BHV-7000

BHV-7000 (now branded Opakalim) is Biohaven's investigational selective Kv7.2/Kv7.3 potassium channel opener, in Phase 2/3 key trials for refractory focal-onset epilepsy [1][2] and with completed Phase 2 trials in major depressive disorder [3][4]. The mechanism is biologically validated: an earlier Kv7 opener, retigabine (Potiga/Trobalt), worked for seizures but was pulled from the market in 2017 after causing blue skin discoloration and retinal pigment changes [5]. Biohaven's bet is engineered selectivity - hit the neuronal Kv7.2/7.3 subtypes that quiet seizure-prone cortex without touching peripheral Kv7 subtypes implicated in retigabine's tissue toxicity. Important framing correction: Biohaven is not alone in the modern Kv7 space. Azetukalner (XEN1101, Xenon Pharmaceuticals) is a directly comparable selective Kv7.2/7.3 opener whose Phase 3 X-TOLE2 trial in focal-onset seizures already read out positive in April 2026 (−53.2% median seizure reduction at 25 mg vs. −10.4% placebo) and is on track for a Q3 2026 NDA filing [10]. If azetukalner is approved first it will set the commercial and efficacy bar BHV-7000 must beat; if it experiences late-stage safety problems, it may also poison the regulatory perception of the entire Kv7 class. The MDD efficacy readout (originally guided 2H 2025 per Biohaven) is the earliest clinical signal for BHV-7000 specifically.

Novel chemical entity, not approved anywhere. Lead program is refractory focal-onset epilepsy in two ongoing key Phase 2/3 trials, NCT06132893 and NCT06309966, each targeting 390 adults on background antiepileptics for a combined program enrollment of approximately 780 [1][2]. Two Phase 2 MDD trials are now complete: NCT06419608 (efficacy, n=336) [3] and NCT06423781 (long-term safety, n=233) [4]. A small Yale-sponsored Phase 1 mechanistic study (NCT07125261, n=5) reads out drug effects via electrodes already implanted as part of NeuroPace's RNS System in chronic focal epilepsy patients [6]. That design is clever: rather than waiting for clinical seizure counts, the team can look directly at interictal epileptiform activity (abnormal electrical spikes between seizures, a measurable surrogate for seizure risk) on drug versus off. No FDA designations (breakthrough, fast track, orphan) for BHV-7000 have been publicly disclosed in Biohaven Ltd.'s 2025 or 2026 10-Ks [7][8]. Biohaven guided MDD key topline for 2H 2025 and first epilepsy Phase 2/3 topline for 2H 2026; readout status as of writeup date (June 2026) should be confirmed from the most recent earnings release [12][13]. The MDD readout is the earliest catalyst for BHV-7000's selectivity thesis.

KCNQ2 and KCNQ3 encode the pore subunits of the Kv7 potassium channel, which together form the neuronal M-current. Plain version: these are tiny gates in the neuron's outer membrane that let potassium ions flow out. When they open, the cell becomes more negatively charged and harder to fire. The M-current acts as a brake on neuronal excitability - turn it up and neurons fire less, turn it down and they fire more [9]. Genetics validate the target cleanly. Loss-of-function mutations in KCNQ2 cause benign familial neonatal seizures and, in more severe forms, KCNQ2 developmental and epileptic encephalopathy. Open Targets gives KCNQ2 an evidence score of 0.842 for benign familial neonatal seizures and 0.724 for developmental epileptic encephalopathy as of the cited access date [11]. Pharmacological validation came from retigabine, which produced roughly 25-30% reductions in seizure frequency as adjunctive therapy in placebo-controlled focal epilepsy trials before its withdrawal [5]. The drug worked. The leading mechanistic hypothesis attributes retigabine's pigmentation to quinone-imine metabolites that bind melanin and to non-neuronal Kv7 activity, though the exact molecular cause remains incompletely characterized in the literature. Biohaven's selectivity story is that BHV-7000's chemical scaffold avoids those metabolic liabilities while preserving Kv7.2/7.3 activation. Publicly, Biohaven has disclosed that BHV-7000 is selective for Kv7.2/7.3 and lacks GABAA activation, and that its Phase 1 program tolerated doses up to 120 mg daily without the typical CNS adverse events (somnolence, cognitive blunting) common to AEDs [12]. What is less publicly visible: specific Kv7.4/7.5 selectivity ratios, full metabolite profiling demonstrating absence of quinone-imine precursors, and chronic-exposure ophthalmology/skin preclinical data - these have not been disclosed in detail in public filings, which is itself a disclosure gap worth flagging.

The two key Phase 2/3 epilepsy trials (NCT06132893, NCT06309966) are adjunctive-therapy studies in adults with refractory focal-onset seizures already on 1-3 antiepileptics [1][2]. Primary endpoint is change from baseline in 28-day average seizure frequency, the FDA-accepted endpoint for AED registration. Each trial targets n=390 with multiple BHV-7000 doses versus placebo, baseline observation followed by titration and maintenance - combined program scale ~780 patients. Standard, well-worn AED design - easy to benchmark against cenobamate, brivaracetam, and the just-released Phase 3 X-TOLE2 data on azetukalner (XEN1101). Running two trials of this size in parallel suggests Biohaven is de-risking with replication and positioning for a faster Phase 3 path on a positive signal. Concerns: refractory focal epilepsy is a competitive recruitment pool right now with azetukalner Phase 3 long-term studies, ganaxolone programs, and several other AEDs running, so enrollment timing matters. The MDD trials (NCT06419608, NCT06423781) use MADRS (Montgomery-Åsberg Depression Rating Scale, a standard 10-item depression severity scale where lower scores mean less depression) change from baseline at week 6 as the efficacy endpoint. The Yale RNS biomarker study (NCT07125261) is the orthogonal layer [6] - n=5 is too small for efficacy claims but large enough to ask whether the drug demonstrably suppresses cortical epileptiform discharges. A positive RNS signal would meaningfully de-risk the larger trials before topline.

Our model estimates a 7% chance this drug is eventually approved. It starts from the historical base rate for Phase 2 drugs in this area (about 24%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is helped by a non-randomized design; it is held back by smaller-than-typical enrollment for this phase, the sponsor's thin or weak approval record, and weak or limited earlier-phase results. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

Five concrete failure modes. (1) Mechanism-adjacent safety: retigabine's pigmentation problems are most commonly attributed to quinone-imine metabolites and non-neuronal Kv7 activity, but the exact molecular cause is not fully established, which makes Biohaven's claim to have 'engineered around it' harder to evaluate quantitatively. Selectivity is only confirmable through long-term human exposure. The MDD safety cohort (n=233) [4] gives some comfort, but it is unclear from public disclosures whether that cohort included formal ophthalmology/dermatology assessments - this is the earliest available retinal safety signal and the writeup cannot confirm what was collected. (2) Same-mechanism competitive risk (azetukalner): Xenon's azetukalner is a directly comparable Kv7.2/7.3 opener whose Phase 3 X-TOLE2 hit primary endpoint with −53.2% median seizure reduction at 25 mg and is targeting Q3 2026 NDA submission [10]. If approved first, azetukalner sets the efficacy and tolerability benchmark BHV-7000 must beat, and Biohaven loses the first-in-the-modern-era narrative. If azetukalner has a late-stage safety failure, BHV-7000 may inherit regulatory caution toward the class. This is the dominant near-term competitive risk and is different in kind from generic AED competition. (3) Different-mechanism AED competition (market share): cenobamate (Xcopri) has responder rates above 40% at top doses, and brivaracetam, lacosamide, and others crowd the refractory focal space. Even with comparable efficacy, BHV-7000 must differentiate on tolerability to take share. (4) Class-typical CNS adverse events: dizziness, somnolence, cognitive blunting are common to drugs that suppress neuronal firing. Phase 1 disclosure suggests BHV-7000 cleared 120 mg daily without these effects [12], but Phase 2/3 doses and chronic exposure remain the real test. (5) Commercial / payer: payers will likely require step-edits (insurer requirements to try and fail cheaper generic drugs before authorizing a new branded therapy) through generics (levetiracetam, lamotrigine) before authorizing a branded Kv7 opener. Biohaven needs differentiated efficacy or a tolerability story to command premium pricing, and the broad refractory focal label fragments across many small subpopulations.

Worth watching, not yet a signal - and the bar to call it a signal has just been raised by azetukalner's Phase 3 win. Three data points to track in order. First, the completed MDD Phase 2 efficacy readout (NCT06419608, MADRS endpoint at week 6) [3] - this is the earliest look at whether the molecule does anything clinically in humans, and the safety database here will tell you whether the selectivity engineering is holding up. Second, the Yale RNS biomarker readout (NCT07125261) [6] - small but mechanistically decisive; if interictal epileptiform activity (abnormal between-seizure electrical spikes) doesn't drop on drug, the epilepsy trials are in trouble. Third, topline from NCT06132893 / NCT06309966 (first topline guided 2H 2026 per Biohaven) - looking for ≥40% responder rate and a clean ophthalmology / urinary / QT package, and now also looking for tolerability that compares favorably to azetukalner's X-TOLE2 profile. Connecting to the sponsor: Biohaven Ltd. (BHVN) is the post-Pfizer-spinout entity that retained the broader pipeline after Pfizer acquired rimegepant/Nurtec [8]. BHV-7000 is a top-three asset alongside troriluzole (spinocerebellar ataxia) and the TYK2 program. Cash position as of March 31, 2026 was approximately $351.8M, supplemented by a $600M non-dilutive royalty/financing agreement with Oberland Capital announced in 2025 ($250M received at closing on April 30, 2025) [13]. That liquidity buys roughly enough runway through the next set of key readouts, but burn-versus-readout cadence makes the MDD and epilepsy topline events high-stakes for the equity. Check back when MDD topline drops - that's when the selectivity thesis gets its first real-world test.