INTRODUCTION: Despite technologic developments, accurate discrimination of ventricular tachyarrhythmia from rapid rhythms of nonventricular origin remains a challenge. We sought to examine the sensitivity and specificity of a dual-chamber arrhythmia detection algorithm, the PARAD algorithm, incorporated in a dual-chamber implantable cardioverter defibrillator, the Defender (ELA Medical). METHODS AND RESULTS: All detailed tachycardia episodes (i.e., with stored atrial and ventricular channel markers and electrograms) retrieved from the Holter memory of the device were analyzed from 95 patients (86 men and 9 women; age 62 +/- 12 years) implanted with the Defender with the algorithm programmed at nominal settings. Over a follow-up of 15 +/- 8 months, 559 detailed sustained tachycardia episodes detected in the tachycardia zone were gathered in 62 patients. Of the 300 ventricular tachycardia (VT) episodes, 298 were appropriately classified as VT by the algorithm. Of the 259 supraventricular tachycardia (SVT) episodes, 231 were appropriately classified as SVT. In 8 patients, 28 episodes of SVT were misclassified as VT: 25 atrial fibrillation episodes (6 patients), 2 sinus tachycardia (1 patient), and 1 atrial tachycardia. Calculated sensitivity and specificity were 99.3% and 89.2% on a per episode basis and 99.8% (95% confidence interval: 97.8 to 100) and 91.6% (95% confidence interval: 86.0 to 97.3) on a per patient basis, respectively. CONCLUSION: In this selected population of patients, the PARAD algorithm was safe and reliable for detection of a wide spectrum of tachyarrhythmias. Its specificity was high, particularly with respect to sinus tachycardia detection, but it must be improved with respect to detection of atrial fibrillation.
INTRODUCTION: Despite technologic developments, accurate discrimination of ventricular tachyarrhythmia from rapid rhythms of nonventricular origin remains a challenge. We sought to examine the sensitivity and specificity of a dual-chamber arrhythmia detection algorithm, the PARAD algorithm, incorporated in a dual-chamber implantable cardioverter defibrillator, the Defender (ELA Medical). METHODS AND RESULTS: All detailed tachycardia episodes (i.e., with stored atrial and ventricular channel markers and electrograms) retrieved from the Holter memory of the device were analyzed from 95 patients (86 men and 9 women; age 62 +/- 12 years) implanted with the Defender with the algorithm programmed at nominal settings. Over a follow-up of 15 +/- 8 months, 559 detailed sustained tachycardia episodes detected in the tachycardia zone were gathered in 62 patients. Of the 300 ventricular tachycardia (VT) episodes, 298 were appropriately classified as VT by the algorithm. Of the 259 supraventricular tachycardia (SVT) episodes, 231 were appropriately classified as SVT. In 8 patients, 28 episodes of SVT were misclassified as VT: 25 atrial fibrillation episodes (6 patients), 2 sinus tachycardia (1 patient), and 1 atrial tachycardia. Calculated sensitivity and specificity were 99.3% and 89.2% on a per episode basis and 99.8% (95% confidence interval: 97.8 to 100) and 91.6% (95% confidence interval: 86.0 to 97.3) on a per patient basis, respectively. CONCLUSION: In this selected population of patients, the PARAD algorithm was safe and reliable for detection of a wide spectrum of tachyarrhythmias. Its specificity was high, particularly with respect to sinus tachycardia detection, but it must be improved with respect to detection of atrial fibrillation.