A Multifactor Approach to Mild Cognitive Impairment

Taha Qarni, BHSc; Arash Salardini, MD


Semin Neurol. 2019;39(2):179-187. 

In This Article


MCI may be diagnosed in several ways:

1. Screening tests: The most widely available form of cognitive data in the clinic derive from screening tests such as the Montreal Cognitive Assessment (MoCA) and Folstein's Mini-Mental Status Examination (MMSE). The former was designed for the diagnosis of MCI by Ziad Nasreddein and colleagues at the Montreal Neurological Institute, and has better psychometric properties compared with the MMSE.[12] It tests frontal functions somewhat better and is thus more applicable to non-Alzheimer's forms of MCI.[13] The MoCA comes in several different languages and has alternate forms to allow for repeat application without having to account for the patient learning the test. The paper forms are available online without cost (www.MoCAtest.org). However, when applying the MoCA, the following caveats have to be considered:

  1. The MoCA test is not just language specific but is also culture specific, and different cut-offs may need to be used for the different versions. For example, the Cantonese version of the test has different cut-offs compared with the English one.[14]

  2. The MoCA test is not education-normalized, except for one point awarded for education of 12 years or less. It is therefore likely to be less accurate for extremes of educational attainment (i.e., both low and advanced educational levels) of individuals tested. Attempts to educationally normalize MoCA have not succeeded in improving specificity without concurrently sacrificing sensitivity.[15]

  3. The specificity and sensitivity of the test is determined by the chosen cut-off. The official cut-off score of 25 or 26 has high sensitivity (80–100%) but low specificity (50–75%) in detecting amnestic MCI. A cut-off of 24/25 has a sensitivity and specificity of 80 and 81%, respectively.[14] The bottom line is that a score of 25 may be thought of as equivocal for the diagnosis of MCI.

  4. Assessing someone's cognition using MoCA depends not only on how well they complete the tasks, but also how they are able (or unable) to accomplish them. For example, it is common when administering the MoCA in our clinic for the individual to absentmindedly identify the rhinoceros as hippopotamus. Similarly, it is much easier for someone who already knows the "trick" of drawing a cube to draw it accurately compared with someone who is figuring out how to do it for the first time. People with anxiety or poor working memory may have greater problems with sentence repetition, and those with lower education may have difficulty with the second abstraction item. Retired individuals who are mostly housebound are less likely to know the exact date, and patients who have stopped driving and are transported by agencies or relatives are less likely to know where they are. A score of 25, where the deducted 5 points all come from the verbal learning task, is much more likely to be real compared with a test which has minor deficits spread across multiple domains.

  5. The MMSE and Dementia Rating Scale have greater limitations regarding the diagnosis of MCI.[16,17] For example, MMSE has few tests of executive and frontal domains, and is best for picking out amnestic MCI and less accurate or reliable for other subtypes of MCI.

2. Neuropsychological testing: There are several different criteria used for the diagnosis of MCI. These are less standardized in Parkinson's dementia and vascular cognitive impairment (VCI). Here, we use a nomenclature used by Jak et al:[18]

  1. The typical criteria for the diagnosis of MCI require one or more cognitive domains to fall 1.5 standard deviation (SD) or more below age- and education-normalized averages.

  2. The liberal criteria use a cut-off of 1 SD below normalized averages.

  3. The conservative criteria require greater than one domain to fall below the 1.5 SD cut-off.

In the clinic, both screening tests and neuropsychological tests are used as adjuncts in the diagnosis of MCI, with the clinical impression and the patient's functional status being the more important parameters used in making the diagnosis.

3. Functional status: Functional status is judged by the ability to perform basic and instrumental activities of daily living. Basic activities of daily living, such as personal hygiene, dressing, eating, toileting, etc., are affected in moderate to severe dementia. Instrumental activities of daily living, such as the ability to balance the checkbook, shop for oneself, use transport independently, prepare a meal, use the phone, etc., are relatively preserved in MCI but are affected in mild dementia.

A formalized way of combining activities of daily living and clinical impression is the CDR. The CDR is a five-point scale that characterizes six functional and cognitive domains. A score of 0.5 is consistent with a diagnosis of MCI. Further information and training is available at https://knightadrc.wustl.edu/cdr/cdr.htm.