Cognitive tests for the detection of mild cognitive impairment (MCI), the prodromal stage of dementia: Meta‐analysis of diagnostic accuracy studies

Abstract

Introduction Mild cognitive impairment (MCI) is regarded as a prodrome to dementia. Various cognitive tests can help with diagnosis; meta‐analysis of diagnostic accuracy studies would assist clinicians in choosing optimal tests. Methods We searched online databases for “mild cognitive impairment” and “diagnosis” or “screening” from 01/01/1999 to 01/07/2017. Articles assessing the diagnostic accuracy of a cognitive test compared with standard diagnostic criteria were extracted. Risk of bias was assessed. Bivariate random‐effects meta‐analysis was used to evaluate sensitivity and specificity. Results Eight cognitive tests (ACE‐R, CERAD, CDT‐Sunderland, IQCODE, Memory Alteration Test, MMSE, MoCA, and Qmci) were considered for meta‐analysis. ACE‐R, CERAD, MoCA, and Qmci were found to have similar diagnostic accuracy, while the MMSE had lower sensitivity. Memory Alteration Test had the highest sensitivity and equivalent specificity to the other tests. Discussion Multiple cognitive tests have comparable diagnostic accuracy. The Memory Alteration Test is short and has the highest sensitivity. New cognitive tests for MCI diagnosis should not be compared with the MMSE.

Cognitive screening instruments to identify vascular cognitive impairment: A systematic review

“…Although its specificity in separating VaD from controls was high (93%‐97%), the MoCA's ability to correctly differentiate VMCI from controls was lower, varying from 75.6% to 79% . This is similar to other studies including a recent systematic review and meta‐analysis, which have shown that while sensitive, the specificity of the MoCA in identifying MCI syndromes is lower, particularly at its established cut‐off of less than 26 out of 30, including for post‐stroke cognitive impairment …”

“…The most accurate for identifying VaD from controls was the ACE‐R (AUC 0.99) when it was used in neurology clinics in Portugal but was only examined in a single study . The MMSE (AUC 0.86‐0.99) had good to excellent accuracy in differentiating VaD from controls but as expected, had lower accuracy in identifying VMCI, consistent with multiple studies showing its poor psychometric performance in identifying MCI . The BMET had excellent accuracy (AUC 0.94) in separating VMCI from controls but no data on its use in VaD were available.…”

Validation study of the Alzheimer's Disease Assessment Scale‐Cognitive Subscale for people with mild cognitive impairment and Alzheimer's disease in Chinese communities

“…Previous studies have shown that the QCSS‐E is an excellent diagnostic battery for both MCI and AD populations among the Chinese older people (sensitivity of 89.6% and specificity of 94.5% for MCI; sensitivity of 90.48% and specificity of 90.48% for AD) . Despite other scales like Montreal Cognitive Assessment (MoCA), Mini‐Mental State Examination (MMSE) are applied more frequently, the sensitivity or specificity of these scales appears to be slightly more moderate compared with that of the QCSS‐E, especially in identifying MCI (MMSE: sensitivity = 66.4%, specificity = 73.5%, MoCA: sensitivity = 81.2%, specificity = 74.0%) . Accordingly, the QSCC‐E was regarded as the superior method for assessing objective cognitive function of all subjects in the current study.…”

Dementia diagnostics in primary care : with a focus on cognitive testing

“…Meta-analysis of cognitive screening for MCI has reported AUC 0.664 for MMSE and AUC 0.811 for MoCA, which can be compared to the MMSE data in our population with an AUC 0.849 (136). This illustrates the challenge with finding consistent results when data originates from different populations.…”