Background: Several reports have demonstrated that prolongation of the QT interval is associated with sudden cardiac death (SCD). However, it is unknown whether any of the components within the QT interval are responsible for its association with SCD. Methods and Results: We examined the association of the individual QT-interval components (R-wave onset to R-peak, R-peak to R-wave end, ST-segment, T-wave onset to T-peak, and T-peak to T-wave end) with SCD in 12 241 participants (54±5.7 years; 26% black; 55% women) from the ARIC study (Atherosclerosis Risk in Communities). The QT interval and its components were measured at baseline (1987–1989) from 12-lead ECGs. SCD cases were adjudicated by a group of physicians through December 31, 2012. During a median follow-up of 23.6 years, a total of 346 cases of SCD were identified. Although prolongation of the QT interval was associated with a 49% increased risk of SCD (hazard ratio, 1.49; 95% confidence interval, 1.01–2.18), only the T-wave onset to T-peak component (per 1-SD increase: hazard ratio, 1.19; 95% confidence interval, 1.06–1.34) was associated with SCD and not any of the other components in separate models. When all of the QT-interval components were included in the same model, T-wave onset to T-peak remained the strongest predictor of SCD (per 1-SD increase: hazard ratio, 1.21; 95% confidence interval, 1.06–1.37). Conclusions: The risk of SCD with the QT interval is driven by prolongation of the T-wave onset to T-peak component. This suggests that shifting the focus from the overall QT interval to its individual components will refine SCD prediction in the community.