Background: Individuals with Brugada syndrome (BrS) in whom a variant can be found in SCN5A, the gene encoding the alpha subunit of the cardiac sodium channel, have a more marked phenotype. However, many SCN5A variants have little apparent effect on the sodium current INa. We asked whether in vitro functional assessment of SCN5A variants could refine the genotype-phenotype correlation.
Methods: A systematic literature search was performed to identify SCN5A variants associated with BrS. Details were extracted for channel function obtained from patch clamp recordings in cell lines expressing mutant channels (peak current, steady-state activation and inactivation, recovery from inactivation and current decay) as were clinical features (cardiac conduction disturbances including atrioventricular block and sinus node dysfunction, the composite of spontaneous ventricular arrhythmias or family history of sudden death (VA/SCD), and spontaneous BrS ECG pattern). Receiver operating characteristic (ROC) curves were constructed to evaluate area under the curve (AUC) and optimal cut-offs for continuous variables.
Results: Ninety-three SCN5A variants were identified for which robust in vitro and clinical characterisation was available.
Variants associated with conduction abnormalities had greater INa reduction (conduction abnormalities 75.5 ± 5.9%, no conduction abnormalities 53.8 ± 5.3%, p<0.01) and a more positive shift in
steady-state activation (conduction abnormalities +7.0 ± 0.7 mV, no conduction abnormalities +2.1 ± 1.7 mV, p<0.01). ROC analysis showed optimal cut-offs of 75% current reduction (AUC 0.65, sensitivity 68%, specificity 58%) and + 8 mV shift in activation (AUC 0.66, sensitivity 89%, specificity 42%). In the dataset as a whole, no association was found between VA/SCD or spontaneous BrS ECG and any functional parameters.
In a sensitivity analysis excluding 25 variants with pedigrees <5 people, a weak relationship was found between peak current and VAs (AUC 0.60, optimal cut-off 100% current reduction, sensitivity 42%, specificity 92%). When metrics were combined, 37/55 (67.3%) variants associated with VA/SCD and 3/13 (23.1%) not associated with VA/SCD had 100% current abolition and/or ≥8 mV activation shift (p<0.01).
Conclusions: There is a modest association between the deleterious effect of SCN5A variants on channel function and clinical phenotype, most markedly with conduction disturbances. Sensitivity analyses suggest these measures may have a potential role for risk stratification although this needs to be validated in datasets containing more precise clinical phenotyping.