SEARCH

SEARCH BY CITATION

Keywords:

  • cognitive impairment;
  • factor analysis;
  • regression;
  • dementia

OBJECTIVES: To explore the associative structure between a screening test for dementia, the Mini-Mental State Examination (MMSE), and a neuropsychological battery for the detection of dementia, the Mental Deterioration Battery (MDB).

DESIGN: A retrospective analysis.

SETTING: Psychology unit of a general hospital in Rome, Italy.

PARTICIPANTS: Three hundred consecutive outpatients and inpatients referred to our hospital on the basis of suspected cognitive impairment and evaluated between January 1999 and March 2000.

MEASUREMENTS: MMSE and MDB.

RESULTS: Of the 300 subjects evaluated by the MMSE score, 142 (47.3%) were considered to be cognitively healthy, and 116 (38.7%) were mildly and 42 (14.0%) moderately impaired. Factor analysis of MDB extracted three factors able to account for 75% of the total variance: a visuospatial factor, verbal memory ability, and a language skill. Using MMSE as an independent variable, a linear regression model could account for the visuospatial and language factors and a cubic regression model for the verbal memory factor. Within the normal MMSE boundaries (24–30), a dramatic decrease of verbal memory could be documented, whereas the slope is less steep in the mild impairment group (16–23).

CONCLUSIONS: Our findings indicate the presence of a warning range within the normal MMSE interval. Thus, the traditional MMSE cutoff values may not be appropriate in detecting early phases of dementia. When patients score about 27 on MMSE, it should be of interest to check whether they fail only on long-term memory tests, because this could be a first signal of a preclinical condition heralding clear dementia (e.g., mild cognitive impairment).