May 2026 Research Update

This month added a clearer picture that two established add-on options, fenfluramine and pharmaceutical-grade cannabidiol (CBD), are repeatedly linked with meaningful seizure reductions across many studies, even though results vary a lot from person to person. The new evidence also reinforces that side effects are often the main tradeoff, and that some side effects seem more likely when CBD is combined with certain other seizure medicines. For fenfluramine, one real-world clinic report suggested some people may respond even at low starting doses, while large pooled analyses suggest seizure benefits are seen across broader groups of developmental epilepsies, with Dravet syndrome often showing a stronger signal than some other syndromes. [1] [2] [3] [4]

In a Japanese clinic experience with 27 people with Dravet syndrome, many patients had fewer convulsive seizures after starting fenfluramine, and diarrhea was the most common side effect reported. The study also described that many responders reduced or stopped other antiseizure medicines, and it did not observe valvular heart disease or pulmonary arterial hypertension in this group during the reported follow-up. Because this was one center and not a controlled trial, it mainly shows what happened in routine care rather than proving exactly how much benefit to expect for any individual. [1]

Two large evidence summaries pooled results for CBD across treatment-resistant epilepsies and developmental epileptic encephalopathies, including Dravet syndrome, and both reported fewer seizures on average with pharmaceutical-grade CBD. One review of randomized placebo-controlled trials reported a stronger seizure-reduction signal in Dravet syndrome than in some other epilepsy groups, and it found that different CBD formulations did not perform the same in the short term (liposomal showed benefit while transdermal did not). Across these reviews, side effects were usually described as mild to moderate, but sleepiness and liver enzyme elevations were reported more often when CBD was taken with valproate or clobazam. [3] [4]

A separate meta-analysis focused on fenfluramine across developmental epilepsies also reported that many patients reached common seizure-reduction thresholds, with a smaller group reported as seizure-free, and it described a stronger signal in Dravet syndrome compared with Lennox-Gastaut syndrome. It also noted that longer-term data, largely from Dravet cohorts, did not show a clear pattern of the benefit wearing off over time, although follow-up and study methods varied widely. This helps place individual clinic reports into a broader context: the medicine often helps, but the amount of help differs across people and study settings. [2]

For families, this month mostly strengthens and organizes what is already known: fenfluramine and pharmaceutical-grade CBD have the most human evidence in this set of papers, and many patients in these studies had fewer seizures. It does not define what will happen for any one person, but it does clarify which side effects were most consistently flagged (for CBD, sleepiness and liver enzyme changes especially with valproate or clobazam; for fenfluramine in one clinic report, diarrhea). The strongest information here comes from pooled analyses of many studies, while the single-center fenfluramine report helps illustrate how treatment can look in day-to-day care, including that some people responded at low initial doses in that clinic. What remains uncertain in these summaries is how best to predict who will respond, and how results compare across different real-world settings and long follow-up periods. [1] [3] [4] [2]

This month included a small real-world report suggesting responsive neurostimulation (RNS), a brain device that detects and responds to seizure activity, can reduce seizures for some people with SCN1A-related epilepsy, but the benefit can be uneven. Some individuals had large improvements, including prolonged seizure freedom or stopping specific seizure types, while others had only small or no meaningful change. The study also highlighted caregiver-noted changes beyond seizure counts, such as fewer clusters or shorter recovery, which may matter to daily life even when seizures are not fully controlled. [5]

In this 6-person retrospective series (5 with Dravet syndrome), most had at least a 50% seizure reduction by last follow-up, but one person did not have meaningful improvement. The leads were most often placed in thalamic targets, and one stimulation side effect (tingling sensation) improved after adjusting the device settings. No major device- or stimulation-related complications were reported in this small group, but the authors emphasized that outcomes likely depend on patient factors and that larger studies are needed. [5]

For patients and caregivers, this is early but patient-centered evidence that RNS is being tried in SCN1A-spectrum epilepsy and that some people may see meaningful improvements, including fewer seizure emergencies and better day-to-day recovery in caregiver reports. Because the report included only 6 people and did not have a comparison group, it cannot tell how often this works or who is most likely to benefit. What it does add right now is a realistic range of outcomes (from none to very large benefit) and a reminder that side effects can sometimes be managed by changing device settings. [5]

This month, multiple preclinical studies aimed at the root cause of Dravet syndrome by boosting SCN1A function or correcting an SCN1A variant, and they reported strong improvements in seizures and survival in mice. At the same time, these studies made it clearer that the hardest part may be delivering the right amount of genetic therapy safely, because too much or the wrong timing can trigger inflammation or other unwanted effects. A separate antisense chemistry study suggested it may be possible to design splice-switching medicines that work at lower doses and last longer in the nervous system, which could matter for future SCN1A-targeted approaches. None of these are human trials, so they show what is possible in models rather than what is proven for people. [6] [7] [8]

One mouse study used an AAV9-delivered adenine base editor to correct the SCN1A R613X variant and reported restored inhibitory neuron function along with fewer spontaneous and heat-triggered seizures. The treated mice also had much higher survival than untreated mice during the study period. This supports the idea that precise correction of a specific SCN1A change can improve core brain cell function in an animal model, although the summary does not include details about longer-term safety or how this would translate to different SCN1A variants in people. [6]

Another mouse study tested a one-time AAV9 gene therapy approach designed to increase the mouse's own Scn1a levels by blocking a natural antisense transcript (AntagoNATs). When given very early in life with combined brain and body delivery, treated mice had better survival and fewer seizures, and Scn1a messenger RNA increased in the cortex. The same study also reported a narrow therapeutic window, where a higher-dose juvenile treatment caused brain inflammation and worse survival, showing that dose and timing may be critical for safety. [7]

A separate preclinical chemistry study focused on how to make splice-switching antisense oligonucleotides more potent and longer-lasting in the brain and spinal cord, and it included an SCN1A-relevant splicing target (exon 20N). The researchers reported that NMA-modified molecules could achieve splicing changes at lower doses than older designs and that some optimized versions showed better tolerability signals in rodents. For Dravet syndrome, this mainly matters as platform work: it suggests one possible route to more efficient SCN1A-targeting medicines, but it does not yet show seizure outcomes in a Dravet animal model in the provided summary. [8]

For families, these studies point to a future direction: treatments that try to address SCN1A directly can produce large seizure and survival improvements in Dravet mouse models, which is encouraging as a proof of concept. What remains uncertain is how these approaches will perform in people, including how durable the benefit is, how to control immune or inflammation risks, and how to find doses that are effective without being harmful. Right now, this evidence mainly matters as research progress rather than something ready for routine care, but it helps clarify which strategies (variant correction, boosting SCN1A expression, improved antisense designs) are being actively tested and what hurdles are showing up early. [6] [7] [8]

This month also included earlier-stage studies exploring new compounds and new ways to deliver familiar medicines, with results that were promising in some models but not yet consistent across tests. A new compound (GM-90663) reduced seizure-like behavior in a Dravet zebrafish model and appeared to act through more than one mechanism, while 6-gingerol showed only modest or model-dependent antiseizure effects in zebrafish and mice. A formulation study developed nanoparticles carrying stiripentol and CBD and reported sustained drug release and a trend toward higher CBD exposure in mice, which could matter for future medicine design but does not yet show seizure outcomes. [9] [10] [11]

In a Dravet zebrafish model (scn1lab knockout larvae), GM-90663 reduced seizure-like movements and also improved a cognitive-like color preference behavior that was abnormal in the model. The study linked these changes to higher serotonin levels through MAO-A inhibition and to changes in sodium channel behavior measured in nerve cells, suggesting a multi-target approach. Safety signals in this model were limited to what could be observed in larvae, and the study reported no adverse effects on development or heart function in zebrafish. [9]

A separate study tested 6-gingerol (a ginger component) in several seizure models and found mixed results, with clearer effects only at higher exposures in some settings. In Dravet-related zebrafish morphants, it did not significantly reduce certain electrical seizure-like signals, and in mice it did not help in some standard seizure tests, while showing benefit in a 6 Hz test at a high dose. This pattern suggests any antiseizure activity may be limited, dependent on the model, or constrained by how much drug reaches the brain. [10]

Researchers also developed nanostructured lipid carriers that package stiripentol together with CBD oil and tested how the particles behaved and how drug levels changed in mice. Compared with commercial products, the nanoparticles did not produce statistically significant differences in mouse drug exposure, but the study reported a trend toward higher CBD bioavailability and a more sustained release profile. For Dravet syndrome, this is a delivery-technology step that could later be paired with clinical testing, but it is not evidence yet that seizure control improves. [11]

For patients and caregivers, these studies are best viewed as early scouting work rather than near-term options. GM-90663 is an example of a new candidate that showed seizure-like improvements in a Dravet zebrafish model and offers clues about possible mechanisms, but it is far from proving benefit in people. 6-gingerol showed limited and inconsistent antiseizure effects across models, which helps set expectations that natural compounds still need rigorous testing and may not translate well. The nanoparticle study mainly matters for the future because it explores whether changing how medicines are packaged could change drug levels and release, which could eventually affect tolerability or convenience if later studies support it. [9] [10] [11]

One population study this month helps place Dravet syndrome within the broader group of very early-onset epilepsies that become drug-resistant. It showed that in children whose seizures started before age 2 and became drug-resistant, many developed resistance within months and most had significant developmental and motor challenges. It also reported that a cause could often be identified using brain imaging and genetic testing, which supports the idea that finding an underlying explanation is common in this severe early-onset group, even though Dravet syndrome itself was only a small part of the cohort. [12]

In the Estonian nationwide cohort of 37 children with early-onset drug-resistant epilepsy, developmental and epileptic encephalopathies were the majority, and most children had global developmental delay or intellectual disability. The study reported that structural brain findings and genetic results together explained many cases, and an etiology was established for most children in the cohort. Dravet syndrome was diagnosed in 2 patients, so the report is more informative about the overall landscape of early severe epilepsy than about Dravet-specific outcomes. [12]

For Dravet families, this study does not add new details about Dravet syndrome treatment, but it adds context about how quickly drug resistance can emerge in severe epilepsies that start in infancy. It also supports that genetic testing and brain imaging often find an underlying cause in this population, which helps explain why so much research and care planning in early-onset epilepsies focuses on identifying etiology. What remains uncertain from this single national cohort is how closely these numbers match other countries and more recent care eras, and it cannot describe the full range of outcomes for Dravet specifically because only 2 patients had Dravet syndrome. [12]

One lab study this month supports the idea that Dravet syndrome may involve early differences in how inhibitory brain cells develop and function, even before considering seizures. Using patient-derived brain organoids enriched for inhibitory interneurons, researchers observed reduced neuronal excitability and synaptic activity compared with an otherwise matched control. The same work reported that fenfluramine partially restored neuronal activity in this model, providing a possible cellular explanation that can be tested further. [13]

In these organoids, the researchers also reported altered activity of gene programs involved in ventral forebrain interneuron development. This matters because inhibitory interneurons are central to keeping brain activity balanced, and SCN1A loss-of-function is expected to disrupt that balance. Because this is an organoid model, it cannot capture the full complexity of a developing child or predict clinical response, but it offers a controlled way to study early developmental vulnerability in Dravet syndrome. [13]

For patients and caregivers, this is mainly research that helps explain Dravet syndrome biology in a more human-relevant lab system than animal models alone. It does not change current care, but it helps clarify why Dravet is often described as a developmental condition as well as an epilepsy, by pointing to early differences in inhibitory neuron programs and activity. It also raises testable questions about whether some treatments might affect brain cell activity in ways that go beyond seizure counts, although this organoid finding does not prove clinical benefit. [13]