Detecting Early Stage Alzheimers Disease in MCI and PreMCI The value of informants

Super Memory Formula

Dementia Reversal Diet

Get Instant Access

John C. Morris1 and Martha Storandt2

In the century since Alzheimer's seminal presentation of the original patient with Alzheimer's disease (AD; Alzheimer et al. 1987), emphasis now is being given to the recognition of early-stage AD in comparison to nondemented aging. As a consequence, there is intense interest in mild cognitive impairment (MCI). The concept of MCI was introduced to characterize older individuals with cognitive deficits that, although abnormal for age, fell short of overt dementia (Flicker et al. 1991). [Related terms such as "age-associated memory impairment" (Crook and Bartus 1986) and "age-associated cognitive decline" (Levy 1994) were proposed as variants of normal cognitive aging, and "cognitive impairment, no dementia" (Graham et al. 1997) overlaps MCI but may characterize a broader array of cognitive dysfunction.] The MCI construct was further developed by Petersen and colleagues to feature memory deficits (amnestic MCI); this group proposed MCI as a transitional state between the cognitive changes of normal aging and AD (Petersen et al. 1999; 2001a). The influential work of Petersen and coworkers resulted in the recognition of MCI as a pathological condition imparting increased risk for AD (Bennett et al. 2002), and it now is a target of treatment trials that aim to prevent the conversion of MCI to clinically diagnosed AD (Petersen et al. 2005).

The MCI construct has engendered controversy, however, because not all individuals with MCI are at increased risk for AD; some individuals meeting criteria for MCI do not progress to AD and some even revert to normal over time (Ritchie et al. 2001; Larrieu et al. 2002; Fisk et al. 2003; Ganguli et al. 2004). Thus MCI is a heterogeneous condition with both neurodegenerative and other etiologies (Ancelin et al. 2006). This heterogeneity contributes to the variability in published outcomes for MCI, particularly for the rates of progression to AD. Other factors contributing to this variability include differences in case ascertainment (e.g., population-based studies versus samples from specialty clinics), nonuniform diagnostic criteria for MCI, and different implementation strategies for the criteria. The criteria for MCI recently have been refined to address its heterogeneous etiologies and identify subtypes without memory deficits or with deficits in multiple cognitive domains (Winblad et al. 2004; Petersen 2004). The major interest in MCI, however, remains focused on its relevance as a prodromal state for AD.

1 Departments of Neurology, Pathology and Immunology, and the Program on Physical Therapy;

2 Departments of Psychology and Neurology; and the Alzheimer's Disease Research Center, Washington University School of Medicine, St. Louis, Missouri, USA

Address correspondence to: John C. Morris, MD, Friedman Distinguished Professor of Neurology, Alzheimer's Disease and Research Center, Washington University School of Medicine, 4488 Forest Park Avenue, Suite 101, St. Louis, Missouri 63108 USA

Both the original and revised criteria for MCI have inherent conceptual difficulties and problems in operationalization that notably limit the value of the MCI construct. The representation of MCI as a "transitional state between the cognitive changes of normal aging and the earliest clinical features of Alzheimer's disease" (Petersen et al. 2005) disregards the basic pathobiology of AD, wherein synaptic and neuronal degeneration occur along a continuum from minimal to extensive. The initial clinical correlates of this progressive neuronal deterioration include mild impairment of memory and other cognitive abilities that may only subtly affect everyday function. That is, the initial clinical manifestation of AD is MCI. The very mild deficits associated with the early symptoms of AD may be insufficient to reach the current clinical detection threshold for many physicians, but nonetheless these deficits represent the earliest clinical features of AD. Stated another way, conceptually MCI is not a risk factor for AD, it already is AD in its earliest symptomatic stage.

In addition to the flawed premise that MCI predates the early symptoms of AD, its characterization remains problematic. The MCI criteria (original and revised) and their application are insufficiently sensitive to distinguish nondemented aging from prodromal AD. Hence, MCI samples include a mixture of both normal and impaired individuals. Observed outcomes (e.g., rates of progression to clinically diagnosed AD) vary as a function of the proportion of nondemented individuals in the sample.

A key factor in the failure to discriminate normal aging from MCI is that MCI criteria characterize "impairment" by comparing an individual's cognitive test performance with normative values (Petersen et al. 2001b; Winblad et al. 2004). Using neuropsychological criteria to define MCI poses methodological problems because there is no consensus on the number and types of cognitive measures needed to characterize each cognitive domain, the quality of the reference normative values may vary, and factors such as education, race, ethnicity, and culture affect neuropsychological test performance (Manly et al. 2005; Teng and Manly 2005). Another problem is the age-related variability in cognitive test performances for nondemented elderly (Morse

1993). This variability contributes to an extensive degree of overlap of performances of normal and MCI groups such that absolute levels of performance on psychometric tests do not reliably distinguish these conditions (Storandt and Hill 1989; Sliwinski et al. 1996).

The criteria for MCI require a comparison of an individual's cognitive performance at a point in time to the performance of (presumably) demographically similar individuals. This interindividual comparison does not indicate whether cognitive function has declined for that individual; it simply places the individual's performance in relation to that of others. Performance below arbitrarily defined cut-points (e.g., 1.5 standard deviations below normative values) determines "impairment." This approach is inconsistent with the definition of dementia, which requires cognitive decline in relation to the individual's previously attained levels (American Psychiatric Association

1994). The clinically relevant information needed to determine that an individual is demented is not how the individual performs in relation to other people but rather that the individual has experienced cognitive loss relative to his/her prior abilities.

Detecting intraindividual change can be accomplished in two ways. One is serial neuropsychological testing. Confounding factors associated with cognitive tests, including statistical issues related to the reliability of the measures (signal-to-noise ratio) and nonlinear patterns of progression, require an observation period of greater than 1 year to infer a change in cognitive status (van Belle et al. 1990; Morris et al. 1993). This duration, of course, is impractical for clinical practice and also for many research programs in which diagnostic classifications are made at time of enrollment.

The other method of detecting cognitive loss involves interviews with someone who knows the individual well (typically the spouse, adult child, other relative, or close friend). The observations of an informant or collateral source permit the individual's current cognitive and functional abilities to be judged in relation to the individual's previous status (Table 1). This longitudinal perspective uses the individual as his/her own control in assessing whether cognitive decline has occurred. Moreover, informant-based observations are "face valid" as they relate to the everyday performance of the individual and are sensitive to even very mild impairment (i.e., avoid the "ceiling" effects of cognitive tests). Informant observations also are unaffected by scale restriction ("floor" effects) and minimize demographic and cultural factors. Most important, informant reports of cognitive change are accurate (Cacchione et al. 2003; Harvey et al. 2005) and have been shown to be highly sensitive to even very mild cognitive decline (McGlone et al. 1990; Morris et al. 1991; Koss et al. 1993; Jorm 1997; Jorm et al. 2000; Ready et al. 2004; Galvin et al. 2005).

The value of informant observations has long been recognized; Alzheimer's case history of the original patient with AD begins with an interview of her husband (Maurer et al. 2000). Standard dementia assessment protocols such as the Dementia Scale (Blessed et al. 1968), the Global Deterioration Scale (Reisberg et al. 1982), the Clinical Dementia Rating (CDR; Hughes et al. 1982; Morris 1993), and the Cambridge Examination for Mental Disorders of the Elderly (Roth et al. 1986) include semistructured informant interviews. Surprisingly, neither the original (Petersen et al. 1999) nor the revised (Winblad et al. 2004) criteria for MCI mandate informant observations; self-reported memory complaints alone are sufficient for a MCI designation. The subjective complaints of memory impairment in normal elderly, however, neither correspond with actual cognitive function nor predict future dementia (Flicker et al. 1993; Jorm et al. 1997; Carr et al. 2000), and the subjective reports of cognitively impaired individuals of course are unreliable (Ganguli et al. 2006). The reliance on interindividual comparison of neuropsychological test performance as a criterion for MCI and the

Table 1. Advantages and disadvantages of informants in dementia diagnosis Advantages

Assess change (longitudinal perspective)

Observations about cognitive abilities are relevant to everyday function (face validity) Culturally fair

Absence of ceiling and floor effects No practice effects

Accurate and sensitive for dementia in its earliest stages


Observant and reliable informant not always available Time for informant interview

Some cultures discourage reporting "negative" information (e.g., dementia symptoms) in elders failure to require informant observations regarding the cognitive and functional status of the individual result in the inclusion of "false positive" and "false negative" cases in MCI samples. Such cases include nonimpaired individuals who selfreport memory complaints or perform poorly on neuropsychological tests and individuals who are experiencing meaningful cognitive decline but still perform above the cutoff values on neuropsychological tests. The current conceptual and methodological approach to MCI thus ensures its heterogeneity and limits its clinical and research utility.

The principle of intraindividual change permits the recognition of the earliest symptomatic stages of AD in MCI individuals (Morris et al. 2001; Storandt et al. 2002), even in individuals who are not yet sufficiently impaired in their neuropsychological test performance to meet criteria for MCI ("preMCI"). We recently reported (Storandt et al. 2006) findings from 388 individuals clinically diagnosed with AD at the CDR 0.5 stage (identical to the CDR stage used for the characterization of MCI individuals for a multicenter trial of donepezil and vitamin E; Grundman et al. 2004). Thirty-two of the 388 individuals met neuropsychological criteria for amnestic MCI, 90 met revised criteria for MCI, and 276 were too minimally impaired cognitively to meet either set of MCI criteria (preMCI). The diagnosis of AD was validated by progression to a CDR 1 or greater stage of dementia with a median survival of about four years for the amnestic and revised MCI groups, comparable to reported rates of progression of 12%-15% per year for MCI to clinically probable AD (Petersen et al. 1999). The preMCI group also progressed to CDR 1 or greater stage of dementia but, because these individuals were less cognitively impaired at baseline, the median survival was about eight years. Additional validation was provided by the neuropathological diagnosis of AD in nine of nine amnestic MCI, 18 of 20 revised MCI, and 43 of 47 preMCI individuals who came to autopsy; overall, AD was confirmed in 92% of the autopsied sample. This neuropatho-logical perspective provides additional evidence that the initial cognitive symptoms that are characterized as MCI in reality represent AD (Markesbery et al. 2006).

Our sample is not unique (Table 2). It is comparable to other reported MCI samples on demographic features, apoE genotype, level of impairment as measured by the MMSE and CDR-SumBoxes, and rate of progression to greater dementia severity. What is different is our clinical detection method that utilizes an informant interview in addition to an examination of the individual. Although informant-based methods may not be appropriate in all settings, an informant interview that can be completed in less than three minutes (Table 3) provides a positive predictive value of 87% in discriminating nondemented aging from dementia at the CDR 0.5 or greater stage (Galvin et al. 2005).

Not all individuals meeting criteria for MCI have AD (or another dementing disorder). The subset of MCI that eventually is recognized as clinicallyprobable AD, however, can be distinguished from non-AD causes of MCI using usual clinical methods, because the phenotype of the subset of MCI caused by AD is identical to that of more overt AD, only milder. The initial deviations from an individual's own baseline that represent cognitive decline caused by AD may not place that individual below some value on a scale determined by group norms. AD can be detected in MCI and even preMCI if the diagnosis is based on intraindividual, not interindividual, comparisons. The sensitivity and accuracy of clinical detection methods suggest that it is time to move beyond the MCI concept in favor of etiologically based classifications, most notably early-stage AD.

Table 2. Demographic and clinical features of MCI and PreMCI samples

Washington University

Mayo Clinic


(Storandt et al. 2006)

(Petersen et al. 1999)

Trial (Grundman et al. 2004)










N = 32

N = 90

n = 276

N = 76

N = 769


69.9 (6.2)

73.3 (8.9)

75.7 (8.9)

80.9 (1.0)

72.9 (7.3)



13.2 (3.0)

12.5 (3.4)

14.2 (3.0)

13.7 (0.4)

14.7 (3.1)









27.5 (2.2)


27.9 (1.6)

26.0 (0.3)

27.3 (1.9)


2.0 (0.9)

2.2 (1.1)

1.6 (1.0)

1.5 (0.2)

1.8 (0.8)

% apoE e4





Legend: Mean values are given +/- standard deviations. M = male, F = female. MMSE = MiniMental State Examination (Folstein et al. 1975), where the range of possible scores is from 30 (best) to 0 (worst). CDR-SB = Clinical Dementia Rating, SumBoxes, where the range of possible scores is from 0 (best) to 18 (worst). Carriers of the apolipoprotein E (apoE) £4 allele for the three samples (Washington University, Mayo Clinic, Multicenter Trial) are given in percentages.

Table 3. Eight-item informant interview to differentiate aging and dementia Report only a change caused by memory and thinking difficulties:

1. Is there repetition of questions, stories, or statements?

2. Are appointments forgotten?

3. Is there poor judgment (e.g., buys inappropriate items, poor driving decisions)?

4. Is there difficulty with financial affairs (e.g., paying bills, balancing checkbook)?

5. Is there difficulty in learning or operating appliances (e.g., television remote control, microwave oven)?

6. Is the correct month or year forgotten?

7. Is there decreased interest in hobbies and usual activities?

8. Is there overall a problem with thinking and/or memory?

(Adapted from Galvin et al. 2005).

Acknowledgements. The authors acknowledge the invaluable contributions of our participants and their families. Drs. Eugene Rubin, James E. Galvin, Gabriel DeErausquin, David Carr, Randall Bateman, Monique Williams, B. Joy Snider, and Consuelo Wilkins, nurse clinicians Mary Coats, Pamela Millsap, Stacy Schneider, Angela Oliver, Christy Tomlinson, and Pamela Jackson and the Clinical and Psychometric Cores provided careful assessments of the participants and are among the many investigators and staff at the Alzheimer's Disease Research Center (ADRC) who contributed importantly to the ideas expressed here. We also are grateful to the ADRC's founding director, Leonard Berg, MD, for continued friendship and support. Supported by National Institute on Aging grants P50 AG05681, P01 AG03991, and P01 AG026276.

William E. Klunk (left) and Chester A. Mathis (right)

Was this article helpful?

0 0
All About Alzheimers

All About Alzheimers

The comprehensive new ebook All About Alzheimers puts everything into perspective. Youll gain insight and awareness into the disease. Learn how to maintain the patients emotional health. Discover tactics you can use to deal with constant life changes. Find out how counselors can help, and when they should intervene. Learn safety precautions that can protect you, your family and your loved one. All About Alzheimers will truly empower you.

Get My Free Ebook

Post a comment