A retrospective cohort study and review of the literature about germline mosaicism in Duchenne/Becker muscular dystrophy prenatal counseling: How to estimate the recurrence risk in clinical settings?

Duchenne muscular dystrophy (DMD) and Becker muscular dystrophy (BMD) are the most common inherited neuromuscular diseases. Following the identification of a pathogenic causative variant in the DMD gene of a proband, potential carriers can be informed of their risk of having offspring with the disease. Germline mosaicism is a variant that is confined to the gonads that can be transmitted to offspring and is usually reported when a non-carrier of a DMD pathogenic variant has two or more offspring carrying the variant in question. On average, one third of cases are the result of a de novo variant, and as DMD and BMD are prone to germline mosaicism, its inclusion in genetic counseling is mandatory. In this retrospective cohort study, we presented clinical data from an unpublished DMD/BMD cohort of 332 families with incidence of germline mosaicism in families with de novo transmission of 8.1%. This is also the first systematic literature review searching PubMed to provide an accurate assessment of the current literature on germline mosaicism in DMD and BMD, including 17 case reports and 20 original studies. The incidence of documented germline mosaicism in de novo event families ranged from 6.0 to 40%, with a mean of 8.3%. The estimated recurrence risk for mothers of a patient with a proven de novo causal variant ranged from 4.3 to 11%, with a mean of 5.8% for a male fetus. By providing an up-to-date and comprehensive overview of the literature, this review aims to improve our understanding of germline mosaicism in DMD and to promote the development of effective strategies and reliable data for occurrence risk assessment in genetic counseling of de novo event families.

Revisiting GDF9 variants in primary ovarian insufficiency: A shift from dominant to recessive pathogenicity?

BACKGROUND: Primary ovarian insufficiency (POI) affects around 2-4% of women before the age of 40. Genetic factors play an important role in POI. The GDF9 gene has been identified as a significant genetic contributor of POI. However, the pathogenicity and penetrance of GDF9 variants remain uncertain. METHODS: A next-generation sequencing approach was employed to investigate the entire coding region of the GDF9 gene in a cohort of 1281 patients with POI or diminished ovarian reserve (DOR). The frequency of each identified GDF9 variant was then compared with that of the general population, taking into account the ethnicity of each individual. RESULTS: By screening the entire coding region of the GDF9 gene, we identified 19 different variants, including 1 pathogenic frameshift variant. In total, 36 patients with POI/DOR (2.8%) carried at least one GDF9 variant. With regard to missense variants, no significant overrepresentation of the most common variants was observed in our POI/DOR cohort in comparison to the general or specific ethnic subgroups. Only one homozygous subject had a frameshift loss of function variant. CONCLUSION: This epidemiological study suggests that the vast majority of heterozygous missense variants could be considered as variants of uncertain significance and the homozygous loss-of-function variant could be considered as a pathogenic variant. The identification of a novel case of a homozygous POI patient with a heterozygous mother carrying the same variant with normal ovarian function strongly suggests that GDF9 syndrome is an autosomal recessive disorder.

Assessing biomarkers of remission in female patients with anorexia nervosa (REMANO): a protocol for a prospective cohort study with a nested case-control study using clinical, neurocognitive, biological, genetic, epigenetic and neuroimaging markers in a French specialised inpatient unit.

BACKGROUND: Anorexia nervosa (AN) is a severe psychiatric disorder associated with frequent relapses and variability in treatment responses. Previous literature suggested that such variability is influenced by premorbid vulnerabilities such as abnormalities of the reward system. Several factors may indicate these vulnerabilities, such as neurocognitive markers (tendency to favour delayed reward, poor cognitive flexibility, abnormal decision process), genetic and epigenetic markers, biological and hormonal markers, and physiological markers.The present study will aim to identify markers that can predict body mass index (BMI) stability 6 months after discharge. The secondary aim of this study will be focused on characterising the biological, genetic, epigenetic and neurocognitive markers of remission in AN. METHODS AND ANALYSIS: One hundred and twenty-five (n=125) female adult inpatients diagnosed with AN will be recruited and evaluated at three different times: at the beginning of hospitalisation, when discharged and 6 months later. Depending on the BMI at the third visit, patients will be split into two groups: stable remission (BMI≥18.5 kg/m²) or unstable remission (BMI<18.5 kg/m²). One hundred (n=100) volunteers will be included as healthy controls.Each visit will consist in self-reported inventories (measuring depression, anxiety, suicidal thoughts and feelings, eating disorders symptoms, exercise addiction and the presence of comorbidities), neurocognitive tasks (Delay Discounting Task, Trail-Making Test, Brixton Test and Slip-of-action Task), the collection of blood samples, the repeated collection of blood samples around a standard meal and MRI scans at rest and while resolving a delay discounting task.Analyses will mainly consist in comparing patients stabilised 6 months later and patients who relapsed during these 6 months. ETHICS AND DISSEMINATION: Investigators will ask all participants to give written informed consent prior to participation, and all data will be recorded anonymously. The study will be conducted according to ethics recommendations from the Helsinki declaration (World Medical Association, 2013). It was registered on clinicaltrials.gov on 25 August 2020 as 'Remission Factors in Anorexia Nervosa (REMANO)', with the identifier NCT04560517 (for more details, see https://clinicaltrials.gov/ct2/show/record/NCT04560517). The present article is based on the latest protocol version from 29 November 2019. The sponsor, Institut National de la Santé Et de la Recherche Médicale (INSERM, https://www.inserm.fr/), is an academic institution responsible for the monitoring of the study, with an audit planned on a yearly basis.The results will be published after final analysis in the form of scientific articles in peer-reviewed journals and may be presented at national and international conferences. TRIAL REGISTRATION NUMBER: clinicaltrials.govNCT04560517.

Diagnostic yield of repeat genetic testing in idiopathic chronic pancreatitis.

Genetic testing is performed for unexplained pancreatitis. The aim of this study was to evaluate the diagnostic value of repeating genetic testing in idiopathic pancreatitis when new predisposing genes are identified. We investigated 330 patients who were initially screened for PRSS1, SPINK1 and CFTR genes. A new analysis was performed by Next-Generation Sequencing (NGS) for PRSS1, SPINK1, CFTR, CTRC, CASR, CPA1, TRPV6 genes and the CEL-HYB1 allele in clinical practice, and patients were included in our cohort study. Additional rare variants were identified in 7.3 % of the patients. Screening for new pancreatitis genes is recommended when initial screening is limited. Routine use of NGS is a useful diagnostic tool in these cases.

Shifting the landscape: Dominant C-terminal rare missense FOXL2 variants in non-syndromic primary ovarian failure etiology.

Pathogenic germline variants in the FOXL2 gene are associated with Blepharophimosis, Ptosis, and Epicanthus Inversus syndrome (BPES) in humans, an autosomal dominant condition. Two forms of BPES have emerged: (i) type I (BPES-I), characterized by ocular signs and primary ovarian failure (POI), and (ii) type II (BPES-II) with no systemic associations. This study aimed to compare the distribution of FOXL2 variants in idiopathic POI/DOR (diminished ovarian reserve) and both types of BPES, and to determine the involvement of FOXL2 in non-syndromic forms of POI/DOR. We studied the whole coding region of the FOXL2 gene using next-generation sequencing in 1282 patients with non-syndromic POI/DOR. Each identified FOXL2 variant was compared to its frequency in the general population, considering ethnicity. Screening of the entire coding region of the FOXL2 gene allowed us to identify 10 different variants, including nine missense variants. Of the patients with POI/DOR, 14 (1%) carried a FOXL2 variant. Significantly, six out of nine missense variants (67%) were overrepresented in our POI/DOR cohort compared to the general or specific ethnic subgroups. Our findings strongly suggest that five rare missense variants, mainly located in the C-terminal region of FOXL2 are high-risk factors for non-syndromic POI/DOR, though FOXL2 gene implication accounts for approximately 0.54% of non-syndromic POI/DOR cases. These results support the implementation of routine genetic screening for patients with POI/DOR in clinical settings.

The value of plasma cell-free DNA levels as biomarker in patients with eating disorders: A preliminary study.

OBJECTIVE: Circulating cell-free DNA (cfDNA) holds promise as a rapid and convenient biomarker for identifying individuals with eating disorders. To investigate this hypothesis, we measured plasma cfDNA in patients with different eating disorders. METHODS: In this study, 110 participants (98 patients with eating disorders divided into 30 patients with bulimia nervosa, 33 patients with anorexia nervosa (AN) Restricting subtype, 35 patients with AN Binge-eating/purging subtype and 12 controls) were enrolled. We measured both cell-free nuclear DNA (cf-nDNA) and cell-free mitochondrial DNA (cf-mtDNA) from plasma using two specific droplet digital PCR assays each, referring to two amplicon sizes. RESULTS: Levels of plasma cf-nDNA and cf-mtDNA showed no significant differences between control participants and those with eating disorders. However, we observed a higher proportion of long cf-nDNA fragments in patients with eating disorders, suggesting its potential as a biomarker for eating disorders. CONCLUSION: This is the first study of cfDNA in patients with eating disorders. Our findings highlight the potential for qualitative exploration of cfDNA, although not of quantitative interest. Full characterization of cfDNA may serve as a valuable biomarker for eating disorders and provide some insights into the hidden mechanisms underlying the chronic development of these conditions. Future studies are needed to confirm or refute this hypothesis.

NOBOX gene variants in premature ovarian insufficiency: ethnicity-dependent insights.

PURPOSE: Premature ovarian insufficiency (POI) affects approximately 1% of women before the age of 40. Genetic contribution is a significant component of POI. The NOBOX gene was considered one of the major genetic causes of POI. However, the pathogenicity and the penetrance of NOBOX variants remain unclear. METHODS: We studied the whole coding region of the NOBOX gene by next generation sequencing in a cohort of 810 patients with POI, and we compared the frequency of each identified NOBOX variant to the general population taking into account the ethnicity of each individual. RESULTS: Screening of the whole coding region of the NOBOX gene allowed us to identify 35 different variants, including 5 loss-of-function variants. In total, 171 patients with POI (25%) carried out at least one NOBOX variant. Regarding missense variants, we observed a significant overrepresentation of the most frequent ones in our 810 POI patients as compared to the general, except for p.(Arg117Trp). However, taking into account the ethnic origin of the individuals, we observed no significant OR difference for p.(Arg44Leu) and p.(Arg117Trp) in African subgroup and for p.(Asp452Asn) in European subgroup. CONCLUSION: This population study suggests that the p.(Arg44Leu) variant could be considered benign variant and that the p.(Asp452Asn) and p.(Arg117Trp) variants could be considered moderate risk pathogenic variants with probably partial and very low penetrance and/or expressivity. In contrast, p.(Gly91Trp) and p.(Gly152Arg) variants could be considered pathogenic variants with a moderate functional impact.

Stratification of the risk of ovarian dysfunction by studying the complexity of intermediate and premutation alleles of the FMR1 gene.

FMR1 premutation female carriers are at risk of developing premature/primary ovarian insufficiency (POI) with an incomplete penetrance. In this study, we determined the CGG repeat size among 1095 women with diminished ovarian reserve (DOR) / POI and characterized the CGG/AGG substructure in 44 women carrying an abnormal FMR1 repeat expansion number, compared to a group of 25 pregnant women carrying an abnormal FMR1 CGG repeat size. Allelic complexity scores of the FMR1 gene were calculated and compared between the two groups. In the DOR/POI cohort, 2.1% of women presented with an intermediate repeat size and 1.9% with a premutation. Our results suggest that the risk of POI is highest in the mid-range of CGG repeats. We observed that the allelic score is significantly higher in POI women compared to the pregnant women group (p-value = 0.02). We suggest that a high allelic score due to more than 2 AGG interspersions in the context of an intermediate number of repetitions could favor POI. Larger studies are still needed to evaluate the relevance of this new tool for the determination of the individual risk of developing POI in women with abnormal number of CGG repeats.

Intellectual Disability and Behavioral Deficits Linked to CYFIP1 Missense Variants Disrupting Actin Polymerization.

BACKGROUND: 15q11.2 deletions and duplications have been linked to autism spectrum disorder, schizophrenia, and intellectual disability. Recent evidence suggests that dysfunctional CYFIP1 (cytoplasmic FMR1 interacting protein 1) contributes to the clinical phenotypes observed in individuals with 15q11.2 deletion/duplication syndrome. CYFIP1 plays crucial roles in neuronal development and brain connectivity, promoting actin polymerization and regulating local protein synthesis. However, information about the impact of single nucleotide variants in CYFIP1 on neurodevelopmental disorders is limited. METHODS: Here, we report a family with 2 probands exhibiting intellectual disability, autism spectrum disorder, spastic tetraparesis, and brain morphology defects and who carry biallelic missense point mutations in the CYFIP1 gene. We used skin fibroblasts from one of the probands, the parents, and typically developing individuals to investigate the effect of the variants on the functionality of CYFIP1. In addition, we generated Drosophila knockin mutants to address the effect of the variants in vivo and gain insight into the molecular mechanism that underlies the clinical phenotype. RESULTS: Our study revealed that the 2 missense variants are in protein domains responsible for maintaining the interaction within the wave regulatory complex. Molecular and cellular analyses in skin fibroblasts from one proband showed deficits in actin polymerization. The fly model for these mutations exhibited abnormal brain morphology and F-actin loss and recapitulated the core behavioral symptoms, such as deficits in social interaction and motor coordination. CONCLUSIONS: Our findings suggest that the 2 CYFIP1 variants contribute to the clinical phenotype in the probands that reflects deficits in actin-mediated brain development processes.

[Towards a generalization of non-invasive prenatal diagnosis of single-gene disorders? Assesment and outlook]. / Vers une généralisation du diagnostic prénatal non-invasif des maladies monogéniques ? État des lieux et perspectives.

OBJECTIVES: The screening of fetal aneuploidies and non-invasive prenatal diagnosis of monogenic diseases (NIPD-MD) both rely on the study of free fetal DNA in maternal circulation, but their respective rise was unequal. Development of NIPD-MD has taken longer as it represents a less attractive commercial dynamic for industry, but also because it usually involves the development of tailored tests specific to each pathogenic variant. METHODS: We have carried out a review of the literature on the various indications and technologies involved in the use of NIPD-MM. We present its current implementation and its development in France. RESULTS: To date, NIPD-MD has been routinely offered in France for several years by the laboratories of the French NIPD-MD network but remains mostly limited to the exclusion of paternal or de novo variants, the exclusion DPNI-MD. Indeed, it is still difficult to study the transmission of maternal variants from circulating free DNA analysis, due to its biological complexity: coexistence and predominance of similar DNA sequences of maternal origin. Different strategies, either direct or indirect, are being evaluated to establish fetal status regardless of the parental origin of the disease or its transmission mode. The emergence of commercial screening solutions for monogenic diseases complements the arsenal of prenatal exploration tools for these diseases. CONCLUSION: The multitude of existing technologies and protocols may complicate the information provided during antenatal consultations, but mastery of know-how and knowledge of ethical issues of NIPD-MD will ensure optimal service and better management of pregnancies at risk of transmitting monogenic disease.

Non-invasive prenatal diagnosis of single gene disorders with enhanced relative haplotype dosage analysis for diagnostic implementation.

Non-invasive prenatal diagnosis of single-gene disorders (SGD-NIPD) has been widely accepted, but is mostly limited to the exclusion of either paternal or de novo mutations. Indeed, it is still difficult to infer the inheritance of the maternal allele from cell-free DNA (cfDNA) analysis. Based on the study of maternal haplotype imbalance in cfDNA, relative haplotype dosage (RHDO) was developed to address this challenge. Although RHDO has been shown to be reliable, robust control of statistical error and explicit delineation of critical parameters for assessing the quality of the analysis have not been fully addressed. We present here a universal and adaptable enhanced-RHDO (eRHDO) procedure through an automated bioinformatics pipeline with a didactic visualization of the results, aiming to be applied for any SGD-NIPD in routine care. A training cohort of 43 families carrying CFTR, NF1, DMD, or F8 mutations allowed the characterization and optimal setting of several adjustable data variables, such as minimum sequencing depth, type 1 and type 2 statistical errors, as well as the quality assessment of intermediate steps and final results by block score and concordance score. Validation was successfully performed on a test cohort of 56 pregnancies. Finally, computer simulations were used to estimate the effect of fetal-fraction, sequencing depth and number of informative SNPs on the quality of results. Our workflow proved to be robust, as we obtained conclusive and correctly inferred fetal genotypes in 94.9% of cases, with no false-negative or false-positive results. By standardizing data generation and analysis, we fully describe a turnkey protocol for laboratories wishing to offer eRHDO-based non-invasive prenatal diagnosis for single-gene disorders as an alternative to conventional prenatal diagnosis.

A systematic literature review and meta-analysis of circulating nucleic acids as biomarkers in psychiatry.

Common mental disorders (CMDs) such as depression, anxiety and post-traumatic stress disorders account for 40% of the global burden of disease. In most psychiatric disorders, both diagnosis and monitoring can be challenging, frequently requiring long-term investigation and follow-up. The discovery of better methods to facilitate accurate and fast diagnosis and monitoring of psychiatric disorders is therefore crucial. Circulating nucleic acids (CNAs) are among these new tools. CNAs (DNA or RNA) can be found circulating in body biofluids, and can be isolated from biological samples such as plasma. They can serve as biomarkers for diagnosis and prognoses. They appear to be promising for disorders (such as psychiatric disorders) that involve organs or structures that are difficult to assess. This review presents an accurate assessment of the current literature about the use of plasma and serum cell-free DNA (cfDNA) as biomarkers for several aspects of psychiatric disorders: diagnosis, prognosis, treatment response, and monitor disease progression. For each psychiatric disorder, we examine the effect sizes to give insights on the efficacy of CNAs as biomarkers. The global effect size for plasma nuclear and mitochondrial cfDNA studies was generally moderate for psychiatric disorders. In addition, we discuss future applications of CNAs and particularly cfDNA as non-invasive biomarkers for these diseases.

How a paternal uniparental isodisomy of chromosome 17 leads to autosomal recessive limb-girdle muscular dystrophy R3.

Uniparental isodisomy is a condition where both chromosomes of a pair are inherited from one parental homologue. If a deleterious variant is present on the duplicated chromosome, its homozygosity can reveal an autosomal recessive disorder in the offspring of a heterozygous carrier. Limb-girdle muscular dystrophy (LGMD) R3 is an autosomal recessive inherited disease that is associated with alpha-sarcoglycan gene (SGCA) variants. We report the first published case of LGMDR3 due to a homozygous variant in SGCA unmasked by uniparental isodisomy. The patient is an 8-year-old who experienced delayed motor milestones but normal cognitive development. He presented with muscle pain and elevated plasma creatine kinase. Sequencing of the SGCA gene showed a homozygous pathogenic variant. Parents were not related and only the father was heterozygous for the pathogenic variant. A chromosomal microarray revealed a complete chromosome 17 copy number neutral loss of heterozygosity encompassing SGCA, indicating paternal uniparental isodisomy.

WITHDRAWN: Clinical interpretation of PRSS1 gene variants in patients with pancreatitis.

Since the description of the PRSS1 gene encoding the cationic trypsinogen as being involved in dominant hereditary pancreatitis, more than 50 PRSS1 gene variants have been reported. Among those that have been classified as pathogenic, some have a high penetrance and others have a low penetrance. Assessing the clinical relevance of PRSS1 variants is often complicated in the absence of functional evidence and interpretation regarding rare variants is not very easy in clinical practice. The aim of this study was to review the PRSS1 variants and to classify them according to their degree of deleterious effect. This classification was based on the results of in vitro experiments and on population data, in comparing the allelic frequency of each variant in patients with pancreatitis and in unaffected individuals. This review should help geneticists and clinicians in charge of patient...s care and genetic counseling to interpret results of genetic studies.

The recurrent TCF4 missense variant p.(Arg389Cys) causes a neurodevelopmental disorder overlapping with but not typical for Pitt-Hopkins syndrome.

TCF4 haploinsufficiency by deletions, truncating variants or loss-of-function missense variants within the DNA-binding and protein interacting bHLH domain causes Pitt-Hopkins syndrome (PTHS). This neurodevelopmental disorder (NDD) is characterized by severe intellectual disability (ID), epilepsy, hyperbreathing and a typical facial gestalt. Only few aberrations of the N-terminus of TCF4 were associated with milder or atypical phenotypes. By personal communication and searching databases we assembled six cases with the novel, recurrent, de novo missense variant c.1165C > T, p.(Arg389Cys) in TCF4. This variant was identified by diagnostic exome or panel sequencing and is located upstream of the bHLH domain. All six individuals presented with moderate to severe ID with language impairment. Microcephaly occurred in two individuals, epilepsy only in one, and no breathing anomalies or myopia were reported. Facial gestalt showed some aspects of PTHS but was rather non-specific in most individuals. Interestingly, the variant is located within the AD2 activation domain next to a highly conserved coactivator-recruitment motif and might alter interaction with coactivator proteins independently from the bHLH domain. Our findings of a recurrent missense variant outside the bHLH domain in six individuals with an ID phenotype overlapping with but not typical for PTHS delineate a novel genotype-phenotype correlation for TCF4-related NDDs.

Comparing Two Neurodevelopmental Disorders Linked to CK2: Okur-Chung Neurodevelopmental Syndrome and Poirier-Bienvenu Neurodevelopmental Syndrome-Two Sides of the Same Coin?

In recent years, variants in the catalytic and regulatory subunits of the kinase CK2 have been found to underlie two different, yet symptomatically overlapping neurodevelopmental disorders, termed Okur-Chung neurodevelopmental syndrome (OCNDS) and Poirier-Bienvenu neurodevelopmental syndrome (POBINDS). Both conditions are predominantly caused by de novo missense or nonsense mono-allelic variants. They are characterized by a generalized developmental delay, intellectual disability, behavioral problems (hyperactivity, repetitive movements and social interaction deficits), hypotonia, motricity and verbalization deficits. One of the main features of POBINDS is epilepsies, which are present with much lower prevalence in patients with OCNDS. While a role for CK2 in brain functioning and development is well acknowledged, these findings for the first time clearly link CK2 to defined brain disorders. Our review will bring together patient data for both syndromes, aiming to link symptoms with genotypes, and to rationalize the symptoms through known cellular functions of CK2 that have been identified in preclinical and biochemical contexts. We will also compare the symptomatology and elaborate the specificities that distinguish the two syndromes.

Non-invasive prenatal diagnosis of single gene disorders by paternal mutation exclusion: 3 years of clinical experience.

OBJECTIVES: Cell-free fetal DNA (cffDNA) analysis is performed routinely for aneuploidy screening, RhD genotyping or sex determination. Although applications to single gene disorders (SGD) are being rapidly developed worldwide, only a few laboratories offer cffDNA testing routinely as a diagnosis service for this indication. In a previous report, we described a standardised protocol for non-invasive exclusion of paternal variant in SGD. Three years later, we now report our clinical experience with the protocol. DESIGN: Descriptive study. SETTING: Multi-centre French. POPULATION: Indications for referral included pregnancies at risk of 25% or 50% of paternally inherited SGD, and pregnancies associated with an increased risk of SGD due to a de novo variant, either from strongly suggestive ultrasound findings or from a possible parental germinal mosaicism in the context of a previously affected child. METHODS: Non-invasive prenatal diagnosis was performed using custom assays for droplet digital PCR. Feasibility, diagnostic performance and turn-around time were evaluated. RESULTS: Mean time for a new assay design and validation was evaluated at 14 days, and mean result reporting time was 6 days. All referred pathogenic variants could be targeted except one located in a complex genomic region. A result was obtained for every 198 referrals except two. CONCLUSION: This service was successfully implemented as a routine laboratory practice. It has been widely adopted by French clinicians and patients for paternal variant exclusion in various disorders. TWEETABLE ABSTRACT: A robust approach to non-invasive prenatal exclusion of paternal pathogenic variant in a diagnosis setting.

Novel UBE3A pathogenic variant in a large Georgian family produces non-convulsive status epilepticus responsive to ketogenic diet.

PURPOSE: To report the effect of the ketogenic diet (KD) on non-convulsive status epilepticus (NCSE) due to Angelman syndrome (AS) in two members of a large Georgian family affected by a novel frameshift variant in the UBE3A gene (NM_000462.3). METHODS: We evaluated two members of this family who were affected with clinical and EEG features of AS. Clinical history with special emphasis on development, seizure type, frequency, and treatment was reviewed. Routine and long-term video EEG monitoring were conducted, particularly during NCSE. A non-fasting inpatient KD protocol was implemented using blended food orally with full administration of 4:1 (fat to non-fat) ratio. Urine ketone bodies (KBs), measured with urine ketone acetone strips readings, reached 150 mg/dL in both patients. RESULTS: Patients had characteristic signs of AS and presented with epilepsy between the age of 2-4 years. As methylation tests were negative, next generation sequencing disclosed a c.2365del variant. For both, NCSE was revealed by cognitive deterioration and did not respond to anti-seizure medication. As recommended, IV pyridoxine, benzodiazepines, and valproic acid were administered, but without success. For both patients, NCSE resolved on the second-third day of KD initiation, before the appearance of ketonuria and resulting in improved communication, mood and sleep. CONCLUSION: KD is safe and effective for the treatment of NCSE due to AS. Resolution before the appearance of ketone bodies points to a possible mechanism of action of KD.