July 2026 Research Update

Across this group of papers, the clearest human evidence came from longer follow-up of established medicines: add-on perampanel and fenfluramine. Overall, seizure reductions were common in the perampanel real-world study, and long-term fenfluramine follow-up suggested global functioning was usually stable or improved, although side effects were frequent. Other reports included expert guidance on managing multiple medicines and early-stage or lab work that points to future treatment directions but does not yet show patient benefit. [1] [2] [3] [4] [5] [6]

In a multicenter, prospective real-world study in China (56 people with Dravet syndrome), add-on perampanel had a 50% or greater seizure reduction in 66.07% at 6 months and 60.71% at 12 months. Seizure freedom was reported in 21.43% at 6 months and 12.50% at 12 months, and about two-thirds stayed on treatment through 12 months. Side effects were reported in 6 people and 3 stopped treatment, and responses did not appear to differ by the type or location of the SCN1A gene change. [1]

In an international open-label extension that followed 412 people who had already been in earlier fenfluramine studies (including Dravet syndrome and Lennox-Gastaut syndrome), 75.5% reported at least one treatment-emergent side effect, while fenfluramine-related serious side effects were reported in 1.2%. Ratings of overall day-to-day functioning by caregivers and clinicians were improved or unchanged in over 90% of assessments. This study did not include a comparison group, so it cannot show how much change was due to fenfluramine versus other factors over time. [2]

Two single-patient case reports described seizure freedom after starting fenfluramine in genetic epilepsies related to an SV2A gene change and to SCN1A-related genetic epilepsy with febrile seizures plus (GEFS+). Because each report describes one child, it is not possible to know how often this happens or which patients are most likely to respond. [7] [8]

An expert consensus panel of 26 North American clinicians focused on practical principles for using multiple medicines in Dravet syndrome, emphasizing Dravet-specific options (stiripentol, cannabidiol, fenfluramine) and attention to drug interactions and monitoring. This work provides guidance rather than new patient outcome data, and some statements did not reach consensus. [3]

Several papers pointed to possible future directions. A conference summary reviewed 13 investigational treatments with early seizure outcome data across epilepsies, including approaches relevant to Dravet syndrome, but it did not provide a single, consistent set of results for Dravet syndrome specifically. [4]

Translational and lab studies suggested additional targets: in mouse models of Dravet syndrome, blocking the cGAS pathway reduced seizure severity and brain inflammation signals, and human epilepsy tissue showed activation of related immune signaling in microglia. Separately, a cell-based platform was developed to measure and modulate Nav1.1 (SCN1A) protein production using antisense tools, which could support future therapy development but does not yet demonstrate clinical benefit. [5] [6]

One small registry study in 10 children with HCN1-related refractory epilepsy (not Dravet syndrome) reported that lamotrigine and oxcarbazepine were linked to worse seizures, while valproate and clobazam were linked to improved seizure frequency, with two children becoming seizure-free. Because this was a small, observational study in a different genetic epilepsy, it mainly serves as background on how treatment response can vary by gene and drug. [9]

For families affected by Dravet syndrome, this set of studies adds real-world and longer-term context to two medicines already used in practice: perampanel (as add-on therapy) and fenfluramine (long-term follow-up after earlier trials). It also highlights that some research is currently guidance-based (expert consensus on combining medicines) or early-stage (lab work and investigational drugs), which may shape future options but is not yet proof of patient benefit, and single-patient reports cannot predict how others will respond.

Across this group of studies, researchers focused on three connected needs in Dravet syndrome: separating Dravet from other epilepsies using brain-wave data, making sense of gene changes that are hard to interpret, and standardizing how Dravet-related information is organized for research and data sharing. Overall, the work suggests these approaches can support classification, but most results are early-stage and need broader testing before they can be relied on in everyday care. [10] [11] [12]

In a proof-of-concept study using electroencephalogram (EEG) recordings from 9 children with Dravet syndrome plus comparison groups, a pre-trained deep-learning model was more accurate at classifying Dravet than a similar model trained without pre-training. The pre-trained model reached 85% accuracy and showed less confusion between classes, but the authors emphasized that larger datasets and prospective validation are needed to know how well it generalizes. [10]

A separate computational study modeled how different SCN1A missense variants might change the stability of the Nav1.1 protein, aiming to help distinguish variants more likely to be harmful from those more likely to be harmless. In their analysis, variants labeled pathogenic in ClinVar tended to be more destabilizing than benign variants, and about 18.5% of variants of uncertain significance looked strongly destabilizing, which could help prioritize which uncertain variants deserve closer study. The authors also reported that destabilization patterns differed by protein region, suggesting that different parts of the protein may lead to different underlying problems, but they framed the method as hypothesis-generating and called for functional validation. [11]

Another project developed a Dravet-focused extension to the Epilepsy and Seizure Ontology, adding Dravet concepts across nine domains (including seizures, development, genetics, and EEG) so that data can be represented in a computable, standardized way. The authors reported public release and examples of reuse in tasks like extracting information from research papers, which may help future Dravet research compare studies more consistently. This work is about organizing knowledge for research rather than changing how an individual child is diagnosed. [12]

In a retrospective cohort of 136 children with developmental and epileptic encephalopathies, about 30.1% had a genetic cause identified, and SCN1A was the most frequently affected gene among those with a genetic etiology. Dravet syndrome accounted for 10 of the genetic electroclinical phenotypes reported, showing how Dravet can appear within a broader group of severe childhood epilepsies where genetic testing may clarify classification. [13]

In a separate retrospective genetic study of fever-sensitive epilepsy, only 30% of 53 individuals had a pathogenic or likely pathogenic variant identified, highlighting that fever-triggered seizures can have multiple genetic explanations and may remain genetically unsolved in many cases. The study also reported ATP6V0C variants in 2 individuals and described a pattern of early fever-triggered motor seizures with EEG that may start normal and later show multifocal abnormalities, plus later neurodevelopmental impairment, features that can overlap with conditions considered in the Dravet differential diagnosis. [14]

For patients and caregivers, these studies mainly add background about how researchers are trying to make Dravet classification more consistent using EEG-based computer models, improved interpretation of SCN1A gene changes, and better organization of Dravet knowledge for research databases. Right now, the EEG model and the SCN1A stability predictions are exploratory and not presented as stand-alone ways to confirm or rule out Dravet, and the ontology work is designed for research standardization rather than direct clinical decisions. Important uncertainties include whether these tools work reliably across hospitals, EEG systems, and diverse patient groups, and how well computational predictions match real-world effects of specific genetic variants. [10] [11] [12] [13] [14]

Across two preclinical mouse studies of SCN1A-related Dravet syndrome, a shared pattern is that brain circuit problems are not fixed at birth. Instead, certain networks show stronger disruptions later in development, including changes in the timing and strength of signals that help brain regions coordinate activity. [15] [16]

In hippocampal slices (a brain area linked to learning and memory), ripple frequency became lower in young adult Scn1a+/- mice compared with typical mice, while juvenile mice did not show this difference. The same work reported that parvalbumin (PV) interneurons, which help synchronize networks, showed weaker alignment of their firing to the ripple rhythm and age-related changes in how their electrical spikes looked in the cell body and along the axon. [15]

A separate study focusing on the thalamus (a key relay for sensory signals) found region-specific changes across development. After seizure onset, excitatory input to the ventral posterolateral (VPL) thalamus decreased and stayed reduced into adulthood, while a nearby sensory thalamus region (ventral posteromedial, VPM) did not show the same excitatory change; inhibitory synapse problems also appeared later and persisted in adult mice only in VPL neurons. [16]

Together, these studies point to a mix of timing-related problems (less coordinated rhythms) and connection-related problems (weaker synapses) that can become clearer with age, and that may affect some brain regions more than others. Because both studies are in mice (including slice experiments), it is still uncertain how directly each specific circuit change maps onto day-to-day symptoms in people. [15] [16]

For patients and caregivers, this research mainly adds background on why Dravet can involve more than seizures, by showing how SCN1A changes can alter brain communication over time in animal models. It suggests that the brain effects may evolve with development and may not be uniform across all regions, but it does not yet provide a test to measure these changes in people or a direct link to specific learning, behavior, or sensory symptoms. More work is needed to confirm which circuit changes also occur in humans and how strongly they relate to everyday functioning.

Across this paper, the main message is that breathing problems in Dravet syndrome (DS) may relate to how the brainstem adjusts breathing when carbon dioxide rises or oxygen falls, which could matter for sudden unexpected death in epilepsy (SUDEP). The evidence discussed is mainly from DS mouse models, and it remains uncertain how well these mechanisms translate to humans or how they could be measured clinically. [17]

In young DS mice, the breathing response to higher carbon dioxide was reduced at lower carbon dioxide levels, suggesting a weaker or altered reflex that normally increases breathing when carbon dioxide builds up. The commentary links this to changes in brainstem circuits that reduce inhibitory control (disinhibition), which can shift baseline activity in breathing-related neurons. [17]

The paper also notes a possible split in older DS mice: those that later died in a SUDEP-like way had a more blunted breathing response to both high carbon dioxide and low oxygen, while mice that survived had responses closer to typical mice. This supports the idea that impaired breathing reflexes could be part of SUDEP risk in at least some cases, but the exact reason why only some animals show this pattern is still unclear. [17]

In lab (in vitro) experiments discussed in the commentary, a drug-like tool that activates the Nav1.1 channel (Hm1a) could normalize activity in a brainstem area involved in carbon dioxide sensing. This is early, preclinical evidence about a possible mechanism, not a tested treatment approach in people. [17]

For patients and caregivers, this work mainly adds background on why breathing can be a key part of SUDEP discussions in Dravet syndrome. Right now, it does not establish a clinical test that can reliably measure these breathing reflex changes in people, and it does not show that changing these reflexes would lower SUDEP risk in patients. The uncertainty is which individuals (if any) have a measurable breathing-reflex problem that meaningfully predicts risk, and whether that could be used safely and accurately in real-world care. [17]

Across this month's evidence, one preclinical animal study suggests Dravet syndrome may affect the eye s retina (the light-sensing tissue at the back of the eye) in multiple ways, not only the brain. The same study also suggests that reducing inflammation-like activity in the retina did not necessarily reverse retinal structure or blood vessel changes, at least over a short treatment period. [18]

In Scn1a +/- rats (a Dravet syndrome model), researchers measured retinal layers, electrical responses to light, and retinal blood vessel patterns. The Dravet model showed thinning of the nerve fiber layer and thickening of another retinal layer, while overall retinal thickness stayed similar. [18]

The study also found reduced retinal oscillatory electrical signals, which can reflect how retinal circuits handle rapid activity, while other major light-response signals were mostly preserved. Fenfluramine treatment reduced signs of microglial activation (a type of immune cell activity in nervous tissue), but it did not correct the measured retinal layer thickness changes or the abnormal blood vessel architecture. [18]

Because these retinal changes were measurable without looking at the brain directly, the authors suggested that retinal testing might be explored as a non-invasive biomarker for tracking disease progression or treatment effects. However, the work was short-term and in an animal model, so it is not yet clear how well these retinal measures translate to people with Dravet syndrome. [18]

For patients and caregivers, this research mainly serves as background science suggesting Dravet syndrome could have body effects beyond seizures, including possible eye-related changes. It does not show that people with Dravet have these exact retinal findings, or that fenfluramine changes vision or eye health in patients. What remains uncertain is whether similar retinal measures can reliably track Dravet syndrome in humans, how they relate to daily symptoms (if at all), and whether longer-term treatment would change structural or blood vessel features.

Across two cross-sectional surveys, caregiving in Dravet syndrome was described as an ongoing, multi-part burden that continues well beyond the time of diagnosis. [19] [20] The hardest parts commonly involved ongoing supervision and complex seizure and medication routines, with knock-on effects on caregivers emotional health, sleep, and ability to work. [19] [20]

In a large multinational European survey of 379 caregivers, burden was described as sustained over time and tied to ongoing seizures, complex treatment routines, and the need for rescue medicines, alongside day-to-day management of behavioral, motor, cognitive, and communication difficulties. [19] Caregivers reported worse emotional well-being, sleep, fatigue, and stress, while physical functioning was described as relatively preserved. [19] Employment was often affected: 26.4% were employed full-time and 25.6% identified as full-time caregivers. [19]

The same European survey reported differences by gender and country. [19] Women reported poorer emotional well-being, lower autonomy, and poorer social well-being than men, while physical well-being did not differ by sex. [19] Emotional well-being and autonomy also differed across countries, with poorer outcomes reported in Poland and more favorable profiles in the Netherlands and Germany. [19]

In a smaller nationwide study in the Czech Republic (44 caregivers), caregiver burden was also described as substantial, especially for medication management, supervision, transportation, and household tasks. [20] Most caregivers were mothers (93.2%, mean age 41.5 years), and about one-third reported mental or psychiatric conditions. [20] Patients averaged 9.5 years and were reported to need extensive daily assistance, and respite care access was described as limited by availability and cost. [20]

For families living with Dravet syndrome, these studies add evidence that caregiver strain is not only an early phase around diagnosis, but can remain high for years and span practical tasks (supervision, transport, household work), seizure and medication routines, and caregiver well-being (sleep, fatigue, stress, emotional health). [19] [20] The research also suggests that experiences can differ between women and men and between countries, which may reflect differences in support systems and resources. [19] Because both studies were surveys at one point in time, they describe what caregivers reported but cannot show what causes changes in burden or which supports reduce it. [19] [20]