Mini‐Mental State Examination trajectories across dementia subtypes in the Swedish Registry for Cognitive/Dementia Disorders and U.S. National Alzheimer's Coordinating Center
Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring · 2026-10 · 18 (4)
DOI: 10.1002/dad2.70531 ↗Abstract
Abstract INTRODUCTION Mini‐Mental State Examination (MMSE) trajectories are widely used to describe cognitive decline in dementia, but estimates may be sensitive to cohort structure, follow‐up density, and modeling assumptions. We compared subtype‐specific MMSE trajectories across two complementary longitudinal data resources. METHODS The analyzed data were from the Swedish Registry for Cognitive/Dementia Disorders (SveDem) and the U.S. National Alzheimer's Coordinating Center (NACC). Participants were included with at least two valid cognitive assessments separated by ≥ 90 days. The final cohorts included 30,115 SveDem participants and 33,429 NACC participants. Subtype‐stratified regression‐spline generalized additive mixed models and linear mixed‐effects models were fitted with adjustment for demographic and clinical covariates. RESULTS Compared with NACC, SveDem participants were older (median age 79 vs. 72 years), had lower baseline MMSE scores (22.3 ± 4.2 vs. 26.5 ± 4.6), and had shorter follow‐up (median 1.9 vs. 3.2 years; mean cognitive assessments 2.9 vs. 4.5). Shared dementia subtypes showed broadly similar relative decline patterns across cohorts. In regression‐spline models, Alzheimer's disease showed faster estimated annualized decline than vascular dementia in both SveDem (−0.79 vs. −0.49 MMSE points/year) and NACC (−1.22 vs. −0.24 points/year). Lewy body dementias and frontotemporal dementia showed intermediate but less consistent patterns. Absolute decline estimates varied by model specification, particularly where later follow‐up was sparse. DISCUSSION Relative subtype‐specific MMSE decline patterns were broadly reproducible, whereas absolute long‐horizon estimates required cautious interpretation. These cross‐cohort reference patterns may support future dementia epidemiology studies using MMSE trajectories to compare subtypes, examine clinical factors or evaluate treatment‐related associations.
EUVIMED