Introduction and Objective: Multimorbidity is common in diabetes and may impact cognitive risk. We used latent class analysis to identify multimorbidity phenotypes and their association with incident dementia. Methods: We included 28,043 adults with diabetes. Twenty-two comorbidity indicators were used to identify multimorbidity classes. We fitted Cox models adjusted for sociodemographic and clinical variables. Results: The incidence of dementia was 4.28 per 1,000 patients/year. Compared with minimally comorbid diabetes, both the inflammatory-psycho-metabolic and the cancer-cardiometabolic high-burden phenotypes showed increased dementia risk, with aHR of 1.69 (95% CI 1.36-2.09) and 1.92 (95% CI 1.59-2.30), respectively. Conclusion: Multimorbidity patterns identify high-risk dementia subgroups and may aid stratification beyond single-disease measures. Disclosure J. Calvo-Marin: Speaker's Bureau; Current; Novo Nordisk, AstraZeneca, Abbott Diabetes, Roche Diabetes Care. F. Ruiz Salazar: Speaker's Bureau; Current; Abbott Diagnostics, AstraZeneca, Novo Nordisk A/S. Advisory Panel; Current; Roche Diagnostics. V. Mora-Gomez: None. M. Lutz: None. G. Torrealba-Acosta: None.
Calvo-Marín et al. (Fri,) studied this question.