New Epilepsy Drugs: Emerging Therapeutic Possibilities

Roughly one in three people with epilepsy still have seizures that existing medications cannot adequately control, a ratio that has barely budged despite the approval of more than 20 new antiseizure medications between 1990 and 2018.¹ That stubborn plateau has pushed researchers toward fundamentally different strategies, from drugs that open potassium channels and neurosteroids that boost inhibitory signaling, to antisense oligonucleotides designed to fix specific genetic defects and delivery systems that bypass the blood-brain barrier entirely. The landscape is shifting fast enough that several of these approaches have already reached clinical trials or early regulatory approval.

Why So Many Drugs and Still So Much Resistance

Most antiseizure medications developed over the past few decades work by tweaking the same handful of targets: sodium channels, calcium channels, or the brain’s main inhibitory chemical messenger, GABA. Those drugs help the majority of patients, but the ones left behind tend to share traits that blunt the effectiveness of the standard toolkit. In some cases, the underlying genetic mutation creates a problem that a broad-spectrum channel blocker simply cannot address. In others, the brain actively pumps drugs back out before they reach the seizure focus. A protein called P-glycoprotein, which normally protects the brain from toxins, gets overproduced in some epileptic tissue and ejects medications at the blood-brain barrier.² Understanding these resistance mechanisms has given researchers new places to intervene, and the result is a pipeline that looks very different from the one that produced levetiracetam or lamotrigine a generation ago.

Potassium Channel Openers

One of the more promising recent additions to the pipeline targets a class of ion channels that most earlier drugs ignored. XEN1101 is a potassium channel opener that enhances the flow of potassium ions out of neurons, making those neurons less likely to fire excessively. In a phase 2b trial of adults with focal epilepsy, the highest dose tested cut monthly seizure frequency by about 53% compared with roughly 18% for placebo, with a clear dose-response pattern across three dosage tiers.³ Those numbers are encouraging because they suggest the drug’s effect grows predictably with dose, a useful property for clinicians trying to balance seizure control against side effects. Phase 3 trials are underway, and potassium channel modulation is now seen as a credible new avenue rather than an academic curiosity.

Serotonergic Drugs and Neurosteroids

Fenfluramine, once known solely as a weight-loss drug before being pulled from the market over heart-valve concerns, has had a surprising second life in epilepsy. Reformulated at lower doses and approved for Dravet syndrome and Lennox-Gastaut syndrome, it works through a combination of serotonin receptor activity and modulation of the sigma-1 receptor. That dual mechanism appears to help rebalance excitatory and inhibitory signaling in the brain, and early evidence suggests it may also reduce the risk of sudden unexpected death in epilepsy (SUDEP) and improve cognitive function in some patients.⁴ Fenfluramine interacts with several serotonin receptor subtypes, and researchers believe this broad serotonergic engagement is part of what makes it effective in conditions that resist simpler drugs.⁵

Ganaxolone takes a different route to a similar destination. It is a synthetic neurosteroid that enhances the activity of GABA-A receptors, the brain’s main inhibitory brake. Unlike benzodiazepines, which also target GABA-A receptors but tend to lose effectiveness over time as the brain adapts, neurosteroids bind at a different site on the receptor and may be less prone to tolerance.⁶ Ganaxolone has been approved for seizures associated with CDKL5 deficiency disorder, a rare genetic condition, and trials in other epilepsy subtypes continue. The neurosteroid approach is attractive partly because the brain already produces its own neurosteroids, so the drugs are, in a sense, amplifying a natural signaling system rather than introducing an entirely foreign chemical.

Antisense Oligonucleotides and Gene-Level Therapies

Perhaps the most conceptually different strategy in the new epilepsy pipeline is the use of antisense oligonucleotides (ASOs), short synthetic strands of genetic material designed to alter how a specific gene is read. In Dravet syndrome, most patients carry a mutation in the SCN1A gene that reduces production of a critical sodium channel protein called Nav1.1. Without enough Nav1.1, inhibitory neurons cannot fire properly, and the brain tips toward runaway excitation. ASOs can fix this at the source by preventing the cell from including a “poison exon,” a stretch of genetic code that marks the SCN1A message for destruction. In mouse models, this approach increased both the functional SCN1A transcript and Nav1.1 protein levels, reduced seizure frequency, and lowered the incidence of sudden death.⁷

Follow-up work has looked more closely at what these oligonucleotides actually do at the cellular level. Treatment restored normal action-potential firing in the specific class of inhibitory neurons most affected in Dravet syndrome, normalized sodium currents, and repaired GABAergic signaling in those cells.⁸ Additional research has shown that ASOs can also correct aberrant poison-exon inclusion caused by deep intronic mutations, variants that standard genetic tests sometimes miss.⁹ Some of these ASO-based therapies have moved into early clinical trials, alongside adeno-associated virus (AAV) gene-therapy approaches that aim to deliver corrected genetic instructions directly into brain cells.¹⁰

The appeal of these gene-level strategies is obvious: rather than managing seizures by constantly suppressing neuronal excitability, they try to fix the underlying defect. The limitations are equally real. ASOs typically need to be injected into the spinal fluid, they require repeat dosing, and so far their clinical evidence comes mostly from animal models and very early human trials. Gene therapies delivered by AAV vectors face questions about long-term safety, immune response, and whether one injection can provide lasting coverage across the brain. But the direction of travel is clear, and for conditions like Dravet syndrome where the genetic cause is well understood, the precision of these tools is hard to match with traditional small-molecule drugs.

How Cannabidiol Actually Works Against Seizures

Cannabidiol (CBD) has been approved for severe childhood epilepsies including Dravet syndrome and Lennox-Gastaut syndrome, but how it controls seizures is still being pieced together. Unlike THC, CBD does not produce a high and does not work primarily through the classic cannabinoid receptors. Instead, it appears to act through multiple targets simultaneously. It blocks certain calcium and sodium channels, dampening the excitatory signals that drive seizures.¹¹ It also antagonizes an orphan receptor called GPR55, which normally promotes excitatory signaling. Blocking GPR55 appears to reduce seizure activity by tipping the balance back toward inhibition.¹²

GPR55 has become an interesting target in its own right. Molecular-modeling studies have found that several CBD-related compounds bind tightly to GPR55, raising the possibility that future drugs could be designed to hit this receptor more selectively and potently than CBD itself does.¹³ CBD also enhances potassium M-currents and interacts with a nucleoside transporter (ENT-1) that influences adenosine signaling, a known natural anticonvulsant pathway. The multimodal nature of CBD’s action is both its strength and its mystery: it works through so many channels that isolating which ones matter most for seizure control has proven difficult. For patients, the practical implication is that CBD’s effectiveness in one epilepsy syndrome does not guarantee it will work in another, since different syndromes may depend on different combinations of those targets.

Disease-Modifying Strategies Beyond Seizure Suppression

Most antiseizure medications manage symptoms. They quiet neurons that are misfiring but do not alter the underlying process that makes the brain prone to seizures. A handful of newer strategies aim to change this by targeting biological pathways involved in epileptogenesis, the gradual process by which a normal brain becomes an epileptic one.

The mTOR pathway, a central regulator of cell growth and metabolism, is overactive in tuberous sclerosis complex (TSC), a genetic condition in which abnormal tissue growths in the brain cause seizures. Everolimus, an mTOR inhibitor originally developed for cancer and organ transplantation, significantly reduced seizure frequency in a phase 3 trial of patients with TSC who had not responded to standard medications.¹⁴ The drug’s effects go beyond seizure suppression; by dialing down mTOR signaling, it may slow or partially reverse the tissue changes that generate seizures in the first place. Animal studies have shown protective effects of mTOR inhibitors across several genetic and acquired epilepsy models, suggesting the pathway could be relevant beyond TSC alone.¹⁵

Neuroinflammation is another area where researchers hope to intervene at the cause rather than the consequence. Persistent brain inflammation is not just a side effect of seizures; it can actively promote neuronal hyperexcitability and contribute to further seizure generation, cell death, and cognitive problems.¹⁶ Several inflammatory signaling pathways are being explored as drug targets, though none has yet produced an approved therapy specifically for epilepsy. The challenge is that inflammation in the brain involves many overlapping cascades, and shutting down the wrong one risks impairing the immune system’s ability to fight actual infections or clean up damaged tissue.

Getting Drugs Where They Need to Go

Even a perfectly designed antiseizure drug is useless if it cannot reach the right part of the brain at the right concentration. One of the less appreciated reasons that oral medications fail in some patients is that overexpression of efflux transporters like P-glycoprotein at the blood-brain barrier actively pumps drugs back into the bloodstream. Laboratory experiments have shown that co-administering compounds that block P-glycoprotein, such as cyclosporin A, can increase drug uptake in brain tissue by more than half.¹⁷ Understanding what drives P-glycoprotein overexpression in epileptic tissue, often seizure activity itself, has opened up the idea of targeting the signaling cascade that turns on transporter production rather than blocking the transporter directly.²

A more radical approach sidesteps the blood-brain barrier altogether. Convection-enhanced delivery (CED) uses a thin catheter and gentle pressure to infuse a drug solution directly into the brain tissue surrounding the seizure focus. This distributes medication more evenly than a simple injection and avoids the systemic side effects of oral dosing.¹⁸ In rat models, chronic CED of the GABA-A receptor activator muscimol into a deep brain structure called the subthalamic nucleus raised seizure thresholds for up to three weeks on continuous infusion, offering proof of concept that localized drug delivery can provide sustained protection.¹⁹ CED is still an experimental technique, and it requires neurosurgery to implant the delivery system, so it would be reserved for the most severe drug-resistant cases. But for patients who are already candidates for brain surgery, having an option that delivers drugs locally rather than removing tissue is an attractive prospect.

Selective Sodium Channel Blockers and AMPA Antagonists

Traditional sodium channel blockers like carbamazepine and phenytoin work broadly, inhibiting sodium flow across multiple channel subtypes. That broad action is effective for many seizure types but creates problems in conditions where one specific sodium channel subtype needs to remain functional. In Dravet syndrome, for example, the deficit is in Nav1.1, yet standard sodium channel blockers tend to suppress Nav1.1 along with everything else, often making seizures worse. Researchers are now screening for compounds that can block other sodium channel subtypes while leaving Nav1.1 intact, or even selectively activate Nav1.1 to compensate for the genetic shortfall.²⁰ This level of selectivity has only become feasible with advances in computational modeling and structural biology that let scientists visualize exactly how drug molecules fit into different channel subtypes.

On the excitatory side of the equation, perampanel represents a class of drugs that blocks AMPA receptors, which mediate most fast excitatory signaling in the brain. Unlike older glutamate-targeting drugs that were too toxic for clinical use, perampanel is a noncompetitive antagonist, meaning it does not compete directly with glutamate at the binding site but instead locks the receptor in a state where it cannot open properly. Preclinical studies showed it could protect against several seizure types in rodent models, and it has since been approved for focal and generalized seizures in many countries.²¹ Perampanel’s success has validated AMPA receptors as a druggable target, and second-generation AMPA antagonists are in development.

Metabolic Approaches and the Ketogenic Connection

The ketogenic diet has been used to treat epilepsy for over a century, but how it works has remained only partly understood. Recent research has pointed to lactate dehydrogenase (LDH), a metabolic enzyme involved in shuttling lactate between cells, as a key piece of the puzzle. Inhibiting LDH tends to hyperpolarize neurons, making them less excitable. Stiripentol, a drug already approved for Dravet syndrome, turns out to act on this same enzyme, which may explain part of its antiseizure effect.²² This insight opens the door to designing drugs that specifically target brain metabolism to suppress seizures, essentially packaging the ketogenic diet’s benefits into a pill. The dietary approach works but is extremely restrictive and hard to maintain, especially for children and adolescents. A drug that mimics its metabolic effects without the dietary burden would be a meaningful advance.

MicroRNAs as Therapeutic Targets

MicroRNAs are tiny molecules that regulate gene expression throughout the body, including the brain. Several of them are consistently altered in epileptic tissue, and manipulating them can change seizure susceptibility. MicroRNA-134, in particular, is found at elevated levels in both animal epilepsy models and in brain tissue removed from people with temporal lobe epilepsy. When researchers used antisense oligonucleotides to reduce microRNA-134 levels in rodents, the animals had higher seizure thresholds and, when status epilepticus was induced, fewer spontaneous seizures afterward.²³ This is particularly intriguing because the treatment appeared to modify the disease trajectory, not just suppress individual seizures.

More broadly, microRNAs influence pathways involved in neuroinflammation, synaptic remodeling, and cell death, all processes that contribute to the progression of epilepsy over time.²⁴ The challenge is delivery: getting a synthetic oligonucleotide into the right brain cells at the right dose is nontrivial, and off-target effects from altering a molecule that regulates hundreds of genes simultaneously remain a concern. Still, the microRNA field has progressed rapidly in other areas of medicine, and the epilepsy-specific work is catching up.

Neuropsychiatric Side Effects Still Shape Drug Choices

For all the excitement about new mechanisms, the reality of epilepsy treatment is that side effects often determine which drug a patient stays on. Psychiatric and behavioral side effects are a particularly common reason patients stop or switch medications. Among widely used drugs, levetiracetam stands out with the highest rate of psychiatric side effects, at around 16 to 22% depending on the study, including irritability, aggression, and mood disturbance.²⁵ Zonisamide also carries above-average psychiatric side-effect rates. By contrast, drugs such as lamotrigine, gabapentin, and carbamazepine are associated with significantly lower rates of these problems.²⁶

This matters for the new pipeline because any emerging drug’s neuropsychiatric profile will weigh heavily in its clinical adoption. A compound that reduces seizures by 50% but causes depression or psychosis in a significant fraction of patients will struggle to find its place. Researchers are increasingly paying attention to EEG-based biomarkers, including the photoparoxysmal response, as a way to screen drug candidates earlier and more efficiently, potentially weeding out compounds with poor tolerability before they reach expensive late-stage trials.²⁷ Better biomarkers could also help match individual patients to the drug most likely to work for them, reducing the trial-and-error prescribing that currently frustrates patients and clinicians alike. In a field where roughly a third of patients remain underserved, making drug development faster and more targeted is not a luxury but a necessity.