Monday, October 16, 2017 -- Objective: To determine if serum magnesium levels are associated with the risk of all-cause dementia and Alzheimer disease. Methods: Within the prospective population-based Rotterdam Study, we measured serum magnesium levels in 9,569 participants, free from dementia at baseline (1997–2008). Participants were subsequently followed up for incident dementia, determined according to the DSM-III-R criteria, until January 1, 2015. We used Cox proportional hazard regression models to associate quintiles of serum magnesium with incident all-cause dementia. We used the third quintile as a reference group and adjusted for age, sex, Rotterdam Study cohort, educational level, cardiovascular risk factors, kidney function, comorbidities, other electrolytes, and diuretic use. Results: Our study population had a mean age of 64.9 years and 56.6% were women.