Background Identification of individuals in the community with left ventricular systolic dysfunction (LVSD) or left ventricular hypertrophy (LVH) may allow earlier initiation of disease-modifying treatment. We performed a comprehensive evaluation of the screening performance of B-type natriuretic peptide (BNP) and N-terminal pro-BNP (NT-proBNP) for LVSD or LVH. Methods In 2,429 subjects without a history of heart failure, myocardial infarction, valvular abnormalities, or a serum creatinine >2.0 mg/dL enrolled in the Dallas Heart Study, measurement of BNP and NT-proBNP and cardiovascular magnetic resonance imaging were performed. Results B-type natriuretic peptide and NT-proBNP were robustly associated with magnetic resonance imaging–defined LVH and LVSD (ejection fraction <55%) among men and women ( P < .0001 for each). In the overall population, neither test discriminated well for LVH or LVSD (area under the receiver operating characteristic curve [AUROC] <0.7). Among women, no differences in AUROC were observed between BNP and NT-proBNP. Among men, AUROCs were similar between BNP and NT-proBNP in the overall population, but among subgroups age 50 or older, or with hypertension, the AUROCs for NT-proBNP (0.73-0.79) were higher than for BNP (0.63-0.69, P < .05 for each comparison). Compared with subjects with isolated BNP elevation (>97.5th percentile), those with isolated NT-proBNP elevation had worse renal function and more LVH and coronary calcium ( P < .05 for each). Conclusions Overall, neither BNP nor NT-proBNP accurately discriminated subjects with LVH or LVSD in this predominately young and healthy population-based cohort. However, among high-risk men, NT-proBNP performed slightly better than BNP and comparably with other routinely used screening tests such as prostate-specific antigen measurement for prostate cancer.