Toward Evidence-Based Management of Migraine

Migraine has long been recognized as a painful and debilitating disorder that profoundly affects the lives of individuals and their families. Population studies conducted over the past decade paint a startling picture of migraine as a major public health problem, suggesting that nearly 1 in 5 women and 1 in 20 men have disabling migraine.^{1 - 2} While precise estimates vary, migraine is clearly responsible for millions of bedridden and reduced-activity days per year and billions of dollars in medical costs and lost productivity.^{2 - 4} Yet most people with migraine are never diagnosed by a physician or treated with prescription medication,^{1 ,5 - 6} and those who do receive care are frequently dissatisfied with the results.⁷

In this context, in 1994, the Agency for Health Care Policy and Research (now the Agency for Healthcare Research and Quality [AHRQ]) initiated a project to evaluate the clinical trial evidence regarding headache management, with a focus on migraine. The resulting technical reports^{8 - 12} and subsequent ancillary reports sponsored by the American Academy of Neurology Education and Research Foundation (AAN-ERF)^{13 - 16} described the evidence on a wide range of management issues in migraine, including acute treatment. The reports on drug treatments for acute migraine^{11 - 12} described hundreds of clinical trials that provided substantial evidence for the efficacy of various agents. However, it is difficult to translate this evidence into clinical practice for several reasons: First, most trials compare an active drug with a placebo and thus are of limited value for informing the choice among available agents; second, most trials report results within single migraine attacks over just a few hours; and third, few trials test the effectiveness of a treatment plan that combines acute and rescue drug treatments or the effectiveness of an algorithmic approach to drug treatment selection based on prior response. These features limit the ease with which the available efficacy studies may be translated into guidelines for effective clinical practice.

These limitations were identified and addressed in 1999 when the US Headache Consortium (USHC), a collaboration of 7 professional organizations (American Academy of Family Physicians, American Academy of Neurology, American Association for the Study of Headache, American College of Emergency Physicians, American College of Physicians, American Osteopathic Association, and National Headache Foundation), developed guidelines for the treatment of migraine^{17 - 22} based on the AHRQ and AAN-ERF technical reports. To make the guidelines more useful practically, the USHC elected to make recommendations not only about specific agents that had been shown to be efficacious but also about the selection and sequencing of drug treatments and the design of an overall treatment plan for individual patients. One issue of particular importance was the role of the step care strategy for acute migraine.

For acute migraine, step care means that individuals with acute attacks are initially treated with the safest, least expensive therapies and progress to the more expensive migraine-specific medications, such as triptans, only when the initial treatment fails. In its most restrictive form, step care applies to each migraine attack, with the result that the use of expensive migraine-specific therapies is always delayed. Is this restrictive form of step care a prudent strategy for minimizing the inappropriate use of triptans and containing health care costs? Or will such a blanket approach lead to prolonged morbidity with little, if any, cost saving? Could such an approach cause secondary harm (eg, due to the patient lapsing from treatment)? What are the alternatives to step care?

The USHC was committed to addressing these and other questions concerning the selection and sequencing of treatments for acute migraine, but at the time it convened, there was no research deemed sufficient to support an evidence-based (ie, grade A or B) recommendation. In arriving at an opinion-based (ie, grade C) recommendation, the USHC relied on general humanistic principles that favor minimizing patient morbidity and on the USHC representatives' collective experience, which suggested that earlier use of migraine-specific therapy was likely to decrease the time a patient experiences pain. By unanimous agreement, the USHC rejected the step care strategy in its most restrictive form. Instead, the USHC recommended the use of migraine-specific therapy for moderate to severe migraine once the physician had established that first-line medications were not effective for a given individual. To minimize the risk of a patient lapsing from treatment, it was further recommended that this approach should be linked to a comprehensive treatment plan to be developed in consultation with the patient.

The study by Lipton and colleagues,²³ in this issue of THE JOURNAL, provides the first controlled trial evidence on the clinical effects of a strict form of step care (which the investigators refer to as "step care within attacks"). This strategy is compared with 2 alternatives: step care across attacks, in which patients who do not respond to first-line medications (here, aspirin plus metoclopramide) in at least 2 of the first 3 attacks are switched to zolmitriptan for attacks 4 to 6; and stratified care, in which treatment is assigned on the basis of the pretrial severity of migraine-related disability, with aspirin plus metoclopramide for patients with little or infrequent disability, and zolmitriptan for those with moderate or severe disability. The investigators found that both stratified care and step care across attacks led to better outcomes than did step care within attacks. There were no significant differences between stratified care and step care across attacks once those patients in the latter group who did not respond to first-line medication were escalated to a migraine-specific agent.

What are the implications of this study for the treatment of patients with migraine? In our judgment, the findings of Lipton and colleagues are wholly consistent with the recommendations of the USHC. First, their study provides convincing evidence that a strict form of step care (step care within attacks) is inferior to both stratified care and step care across attacks. This supports the USHC's rejection of the strict form of step care as a clinical policy. Second, the study showed that delaying the use of zolmitriptan until the fourth attack did not diminish the responsiveness of patients to the drug, nor did this approach lead to significantly higher numbers of dropouts than were observed with the stratified approach. Although delaying the use of migraine-specific agents in such circumstances means that some patients will experience avoidable pain during the short trial period, the study by Lipton et al suggests that the step care across attacks approach may not be an unreasonable strategy for minimizing the unnecessary long-term use of these expensive and not entirely benign medications.

This study, while of major clinical significance, leaves several important questions unanswered. For example, do the results apply to other triptans and to other migraine-specific agents, such as dihydroergotamine? How effective would it be to stratify treatment based on the severity of individual attacks, treating mild attacks with first-line therapy, and moderate to severe attacks with migraine-specific medication? From a slightly broader perspective, how might the implementation of evidence-based treatment guidelines—perhaps through the development of generally practical headache management programs—impact the care of patients with migraine? In this larger context, what are the cost and cost-effectiveness implications of alternative medication sequencing strategies?

Choices about the selection and sequence of drug treatments are at the core of everyday clinical care of patients with migraine but have been largely ignored in clinical research to date. The study of real-world interventions is difficult because these interventions are more complex than those tested in traditional drug trials, larger sample sizes are required, and funding may be more difficult to obtain. Lipton et al are to be commended for recognizing the need for this type of research, for addressing the design challenges in a rigorous fashion, and for significantly strengthening the evidence base for clinical care of patients with migraine. The next step is to translate this evidence into practice.