BACKGROUND: Data on inappropriate implantable cardioverter-defibrillator (ICD) therapy and effects of programming by heart rate are lacking. OBJECTIVE: We aimed to characterize inappropriate ICD therapy and assess the effects of novel programming by heart rate. METHODS: Incidence and causes of inappropriate therapy by heart rate range (below or above 200 bpm) were assessed. Predictors of inappropriate therapy and effects of programming by heart rate were evaluated with multivariate Cox regression models. Crossovers were excluded. RESULTS: Inappropriate therapy occurred in 9.2% of the total patient population, with 19% of patients randomized to study arm A, 3.6% in arm B, and 4.7% in arm C. Inappropriate therapies <200 bpm were attributable to supraventricular tachycardia (SVT)/sinus tachycardia (78%) or atrial fibrillation/flutter (20%). Inappropriate therapy ≥200 bpm occurred because of SVT (47%), atrial fibrillation/flutter (41%), or electromagnetic interference (13%). Conventional ICD programming was associated with more inappropriate therapy <200 bpm than high-rate or delayed therapy, as were younger age, history of atrial arrhythmia, advanced New York Heart Association functional class, ICD versus cardiac resynchronization therapy with defibrillator, and absence of diabetes. High-rate and long-delay therapy significantly reduced the risk of inappropriate therapy in the <200 bpm range. Long delay was associated with further reduction of fast (≥200 bpm) inappropriate therapy (P = .032) and a reduction in subsequent inappropriate episodes (P = .006). CONCLUSION: In MADIT-RIT, inappropriate ICD therapy is most frequent at rates below 200 bpm and can be predicted, and effectively prevented, with high-rate cutoff programming. Long-delay therapy effectively reduces fast inappropriate therapy ≥200 bpm and subsequent events. [ CLINICAL TRIAL REGISTRATION: http://clinicaltrials.gov/ct2/show/NCT00947310].