We report the case of a 69-year-old man with idiopathic dilated cardiomyopathy, biventricular implanted cardioverter–defibrillator (ICD) and a spontaneous ventricular tachycardia during device follow-up that was not automatically treated by the ICD. Four years before a dilated left ventricle (end-diastolic diameter, 92 mm) with a severely reduced ejection fraction (EF 20 %), congestive heart failure (NYHA class III), left bundle branch block (QRS duration 200 ms) and syncope during sustained ventricular tachycardia (VT) were noted. Subsequently, a biventricular ICD (Sorin Ovatio CRT-D, leads: RA: Medtronic 4076; RV: Guidant 0185; LV: Medtronic 4194) had been implanted for secondary prevention of sudden cardiac death (SCD) and for cardiac resynchronization. On the day after implantation an electrical storm had occurred and six effective shocks had been delivered. Amiodarone therapy was initiated and no further sustained ventricular arrhythmias occurred. However, due to skeletal muscle myopotential oversensing ventricular episodes were inadequately diagnosed by the ICD. Therefore, two further defibrillation testings were performed to ensure proper sensing of VF after stepwise increasing of the sensing threshold. The last test was performed with a sensing of 2.2 mV, and permanent ICD sensitivity was set to 1.6 mV. Sensing (RA: 4 mV; RV: 15 mV) and pacing thresholds (RA: 0.75 mV/0.35 ms; RV: 0.75 mV/0.35 ms; LV: 0.75 mV/0.35 ms) remained stable during the following years and no further noise signals were detected. At device follow-up 4 years after implantation, a sustained polymorphic VT (*200 bpm) occurred spontaneously (Fig. 1) during determination of ventricular sensing. The tachycardia was hemodynamically relevant with fainting of the patient, but ICD therapy was not delivered. The ICD even performed ventricular backup-pacing during the VT. ICD detection settings were: slow VT—120/bpm, 12 cycles; VT—150/bpm, 12 cycles; FVT—200/bpm, 10 cycles; VF—240/bpm. Intracardiac electrogram (IEGM) showed undersensing of the ventricular tachycardia due to the previously programmed low sensitivity. An internal manual shock was delivered that terminated the VT. Revision of the ICD was performed on the next day. Assuming that the combination of an integrated bipolar lead with the sensing algorithm and/or signal processing of the ICD manufacturer was responsible for the previous oversensing of myopotentials, only generator exchange (Boston Scientific Cognis 100-D CRT-D) was targeted at first. For intraoperative testing, ventricular sensing was set to 1.2 mV and a polymorphic fast VT was induced by T-wave shock. Now, there was no undersensing but underdetection of the induced arrhythmia, i.e. the local V–V interval detected by the ICD was clearly longer than the apparent cycle length in the ECG resulting in failure to diagnose VT (see Fig. 2). A manually applied ICD shock with 41 J terminated the arrhythmia. Programmed ventricular stimulation using the implanted RV electrode (apical position) revealed a locally long effective refractory period (ERP) of 400 ms (basic cycle length 500 ms, 3 times threshold). A new pace-sense electrode (Medtronic 5076) was placed in septal position and an ERP of 300 ms could be measured there. During ICD testing, the induced ventricular arrhythmia was appropriately detected using a I.-O. Lee (&) C. Ukena M. Bohm A. Buob H.-R. Neuberger Klinik fur Innere Medizin III (Kardiologie, Angiologie, Internistische Intensivmedizin), Universitatsklinikum des Saarlandes, 66421 Homburg/Saar, Germany e-mail: illkyu@hotmail.com