The relation of bundle branch block (BBB) with adverse outcome is controversial. We hypothesized that increased QRS duration is an independent predictor of cardiovascular (CV) mortality in a cross-sectional US population. This is a retrospective cohort study on prospectively collected data to assess the relationship between QRS duration on routine ECG and CV mortality. Participants included 8,527 patients with ECG data available from the National Health and Nutrition Examination Survey data set, representing 74,062,796 individuals in the United States. Mean age was 60.5 ± 13.6 years. Most subjects were white (87%) and women (53%). During the follow-up period of 106,244.6 person-years, 1,433 CV deaths occurred. Multivariate analysis revealed that the highest quartile of QRS duration was associated with higher CV mortality than lowest quartile (hazard ratio [HR] 1.3, 95% confidence interval [CI] 1.01 to 1.7, p = 0.04) after adjustment for established risk factors. Both left BBB (HR 2.4, 95% CI 1.3 to 4.7, p = 0.009) and right BBB (HR 1.90, 95% CI 1.2 to 3.0, p = 0.008) were significantly associated with increased CV mortality. The addition of the QRS duration in 10-millisecond increments to the Framingham Risk Score model resulted in 4.4% overall net reclassification improvement (95% CI 0.02 to 0.04; p = 0.00006). In conclusion, increased QRS duration was found to be an independent predictor of CV mortality in this cross-sectional US population. A model including QRS duration in addition to traditional risk factors was associated with improved CV risk prediction.