The effect of irradiation during computed tomography (CT) imaging on implantable cardioverter-defibrillators (ICDs) has not been fully evaluated in various settings. The purposes of this study were to evaluate the occurrence of electromagnetic interference (EMI) during CT irradiation in various clinically available ICDs with phantom experiments and to determine the potential risks related to irradiation during CT imaging. Five types of clinically available ICDs from five manufacturers were tested. An ICD was combined with an electrocardiogram (ECG) simulator, mounted in a chest phantom, and subjected to CT imaging. Each ICD was irradiated at the maximal power level (tube voltage, 135 kVp; tube current, 510 mA; rotation time, 1.5 s). EMI was defined as oversensing, ventricular tachycardia/ventricular fibrillation (VT/VF) detection, noise, or shock delivery during CT imaging. For ICDs in which EMI was observed, EMI was then evaluated under 144 different irradiation conditions (tube voltage [four patterns from 80-135 kVp], tube current [six patterns from 50-550 mA], and rotation time [six patterns from 0.35-1.5 s]). Testing was also performed during irradiation at the typical doses in three clinical settings and in two settings with inappropriate irradiation of ICDs due to incorrect setup. Among the five ICDs, a shock was delivered by one ICD manufactured by Medtronic (Minneapolis, MN, USA) due to oversensing during irradiation, which occurred at the maximal power level. No oversensing was observed in other ICDs. In the malfunctioned ICD, oversensing was observed in 134 of 144 irradiation patterns, even at a low power in the ICD. The VF-detection criterion was fulfilled in 20 of 134 tests and was significantly associated with tube voltage, tube current, ration time, and tube voltage × rotation time interaction. Although oversensing was observed in three clinical settings (typical chest CT, CT coronary angiography after coronary artery bypass graft, and dynamic assessment for pleural tumors) and one situation during an incorrect scan range on the chest for head perfusion CT, they were not recognized as tachycardia beats. Oversensing was observed when scans were incorrectly set over the ICD during bolus tracking of contrast-enhanced CT. Maximal power CT imaging induced VT/VF detection and shock delivery in one model of ICD placed in a chest phantom. VT/VF detection was observed when tube voltages were high and irradiation times were longer. Oversensing can occur during inappropriate CT imaging, particularly when slices are positioned over the ICD.