Abstract
Diagnostic Yield of Genetic Testing in Epileptic Encephalopathy in Childhood
Mercimek-Mahmutoglu S, Patel J, Cordeiro D, Hewson S, Callen D, Donner EJ, Hahn CD, Kannu P, Kobayashi J, Minassian BA, Moharir M, Siriwardena K, Weiss SK, Weksberg R, Snead OC III. Epilepsia 2015;56(5):707–716.
OBJECTIVE: Epilepsy is a common neurologic disorder of childhood. To determine the genetic diagnostic yield in epileptic encephalopathy, we performed a retrospective cohort study in a single epilepsy genetics clinic. METHODS: We included all patients with intractable epilepsy, global developmental delay, and cognitive dysfunction seen between January 2012 and June 2014 in the Epilepsy Genetics Clinic. Electronic patient charts were reviewed for clinical features, neuroimaging, biochemical investigations, and molecular genetic investigations including targeted next-generation sequencing of epileptic encephalopathy genes. RESULTS: Genetic causes were identified in 28% of the 110 patients: 7% had inherited metabolic disorders including pyridoxine dependent epilepsy caused by ALDH7A1 mutation, Menkes disease, pyridox(am)ine-5-phosphate oxidase deficiency, cobalamin G deficiency, methylenetetrahydrofolate reductase deficiency, glucose transporter 1 deficiency, glycine encephalopathy, and pyruvate dehydrogenase complex deficiency; 21% had other genetic causes including genetic syndromes, pathogenic copy number variants on array comparative genomic hybridization, and epileptic encephalopathy related to mutations in the SCN1A, SCN2A, SCN8A, KCNQ2, STXBP1, PCDH19, and SLC9A6 genes. Forty-five percent of patients obtained a genetic diagnosis by targeted next-generation sequencing epileptic encephalopathy panels. It is notable that 4.5% of patients had a treatable inherited metabolic disease. SIGNIFICANCE: To the best of our knowledge, this is the first study to combine inherited metabolic disorders and other genetic causes of epileptic encephalopathy. Targeted next-generation sequencing panels increased the genetic diagnostic yield from <10% to >25% in patients with epileptic encephalopathy.
Commentary
Discovering the etiology of epileptic encephalopathies of childhood has long been the code pediatric neurologists and the families of affected children have tried to break. These disorders, fraught with intractable seizures and developmental delay, are all the more frustrating when the reason for the disease remains unknown. The traditional approach of a good history, physical exam, neuroimaging, and electroencephalography reveals an answer for many, yet a significant number remain cryptogenic. Biochemical and basic genetic (i.e., karyotype/microarray) testing have long been the next level of investigation, but the advent of next-generation gene sequencing has many hopeful that diagnostic answers can be achieved with a single test—saving time, money, and potentially developmental potential. Next-generation sequencing comes with significant financial burden, many times not covered by insurance, so defining the positive yield compared with traditional diagnostic methods is paramount. Likewise, for those children whose symptoms don't present a high suspicion for a well-defined disorder, is it better to cast many small nets (i.e., multiple biochemical tests) or one big net (i.e., multi-gene sequencing) to reach a diagnosis?
Mercimek-Mahmutoglu et al. begin to investigate these questions with their retrospective review of genetic testing yield in a cohort of children with epileptic encephalopathy. The cohort was referred from primarily neurologists and inpatient hospital units and evaluated in a genetic epilepsy clinic, thus this cohort represents the population without defined etiology after routine evaluation. The timing of genetic testing in relationship to routine evaluation is important, since common epileptic encephalopathies such as West syndrome have defined etiologies in 55% after routine evaluation (1). The authors took a real-world approach to investigating the cohort, with directed biochemical testing based on symptoms, followed by targeted gene sequencing when initial testing was negative. Through this method, 28% of patients had an identified genetic etiology—15% through biochemical testing and 13% through gene sequencing. Still, 72% remained undiagnosed. The gene-sequencing panels were not uniform, ranging from 38 to 327 genes investigated. Whether including more genes would have increased the yield of gene sequencing remains unclear. While 93% of the genetically diagnosed patients were identified using 38-gene panels, only 16% were investigated with a 327-gene panel, of which one had an identified genetic etiology not included in the smaller panel. Thus, potentially, another 7% of patients may have been diagnosed with the larger panel. At what point increasing the number of genes evaluated outweighs the potential for useful diagnostic results remains unknown. The largest study of whole exome sequencing (n = 264) showed a diagnostic yield of 11%, though many smaller studies found yields as high as 72% (2–5). As such, evaluating more genes may not always provide greater diagnostic yield. Extensive gene sequencing can often discover variants of unknown significance, which will require yet more testing to define their meaning—thus adding to the financial burden of the test. In addition, interpreting the meaning of variants of unknown significance requires experience in both genetics and diagnosis of epileptic encephalopathies to avoid misdiagnosis—for example, not every patient with SCN1A mutations has Dravet syndrome.
Despite the potential financial burden of multi-gene testing, the importance of finding an etiology early should not be overlooked, especially in those disorders where early treatment can positively impact outcome (i.e., inherited metabolic disorders, SCN1A, KCNQ2). The authors describe one case of GLUT-1 deficiency that was not diagnosed until 11.5 years of age, potentially losing valuable time in neurodevelopment because effective therapy (i.e., ketogenic diet) was not applied earlier. The presentation of intractable absence with normal development initially was not classic GLUT-1 deficiency, but this underscores the role early gene testing has on identifying etiologies in those patients with atypical intractable epilepsy phenotypes. Increasingly, milder phenotypes of epileptic encephalopathies and novel syndromes caused by gene mutations are being identified through gene sequencing, thus the positive yield will likely only increase with time.
It is noted that the goal of this article was not to define the value of next-generation sequencing in isolation, but instead to assess the diagnostic yield of all genetic testing done in the cohort. When first-line directed biochemical testing and array comparative genomic hybridization failed to provide an etiology, second-line targeted next-generation sequencing was completed. Among the eight patients with inherited metabolic disorders identified through biochemical testing, six had gene mutations that would have identified their disease. Of the 23 patients with noninherited metabolic disorders (non-IMD), 19 had mutations that could be identified with a gene panel. We are left to ponder whether reversing the order of operations should be considered; completing multi-gene sequencing panels initially, followed by directed biochemical testing to confirm the diagnosis. The time between establishing a clinical diagnosis of epileptic encephalopathy and determining the cause is not explicitly stated, but at least half of the children with inherited metabolic causes were diagnosed with biochemical testing within months of presentation. While turnaround time for next-generation sequencing results continues to improve, for these patients biochemical testing likely shortened time to diagnosis. At the same time, several patients, many in the non-IMD group, had a time to diagnosis >6 months, which suggests early gene sequencing could have afforded more rapid diagnosis. While there is still not enough evidence to suggest extensive gene testing should be a first-line test, there may be a place for rapid turnaround testing (i.e., 1 to 2 weeks) offered by several companies for selected genes in which therapeutic intervention may positively impact outcome. We also don't know the number of biochemical tests, repeated MRIs, or other invasive diagnostic evaluations that could have been limited or avoided altogether had gene testing been done earlier. As multi-gene testing becomes more commonplace and affordable, its place in the diagnostic algorithm will become clearer.
Without doubt, the yield of genetic testing in patients with epileptic encephalopathy for which routine evaluation is unrevealing is clinically significant. Next-generation sequencing provides additional diagnostic yield in patients with cryptogenic epileptic encephalopathy. With shortened time to produce a result and decrease in cost, multi-gene sequencing may one day usurp the multitude of diagnostic tests that are currently employed to diagnosis the causes of epilepsy. Gene sequencing is very likely to aid in earlier diagnosis and improved outcome thanks to earlier appropriate treatment, especially in those patients where clinical symptoms do not suggest a well-defined epilepsy syndrome.