Consensus Statement: Cardiovascular Safety Profile of Triptans (5-HT_1B/1D Agonists) in the Acute Treatment of Migraine

Abstract

Background.—Health care providers frequently cite concerns about cardiovascular safety of the triptans as a barrier to their use. In 2002, the American Headache Society convened the Triptan Cardiovascular Safety Expert Panel to evaluate the evidence on triptan-associated cardiovascular risk and to formulate consensus recommendations for making informed decisions for their use in patients with migraine.

Objective.—To summarize the evidence reviewed by the Triptan Cardiovascular Safety Expert Panel and their recommendations for the use of triptans in clinical practice.

Participants.—The Triptan Cardiovascular Safety Expert Panel was composed of a multidisciplinary group of experts in neurology, primary care, cardiology, pharmacology, women's health, and epidemiology.

Evidence and Consensus Process.—An exhaustive search of the relevant published literature was reviewed by each panel member in preparation for an open roundtable meeting. Pertinent issues (eg, cardiovascular pharmacology of triptans, epidemiology of cardiovascular disease, cardiovascular risk assessment, migraine) were presented as a prelude to group discussion and formulation of consensus conclusions and recommendations. Follow-up meetings were held by telephone.

Conclusions.—(1) Most of the data on triptans are derived from patients without known coronary artery disease. (2) Chest symptoms occurring during use of triptans are generally nonserious and are not explained by ischemia. (3) The incidence of serious cardiovascular events with triptans in both clinical trials and clinical practice appears to be extremely low. (4) The cardiovascular risk-benefit profile of triptans favors their use in the absence of contraindications.

BACKGROUND AND OBJECTIVE

Migraine, which affects approximately 1 in 10 individuals,¹ is pervasive and often debilitating.² It is a brain disorder that manifests as attacks of often unilateral throbbing headache lasting from 4 hours to 3 days with associated symptoms such as nausea; sensitivity to light, sound, and head movement; and focal neurologic symptoms or signs.³ Migraine restricts or prevents normal activities, and patients are often bedridden until symptoms subside. In the 1999 US population-based American Migraine Study II,^1,4 91% of migraineurs indicated that severe headaches cause at least some disability; and more than half (53%) reported substantial impairment in activity or the need for bed rest with severe headaches.

The US Headache Consortium recommends use of prescription medicines for patients whose migraines cause moderate or severe pain.⁵ Triptans, selective 5-HT_1B/1D agonists, constitute the mainstay of migraine-specific prescription therapy in the United States, where 7 triptans are currently marketed (almotriptan, eletriptan, frovatriptan, naratriptan, rizatriptan, sumatriptan, and zolmitriptan). The efficacy of triptans in the acute treatment of migraine is well established in controlled trials.⁶ The US Headache Consortium's recommendations and the availability of effective pharmacotherapy notwithstanding, fewer than half of migraineurs in the United States receive prescription medicines for headache, and less than one third of patients are very satisfied with their acute therapy.⁷

Research suggests that concerns about cardiovascular safety limit the use of triptans.⁸ Two large studies evaluated the association between adverse vascular events in migraine patients using triptans.⁹ These retrospective surveys utilized two large patient databases in the United States (United Healthcare/Ingenix Research Database n=130 411, migraine cohort) and United Kingdom (General Practice Research Database n = 63 575, migraine cohort) with age- and sex-matched control populations. Information regarding diagnoses, medical services, and prescriptions was available from each of these databases. In neither study was there evidence of an increased risk of adverse vascular events or death in migraine patients using triptans. Therefore, cardiovascular risks should be viewed in the context of their low probability and in the context of the benefits of treatment. In 2002, the American Headache Society convened The Triptan Cardiovascular Safety Expert Panel to evaluate the scientific and clinical data on triptan-associated cardiovascular risk and to formulate consensus recommendations to guide health care providers in making informed decisions for their use in patients with migraine. This article summarizes the evidence reviewed by the panel, their conclusions about the evidence, and their recommendations for the use of triptans in clinical practice.

METHODS

The Triptan Cardiovascular Safety Expert Panel was composed of a multidisciplinary group of experts in neurology, primary care, cardiology, pharmacology, women's health, and epidemiology. The panel held one open meeting during which several panel members presented pertinent issues relevant to their area of expertise (eg, cardiovascular pharmacology of triptans, epidemiology of cardiovascular disease, cardiovascular risk assessment, migraine) as a prelude to group discussion and formulation of consensus conclusions and recommendations. Follow-up meetings were held by telephone.

The panel identified 3 questions to be answered in developing consensus views on the cardiovascular safety of triptans and their use in clinical practice: (1) What are the incidence and characteristics of serious and nonserious cardiovascular events associated with the use of triptans? (2) What is the mechanism of cardiovascular events associated with the use of triptans? (3) How should triptans be used in clinical practice in view of currently available evidence regarding their risks and benefits?

Information reviewed by the panel members in attempting to answer these questions included (1) epidemiologic data on cardiovascular disease and migraine, (2) clinical trials data on cardiovascular adverse events with triptans, (3) postmarketing surveillance data on spontaneous reports of cardiovascular adverse events with triptans, (4) pharmacologic and pharmacodynamic studies relevant to cardiovascular safety of triptans, and (5) methods of evidence-based clinical assessment of cardiovascular risk. The following sections summarize key evidence discussed by the panel in establishing a consensus view on each of the 3 questions above. Where relevant, the sections describe, in boxed text, the panel's consensus view of the evidence rated according to an evidence classification scheme (Table 1) adapted from one developed by the American Academy of Neurology.¹⁰ The panel limited their inquiry and conclusions to cardiovascular safety and did not systematically consider cerebrovascular safety, which has been discussed elsewhere.^11,12

Table 1. —Evidence and Recommendation Classification Scheme*

Rating of Evidence
Class I: Evidence provided by a prospective, blinded study of a broad spectrum of persons who may be at risk for developing an adverse cardiovascular outcome. The study measures the predictive ability of a risk factor (ie, triptan use) using an independent criterion standard for case definition.
Class II: Evidence provided by a prospective study of a narrow spectrum of persons at risk for an adverse cardiovascular outcome or by a retrospective study of a broad spectrum of persons at risk for an adverse cardiovascular outcome compared with a broad spectrum of controls. The study measures the prognostic accuracy of the risk factor (ie, triptan use) using an acceptable independent criterion standard for case definition. The risk factor is measured in an evaluation that is masked to outcome.
Class III: Evidence provided by a retrospective study where either the persons at risk for an adverse cardiovascular outcome or the controls are of a narrow spectrum. The study measures the predictive ability using an acceptable independent criterion standard for case definition. The risk factor (ie, triptan use) is measured in an evaluation that is masked to the outcome.
Class IV: Any design where the predictor is not applied in a masked evaluation or evidence provided by expert opinion or case series without controls.
Rating of Recommendation or Conclusions
Level A: Evidence establishes an association or lack of association between triptan use and an adverse cardiovascular outcome in migraineurs. Requires at least 1 convincing class I study or at least 2 consistent, convincing class II studies.
Level B: Evidence indicates that there is probably an association or lack of association between triptan use and an adverse cardiovascular outcome in migraineurs. Requires at least 1 convincing class II study or at least 3 consistent class III studies.
Level C: Evidence indicates that there is possibly an association or lack of association between triptan use and an adverse cardiovascular outcome in migraineurs. Requires at least 2 convincing and consistent class III studies.
Level U: Data are inadequate or conflicting. Given current knowledge, it is not possible to establish whether or not an association between triptan use and an adverse cardiovascular outcome in migraineurs exists.

• *Adapted from Practice Guidelines: Official AAN Practice Statements.¹⁰

SYNOPSIS OF THE EVIDENCE

Incidence and Characteristics of Triptan-Associated Serious and Nonserious Cardiovascular Events.— The panel evaluated data from both clinical trials and postmarketing surveillance (which includes case reports).

Adverse Events in Placebo-Controlled Clinical Trials.— Chest tightness, heaviness, pain, or pressure were reported in approximately 1% to 7% of patients taking therapeutic doses of triptan tablets in placebo-controlled clinical trials.^13–24 These events uniformly occurred at a frequency either comparable to or slightly (ie, 1% to 2.5% absolute percentage) higher than that with placebo regardless of the triptan. Patients reporting chest symptoms often experienced similar sensations in other body areas, such as the face and limbs. Triptan-associated chest symptoms were typically mild and transient. In one open-label almotriptan clinical trial, a possible ischemic event was reported after the use of almotriptan, but no angiographic evidence of ischemia was apparent.²⁵

These data should be interpreted in view of characteristics of the patient population in migraine clinical trials. Generally, controlled clinical trials with triptans excluded patients with cardiovascular risk factors including known ischemic heart disease, symptoms or signs consistent with ischemic heart disease, cardiac arrhythmias requiring medication, and supine diastolic blood pressure >95 mm Hg and/or systolic blood pressure >160 mm Hg. Thus, the clinical trials data cannot be generalized to migraine sufferers with cardiovascular risk factors.

Triptans are associated with a modestly elevated incidence of chest symptoms (ie, triptan sensations) relative to placebo in well-controlled clinical trials that excluded patients with significant cardiac risk factors or known ischemic heart disease. The chest symptoms in clinical trials were generally transient, mild, and nonserious. Class II evidence, Level A conclusion.

Long-term Open-Label Studies.— In addition to these placebo-controlled trials, which were primarily conducted to assess triptan efficacy, several long-term studies have been conducted. The cardiovascular safety profile of triptans was studied in 1- to 2-year, open-label studies in which 12 339 migraineurs treated an average of 1.28 migraines per patient per month with sumatriptan injection 6 mg,²⁶ 275 migraineurs treated 11 501 attacks with sumatriptan tablets 100 mg,²⁷ 496 migraineurs treated 14 373 migraines with frovatriptan 2.5 mg,¹⁹ 25 501 patients treated 70 537 migraines with rizatriptan 10 mg,²⁸ 2058 patients treated 31 579 migraines with zolmitriptan 5 mg,²⁹ 417 patients treated 15 301 migraines with naratriptan 2.5 mg,³⁰ and 762 patients treated 13 751 migraines with almotriptan 12.5 mg.²⁵ (One-year data for eletriptan have not been published at the time of this writing.) No serious cardiovascular adverse events (as determined by study investigators) or cardiovascular deaths were attributed to triptan use. Chest symptoms (usually pain or pressure) were reported by approximately 3% to 10% of patients in each study. These studies excluded patients with known or suspected ischemic heart disease, and the rizatriptan study excluded patients at risk of having undiagnosed heart disease (ie, patients with diabetes, smokers, postmenopausal women, or those with a family history of heart disease).

Chest symptoms sometimes reported after use of triptans were not associated with serious cardiovascular outcomes in long-term, open-label clinical trials that excluded patients with known ischemic heart disease. Class IV evidence, Level U conclusion

Study in Patients With Known or Suspected Coronary Artery Disease.— The clinical studies described above excluded patients with known or suspected coronary artery disease. To evaluate triptan safety in patients at higher risk for serious cardiovascular events, frovatriptan 2.5 mg was administered to 75 patients with documented coronary artery disease or a Framingham Coronary Artery Disease Risk Prediction Score of at least 14 (http://www.framingham.com) in a randomized, double-blind, parallel-group, placebo-controlled, multicenter study.¹⁹ The incidences of ischemic/arrhythmic episodes as assessed by 24-hour Holter electrocardiogram (ECG) monitoring; clinically meaningful ECG changes measured 2, 4, 6, and 24 hours postdose; change in blood pressure; and change in troponin T levels did not differ between the frovatriptan group and the placebo group (with the exception of a higher incidence of ECG abnormalities with placebo compared with frovatriptan 4 hours postdose), but the mean number of ischemic/arrhythmic episodes per patient was greater with frovatriptan. None of the patients in the frovatriptan group reported treatment-emergent chest pain or palpitations. It is impossible to draw conclusions from this study, which was underpowered to assess the incidence of ECG changes and cardiac adverse events in patients given triptans.

Data Reported During Postmarketing Surveillance.— The panel used the US Food and Drug Administration's (FDA's) Adverse Event Reporting System in the Office of Postmarketing Drug Risk Assessment as the source of postmarketing surveillance data for their assessment of triptan safety.³¹ The Adverse Event Reporting System is a computerized database comprising adverse events reported by health care professionals, pharmaceutical companies, and consumers after a drug is marketed. Postmarketing surveillance data are derived from several sources including spontaneous reports from clinical practice, case reports in the published literature, and postmarketing clinical studies.

Postmarketing surveillance data can provide important information about the occurrence of rare adverse events that go undetected in clinical trials, or the adverse events that occur with a medicine as it is used in clinical practice, outside the strict confines of a controlled clinical trial. Postmarketing surveillance data should be interpreted, however, in the context of their limitations (Table 2). For example, adverse event reports derived from clinical practice are often incomplete and lack information crucial for attempting to determine whether a medication caused the adverse event. Other limitations include the absence of a control group that is not exposed to medication and the fact that adverse events, reports of which are not actively solicited, are typically underreported.

Table 2. —Strengths and Limitations of Postmarketing Surveillance Data

Strengths
Large number of patient exposures facilitates detection of extremely rare events
Provides indication of adverse events occurring during actual clinical use
Limitations
Inability to calculate a precise incidence of adverse events because of
• lack of information about total number of exposures to a drug
• lack of information about total number of patients experiencing adverse events with drug
Limited ability to establish cause of adverse event because of incomplete reporting

Perhaps most importantly, postmarketing surveillance data cannot be used to calculate incidence rates of adverse events because of lack of accurate information about both the total number of patient exposures to a drug and the total number of patients experiencing a particular adverse event.

The panel reviewed, for each marketed triptan, all adverse event reports received by the FDA from November 1, 1997, through February 28, 2002, that described a serious or nonserious adverse event possibly referable to the cardiovascular system regardless of the suspected cause of the event.

Almotriptan tablets were introduced in the United States in June 2001. During postmarketing surveillance between June 2001 and February 2002, 2 adverse events affecting any body system were reported. Both of these adverse events involved the cardiovascular system. Neither was fatal.

Naratriptan tablets were introduced in the United States during the first quarter of 1998. During postmarketing surveillance between the first quarter of 1998 and February 2002, 227 adverse experiences affecting any body system were reported. Of these 227 adverse experiences, 26 involved at least one cardiovascular adverse event. None of these events were fatal.

Rizatriptan, available in an oral tablet and oral wafer form, has been marketed in the United States since late 1998. During postmarketing surveillance between late 1998 and February 2002, 472 adverse experiences affecting any body system were reported. Of these 472 adverse experiences, 80 involved at least one cardiovascular adverse event. Two of the events were fatal.

Sumatriptan, available in oral tablet, subcutaneous injection, and intranasal spray forms, has been marketed in the United States since 1992. The postmarketing surveillance data for sumatriptan reflect the period between November 1, 1997, and February 2002, during which time 1729 adverse experiences affecting any body system were reported. Of these, 134 (including 11 fatalities) involved at least one cardiovascular adverse event with sumatriptan injection; 121 (including 8 fatalities) involved at least one cardiovascular adverse event with sumatriptan tablets; 20 (including no fatalities) involved at least one cardiovascular adverse event with sumatriptan nasal spray; and 40 (including 6 fatalities) involved at least one cardiovascular adverse event with an unknown (ie, unrecorded) sumatriptan formulation.

Zolmitriptan, available in an oral tablet and oral rapid-melt form, has been marketed in the United States since late 1997. During postmarketing surveillance between late 1997 and February 2002, 719 adverse experiences affecting any body system were reported. Of these 719 adverse experiences, 138 involved at least one cardiovascular adverse event, 8 of which were fatal.

The panel noted numerous deficiencies in information reported to the FDA that render these data difficult to interpret. For example, the time between dosing with the triptan and the adverse event was often not reported; therefore, inferences regarding a causal relationship between the drug and the adverse event cannot be made. Furthermore, reports often did not indicate the outcome of an event (eg, death, hospitalization). Finally, the information in the FDA reports could not be medically verified in the majority of cases. In some instances, however, the timing between triptan administration and occurrence of a serious cardiovascular event was consistent with a causal role of the triptan.

Given the widespread use of triptans, the risk of serious cardiovascular adverse events during postmarketing surveillance appears to be very low. While the risk of a serious cardiovascular event during triptan use appears to be very small, it cannot be dismissed. Serious cardiovascular events, some of which resulted in death, have been reported in association with triptans during postmarketing surveillance. The causal association of triptan use with serious cardiovascular adverse events is difficult to determine based on the postmarketing surveillance data alone. Class IV evidence, Level U conclusion

Data From Prescribing Information.— The Complete Prescribing Information for triptans can include data that are not captured in published reports of clinical studies or that are described incompletely in postmarketing surveillance records. The Complete Prescribing Information for all triptans, except almotriptan and naratriptan, indicates that isolated serious cardiovascular events including myocardial infarction have been reported within a few hours of triptan administration.^19,32–36 In some cases, these events were observed in patients without risk factors for coronary artery disease or in patients with documented absence of coronary artery disease. Considering the extent of use of the triptans, the incidence of serious cardiovascular events is extremely low.

Pharmacologic and Pharmacodynamic Studies.— A limited series of pharmacologic and pharmacodynamic studies has been conducted to elucidate possible mechanisms of triptan-associated chest symptoms and other cardiovascular adverse events. The panel discussed this evidence with particular attention to studies assessing a possible vascular mechanism.

Receptor Pharmacology.— Triptans are agonists at serotonin (5-HT) receptors, numerous classes and subtypes of which have been identified (eg, 5-HT₁, 5-HT₂, 5-HT₃, 5-HT₄, 5-HT₅, 5-HT₆, 5-HT₇ classes; 5-HT₁ receptor subtypes: 5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1E, 5-HT_1F). Selective for the 5-HT₁ class of receptors—specifically, for 5-HT_1B, 5-HT_1D, and 5-HT_1F—triptans minimally affect other 5-HT receptor classes.³⁷ Furthermore, triptans have low or no affinity for other families of receptors including α- or β-adrenergic, cholinergic, or dopaminergic receptors.

Coronary arteries contain 5-HT receptors, particularly the 5-HT_1B and 5-HT₂ subtypes. Serotonin (the endogenous ligand for serotonin receptors) contracts human coronary arteries primarily by stimulating the 5-HT₂ receptor; triptans have low or no affinity for 5-HT₂ receptors.

Studies in Human Isolated Coronary Artery.— Triptans can contract human isolated coronary arteries primarily via a 5-HT₁-dependent mechanism with variable participation of the 5-HT_2A receptor and to a much smaller degree than does serotonin.^37–40 In early in vitro studies, the magnitude of human coronary-artery contraction produced by sumatriptan was one fifth that produced by serotonin.^37,38 While serotonin-induced contractions were inhibited by administration of a 5-HT₂ antagonist, sumatriptan-induced contractions were not. Like sumatriptan, other triptans, including rizatriptan, frovatriptan, and eletriptan, have been shown in studies conducted in separate laboratories to cause small contractions of human isolated coronary arteries.^40–44

The triptans do not appear to differ from one another in their ability to contract human isolated coronary arteries, a finding consistent with their similar pharmacologic profiles. In an important study directly comparing the effects of several triptans in the same in vitro preparation, sumatriptan, zolmitriptan, rizatriptan, and naratriptan produced comparable, small contractions of isolated coronary arteries of organ donors dying of noncardiac causes.⁴⁵ For each triptan, the clinically effective maximum plasma concentration was less than 40% of that required to evoke half of the maximal coronary artery contraction elicited by the drugs. This finding suggests that, at plasma concentrations achieved in ordinary clinical use, these 5-HT_1B/1D agonists are associated with minimal or no contraction of nondiseased coronary arteries. These findings were later extended to eletriptan.⁴⁰

While the effects of the triptans on isolated nondiseased human coronary arteries are well defined, their effects on diseased human coronary arteries have not been systematically studied. In one study, decreasing endothelial integrity did not increase the contractility of the human isolated coronary artery to sumatriptan.⁴⁶

Angiography.— In vivo data are consistent with the studies in isolated coronary arteries. In angiographic studies undertaken in migraineurs with less than 50% coronary artery stenosis, sumatriptan (2 studies),^47,48 naratriptan (1 study),⁴⁹ and eletriptan (1 study)⁵⁰ administered at subcutaneous or intravenous doses that produced maximum plasma concentrations equaling or exceeding those produced by therapeutic tablet doses did not clinically significantly reduce mean coronary artery diameter. Rather, triptans variably affected coronary arteries. Often, coronary artery diameter did not change, whereas on other occasions, small increases or decreases were observed. With the exception of chest pain coinciding with possible catheter-induced proximal right segmental coronary artery spasm in the eletriptan study,⁵⁰ no significant changes in ECG recordings and no chest symptoms were reported in any of the studies.

The angiographic effects of sumatriptan injection 6 mg in patients with coronary artery disease, as reflected by at least 50% stenosis in at least one coronary vessel segment, were assessed in a placebo-controlled, parallel-group study.²⁷ No statistically significant differences between the sumatriptan group (n = 11) and the placebo group (n = 5) were observed for percentage change from baseline at 15 minutes or 30 minutes postdose in mean, minimum, or maximum diameters in either stenosed or nonstenosed segments.

A recent study evaluated the coronary vasoconstrictive effect of intravenous eletriptan at supratherapeutic plasma concentrations (three times the C_max of the 80-mg oral dose) and subcutaneous sumatriptan 6 mg versus placebo.⁵¹ The study population involved 60 patients with no clinically or angiographically significant coronary artery disease (<20% focal stenosis and without diffuse irregularities) who were undergoing a diagnostic coronary angiogram. The maximum mean change in coronary artery diameter for eletriptan was −22% (95% CI: −26%, −19%), for sumatriptan was −19% (95% CI: −22%, −16%), and for placebo was −16% (95% CI: − 20%, −12%). These results demonstrate that in patients with normal coronary arteries, eletriptan, administered at plasma concentrations greater than three times the C_max of an oral 80-mg dose, results in only a mild and clinically insignificant degree of coronary vasoconstriction, similar to that in patients receiving a standard dose of subcutaneous sumatriptan or placebo.

These prospective angiographic studies are supplemented by the results of a case report of the effects of sumatriptan on the coronary arteries of a 53-year-old woman with a 35-year history of migraine.⁵² The patient typically experienced mild chest symptoms with use of sumatriptan injection. Unknown to physicians at the time they performed diagnostic angiography in this woman, she had injected herself with sumatriptan 6 mg approximately 30 minutes before the procedure. Over a 15-minute observation period, the coronary angiogram showed no abnormalities despite the mild chest symptoms that followed her injection.

In vitro and in vivo studies of the human coronary artery show that triptans had small, variable effects on human coronary arteries. Triptans were comparable in their effects on human coronary arteries. When small contractions did occur, clinical signs or symptoms of ischemia did not accompany them. The balance of information supports the conclusion that, at therapeutically relevant plasma concentrations, triptans have minimal effects on coronary arteries. Class I through IV evidence, Level A conclusion.

Positron Emission Tomography.— Angiographic studies are consistent with the results of a positron emission tomography (PET) study.⁵³ Positron emission tomography scans of the heart were obtained before and 10 minutes after subcutaneous injection of either sumatriptan or placebo in 19 women, ages 33 to 62 years, at low risk of coronary artery disease, in a double-blind crossover study. Neither sumatriptan (6 mg subcutaneously) nor placebo significantly affected global or regional myocardial perfusion. The mean percentage change from baseline in global myocardial perfusion was +6.6% (standard deviation [SD], 18.8) for sumatriptan compared with +9.5% (SD, 18.0) for placebo. Therefore, myocardial perfusion does not appear to be affected adversely by sumatriptan.

Four of the women developed chest tightness (n = 1) or neck tightness (n = 3) after administration of sumatriptan. The reports of chest or neck tightness in the absence of an effect on myocardial perfusion suggest that ischemia does not account for the symptoms in these cases.

Rechallenge Study With Electrocardiographic Measurements.— Electrocardiograms were measured in a study of 20 patients who had previously experienced chest symptoms after administering intravenous, subcutaneous, or oral sumatriptan.²⁷ When these patients were not experiencing a migraine, they were rechallenged with a subcutaneous placebo injection that was followed 1 hour later with a subcutaneous injection of sumatriptan. Three patients (15%) experienced chest symptoms after administration of sumatriptan. Of these 3 patients, 2 also experienced symptoms after placebo. Electrocardiograms revealed no evidence of myocardial ischemia in any patient after sumatriptan or placebo regardless of whether chest symptoms were experienced. The absence of ECG changes in patients treated with sumatriptan who develop chest discomfort suggests that myocardial ischemia was not the cause of chest symptoms in these patients.

Data obtained using a variety of methodologies including PET, ECG, and angiography demonstrate that chest symptoms occurring after use of triptans in some patients are rarely accompanied by evidence of myocardial ischemia. These studies generally involved small numbers of patients who did not have coronary artery disease; therefore, the ability to generalize the results is limited. Class I through II evidence, Level A conclusion.

HYPOTHESIZED NONISCHEMIC MECHANISMS FOR CHEST SYMPTOMS

Several nonischemic mechanisms have been proposed for the chest symptoms that occur in some patients after use of triptans.

Generalized Vasospastic Disorder.— It has been hypothesized that patients with migraine have a generalized vasospastic disorder, 2 manifestations of which are variant angina and migraine.^54–55 This hypothesis is supported by results of a study showing that migraine occurred more often in patients with variant angina than in those without variant angina.⁵⁶ In a study of patients hospitalized in a coronary care unit, 26% of patients with variant angina also had migraine, whereas only 10% of matched controls with no known coronary artery disease and 6% of a control group of patients undergoing coronary arteriography because of suspected ischemic chest pain had migraine.⁵⁴ The authors noted that both variant angina and migraine are characterized by abnormal vascular tone and postulated that they share a common pathophysiology. Others have also observed that migraine and angina may co-occur regardless of whether migraines are medicated with serotonergic drugs,^55–57 but the association between migraine and angina is not consistently observed.⁵⁸

Abnormal Esophageal Motility.— Esophageal mechanisms also have been hypothesized to underlie the chest symptoms sometimes reported with triptans.⁵⁹ In support of this hypothesis, a supratherapeutic 16-mg subcutaneous injection of sumatriptan increased the amplitude and duration of esophageal contractions as well as abnormal esophageal motility, but did not cause electrocardiographic abnormalities in 24 healthy volunteers.⁵⁵ Changes in esophageal function after sumatriptan injection were observed in every subject in this study. Four (80%) of 5 subjects reporting chest pain after sumatriptan injection showed clinically abnormal esophageal motility, whereas only 8 (42%) of 19 subjects with no chest symptoms showed clinically abnormal motility.

Pulmonary Mechanisms.— Triptan-associated chest symptoms also have been hypothesized to arise from a direct effect on the pulmonary vasculature.⁶⁰ Little work has been conducted to test this hypothesis, although the demonstration that triptans affect pulmonary arterial pressure and the discovery of 5-HT_1B/1D receptors on human pulmonary arteries are consistent with this possibility.^61,62

Alterations in Skeletal Muscle Energy Metabolism.— Triptan-associated reductions in the oxygen stores of skeletal muscles may contribute to chest symptoms and similar side effects such as heaviness of the limbs.⁶³ In a study of the effects of sumatriptan on muscle energy metabolism during isometric exercise, subjects with no side effects after a 6-mg subcutaneous injection showed no differences from pre-exerecise to postexercise in skeletal muscle energy metabolism, while subjects experiencing side effects such as heaviness of the limbs showed a transient reduction in oxygen storage as measured by mitochondrial functioning after exercise.⁶³

Central Sensitization of Pain Pathways.— Heightened sensory sensitivity may explain the chest symptoms that occur during migraine in some patients. Research shows that heightened sensory sensitivity in the form of cutaneous allodynia (ie, experience of pain with nonpainful stimuli) spreads over the course of a migraine from a small localized cranial area ipsilateral to the headache to other body regions including the contralateral head and forearms.⁶⁴ A similar heightened sensitivity to pain has been observed in patients with noncardiac chest pain, which is similar to anginal pain but typically originates from the esophagus. In one study, patients with noncardiac chest pain compared with healthy volunteers had a lower baseline esophageal pain threshold and demonstrated more marked and prolonged reductions in upper esophageal pain threshold in response to acid infusion into the lower esophagus.⁶⁵

A lower pain threshold among migraineurs compared with nonmigraineurs may explain the finding that migraineurs were significantly more likely than nonmigraineurs to report chest pain in a large study conducted before triptans became available.⁶⁶ Two cohorts of patients—the first comprising 46 619 patients with a migraine diagnosis based on self-reported symptoms and the second comprising 32 669 patients with a migraine diagnosis based on self-reported physician diagnosis of migraine—who enrolled in the northern California Kaiser Permanente Medical Care Program in 1971 through the first half of 1973 were studied. Presence or absence of chest pain was assessed by asking patients at a health care visit whether they had experienced during the past year pain, pressure, or a tight feeling in their chest that either hurt in the middle under the breastbone or forced them to stop walking. Patients were followed to assess the occurrence of myocardial infarction through December 31, 1987, death, hospitalization for myocardial infarction, or termination of Kaiser Permanente membership, whichever occurred first. Results show that in both cohorts, chest symptoms were reported significantly more frequently by migraineurs than by enrollees without migraine (Table 3). Migraineurs were also significantly more likely than nonmigraineurs to report other symptoms, including stomach pain, heartburn, painful joints, and neck and back pain. Notably, however, the chest symptoms reported by migraineurs in this study did not predict the occurrence of myocardial infarction: migraineurs were no more likely to experience a myocardial infarction than were enrollees without migraine.

Table 3. —Percentage of Enrollees With Self-reported Chest Symptoms*

	Women		Men
	Migraine	No Migraine	Migraine	No Migraine
Cohort 1†	31.3‡	13.4	33.2‡	12.5
Cohort 2§	20.8‡	13.1	21.4‡	14.6

• *Data from Sternfeld et al.⁶⁶
• †In cohort 1 (N = 46 619), migraine status was defined by symptoms.
• ‡P < .01.
• §In cohort 2 (N = 32 669), migraine status was defined by self-reported physician diagnosis.

These data show a dissociation between chest symptoms and cardiovascular events in patients with migraine: chest symptoms but not myocardial infarction occurred at an excess frequency in migraineurs compared with nonmigraineurs. These findings are inconsistent with an ischemic mechanism for the chest symptoms sometimes reported by migraineurs. An important limitation of this study is that the diagnosis of migraine was not reliably validated, and the study was completed before the IHS diagnostic criteria were available.

Several nonischemic mechanisms for triptan-associated chest symptoms have been proposed. Although evidence to date is insufficient to provide definitive support for any one of these mechanisms, they warrant further exploration. Class I through IV evidence, Level U conclusion

APPROPRIATE USE OF TRIPTANS IN CLINICAL PRACTICE

In discussing application of the evidence to recommendations for the use of triptans in clinical practice, the panel concluded that there is insufficient evidence about the determinants of triptan-associated cardiovascular adverse events to support a definitive algorithmic approach to prescribing (or not prescribing) them. While the panel considered the determinants of cardiovascular adverse events with triptans to be poorly defined, they unanimously agreed that the risk of triptan-associated cardiovascular adverse events appears to be extremely low among patients meeting the typical inclusion and exclusion criteria for triptan clinical trials or given triptans in a manner consistent with the US Prescribing Information—that is, in patients without known coronary artery disease.

Importantly, a heightened risk of serious cardiovascular adverse events in patients with coronary artery disease did not constitute the reason that the panel confined their conclusions to those without known coronary artery disease. In fact, the panel cited a lack of compelling evidence that serious cardiovascular adverse events with triptans are more likely to occur in patients with coronary artery disease than in those without known coronary artery disease. Isolated, serious, cardiovascular adverse events have occurred in both groups of patients. The panel restricted their conclusions about the cardiovascular safety of triptans to those without coronary artery disease because the vast majority of data reviewed by the panel applied to this patient population. Those data show that, among patients without known or suspected coronary artery disease, the safety profile of triptans is well defined and appears to reflect a very low risk of serious cardiovascular adverse events.

CONCLUSIONS

Considering the pharmacologic, epidemiologic, and clinical evidence in aggregate, the panel arrived at the following conclusions: (1) Chest symptoms occurring during use of triptans are usually nonserious and usually not attributed to ischemia. (2) While serious cardiovascular adverse events have occurred after use of triptans, their incidence in both clinical trials and clinical practice appears to be extremely low. (3) The cardiovascular risk-benefit profile of triptans favors their use in the absence of contraindications. (4) Most clinical trials and clinical practice data on triptans are derived from patients without known coronary artery disease. These data support the conclusion that, in patients at low risk of coronary artery disease, triptans can be prescribed confidently without the need for prior cardiac status evaluation.

Acknowledgments

Acknowledgments:  The American Headache Society provided financial support for this initiative.