Three integrative cases — neonatal seizures, progressive ataxia, and a CP mimic — synthesizing testing, interpretation, and treatment.
Tags: Clinical Decision-Making · Neurogenetics
Presentation. Baby M is a 3-day-old term neonate who presents with focal tonic seizures, with apnea and desaturation, beginning on day 2 of life; they have not stopped on phenobarbital or levetiracetam. Continuous EEG monitoring reveals a burst-suppression pattern — prolonged periods of near-flat voltage suppression interrupted by high-amplitude bursts of mixed epileptiform activity. This is the most severe EEG signature of a neonatal epileptic encephalopathy: the suppression reflects pathological loss of the background activity that should be present even in a sleeping neonate, and its presence (rather than the seizures alone) is what defines the encephalopathy and predicts developmental impairment.
Why this matters for the differential. Burst-suppression in a term neonate has two broad explanations that must be separated early: an acquired static insult (hypoxic-ischemic injury, stroke, infection) versus a genetic/metabolic cause. The reassuring birth history and normal MRI below push strongly toward the latter — and within the genetic causes, burst-suppression specifically points to the early-infantile developmental and epileptic encephalopathies (the Ohtahara-spectrum phenotype) rather than benign familial neonatal epilepsy.
Birth history. Reassuringly uncomplicated: spontaneous vaginal delivery at 39 weeks with APGAR scores of 8 at one minute and 9 at five minutes. Normal APGARs and absence of a sentinel event make significant perinatal hypoxic-ischemic injury unlikely — an important point, because HIE is the diagnosis most often invoked (sometimes incorrectly) for any sick seizing neonate.
Exam. Baby M demonstrates:
The distinctive features are nonspecific here. They are worth noting but do not narrow the differential the way they would in a recognizable multiple-anomaly condition — in the neonatal epilepsies the face is usually unremarkable, and a normal-appearing infant should never lower suspicion for a channelopathy.
Family history provides important clues that must not be overlooked: Baby M's maternal grandfather, 66, has an action tremor and falls and carries a diagnosis of Parkinson disease, and his mother, 31, has had irregular menstrual cycles since her late twenties. Holding these in mind is the discipline this case rewards — an older man with late-onset tremor and falls and his daughter with ovarian dysfunction is a pattern, not noise, and it will matter later.
The clinical team now faces a critical testing decision — whether to order a targeted epilepsy gene panel or proceed directly to rapid trio whole exome sequencing.
Key Points
Why trio WES here. The team chooses rapid trio whole exome sequencing over a targeted epilepsy panel for a specific reason: in a critically ill neonate the speed of a confident answer changes therapy, and the trio design (proband plus both parents) lets the lab phase variants and immediately confirm de novo status. Because the great majority of severe neonatal epileptic encephalopathy is caused by de novo dominant variants, demonstrating that a variant is absent in both parents supplies PS2 — Strong under the generic ACMG rules, though ClinGen scales it by how specific the phenotype is — in the same report that finds the variant, rather than requiring a second round of parental Sanger confirmation that costs days. Rapid sequencing is now fast enough to change acute care: in Project Baby Bear, rapid genome sequencing diagnosed about 40% of critically ill infants with a median time to result of 3 days, and in early-onset epilepsy sequencing finds a cause in about 37% (Stefanski 2021). A panel would have been a reasonable, often cheaper alternative, but it returns a variant without parental context and is fixed to its gene list.
The key test. Rapid trio WES returns on day 12 of life and identifies a de novo heterozygous pathogenic variant in KCNQ2: c.881C>T, p.Ala294Val.
Mechanism — and why it predicts the drug. The variant sits in the S6 segment of the pore domain of the Kv7.2 subunit. Kv7.2 and Kv7.3 co-assemble as heterotetramers to carry the neuronal M-current, a slow, non-inactivating potassium current that acts as a brake on repetitive firing near threshold. Most KCNQ2-DEE variants reduce M-current (loss of function), but the exact variant matters. Baby M's variant has been studied directly (Abidi 2015): the p.Ala294Val subunit alone carries no current; in the heterozygous configuration, with wild-type Kv7.2 and Kv7.3, current falls by only about 30% — less than the severity of the disease would suggest — but the channels are mislocalized away from the axon initial segment, where they normally brake firing. At the same residue, p.Ala294Gly causes self-limited familial neonatal epilepsy, so the gene name, or even the position, does not tell you which child you have. With the brake reduced where it matters, neurons fire in runs — clinically, seizures. This is the mechanistic basis for treatment:
ACMG classification: Pathogenic. p.Ala294Val is a recurrent variant classified as pathogenic in ClinVar by multiple submitters with no conflicts (VCV000205886); Abidi 2015 described seven patients with it, all with early-onset epileptic encephalopathy. The evidence codes are worth reading as a logic chain, not a checklist:
Two Strong plus a Moderate already clear the Pathogenic threshold; PM1/PP3 reinforce rather than carry the call. Note that PS3 and PP3 are partly correlated (both speak to deleteriousness), so they should be weighted thoughtfully rather than treated as fully independent.
Management. Carbamazepine is initiated as a mechanism-matched first-line agent rather than escalating broad-spectrum anticonvulsants. Seizures resolve completely within 48 hours and phenobarbital is weaned — the kind of rapid, complete response that both treats the infant and corroborates the diagnosis.
Second diagnosis (family). The retained family-history clue now pays off. The grandfather's action tremor and falls, labelled Parkinson disease, and the mother's irregular cycles prompt targeted FMR1 testing of the mother, revealing an FMR1 premutation of 89 CGG repeats (normal <45, gray zone 45–54, premutation 55–200, full mutation >200). The premutation — not the full mutation — is what causes FXTAS via a toxic gain-of-function of the expanded CGG-containing mRNA (RNA toxicity). FXTAS develops in about 40% of male premutation carriers, usually after 60, which fits the grandfather's 'Parkinson disease' at 66; the mother, at 31, has about a 20% risk of fragile X-associated primary ovarian insufficiency (FXPOI), which fits her irregular cycles, and a lower risk of FXTAS later. Critically, this is mechanistically unrelated to Baby M's KCNQ2 disorder, and Baby M inherited his mother's normal FMR1 allele.
The teaching point is a trap to avoid: a confirmed diagnosis in the proband can anchor the team and shut down inquiry into a relative's unexplained symptoms. Here, two independent molecular diagnoses coexist in one family — and the maternal premutation carries its own counseling implications entirely separate from the index seizure case: her own FXPOI risk, and expansion risk in future pregnancies — maternal alleles of 80–89 repeats expanded to a full mutation in about 58% of transmissions (Nolin 2003), so about 29% per pregnancy before AGG interruption testing, which can refine the estimate. More than one diagnosis is not rare even within a single patient: Posey 2017 found more than one molecular diagnosis in about 4.9% of patients with a diagnostic exome, and families hold more.
Key Points
Presentation. Alex is a 14-year-old previously healthy teenager referred to the neurogenetics clinic with an 18-month history of progressive gait unsteadiness, frequent falls, and declining school performance. His parents first noticed difficulty during football practice when he began stumbling without apparent cause. Over the following months, his handwriting deteriorated and his speech became mildly slurred. The progressive trajectory over the better part of two years is itself a discriminator: it argues against a static or acute process and toward a degenerative, often inherited, disorder.
The most informative finding is the combination of cerebellar signs with absent reflexes plus large-fiber sensory loss. Pure cerebellar disease leaves reflexes intact; areflexia here signals a coexisting sensory/peripheral neuropathy and dorsal root ganglion (posterior column) involvement. A mixed cerebellar-plus-sensory-axonal picture with areflexia in a teenager is a specific, almost fingerprint pattern. (Alex's extensor plantar responses complete the picture — areflexia with upgoing toes is a classic, seemingly paradoxical combination that points hard at this diagnosis.)
Family history is significant: Alex's parents are first cousins of Pakistani origin, and his 17-year-old brother has been clumsy for two years. Consanguinity raises the prior probability of a rare autosomal recessive condition by making it more likely that a child inherits two copies of the same ancestral allele (identity by descent); an affected sibling with unaffected parents fits recessive inheritance and makes it the leading hypothesis.
The consanguinity plus the mixed cerebellar-sensory-areflexic phenotype now focuses the recessive ataxia differential, which must be evaluated systematically before committing to a test.
Key Points
The diagnostic pitfall. WES is ordered first, a defensible choice given the genuinely broad recessive-ataxia differential and consanguinity. It returns negative — no pathogenic or likely pathogenic variant in any ataxia gene. The decisive move is interpreting that negative correctly: a negative exome does not exclude Friedreich ataxia, because the disease is caused not by a coding point mutation but by an expanded GAA repeat in the first intron of FXN. Short-read exome sequencing both targets coding exons (often skipping deep-intronic regions) and, more fundamentally, cannot span or accurately count long, homogeneous repeat tracts — the reads are shorter than the expansion and map ambiguously. The phenotype here is too characteristic to abandon on a test that is mechanistically blind to the relevant mutation type.
The key test. Targeted FXN GAA repeat analysis confirms the diagnosis: Alex has biallelic GAA expansions of roughly 850 and 920 repeats (normal ≤33; premutation 34–65; full penetrance ≥66; most affected alleles carry 600–1,200 repeats). The analysis is a combination of methods: standard PCR shows normal-sized alleles (it fails to amplify across long expansions), repeat-primed PCR confirms that an expansion is present but does not size long alleles, and long-range PCR or Southern blot sizes them. The shorter of the two alleles correlates best with age at onset.
Mechanism — why the repeat causes the disease. The intronic GAA expansion does not change the frataxin protein sequence; it silences transcription of FXN (through heterochromatin formation/R-loop effects), so patients make too little of a normal protein. Frataxin is a mitochondrial protein essential for iron-sulfur cluster biogenesis; its deficiency impairs key respiratory-chain and aconitase enzymes and promotes oxidative stress. This loss-of-expression mechanism explains two clinical facts: longer GAA tracts cause earlier, more severe disease (more silencing), and the most vulnerable tissues are the high-energy, post-mitotic ones — dorsal root ganglia, spinocerebellar tracts, and cardiomyocytes.
Diagnosis: Friedreich ataxia, confirmed. Echocardiography reveals concentric left ventricular hypertrophy consistent with early hypertrophic cardiomyopathy. This is not incidental: hypertrophic cardiomyopathy develops in about two thirds of patients, and cardiac causes accounted for 59% of deaths in Tsou 2011 — cardiac, not neurological, disease is the leading cause of premature death in FRDA, which is why the murmur on the screening exam earned an echo.
Cascade testing. The 17-year-old sibling undergoes targeted FXN testing (the correct, inexpensive, mechanism-specific test — not a repeat exome). He is confirmed to carry biallelic expansions. At 17 he is within the omaveloxolone label, and catching him at an earlier stage allows surveillance and treatment to begin before advanced cardiomyopathy is established, illustrating why cascade testing within a family should use the known molecular target rather than re-running a broad, blind assay.
Key Points
Presentation. Priya is a 5-year-old girl referred from the cerebral palsy clinic for genetic evaluation. She was diagnosed with spastic diplegic cerebral palsy at age 2 based on gross motor delay, lower-extremity hypertonia, and persistent toe-walking. The early CP label is understandable — these features are real and common in CP — but the diagnosis was made on a cross-sectional snapshot, and CP is fundamentally a diagnosis about trajectory. The story only becomes diagnostic when the course is observed over time.
Birth history is unremarkable: term, uncomplicated spontaneous vaginal delivery, normal APGARs, no perinatal risk factors and no evidence of hypoxic-ischemic encephalopathy. CP usually has an identifiable antenatal or perinatal risk profile; its absence here weakens (without alone excluding) an acquired static etiology and should lower the threshold to look for a mimic.
Red flags — and why each one breaks the CP definition. The referring neurologist is concerned about features inconsistent with CP:
Taken together, progression + diurnal fluctuation + normal MRI + an affected relative is not atypical CP — it is a coherent argument against CP and for a treatable genetic mimic.
Key Points
Why treat before sequencing. With diurnal fluctuation and regression, the pretest probability of DRD is high enough that the most efficient and humane next step is a therapeutic trial of levodopa with carbidopa rather than waiting on genetics — in a child under 6 years, GeneReviews suggests 1–10 mg/kg/day of levodopa in divided doses, titrated over weeks. In DRD a small dose produces a large, sustained effect because the defect is in dopamine synthesis, not in the neurons or receptors themselves — the nigrostriatal pathway is structurally intact and merely substrate-starved, so replacing the downstream product restores function. (The same trial doubles as a safety net: levodopa-responsiveness is also seen in the dopamine-synthesis enzyme deficiencies, so a robust response is meaningful even before the exact gene is known; a partial response does not exclude a dopamine-synthesis defect, tyrosine hydroxylase deficiency in particular.)
Response. Within a week Priya's stiffness eases and she walks a few steps; by six weeks she walks without the walker and no longer declines through the afternoon. A response this complete and durable is, in practice, near-diagnostic of DRD. Unlike Parkinson disease, where a good initial response is followed over years by wearing-off, on–off fluctuations and dyskinesias, DRD stays responsive to low doses for decades without these motor complications (transient dyskinesias at the start settle with a lower dose).
Genetic confirmation. GCH1 sequencing with deletion/duplication analysis (some families carry exon deletions that sequencing alone misses) reveals a heterozygous pathogenic missense variant, NM_000161.3 c.607G>A p.(Gly203Arg), inherited from her asymptomatic father. GCH1 encodes GTP cyclohydrolase 1, the rate-limiting enzyme for tetrahydrobiopterin (BH4) synthesis; BH4 is the obligate cofactor for tyrosine hydroxylase, the rate-limiting step in dopamine production. A single defective allele lowers BH4 enough to throttle striatal dopamine — a dosage effect on a rate-limiting cofactor — which is why a heterozygous variant suffices to cause disease.
ACMG classification: Pathogenic (ClinVar VCV000381665: pathogenic, multiple submitters, no conflicts). What this one family adds is worth reading separately from what the laboratories already know:
These family-level codes alone would not reach Likely Pathogenic; the Pathogenic call rests on the laboratories' aggregated evidence across unrelated patients, which is why the ClinVar review status (several concordant submitters, no conflicts) matters when you read the report. A nuance worth teaching: her father carries the same variant and has no symptoms. That is reduced penetrance, a known property of GCH1-related DRD, not evidence against pathogenicity — benign evidence from healthy carriers should not be applied in a disorder known to be incompletely penetrant.
Cascade testing and a counseling nuance. Her 8-year-old brother should be tested for the variant and, if he carries it, offered the same levodopa trial; toe-walking may be his only sign. Her father needs no treatment now, but some adult carriers develop a levodopa-responsive parkinsonism, so he should know what to report. Each child of a carrier has a 50% chance of inheriting the variant. GCH1-DRD is autosomal dominant with reduced penetrance that is notably sex-biased — females are affected more often and more severely than males, so an obligate male carrier can be mildly affected or even asymptomatic. This skewed penetrance can mimic skipped generations or an apparently 'sporadic' case and is a classic counseling pitfall; family members should be assessed (and offered the levodopa trial) on the basis of carrier status, not on how striking their symptoms appear.
The big-picture lesson. A meaningful minority of children carrying a CP label have an identifiable genetic etiology — exome sequencing finds a cause in about 31% of children with CP overall and about 35% in cryptogenic CP (Gonzalez-Mantilla 2023; Srivastava 2022) — and a subset of those — DRD foremost among them — are treatable. DRD is often called one of the most rewarding diagnoses in child neurology precisely because an inexpensive, well-tolerated medication can convert lifelong disability into near-normal function. The case's enduring message is procedural: when a CP diagnosis shows progression, diurnal fluctuation, a normal MRI, or a suggestive family history, reopen the differential.
For detailed coverage of GCH1 and other genetic dystonias, see the Genetic Dystonias module.
Key Points
1. A 5-day-old neonate presents with multifocal seizures refractory to phenobarbital and levetiracetam. EEG shows burst-suppression. Brain MRI is normal. Metabolic workup is unremarkable. The team considers ordering genetic testing. What is the MOST important advantage of rapid trio WES over a singleton epilepsy gene panel in this scenario?
The key advantage of trio WES (proband + both parents) is the ability to identify de novo variants — variants present in the child but absent in both parents. In neonatal epileptic encephalopathy, the majority of causative variants arise de novo, and confirming de novo status provides strong ACMG evidence for pathogenicity (PS2; Strong by default, scaled by phenotype specificity under ClinGen guidance). A singleton gene panel can identify the same variants but cannot determine whether they are de novo without separate parental testing, which adds time and complexity. Both trio WES and gene panels can detect variants in channelopathy genes; neither can detect repeat expansions. The speed of de novo confirmation in a critically ill neonate can directly impact treatment decisions.
2. A neonate with KCNQ2 epileptic encephalopathy is started on phenobarbital and then levetiracetam, and the seizures continue. The team switches to carbamazepine based on the genetic diagnosis, and seizures resolve within 48 hours. Which concept does this case BEST illustrate?
This exemplifies precision medicine in epilepsy: identifying the specific genetic cause (KCNQ2 channelopathy) directly guides treatment selection. KCNQ2 encephalopathy variants disrupt the M-current that normally dampens repetitive neuronal firing. Sodium channel blockers (carbamazepine, oxcarbazepine, phenytoin) work because they reduce the hyperexcitability downstream of the M-current deficit: in 15 infants with KCNQ2 encephalopathy, carbamazepine controlled seizures in 6 and phenytoin in 5, often after other drugs had failed (Pisano 2015). This is distinct from pharmacogenomics (which concerns drug metabolism variants) — here the genetic diagnosis determines which drug mechanism is appropriate for the specific disease pathophysiology.
3. A 16-year-old with progressive gait ataxia, absent knee reflexes, pes cavus, and scoliosis has a normal whole exome sequencing result. His echocardiogram shows concentric left ventricular hypertrophy. His parents are first cousins. The neurologist orders targeted FXN GAA repeat testing, which reveals 750 and 900 GAA repeats on the two alleles. Which therapeutic intervention has been FDA-approved for this condition?
Omaveloxolone (Skyclarys) was approved by the FDA in February 2023 as the first treatment for Friedreich ataxia in patients aged 16 and older. It works by activating the NRF2 transcription factor, which upregulates antioxidant gene expression to counteract the oxidative stress caused by frataxin deficiency in mitochondria. In the MOXIe trial it slowed worsening on the modified Friedreich Ataxia Rating Scale (mFARS) by about 2.4 points over 48 weeks versus placebo (Lynch 2021). Liver enzymes (ALT, AST) and bilirubin are checked before treatment, monthly for 3 months and periodically thereafter, with BNP and lipids at baseline and periodically. The case also reinforces that WES cannot detect the GAA trinucleotide repeat expansion — about 96% of patients have biallelic GAA expansions — so targeted FXN repeat analysis is required.
4. A 4-year-old girl diagnosed with 'spastic diplegic cerebral palsy' at age 2 has been declining functionally — she previously walked with a walker but now needs a wheelchair. Her brain MRI is normal. Her mother notes she is 'a different child in the morning versus the evening.' Before ordering any genetic testing, the treating neurologist should:
The combination of progressive motor decline (incompatible with true CP, which is non-progressive), normal brain MRI, and clear diurnal fluctuation (better in morning, worse by evening) are red flags pointing strongly to dopa-responsive dystonia (DRD/Segawa disease). A therapeutic trial of low-dose levodopa should be started immediately and should NOT be delayed while awaiting genetic testing. The dramatic and sustained response to levodopa is itself virtually diagnostic. DRD is considered one of the most rewarding diagnoses in child neurology because an inexpensive, well-tolerated medication can transform a wheelchair-dependent child into an independently walking one. Exome sequencing finds a genetic cause in about 31% of children with CP overall and about 35% in cryptogenic CP (Gonzalez-Mantilla 2023; Srivastava 2022), and DRD is the quintessential treatable mimic.
5. Trio WES for a critically ill neonate identifies a de novo pathogenic STXBP1 variant. Separately, targeted FMR1 testing of the mother, prompted by her father's late-onset tremor and falls, shows an FMR1 premutation (85 CGG repeats). This dual finding illustrates which important principle in clinical neurogenetics?
Independent genetic findings can coexist in one family, and more than one diagnosis is not rare even within a single patient: Posey 2017 found more than one molecular diagnosis in about 4.9% of patients with a diagnostic exome. In this scenario, the infant's seizures are caused by the de novo STXBP1 variant, completely independent of the mother's FMR1 status, while the mother's FMR1 premutation (55–200 CGG repeats) confers her own set of risks — fragile X-associated primary ovarian insufficiency (FXPOI), a later risk of fragile X-associated tremor/ataxia syndrome (FXTAS), and the risk of transmitting a full mutation: maternal alleles of 80–89 repeats expanded to a full mutation in about 58% of transmissions (Nolin 2003), so about 29% per pregnancy before AGG interruption testing — not every transmission, and there is no treatment to start. Note that an exome does not detect FMR1 CGG repeat expansions; the premutation is found by targeted FMR1 testing. Each finding must be evaluated independently, and establishing one genetic diagnosis in a family should not prevent investigation of unexplained symptoms in other family members.
6. A clinical geneticist is evaluating a child with neonatal epileptic encephalopathy. The WES report shows a de novo KCNQ2 missense variant classified as Likely Pathogenic. The variant is in the S4 voltage-sensor domain, absent from gnomAD, and computational tools predict it is deleterious. What additional evidence would MOST likely upgrade this variant to Pathogenic?
The evidence listed already reaches Likely Pathogenic: confirmed de novo status (PS2, Strong, 4 points), a critical functional domain (PM1, Moderate, 2), absence from population databases (PM2_Supporting, 1) and concordant computational predictions (PP3, 1) total 8 points, inside the Likely Pathogenic range of 6 to 9 (Tavtigian 2018; Richards 2015 combining rules). Well-established functional studies showing that the variant alters Kv7.2 channel behavior (PS3, Strong, 4 points) would bring the total to 12, Pathogenic (10 or more). A population frequency of 0.01% is benign-direction evidence, parental status is already counted in PS2, and a microarray addresses a different question.