KCNQ2 neonatal epilepsy: Impact of prompt diagnosis and treatment, and early predictors of outcome severity Authors: Trupti Jadhav, Sophie E Bouffler, Emily Innes, Michael Fahey, Matthew Hunter, Kavitha Kothur, Sebastian Lunke, Matthew Lynch, Emma Macdonald‐Laurs, Elizabeth Emma Palmer, Chirag Patel, Jason Pinner, Kate Riney, Rani Sachdev, Sarah A Sandaradura, Ingrid E Scheffer, Zornitza Stark, Katherine B Howell Published: 2026 ABSTRACT AbstractObjectiveTo determine whether prompt genetic diagnosis in children with KCNQ2 neonatal epilepsy enabling targeted therapy is associated with improved outcomes, and identify early predictors of developmental outcomes.MethodsThirty‐seven children with KCNQ2 neonatal epilepsy were recruited from five pediatric centers. We reviewed demographic, clinical, EEG, and genetic data. We determined differences in outcomes between individuals with prompt (greater than 30 days from seizure onset) and later genetic diagnosis, and we identified neonatal factors associated with developmental outcome.ResultsBaseline characteristics were similar between children with prompt (n = 6, median age at genetic diagnosis 15 days) and later (n = 31, median age 309 days, p < .05) diagnosis. All with prompt diagnosis received sodium channel blocking (SCB) anti‐seizure medication (ASM) in the neonatal period compared with 15/31 (48%) in the later diagnosis group. Children with prompt diagnosis had higher rates of seizure freedom at age 12 months than those with later diagnosis (6/6 [100%] vs. 17/31 [54%]; p .049], and lower number of emergency department representations (median 0 vs. 2), and hospital readmissions (median 0 vs. 1). Factors in the neonatal period associated with abnormal developmental outcome included neurological abnormalities (e.g., abnormal tone) and markedly abnormal neonatal EEG background (11/11 [100%] with markedly abnormal EEG vs. 11/24 [46%] with normal to moderately abnormal EEG).SignificancePrompt genetic diagnosis was associated with targeted therapy, resulting in improved seizure control and reduced hospital representation. Clinical features present in the neonate assist in predicting outcome severity, which is critically important in counselling families receiving a KCNQ2 diagnosis soon after seizure onset.Plain Language SummaryIn KCNQ2 neonatal epilepsy, sodium channel blocking antiseizure medicines are recommended, but the benefits of starting treatment early have been uncertain. Our findings show that prompt genetic diagnosis enabled early targeted treatment, with potential to improve outcomes. Specifically, prompt genetic diagnosis was associated with improved seizure control and reduced hospital visits compared with delayed diagnosis. However, a prospective, long‐term study is needed to determine whether early treatment also improves developmental outcomes. Predicting outcome severity in newborns remains challenging, although abnormal neurological examination and markedly abnormal EEG in the newborn period were linked to abnormal developmental outcomes.Keywords: KCNQ2, neonatal epilepsy, outcome, prompt genetic diagnosis, sodium channel blocking antiseizure medication Key points Sodium channel blocking antiseizure medication (SCB ASM) is indicated in KCNQ2 neonatal epilepsy, but the benefit of prompt commencement is not clear. Sodium channel blocking antiseizure medication (SCB ASM) is indicated in KCNQ2 neonatal epilepsy, but the benefit of prompt commencement is not clear. Prompt KCNQ2 diagnosis and SCB ASM treatment improved seizure control and reduced hospital representation compared with delayed diagnosis. Prompt KCNQ2 diagnosis and SCB ASM treatment improved seizure control and reduced hospital representation compared with delayed diagnosis. A prospective, longitudinal study will be required to determine if prompt diagnosis and treatment with SCB ASM improve development. A prospective, longitudinal study will be required to determine if prompt diagnosis and treatment with SCB ASM improve development. Predicting outcome severity in the neonatal period can be difficult in KCNQ2 neonatal epilepsy. Predicting outcome severity in the neonatal period can be difficult in KCNQ2 neonatal epilepsy. Abnormal neurologic examination and markedly abnormal EEG in the neonatal period were associated with abnormal developmental outcome. Abnormal neurologic examination and markedly abnormal EEG in the neonatal period were associated with abnormal developmental outcome. INTRODUCTION KCNQ2, encoding the voltage‐gated potassium channel Kv7.2, is one of the most common genetic causes of neonatal epilepsy. 1 KCNQ2 neonatal epilepsies range in severity from self‐limited (familial) neonatal epilepsy (SeLNE) to developmental and epileptic encephalopathy (DEE), with developmental outcomes ranging from normal to profound impairment. 2 , 3 , 4 , 5 At present, outcome severity can be challenging to predict in the neonatal period. We therefore need a deeper understanding of factors that determine outcome severity to critically inform prognostic and genetic counseling. Rapid (weeks, typically 2–3) and ultrarapid (days, typically 3–5) genomic testing is increasingly used in the neonatal intensive care unit. 6 , 7 , 8 , 9 Genomic testing has a high yield in neonates with epilepsy, and diagnosis frequently impacts management. 10 , 11 Diagnosis of KCNQ2 neonatal epilepsy enables implementation of targeted treatments with sodium channel blocking (SCB) antiseizure medication (ASM), which can improve seizure control. 12 While there are case reports of favorable outcomes with prompt genetic diagnosis and implementation of SCB ASM, it is not clear whether this improves the ongoing risk of seizures, developmental, and other health‐related outcomes. 4 , 12 , 13 , 14 In this study, we aimed to determine whether prompt genetic diagnosis in infants with KCNQ2 neonatal epilepsy, enabling early administration of targeted therapy, is associated with improved outcomes, and to identify early predictors of developmental outcomes. METHODS We undertook a retrospective study of children with KCNQ2 neonatal epilepsy at five Australian tertiary pediatric centers involved in the Australian Genomics Acute Care study of ultrarapid trio genomic sequencing (tGS). 6 , 8 Patients with prompt diagnosis after seizure onset at the participating centers were identified from the Acute Care Genomics study database, while those diagnosed at any time via alternative testing pathways were identified through research and clinical databases at each center. Patients with later diagnosis were an older cohort, some being born prior to genetic testing being routine in clinical practice and others at a time where genetic testing was available but often took many months; these patients had a range of genomic technologies performed (e.g., gene panel, exome sequencing) as singleton or trio. We included patients with a pathogenic/likely pathogenic KCNQ2 variant, onset of seizures in the neonatal period, and available seizure and developmental data to at least age one year. 15 We excluded children with dual diagnoses or KCNQ2‐containing chromosome 20q deletions (due to possible influence of other genes on phenotypes and/or outcomes), and those with KCNQ2 variants known to result in gain‐of‐function (as gain‐of‐function variants have a different presentation and treatment implications to variants that cause KCNQ2 neonatal epilepsy). Medical records were reviewed for demographic, genetic findings (variant, age at molecular diagnosis), clinical features (including seizure onset age and semiology, ASM use, neurological findings, and developmental progress), and details of hospital use (emergency department presentations, hospital admissions). Neonatal EEGs were reviewed for background abnormalities and interictal epileptiform discharges (IEDs). EEG background was classified according to the neonatal EEG background classification described by Shellhaas et al. 16 (normal vs. mildly, moderately, or markedly abnormal). With this classification, markedly abnormal EEG backgrounds include burst‐suppression patterns, moderately abnormal backgrounds were discontinuous (but not burst suppression) and had a paucity of background features and reduced state differentiation, and mildly abnormal backgrounds were those with excessive discontinuity or asynchrony in quiet sleep but not other states. Frequency of IEDs was categorized as nil compared with 0%–25%, 25%–50%, or >50% of the whole EEG recording (typically approximately one‐hour recording). Epilepsy syndromes were determined based on the ILAE classification, with individuals assigned to SeLNE or KCNQ2‐DEE syndromes. 4 Individuals who fit criteria for SeLNE other than having a more abnormal EEG background than is characteristic for this syndrome (i.e., at least moderately abnormal EEG background) were assigned to an ‘intermediate’ epilepsy syndrome category. Developmental outcomes were categorized as normal, mild‐moderately abnormal, or severe‐profoundly abnormal, at age 12 months, and at the most recent review, based on a developmental quotient inferred from motor and language developmental milestones attained. We considered DQ >70 to be normal, DQ 35–69 to be mild‐moderately abnormal, and DQ <35 severe‐profoundly abnormal. Medical literature and the ClinVar database were searched to identify whether each individual's variant had been previously reported, and the associated phenotype in previously reported individuals. 17 We determined differences in health resource use in the first year of life and developmental and epilepsy outcomes at age 12 months between children with prompt (defined as <30 days from seizure onset) and later (>30 days from seizure onset) genetic diagnosis. We identified clinical, EEG, and genetic factors in the neonatal period associated with developmental outcome (normal vs. abnormal) at the last review. Chi‐square test and Fisher's exact test were used to determine differences between groups for categorical variables, and the Mann–Whitney U test was used to compare median values. We corrected for multiple comparisons by using the Benjamini–Hochberg procedure, controlling the false discovery rate at 5%. We then used a heat map to represent neonatal factors associated with developmental outcomes (categorical variables) as participant numbers were too small to conduct a multivariate analysis. RESULTS Cohort We identified 49 children with KCNQ2 epilepsy. Of these, 12 were excluded for the following reasons: presumed KCNQ2 epilepsy based on being affected relatives of other individuals with confirmed pathogenic KCNQ2 variants, but not tested themselves (4), postneonatal seizure onset and/or known gain‐of‐function mutation (5), 20q deletion (1), dual genetic diagnosis (1), and inadequate data (1). Thirty‐seven children (17 [45%] male) were included in this study (Table 1). Children were a median of 3 days (interquartile range [IQR] 1–4 days) at seizure onset and were followed for a median of 3.8 years (IQR 1.3–7 years). KCNQ2 variants were missense in 29 (78%) and protein truncating in 8 (21%). Eleven (27%) had a family history of neonatal or infantile epilepsy (3 probands had confirmed KCNQ2 variants in affected family members and family members were not tested for 8 probands). TABLE 1Phenotypic and genetic data of 37 children with KCNQ2 neonatal epilepsy.Cohort, N = 37Prompt diagnosis, N = 6Later diagnosis, N = 31 p ValueDemographicsFamily history neonatal/infantile epilepsy11 (27%)011 (35%)0.08SeLFNE phenotype (family history and normal EEG)4 (10%)04 (13%)Genetic dataVariant typeMissense29 (78%)5 (83%)24 (64%)0.74Nonsense/frameshift/splice site8 (21%)1 (16%)7 (22%)Variant inheritanceDe novo22 (59%)6 (100%)16 (51%)Inherited3 (8%)03 (9%)Unknown (but +ve FHx)8 (21%)08 (25%)Unknown (no FHx)4 (10%)04 (12%)Baseline neonatal dataAge seizure onset (days), Median (IQR)3 (1–4)1 (1–2)3 (1–4.5)0.87No. of seizure days in first month (n = 35)<717 (45%)1 (16%)16 (51%)0.17>718 (48%)5 (83%)13 (41%)No. of >hourly seizure days in first month (n = 23)<719 (51%)5 (83%)14 (45%)1.0>74 (10%)04 (12%)EEG background (n = 35)Normal7 (18%)07 (22%)0.08Mildly abnormal9 (24%)2 (33%)7 (22%)Moderately abnormal8 (21%)3 (50%)5 (16%)Markedly abnormal11 (29%)011 (35%)EEG IEDS (n = 29)Nil6 (16%)06 (19%)0.67<50%10 (27%)2 (33%)8 (25%)>50%13 (35%)3 (50%)10 (32%)Neurologic stateAny abnormality21 (56%)5 (83%)16 (51%)0.15Visual attention not appropriate7 (18%)1 (16%)6 (19%)0.87Neurologic examination abnormal14 (37%)2 (33%)12 (38%)0.80Supplemental feeding required18 (48%)5 (83%)13 (41%)0.06Intubation required6 (16%)2 (33%)4 (12%)0.21Abbreviations: FHx, family history; IEDs, interictal epileptiform discharges; IQR, interquartile range; No., number; SeLFNE, self‐limited familial neonatal epilepsy. Phenotypic and genetic data of 37 children with KCNQ2 neonatal epilepsy. Abbreviations: FHx, family history; IEDs, interictal epileptiform discharges; IQR, interquartile range; No., number; SeLFNE, self‐limited familial neonatal epilepsy. Prompt vs. later genetic diagnosis Genetic diagnosis was prompt in six children (16%) (median 15 days, IQR 13.5–15 days) and later in 31 (84%) (median 10 months, IQR 4–55 months) (Table 2). Baseline characteristics (Table 1) were similar between children with prompt and later diagnosis, but last review was more recent in the prompt (median 16 months) compared with the later group (median 50 months). At presentation, the EEG background in the later diagnosis group ranged from normal to markedly abnormal, whereas those in the prompt diagnosis group had a mildly or moderately abnormal EEG background. Notably 3/6 (50%) neonates in the prompt group had IEDs occupying >50% of the EEG at presentation, compared with 10/31 (32%) in the later group. TABLE 2Intervention and outcome data of 37 children with KCNQ2 neonatal epilepsy.Cohort, N = 37Prompt diagnosis, N = 6Nonprompt diagnosis, N = 31 p Value (unadjusted) p Value (adjusted for multiple comparisons)Intervention dataAge at genetic diagnosis (days), median (IQR)233 (54–1133)15 (13.5–15)309 (131–1674) <0.001 0.007 SCB use (n = 30)Given SCB in neonatal period21 (56%)6 (100%)15 (48%) 0.004 0.017 Given SCB b/w age 1–12 months6 (16%)06 (19%)No SCB in first year10 (27%)010 a (32%)No SCB used ever7 (18%)07 (22%)Age at first SCB use (days), median (IQR)14 (9–78)11 (7–15)16 (10–122)0.230.25PB used in neonatal period29 (78%)6 (100%)23 (74%)Outcome data at age 12 monthsSeizuresBecame seizure‐free without SCB7 (18%)07 (22%)Age first seizure‐free (days), median (IQR)42 (13–180)26 (22–28)60 (12.5–255)0.2070.245Seizure recurrence (after initial 3 m seizure‐free)9 (24%)2 (33%)9 (29%)Seizure‐free at age 12 m23 (62%)6 (100%)17 (54%) 0.03 0.049 No. of seizures in first year<5019 (51%)3 (50%)16 (51%)50–2004 (10%)3 (50%)1 (3%)>20011 (29%)011 (35%) 0.006 0.018 ASMTotal number ASM, median (range)4 (1–8)5.5 (3 to 7)4 (1–8)0.200.245Hospital resource useDuration of neonatal admission (days), median (IQR)21 (10–32)25 (21.5–34)19.5 (9–32)0.260.26No. of re‐presentations to ED2 (0–5)0 (0–2) b 2 (0–5) 0.007 0.018 No. of readmissions (n = 21)1 (0–2)0 (0–1) b 1 (0–4) 0.022 0.048 Duration of readmissions (days), median (IQR)7 (2–31)0 (0–2)7 (2–31) <0.001 0.007 Outcomes at most recent reviewDevelopmentNormal15 (40%)2 (33%)13 (45%) 0.03 0.049 Mild‐mod DD8 (21%)4 (66%)4 (12%)Sev‐prof DD14 (37%)014 (45%)Seizures (n = 34)Ongoing seizures14 (37%)014 (45%)0.120.173Seizure‐free20 (54%)6 (100%)14 (45%) Note: Bolded values indicate statistically significant difference between normal and abnormal development groups.Abbreviations: ASM, antiseizure medication; DD, developmental delay; ED, emergency department; IQR, interquartile range; No., number; PB, phenobarbitone; SCB, sodium channel blocking ASM. a 7/10 were seizure‐free at 12 months without using SCB. b One participant had two ED presentations and one admission, which were unrelated to seizures (due to COVID‐19 infection and another viral illness). Intervention and outcome data of 37 children with KCNQ2 neonatal epilepsy. Note: Bolded values indicate statistically significant difference between normal and abnormal development groups. Abbreviations: ASM, antiseizure medication; DD, developmental delay; ED, emergency department; IQR, interquartile range; No., number; PB, phenobarbitone; SCB, sodium channel blocking ASM. 7/10 were seizure‐free at 12 months without using SCB. One participant had two ED presentations and one admission, which were unrelated to seizures (due to COVID‐19 infection and another viral illness). SCB ASMs (carbamazepine, lacosamide, lamotrigine, oxcarbazepine, phenytoin) were used as a regular medication in the first year of life in 27 infants, with 21 receiving them during the neonatal period. The proportion of infants administered regular SCB ASM in the neonatal period was higher in infants with prompt diagnosis [6/6 (100%) vs. 15/31 (48%)]. The number of other ASMs (e.g., phenobarbitone, levetiracetam, but also including vitamins such as pyridoxine) used before receiving a regular SCB ASM was similar between the two groups, with a median of 4 (range 2–7) in the prompt diagnosis group and a median of 3 (range 1–6) in the later diagnosis group. All children had focal seizures at presentation. Those with prompt diagnosis had lower seizure burden in the first year of life compared with those with later diagnosis [0/6 (0%) vs. 11/31 (35%), p = 0.018]. There were also higher rates of seizure freedom at age 12 months in the prompt diagnosis group [6/6 (100%) vs. 17/31 (54%), p = 0.049]. Of the 31 children with a later diagnosis, three developed epileptic spasms in infancy; two of whom had not received SCB ASM prior to spasm onset. None in the prompt diagnosis group developed epileptic spasms. At age 12 months, the number of ASMs being taken was no different between the promptly diagnosed infants (median 1, range 0–2) and those diagnosed later (median 1, range 0–6) (p = .245). Three of the six promptly diagnosed children weaned SCB ASM after age 12 months; seizures recurred in two at ages 12 and 13 months and were controlled with SCB recommencement. The duration of the initial hospital admission was comparable between the two groups (later diagnosis median 19.5 days vs. prompt median 25 days). In contrast, patients with later diagnosis had more emergency department re‐presentations (median 2 vs. 0, p = .018), hospital readmissions (median 1 vs. 0, p = .022), and longer duration of hospital readmissions (median 7 vs. 0 days, p = .007) than those with prompt diagnosis. No patient with a prompt diagnosis had severe‐profound developmental delay at last review, compared with 13/31 (42%) individuals with later diagnosis (p = .049). Four patients had SeLNE phenotypes at presentation (i.e., positive family history, normal neonatal EEG, normal neurologic examination), all in the later diagnosis group. We repeated analyses with these four patients excluded (later diagnosis group n = 27) and saw no difference in all the results. We also undertook a subanalysis of the 15 infants in the later diagnosis group who had early use of SCB ASM. At age 12 months, 10/15 (67%) were seizure‐free. Following the initial admission, these infants had 0–1 days of hospital readmission in the first year of life. Epilepsy syndrome and molecular‐syndrome classification Epilepsy syndromes were not always able to be definitively diagnosed in the neonatal period. On review of all available data, 14 children had SeLNE (1 prompt, 13 later), 20 had DEE (3 prompt, 17 later), and 3 had an intermediate phenotype (2 prompt, 1 later). There was no difference in epilepsy syndromes between those with prompt or later diagnosis (p = .39). DEE or intermediate phenotypes were seen in 21/29 (72%) of individuals with missense variants, and 2/8 (25%) with protein‐truncating variants. Missense variant location was intracellular (IC) in 7/29 individuals, extracellular (EC) in 2/29, and transmembrane (TM) in 20/29 (S3 in 1, S4 in 4, S5 in 4, pore‐forming loop in 6, and S6 in 5). We did not see a difference between missense variant location and epilepsy syndrome (DEE = 3 IC, 1 EC, 11 TM [including S3 in 1], intermediate = 1 IC, 2 TM, and SeLNE = 3 IC, 1 EC, 7 TM [including pore‐forming loop in 3, and S6 in 2]). Similarly, for the 8 individuals with protein‐truncating variants, we did not see a difference in epilepsy syndrome for truncations in the C‐terminal segments compared with those occurring more proximally (i.e., N‐terminal and S1‐S6 regions) (DEE = 2 C‐terminal, 2 proximal, intermediate = 1 C‐terminal, SeLNE = 2 C‐terminal, 1 proximal). Variants in 30 patients were previously reported in the medical literature or ClinVar database; 4 variants have been previously associated with both SeLNE and DEE phenotypes, and the remainder with only one of SeLNE or DEE. Eleven individuals with SeLNE in our cohort had previously reported variants; seven patients had a variant previously associated with DEE (and two had variants associated with both DEE and SeLNE). One individual with DEE had a variant previously associated with SeLNE (and two had variants associated with both SeLNE and DEE). The discrepancy between the previously reported epilepsy syndrome for a particular variant and that reported in patients in our cohort with the same variant did not vary by prompt or later diagnosis (p = .50). Of 5 patients in our cohort with prompt diagnosis and variants that had previously been associated with DEE, 3 had DEE, 1 had an intermediate phenotype, and 1 had SeLNE. Of 19 patients with later diagnosis and variants previously reported in people with DEE, 11 had DEE, 1 had an intermediate phenotype, and 7 had SeLNE. Early predictors of developmental outcome At the most recent review, 15 (41%) patients had normal development, and 22 (59%) had abnormal development (8 mild‐moderate, 14 severe‐profound). Only children with abnormal developmental outcomes had a markedly abnormal neonatal EEG background (11/22 (50%) vs. 0/15 (0%), p 0.027) (Figure 1, Table 3). However, the neonatal EEG background was moderately abnormal in 2/15 (13%) with normal developmental outcomes. Eleven of the 13 (85%) patients with IEDs occupying >50% of their neonatal EEG had abnormal developmental outcomes. No patient with severe‐profound developmental impairment had a normal neonatal EEG, although one child with a normal EEG had mild–moderate developmental impairment at last review. Heat map showing electroclinical factors in the neonatal period which were associated with developmental outcome. Shading of cells: Light gray = normal (or not more than mildly abnormal for electrical features), orange = abnormal, dark gray = missing data. TABLE 3Factors at initial presentation associated with developmental outcomes in 37 children with KCNQ2 neonatal epilepsy.Normal development, N = 15Abnormal development, N = 22 p Value, normal vs. impaired (unadjusted) p Value, normal vs. impaired (adjusted)Genetic dataVariant typeMissense11 (73%)18 (81%)0.660.85Nonsense/frameshift/splice site4 (26%)4 (18%)Variant inheritanceDe novo8 (53%)14 (63%)0.770.85Inherited2 (13%)1 (4%)Unknown (but positive FHx)4 (26%)4 (18%)Unknown (no FHx)1 (6%)3 (13%)Neonatal dataAge seizure onset (days), median (IQR)3 (2–4)2.5 (1–4)0.560.85No of seizure days in first month (n = 35)<711 (73%)6 (27%) 0.01 0.027 >74 (26%)14 (63%)No of >hourly seizure days in first month (n = 23)<139 (60%)12 (54%)0.230.41>1302 (9%)EEG background (n = 35)Normal6 (40%)1 (4%) 0.005 0.027 Mild abnormal5 (33%)4 (18%)Mod abnormal2 (13%)6 (27%)Marked abnormal011 (50%)EEG IEDS (n = 29)Nil5 (33%)1 (4%) 0.01 0.027 <50%5 (33%)5 (22%)>50%2 (13%)11 (50%)Neurologic stateAny abnormality5 (33%)16 (72%) 0.01 0.027 Visual attention not appropriate07 (31%) 0.006 0.027 Neurologic examination abnormal014 (63%) <0.001 0.016 Supplemental feeding required5 (33%)13 (59%)0.180.40Intubation required2 (13%)4 (18%)0.670.85Intervention dataAge at genetic diagnosis (days), med (IQR)110 (50–309)425 (131–1179)0.200.40SCB use (n = 30)Given SCB in neonatal period8 (53%)12 (54%)0.800.85Age at first SCB use (days), median (IQR)14 (8.5–77)18 (9.5–61)0.900.90Phenobarbitone used in neonatal period (n = 29)12 (80%)17 (77%)0.800.85 Note: Bolded values indicate statistically significant difference between normal and abnormal development groups.Abbreviations: FHx, family history; IEDs, interictal epileptiform discharges; IQR, interquartile range; SCB, sodium channel blocking antiseizure medication. Factors at initial presentation associated with developmental outcomes in 37 children with KCNQ2 neonatal epilepsy. Note: Bolded values indicate statistically significant difference between normal and abnormal development groups. Abbreviations: FHx, family history; IEDs, interictal epileptiform discharges; IQR, interquartile range; SCB, sodium channel blocking antiseizure medication. A heat map displaying the factors associated with developmental outcomes in each child (Figure 1) highlights that some aspects of the neurologic status in the neonatal period were also associated with normal versus abnormal developmental outcomes. All individuals with normal developmental outcome, including the two with moderately abnormal neonatal EEG background, had normal visual attention and neurologic examination in the neonatal period, whereas 14/22 (64%) with abnormal developmental outcomes had either or both abnormal visual attention or neurologic examination. In contrast, the need for endotracheal intubation or supplemental feeding (i.e., nasogastric tube) were not associated with developmental outcome. We saw no association between developmental outcome and age at genetic diagnosis, variant type (missense vs. truncation), number of days with seizures in the neonatal period (categorized as < or >7 days), or age at first SCB use. Family history of neonatal seizures was associated with a range of developmental outcomes (normal 6, mild–moderately abnormal 2, severe‐profoundly abnormal 3). Genetic testing of parents and/or other affected family members was only performed in 3/11 individuals with a positive family history. This means we are not able to determine whether KCNQ2 was the cause of the neonatal seizures in relatives of the probands in all these families, nor whether any affected parents were mosaic for the variant seen in germline form in the proband. We did not observe differences in neonatal neurologic examination parameters between patients with mild‐moderate and severe‐profound developmental impairment. DISCUSSION Our study demonstrates that prompt genetic diagnosis in children with KCNQ2 neonatal epilepsy enables early use of SCB ASM, which may improve outcomes. Infants with prompt diagnosis had lower hospital resource use in the first year of life and higher rates of seizure freedom at age 12 months, compared with infants who received a later genetic diagnosis. These findings highlight the benefits of rapid genetic diagnosis, which allows management to be targeted with the potential to improve outcomes. We also identify electroclinical features in the newborn period that assist in predicting normal versus abnormal developmental outcome, which is critical in counseling families receiving a KCNQ2 epilepsy diagnosis early in life. The positive impact of SCB ASM in improving seizure control in KCNQ2 neonatal epilepsy is well‐established, and their preferential use over other ASM is recommended. 3 , 4 , 12 However, while SCB use has been reported to shorten the duration of NICU admission, the magnitude of benefit on seizures, need for other ASM, developmental outcomes and hospital resource use beyond the initial neonatal hospitalization are incompletely understood. 18 Further, while rapid genetic diagnosis is increasingly possible, the longer‐term impacts of implementing prompt, optimal therapy were not clear. Overall, prompt diagnosis (and implementation of SCB therapy) was associated with earlier seizure control and higher rates of seizure freedom compared with later diagnosis. Seizures were controlled in all six patients at age 12 months compared with approximately half of individuals with later diagnosis. Our study also demonstrated that seizure freedom was possible with SCB monotherapy in all the promptly diagnosed patients, meaning that prompt diagnosis may enable early cessation of (or lack of need for) other ASM. Further, we demonstrated reduced health system resource use, with no seizure‐related re‐presentations to hospital in the prompt diagnosis group in the first year of life. Emergency department presentations and hospital readmissions are not only extremely stressful for families, but they also carry important economic costs to the health system. The average cost of hospital admission in Australia is AU$2884/day, and the cost of tGS, whether ultrarapid (turn‐around‐time [TAT] 3–5 days) or rapid (TAT 3 weeks), is AU$15 000 and AU$8900 respectively. 19 , 20 , 21 A reduction in days of readmission by approximately 5 and 3 days per patient respectively would therefore recoup the costs of genomic testing. This was achieved in the first year of life, with a median of zero readmission days in those with prompt diagnosis compared with seven days in the later diagnosis group. We did not quantify additional direct costs, such as diagnostic investigations avoided or the cost of ASM prescriptions. Further cost‐savings relating to prompt diagnosis have not been analyzed, but are also important to consider, including indirect costs to families, such as lost work productivity. We could not determine the impact of prompt genetic diagnosis on development, as the high phenotypic variability and our relatively crude measures of developmental outcome meant that it was not possible to determine whether the outcome deviated from that expected for any individual child. Nevertheless, we saw no severe‐profoundly impaired children in the prompt diagnosis group, despite their clinical features at onset and duration of initial hospitalization being similar to the later diagnosis group. This absence of severe‐profoundly impaired children cannot be attributed to the assumption that the prompt diagnosis group had SeLNE; ultimately, only one met criteria for this syndrome. Of the remaining five, none had inherited variants or a family history of epilepsy, none had a normal interictal EEG at baseline, 3/6 had IEDs occupying >50% of neonatal EEG, 2/6 had an abnormal neonatal neurologic examination, and 2/6 had seizure recurrence after SCB were weaned after infancy (indicating their epilepsy had not resolved as would be expected in SeLNE). The prompt and later diagnosis groups were also similar in the duration of initial admission, suggesting the groups were not dissimilar in terms of severity in the neonatal period. However, none with a prompt diagnosis had a markedly abnormal neonatal interictal EEG, so our prompt group is likely to have been missing infants at the most severe end of the KCNQ2 DEE spectrum. When considering early predictors of outcome, we did not see a higher rate of early use of SCB ASM in those with normal compared with abnormal developmental outcomes. This does not imply lack of benefit with prompt treatment, however. The categorization of normal versus abnormal development is likely too crude, and looking at this variable in isolation from related factors like total number of ASM and baseline severity may mask benefit. Further study, with quantitative and serial measurement of development in babies with KCNQ2 epilepsies, is needed to understand the impact of prompt diagnosis and early use of SCB ASM monotherapy on developmental outcomes. Our observation of abnormal developmental outcomes in some promptly diagnosed and treated individuals is consistent with the knowledge that KCNQ2 neonatal epilepsies have a developmental encephalopathy component that is independent of the epileptic encephalopathy. These observations reinforce the need for better treatments, ideally targeted at the underlying cause. 13 Availability of ultrarapid and rapid genomic sequencing remains limited around the world, which is a limitation to the generalizability of our findings. However, given recent updates to the ILAE neonatal seizure consensus guidelines, 22 which now state that SCB ASM may be preferred for neonates with presumed channelopathy (presumption based on one or more of a number of clinical features, including no other identified etiology for seizures), SCB ASM will likely be utilized earlier in infants with KCNQ2‐epilepsy, prior to genetic confirmation. Early use of SCB ASM, rather than the genetic diagnosis itself, is likely the main driver of favorable seizure outcome (although other contributory factors guided by the genetic diagnosis are possible, given that in our cohort, all 6 promptly diagnosed infants were seizure‐free at age 12 months compared with only 10/15 infants with later diagnosis but early SCB ASM use). However, we believe it likely that the genetic diagnosis will yield additional benefit, such as providing confidence in early cessation of other ASM and therefore minimizing overall ASM burden. The spectrum of severity of KCNQ2 neonatal epilepsy is broad, and prediction of outcome severity soon after seizure onset can be difficult. Some factors have been reported to predict outcomes, such as a burst‐suppression pattern on EEG which is associated with poor outcomes, and a family history of KCNQ2‐SeLNE and variant location (S2‐3 domain more likely SeLNE, S5‐6 domain more likely DEE), which are associated with favorable outcomes. 4 , 23 De novo variants may be associated with both SeLNE and DEE. 3 , 24 However, these variant factors are incompletely predictive as, in our cohort, we saw children with de novo variants with normal outcomes and children with inherited variants whose outcomes were abnormal. Specific genetic variants may not be predictive, as variability in phenotypic severity has been reported for a number of recurrent variants, and discordance with the previously reported phenotype for a particular variant was observed in numerous individuals with recurrent variants in our cohort (e.g., SeLNE where variant previously associated with DEE). 25 , 26 Machine learning models trained on the genetic variant have to date had greater success with predicting benign versus pathogenic variants, than predicting normal versus abnormal outcome, with authors concluding that these tools are potentially useful in conjunction with phenotypic analysis, rather than in isolation. 27 With prompt diagnosis an increasing reality, more precise prognostication is critical for families, to guide appropriate medical management (e.g., around duration of ASM use, and likely need for therapies to maximize developmental outcomes), reproductive counseling, and to support parental wellbeing in the early months of life, as uncertainty is an important contributor to poor psychological health. 28 The clinical features in the neonatal period that were most predictive of poor outcome were abnormalities of visual attention or neurologic examination. They were both highly specific, although not sensitive, as they occurred in only 31% and 63% of those with abnormal developmental outcomes, respectively. The need for intubation or supplemental feeding did not predict outcome. This may be because they were related to sedation from ASM or seizure characteristics (e.g., apnea with seizures), rather than solely the underlying severity of the disease. Like others, we found that a markedly abnormal neonatal EEG was universally associated with abnormal developmental outcome. 29 However, we saw normal developmental outcomes in two infants with IEDs occupying >50% of the EEG and/or moderately abnormal EEG background, which, in the classification by Shellhaas et al. that we applied, encompassed features such as interburst intervals of up to 15 s, poor state differentiation and a paucity of background features. 16 We considered whether the EEG abnormalities in these infants could be explained by ASM use, as both received phenobarbitone and one a midazolam infusion in the neonatal period (although whether they were on these ASM at the time of their neonatal EEG was difficult to determine). However, while such ASM could explain background abnormalities, they do not explain the frequent IEDs. Both infants had normal visual attention and neurologic examination, which could potentially be reassuring factors in the context of prominent EEG abnormalities, although our numbers are too small to conclude this definitively. Overall, we found that normal, mildly and moderately abnormal EEG background did not consistently predict normal or abnormal outcome, nor did the proportion of EEG occupied by IEDs, although there was a higher likelihood of poor outcome with more severe EEG abnormalities. Overall, our findings enable greater prediction of abnormal outcomes, rather than confidently identifying neonatal factors associated with normal outcomes. We did not identify any factors that clearly differentiated mild‐moderately abnormal developmental outcomes from severe‐profoundly abnormal outcomes. Our study was limited in addressing this, by virtue of small cohort size, short follow‐up duration, and lack of quantitative developmental data. Thus, there remains an urgent need for identification of more sensitive predictors of outcome, potentially through more detailed prospective study of a larger patient group with early molecular diagnosis. We acknowledge that our study is limited by its small number of promptly diagnosed patients, retrospective nature, and the inability to control for variability in clinical practice. This may influence the available data (e.g., EEG acquisition methods varied, with variable duration of wake and sleep recordings, which could impact IED frequency), and impact clinical and health economic outcomes over and above the underlying severity of the disease. Notably, given the infants with a prompt diagnosis were born in more recent years (reflecting the availability of rapid and ultrarapid genomic testing), we did not see major differences in first‐ or second line ASM use (e.g., phenobarbitone use similar in the prompt and later diagnosis group) or supportive treatments used in the neonatal unit. This highlights that other major changes in neonatal practice over time are not likely to be determining the differences that we found between groups. Our findings emphasize that prompt diagnosis of KCNQ2 neonatal epilepsy and early implementation of SCB ASM can enable seizure freedom, typically with monotherapy, and reduce the need for hospital readmission in the first year of life. In addition to the health economic benefits, the clinical benefit to patients and families of prompt molecular diagnosis cannot be overestimated. AUTHOR CONTRIBUTIONS TJ, SEB, IES, ZS, KBH: study concept and design. All authors: acquisition, analysis, or interpretation of data; critical revision of the manuscript for important intellectual content. TJ, SEB, IES, KBH, ZS: drafting of the manuscript. ZS, SL, KBH: obtained funding. ZS, KBH: study supervision. CONFLICT OF INTEREST STATEMENT KBH is supported by a Melbourne Children’s Campus clinician–scientist fellowship, has received research funding from the National Health and Medical Research Council, the Medical Research Futures Fund, UCB Australia, Praxis Precision Medicines and RogCon Biosciences, Inc., has served on an advisory board for UCB Australia, and is a member of the Scientific and Medical Board for SCN2A Asia‐Pacific. MCF has received research funding from the National Health and Medical Research Council, the Medical Research Futures Fund, and the Cerebral Palsy Alliance and is a consultant for Fenix Innovation. The Murdoch Children’s Research Institute is supported by the Victorian Government’s Operational Infrastructure Support Program. EML is supported by a Melbourne Children’s Campus clinician–scientist fellowship and Australia New Zealand Child Neurology Society Kate Sinclair Memorial scholarship. IES is supported by an Australian National Health and Medical Research Council Investigator Grants (GNT2033247, GNT1172897), Synergy Grant (GNT2010562) and Centre for Research Excellence (GNT2006841), and Medical Research Futures Fund (MRF2007707). IES has served on scientific advisory boards for Biocodex, BioMarin, Chiesi, Eisai, Encoded Therapeutics, Knopp Biosciences, Longboard Pharmaceuticals, Takeda Pharmaceuticals, UCB; has received speaker honoraria from Akumentis, Biocodex, BioMarin, Chiesi, Eisai, GlaxoSmithKline, Liva Nova, Nutricia, Stoke Therapeutics, Zuellig Pharma; has received funding for travel from Biocodex, BioMarin, Eisai, Encoded Therapeutics, GlaxoSmithKline, Stoke Therapeutics, UCB; has served as an investigator for Anavex Life Sciences, Biohaven Ltd., Bright Minds Biosciences, Cerebral Therapeutics, Cerecin Inc., Cereval Therapeutics, Encoded Therapeutics, EpiMinder Inc., ES‐Therapeutics, GW Pharma, Longboard Pharmaceuticals, Marinus, Neuren Pharmaceuticals, Neurocrine BioSciences, Ovid Therapeutics, Praxis Precision Medicines, Shanghai Zhimeng Biopharma, SK Life Science, Supernus Pharmaceuticals, Takeda Pharmaceuticals, UCB, Ultragenyx, Xenon Pharmaceuticals, Zogenix, Zynerba; and has consulted for Atheneum Partners, Biohaven Pharmaceuticals, Care Beyond Diagnosis, Cerecin Inc., Eisai, Epilepsy Consortium, Longboard Pharmaceuticals, Praxis, Stoke Therapeutics, UCB, Zynerba Pharmaceuticals; and is a Non‐Executive Director of Bellberry Ltd. and a Director of the Australian Academy of Health and Medical Sciences. She may accrue future revenue on pending patent WO61/010176 (filed: 2008): Therapeutic Compound; has a patent for SCN1A testing held by Bionomics Inc. and licensed to various diagnostic companies; has a patent molecular diagnostic/theranostic target for benign familial infantile epilepsy (BFIE) [PRRT2] 2011904493 & 2012900190 and PCT/AU2012/001321 (TECH ID:2012‐009). The remaining authors report no disclosures or conflicts of interest. 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. ETHICS STATEMENT The study was approved by the Melbourne Health Human Research Ethics Committee as part of the Australian Genomics Health Alliance protocol: HREC/16/MH/251. ACKNOWLEDGMENTS We thank participants and their families, members of the Acute Care Flagship of Australian Genomics, and treating clinicians. The Acute Care Genomics program was funded by the Medical Research Futures Fund, Genomics Health Futures Mission (GHFM76747), Royal Children’s Hospital Foundation grant (2020‐1259), and Queensland Genomics. In‐kind support was provided by Australian Genomics (National Health and Medical Research Council grants GNT1113531 and GNT2000001) and the Sydney Children’s Hospital Network. Open access publishing facilitated by The University of Melbourne, as part of the Wiley ‐ The University of Melbourne agreement via the Council of Australasian University Librarians. Jadhav T, Bouffler SE, Innes E, Fahey M, Hunter M, Kothur K, et al. KCNQ2 neonatal epilepsy: Impact of prompt diagnosis and treatment, and early predictors of outcome severity. Epilepsia Open. 2026;11:949–961. 10.1002/epi4.70266 Zornitza Stark and Katherine B. Howell contributed equally to this work. DATA AVAILABILITY STATEMENT The data that support the findings of this study are available from the corresponding author upon reasonable request. REFERENCES Associated Data Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.