Annals of Internal Medicine | 2019
Scam Awareness Related to Incident Alzheimer Dementia and Mild Cognitive Impairment
Abstract
The persistent difficulty of determining who is at high risk for adverse cognitive outcomes in old age has led to intense interest in identifying behavioral changes that predict the onset of Alzheimer dementia and its precursor, mild cognitive impairment (MCI), earlier than standard neuropsychological measures currently do (1, 2). Early detection of persons at risk for MCI in particular might enable better targeting of potential disease-modifying therapies. Despite awareness of the need to identify predictors of dementia and MCI, the aspects of behavior to target are unclear. Evidence suggests that impairment in complex behaviors (such as decision making, particularly financial) is associated with progression from MCI to Alzheimer dementia, but little is known about their association with the transition from normality to MCI (37). Scam awareness is a key component of decision making in old age. Fraudsters frequently target and victimize older persons, and an awareness of deceptive tactics and behaviors that increase susceptibility to scams or exploitation is essential for sound decision making (813). Emerging findings indicate that older persons, particularly those with MCI or other cognitive syndromes, are vulnerable to decreased scam awareness (8, 14). Further, even subtle changes in cognition related to agethose observed among cognitively intact personshave a deleterious effect on scam awareness (15). Together, these findings suggest that decreased scam awareness may be an early manifestation of pathologic cognitive aging and a harbinger of adverse cognitive outcomes. Here, we test the hypothesis that low scam awareness is associated with increased risk for incident MCI and Alzheimer dementia using data from more than 900 participants in a community-based study of aging (16). A subset of participants who died provided autopsy data to further explore whether low scam awareness in old age is a consequence of accumulating Alzheimer disease (AD) pathology in the brain. Methods Participants Data came from an ongoing clinical pathologic study of aging (16), the Rush Memory and Aging Project. The study was approved by the Institutional Review Board of Rush University Medical Center. Enrollees have no known dementia, receive annual clinical evaluations, and agree to organ donation. Each participant provided written informed consent and a document of anatomical gift. The Memory and Aging Project started in 1997, and assessment of scam awareness was introduced in 2010. At the time of these analyses, 2031 participants had enrolled and completed the baseline evaluation. Figure 1 details the assembly of the final analytic sample of 935 older persons free of dementia. Figure 1. Flow chart showing the assembly of the analytic cohort. Assessment of Scam Awareness Scam awareness was assessed using a measure that addresses knowledge of tactics used to deceive older persons and willingness to engage in behaviors that may increase risk for falling prey to financial scams and other forms of exploitation (14, 15, 1719). We focused on these behaviors because exploitation of elders is a growing public health problem (813, 20). Participants were asked to rate their agreement on a 7-point Likert scale (Appendix Table) with 5 statements that assess openness to sales pitches (items 1, 2, and 5), interest in potentially risky investments (item 3), and awareness of heightened vulnerability due to older age (item 4). Items in this measure have moderate internal consistency, with an intraclass correlation coefficient of 0.63. In prior work from this cohort, lower scam awareness was associated with older age, lower cognition, lower financial and health literacy, lower psychological well-being, and poorer decision making (14, 15, 17, 18). The average of ratings across the 5 items is the total score, and higher scores indicate lower scam awareness. Appendix Table. Items Used to Assess Scam Awareness and Descriptive Data on the Percentage of Respondents Who Responded Incorrectly to Each Item* Assessment of Cognition Cognitive function was assessed annually using 21 performance tests (15, 16, 2123). Results from 19 tests were used to create a composite measure of global cognition. These included 7 measures of episodic memory (immediate and delayed recall of Logical Memory Story A and the East Boston Story, Word List Memory, Word List Recall, and Word List Recognition), 3 measures of semantic memory (Boston Naming Test, Verbal Fluency, and Word Reading), 3 measures of working memory (Digit Span Forward, Digit Span Backward, and Digit Ordering), 4 measures of perceptual speed (Symbol Digit Modalities Test, Number Comparison, and 2 indices from a modified Stroop Neuropsychological Screening Test), and 2 measures of visuospatial ability (Standard Progressive Matrices and Judgment of Line Orientation). Raw test scores were converted to Z scores using the baseline mean and SD of the cohort, and Z scores from all 19 tests were averaged to yield the composite measure of global cognitive function (16, 22, 23). Clinical Diagnosis of Alzheimer Dementia and MCI All participants had structured annual clinical evaluations that included the cognitive performance tests described in the previous paragraph, medical history interviews, and in-person neurologic examinations (16). Clinical classification of cognitive impairment and dementia followed a 3-step process. First, 11 cognitive tests were scored by computer, and an education-adjusted rating of impairment was provided for 5 cognitive domains (16, 21). Second, impairment ratings were reviewed by a neuropsychologist blinded to scam awareness and other information except cognitive data, education, sensorimotor function, and motivation. The neuropsychologist rendered a judgment on impairment. Third, an experienced clinician reviewed the cognitive data, neuropsychologist s ratings, medical history, and neurologic examination results and rendered a decision regarding dementia and its likely cause. Clinical diagnosis of Alzheimer dementia was based on criteria from the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer s Disease and Related Disorders Association; these criteria include a history of cognitive decline with impairment in memory and at least 1 other cognitive domain (16, 24). Mild cognitive impairment was defined as cognitive impairment that did not meet criteria for dementia (14, 16, 22). Neuropathologic Indices of AD Brain autopsy followed standard procedures (25, 26). -Amyloid immunoreactive plaques were labeled with N-terminusdirected monoclonal antibodies (4G8 [Covance], 1:9000; 6F/3D [Dako North America], 1:50; or 10D5 [Elan Pharmaceuticals], 1:600). Neurofibrillary tangles were labeled with an antibody specific for phosphorylated tau (AT8 [Innogenetics], 1:1000). Computer-assisted image analysis quantified the areas occupied by -amyloid and tau tangles, and regional measures were averaged to yield composite measures of -amyloid burden and tangle density. Statistical Analysis Spearman correlations and t tests were used to describe the bivariate relationships between demographic variables and scam awareness at baseline. Cox proportional hazards models adjusted for age, sex, and education were used to examine the associations of scam awareness with incident Alzheimer dementia and MCI. Scores on the scam awareness measure served as the predictor, and time in years until the first incident event (Alzheimer dementia and MCI, separately) was the outcome. Thus, the increase in log hazard ratio (HR) is proportional with every 1-unit increase (1.4 SD) in scam awareness. Participants who had no event or died before an event were right-censored at the last evaluation. We assessed the proportional hazards assumption by using cumulative sums of Martingale residuals. We did sensitivity analyses that further examined the influence of baseline global cognitive function, extremes of age, and item-level associations. Mixed-effects models were used to examine the association of scam awareness with the starting level of and rate of decline in global cognitive function. Finally, among participants who had an autopsy (n= 264), linear regression models were used to examine the association between scam awareness and 2 molecularly specific markers of AD pathology. Role of the Funding Source This research was funded by the National Institute on Aging, which had no role in the design or conduct of the study, analysis of the data, or decision to submit the manuscript for publication. Results Descriptive Properties of the Sample Table 1 provides descriptive data on the analytic cohort. The mean scam awareness score was 2.8 (SD, 0.7), with higher scores indicating lower scam awareness. In bivariate analyses, lower scam awareness was associated with older age (P< 0.001), lower education (P< 0.001), and lower global cognition (P< 0.001) at baseline. Table 1. Baseline Characteristics of the Analytic Group and Participants Who Did and Did Not Develop Alzheimer Dementia* Incidence rates for Alzheimer dementia and MCI were higher among participants with lower scam awareness (Table 2). Rates of incident Alzheimer dementia ranged from 11 cases per 1000 person-years for high scam awareness (10th percentile) to 42 cases per 1000 person-years for low awareness (90th percentile); those of MCI ranged from 38 cases per 1000 person-years (10th percentile) to 91 cases per 1000 person-years (90th percentile). Appendix Figure 1 shows incident Alzheimer dementia and MCI by percentile of scam awareness score. Table 2. Incidence Rates of Alzheimer Dementia and MCI for Different Percentiles of Scam Awareness Appendix Figure 1. KaplanMeier curves for incident Alzheimer dementia (top) and MCI (bottom). MCI = mild cognitive impairment. Scam Awareness and Risk for Alzheimer Dementia During a mean of 5.7 years (SD, 2.4) of observation (median, 6.0 years; range, 2 to 9 years), 151 persons developed incident Al