Objectives: To assess the relationship between changes of frailty status and intervening hospitalizations, using information of the GAZEL cohort, matched with the data of the French National Health Data System. Design: Observational cohort study. Participants: Community-dwelling adults of the GAZEL cohort (n = 12145; aged between 58 and 73 years in 2012). Measurements: Frailty was determined with the Strawbridge questionnaire in 2012, 2013 and 2014. Data regarding hospitalizations (notably their number, length of stay, emergency department use, and main diagnosis) were collected from the French National Health Data System. The relationship between intervening hospitalizations and changes of frailty status over time was assessed with multivariate Markov models. Results: The prevalence of frailty was 14% in 2012 and 2013 and 17% in 2014. A total of 2715 changes in frailty status were observed from 2012 to 2014. At least one hospitalization was recorded for 1453 people (12%) between the 2012 and 2013 questionnaires, and 1472 (13%) between the 2013 and 2014 questionnaires. No association was found between intervening hospitalizations and changes of frailty status (aHR 1.14 [0.97-1.35] for robust to frail transition and aHR 0.89 [0.73-1.08] for frail to robust transition). However, repeated hospitalizations, hospitalizations after emergency department use, surgery and several diagnosis groups were significantly associated with transitions towards frailty or its recovery. Conclusion: Hospitalizations encompass a wide range of clinical situations, some of them being associated with incident frailty. An early recognition of these situations could help to better prevent and manage frailty in the early old age.