RESUMO
OBJECTIVE: To evaluate how much the Addenbrooke's Cognitive Examination-revised (ACE-R) improves the estimate of cognitive ability from the Mini-Mental State Examination (MMSE) in people with Alzheimer disease (AD). METHODS: We examined itemized data in people with AD who were on the Scottish Dementia Research Interest Register drawn from eight centres across Scotland, covering 75% of the Scottish population. ACE-R items that comprise the MMSE and those that did not (non-MMSE items) were summed separately. We residualized MMSE total on non-MMSE total and vice versa to derive a measure of the variance unique to each. RESULTS: Five hundred and one (258 male, 243 female) participants, mean age 75.7 (range 52-94) years were on the register, of whom 329 (160 men, 169 women) had AD. Of those with AD, 309 had a mean MMSE of 20.5 and mean ACE-R of 57.5 measured with Pearson r = 0.92 between MMSE and ACE-R totals, and the regression equation ACE-R score = 3.0 × MMSE - 4.1. The unique non-MMSE items score correlated with ACE-R total r = 0.40 (16% of ACE-R variance). CONCLUSIONS: The ACE-R and MMSE total scores are highly correlated. In this clinical sample of people with established AD, for an MMSE score of 24, the predicted ACE-R score was 67.9 with 95% confidence intervals of 61.6-75.4. The extra non-MMSE ACE-R items improve estimates of cognitive ability by 16%.
