To investigate the prevalence of frailty in a Dutch elderly population and to identify adverse health outcomes associated with the frailty phenotype independent of the comorbidities. Cross-sectional and longitudinal analyses within the Rotterdam Study (the Netherlands), a prospective population-based cohort study in persons aged ≥55 years. Frailty was defined as meeting three or more of five established criteria for frailty, evaluating nutritional status, physical activity, mobility, grip strength and exhaustion. Intermediate frailty was defined as meeting one or two frailty criteria. Comorbidities were objectively measured. Health outcomes were assessed by means of questionnaires, physical examinations and continuous follow-up through general practitioners and municipal health authorities for mortality. Of 2,833 participants (median age 74.0 years, inter quartile range 9) with sufficiently evaluated frailty criteria, 163 (5.8 %) participants were frail and 1,454 (51.3 %) intermediate frail. Frail elderly were more likely to be older and female, to have an impaired quality of life and to have fallen or to have been hospitalized. 108 (72.0 %) frail participants had ≥2 comorbidities, compared to 777 (54.4 %) intermediate frail and 522 (44.8 %) non-frail participants. Adjusted for age, sex and comorbidities, frail elderly had a significantly increased risk of dying within 3 years (HR 3.4; 95 % CI 1.9-6.4), compared to the non-frail elderly. This study in a general Dutch population of community-dwelling elderly able to perform the frailty tests, demonstrates that frailty is common and that frail elderly are at increased risk of death independent of comorbidities.