SCN5A mutations in 442 neonates and children: genotype-phenotype correlation and identification of higher-risk subgroups.

Alban-Elouen Baruteau, Florence Kyndt, Elijah R Behr, Arja S Vink, Matthias Lachaud, Anna Joong, Jean-Jacques Schott, Minoru Horie, Isabelle Denjoy, Lia Crotti, Wataru Shimizu, Johan M Bos, Elizabeth A Stephenson, Leonie Wong, Dominic J Abrams, Andrew M Davis, Annika Winbo, Anne M Dubin, Shubhayan Sanatani, Leonardo Liberman, Juan Pablo Kaski, Boris Rudic, Sit Yee Kwok, Claudine Rieubland, Jacob Tfelt-Hansen, George F Van Hare, Béatrice Guyomarc'h-Delasalle, Nico A Blom, Yanushi D Wijeyeratne, Jean-Baptiste Gourraud, Hervé Le Marec, Junichi Ozawa, Véronique Fressart, Jean-Marc Lupoglazoff, Federica Dagradi, Carla Spazzolini, Takeshi Aiba, David J Tester, Laura A Zahavich, Virginie Beauséjour-Ladouceur, Mangesh Jadhav, Jonathan R Skinner, Sonia Franciosi, Andrew D Krahn, Mena Abdelsayed, Peter C Ruben, Tak-Cheung Yung, Michael J Ackerman, Arthur A Wilde, Peter J Schwartz, Vincent Probst

To clarify the clinical characteristics and outcomes of children with SCN5A-mediated disease and to improve their risk stratification. A multicentre, international, retrospective cohort study was conducted in 25 tertiary hospitals in 13 countries between 1990 and 2015. All patients ≤16 years of age diagnosed with a genetically confirmed SCN5A mutation were included in the analysis. There was no restriction made based on their clinical diagnosis. A total of 442 children {55.7% boys, 40.3% probands, median age: 8.0 [interquartile range (IQR) 9.5] years} from 350 families were included; 67.9% were asymptomatic at diagnosis. Four main phenotypes were identified: isolated progressive cardiac conduction disorders (25.6%), overlap phenotype (15.6%), isolated long QT syndrome type 3 (10.6%), and isolated Brugada syndrome type 1 (1.8%); 44.3% had a negative electrocardiogram phenotype. During a median follow-up of 5.9 (IQR 5.9) years, 272 cardiac events (CEs) occurred in 139 (31.5%) patients. Patients whose mutation localized in the C-terminus had a lower risk. Compound genotype, both gain- and loss-of-function SCN5A mutation, age ≤1 year at diagnosis in probands and age ≤1 year at diagnosis in non-probands were independent predictors of CE. In this large paediatric cohort of SCN5A mutation-positive subjects, cardiac conduction disorders were the most prevalent phenotype; CEs occurred in about one-third of genotype-positive children, and several independent risk factors were identified, including age ≤1 year at diagnosis, compound mutation, and mutation with both gain- and loss-of-function.