Background frailty and disability are very common in older adults; they share some risk factors and pathophysiological mechanisms. Yet, they are different clinical entities. Objectives this study aimed to explore a potential hierarchical relationship between frailty and disability along the continuum of the disablement process. Design prospective cohort study. Setting the French Three-City (3C) study. Subjects the sample included 943 participants aged 75 and older. Methods the Fried frailty phenotype, Instrumental Activities of Daily Living (IADL) and basic Activities of Daily Living (ADL) were used. We distinguished between four mutually excluding groups: (i) robust (no frailty and no disability); (ii) pure frailty (no disability); (iii) frailty with IADL disability (no ADL disability) and (iv) frailty with IADL and ADL disabilities. We used Cox’s regression models to study the 4-year mortality risk associated with each status. Results Eight-two per cent of participants were classified according to the assumed hierarchy: 61.3% was robust, 5.4% frail, 10.5% frail and IADL-disabled and 4.8% frail, IADL and ADL-disabled. An extra group of 17% was identified with IADL-disabled individuals without frailty. This extra group was similar to pure frailty in terms of characteristics and risk of death, placing them along the continuum at an intermediate stage between robustness and the two most disabled sub-groups. Conclusions our findings suggest that including frailty along the continuum could be relevant to describe the whole disablement process. Frailty would occur upstream of the process and might be relevant to identify an opportune time window, where specific monitoring and clinical interventions could be implemented in order to interrupt the process at a potentially more reversible stage.< Réduire