BACKGROUND: The European Society of Clinical Nutrition and Metabolism (ESPEN) has developed a consensus definition of malnutrition. This study aimed to determine the prevalence of malnutrition according to the ESPEN definition in otherwise healthy community-dwelling older women and to explore its value for predicting long-term mortality in this population. METHODS: This prospective population-based cohort study included 181 women (age ≥75 years) from a subsample of the EPIDémiologie de l'OStéoporose (EPIDOS) study participants from Toulouse. Inclusion criteria were the availability of the data on variables required to apply the ESPEN definition and survival after 7 years of follow-up. Primary outcome was mortality at 12-year follow-up; main covariates were malnutrition assessment according to the ESPEN consensus and its components (unintentional weight loss, BMI, and FFMI). Body composition was assessed by dual-energy X-ray absorptiometry at baseline and at 7-year follow-up. Kaplan-Meier survival curves and adjusted Cox regressions were performed. Analysis was adjusted for age, hypertension, diabetes mellitus, and coronary heart disease as potential confounders. RESULTS: Complete data were available for 179 of the 181 women in the EPIDOS-Toulouse cohort (83.1 ± 2.2 years) and 13 (7.3%) fulfilled the ESPEN definition for malnutrition at 7-year follow-up. Malnutrition was associated with increased risk of mortality (adjusted HR = 4.4 [95%CI: 1.7-11.3]). Among the ESPEN components, only BMI was associated with increased mortality (adjusted HR=0.6 [95%CI: 0.4-0.9]). CONCLUSIONS: Although malnutrition prevalence according to the ESPEN definition was relatively low (7.3%) in this sample of otherwise healthy community-dwelling older French women, malnutrition was associated with 4.4-fold higher mortality risk at 12-year follow-up.