Epilepsy with Myoclonic-Atonic Seizures

1. Disease Information

EMAtS is a childhood-onset generalized epilepsy syndrome defined by the presence of myoclonic-atonic (and often pure atonic, myoclonic, absence, and generalized tonic-clonic) seizures, arising in a previously typically-developing young child. It sits in that fascinating middle zone between the "self-limited" epilepsies (kid grows out of it, no scars) and the "developmental and epileptic encephalopathies" (the seizures themselves chew up development) — and which side a given child lands on is genuinely hard to call at onset. The 2025 Lancet Neurology review put it plainly:

"Although two-thirds of children attain remission from seizures without cognitive or behavioural sequelae, some continue to have drug-resistant seizures, intellectual disability, and behavioural problems." — Guerrini, Scheffer & Balestrini, Lancet Neurol 2025;24:348–360, PMID:40120618

Key identifiers: - Orphanet: ORPHA:1942 (solid) - OMIM: 616421 (MYOCLONIC-ATONIC EPILEPSY; MAE — the GABRG2-associated molecular entry). Note OMIM treats this as a molecularly-defined slot, not the whole clinical syndrome. - MONDO: search surfaced MONDO:0014633 (MalaCards) but I could not verify this against OAK — ⚠️ run runoak -i sqlite:obo:mondo before trusting it. The classic "myoclonic-astatic epilepsy" MONDO node may differ; verify before setting disease_term. - ICD-10: G40.4 (other generalized epilepsy and epileptic syndromes); ICD-11: 8A61.x (generalized epilepsies) - MeSH: "Myoclonic-Astatic Epilepsy" / Doose syndrome (Epilepsies, Myoclonic subtree)

Synonyms: Doose syndrome; myoclonic-astatic epilepsy (MAE); myoclonic astatic epilepsy of early childhood; epilepsy with myoclonic-atonic seizures (EMAtS/EMAS). "Astatic" and "atonic" are used interchangeably in the drop-attack sense.

Data derivation: Almost entirely disease-level aggregated — small-to-medium retrospective clinical cohorts and case series, not EHR-mined patient records. The largest recent evidence base is multicenter retrospective cohorts (dozens to a few hundred children).

2. Etiology

Here's the honest headline: most cases are still "genetic, cause unknown." Doose himself pegged it as idiopathic/genetic generalized epilepsy, and that framing has held up — a substantial fraction of kids have a family history of epilepsy or febrile seizures, consistent with a complex/polygenic background rather than one broken gene. Layered on top of that polygenic soup is a growing list of monogenic causes that produce an EMAtS-like picture.

Primary causal factors: - Genetic (monogenic subset): SLC6A1 is the standout MAE gene — loss of function in the GABA transporter GAT-1. GABRG2, SCN1A, SCN1B, SLC2A1 (GLUT1), STX1B, CHD2, SYNGAP1, KCNA2, and others show up across cohorts. The 2015 discovery paper:

"Targeting resequencing of 644 individuals... six SLC6A1 mutations in seven individuals, all of whom have epilepsy with myoclonic-atonic seizures (MAE)... pathogenic mutations occurred in 6/160 individuals with MAE, accounting for ∼4% of unsolved MAE cases." — Carvill et al, Am J Hum Genet 2015, PMID:25865495 - Genetic (polygenic): the majority — inferred from twin/family aggregation, no single Mendelian locus. - Metabolic: ~5% are GLUT1 deficiency (SLC2A1) — this one matters clinically because it's treatable with the ketogenic diet, so it must be actively excluded. - Environmental/infectious: none established as causal. This is not an acquired or structural epilepsy — normal MRI is part of the definition.

Risk factors: young age (2–5 yr window), male sex, and prior febrile seizures (~25% of kids). Family history of epilepsy is a susceptibility signal.

Protective factors / gene-environment interactions: not well characterized. No protective alleles or dietary/lifestyle protective factors are documented. GLUT1's ketogenic-diet responsiveness is the closest thing to a gene-treatment interaction, but that's therapeutic, not preventive.

3. Phenotypes

The defining move is the myoclonic-atonic seizure — a symmetric myoclonic jerk (often trunk/shoulders/arms) immediately followed by loss of tone, producing a drop attack (falls, head nods, buckling knees). But EMAtS is a seizure buffet, and different types dominate at different points. From the Japanese Doose cohort (Nickels-style breakdown), PMID:32913952:

At onset: generalized tonic-clonic 41%, tonic seizures 38%, myoclonic 24%, myoclonic-atonic 14%. During course: myoclonic 48%, absence 45%, atonic 24%, nonconvulsive status epilepticus 14%.

Characteristics: onset 6 months–6 (some say 8) years, peaking 2–4 yr; development typically normal before onset in ~two-thirds; severity highly variable (self-limited → drug-resistant DEE); course episodic/fluctuating, sometimes with "stormy" onset periods of near-continuous drops. From epilepsydiagnosis.org: "Developmental stagnation or regression is typically seen during the phase of active seizures."

Quality of life: driven by drop attacks (injury risk, helmet use), cognitive/behavioral load, and drug resistance in the unlucky third. No EMAtS-specific EQ-5D/SF-36 data surfaced — flag as not available.

4. Genetic / Molecular Information

The molecular story is a "many roads into the same town" situation, and the roads mostly run through GABAergic inhibition and ion channels.

Marquee gene — SLC6A1 (GAT-1, HGNC verify hgnc:11042): - Encodes the sodium/chloride-dependent GABA transporter type 1, which vacuums GABA back out of the synaptic and extrasynaptic space. - Variant classes: missense (most), nonsense, frameshift, splice, and whole-gene/translocation — converging on loss of function (reduced GABA reuptake, but also protein misfolding/destabilization and ER retention for some missense alleles). - ~4% of unsolved MAE (Carvill 2015); most common single-gene MAE cause. De novo dominant, mostly.

Other genes across cohorts: - GABRG2 (GABA-A receptor γ2 subunit) — OMIM 616421's assigned gene; GABAergic again. - SCN1A / SCN1B (sodium channels) — overlap with the Dravet spectrum; a caution flag for named-entity confusion. - SLC2A1 (GLUT1) — ~4–5%; loss of function → CNS glucose-transport failure; the treatable one. - STX1B — syntaxin-1B, presynaptic vesicle fusion; haploinsufficiency causes MAE-like epilepsy. - CHD2, SYNGAP1, KCNA2, HNRNPU and, in the newest cohort (PMID:41523187), a widened net: ANKRD11, CSNK2B, NEXMIF, POLR3B, plus novel associations KMT2E, POGZ, SHANK3, YWHAG. That cohort's yield:

"15/39 patients (38.5%) who underwent next-generation sequencing had pathogenic variants."

So NGS yield in a well-selected modern cohort is roughly a third to 40%, but the classic candidate-gene panels (SCN1A/GABRG2/SLC2A1) are individually low-yield.

Chromosomal: microdeletions (e.g., involving SCN1A, STS) and a reported 4q21.22-q21.23 microduplication; balanced translocations disrupting SLC6A1.

Epigenetics / modifiers: no established EMAtS-specific methylation signature or modifier gene — not available. The polygenic background is effectively the modifier layer, but it's uncharacterized at the locus level.

Ontology anchors: GO:0015812 (GABA transport), GO:0051932 (GABAergic synaptic transmission), GO:0007214 (GABA signaling pathway), GO:0005328 (neurotransmitter:sodium symporter activity).

5. Environmental Information

Short section, and that's the finding: EMAtS has no established environmental, lifestyle, toxic, or infectious cause. Febrile seizures precede it in ~25%, but fever is a trigger/marker of susceptibility, not an environmental etiology. No occupational, dietary, or pollution links. Normal neuroimaging and no acquired insult are baked into the diagnostic definition. Mark §5 = not applicable / not available.

6. Mechanism / Pathophysiology

The through-line is failure of GABAergic inhibition in the thalamocortical circuitry, tipping the cortex toward generalized hypersynchronous discharge.

Causal chain (canonical, SLC6A1 exemplar): 1. Trigger: loss-of-function variant in GAT-1 (SLC6A1) → GABA not efficiently cleared from synaptic/extrasynaptic space. (Counterintuitively, more ambient GABA can be pro-seizure here because tonic GABA-A currents and receptor desensitization dysregulate thalamocortical rhythms — the same paradox seen in absence epilepsy.) 2. Cellular: GAT-1 lives mainly on astrocytes and GABAergic nerve terminals, so the defect degrades the astrocyte-neuron GABA homeostasis loop. From the mechanism literature: "GAT-1... is expressed mainly in astrocytes and the terminals of GABAergic neurons, where it regulates GABA levels in the synaptic and extrasynaptic compartments." 3. Circuit: dysregulated tonic inhibition → abnormal thalamocortical oscillation → generalized 2–4 Hz spike-and-slow-wave / polyspike-wave discharges (the EEG hallmark). Gat1-null mice recapitulate spontaneous spike-wave discharges — a nice cross-species anchor (PMID:25865495). 4. Clinical output: the spike drives the myoclonic jerk, the trailing slow wave drives the atonic drop. Absence and GTC seizures emerge from the same generalized-network instability.

GLUT1 branch: SLC2A1 LOF → impaired glucose flux across the blood-brain barrier → chronic cerebral energy deficit → seizures + movement/cognitive features. Mechanistically distinct (an energy-metabolism failure, not a channel/transporter-of-GABA failure), which is exactly why the ketogenic diet — supplying ketone bodies as an alternate brain fuel — works so well for it.

Involved cell types / regions: CL:0000617 (GABAergic neuron), CL:0000127 (astrocyte), CL:0000498 (inhibitory interneuron); UBERON:0000956 (cerebral cortex), UBERON:0001897 (thalamus), UBERON:0002037 (cerebellum, for the ataxia thread). No immune, fibrotic, or neurodegenerative mechanism — this is a channelopathy/synaptopathy of inhibition, a good conformance candidate for your epilepsy_excitation_inhibition_imbalance module (#Excitation-Inhibition Imbalance).

7. Anatomical Structures Affected

• Organ/system: central nervous system, generalized — no focal lesion. Primary structure is the thalamocortical network (cortex + thalamus). Secondary: cerebellar circuits (ataxia).
• Tissue/cell: cortical and thalamic neurons, GABAergic interneurons, and astrocytes (GAT-1 expression site).
• Subcellular (GO Cellular Component): GO:0045202 (synapse), GO:0043195 (terminal bouton / presynaptic terminal), GO:0005886 (plasma membrane — where GAT-1/channels sit), GO:0098982 (GABA-ergic synapse). For GLUT1: GO:0005886 at the BBB endothelium.
• Localization/laterality: bilateral, symmetric, generalized by definition — the EEG discharges are bisynchronous. Persistent focal spikes argue against the diagnosis.

8. Temporal Development

• Onset: early childhood, 6 mo–6 yr, peak 2–4 yr; typically abrupt/subacute ("usually begins abruptly, with frequent generalized seizures... between 2–6 years of age"), sometimes heralded by febrile or afebrile GTC seizures.
• Active phase: often a "stormy" 1–3 year period of frequent daily drops, sometimes with episodes of nonconvulsive status ("minor epileptic status") that dents cognition transiently.
• Course pattern: episodic/fluctuating during the active phase, then commonly self-limited — the "two-thirds remit" figure. From the outcomes cohort (PMID:41523187): "61.7% achieved seizure freedom after mean 5.1 years." From epilepsydiagnosis.org: "Two thirds of children achieve epilepsy remission, usually within 3 years of epilepsy onset."
• Critical window: the active-seizure phase is the intervention window — controlling drops and status early (right drug/diet, avoiding aggravators) appears to protect development. Interestingly, the outcomes cohort found "'Stormy' onset did not predict worse prognosis" — it's the baseline developmental delay, not the seizure intensity, that flags trouble.

9. Inheritance and Population

Epidemiology: - Incidence: ~16.4 per 100,000 children (one population estimate). - Share of childhood epilepsy: 1–2.2% of childhood-onset epilepsies; ~5.5% of generalized epilepsies in 1–9 year olds. - Prevalence: not precisely known (rare disease; Orphanet lists it as rare). - Normalized for your Prevalence slots: ANNUAL_INCIDENCE, rate_per_100000: 16.4, population "children," prevalence_class ~BAND_1_5_PER_10000 if reasoning from incidence + short active duration (⚠️ but incidence ≠ prevalence — keep them in separate records; don't cross the streams).

Sex ratio: male predominant, ~2:1 to 3:1 (M:F); Orphanet cites 2.7–3.1:1. The Japanese cohort was 21:8 (~2.6:1); the outcomes cohort was 26.7% female (~2.75:1 M).

Inheritance (genetic subset): - Pattern: mostly complex/polygenic; monogenic cases are usually autosomal dominant, de novo (SLC6A1, GABRG2, STX1B, SLC2A1, SCN1A). - Penetrance/expressivity: highly variable expressivity even within a gene — SLC6A1 alone spans MAE, milder GGE, and focal epilepsy with intellectual disability. - Anticipation / germline mosaicism / founder effects / carrier frequency: not established for EMAtS specifically — not available (de novo dominant biology makes classic carrier-screening framing largely N/A). - Consanguinity: not a notable feature (dominant/de novo, not recessive).

Demographics: no strong ethnic enrichment reported; described across European, North American, and Asian cohorts.

10. Diagnostics

EMAtS is a clinical-electroencephalographic diagnosis of inclusion + exclusion — there's no single confirmatory test, and genetics is confirmatory only in the monogenic subset.

Core clinical + EEG criteria (Ren et al 2021 modification, PMID:34883415; per ILAE 2022 nosology): 1. Normal development/cognition before onset; 2. Onset ~6 mo–6 yr (peak 2–4); 3. Myoclonic-atonic seizures mandatory (plus atonic/myoclonic drop attacks); 4. Generalized 2–3 Hz (up to ~4 Hz) spike-wave / polyspike-wave on EEG, without persistent focal spikes; 5. Exclusion of other myoclonic epilepsies (Dravet, LGS, epileptic spasms, progressive myoclonic epilepsies).

Tests: - EEG (the workhorse): normal or theta-rich background early; generalized 2–4 Hz spike/polyspike-wave; characteristic biparietal/central theta rhythm (seen in ~69% of the Japanese cohort). ⚠️ Predictors of poor outcome: "slow (<2.5Hz) spike wave or generalized paroxysmal fast activity on EEG" (the latter smells more like LGS). - MRI: normal (part of the definition; abnormal imaging → reconsider). - Genetic testing: gene panel or exome/genome sequencing is now recommended, given ~⅓–40% yield and management implications (SLC6A1, GLUT1). Single-gene testing is low-yield except targeted GLUT1 workup. - CSF glucose / CSF:blood glucose ratio (± SLC2A1 sequencing): to catch GLUT1 deficiency — cheap, high-stakes, don't skip it. - Metabolic/lactate workup if a progressive myoclonic epilepsy or mitochondrial mimic is on the table.

Differential diagnosis (the "rule these out" list): Dravet syndrome (SCN1A, but febrile/hemiclonic, worse trajectory), Lennox-Gastaut syndrome (tonic seizures in sleep, slow <2.5 Hz spike-wave, GPFA), epilepsy with eyelid myoclonia, myoclonic epilepsy in infancy, and progressive myoclonic epilepsies.

LOINC/ontology: EEG → the electrophysiology bucket; MAXO diagnostic terms exist for EEG (verify). No validated blood biomarker.

11. Outcome / Prognosis

The prognosis is genuinely bimodal, and that bimodality is the most clinically important thing about this disease.

• Seizure remission: ~two-thirds remit, often within ~3 years of onset; 61.7% seizure-free after mean 5.1 yr in the outcomes cohort (PMID:41523187).
• Cognition: roughly 40–44% keep normal cognition; the outcomes cohort reported "58.3% had intellectual disability; 43.7% had normal cognition," and "38.3%" drug-resistant. The Japanese cohort was a bit rosier (41% normal IQ).
• Mortality: low; not a classically high-mortality epilepsy, though drug-resistant DEE carries the usual SUDEP and injury risks. No EMAtS-specific mortality rate surfaced — flag as limited data.
• Morbidity: drop-attack injuries (helmets), behavioral comorbidity (ADHD ~40%, the most common), learning problems.

Prognostic indicators (from PMID:41523187):

"Global developmental delay at epilepsy onset was associated with drug resistance and with intellectual disability." - Early dual-domain (motor + language) delay → worse outcome. - Identified monogenic aetiology correlated with higher ID rates (i.e., a positive genetic finding tends to flag the harder-course kids). - "Stormy" onset did NOT predict worse prognosis — counterintuitive but repeatedly noted. - Tonic seizures, GPFA, and slow (<2.5 Hz) spike-wave lean toward the LGS-like, worse-outcome end.

12. Treatment

Treatment is broad-spectrum antiseizure meds + ketogenic diet, with a hard rule about which drugs to avoid because they make generalized epilepsies worse.

First-line pharmacotherapy: - Valproate / valproic acid — the consensus first-line (CHEBI:39867). Japanese cohort: "[valproate] was efficacious in 23 patients (79%)." An international Delphi consensus endorsed valproate + clobazam first-line. - Clobazam (benzodiazepine; CHEBI:31413 verify) — first-line partner. - Ethosuximide (CHEBI:4887 verify) — good for the absence component. - Levetiracetam, lamotrigine, topiramate, zonisamide, clonazepam — common add-ons (clonazepam for myoclonus).

Ketogenic diet — the star second-line (and arguably should be earlier), MAXO:0000088 (dietary intervention) / consider a ketogenic-diet-specific MAXO term (verify). International consensus: "the ketogenic diet identified as the optimal second-line treatment." Mandatory and curative-ish if GLUT1 is the cause. Case data show seizure freedom at ~2.5:1 ratio with BHB 4–7 mmol/L.

⚠️ Contraindicated / aggravating (drop-attack worseners): - Carbamazepine, oxcarbazepine, phenytoin, vigabatrin (and often gabapentin) — these can worsen myoclonic/atonic/absence seizures in generalized epilepsy. This is a genuine "first, do no harm" curation point.

Precision / emerging: - GLUT1 (SLC2A1): ketogenic diet is targeted therapy. - SLC6A1: antisense oligonucleotide and gene-based programs are in preclinical/early development (a real "personalised treatment" frontier the Lancet review flags). - Supportive: injury prevention (helmets), developmental/behavioral support, ADHD management.

MAXO anchors: pharmacotherapy (NCIT:C15986 for the therapeutic-agent pattern), MAXO:0000088 dietary intervention, MAXO:0000950 supportive care.

13. Prevention

Not a preventable disease in the classic sense — no primary prevention (no vaccine, no modifiable exposure). What exists: - Secondary prevention: early recognition + prompt broad-spectrum treatment and early GLUT1 exclusion to start the ketogenic diet before energy-deficit damage accrues — this is the highest-value "prevention" lever. - Tertiary prevention: avoiding aggravating drugs, controlling nonconvulsive status, injury protection, developmental/behavioral support to limit encephalopathic sequelae. - Genetic counseling: relevant for the monogenic subset (mostly de novo → low sibling recurrence, but reproductive counseling still warranted); NSGC/ACMG framing. - Population screening / immunization / public-health interventions: not applicable.

14. Other Species / Natural Disease

• Taxonomy: the meaningful non-human data is the mouse (Mus musculus, NCBITaxon:10090) — specifically the Gat1 (Slc6a1)-null mouse, which shows spontaneous spike-wave discharges and shared electrophysiology with human MAE (Carvill 2015). More recent patient-derived SLC6A1 knock-in mice (e.g., S295L) model the electrobehavioral phenotype and drug sensitivity.
• Orthologs: Slc6a1, Gabrg2, Scn1a, Slc2a1 are all conserved in mouse/rat/zebrafish.
• Natural disease in animals: no recognized spontaneous "Doose syndrome" in companion animals or wildlife (OMIA search: nothing specific) — not applicable. Zoonosis: N/A.

15. Model Organisms

• Primary model: mouse — Slc6a1/Gat1 knockout and patient-variant knock-in lines are the workhorses for the GABA-transporter mechanism (spontaneous SWDs, tremor, behavioral phenotypes). Evidence source = MODEL_ORGANISM for any of this.
• GABRG2 mouse models (e.g., R43Q knock-in) reproduce generalized spike-wave/absence-type activity.
• Zebrafish (Danio rerio, ZFIN) and iPSC-derived neurons are emerging for SLC6A1 functional screening (trafficking/surface-expression assays) — good IN_VITRO evidence for the misfolding/LOF mechanism.
• Recapitulation vs limitations: mice capture the EEG SWD signature and some behavior well, but the human myoclonic-atonic drop attack and the developmental-cognitive trajectory don't map cleanly onto rodents — a HUMAN_MODEL_MISMATCH flag is apt if you curate model evidence for the cognitive phenotype.
• Resources: MGI, IMPC/KOMP for the knockout lines; Cellosaurus/ATCC for patient iPSC lines.

Curation notes & flags for you

A few things worth doing before this lands in kb/disorders/: - ⚠️ Verify the MONDO ID with OAK — MONDO:0014633 came from MalaCards and I couldn't confirm it. The seed-cache-in-both-checkouts dance from your memory ([new-mondo-term-ols-cache-miss]) probably applies. - ⚠️ Every HPO/CHEBI/GO/CL term above is a suggestion — run just validate-terms-file before trusting labels (the myoclonic-atonic HP term especially). - ⚠️ NEC watch: this disease is exactly the high-risk profile — shared genes with Dravet (SCN1A) and phenotype overlap with LGS. When you pull DR content, confirm the report is about EMAtS and not a SCN1A-Dravet paper wearing a trenchcoat. - This is a clean conformer for your epilepsy_excitation_inhibition_imbalance module (#Excitation-Inhibition Imbalance), with the GABAergic-inhibition-failure arm as the disease-specific substitution. - The avoid-carbamazepine/phenytoin/vigabatrin point is a "side-effect-as-mechanism" nugget worth encoding as a treatment caution, not just prose.

Best single anchor references (all real, PMIDs verified by title): - Guerrini, Scheffer, Balestrini. Lancet Neurol 2025 — PMID:40120618 (the current authoritative review) - Genetic aetiologies/outcomes cohort, Brain Commun 2025/2026 — PMID:41523187 - Carvill et al, Am J Hum Genet 2015 (SLC6A1 discovery) — PMID:25865495 - Ren et al, Eur J Paediatr Neurol 2021 (diagnostic criteria) — PMID:34883415 - Doose clinical/genetic cohort, 2020 — PMID:32913952 - Zuberi et al, ILAE nosology, Epilepsia 2022 — PMID:35503712 (neonate/infant; the childhood-onset companion by Specchio et al covers EMAtS formally — pull that PMID when you cite the exact ILAE definition)

Sources: - Lancet Neurology 2025 review - Genetic aetiologies/outcomes cohort (PMC12782104) - Carvill et al, SLC6A1 (PMC4570550) - Clinical & genetic characteristics of Doose syndrome (PMC7469791) - Ren 2021 diagnostic criteria modification - ILAE 2022 nosology (Zuberi et al) - epilepsydiagnosis.org — EMAtS overview - Orphanet ORPHA:1942 - OMIM 616421 — MAE - SLC6A1 — Epilepsiome/Beyond the Ion Channel - Treatment effectiveness review — NCBI Bookshelf NBK581164 - Astrocytic GAT-1 deficit mechanism (PMC9472560)

That's the whole organism, dissected. Want me to kick off the actual /curate run for this one on the curate/doose-mae branch, or turn this into a first-pass disorder YAML skeleton so you can start filling evidence blocks?