A risk score model for predicting intraprocedural cardiac injury during implantable cardioverter defibrillator implantation: Insights from the National Cardiovascular Data Registry.

Background: Individual risk factors of intraprocedural cardiac injury (cardiac perforation and tamponade) during implantable cardioverter defibrillator (ICD) placement have been documented. However, the prognostic impact of their coexistence has not been explored.

Objective: To develop a risk score model to identify patients at risk for intraprocedural cardiac injury.

Methods: We identified 438 679 patients from National Cardiovascular Data Registry (NCDR)-ICD who underwent de novo ICD implantation between 2010 and 2015, split randomly into a derivation cohort (n = 220 000) and a validation cohort (n = 218 679). The generalized estimating equations (GEEs) analysis with quasilikelihood under the independence model criterion goodness-of-fit statistics were used to identify the predictors of intraprocedural cardiac injury and a risk scoring model was developed. Model discrimination was assessed by receiver-operator characteristic curve and C-statistic.

Results: The risk of intraprocedural cardiac injury in the overall cohort was 0.13%. GEE analysis yielded seven variables (points in parentheses) that were strongly associated with intraprocedural cardiac injury: age, greater than 75 years (1), female gender (1), body mass index, less than 18.5 kg/m (1), hypertension (1), chronic lung disease (1), left bundle branch block (1), and continued warfarin use (1). Only prior history of coronary artery bypass grafting (CABG) (-1) was associated with reduced risk. A risk scoring system was developed that had good discrimination with a C-statistic of 0.72. The risk of intraprocedural cardiac injury increased with the increase in risk score from low risk (0.03%) to high risk (1.37%).

Conclusion: A practical risk score model can stratify patients into high- and low-risk groups for cardiac perforation or tamponade before undergoing ICD implantation.