Brief Cognitive Tests Used in Primary Care Cannot Accurately Differentiate Mild Cognitive Impairment from Subjective Cognitive Decline

Abstract

Background: Differentiating mild cognitive impairment (MCI) from subjective cognitive decline (SCD) is important because of the higher progression rate to dementia for MCI and when considering future disease-modifying drugs that will have treatment indications at the MCI stage.

Objective: We examined if the two most widely-used cognitive tests, the Mini-Mental State Examination (MMSE) and clock-drawing test (CDT), and a test of attention/executive function (AQT) accurately can differentiate MCI from SCD.

Methods: We included 466 consecutively recruited non-demented patients with cognitive complaints from the BioFINDER study who had been referred to memory clinics, predominantly from primary care. They were classified as MCI (n = 258) or SCD (n = 208) after thorough neuropsychological assessments. The accuracy of MMSE, CDT, and AQT for identifying MCI was examined both in training and validation samples and in the whole population.

Results: As a single test, MMSE had the highest accuracy (sensitivity 73%, specificity 60%). The best combination of two tests was MMSE < 27 points or AQT > 91 seconds (sensitivity 56%, specificity 78%), but in logistic regression models, their AUC (0.76) was not significantly better than MMSE alone (AUC 0.75). CDT and AQT performed significantly worse (AUC 0.71; p < 0.001-0.05); otherwise no differences were seen between any combination of two or three tests.

Conclusion: Neither single nor combinations of tests could differentiate MCI from SCD with adequately high accuracy. There is a great need to further develop, validate, and implement accurate screening-tests for primary care to improve accurate identification of MCI among individuals that seek medical care due to cognitive symptoms.

Keywords: AQT; Mini-Mental State Examination; clock drawing test; cognitive screening; diagnostic accuracy; mild cognitive impairment; subjective cognitive decline.

Fig.1

Enrollment process. The eligible population was defined as those who met the inclusion criteria (see Methods). Participation in other studies was not an exclusion criterion, but because of multiple investigations at the same time, some patients could not be included in the present study. “Any missing data” was defined as missing data for MMSE, CDT, AQT, SCD/MCI classification, or years of education.

Fig.2

ROC curves from logistic regression models for identifying MCI. A shows the ROC curves of the individual cognitive tests adjusted for age and education in logistic regression models with classification (SCD or MCI) as outcome variable. B shows the different combinations of tests, also adjusted for age and education. See Table 3 for more details. AUC, Area under the Curve; MMSE, Mini-Mental State Examination; AQT, A Quick Test of Cognitive Speed; CDT, Clock Drawing Test.