Broadening the phenotype associated with pathogenic variants in the FGF12 gene: From developmental and epileptic encephalopathy to drug‐responsive epilepsy with favorable cognitive outcome

Abstract

The fibroblast growth factor 12 (FGF12) gene encodes a protein interacting with voltage‐gated sodium channels. Two variants, p.(Arg52His) and p.(Gly50Ser), have repeatedly been associated with developmental and epileptic encephalopathy‐47 (DEE47; Mendelian Inheritance in Man #617166) with poor outcome. We aim to refine the electroclinical phenotype and outcomes of 10 unpublished patients (2–38 years old) with these recurrent pathogenic variants in the FGF12 gene without DEE (p.[Arg52His], n = 4; p.[Gly50Ser], n = 6). The patients with p.(Gly50Ser) showed later and more explosive epilepsy onset, whereas p.(Arg52His) cases had gradual onset. All developed epilepsy before 5 months, with 70% achieving seizure remission by 6 months with antiseizure medication (ASM), leading to good neurodevelopmental outcomes (median follow‐up = 6.8 years). In contrast, the patients with mild intellectual disability had persistent epilepsy despite ASM. Additionally, patients with favorable neurodevelopmental outcomes and FGF12 pathogenic variants showed no signs of cerebellar atrophy. Moreover, we did not find a clear correlation between treatment with sodium channel blockers, its timing, and neurodevelopmental outcome. Here, we expand the phenotypic spectrum of FGF12 pathogenic variants and underscore cases with favorable neurodevelopmental outcomes.

1. INTRODUCTION

The FGF12 gene, formerly known as FHF1, encodes a small cytoplasmic protein that interacts with the C‐terminal domain of the alpha subunit of voltage‐gated sodium channels (Na_Vs) Na_V1.2, Na_V1.5, and Na_V1.6. ¹ , ² , ³ This interaction is critical for regulating neuronal excitability by delaying the fast inactivation of these channels. ¹ , ² , ³ , ⁴

The pathogenic variants of FGF12 are associated with developmental and epileptic encephalopathy‐47 (DEE47; Mendelian Inheritance in Man #617166). The term DEE refers to early onset genetic epilepsies related to severe cognitive and behavioral impairments above and beyond what might be expected from the underlying pathology alone. ⁵ , ⁶

This phenotype is mainly caused by two de novo heterozygous missense variants: NM_004113.6:c.155G>A;NP:004104.3:p.(Arg52His) ⁷ , ⁸ usually referred as p.(Arg52His) and NM_004113.6:c.148G>A;NP:004104.3:p.(Gly50Ser) ⁸ , ⁹ usually referred as p.(Gly50Ser).

The most extensive published FGF12 case series, by Trivisano et al., ⁸ gathered 17 patients including 14 patients carrying the p.(Arg52His) variant and two patients with the p.(Gly50Ser) variant.

With the growing accessibility of genetic testing, more patients with nonstructural epilepsy, especially early onset cases with uncertain prognoses, now undergo testing. This has significantly improved diagnosis rates and streamlined clinical management and parental counseling. However, the phenotypic variability of patients with these recurrent variants in the FGF12 gene remains poorly characterized.

This study reports 10 unpublished children presenting with a recurrent p.(Arg52His) (n = 4) or p.(Gly50Ser) (n = 6) variant in the FGF12 gene. We aim to extend the phenotypic spectrum associated with FGF12 variants and enhance our understanding of genotype–phenotype correlations.

2. MATERIALS AND METHODS

This national retrospective multicentric study examined patients with FGF12 variants and early onset epilepsy without a DEE phenotype.

Data from neurology departments across six French cities (Besançon, Lyon, Marseille, Nantes, Nice, and Paris) included demographic data, family history, seizure characteristics (age at onset, semiology, duration, efficiency of antiseizure medications [ASMs]), electroencephalographic (EEG) findings, brain imaging results, clinical examination, and the neurodevelopmental trajectory of the patients. Genetic analyses were conducted with informed consent. Clinical care data were repurposed for research. The genetic analyses were carried out in two specialized centers. Patients C, F, and G were analyzed using a 172‐gene epilepsy panel (SureSelect XT All Exons v8, Agilent), with sequencing performed on a NovaSeq6000 system (Illumina). The remaining patients were tested using a 190‐gene panel for monogenic epilepsies (KAPA Hyper Prep kit, Roche), sequenced on a NextSeq500 platform (Illumina). Variant calling and interpretation were performed using validated pipelines (Dragen and GATK, respectively) and single nucleotide variant (SNV) in the FGF12 gene (NM_004113.6; NP_004104.3), classified according to American College of Medical Genetics and Genomics (ACMG) criteria. ¹⁰ Technical details are available upon request. All variants were identified with sufficient read depth and allelic balance, consistent with heterozygosity. No evidence of somatic mosaicism was observed. We cannot formally exclude germline mosaicism, although there was no family history of epilepsy in any case except for Patients E and I (mother–daughter pair).

3. RESULTS

We report eight unrelated patients, along with two related patients (Patient E and her daughter, Patient I), with early onset epilepsy without DEE and pathogenic FGF12 SNVs. Clinical, EEG, neuroimaging, and genetic details are provided in Table 1. At the last follow‐up, the median age was 7.6 years (range = 2–38 years), and 70% were female. Figure 1 summarizes ASM usage, epilepsy onset, intellectual outcomes, and seizure resolution when applicable.

TABLE 1.

Phenotypic and genotypic information for the 10 patients with pathogenic FGF12 variants.

Characteristic	Patient
	A	B	C	D	E	F	G	H	I	J
Gender/age at last examination	M/4 years 4 months	F/22 years	F/2 years	F/2 years 1 month	F/38 years	M/4 years	F/8 years 5 months	F/6 years 10 months	M/9 years 10 months	F/10 years 4 months
FGF12 variant	c.155G>A p.(Arg52His) Likely de novo	c.155G>A p.(Arg52His) De novo	c.155G>A p.(Arg52His) De novo	c.155G>A p.(Arg52His) De novo	c.148G>A p.(Gly50Ser) De novo	c.148G>A p.(Gly50Ser) De novo	c.148G>A p.(Gly50Ser) De novo	c.148G>A p.(Gly50Ser) De novo	c.148G>A p.(Gly50Ser) Inherited from the mother (Patient E)	c.148G>A p.(Gly50Ser) De novo
Epilepsy
Epilepsy onset	3 days	30 days	38 days	48 days	3 months	3 months 4 days	3 months 11 days	4 months	4 months	4 months 9 days
Seizure type	TS, GTCS	TS, FS	TS	FS, GTCS	GTCS	TS, FS	TS, FS, AS	TS, FS, GS	GTCS	TS, FS, GTCS
Onset seizure semiology	TS	TS	TS, cyanosis	Staring trismus, epistaxis, and apnea	GTCS	Staring Oral automatism (chewing) Flexion of right arm and hypertonia of left arm Bilateral TS (upper limbs)	Staring Oral automatism (chewing) Cluster of short AS	Staring, pallor, head and eyes deviation to the left, left hemicorporeal clonus, ±generalization	Hypersalivation, GTCS	TS, GTCS
Recurrent seizure semiology	Seizure‐free period of almost 2 years 2.1 years old: cluster of GTCS during viral febrile episode	Childhood: oculocephalic deviation, groaning, gestural automatism, fear; followed by a generalization with right‐side predominance Adulthood: loss of contact	TS	2 months: cluster of seizures with staring, oral automatism (chewing), trismus, slight tonic–clonic seizure	GTCS	n.a.	5 years old: brief TS	Hypertonia with elevation of upper limbs, tremors, followed by staring, oral automatism (chewing)	GTCS	Staring, TS
Epilepsy type	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy	Generalized epilepsy	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy	Combined focal and generalized epilepsy
SE (frequency)	Rare, responsive to PHE	Rare (2 SE episodes at 18 years)	Rare, responsive to PHE	No	No	No	At diagnosis, responsive to PHE	At diagnosis	No	Rare (2 SE episodes), responsive to PHE
Resolution/age	Yes/3 years 5 months	No	No	Yes/4 months	Yes/22 years	Yes/5 months	Yes/5 years	Yes/6 years	Yes/6 years 9 months	No
Frequency	Monthly (between 2.1 and 3.4 years old)	Childhood: unknown Adult period: sporadic	Monthly	Sporadic	Sporadic	Sporadic	Sporadic	Sporadic	Sporadic	Sporadic
EEG
Interictal EEG	Normal	Normal	Initially normal 2 years old: Wake: slow BG at 4.3 Hz Sleep: right frontal spike	Normal BG in wake and sleep Sleep: right posterior central SW and rare spikes	Usually normal Postictal EEG (22 years): right temporo‐occipital SW	Wake: normal Sleep: frontal SW	Normal	Normal BG in wake and sleep Sleep: left frontotemporal spikes activation	Normal	Normal BG in wake and sleep Left temporal SW
Ictal EEG	n.a.	Left spikes and fast activities with contralateral diffusion	Diffuse flattening and slowing	n.a.	n.a.	Left temporal rhythmic SW and spikes	Right temporal and bifrontal progression; 1 GTCS	n.a.	Bitemporal spikes	n.a.
ASM
SCB (CBZ/PHE)/response	CBZ: yes PHE: yes	CBZ, PHE, LMT: partially responsive	CBZ: partially responsive PHE: yes	CBZ: yes	PHE: yes LMT: yes	CBZ: yes	CBZ: yes PHE: yes	OXC: yes	LMT	PHE: yes
Other ASM used	LVT	PB, VGB, CBZ, LVT, VPA, PER, BRV, LCS	VPA, CLB if fever, VGB, LVT	n.a.		n.a.	n.a.	LVT, VGB	VPA, LVT	VPA, TPM
Seizure trigger	Fever, accidental cranial trauma with low kinetic	None	Fever	Vaccination, accidental cranial trauma with low kinetic	Sleep deprivation	None	None	None	Fever	Sleep deprivation
Brain MRI
1st brain MRI	1 month: normal	8 years: left hippocampus atrophy	2 months: normal	3 months: normal	n.a.	3 months: normal	3 months: normal	4 months: normal	4 months: normal	4 months: normal
2nd brain MRI	n.a.	17 years: bilateral polymicrogyria	n.a.	n.a.	23 years: normal	n.a.	4 years 5 months: normal	20 months: left temporal pole atrophy with subcortical T2 hypersignal	3 years 5 months: normal	4 years: normal
Neurodevelopment
Walking age	13 months	Before 18 months	19 months	17 months	Before 18 months	15 months	13 months	13 months	13 months	12 months
ID	No	Mild WISC‐IV at 16 years: IQ score = 75	Mild Motor and language delay	No	No	No BSID‐IV at 2.6 years: normal	No WISC‐V at 8.3 years: IQ score = 116 VCI = 108, VSI = 108, FRI = 115, WMI = 107, PSI = 114	No	No	Mild Language delay WIPPSI‐IV at 6.2 years old: VCI = 78, VSI = 69, FRI = 59, WMI = 75, PSI = 65
School/work	Normal schooling	Adapted work	n.a.	n.a.	Work as a nurse	Normal schooling	Normal schooling	Normal schooling	Normal schooling	Adapted schooling
ASD/other symptoms	No	No	Yes	No	No	No	No	No	No	No

Abbreviations: AS, atonic seizure(s); ASD, autism spectrum disorder; ASM, antiseizure medication; BG, background activity; BRV, brivaracetam; BSID‐IV, Bayley Scales of Infant Development, fourth edition (note: with cognitive, motor, and language subtests); CBZ, carbamazepine; CLB, clobazam; F, female; FRI, fluid reasoning index; FS, focal seizure(s); GS, generalized seizure(s); GTCS, generalized tonic–clonic seizure(s); ID, intellectual disability; IQ, intelligence quotient; LCS, lacosamide; LMT, lamotrigine; LVT, levetiracetam; M, male; MRI, magnetic resonance imaging; n.a., not available; OXC, oxcarbazepine; PB, phenobarbital; PER, perampanel; PHE, phenytoin; PSI, processing speed index; SCB, sodium channel blocker; SE, status epilepticus; SW, slow wave; TPM, topiramate; TS, tonic seizure(s); VCI, verbal comprehension index; VGB, vigabatrin; VPA, valproic acid; VSI, visual spatial index; WISC‐IV, Wechsler Intelligence Scale for Children, fourth edition; WMI, working memory index.

FIGURE 1.

Chronology of antiseizure medication (ASM), the beginning of epilepsy, and intellectual outcome of the 10 patients with pathogenic FGF12 variants. Only ASMs used for >1 month are reported. Patients A–D present the p.(Arg52His) variant; Patients E–J present the p.(Gly50Ser) variant. The introduction age of the actual ASM or principal ASM is indicated next to it. The first seizure age for each patient is shown under the first (red) star from the left. Seizure remission (when appropriate, i.e., all patients except Patients B, C, and J) is indicated with a second (blue) star from the left. The age at the last examination and intellectual disability status are shown on the right. The precise introduction dates of ASM before the age of 6 years for Patient B are unknown. Abbreviations: BRV, brivaracetam; CBZ, carbamazepine; CLB, clobazam; LCS, lacosamide; LEV, levetiracetam; LMT, lamotrigine; mo, months; OXC, oxcarbazepine; PB, phenobarbital; PER, perampanel; PHE, phenytoin; TPM, topiramate; VGB, vigabatrin; VPA, valproate; y, years.

3.1. Pathogenic variants in the FGF12 gene

Four patients (Patients A–D) carried the p.(Arg52His) variant (chr3[hg19]:g.192053223C>T;NM_004113.6:c.155G>A;NP:004104.3:p.[Arg52His]) and six patients (Patients E–J) the p.(Gly50Ser) variant (chr3[hg19]:g.192053230C>T;NM_004113.6:c.148G>A;NP:004104.3:p.[Gly50Ser]), both absent from the Genome Aggregation Database and previously linked to epilepsy phenotype in the ClinVar database. All patients were heterozygous for one of these two variants; no cases of somatic mosaicism have been identified.

In silico prediction tools supported their pathogenicity (Combined Annotation Dependent Depletion scores: p.[Arg52His] 29.9, p.[Gly50Ser] 28.3). Both variants are in proximity (two amino acids) and located in the FGF family domain of the protein. In the present study, nine of 10 variants arose de novo. Patient I inherited a variant from his symptomatic mother. Patient A's variant was likely de novo, although healthy mother test was unavailable. Together, these elements allowed us to classify these variants as pathogenic (class ACMG‐5).

3.2. Epilepsy

Seizure onset ranged from 3 days to 4 months, with earlier onset in Patients A–D (p.[Arg52His], 3–48 days) and later onset for Patients E–J (p.[Gly50Ser], 3–4 months). Two patients with the p.(Gly50Ser) variant had stormy onset with status epilepticus, whereas those with the p.(Arg52His) variant showed gradual onset. The diagnosis of epilepsy was delayed >1 month for Patients A and B, due to normal initial clinical examinations and EEGs, and misdiagnosis as gastroesophageal reflux.

Tonic seizure was the most frequent initial seizure type (7/10) with one patient presenting only generalized tonic–clonic seizures. During follow‐up, 90% experienced both focal and generalized seizures. Detailed seizure semiology is provided in Table 1. Seizure frequency was variable across the cohort. Most patients exhibited sporadic seizures, with occasional clusters. Two patients (Patients E and F) presented with status epilepticus at epilepsy onset. Seizure frequency data have been added to Table 1 when available. All patients, except Patient E (older patient), were classified as having combined focal and generalized epilepsy. No consistent International League Against Epilepsy (ILAE) epilepsy syndrome could be identified across the cohort.

3.3. EEG findings

EEGs were performed in all patients, particularly at epilepsy onset and during seizure clusters. All recordings were routine EEGs; no prolonged or video‐EEG monitoring was conducted. Interictal EEGs revealed epileptiform discharges in six of 10 patients, predominantly focal (5/6 patients). Four patients (Patients A, B, G, and I) had normal interictal recordings. Ictal EEGs were available for five patients and demonstrated focal seizure onset in four of them, with temporal lobe involvement observed in three cases (Patients F, G, and I), as illustrated in Figure S1.

3.4. Antiseizure medication

Seizure clusters or status epilepticus were successfully treated with phenytoin (PHE; or fosphenytoin) for four patients (Patients A, C, G, and J). The ASM chronology is shown in Figure 1 (except for Patient B, due to missing data). Across the cohort, a total of 13 different ASMs were prescribed. The number of ASMs trialed per patient ranged from 2 to 6. The most used ASMs were carbamazepine (CBZ; used in six patients), PHE (used in six), and lamotrigine (LMT; used in three). Oxcarbazepine (OXC) was used in one patient. Seizure remission, defined as seizure freedom for at least 20 months, was achieved in seven patients. Among them, four were treated with CBZ, one with OXC, one with PHE followed by LMT, and one with LMT alone. Patients who did not achieve seizure remission (n = 3) had trialed a higher number of ASMs (range = 4–6), including multiple sodium channel blockers (SCBs).

3.5. Brain magnetic resonance imaging

Two patients presented with abnormal brain magnetic resonance imaging without cerebellar abnormalities. Patient B showed left hippocampal atrophy at 8 years and later bilateral polymicrogyria at 17 years; this patient also had persistent epilepsy and mild intellectual disability. Patient H exhibited left temporal pole atrophy with subcortical T2 hyperintensity at 20 months, despite a normal neurological examination and cognitive development at 6 years 10 months. To our knowledge, hippocampal atrophy has not been previously reported in FGF12‐related epilepsy and may reflect secondary epileptogenic changes. The clinical significance of the findings in Patient H remains unclear.

3.6. Neurodevelopmental assessment and outcome

Nine of 10 patients had normal initial neurological examinations; Patient D showed transient nystagmus without structural or EEG abnormalities.

Neurodevelopmental trajectories were normal in seven patients (2.1–38 years, median = 6.8), including two with the p.(Arg52His) variant and five with the p.(Gly50Ser) variant.

Seizure remission before age 1 year was achieved in all with normal outcomes (seven patients) and was associated for six of them with prompt SCB treatment (CBZ, n = 5; OXC, n = 1). Intellectual disability (Patients B, C, and J) was associated with uncontrolled epilepsy without prolonged remission (note that one and possibly two of these patients received prompt CBZ treatment). Autistic features were noted only in Patient C. We did not observe severe cognitive and behavioral impairments above and beyond the uncontrolled epilepsy, permitting us to classify theses three patients as DEE (despite them having a neurodevelopmental disorder).

3.7. Genotype–phenotype correlations

Patients A–D (p.[Arg52His]) had earlier seizure onset (<48 days) than Patients E–J (p.[Gly50Ser], >3 months). The p.(Gly50Ser) variant was generally associated with normal neurodevelopment and clinical examinations, except for Patient J.

4. DISCUSSION

We report 10 additional cases with recurrent pathogenic FGF12 variants, p.(Arg52His) (n = 4) and p.(Gly50Ser) (n = 6). Notably, the majority (7/10) demonstrated normal neurodevelopmental outcomes. Only three patients—two with p.(Arg52His) and one with p.(Gly50Ser)—exhibited mild intellectual disability. Importantly, none met the diagnostic criteria for DEE. In accordance with the ILAE definition, ⁵ , ⁶ DEE refers to severe developmental impairment due to both the underlying genetic cause and the epileptic activity itself. Although three of our patients presented with persistent seizures and mild intellectual disability, their neurodevelopmental profiles did not meet the criteria for DEE. These patients had mild and nonprogressive cognitive impairments, no significant behavioral comorbidities, and in only one case (Patient C), limited autistic traits. We did not observe cognitive or behavioral impairment beyond what could be explained by ongoing epilepsy. This distinction justifies excluding them from the DEE classification and supports the broader phenotypic spectrum associated with FGF12 variants. Favorable neurological outcomes were linked to well‐controlled epilepsy and, to a lesser extent, later epilepsy onset. The most used ASM was CBZ, which was associated with most of cases of epilepsy remission (four patients out of six).

In the FGF12‐related epilepsy cohort by Trivisano et al., ⁸ two patients displayed phenotypes inconsistent with DEE. One carrying the p.(Arg52His) variant had normal neurodevelopment, and the other with the p.(Gly50Ser) variant presented a non‐DEE phenotype. Both were treated with ASM, including SCBs (CBZ or LMT). Aligning with Trivisano et al., we confirmed that p.(Arg52His) variant is associated with an earlier epilepsy onset, whereas p.(Gly50Ser) variants correlate with better ASM response and more favorable outcomes. However, in contrast with prior data, none of our patients fulfilled the criteria for DEE, and seven of the 10 exhibited normal neurodevelopmental outcomes. Our series also includes the largest subgroup of patients with the p.(Gly50Ser) variant to date (n = 6 vs. n = 2 in Trivisano et al.), reinforcing the idea that this variant may be associated with a more favorable phenotype. Together, these findings broaden the phenotypic spectrum associated with FGF12 variants and suggest that DEE is not the inevitable outcome, even in the context of early onset epilepsy.

Ideally, studying patients with an identical de novo variant should yield a more consistent phenotype. However, our findings suggest that identical FGF12 variants, such as p.(Arg52His) or p.(Gly50Ser), can underlie at least two distinct clinical entities, with outcomes ranging from favorable neurodevelopment to severe DEE.

The clinical variability observed in pathogenic variants of the same gene can often be attributed to differences in variant type (truncation vs. missense) or their position within the protein, resulting in distinct functional consequences at the cellular level. For instance, KCNQ2 variants are associated with a broad spectrum of clinical presentations, from self‐limited epilepsy to DEE. ¹¹ , ¹² Importantly, the phenotype correlates closely with the variant's effect on the channel, and the same variant has not been reported in both DEE and self‐limited epilepsy. This may similarly apply to FGF12. Therefore, future functional studies are needed to better understand the phenotype variability of FGF12 variants.

One potential explanation for milder presentations may be somatic mosaicism, as observed with KCNQ2 variants. ¹³ Although this cannot be ruled out, the finding that the mutation is present in the heterozygous state does not support this. Moreover, somatic mosaicism can be ruled out in the patient who inherited the variant from his mother.

Significant phenotypic variability in patients with similar variants may also result from unidentified environmental or genetic modifiers. FGF12, a member of the fibroblast growth factor homologous factor family, modulates sodium channels Na_V1.2 (SCN2A), Na_V1.5 (SCN5A), and Na_V1.6 (SCN8A), but not Na_V1.1 (SCN1A). ¹ , ² , ³ , ⁴ , ⁶ Functional studies of p.(Gly50Ser) and p.(Arg52His) variants revealed a complex gain‐of‐function effect on Na_V1.2 and Na_V1.6 channels, increasing neuronal excitability. ⁶ , ¹² Despite advances in electrophysiology, understanding the molecular mechanisms underlying phenotypic variability for identical de novo variants remains challenging, with modifying genetic factors being a central hypothesis.

Moreover, long‐read whole genome sequencing enables the detection of DNA structural variants (SVs) that other genetic analysis techniques could not detect. Ohori et al. recently identified biallelic pathogenic SVs in the FGF12 gene, leading to a loss of function in two patients with severe DEE. ¹⁴ This work allows us to consider new research projects, such as performing genome sequencing and searching for SVs in patients with severe DEE and heterozygous SNVs in the FGF12 gene.

Another potential factor influencing outcomes is the role of ASMs, particularly SCBs like CBZ, OXC, and PHE. Although we found no consistent correlation between early SCB use and neurodevelopmental outcomes, seizure control appeared to be a key factor; all seven patients with normal development were seizure‐free at last follow‐up, whereas the three with persistent epilepsy exhibited mild ID. Most patients with favorable outcomes received SCBs early, but this was not universally predictive. Thus, although early SCB administration alone may not determine neurodevelopmental prognosis, it could contribute to improved seizure control, which in turn may support better cognitive outcomes. These observations warrant further investigation in larger prospective cohorts. This aligns with previously reported cases of FGF12‐related epilepsy. However, our findings highlight a key distinction; the three patients with mild ID had ongoing epilepsy despite ASM, whereas all seven patients with normal neurodevelopment were seizure‐free.

The potential impact of SCB therapy on disease progression is essential for future research. Both CBZ and OXC are indicated for treating early onset epilepsy with negative brain magnetic resonance imaging. ¹⁵ , ¹⁶ In our cohort, tonic seizures were the most common seizure type, strongly suggesting genetic epilepsy, as noted by Cornet et al., ¹⁷ and these are often well controlled by SCBs. We believe that the early, widespread use of SCBs could affect the neurodevelopmental trajectory and epilepsy history of these patients.

Here, we underscore the relevance of the FGF12 variant as a candidate in early onset epilepsy cases, whether associated with normal neurodevelopmental outcomes or global neurodevelopmental disorders, including autism spectrum disorder. A limitation of our study is the relatively short follow‐up period for several patients, particularly in the younger age group. Consequently, the possibility of later emerging intellectual disability, learning difficulties, or other neurological comorbidities cannot be entirely excluded. In addition, one patient in our cohort experienced seizure recurrence in adulthood, suggesting that long‐term seizure control may not be guaranteed. Although the natural history of FGF12‐related epilepsy remains incompletely defined due to the limited number of reported cases, the absence of early severe and global developmental impairments in our cohort nonetheless supports a distinction from the classic DEE phenotype. Larger longitudinal studies will be necessary to confirm this differentiation and to further characterize long‐term outcomes.

Future descriptions will likely expand to include cases of intellectual disability without epilepsy, possibly associated with autistic disorders, as has already been described for the SCN2A and SCN8A pathogenic variants.

5. CONCLUSIONS

Altogether, this study contributes to refining the phenotypic spectrum associated with FGF12 variants and suggests that FGF12 should be considered as a strong susceptibility gene. We highlight cases of sporadic good neurodevelopmental outcome epilepsy in contrast to the DEE phenotype. However, FGF12 pathogenic potential likely depends on additional modifiers, supporting a multihit model in the pathophysiology of early onset epilepsy and neurodevelopmental disorders. This work aims to enhance the awareness among practitioners (geneticists, pediatric neurologists, etc.) about the significant clinical variability of FGF12 variants. Early therapeutic interventions with SCBs such as CBZ may influence this variability. Further functional studies will be essential to elucidate the mechanisms driving the phenotypic diversity observed in patients with identical recurrent variants of FGF12.

AUTHOR CONTRIBUTIONS

Béatrice Desnous conceived the study. Clément Pierret, Florence Riccardi, and Béatrice Desnous collected and curated the data. Clément Pierret, Florence Riccardi, Mathieu Milh, and Béatrice Desnous interpreted the results and drafted the manuscript with input from all authors. All authors critically revised the manuscript and approved the final version.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to report. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Supporting information

Figure S1.

Pierret C, Riccardi F, Neveu J, Alesandrini M, Altuzarra C, Boulogne S, et al. Broadening the phenotype associated with pathogenic variants in the FGF12 gene: From developmental and epileptic encephalopathy to drug‐responsive epilepsy with favorable cognitive outcome. Epilepsia. 2025;66:e158–e168. 10.1111/epi.18495

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.