Automatic identification of implantable cardioverter-defibrillator lead problems using intracardiac electrograms

Implantable Cardioverter-Defibrillator (ICD) lead problems can cause inappropriate painful shocks or inappropriately withheld lifesaving shocks. ICDs contain storage for detected spontaneous episodes that can be analyzed to characterize lead performance. The goal of this project was to develop an automatic lead problem identification algorithm using stored episode data. The algorithm combines sensed RR interval patterns and electrogram (EGM) characteristics to identify non-cardiac (NC) oversensing (OS) problems (e.g. lead failures) and cardiac (C) OS problems (e.g. T-wave OS). Stored episodes from 59 patients with OS and 147 patients with no OS were used for evaluation. The sensitivities to identify the lead problems were 97.7% for NC-OS and 86.7% for C-OS with a specificity of 98.0%. Analysis of stored episodes with EGM maybe used to identify ICD lead problems with very high sensitivity and specificity.