BMS-927711 for the acute treatment of migraine

Calcitonin gene-related peptide (CGRP) is a well-studied neuropeptide found at the very centers of the migraine processes, both centrally and peripherally (1,2). Its role in migraine pathophysiology was suggested more than 20 years ago (3,4) and, since then, our knowledge of the peptide has increased substantially, leading to a robust interest in targeting CGRP to treat migraine. At present, CGRP remains the most actively evaluated, and probably best validated target for currently in-development migraine medications (5). CGRP receptor antagonists disrupt the interaction of CGRP with its receptor and are being developed primarily for the acute treatment of migraine. Free CGRP and CGRP receptors can also be targeted using monoclonal antibodies which are being developed for the preventive treatment of episodic and chronic migraine (6).

Four distinct CGRP receptor antagonists (the ‘gepants’) had proof of efficacy for the acute treatment of migraine. Of them, olcegepant (BIBN4096BS) was discontinued because of difficulties in developing an oral formulation (7); telcagepant (MK-0974) was discontinued because of concerns of liver toxicity after frequent use (6,8); MK-3207, a molecule that was significantly more potent than telcagepant (9), was also discontinued because of concerns of liver toxicity (10); and BI44370A had efficacy demonstrated in a Phase 2 clinical study (11). In addition to demonstrating proof of efficacy, CGRP receptor antagonist clinical trials demonstrated the tolerability of the class with acute dosing and that, as opposed to triptans, their use is not associated with vasoconstriction (12–14).

A fifth ‘gepant’ just joined the club by its own merits. In the current issue of Cephalalgia, Marcus and colleagues report the result of a large Phase 2b study testing BMS-927711 for the acute treatment of migraine (15). The study used an adaptive design to test six doses of BMS-927711 (10, 25, 75, 150, 300 and 600 mg) against placebo. Sumatriptan 100 mg was used as an active comparator. Endpoints were those recommended by the International Headache Society (16). The primary efficacy endpoint was the proportion of pain-free subjects at 2 hours post dose. In addition to testing the primary endpoint for statistical significance compared with placebo, the authors attempted to define a measure of clinical relevance, or clinical response of at least 15% greater than the response of placebo.

Several of their findings are worth mentioning and commenting on. First and foremost: the drug was clearly effective at multiple doses, meaning that all CGRP receptor antagonists tested to date have demonstrated efficacy in Phase 2 and Phase 3 studies. For the primary endpoint (pain-free at 2 hours), doses of 75 mg, 150 mg and 300 mg (as well as sumatriptan) were superior to placebo. Efficacy was numerically inferior to sumatriptan for this endpoint for all doses (although the study was not powered for direct comparisons). The drug also passed the bar of the secondary endpoints where most (and sometimes all) doses were superior to placebo, and most effective doses were also numerically superior to sumatriptan. Accordingly, certain doses of BMS-927711 seem to deliver levels of efficacy that are similar to those delivered by the highest dose of an effective triptan. Overall, the efficacy of the drug increased up to the dose of 75 mg when a plateau effect seemed to have been reached for doses between 75 mg and 300 mg. The highest tested dose (600 mg) demonstrated no additional benefit over the doses of 75 mg and 150 mg.

Although the drug clearly achieved statistical significance, it did not achieve ‘clinical significance’ (response of at least 15% greater than the response of placebo) for any of the reported endpoints. This apparent discrepancy could be explained by the design of the study. Adaptive designs using Bayesian models allow better characterization of the dose response with optimal use of the sample size. However, random imbalances happening at the initial phases of a study may lead to hierarchical decisions that magnify these imbalances. Being very aware of this problem, the authors first used fixed randomization schedule to load a sizeable number of patients into each treatment arm before allowing group changes. Nonetheless, random biased response seemed to have occurred in the groups receiving 300 mg and 600 mg. As a consequence, the Bayesian algorithm smoothed the dose-response curve, resulting in fewer patients being allocated to the groups with better efficacy performance (75 mg and 100 mg). The process can be understood in the light of the stock market. On occasion, equity values inexplicably fluctuate more than expected, which triggers a cascade of computer-generated orders that, in turn, introduce a force that moves the equity price further. Accordingly, although adaptive designs distribute sample more efficiently (and are perfect for dose ranging studies), they do so at the cost of increasing the variance of response, which could explain the lack of ‘clinical significance’.

The most effective doses of the drug also significantly improved photophobia and phonophobia. The fact that statistical significance was not achieved for nausea only reflects the fact that Phase 2 studies are powered for the primary endpoint (headache), not associated symptoms.

The tolerability of BMS-927711 was reported as being placebo-like, which once more reinforces the tolerability of the class for acute dosing (17–19). As with the other CGRP receptor antagonists, a pattern of side effects could not be identified and this is different from what is seen for other classes such as ergot derivatives and triptans (where certain adverse events, such as chest tenderness and muscle tightness seem to be class-specific). No serious adverse event was reported in the trial.

What remains to be better described in future studies is the safety of the drug, especially in terms of liver toxicity, which, so far, has been the Achilles’ heel of the class, to the point that other strategies (using monoclonal antibodies targeting CGRP) are now being tested for prevention. Antibodies, not being metabolized by the liver, are less likely to cause drug-mediated liver toxicity (20). Although in the study by Marcus there were no overt signs of liver toxicity, this needs to be interpreted with tremendous caution. First, this was a single-dose study with only one treated attack per patient. In other studies liver toxicity was suggested after frequent use of medication, in paradigms that resemble how patients would use the medication in real life (several times per month, sometimes at non-recommended doses) (6,8). Second, to be in the study, patients could not be using drugs metabolized by CYP3A4. Both decisions (single dose and avoidance of drugs metabolized at the CYP3A4) were very prudent and certainly aimed to not expose patients to unnecessary risk in this phase of development. The downside is that, as the authors acknowledge, there is very little that can be said about the safety of the drug at this point.

What is next for the drug in a hypothetical Phase 3? The efficacy of a single dose of BMS-927711 has been demonstrated. In this regard, I would expect Phase 3 to focus on defining consistency of efficacy when treating multiple attacks. Tolerability also seems to be consistent with expectations for the class, and I would expect Phase 3 to reinforce the finding. Accordingly, Phase 3 would be mainly about defining the safety of the drug, especially its effect on the liver in situations of frequent dosing and in patients using medications that are metabolized by the CYP3A4. A drug that has sumatriptan-like efficacy, that may be better tolerated, and without vasoconstrictive properties would be an incredible addition to the migraine treatment arsenal, clearly addressing current unmet needs. In particular, patients with existing vascular disease or vascular risk factors currently have only very limited choices of acute migraine medications. And, after so many years of disappointing development we should cross our fingers and hope that the gepants find their way toward approval for the acute therapy of migraine.