Composite endpoints for detecting meaningful changes in trials of headache treatments

In their recent commentary ‘Detecting meaningful changes in trials of headache treatments: Which outcome measure is best?’, Loder and Rizzoli (1) note that measuring headache pain or other symptoms of migraine in isolation is potentially inadequate to summarize a patient’s experience with medication. In this regard, it is worth noting that, for registration of new migraine agents in the United States, the Food and Drug Administration (FDA) requires demonstration of pain freedom (or relief) and absence of each associated symptom (phonophobia, photophobia, and nausea)—that is, statistical significance on four separate endpoints. One approach that has been taken to address the issue raised by Loder and Rizzoli and that is also relevant to the FDA requirement is to combine responses on multiple endpoints in a single composite outcome measure. Rodgers et al. (2) recently evaluated ‘total migraine freedom’ as a potential primary endpoint to assess acute treatment in migraine. This composite measure is defined as pain freedom and absence of associated symptoms (photophobia, phonophobia and nausea). Using a single composite endpoint requiring patients be free of all four symptoms facilitates comparisons between treatments and helps limit problems in determining ‘meaningfulness’ if, for instance, one drug is better at alleviating nausea and another better at alleviating photophobia. As demonstrated by Rodgers et al., the composite endpoint as well as being clinically meaningful is statistically more efficient (requires a smaller sample size) compared with the approach of having four separate co-primary endpoints.

Loder and Rizzoli further comment that measures based on efficacy do not incorporate information about adverse events that can be important contributors to the patient’s overall experience of treatment. One approach to address this is to develop composite measures that incorporate efficacy and tolerability. For example, Dodick et al. (3) have proposed ‘sustained pain freedom over 24 hours with no adverse events’ as a composite measure of efficacy and tolerability that captures key elements of treatment effects relevant to patients. The usefulness of this endpoint was investigated in a post hoc analysis of data from a trial that compared two acute migraine treatments with different mechanisms of action—a triptan and a calcitonin gene-related peptide (CGRP) receptor antagonist (4). While the two agents appeared similar on efficacy measures, the composite efficacy plus tolerability measure suggested an overall advantage for the CGRP receptor antagonist.

Certainly, composite measures are not without limitations but they warrant consideration in the discussion of which outcome measure is best for detecting meaningful changes in trials of headache treatments.