a challenge to all investigations due to the subjective nature and how it is affected by each individual’s psychological states and recall. While randomization, blinding and standardized instruments are used in trials, these approaches do not fully address this challenge.

Due to the quantity of literature captured by this review, studied endpoints were limited to those relating to pain freedom and relief, function, and harms. Migraine is defined by a set constellation of symptoms, including not only headache, but also photophobia, phonophobia, nausea and vomiting, and others. Unfortunately, we were unable to report on these outcomes. For some patients, the pain component may not be as bothersome as the non-pain symptoms of migraine. Recognizing this, most bothersome symptom (MBS) has been suggested as a preferred endpoint. In migraine studies, there is a shift to use more patient-centric endpoints that reflect the quality of life impacted by migraine and its return to normal by acute treatment rather than only pain freedom or pain improvement. Total migraine freedom, or absence of all migraine-related symptoms including pain and all associated symptoms and return to baseline, is another important patient-centric endpoint. Future studies should emphasize these patient-centric endpoints. Patients have also indicated that speed of onset of acute treatment is highly important.^206–208 The available literature did not use these endpoints consistently, and hence the current analysis cannot address them. Consequently, future studies can compare the time it takes to reach to pain freedom, total migraine freedom, and MBS freedom as clinically meaningful endpoints. As mentioned in the limitations section, comparing outcomes relating to non-pain symptoms of migraine such as photophobia, phonophobia, nausea, and vomiting, is also important to reflect the entirety of the patient experience.

The harms outcomes that were captured in this review are those relating to adverse effects seen during the immediate exposure period. Consequently, this review does not capture harms that may arise with frequent or long-term intermittent use of these treatments. The inability to capture such harms is due to limitations in study design as the majority of trials evaluating acute treatment of migraine evaluate the efficacy and harms of the intervention during one or a few attacks. For example, telcagepant was studied against placebo in 6 RCTs between 2007 and 2012 and included a total of 6,021 subjects.^69,108–112 These studies showed significant improvement of the drug at all endpoints including restored function at 2 hours, being pain free at 2 hours, pain relief at 2 hours, sustained pain free at 1 day, sustained pain free at 1 week, sustained pain relief at 1 day, and sustained pain relief at 1 week. When studies shifted from intermittent acute use to a more chronic use of telcagepant, hepatoxicity was noted and this halted further research into this medication.²⁰⁹ This example also speaks to the importance of being mindful of adverse effects which may not yet be known in the newer treatment options. Postmarketing data will be important to monitor for adverse effects of the newer therapeutic options. Additionally, harms from treatments are less likely to arise when taken under the controlled parameters of a trial verses in real world circumstances where individuals may be using interventions outside the parameters of the recommended dosing. Future studies should also routinely report on MOH as an outcome. The International Classification of Headache Disorders, 3rd edition (ICHD-3) has established that frequent use of acute therapies for migraine can lead to increase in migraine attack frequency over time, and this risk varies depending on the acute treatment. Furthermore, medication overuse has been associated with significant disability, as evidenced by the 2016 Global Burden of Disease study,²¹⁰ where it was listed amongst the top 20 causes of years lived with disability worldwide. Future studies on the acute treatment of migraine can compare relative risks of MOH with different classes of acute treatments so that this important problem can be better addressed.

The included studies were also conducted in different settings, from the ED to outpatient to inpatient environments, which all have implications regarding the type of migraine attack being treated. It is generally accepted that some acute treatments work best when taken early in an attack whereas others can still work in more refractory situations such as those that prompt ED visits and inpatient stays. Given that patients may not respond to the initial acute treatments that they receive, future research should look at treatment efficacy in patients that have failed an initial acute treatment as this is clinically relevant information.

Available evidence for opioids in the acute treatment of migraine is low or insufficient. Despite this, they continue to be frequently prescribed. Research to evaluate why opioids are being prescribed may help provide strategies to address the opioid epidemic. One study found that physicians who were more likely to prescribe an opioid for a migraine headache were also more likely to prescribe an opioid to a patient with back pain.²¹¹ This finding may suggest that physicians are lumping different types of acute pain together and not considering the nuances of the individual conditions and the evidence for treatment efficacy specific to those conditions. Factors that may affect opioid prescribing patterns should be studied.

With the advent of noninvasive neuromodulation, further research, including comparative studies with medications, to truly clarify their role as acute therapies for migraine are needed. These devices are being recommended by headache specialists in clinical practice more frequently given their safety profile despite lack of repeated, large-scale studies confirming that they are effective acute treatments. Cost and lack of insurance coverage is a current barrier with these devices; hopefully with additional studies establishing their role as acute therapies for migraine this hurdle can be made easier for patients.

Additionally, it is important to also note that while it is accepted that behavioral pain therapies such as cognitive behavioral therapy, mindfulness based stress reduction, and others can play an important role in the treatment migraine, there was a striking lack of evidence when a critical appraisal of the available data was done. This lack of rigorously designed, high quality clinical trials involving behavioral pain therapies leads clinicians to rely more on pharmacological options for migraine, accepting their risks and adverse effects. The lack of scientific evidence for behavioral pain management of migraine remains a significant limitation. Risk mitigation, using an integrated approach that combines medications and behavioral pain management, can only be successful if we have an improved evidence base. This lack of scientific evidence for behavioral pain management in migraine can be addressed with future studies that help strengthen this body of scientific literature.

This review focused on abortive, not preventive migraine treatment. However, there is a current paradigm shift regarding migraine and when to start preventive treatment. Rather than simply considering prevention based on number of migraine attacks per week or frequency of acute medication consumption, there is a shift to consider migraine-associated disability as an important determinant as to whether an individual should be placed on a preventive medication, regardless of attack frequency. It would be very helpful for shared decision making regarding abortive treatment to know whether people who are on preventive therapy have an improved response to a particular acute treatment.

Disparities based on race and socio-economic status exists in the acute treatment of migraine. Future research to identify the disparities, identify determinants that contribute to these disparities, and explore strategies to overcome these are needed. Cost of acute treatments affect access and may contribute to disparities in prescribing patterns for patients depending on race and socio-economic status including insurance coverage.

Finally, it is important to note that multiple interventions, including BI44370, telcagepant, dapitant, lanepitant, selurampanel, tezampanel, tonabersat, and flunarizine, are not FDA approved or not available in the United States.

Conclusion

A number of acute treatments for episodic migraine exist with varying degrees of evidence. In addition to already established effective treatments, such as triptans, NSAIDs, antiemetics, and ergot alkaloids, newer treatments such as gepants and ditans are associated with improved outcomes in pain and function in acute treatment of episodic migraine. Opioids have low or insufficient strength of evidence for acute treatment of migraine. Despite increasing literature pertaining to migraine, the evidence base for many interventions in migraine remains limited. Selection of acute treatments for migraine must be individualized based on adverse effect profile and patient characteristics such as relevant comorbidities. Continued research is required to assess the comparative effectiveness and harms of several pharmacological and nonpharmacologic treatments.