Hypertension, diabetes and depression as modifiable risk factors for dementia: a common data model approach in a multicenter population-based cohort

Background Dementia represents a major public health challenge, with age as its primary non-modifiable risk factor. Several modifiable conditions, including hypertension, diabetes and depression, have been identified as potential targets for prevention. Objective To evaluate the impact of hypertension, diabetes, depression and their interactions on the onset of dementia within the Italian National Health Service, using a common data model approach. Methods This population-based cohort study, part of the PREV-ITA-DEM project, was conducted across five Italian regions and cities participating in the NeuroEpiNet network. Individuals aged ≥ 50 years without prior diagnoses of dementia, depression, diabetes or hypertension were followed from cohort entry (2011–2013) until dementia onset, death, emigration or study end (2019–2022). Exposures were treated as time-dependent and identified through validated algorithms applied to health administrative databases. Associations were estimated using competing risks regression models adjusted for confounders, with mortality treated as the competing event. A methodological framework based on a common data model ensured consistency across sites through a shared protocol, standardized local databases and uniform analytic scripts. Site-specific results were pooled using meta-analytic techniques. Results The final cohort included over 3 million individuals (Bologna and Trapani cities; Lazio, Piemonte and Toscana regions). Pooled estimates showed an association between the risk factors and dementia onset with sub-hazard ratio of 4.54 (95%CI 3.35-6.16) for depression, 1.32 (95%CI 1.09-1.61) for diabetes and 1.11 (95%CI 0.99-1.24) for hypertension. No significant interaction were observed between exposures. Sensitivity analysis with a 3 year of lag between exposure and outcome confirmed the association with attenuated effect sizes. Conclusion Depression showed the strongest association with dementia onset, followed by diabetes and hypertension. These findings support the use of common data models for multi-regional analyses and highlights the importance of early interventions on psychiatric and metabolic risk factors.