Comparative methods for determining MoCA, MMSE, and CDR cut-offs in Alzheimer's disease: insights from the ADNI study

Accurate cut-off values for the Montreal Cognitive Assessment (MoCA), Mini-Mental State Examination (MMSE), and Clinical Dementia Rating (CDR) are essential for detecting and staging Alzheimer's disease-related cognitive impairment. Prior studies often relied on a single statistical method, potentially misaligning thresholds with clinical priorities.

We analyzed cognitive assessments from 883 Alzheimer's Disease Neuroimaging Initiative participants classified as cognitively normal, mild cognitive impairment (MCI), or clinical impairment (CI). Optimal thresholds were identified using Euclidean distance, Index of Union, maximum product of sensitivity and specificity, and Youden's J Index. Sensitivity, specificity, and area under the curve were calculated.

For MCI, MoCA cut-offs were 24 (three methods) and 25 (Youden's J Index). For CI, MoCA cut-offs were 17 (Youden's J Index) and 18 (others). MMSE (28, 24) and CDR (0.5, 1) cut-offs were consistent.

MoCA thresholds vary by method, requiring context-specific selection, while MMSE and CDR cut-offs remain stable.