The age-associated memory impairment construct revisited. Comments and recommendations of a french-speaking workgroup

This article reports the comments and recommendations of a French-speaking workgroup concerning the controversial ‘age-associated memory ilmpairment’ (AAMI) construct, proposed by Crook et al. (1986) to describe the memory difficulties associated with ageing. This construct's relevance and validity seem doubtful, and our workgroup met to discuss (i) the proposed causal link between age-associated memory changes and biological cerebral ageing, (ii) the psychometric criteria which could improve objective evaluation of age-related memory impairment (the initial AAMI definition criteria being inadequate), (iii) the problems associated with, and the clinical realities and implications of, ‘memory lcomplaint’ in the elderly, (iv) how to improve definition and evaluation of the psychoaffective factors contributing to a decrease in memory performance, and (v) the specificity (or lack thereof) of this construct. The following conclusions were reached: (i) no definite link between age-associated memory changes and biological cerebral ageing has been demonstrated in either humans or animals, and therefore remains a hypothesis; (ii) objective evaluation of age-related memory impairment could be improved by comparing subjects with both more appropriately defined, education-matched young subjects (age: 25–34) and education-matched subjects of the same age. No agreement was reached concerning the validity of existing global tests, or concerning which and how many of them should be used to detect AAMI; however, both verbal and non-verbal tests should be employed and more specific memory tests with adequate validity need to be developed. Specific tests were proposed to improve detection of decreased memory performance; (iii) subjective memory complaints in the elderly are not exclusively dependent on decreased memory performance and have multiple and complex determinants—the role of certain psychoaffective factors, such as anxiety, has been relatively underestimated; (iv) improved detection of the many factors contributing to decreased memory performance could be achieved by better patient screening, and (v) AAMI cannot currently be considered a specific disease entity. Should the AAMI construct be used to select patients for memory-enhancer drug trials, our workgroup proposed classifying elderly subjects into five groups according to the presence or absence of memory complaint, and memory performance compared with education-matched young and education- and age-matched subjects: (i) normal elderly subjects, (ii) subjects with purely subjective memory complaint (no objective impairment), (iii) subjects with memory complaint and objective impairment compared with young but not with age-matched subjects (score between one standard deviation above and below the mean of age- and education-matched controls, ie age-consistent memory impairment), (iv) subjects with memory complaint and objective impairment compared with age-matched controls (score between one and two standard deviations below the mean of age- and education-matched controls, ie late life forgetfulness), and (v) subjects with memory test performance below two standard deviations below the mean of their age-and education-matched controls, in whom organic pathology is likely in the absence of major psychoaffective disturbance.