Incident heart failure (HF) continues to pose a common and serious problem. We sought to examine the value of echocardiographic predictors of new HF in a community-based elderly population at risk for HF, independent of and incremental to clinical evaluation.Asymptomatic patients ≥65 years old, with ≥1 HF risk factor (hypertension, type 2 diabetes, or obesity) were recruited from the community; patients with valve disease, reduced ejection fraction (EF), and atrial fibrillation (AF) were excluded. Patients underwent standard clinical evaluation including calculation of the Charlson co-morbidity score and a comprehensive echocardiography including global longitudinal strain (GLS). Functional capacity was assessed by 6 min walk test. New HF and cardiovascular death were assessed after a mean follow-up of 14 ± 4 months by three independent cardiologists using Framingham criteria. Of 410 subjects (median age 70 years; 48% men), the prevalence of stage B HF was 13% [by LV hypertrophy (LVH)], 12% (by abnormal E/e'), 33% (by impaired GLS), and 31% [by left atrial enlargement (LAE)]. New HF symptoms developed in 49, and 2 died of cardiovascular causes, giving an event rate of 104/1000 person-years. These patients were older (P = 0.012), had a higher Charlson co-morbidity score (P < 0.001), larger LV mass and left atrium, higher E/e', and lower GLS (P < 0.05). LAE, LVH, abnormal GLS, and E/e' were independent predictors of new HF. In sequential models, LV mass and GLS added incremental information to clinical parameters. GLS significantly reclassified individuals (P = 0.002), but no reclassification improvement was identified using LV mass index, E/e', and left atrial volume index.Echocardiographic assessment (especially GLS and LV mass) provides incremental value in predicting incident HF.