STOK Zorevunensen: $1.8B Bet on First Dravet Modifier

Dravet syndrome is one of the most devastating pediatric epilepsies. Children develop normally for the first few months, then start having prolonged febrile seizures that do not stop. By age two, the seizures are relentless. Cognitive development stalls. Up to 20% of these children die before adulthood from SUDEP (sudden unexpected death in epilepsy), prolonged seizures, or seizure-related accidents. Every drug on the market, from Epidiolex to Fintepla, only manages the seizures. None addresses the genetic cause.

Stoke Therapeutics (NASDAQ: STOK) is building the first drug that does. Zorevunersen is an antisense oligonucleotide delivered by intrathecal injection that upregulates the SCN1A gene to produce more NaV1.1 sodium channel protein, the very protein that Dravet patients lack. The Phase 1/2a open-label data showed a median 84.8% reduction in major motor seizure frequency at the 70mg dose, sustained through three years of follow-up. Cognition and behavior improved alongside the seizure reduction.

The Phase 3 EMPEROR trial (NCT06872125) is the first randomized, sham-controlled test of this mechanism. Stoke completed enrollment of 162 patients in June 2026, ahead of schedule. The data readout is expected in Q3 2027. At a market cap of $1.82 billion (stock at $29.31 as of July 25 per Polygon), the market is pricing in a meaningful probability of success. The market is underpricing it. Here is why.

The Disease and the Unmet Need

Dravet syndrome is caused by loss-of-function mutations in one copy of the SCN1A gene, which produces NaV1.1, a sodium channel protein critical for inhibitory interneuron function in the brain. When NaV1.1 is insufficient, inhibitory signaling fails and seizures propagate unchecked. The condition affects an estimated 38,000 patients across the US, UK, EU-4, and Japan, with roughly 16,000 in the US alone.

Current treatment is purely symptomatic. Epidiolex (cannabidiol) from Jazz Pharmaceuticals generated $1.1 billion in revenue in 2025, growing 9% year over year. Fintepla (fenfluramine), acquired by UCB in the $1.9 billion Zogenix deal, targets the same seizure symptom through a serotonergic mechanism. Stiripentol rounds out the armamentarium. All three reduce seizure frequency. None restores NaV1.1 production. Up to 57% of Dravet patients on the best available anti-seizure medicines still do not achieve a 50% reduction in seizure frequency.

The opportunity is not to replace these drugs. It is to add a disease-modifying therapy on top of them.

The Mechanism: TANGO and Antisense Upregulation

Stoke’s TANGO (Targeted Augmentation of Nuclear Gene Output) platform is the inverse of classic antisense. Most antisense drugs, like Ionis’s portfolio, knock down a target protein. TANGO antisense oligonucleotides bind to non-coding regions of mRNA to increase productive gene expression. In Dravet, where one copy of SCN1A is mutated, zorevunersen boosts the output of the remaining wild-type copy to compensate.

This is mutation-agnostic for loss-of-function variants. Whether the patient has a nonsense mutation, a frameshift, or a deletion in one SCN1A allele, the drug increases NaV1.1 output from the intact copy. That matters because Dravet is genetically heterogeneous. A therapy that works regardless of the specific mutation addresses the full addressable population, not a biomarker-selected subset.

The delivery is intrathecal, meaning the drug is injected into the spinal canal via lumbar puncture. This is the same route used for Biogen’s Spinraza (nusinersen) in spinal muscular atrophy, another antisense drug for a pediatric neurological disease. It requires a procedural visit but achieves direct CNS exposure without the blood-brain barrier problem that defeats most oral neuro drugs. The sham control in EMPEROR uses a needle prick without drug injection, which is the correct design for a procedure-dependent therapy.

The Phase 1/2a Data That Got Them Here

The open-label Phase 1/2a studies are the foundation of the bull case. Patients treated with 70mg zorevunersen (2 or 3 initial doses, n=10) achieved a median 84.8% reduction in major motor seizure frequency at 3 months post-dose, with a median increase of 8 seizure-free days per 28-day period. These reductions were sustained through three years in the open-label extension, with 77% of eligible patients remaining on treatment.

The cognitive and behavioral data is what makes zorevunersen different from every other Dravet drug. Vineland-3 subdomain scores improved by 4.3 to 9.7 raw points across eight key domains, including 7.6 points in expressive communication and 6.1 points in receptive communication. Caregivers identified a 1-3 point change as clinically meaningful. These are improvements in neurodevelopment, not just seizure counts.

The safety profile is the key open question. CSF protein elevations occurred in 42% of Phase 1/2a patients and 86% of open-label extension patients. These were generally asymptomatic and monitored, but intrathecal delivery carries procedural risk from lumbar puncture. The EMPEROR trial will determine whether the safety profile holds in a larger, randomized, sham-controlled setting.

The Phase 3 EMPEROR Trial: What to Watch

EMPEROR (NCT06872125) is a global, double-blind, sham-controlled study enrolling 162 patients ages 2 to under 18 with Dravet syndrome and confirmed SCN1A loss-of-function variants. Patients are randomized 1:1 to zorevunersen via intrathecal administration or sham for a 52-week treatment period.

The primary endpoint is percent change from baseline in major motor seizure frequency at Week 28. This is the headline number investors will parse first. The Phase 1/2a open-label data showed an 84.8% median reduction at the 70mg dose. If the randomized controlled trial shows anything close to that against sham, the stock will re-rate dramatically.

Key secondary endpoints measure durability of seizure reduction and cognition/behavior via Vineland-3 subdomains (expressive communication, receptive communication, interpersonal relationships, coping skills, personal skills). The study remains blinded through Week 52 to capture these secondary endpoints, even though the primary endpoint is measured at Week 28.

Three things to watch in the readout. First, the primary endpoint effect size. A 30% reduction vs sham would be clinically significant given that 57% of patients fail to achieve 50% reduction on current medicines. A 50% or greater reduction would be transformative. Second, the cognitive and behavioral secondary endpoints. If Vineland-3 improvements replicate the open-label signal, zorevunersen is not just a seizure drug but a neurodevelopmental therapy, which expands the pricing and market opportunity. Third, the safety profile, particularly CSF protein elevations and any lumbar puncture-related adverse events. If safety is clean, the path to approval is straightforward.

The Competitive Picture and Valuation

The Dravet market is currently dominated by Epidiolex ($1.1 billion in 2025 revenue, Jazz Pharmaceuticals) and Fintepla (UCB). Both are oral drugs. Zorevunersen would not replace them. It would be used alongside them as the first disease-modifying therapy. The competitive question is whether an intrathecal antisense drug with a 70mg dosing regimen can capture share in a market where oral drugs are the standard.

For context on what a disease-modifying rare disease therapy is worth, consider the Ultragenyx UX111 gene therapy for Sanfilippo syndrome (analysis here), another one-time-administration (conceptually) rare disease drug. Or consider Biogen’s Spinraza, which generated over $2 billion at peak as an intrathecal antisense drug for SMA. Spinraza proved that intrathecal antisense delivery is commercially viable for a pediatric neurological disease. Zorevunersen is the same drug class in a different indication.

Stoke’s market cap of $1.82 billion at $29.31 per share (Polygon, July 25) reflects a company with $391.7 million in cash and marketable securities as of December 2025, plus a $165 million upfront payment from Biogen for ex-US rights, plus up to $385 million in milestones. The Biogen partnership de-risks the commercial path: Biogen holds rest-of-world rights while Stoke retains US, Canada, and Mexico. The cash runway extends into 2028, past the expected NDA submission and potential approval in early 2028.

The comp that matters is Jazz Pharmaceuticals (JAZZ), which trades at roughly $130 per share with a $8-9 billion market cap and $1.1 billion in Epidiolex revenue. If zorevunersen is approved as the first disease-modifying Dravet therapy and captures even 20% of the addressable US patient population at a Spinraza-like price ($750,000 per year for the first year, $375,000 annually thereafter), the revenue potential exceeds $1 billion annually. Stoke at $1.82 billion is pricing in roughly one-third of that scenario at a 50% probability of success.

Analyst consensus targets STOK at $41 on average, with a high of $60, per aggregated analyst estimates. The stock trades at $29.31. The spread between the current price and the average target implies the market is pricing in a probability of Phase 3 success below what the open-label data and Breakthrough Therapy designation suggest.

The Risk: Binary and Unforgiving

This is a single-asset company. If EMPEROR fails the primary endpoint, STOK loses 60-70% of its value overnight. The open-label Phase 1/2a data is encouraging but open-label data overestimates treatment effects. The placebo response in Dravet trials is real, and the sham-controlled design will reveal the true effect size. The 84.8% open-label reduction will not replicate exactly in a controlled trial, and the question is how much of the effect is drug versus placebo, regression to the mean, and patient selection.

The secondary risk is regulatory. Zorevunersen has Breakthrough Therapy designation and rare pediatric disease designation, both positive signals. But the FDA has limited precedent with antisense upregulation therapies. The agency will scrutinize the safety profile, particularly the CSF protein elevations and the intrathecal delivery route. Any safety signal in the larger EMPEROR population could delay approval or require a REMS program.

The competitive risk is longer-dated. Gene therapy approaches to Dravet, including AAV-delivered SCN1A replacement, are in early development. If a one-time gene therapy proves safe and effective, it would obsolete a chronic intrathecal antisense regimen. But gene therapy for SCN1A is years behind zorevunersen, and Stoke’s rolling NDA submission in Q1 2027 puts it first in line for a disease-modifying approval.

The Verdict

Stoke Therapeutics at $1.82 billion is a calculated bet that the Phase 1/2a open-label data translates to a statistically significant Phase 3 result. The mechanism is biologically rational: upregulating NaV1.1 from the wild-type SCN1A allele addresses the root cause of Dravet in a mutation-agnostic way. The cognitive and behavioral improvements in open-label data differentiate this from every symptomatic anti-seizure drug. The Biogen partnership and $391.7 million cash position fund the company through the catalyst.

The risk is binary and the timeline is Q3 2027. If EMPEROR hits, STOK is worth $50-60 based on analyst consensus and the Spinraza comp. If it misses, the stock loses two-thirds of its value. I am positioning for a positive readout based on the strength and durability of the open-label data, the Breakthrough Therapy designation, and the biological logic of the mechanism. The market is pricing this as a coin flip. The odds are better than that, but this is a position-size bet, not a portfolio cornerstone. For investors who want to understand how to position for binary trial readouts, our guide to trading FDA catalysts covers the mechanics.

The next major checkpoint is the rolling NDA submission in Q1 2027, which will signal Stoke’s confidence in the emerging Phase 3 data. Watch for that submission and any interim safety disclosures. This is one of the rare disease catalysts I am most convicted on for 2027.