Aims: To examine whether age at type 2 diabetes onset is an independent predictor of dementia risk.
Methods: Retrospective cohort drawn from healthcare administrative records of all inhabitants within Romagna's catchment area, Italy, with an estimated onset of type 2 diabetes in 2008-2017 and aged ≥ 55, with follow-up until 2020. Time to dementia or censoring was estimated with the Kaplan-Meier method, using diabetes onset as the time origin.