A simple Mini‐Mental State Exam subscale formula to help differentiate Alzheimer's disease from dementia with Lewy bodies

Abstract

AbstractBackgroundAlthough there are clinical criteria for the diagnosis of Alzheimer's disease (AD) and dementia with Lewy bodies (DLB), the two conditions may sometimes be difficult to differentiate. Prior research has documented that the two conditions often have striking neuropsychological differences, and there have been a number of reports promoting the ability of simple screening tests to help the diagnostic accuracy of the clinician. These screening tests, however, have been limited in their specificity. Building upon previous research from our group and others, we now report a simple manipulation of the Mini‐Mental State Exam (MMSE) subscale scores that greatly improves the specificity of the screening test.MethodWe studied all AD and DLB patients who visited our dementia clinic during an 18‐month period, including both new and follow‐up visit patients. All who had an MMSE administered to them were included. All MMSE scores were included. The first MMSE recorded for each patient was studied.ResultWe studied the MMSEs from 136 AD and 24 DLB patients. The mean patient age was 77.8 ± 11.1 for AD, 78.3 ± 8.7 for DLB. Mean MMSE score was 20.0 ± 6.0 for AD, 21.8 ± 5.1 for DLB; MMSE range 1‐30 for AD, 10‐28 for DLB. Many different formulae were studied using the MMSE subscales of Memory (3 points max), Attention (5 points max), and Pentagon‐copying (1 point max) to determine which was the best to differentiate AD from DLB. The simple formula of the Pentagon score minus the Memory score (P‐M, range 1 to ‐3) provided the best discrimination between the AD and DLB patient groups. For P‐M scores = 1, the specificity of the equation to identify AD was 0.92. For P‐M scores < 0, the specificity to identify DLB was 0.85. The sensitivities for the P‐M equation were 0.43 for AD (scores = 1) and 0.46 for DLB (scores < 0).ConclusionThe simple MMSE subscale formula of Pentagon score minus Memory score (P‐M) provided good specificities to identify AD patients (0.92) and LBD patients (0.85) within our cohort of AD and DLB patients. A weakness of the P‐M equation was relatively weak sensitivity.