Carisbamate

SK Life Science is running a Phase 3 trial of carisbamate (YKP509, originally RWJ-333369) as add-on therapy for drop seizures in Lennox-Gastaut syndrome (LGS), a severe childhood-onset epilepsy in which patients fall to the ground multiple times daily and respond poorly to standard antiseizure drugs [1]. NCT05219617 enrolled 252 children and adults and uses the same regulatory-validated endpoint that supported approvals for rufinamide, cannabidiol, and fenfluramine in this indication [2]. The asset has a complicated history: Johnson & Johnson developed it for partial-onset seizures more than a decade ago and walked away after the FDA flagged efficacy and skin-rash concerns [3]. SK Life Science, the Korean drugmaker behind cenobamate (Xcopri), reacquired it and repositioned it into a refractory pediatric epilepsy market where the bar for adjunctive efficacy is lower and unmet need is high [4]. A clean readout would give SK a second branded neurology product to sit alongside cenobamate; a failure would mark the second Phase 3 burial for this molecule.

Carisbamate is a novel small molecule that has never been approved anywhere in the world. It is currently in Phase 3 (NCT05219617, recruiting), with completed Phase 1 PK and bridging studies in LGS patients (NCT04062981, NCT03731715) and a food-effect study (NCT04520360) supporting the registrational program [2]. No breakthrough therapy, fast track, or RMAT designations are listed in the public record for this LGS indication. Orphan Drug Designation (ODD) status for the LGS indication is not confirmed in publicly searchable FDA orphan-designation listings as of mid-2026; LGS itself plausibly qualifies (US prevalence <50,000) and SK has had opportunity to apply, but neither an ODD grant nor a public application disclosure has been verified for this asset (this is a notable gap and should be confirmed against SK's most recent investor materials). The compound carries baggage. J&J ran a large partial-onset seizure program for RWJ-333369 in the mid-2000s and submitted an NDA that the FDA declined in 2008, citing efficacy and a serious skin-rash signal [3]. SK Life Science licensed the molecule from J&J in 2010 and shifted strategy: instead of competing in the crowded focal epilepsy market, target a refractory orphan indication where adjunctive antiseizure efficacy is the gold standard. The Phase 3 trial began dosing in 2022. Public guidance on topline readout has not been disclosed; given the 2022 start and recruitment status as of mid-2026, readout in 2026-2027 is plausible but not confirmed by the sponsor.

Carisbamate is classified as a sodium channel modulator [5]. Sodium channels are the molecular gates that let neurons fire - when they open at the wrong time or stay open too long, neurons fire in runaway bursts that produce seizures. Carisbamate dampens this by stabilizing the inactive state of voltage-gated sodium channels, with additional activity at other neuronal ion channels that contributes to its anticonvulsant effect in animal models [1]. The mechanism is heavily validated for epilepsy in general: lamotrigine, carbamazepine, oxcarbazepine, lacosamide, and SK's own cenobamate all hit sodium channels and all work in focal seizures. Whether this class works in LGS specifically is a more open question. LGS involves diffuse network dysfunction with multiple seizure types, not just cortical hyperexcitability, and the approved LGS drugs are mechanistically diverse - rufinamide is a sodium channel modulator, cannabidiol acts via GPR55 antagonism and sodium/calcium channel modulation with additional activity at TRPV1 and adenosine receptors, fenfluramine works through serotonin and sigma-1 receptors. Carisbamate showed activity across multiple preclinical seizure models (maximal electroshock, pentylenetetrazol, kindling) and produced signal in early focal epilepsy trials before the J&J program collapsed [5,6]. The LGS rationale rests less on a unique mechanism than on the empirical observation that broad-spectrum sodium channel drugs frequently reduce drop seizures (rufinamide is the existing proof point), combined with the pharmacokinetic profile SK is targeting in pediatric dosing.

NCT05219617 is a multicenter, double-blind, placebo-controlled, parallel-group Phase 3 trial enrolling 252 children and adults with a confirmed LGS diagnosis who are inadequately controlled on one to four background antiseizure medications [2]. The primary endpoint is percentage change from baseline in the 28-day frequency of countable drop seizures (tonic, atonic, and tonic-clonic) during the double-blind treatment period [2]. This is the same endpoint construction the FDA accepted for rufinamide (Banzel), cannabidiol (Epidiolex), and fenfluramine (Fintepla) - a regulatory-friendly choice with established precedent. Placebo control is the standard comparator in adjunctive LGS trials and there is no head-to-head against an active drug, which is typical for orphan epilepsy programs. Enrollment target of 252 is well-powered for the magnitudes of effect seen in prior LGS approvals (rufinamide showed roughly 30% median seizure reduction; cannabidiol around 40%). The age range is broad and the inclusion criteria allow patients on multiple background drugs, reflecting real-world LGS practice but also introducing variability that can blunt signal. Recruitment status as of mid-2026 is active. The main design concern: no enrichment or biomarker selection, which means the trial bets entirely on the molecule's broad antiseizure activity translating into drop seizure reduction across a heterogeneous patient population. Given the 2008 FDA complete response letter on the partial-onset program, an FDA Peripheral and Central Nervous System Drugs Advisory Committee meeting on any LGS NDA is plausible - the prior CRL plus a pediatric population frequently triggers AdComm review, though FDA has not signaled this publicly.

Our model puts the odds of this drug eventually getting approved at 16%. It starts from the historical approval rate for Phase 3 drugs in this area - about 51% - then adjusts that number using ten facts about the trial and its sponsor. The biggest downward pulls are the sponsor's weak approval track record, limited earlier-phase results, a randomized trial design, and the use of a comparator arm. Most other factors landed near average, so they didn't move the estimate much from there.

Three concrete failure modes. First, the J&J skin-rash signal. The prior RWJ-333369 program documented serious cutaneous adverse events, including cases that raised Stevens-Johnson syndrome (a rare but potentially life-threatening skin and mucous membrane reaction) concerns at FDA review [3]. SK's program will have to demonstrate either a more favorable rash rate at the LGS dosing regimen or a manageable risk profile in a pediatric population - a high bar given LGS patients are also frequently on lamotrigine, itself a rash-causing drug. Second, efficacy in a heterogeneous LGS population. Carisbamate's anticonvulsant signal in J&J's focal epilepsy program was real but modest, and the FDA's complete response letter cited efficacy concerns alongside safety [3]. LGS drop seizures can respond to broad-spectrum sodium channel drugs, but the effect size needs to clear placebo by a meaningful margin in 252 patients with varied background therapy. Third, commercial positioning even if approved. LGS has four branded adjunctive options (rufinamide, clobazam, cannabidiol, fenfluramine) plus broad-spectrum antiseizure drugs used off-label. Cannabidiol and fenfluramine have generated strong real-world adoption and pediatric neurologist familiarity. Carisbamate enters as a fifth-line option with no compelling differentiation story unless the Phase 3 data show a separation on a clinically meaningful subgroup or safety profile. Payer access in pediatric epilepsy is generally not the binding constraint, but pricing power against established competitors is. Composition-of-matter patent protection on carisbamate dates to the late 1990s/early 2000s J&J filings and would be substantially expired at a 2027-2028 launch; SK's IP runway likely rests on formulation, dosing-regimen, and method-of-use patents plus regulatory exclusivities - pediatric exclusivity under PREA (an additional 6 months) and ODD (7 years, if granted) would be the primary commercial moats, making ODD status materially important.

Worth watching, not a high-conviction signal. The trial structure and endpoint are regulatory-clean, and SK Life Science has demonstrated it can move a CNS asset through FDA approval with cenobamate. But this is a recycled molecule with a documented prior Phase 3 failure, an unenriched trial design, and a crowded competitive position in an orphan indication. The data point that would convert this into a real signal: a topline that shows a 30%+ median drop seizure reduction with a separation from placebo wider than 15 percentage points, paired with a rash rate at or below the lamotrigine background. Anything short of that produces a marginal label and a tough launch against cannabidiol and fenfluramine. Check back when SK provides topline guidance - the company has historically waited until 6-12 months before readout to telegraph timing. The earnings call to watch is SK Biopharmaceuticals' quarterly update, where the carisbamate Phase 3 progress is typically referenced alongside cenobamate revenue [4]; SK Biopharmaceuticals' top line is heavily concentrated in cenobamate, which means pipeline investment is sensitive to cenobamate growth - a slowdown in Xcopri sales would tighten the spending posture around carisbamate launch readiness. If SK acquires a rash-safety dataset from an interim DSMB (Data Safety Monitoring Board - an independent committee that reviews interim safety data) review and chooses to disclose, that would be a near-term tell. For a portfolio, this sits as a low-probability binary readout with modest upside on approval and limited downside on failure since SK's value sits primarily in cenobamate.