Discussion
Device interrogation showed stable overall capture thresholds and impedances. Lead dislocation was excluded by means of chest X-ray showing normal position of atrial and ventricular leads, fully similar to previous X-rays. These findings support the hypothesis that the found interference signal was of external origin.
Upon closer examination of the ICD-interrogation showed a few instances of different interpretations of the interference signal by the ICD (figure 1).
The main concerns that remained were:
  1. (Near-)Syncope caused by asystole / bradycardia due to inhibition of pacing
  2. Oversensing induced inappropriate tachy therapy or even shocks
The chosen solution was programming separate levels of sensitivity for the detection of bradycardia and tachycardia providing different responses to the same noise signal. The sensitivity threshold for bradycardia detection was elevated to a level above the amplitude of the recurrent noise signal (i.e. 1mV). This significantly reduces the risk inappropriate inhibition of ventricular pacing resulting in asystole and danger of fainting. In addition, the sensitivity threshold for tachycardia detection was kept below the amplitude of the noise signal. By doing so the noise could still be monitored for change in amplitude or duration. And raising the sensitivity threshold for detecting tachycardia would also have introduced the danger of undersensing ventricular arrhythmias, especially ventricular fibrillation.
In this particular case, the SecureSenseā„¢ algorithm was successful in distinguishing noise from true high rate ventricular episodes and thereby provided the ability to withhold tachycardia therapy in the presence of the recurrent interference signal. To further reduce the risk of inappropriate shocks the incount for ventricular tachycardia was increased to 50 intervals.