Timing of cardioverter-defibrillator implantation in patients with cardiac laminopathies-External validation of the LMNA-risk ventricular tachyarrhythmia calculator

BACKGROUND LMNA genotype-positive patients have high risk of experiencing life-threatening ventricular tachyarrhythmias (VTAs). The LMNA-risk VTA calculator published in 2019 has not been exter-nally validated. OBJECTIVE The purpose of this study was to validate the LMNA-risk VTA calculator. METHODS We included LMNA genotype-positive patients without previous VTAs from 2 large Scandinavian centers. Patients under-went electrocardiography, 24-hour Holter monitoring, and echocar-diographic examinations at baseline and repeatedly during follow-up. Validation of the LMNA-risk VTA calculator was performed using Harrell's C-statistic derived from multivariable Cox regression anal-ysis. RESULTS We included 118 patients (age 37 years [IQR 27-49 years]; 39 [33%] probands; 65 [55%] women; 100 [85%] with non-missense LMNA variants). Twenty-three patients (19%) experi-enced VTA during 6.1 years (interquartile range 3.0-9.1 years) follow-up, resulting in 3.0% (95% confidence interval 2.0%- 4.5%) yearly incidence rate. Atrioventricular block and reduced left ventricular ejection fraction were independent predictors of VTAs, while nonsustained ventricular tachycardia, male sex, and non-missense LMNA variants were not. The LMNA-risk VTA calculator showed 83% sensitivity and 26% specificity for identifying patients with VTAs during the coming 5 years, and a Harrell's C-statistic of 0.85, when applying >= 7% predicted 5-year VTA risk as threshold. The sensitivity increased to 100% when reevaluating risk at the time of last consultation before VTA. The calculator overestimated arrhythmic risk in patients with mild and moderate phenotype, particularly in men. CONCLUSION Validation of the LMNA-risk VTA calculator showed high sensitivity for subsequent VTAs, but overestimated arrhythmic risk when using >= 7% predicted 5-year risk as threshold. Frequent reevaluation of risk was necessary to maintain the sensitivity of the model.