As physician screening for dementia becomes more widespread, an increasing number of elderly patients will be identified who do not meet the diagnostic criteria for dementia but nonetheless have significant memory impairment. These patients, who have received a variety of diagnoses over the years, may have a number of different conditions. The most benign among these are “age-consistent cognitive decline” and “age-associated memory impairment.” These ailments are characterized by patient complaints of failing memory but cognitive test performance that is within normal limits for their age (although typically worse than average for younger people). Such patients should be screened for depression, because memory complaints that exceed memory impairment are common among depressed patients.^1,2 Most patients without depression respond well to reassurance and the prescription of potentially neuroprotective health measures, such as proper diet (including micronutrient intake), limited alcohol use, control of hypertension, etc.

A condition that has come to be known as “mild cognitive impairment” (MCI) is a greater concern. To be considered for a diagnosis of MCI, patients must not have severe or pervasive cognitive impairment and their score on the Mini-Mental State Examination (MMSE) must be 24 or higher. In addition, they must not have impaired ability to carry out daily activities, and their Clinical Dementia Ratings (CDRs) must be no higher than 0.5 (questionable dementia).³ However, these patients must have complaints of worsening memory (usually corroborated by family members), as well as objective memory impairment. The usual criterion for the latter is performance at a level that is at least 1.5 standard deviations below average for age and education on neuropsychologic tests of new learning and memory. In the absence of a known cause for the isolated memory impairment (e.g., stroke), the diagnosis of MCI should be made.

A recent study³ at Mayo Clinic, Rochester, Minn., found that patients with MCI performed like normal elderly persons on the MMSE and tests of general intelligence. However, they performed much worse than normal study subjects and only slightly better than patients with Alzheimer's disease on tests of verbal and spatial memory. During four years of follow-up, the average decline on MMSE was approximately 1 point per year in patients with MCI, compared with approximately 3 points per year in the patients with Alzheimer's disease and zero points per year in normal elderly persons.

Persons with MCI appear to be at increased risk for developing Alzheimer's disease. In the Mayo Clinic study,³ the annual incidence of Alzheimer's disease was 10 to 12 percent in the MCI group, compared with 1 to 2 percent in normal elderly persons. Thus, almost one half of all patients with MCI develop Alzheimer's disease within four years. Several other studies, using somewhat different diagnostic criteria and research methods, support this incidence rate.^4–6

A major focus of current research on MCI is determining which patients will progress to Alzheimer's disease and which will not. In the large prospective Canadian Study of Health and Aging,⁵ 59 percent of those who had “cognitive impairment, no dementia” at entry (essentially, MCI) were deceased five years later, and another 19 percent had developed a dementia syndrome. Three variables at entry were significant predictors of which patients would be demented at the five-year follow-up: higher age, informant's report of memory impairment and lower MMSE score. However, the best statistical algorithm that could be developed for the diagnosis of dementia had insufficient sensitivity and specificity to be of clinical use.

It remains unclear whether the pattern of cognitive deficits among patients with MCI can predict who will develop Alzheimer's disease. Bowen and colleagues⁴ found that none of the nine neuropsychologic tests administered to elderly patients with isolated memory loss predicted progression to dementia four years later. On the other hand, Tierney and co-workers⁶ found that a statistical model that included tests of mental tracking and delayed recall as well as an informant's report of cognitive decline was highly predictive of conversion from MCI to probable Alzheimer's disease within two years (area under the receiver operating characteristic curve = 0.955).

MCI patients with the ε4 allele of the apolipoprotein-E gene (ApoE) appear to be at especially high risk for converting to Alzheimer's disease. Tierney and colleagues⁷ studied 143 patients without dementia who had “memory problems that interfered with daily functioning.” All had scores of 24 or more on the MMSE, or 123 or higher on the Dementia Rating Scale.⁸ Of the 107 patients from whom DNA samples were available for analysis, 29 developed Alzheimer's disease within two years. ApoE genotype, together with age and education, had only 17 percent sensitivity and 95 percent specificity for development of Alzheimer's disease. However, when delayed recall and recognition performance were included in the statistical algorithm, the conversion from MCI to Alzheimer's disease could be predicted within 83 percent sensitivity and 96 percent specificity.

Among the very earliest sites of neurofibrillary pathology in Alzheimer's disease are the hippocampus and entorhinal cortex.^9,10 The resulting atrophy of the hippocampal formation can be detected and measured with magnetic resonance imaging (MRI) scans of the brain, even in very early Alzheimer's disease.^11,12 Recently, Jack and co-workers¹³ have shown that hippocampal volume on brain MRI is a significant predictor of subsequent conversion from MCI to Alzheimer's disease. Approximately 10 percent of patients with MCI who had standardized hippocampal volume scores in the normal range developed Alzheimer's disease within three years of their scans. This contrasts with approximately 25 percent of those with moderately atrophic hippocampi and approximately 50 percent of those with extremely small hippocampi.

The goals of treatment for MCI are twofold: (1) to improve or maximize functioning, and (2) to delay or prevent conversion to Alzheimer's disease. At present, most attempts at treatment involve environment modifications and “memory prostheses.” This approach is not so much an attempt to restore or improve memory as it is to compensate for lack of memory. Strategies include instruction and practice in the use of date books, calendars, reminder notes, compartmentalized pill boxes, electronic cuing devices, automatic telephone dialers and the like. To date, there is little empiric evidence for the success of these approaches in patients with MCI.

A variety of medications (including non-steroidal anti-inflammatory drugs) hormones (especially estrogen), vitamins (e.g., vitamin E) and herbal preparations (especially gingko biloba) have been advocated as treatments for memory loss in the elderly and preventers of Alzheimer's disease. While some of these agents hold promise, robust effects from carefully executed, well-controlled clinical trials are nonexistent. The three acetylcholinesterase inhibitors labeled for use in Alzheimer's disease—tacrine (Cognex), donepezil (Aricept) and rivastigmine (Exelon)—have not yet been systematically tested in patients with MCI.

The Alzheimer's Disease Cooperative Study, a program of the National Institute on Aging, is currently conducting a double-blind, placebo-controlled trial of donepezil and vitamin E (α-tocopherol) in preventing or delaying the onset of Alzheimer's disease in persons with MCI. Participants are being seen at more than 60 clinical research sites across the United States and Canada. More information about this study may be obtained by telephoning 888-455-0655 or visiting the following Web site:http://www.memorystudy.org.