When to Stress Over Triptans: A Markov Analysis of Cardiovascular Risk in Migraine Treatment

Authors

  • Lori A. Orlando MD,

  • David B. Matchar MD, FACP


  • From the Duke University Medical Center and the Durham VA; and the Center for Clinical Health Policy Duke University and the Durham VA.

Address all correspondence to Lori A. Orlando, MD, First Union Tower St. 220, 2200 W. Main St., Durham, NC 27710.

Abstract

Background.—Migraines affect 10% of the U.S. population and the episodes are frequently associated with significant disability. Triptans, 5HT1 receptor agonists, can be highly effective in treating pain and reducing disability. However, reports of cardiac events associated with triptan ingestion have led to concerns about its use in the face of possible cardiac disease.

Objective.—Should a patient without known cardiovascular disease (CAD) and moderately severe to severe migraines undergo cardiovascular testing prior to the initiation of triptan therapy?

Design.—A Markov model of migraine and cardiac disease using DATA 4.0. Three strategies were compared: (1) use triptans without further evaluation (TREAT); (2) test, then treat if negative (TEST); and (3) avoid triptans (NOTRIPTAN). Triptans were prohibited if a cardiac event occurred.

Data.—Model inputs were derived from the literature and subjected to sensitivity analyses across all possible values.

Time Horizon.—Markov cycle is 1 week.

Outcomes.—The primary outcomes of interest were quality-adjusted life expectancy, in years (QALYs) and the impact of various cardiovascular risk levels on the preferred strategy.

Results.—For the base case results were TREAT 19.4 QALYs, TEST 19.2, NOTRIPTAN 19.1. When altering CAD probability: TREAT dominated from 0 to 87%, TEST 87% to 97%, and NOTRIPTAN above 97%. Results were robust during sensitivity analyses.

Conclusions.—This analysis suggests that even for individuals with a relatively high risk of CAD it is not beneficial to perform cardiac testing, nor to avoid triptans. The exact level of cardiac risk at which testing should be considered is probably at or above 87%.

Ancillary