ICD treatment has been shown to be more effective than antiarrhythmic drugs in the prevention of sudden cardiac death, due to malignant ventricular arrhythmias.1 Despite recent advances in device technology, inappropriate therapy (antitachycardia pacing or shock) still remains an unsolved problem. Electromagnetic interference is a cause of inappropriate therapy and EGMs stored by the devices usually show a pattern of electrical noise.

Case Report
       In a 70-year-old male patient with prior myocardial infarction and an ejection fraction of 25%, an ICD (Medtronic GEM DR 7275, Minneapolis, Minnesota) had been implanted due to spontaneous and inducible ventricular tachycardia while receiving amiodarone.
       He experienced the first ICD shock without any warning symptoms after switching the bathroom lights on at his vacation home six months later. The interrogation of the episode revealed that electrical noise was recorded in both atrial and ventricular channels before the discharge of the ICD (Figure 1).

Figure 1.

Telemetry of the episode. In both channels noise is recorded. In the ventricular channel the noise is interpreted as ventricular fibrillation and a 15 Joules shock is delivered. Top: Atrial channel, middle: ventricular channel, lower: event marker channel. AR = atrial refractory; AS = atrial sense; VS = ventricular sense; FS = fibrillation sense; FD = fibrillation detected; VFRx1 defibrillation: first ventricular fibrillation therapy; CE = end of charge; VR = ventricular refractory; CD = cardioversion-defibrillation charge delivered.

In the ventricular channel, the mean cycle length of the noise was 120 ms (500 bpm). The VF zone of the device was programmed at a cycle length < 330 ms. The noise was intepreted by the ventricular channel as ventricular fibrillation and a 15 Joules shock was delivered. There were no indications of lead or device dysfunction (pacing lead impedance: atrial 444 ohms, ventricular 523 ohms, shocking 45 ohms and battery voltage 3.00 volt). Subsequently, the control of the house electrical appliance identified an electrical current leakage from the electrical switch that was not grounded in the bathroom. Alternating current (AC) in Greece operates at a voltage of 220 at 50 Hz. The defect was repaired immediately. We also decreased the ventricular sensitivity from 0.3 mV (nominal value) to 0.45 mV. During ICD implantation the defibrillation threshold was tested at a sensitivity level of 1.2 mV, providing assurance that a VF episode could be detected at the present (0.45 mV) ventricular sensitivity level. The patient did not experience any inappropriate discharge during one-year follow-up.

Discussion
       Inappropriate ICD shock in a conscious patient, apart from being painful, can also cause proarrhythmia.² Inappropriate therapy occurs in 18–24% of patients with ICDs.³ Weber et al. reported 39% of inappropriate therapy due to atrial fibrillation, 30% to sinus tachycardia, and 21% to ventricular oversensing.³ Potential causes of ventricular oversensing are ventricular lead fracture or isolation defect, sensing of T-wave (double counting), sensing of P-wave in atrial flutter or tachycardia (farfield sensing), myopotentials, insufficient contact of the ventricular lead with myocardium and to electromagnetic interference.⁴ Cases of electromagnetic interference have been previously described, caused by electronic anti-theft surveillance devices, slot machines, electrical razors, during showering or touching a washing machine and external alternating current leakage, as in our case.^5–9 In conclusion, electromagnetic interference due to external alternating current leakage was the cause of the inappropriate ICD shock in this patient.