Executive Summary

Publication Details

BACKGROUND

Migraine is a common and disabling health problem among adult Americans. Surveys from the US and elsewhere suggest that 6% of men and 15%–17% of women experience migraine headaches (Stewart, Shechter, and Rasmussen, 1994). These headaches result in significant disability and work loss; estimated aggregate indirect costs to employers in the US for reduced productivity due to migraine range from 6.5 to 17 billion dollars annually (Osterhaus, Gutterman, and Plachetka, 1992).

Patterns of medical care for the treatment of migraine are highly variable. A substantial proportion of migraineurs never consult a physician about their headaches. Among those who do seek medical attention, many do not continue with the course of treatment prescribed by their physician (Edmeads, Findlay, Tugwell, et al., 1993), citing the availability of non-prescription medications and negative side effects associated with prescription medications among the reasons for self-treatment.

The pathophysiology of migraine is poorly understood, but recent advances in neuroimaging, neurobiology, genetics, and pharmacology have resulted in a changing view of migraine pathogenesis, from one favoring vascular and muscle tone as primary causes to one involving a primary neuronal event producing secondary vascular changes. There continues to be controversy among headache researchers and clinicians over whether tension-type headache and migraine are part of a single spectrum of headache disorders or different conditions with distinct etiologies. The diagnostic distinction between the two types of headache is, however, generally assumed in clinical trials, most of which have focused on one or the other diagnosis.

A substantial body of high-quality evidence exists describing the effectiveness of various drugs for the treatment of acute migraine. Synthesis and dissemination of this information may help correct the underuse or misuse of acute drug treatment strategies for migraine headache.

SCOPE OF REPORT

The objective of this Technical Review is to provide a comprehensive review and analysis of published reports of randomized controlled trials (RCTs) and other prospective, comparative clinical trials of self-administered drug treatments for the acute, episodic treatment of migraine headaches. The report is restricted to trials of treatments that can be administered by a patient in a non-clinical setting (i.e., at home or work). There are many such agents, ranging from over-the-counter analgesics to migraine-specific prescription drugs; most are taken orally, but some are administered by intranasal insufflation, subcutaneous injection, or rectal suppository. Studies of parenteral treatments ordinarily administered in a clinical setting will be covered in a companion report (Parenteral drug treatments for acute migraine headache). The present report does not cover all analgesics or other drugs that may be used for the treatment of migraine, but only those that have been studied in controlled trials among a population of migraineurs. These include aspirin, acetaminophen, non-steroidal anti-inflammatory drugs (NSAIDs) (diclofenac sodium, diclofenac-K, flupirtine maleate, flurbiprofen, ibuprofen, ketoprofen, lysine acetylsalicylate, mefenamic acid, naproxen, naproxen sodium, piroxicam, pirprofen, proquazone, and tolfenamic acid), ergot alkaloids (ergotamine tartrate (ET) and ET combinations), dihydroergotamine (DHE), sumatriptan and other 5HT1-receptor antagonists, opiate analgesics, isometheptene combinations, and domperidone.

METHODOLOGY

The literature review addressed the question “What are the effects on headache pain and the tolerability of self-administered drug treatments for acute migraine headache compared to placebo, alternative drug treatments, and non-drug therapies?”

To be considered for this review, studies were required to be prospective, controlled trials of self-administered drug treatments aimed at the relief of symptoms of individual episodes of headache in patients with migraine. Although the use of a specific set of diagnostic criteria (e.g., Ad Hoc Committee on the Classification of Headache, 1962; Headache Classification Committee of the International Headache Society, 1988) was not required, diagnoses were required to be based on at least some of the distinctive features of migraine, e.g., nausea/vomiting, severe head pain, throbbing character, unilateral location, phono/photophobia, or aura.

Studies were included only if allocation to treatment groups was randomized or pseudo-randomized (based on some non-random process unrelated to the treatment selection or expected response); concurrent cohort comparisons or other subexperimental designs were excluded. Control groups could comprise placebo, no intervention, usual care, or a specified alternative drug or non-drug treatment.

Relevant controlled trials were identified by searching MEDLINE (January 1966 through December 1996) using the MeSH term “headache” (exploded) and a published strategy for identifying randomized controlled trials. Additional search strategies included computerized bibliographical searching of PsycINFO and CINAHL databases; retrospective and prospective hand-searching of the journals Headache, Cephalalgia, and Headache Quarterly; searching the reference lists of review articles and included studies; searching books related to headache; and consulting experts in the field. We also searched a database of randomized trials in pain relief which is now part of the Cochrane Controlled Trials Register (1997).

Studies identified by the literature search were screened for further review based on criteria focusing on patient population, intervention, study design, and type of outcome data reported.

Studies passing the initial screen were reviewed for methodological quality based on the following considerations: the use of random allocation; description of an adequate method of concealment of allocation; the use of double-blinding; description of an adequate method of blinding; and a description of drop-outs sufficient to determine the number of patients in each treatment group entering and completing the trial. Each trial could score between 0 and 5 points, with higher scores indicating higher quality in the conduct or reporting of the trial.

Efficacy and adverse events data were abstracted from the original reports onto specially designed forms. We collected trial data on short-term symptomatic outcomes related to head pain (severity/intensity, relief, and duration), other symptoms (nausea, vomiting, photophobia, phonophobia), functional status (disability), and quality of life. We did not consider physiological or other measures not directly relevant to the patients’ symptomatic experience.

We required that outcome data be obtained directly from the patient, not judged by the treating physician or study personnel. We required that data be recorded at or near the time of symptoms. We chose to analyze 2-hour data whenever they were reported. If 2-hour data were not reported, we chose the measurement closest to two hours.

Among outcomes related to headache pain, we generally preferred those that measured headache relief or change in headache intensity, since these are more comparable among patients with different baseline pain intensity scores (Scott and Huskisson, 1976). If these outcomes were not available, we analyzed data on post-treatment headache intensity. If no headache relief or pain intensity data were reported, then we used data on functional disability or headache duration as a proxy.

For outcomes measured on a dichotomous scale (success/failure), we required that the threshold for distinguishing between success and failure be clinically significant; for example, we interpreted a 50% or more decrease in severity or a reduction from moderate-severe to mild-none (two of the most common definitions) as meeting this criterion. Some studies also reported results for a more stringent criterion, complete resolution of symptoms. We did not consider a reduction in severity from mild to none as meeting our criterion.

Dichotomous outcomes meeting our definition of a clinically significant threshold were reported as proportions (or response rates for each treatment) which may be directly compared (difference in proportions). We also used these proportions to calculate odds ratios.

When outcome data were provided on an ordinal scale (e.g., for headache relief: none, mild, moderate, near complete, complete), we selected a threshold based on the definition for improvement (discussed above) and converted these data into a dichotomous outcome. When categorical data could not be split into dichotomous outcomes meeting our a priori definition and no continuous data were reported, we assigned a numeric score to each category and analyzed the results as continuous data.

When outcomes were reported on a continuous scale (e.g., a visual analog pain scale) and variance estimates were also available, we calculated an effect size, or standardized mean difference.

When several efficacy estimates were available from comparable trials for a given treatment comparison, we tested these for homogeneity and combined them using an inverse variance weighted method to yield a meta-analytic summary estimate.

Adverse events

The incidence of adverse events was recorded and the proportion of patients experiencing adverse events calculated for each treatment group, whenever possible. The difference between rates of adverse events was calculated along with a 95% confidence interval (CI) for the difference. A 95% CI that excludes zero suggests that the rates are significantly different between groups. The identity and rates of specific adverse events reported were summarized for each study.

SUMMARY OF FINDINGS

The evidence supporting different agents or classes of agents varies greatly in quantity and quality. This section summarizes the results of our analysis, and the next section describes the most urgent priorities for future research.

NSAIDs and other non-opiate analgesics

Our analysis included 33 controlled trials involving the following agents in this class: acetaminophen, aspirin, diclofenac sodium, diclofenac-K, flupirtine maleate, flurbiprofen, ibuprofen, ketoprofen (pr), lysine acetylsalicylate, mefenamic acid, naproxen, naproxen sodium, piroxicam (sl), pirprofen (pr and po), proquazone, and tolfenamic acid. Except where otherwise noted, all study drugs were administered orally. Diclofenac-K, flupirtine maleate, pirprofen, and proquazone are not available in the US.

The principal findings of the analysis were these:

Although there were not many studies of any one agent, the placebo-controlled trials reviewed in this report were remarkably consistent in demonstrating the efficacy of this class of drugs for the relief of acute migraine attacks. There were three positive placebo-controlled studies of aspirin, two of ibuprofen, two of tolfenamic acid, and one each of diclofenac-K, flurbiprofen, naproxen, naproxen sodium, piroxicam, pirprofen and proquazone. In addition, one study each of diclofenac sodium, ketoprofen, naproxen, and pirprofen provide point estimates favoring the active drug over placebo with confidence intervals that did not exclude null effect. There were no studies that had point estimates for efficacy measures that favored placebo.

The only study comparing acetaminophen with placebo found no significant difference in efficacy between the two.

Three trials made direct comparisons of one agent in this class with another. With one exception showing tolfenamic acid superior to acetaminophen, no significant differences were observed among these agents.

A series of studies examined the effect of adding an antinauseant (domperidone or metoclopramide) or caffeine to the main analgesics reviewed in this report. With the exception of a single study which found that tolfenamic acid + metoclopramide was better than tolfenamic acid alone at reducing headache severity, these studies demonstrated that the combination agents offered no significant advantage over the analgesics alone.

Comparisons with other drugs commonly used for the acute treatment of migraine demonstrated few important differences. The aspirin-containing opiate compounds Doleron® and Doleron novum® were found to be superior to aspirin alone in two trials. Ergotamine was shown to be superior to aspirin in two of three trials comparing the two. However, no significant differences were observed between ergotamine and ketoprofen, naproxen sodium, or tolfenamic acid, or between ergotamine + caffeine and naproxen sodium. Furthermore, in two trials of ergotamine-containing compounds, Cafergot Comp.® was not significantly more effective than pirprofen for complete relief at 30 minutes, and Migwell® was significantly worse than naproxen sodium for headache severity.

Aspirin was well tolerated in nine trials and was shown not to have significantly more adverse events than placebo, tolfenamic acid, aspirin + metoclopramide, acetaminophen + codeine, or Doleron®. Aspirin was associated with fewer adverse events than ergotamine.

Trials of other NSAIDs demonstrated the well-known adverse events of gastric irritation/discomfort, nausea, and vomiting. In comparison to ergotamine and ergotamine-containing compounds, however, NSAIDs were consistently associated with lower overall adverse event rates and, in particular, with lower rates of nausea and vomiting.

The addition of caffeine or an antinauseant did not increase the total number of adverse events, but neither is there consistent evidence that adding an antinauseant reduces the adverse gastrointestinal events typically associated with NSAID use.

Ergot alkaloids

Our analysis included 23 separate controlled trials of ergotamine or ergotamine-containing compounds. These trials were conducted over four decades, during which the diagnosis of migraine has evolved through two sets of diagnostic criteria. Differences in the specifications for subjects, trial design, and quality, as well as the quality of reporting, make these trials a challenge to interpret. The main findings of the review were:

Five placebo-controlled trials of ergotamine tartrate ranged from finding no effect to finding large differences in favor of ergotamine. Only three studies provided sufficient information to calculate a quantitative estimate of the efficacy of ergotamine compared to placebo, and two of them suggested clinically important benefits from ergotamine. The three trials differed in the dose of ergotamine used, the route of administration, and the definition of migraine used for inclusion, so meta-analysis of the estimates was not attempted.

Combinations of ergotamine tartrate + caffeine did not demonstrate significant differences in headache relief compared to placebo. Two additional ergotamine-containing proprietary combinations (Cafergot Comp.® and Cafergot® P-B) were found to be superior to placebo and, in the case of Cafergot® P-B, superior to ergotamine + caffeine. Another combination (ergotamine + caffeine + cyclizine [Migwell®]) was not found to be any better than placebo in treating nausea.

Compared to the number of placebo-controlled trials, the number of trials comparing different ergotamine-containing compounds was relatively small. One study found Cafergot® P-B superior to Cafergot® in relieving headache pain severity. Ergotamine + metoclopramide was found to improve headache duration in one study and to reduce nausea in another when compared to ergotamine tartrate alone. Otherwise, no significant differences were shown among ergotamine tartrate, Cafergot®, Cafergot Comp.®, and ergostine.

Given equivocal efficacy in placebo-controlled trials, the comparisons of ergotamine-containing compounds with other agents are difficult to interpret. Two of three comparisons of ergotamine with aspirin alone found ergotamine significantly better in terms of headache relief. However, ergotamine was not significantly different from two aspirin-containing compound drugs (Doleron® and Doleron novum®), three NSAIDs (ketoprofen, naproxen, tolfenamic acid), metoclopramide, or an isometheptene combination (Midrin®/Midrid®). However, only one of these mostly small negative trials ruled out the possibility of a clinically significant benefit to ergotamine in comparison to the alternative drug.

Ergotamine + caffeine was not significantly different from DHE nasal spray or naproxen but it was inferior to oral sumatriptan and an isometheptene combination (Midrin®) in reducing headache severity. Cafergot Comp.® was not significantly different from pirprofen in providing complete headache relief at 30 minutes. The ergotamine + caffeine + cyclizine combination (Migwell® or Migril®) was shown to be worse than naproxen sodium but not significantly different from Migraleve® at reducing headache severity. The combination of ergotamine + metoclopramide appeared to reduce headache duration compared with metoclopramide alone.

There was consistent evidence for increased incidence of nausea and vomiting with ergotamine tartrate compared with placebo groups. Most of the ergotamine combinations -- including ergotamine + caffeine, Migwell®/Migril®, Cafergot® P-B, Cafergot Comp.®, and ergotamine + metoclopramide -- appeared to result in rates of nausea and vomiting lower than those associated with ergotamine alone.

Ergotamine and ergotamine-containing compounds consistently demonstrated a higher incidence of adverse events than placebo or comparator drugs including sumatriptan, isometheptene combinations, NSAIDs, and dextropropoxyphene compounds. Nausea and vomiting were the most commonly observed adverse events.

DHE Nasal Spray

Our analysis included 12 controlled trials of intranasal dihydroergotamine (DHE). The principal findings were as follows:

DHE nasal spray showed generally consistent evidence for efficacy in nine separate placebo-controlled trials. Meta-analysis of homogeneous results from four trials showed a statistically significant benefit in favor of DHE; the magnitude of the benefit was, however, small to moderate. Three comparisons of different doses of DHE were inconclusive.

In the only two studies comparing DHE nasal spray other active treatments, no significant difference was observed between DHE and Cafergot® for headache relief, and subcutaneous sumatriptan was found to be significantly better than DHE at providing headache relief and complete relief at 2 hours.

Adverse events reported in association with DHE were generally described by investigators as mild to moderate and were clearly related to the intranasal route of administration. The rate of adverse events was similar to that observed with Cafergot® in one comparative study and significantly lower than that observed with subcutaneous sumatriptan in another.

Subcutaneous sumatriptan

Our analysis included 15 controlled trials of subcutaneous sumatriptan. The main findings were:

Subcutaneous sumatriptan, in the 14 placebo-controlled studies reviewed, was consistently shown to be effective in relieving moderate to severe migraine headache pain at 1 and 2 hours. Limited data also indicated that sumatriptan was significantly more effective than placebo at providing complete headache relief at 1 and 2 hours.

Sumatriptan as shown in one trial to be equally effective in relieving migraine with aura and migraine without aura. The same study provided data suggesting that sumatriptan is effective in providing headache relief even when taken more than 4 hours after the onset of symptoms. At the other end of the spectrum, however, the evidence from another trial suggests that sumatriptan is ineffective when taken during the migraine aura, before the onset of head pain, or while head pain is still mild.

Two studies suggested that there is no additional benefit to a second dose of sumatriptan for patients who do not adequately respond to the first. Moreover, a higher number of patients experienced adverse events on a two-dose regimen than on a single-dose regimen, thus supporting the view that a second dose of sumatriptan has little benefit and may be harmful.

Some studies measured 24-hour headache outcomes and found generally higher rates of recurrent headache among patients initially treated with sumatriptan than those initially treated with placebo. Confounding from subsequent rescue medication use in placebo patients makes interpretation of these results difficult. The studies also failed to provide guidance on how to treat recurrent headaches.

Adverse events reported in association with the use of sumatriptan were mostly described as mild and transient.

Two trials compared sumatriptan with another commonly used anti-migraine drug, DHE, given either subcutaneously or intranasally. These trials reported similar results, with land 2-hr data on headache relief and complete relief favoring sumatriptan, while 2–24 hr recurrence rates favored DHE. Nausea and vomiting were more frequent among DHE patients.

Oral sumatriptan

We analyzed a total of nine controlled trials of oral sumatriptan (two comparisons with other active treatments, one comparison with another active treatment and placebo, and six comparisons with placebo alone). The main results of this analysis were:

Oral sumatriptan, in the seven placebo-controlled trials we reviewed, was significantly more effective than placebo at providing headache relief and complete relief at 2 and 4 hours. The 100-mg dose of sumatriptan has the strongest support in clinical trials. The rates for headache relief and complete relief for oral sumatriptan were slightly lower than those for subcutaneous sumatriptan.

Though the 100-mg dose of sumatriptan has the strongest support, doses of 25 and 50 mg have also been shown to be effective in two placebo-controlled studies. Neither of the studies involving these lower doses was powered to compare the 25-,50-, and 100-mg doses directly.

Two trials compared sumatriptan (100 mg) with a combination of aspirin (900 mg) or its equivalent, lysine acetylsalicylate (LAS) (1620 mg), and metoclopramide (10 mg). Overall there was not a significant difference between the two treatments in providing headache relief at 2 hours, but sumatriptan was significantly more effective than aspirin + metoclopramide at providing complete relief of headache at 2 hours.

The single study comparing sumatriptan (100 mg) with Cafergot® found that sumatriptan was significantly more effective than Cafergot® at providing headache relief and complete relief at 2 hours.

Oral sumatriptan (100 mg), like the subcutaneous form of the drug, was shown in one trial to be equally effective in relieving migraine with and without aura. The same study also showed that the duration of the attack before treatment did not affect the headache relief provided by sumatriptan: patients treating an attack more than 4 hours after the onset of symptoms had results that were similar to those treating an attack within 4 hours of the onset of symptoms.

In one large trial, the use of a second 100-mg dose of sumatriptan, taken 2 hours after the first, did not significantly increase the percentage of patients achieving headache relief, nor did it appear to prevent or delay headache recurrence. Another study suggested that a dose of oral sumatriptan, taken 4 hours after initial treatment with subcutaneous sumatriptan (6 mg), does not prevent, but may delay, headache recurrence.

Two trials examined, in a controlled fashion, the question of whether oral sumatriptan (100 mg) is an effective treatment for recurrent headache. Both found that oral sumatriptan was significantly more effective than placebo at relieving recurrent headache pain.

Adverse events were generally more common among patients taking sumatriptan (100, 50, or 25 mg) than among patients using placebo. Sumatriptan (100 mg) was also associated with a significantly higher rate of adverse events than aspirin/LAS + metoclopramide. The single study comparing sumatriptan (100 mg) with Cafergot® found no significant difference between the two treatment groups in the overall percentage of patients reporting adverse events.

The adverse events reported in association with oral sumatriptan were universally characterized as minor and transient in the literature.

Intranasal snmatriptan

The literature review identified two publications reporting the results of three independent, placebo-controlled trials of intranasal sumatriptan. The main results were:

Three trials showed that intranasal sumatriptan at doses of 20- and 40-mg was significantly more effective than placebo at providing headache relief and complete relief at 2 and 4 hours. Administering sumatriptan as a divided dose in both nostrils did not result in consistently higher response rates than administering the drug as a single dose in one nostril. Likewise, efficacy appeared to be similar whether a 40-mg dose was administered at one time or as a divided dose with the two administrations separated by 15 minutes. In general, response rates to the 20- and 40-mg doses of sumatriptan administered by the intranasal route were equivalent to those reported for subcutaneous sumatriptan and higher than those reported for oral sumatriptan.

Lower doses tested from 1 to 10 mg suggested a dose response relationship but were not clearly demonstrated to be more effective than placebo.

Very little data were available on the rate of headache recurrence after treatment with intranasal sumatriptan.

Significantly more sumatrlptan patients than placebo patients reported adverse events (37% vs. 13%), most commonly “taste disturbances.”

Other oral 5HT1D-receptor agonists (rizatriptan and zolmitriptan)

Two relatively small studies each provided support for the effectiveness of a new 5HT1D-receptor agonist drug (rizatriptan and zolmitriptan) compared to placebo. Although not directly compared to oral sumatriptan, these drugs were tested in studies that were similar in design to the oral sumatriptan trials and measured many of the same outcomes. The response rates observed with the active drugs were somewhat higher than those for sumatriptan, while placebo group response rates were similar to the sumatriptan studies, indicating comparable populations. Several other, similar compounds are being developed and tested; it is likely that many clinical trials of these new agents will be forthcoming in the near future.

Opiate analgesics

Our analysis included eleven controlled trials of opiate analgesics. The main results were:

Codeine-containing combination analgesics -- including the combination of acetaminophen + codeine and proprietary combinations including acetaminophen, codeine, and doxylamine (Mersyndol®) or buclizine (Migraleve®) -- were compared to placebo in seven trials. These trials studied varying doses of slightly different agents, used different outcome measures, and reached mixed conclusions. Though no formal meta-analysis was possible, on the whole the evidence suggests that these acetaminophen + codeine-containing agents provide statistically and clinically significant migraine relief.

A single comparison of acetaminophen + codeine + doxylamine with acetaminophen + codeine found no significant difference between the two in their ability to reduce pain intensity.

A single trial comparing acetaminophen + codeine with aspirin found that there was no significant difference between the two as far as headache relief at 2 hours was concerned. Aspirin was statistically significantly better than acetaminophen at reducing pain intensity between 0 and 2 hours, but the magnitude of the effect was small.

A single trial comparing Migraleve® with an ergotamine + cyclizine + caffeine combination (Migril®) found that there was no significant difference between the two for headache severity. Mean duration of nausea and of vomiting were slightly shorter with Migraleve®, but there were no significant differences between the two treatments as far as the severity of nausea and of vomiting were concerned.

In two placebo-controlled trials, intranasal butorphanol demonstrated consistent results indicating clinically and statistically significant efficacy at 2 hours. It has not been compared with other currently available home treatments for migraine.

Two trials comparing aspirin + dextropropoxyphene + phenazone combinations (Doleron®, Doleron novum®) with aspirin alone found that these compounds were significantly more effective than aspirin at providing complete relief of headache at 30 min. However, the dose of aspirin used was relatively low (500 mg in both trials). No other pain outcomes were measured over the entire study population.

The same two trials found that there was no significant difference between Doleron® and Doleron novum®, on the one hand, and ergotamine tartrate for complete relief at 30 min. Doleron novum® was significantly better than ergotamine at controlling nausea and vomiting.

The orally administered opiate compounds examined in this report were associated with only slightly higher rates of adverse events than was placebo, and were comparable to aspirin and better than ergotamine in this respect. Adverse events were much more frequently reported with intranasal butorphanol than with placebo or with oral opiate analgesics.

Isometheptene/Midrin®

The literature review identified six controlled trials of isometheptene and Midrin®/Midrid®. Our analysis of those trials yielded the following results:

Placebo-controlled trials of isometheptene and Midrin®/Midrid® suggest that both agents have a modest effect in relieving headache. In two placebo-controlled trials of isometheptene, the drug attained borderline significance for headache relief. Two of three placebo-controlled trials of Midrin®/Midrid® showed modest, but statistically significant headache relief, and the third was compatible with a modest effect (odds ratio favoring Midrin® that is non-significant).

The single study comparing Midrin® with isometheptene alone failed to detect a significant difference in headache relief. A separate study compared Midrin® with another of its constituents, acetaminophen; this study, too, failed to demonstrate a significant benefit to the combination drug.

No firm conclusions could be drawn from the only trial comparing Midrid® and ergotamine. However, another trial demonstrated that Midrid® was statistically superior to the combination of ergotamine tartrate + caffeine in its ability to decrease headache intensity. Moreover, Midrid® was associated with significantly less nausea and vomiting than ergotamine + caffeine.

Adverse events associated with isometheptene and Midrin®/Midrid® were mild, transient, and not significantly more frequent than with placebo or with the comparator drugs considered in this report.

Domperidone taken during the migraine prodrome

Domperidone (not commercially available in the US) was shown in two small trials to be effective for aborting or preventing migraine attacks in patients with migraine with aura when taken at the onset of premonitory symptoms. One trial showed a clinically important difference compared with placebo while another showed evidence of a dose-response relationship. Neither study observed any adverse events.

FUTURE RESEARCH NEEDS

Further research is required into the safety and efficacy of currently available self-administered drugs if their use for the short-term treatment of acute migraine is to be optimized. The following recommendations may be made:

Conduct and reporting of trials

  • (1) The generalizability of the results of trials conducted among patients in headache specialty centers to a primary care population is uncertain. More trials should be conducted among patients recruited from general practice settings.
  • (2) The diagnosis of migraine -- even when made according to specific criteria such as the IHS criteria for migraine with aura and migraine without aura -- encompasses a wide range of symptomatology. Researchers should be as precise as possible in describing any operational inclusion or exclusion criteria they employ in addition to headache diagnosis, such as headache frequency, severity, and chronicity.
  • (3) Similarly, future studies should indicate whether patients were allowed to take migraine prophylactic medication during the trial.
  • (4) More head-to-head comparisons of acute migraine treatments should be performed in order to help clinicians and patients make informed choices among the many available therapies. Such comparative trials are particularly important for older drugs and drugs approved for treating pain in general.
  • (5) Future trials should use common scales for measuring pain outcomes, if possible. The IHS recommends the use of a four-point verbal scale or VAS to measure headache severity, and the use of the number of attacks resolved within two hours as the primary measure of efficacy (International Headache Society Committee on Clinical Trials in Migraine, 1991). The consistent adoption of these recommendations in trials of acute migraine drugs would greatly facilitate future meta-analyses.
  • (6) Future trials should measure pain relief at several different times post-intervention (e.g., 30 minutes, and 1, 2, 4 and 24 hours), especially when the agents being compared have different speeds of onset.
  • (7) Future trials should expand the scope of the clinically relevant outcomes measured. Obtaining data on 24-hour headache outcomes will improve the understanding of headache recurrence, while even longer-term outcomes, such as measuring a headache index over several weeks, can assess the impact of an acute drug treatment strategy on the disease-specific measures (e.g., headache index, frequency, severity) and overall quality of life.
  • (8) Adverse events should be reported for all patients taking study medication, whether or not they completed the trial. The number of patients experiencing adverse events in each treatment group should be reported, and the specific adverse events reported should be described.

New directions for research

  • (9) Treatment protocols should be devised and tested to achieve the dual goals of rapid and durable relief of headache by combining agents that differ in speed of onset and duration of response. For example, a fast-acting agent like subcutaneous sumatriptan might be effectively combined with a slower-acting NSAID.
  • (10) Analgesic overuse or “rebound” headache is recognized by most headache researchers and clinicians, and a given drug’s capacity to produce rebound headache is an important limiting factor in its overall utility. The rebound phenomenon has not been carefully studied, and the short-term trials reviewed in this report are ill-suited for this purpose. Future, longer-term trials should be designed to investigate the rebound potential of the various drugs reviewed in this report, and should seek to establish the dosage and frequency of administration that are likely to result in rebound.
  • (11) Adverse events associated with long-term use of some analgesics have been identified (e.g., renal failure associated with NSAIDs). The short-term clinical trials reviewed in this report do not effectively capture these adverse events. New approaches to ascertaining long-term or rare complications of treatment are needed.
  • (12) The real-life management of headache disorders often involves multiple simultaneous interventions including an acute drug treatment plan (which might include an initial and rescue medication), a preventive drug treatment plan, behavioral therapy or other self-management education, or skill training. Such multidisciplinary interventions, usually delivered in specialty clinics, have been reported to show dramatic response rates in uncontrolled studies. Testing these bundled multiple interventions in a prospective, controlled trial may provide more understanding of how to care for patients with chronic headache disorders than the short-term, single intervention acute drug studies reviewed here.

Further study of the agents reviewed in this report

NSAIDs

  • (13) Long-acting NSAIDs have been purported to be effective while withdrawing analgesics in suspected rebound headaches. A controlled trial of analgesic withdrawal, coupled with a long-acting NSAID, and compared to continuing suspected rebounding agent could provide support for rebound headache while suggesting an effective treatment. Such a trial would require particular attention to dropout rate in the analgesic withdrawal group.

Ergot alkaloids and DHE

  • (14) The development of an evidence base for the commonly used older drugs (e.g., ergotamine tartrate and ergotamine-containing compounds, isometheptene compounds, and some over-the-counter analgesics) should be a priority for future clinical research. Clinical trials of new drug treatments could go some way towards meeting this goal by including both comparator treatments and placebo groups.
  • (15) Further studies are needed to determine the effectiveness of ergotamine among patients with migraine diagnoses as defined by current diagnostic criteria. The older studies supporting efficacy of ergotamine used less specific criteria and their clinical utility is uncertain.
  • (16) The optimal dose of ergotamine is uncertain. Dose-ranging studies to determine the relative efficacy and risk of adverse effects at various dosage levels between 1 mg and 5 mg may further clarify the clinical utility of ergotamine.

Sumatriptan (sc, po, in)

  • (17) Further trials should be conducted using lower doses of oral sumatriptan (50 mg, 25 mg), which are commercially available in the US.
  • (18) The rate of headache recurrence is an important clinical question for all three formulations of sumatriptan. In order to illuminate the phenomenon of headache recurrence, future studies should agree on a single definition of recurrence, collect data on recurrence at 24 hours or beyond, and standardize the use of rescue medication and non-study medication in such a way that recurrence rates can be examined among patients who do and do not use such medication. Future dose-ranging studies should break down headache recurrence data by dose.
  • (19) Research protocols should be devised to test the efficacy of other anti-migraine agents (e.g., NSAIDs) for the treatment of headache recurring after treatment with sumatriptan.

Opiate analgesics

  • (20) Opiate analgesics are frequently used clinically as a rescue medication when other drugs fail to abort or alleviate migraine attacks. Trials designed to test the efficacy of opiate analgesics in this role should be conducted.
  • (21) Longer-term studies that examine the efficacy of headache management strategies incorporating opiate analgesics in terms of headache frequency, severity, and duration, and that also consider disability, should be performed. Such studies could examine the problems of analgesic “rebound” headache, dependence, tolerance, and drug-related adverse events that may arise with long-term use of opiate analgesics.