The Challenge of Characterizing Normal Brain Aging in Relation to Alzheimer’s Disease

Abstract

IDENTIFYING the earliest stages of Alzheimer’s disease (AD) is receiving new impetus with the advent of approved drugs for symptomatic therapy and with promise for the development of interventions that may slow disease progression or even delay its onset. From a clinical perspective, cognitive states that previously were considered part of the spectrum of normal aging increasingly are viewed as possible prodromal stages of AD (3). These conditions, variously described as “age-associated memory impairment,” “cognitive impairment, no dementia,” and “mild cognitive impairment,” are being studied to elucidate factors that predict either cognitive stability (5) or development of overt dementia (11). One such factor is a history of diminished cognitive function in usual activities, as elicited by informant-based clinical methods that are sensitive to the detection of very early stages of AD (7,9), including high-functioning individuals who still are capable of normal performance on mental status tests and psychometric measures. Detection of individuals with very mild dementia has implications not only for understanding the initial clinical manifestations of AD but also for concepts about normal aging. For example, because very mildly impaired persons almost certainly contaminate samples of putatively nondemented older adults, the effect of age on cognitive performance is overestimated (10,15, 16). Exclusion of persons with mild cognitive impairment from normative samples is necessary to permit the true abilities of successfully aging individuals to be determined and to rethink assumptions about “age-associated” cognitive decline. Successful aging also is important to study from a neuropathological perspective. It has been generally accepted that the distinctions between AD are quantitative rather than qualitative. Neuropathological criteria for the diagnosis of AD are based on ageadjusted densities of neocortical amyloid plaques (6,8), reflecting numerous reports that the presence of plaques increases with age in apparently nondemented individuals. However, the finding of numerous neocortical plaques with normal aging is suspect because few of these studies used measures to exclude the possibility of very mild cognitive impairment. The presence of plaques in apparently nondemented individuals may instead indicate unrecognized early stage AD. My colleagues (Leonard Berg, Daniel McKeel, J. Philip Miller, Joseph Price, Eugene Rubin, and Martha Storandt, among others) and I find that when clinical measures are used to ensure that only brains from clearly cognitively normal persons are studied, neocortical plaques are absent or only negligibly present in patients up to 90 years of age (10). The same observation holds for the presence of neocortical neurofibrillary tangles, which essentially are absent in truly nondemented aging. These findings challenge the notion that aging and AD form a neuropathological continuum. A different situation is found in the medial temporal lobe. The elegant study by Braak and Braak in this issue confirms that, after about 70 years of age, very few brains fail to demonstrate at least some neurofibrillary change in the transentorhinal region. By age 90 years, all 70 patients in their series displayed these changes. In contrast, a substantial number of elderly persons in the same series did not develop plaques; even in the group of persons 90 years or older, 15 of the 70 brains still were devoid of plaques. The results from Braak and Braak correspond with data from our case material as studied by Joseph L. Price (12‐14). These data indicate that plaques are not an inevitable part of aging (at least into the late 80’s), whereas tangles occur to at least some extent in virtually all people older than 55 years (demented and nondemented) but are limited to entorhinal cortex, hippocampus, and other vulnerable medial temporal lobe structures. [Differences in methodology, especially the examination of serial sections through the hippocampus and parahippocampal gyrus with both silver and immunohistochemical staining methods, may explain, in part, the earlier age at which we and others (1) detect the uniform presence of tangles as compared with the older age in the