Inappropriate implantable cardioverter- defibrillator(ICD)shocks continue to be a major source of distress to patients and a drain on the health care system. Expanding indications for ICD implantation include a large portion of patients with heart failure. This study investigated the relation between inappropriate ICD shocks and the severity of heart failure symptoms. Predictors of the time to first inappropriate ICD therapy were investigated in 230 consecutive patients implanted in 2001 and 2002. Thirty- two patients received 42 inappropriate shocks during a median follow- up of 501 days. Inappropriate shocks were due to atrial fibrillation(AF) or tachycardia(n=31), other supraventricular tachycardias(n=6), sinus tachycardia(n=3), and noise or double counting(n=2). The time to first inappropriate ICD shock was earliest in patients with advanced classes of heart failure(1- and 2- year shock- free survival of 79% and 70% for patients in New York Heart Association [NYHA] class III or IV vs 92% and 88% for patients in NYHA class I or II, respectively, p=0.02). After correcting for age, gender, the presence of coronary artery disease, the presence of AF, the use of β blockers, and indication for ICD implantation in a Cox regression model, advanced heart failure(NYHA class III or IV) remained an independent predictor of first inappropriate ICD shocks(hazard ratio 2.7, p=0.01). Other predictors of the time to first inappropriate ICD shock included the presence of AF as the baseline rhythm at the time of the ICD implantation and the absence of coronary disease. In conclusion, inappropriate ICD shocks are predominantly due to AF. Advanced heart failure is an independent predictor of the time to first inappropriate ICD shocks. The effect of ICD programming and antiarrhythmic drug therapy on the incidence of inappropriate shocks deserves further investigation. Inappropriate implantable cardioverter-defibrillator (ICD) shocks continue to be a major source of distress to patients and a drain on the health care system. Expanding indications for ICD implantation include a large portion of patients with heart failure. ICD shocks and the severity of heart failure symptoms. Predictors of the time to first inappropriate ICD therapy were investigated in 230 consecutive patients implanted in 2001 and 2002. Thirty- two patients received 42 inappropriate shocks during a median follow- up of 501 days. shocks were due to atrial fibrillation (AF) or tachycardia (n = 31), other supraventricular tachycardias (n = 6), sinus tachycardia (n = 3), and noise or double counting ICD shock was earliest in patients with advanced classes of heart failure (1- and 2-year shock-free survival of 79% and 70% for patients in New York Heart Association [NYHA] class III or IV vs 92% and 88% for patients in NYHA class I or II, respectively, p = 0.02). After correcting for age, gender, the presence of coronary artery disease, the presence of AF, the use of β blockers, and indication for ICD implantation in a Cox regression model, advanced heart failure (NYHA class III or IV) remained an independent predictor of first inappropriate ICD shocks (hazard ratio 2.7, p = 0.01). Other predictors of the time to first inappropriate ICD shock included the presence of of AF as the baseline rhythm at the time of the baseline and at the absence of coronary disease. In conclusion, inappropriate ICD shocks are predominantly due to AF. Advanced heart failure is an independent predictor of the time to first inappropriate ICD shocks. The effect of ICD programming and antiarrhythmic drug therapy on the incidence of inappropriate shocks deserves further investigation.