Three-year follow-up of Mini-Mental State Examination score in community residents aged 85 and over.

The objective of this study was to describe over time the course of cognitive function of elderly without cognitive impairment and of elderly with different stages of impairment, and to assess if the change in cognitive function was dependent on the initial level of function. The Mini-Mental State Examination (MMSE) was used at two time points. The first assessment (MMSE-1) was part of a community-based study and was obtained from 871 subjects. For the second assessment (MMSE-2) a sample of 166 subjects was drawn from the subjects alive at follow-up who had an MMSE-1 score. This sample was stratified by MMSE-1 score to avoid oversampling of subjects with high MMSE-1 scores. A second MMSE score was obtained from 134 elderly, whereas 18 subjects refused participation and 14 subjects were not traceable. The median age at first assessment was 89 years (25th percentile 87, 75th percentile 92), the mean follow-up period (S.D.) was 3.3 (0.5) years. The median change in MMSE score was minus 4 points (95% confidence interval (CI) -7 to -2) and the slope of the regression line of MMSE-2 on MMSE-1 was 1.1 (95% CI 0.9-1.3). It is likely that the slope was underestimated due to a floor effect, regression to the mean and missing observations. However, the probability of decline decreased if MMSE-1 was higher. Nevertheless, the probability ranged from 27 to 59% for subjects with the highest MMSE-1 scores aged 85 and 95 years respectively.(ABSTRACT TRUNCATED AT 250 WORDS)