Olcegepant

The clinical proof-of-concept arrived in Olesen, Diener, Husstedt et al., N Engl J Med 2004, 350:1104–1110 — the single trial that established CGRP-receptor antagonism as a clinically validated mechanism for acute migraine termination. The trial randomised 126 migraine patients to a single intravenous infusion of placebo or BIBN 4096 BS at one of six escalating doses (0.25, 0.5, 1, 2.5, 5, or 10 mg) administered over 10 minutes during a moderate-to-severe migraine attack. The 2.5 mg dose produced a 2-hour pain-response rate of 66% versus 27% for placebo (p = 0.001); the pooled active-dose group produced a 60% response rate; the 2.5 mg dose was identified as the minimum effective dose without further potency gain from higher doses. The trial was small by modern migraine-trial standards and used a parallel-group dose-finding rather than a definitive efficacy design, but the separation against placebo was unambiguous and the safety profile was clean — paresthesias were the most common adverse event (25% on 2.5 mg vs 12% on placebo), with no serious adverse events.

The trial mattered for two reasons beyond its specific result. First, it was the first demonstration in humans that selective CGRP-receptor antagonism could acutely terminate migraine — before this paper, the CGRP-receptor-blockade hypothesis was supported only by mechanistic and animal data, and the regulatory case for an entire therapeutic class rested on a clinical-translation question that this trial answered. Second, it confirmed that the mechanism did not require triptan-like cerebrovascular vasoconstriction to produce migraine relief, which licensed the development of CGRP-pathway pharmacology for patients in whom triptan use was contraindicated by cardiovascular risk. The Olesen 2004 trial is the load-bearing publication on which the modern CGRP-pathway class rests; everything that followed in the gepant and anti-CGRP-antibody pipelines presupposed its result.

The clinical pharmacokinetics of BIBN 4096 BS were characterised in parallel by Iovino, Feifel, Yong et al., Cephalalgia 2004, 24:645–656 in a single-rising-dose Phase 1 study of 55 healthy volunteers across the 0.25–10 mg IV range. The terminal plasma half-life was approximately 2.5 hours and the molecule had very poor oral bioavailability; the development programme could not produce a formulation that achieved therapeutic exposure by an oral route. The IV route that produced the Olesen 2004 result was therefore the only route the molecule could practically use, which made BIBN 4096 BS unviable as an outpatient acute-migraine therapeutic in the way the field needed.

Boehringer Ingelheim discontinued the BIBN 4096 BS clinical-development programme without bringing the molecule to a Phase 3 pivotal or to regulatory submission in any jurisdiction. The proximate reason was the intravenous-only formulation; the parallel competitive context was the emergence of orally bioavailable CGRP-receptor antagonists at Merck (telcagepant, MK-0974, and the follow-on MK-3207), which appeared to obviate the need for a self-administered IV migraine therapeutic. Olcegepant's role in the field shifted from clinical candidate to research-tool compound and class-validating proof-of-concept — every subsequent CGRP-receptor-antagonist development programme cited the Doods 2000 preclinical characterisation and the Olesen 2004 NEJM trial as the foundation on which it was built.

The first-generation oral gepants that followed olcegepant then ran into a different failure mode. Telcagepant advanced through Phase 3 trials in acute migraine and into preventive-indication Phase 2 trials by 2009 before being discontinued for elevations in liver transaminases observed across the chronic-dosing programmes; the follow-on MK-3207 was discontinued at Phase 1 for the same hepatic signal. Mechanistic re-analysis in Smith, Rowe, Watkins et al., Toxicological Sciences 2020, 177:84–93 established that the hepatotoxicity was driven by a reactive-metabolite pathway that was a structural feature of telcagepant and MK-3207's bioactivation — the paper calculates a reactive-metabolite body burden of 14,560 for telcagepant and 14,720 for MK-3207, versus 4,600 for the second-generation gepant ubrogepant. The hepatic injury was not a consequence of CGRP-receptor inhibition itself: α-CGRP knockout mice do not develop liver pathology, and the anti-CGRP monoclonal antibodies that engage the same biology have not raised hepatic safety signals. The hepatotoxicity story therefore belongs to telcagepant and MK-3207 as molecules, not to the CGRP-receptor antagonist class as pharmacology — a distinction load-bearing for how the second-generation programme was rebuilt.

The second-generation gepants — ubrogepant, rimegepant, atogepant, and zavegepant — were engineered with explicit structural attention to the reactive-metabolite pathway that ended the Merck programme, and the clinical-trial liver-safety profiles across the four molecules have remained clean. The four FDA-approved oral or intranasal gepants are the descendants of the chemistry programme that olcegepant opened in 2000 and that the Olesen 2004 NEJM trial clinically validated. The class redemption is detailed in the CGRP-pathway migraine therapeutics entry; the broader failed-trial pattern across the modern peptide field is catalogued in the failed peptide trials archive; the clinical decision-architecture across the modern migraine pharmacology is developed in the migraine and peptides dossier.

The honest framing of olcegepant in 2026 is the one this entry leads with. The molecule succeeded on the question it was designed to answer — does selective CGRP-receptor antagonism acutely terminate migraine in humans — and failed on the question of whether that proof-of-concept could be turned into a marketable therapeutic at the formulation the field needed. The mechanism it validated then survived a second failure (the telcagepant and MK-3207 hepatotoxicity discontinuations) before producing the FDA-approved gepants of 2019–2023. The arc from BIBN 4096 BS to Zavzpret is a worked example of a pharmacological hypothesis that took two decades and two intermediate development failures to reach approved medicines. The comparison to other "failed development that opened a class" stories in the corpus is instructive — the dihexa entry walks an HGF/c-Met-system small molecule whose closest clinical descendant failed its definitive Alzheimer's trial in 2024; that class has not been redeemed. The difference between a redeemed and an unredeemed failed development is not whether the original molecule worked — it is whether the mechanism the original molecule validated supported a second-generation programme that addressed the failure mode without abandoning the target. Olcegepant itself remains a research-only compound and is not a therapeutic in any jurisdiction.