Frailty is associated with incident Alzheimer's disease and cognitive decline in the elderly

/Home/Alzheimer's Risk Factors/Alzheimer's Risk Factors - Frailty and Age

Objective: To assess the association between frailty and incident Alzheimer's disease (AD) and cognitive decline. Frailty is common in older persons and associated with adverse health outcomes.

Methods: Study subjects included 823 older persons without dementia who participated in the Rush Memory and Aging Project, a longitudinal study of aging, and underwent annual assessments of frailty, cognition, and diagnostic evaluation for AD.

Results: During a 3-year follow-up, 89 of 823 participants developed AD. In a proportional hazards model, both baseline level of frailty and annual rate of change in frailty were associated with an increased risk of incident AD. Each additional one tenth of a unit increase on the frailty scale at baseline was associated with >9% increased risk of AD (hazard ratio: 2.44; 95% confidence interval (CI): 1.49, 3.37); each one tenth of a unit increase in annual rate of change in frailty was associated with a 12% increased risk of AD (hazard ratio: 3.30; 95% CI: 1.52, 7.13). These results were unchanged in analyses controlling for vascular risk factors and vascular diseases. Results were similar with a categorical measure of frailty instead of a continuous measure. Further, linear mixed-effects models showed that the level of and rate of change in frailty were also associated with the rate of cognitive decline.

Conclusion: Increasing frailty is associated with incident AD and the rate of cognitive decline in older persons. These findings suggest that frailty and AD may share similar etiologies.

Key Words: aging • frailty • cognitive decline • Alzheimer's disease • dementia

Abbreviations: AD = Alzheimer’s disease; PD = Parkinson’s disease; CI = confidence interval; BMI = body mass index; SD = standard deviation.


Buchman AS, Boyle PA, Wilson RS, Tang Y, Bennett DA. Frailty is associated with incident Alzheimer's disease and cognitive decline in the elderly. Psychosom Med 2007;69:483–489.

INTRODUCTION

Frailty is a multidimensional construct that represents an age-related reduction in physiologic reserve and resistance to stressors; it is widely recognized that frailty is associated with adverse health outcomes (1–4). Limited data are available about the nature of the biological processes that underlie frailty, probably because of their complex and multifactorial nature. In the absence of direct biological measures, investigators have developed operational definitions of frailty that aim to identify older persons vulnerable to adverse health outcomes (1–4). Although frailty is a heterogeneous syndrome, numerous features including physical function (e.g., impaired strength and gait), metabolism (e.g., body composition), and psychological components (e.g., fatigue) have been endorsed by many investigators as core components for classifying frail older persons (1,2).

Although the classification system proposed by Fried et al. remains the most widely used construct, there is considerable debate regarding the extent to which other common age-related conditions such as cognitive impairment and Alzheimer’s disease (AD) should be included in the definition of the syndrome (1,2). For example, cross-sectional data demonstrate a relationship between frailty and cognition (1–4). Although the clinical hallmark of AD is progressive loss of memory and other cognitive abilities, several studies have shown that persons with AD also exhibit changes in mobility and body composition (5–7), suggesting that many older persons with AD may be frail. Recent data suggest that changes in the motor system including reduced strength and walking speed and changes in body composition can antedate the onset of AD by a number of years (5–8). These findings raise the possibility that frailty, whose core features may include loss of strength, mobility, and muscle bulk (1–4), may be associated with the development of AD.

We used data from the Rush Memory and Aging Project, a longitudinal study of common chronic conditions of aging, to examine the association of frailty with incident AD and rate of cognitive decline in older persons. All participants underwent annual structured evaluations of frailty and clinical evaluations for AD and other causes of dementia. Four clinical components (strength, gait, body composition, and fatigue) previously used to identify frailty, were converted to standardized scores and the results were averaged to yield a continuous composite measure of frailty. We then used Cox proportional hazards models to examine the associations of the baseline level and the annual rate of change in frailty with the risk of incident AD. In secondary analyses, we repeated these analyses after adding covariates that might account for the association of frailty and risk of AD. Because prior research on frailty had used categorical measures, we also constructed a categorical measure of frailty and examined its association with risk of AD. Finally, we used linear mixed-effects models to examine the relationship between the change in frailty and the rate of cognitive decline.


METHODS


Participants

All participants were from the Rush Memory and Aging Project, a longitudinal clinical-pathologic investigation of chronic conditions of old age (9). Participants were recruited from >40 residential facilities across the metropolitan Chicago area, including subsidized senior housing facilities, retirement communities, and retirement homes, in addition to social service agencies and Church groups. Participants agreed to annual detailed clinical evaluations; all evaluations were performed at the parent facility or the participants’ homes to reduce burden and enhance follow-up participation. In addition, all participants signed an anatomical gift act donating their entire brain and spinal cord, as well as selected nerves and muscles, at the time of their death to Rush investigators. The study was in accordance with the latest version of the Declaration of Helsinki and was approved by the Institutional Review Board of Rush University Medical Center.

Each person underwent a uniform structured clinical evaluation, which included a medical history, neurological and physical examination, and assessment of cognitive function. Follow-up clinical evaluations, identical in all essential details to the baseline examination, were performed at 1-year intervals by examiners blinded to previously collected data. At the time of these analyses, 1085 participants had completed the baseline evaluation. Eligibility for these analyses required the absence of clinical dementia based on the clinical evaluation as well as a valid frailty score from the baseline evaluation and at least one valid frailty follow-up score to allow calculation of change in frailty using ordinary least squares. The participants considered in these analyses entered the study from October 1997 through March 2006; 67 met the criteria for dementia at baseline; 186 were excluded because they lacked follow-up examination (125 have not been in the study long enough for follow-up evaluation; 24 died before their first follow-up examination; and 37 were missing a valid follow-up frailty measure)—leaving 832 persons with a valid frailty measure and who were free of dementia at baseline. Because the primary goal of these analyses was to determine the risk of developing AD, we excluded an additional nine (1.1%) participants who developed non-AD dementias during follow-up. Those persons included for analyses (n = 823) and those excluded (n = 195) showed similar age, gender, education, frailty, body mass index (BMI), and vascular risk factors (p > .09). Those excluded had lower global cognition whereas those included were sicker based on vascular diseases (p < .03).

The group of 823 participants included in the following analyses had ≥1 follow-up evaluations (2.9 ± 1.8 (standard deviation, SD) follow-ups; range = 1–8), with >75% having ≥2 follow-up evaluations. There was missing data from 3.4% (83 of 2413) of the total possible follow-up visits. Their age was 80.4 ± 6.90 years; years of education was 14.4 ± 3.03; and Mini-Mental State Examination (MMSE) score was 27.9 ± 2.10; 74.6% were women and 91.3% were white and non-Hispanic.

Cognitive Function Testing

Cognitive function was assessed at each evaluation via a battery of 21 tests. The MMSE (10) was used to describe the cohort. Scores on 19 tests were used to create a composite measure of global cognitive function: immediate and delayed recall of story A from Logical Memory (11), immediate and delayed recall of the East Boston Story (12,13), Word List Memory, Word List Recall, Word List Recognition (14), a 15-item version of the Boston Naming Test (15), Verbal Fluency (14), a 15-item reading test (13) Digit Span Forward, Digit Span Backward (11), Digit Ordering (16), Symbol Digit Modalities Test (17), Number Comparison (18), two indices from a modified version of the Stroop Neuropsychological Screening Test (19), a 15-item version of Judgment of Line Orientation (20), and a 16-item version of Standard Progressive Matrices (21). One additional test, Complex Ideational Material (22), was used for diagnostic classification purposes only. To compute the composite measure of global cognitive function, the raw scores on each of the individual tests were converted to z scores, using the baseline mean and SD of the entire cohort, and the z scores of all 19 tests were averaged. Psychometric information on this composite measure is contained in previous publications (23).

Clinical Diagnoses

Subjects underwent a uniform structured clinical evaluation including a medical history, neurologic examination, and cognitive performance testing (9). Cognitive tests were reviewed by an experienced neuropsychologist. Participants were evaluated in person by a physician, who used all available cognitive and clinical testing results, to diagnose dementia and other common neurologic conditions affecting cognitive function. The diagnosis of dementia followed the criteria of the joint working group of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association (24). These require a history of cognitive decline and evidence of impairment in two or more domains of cognition—one of which must be memory—for classification as Alzheimer’s disease. Probable AD refers to persons meeting these criteria who do not have another condition thought to be contributing to dementia. Possible AD refers to persons meeting these criteria who also have another condition (e.g., stroke) contributing to dementia. The diagnoses of parkinsonism and Parkinson’s disease are made according to the clinical criteria recommended by the Core Assessment Program for Intracerebral Transplantation (25). The diagnosis of Dementia with Lewy Bodies is adapted from the recommendations of the Report of the Consortium on DLB International Workshop (26). The diagnosis of major depression is based on DSM-III-R criteria supported by a subset of items from the Diagnostic Interview Schedule (27). Using this approach, we found >90% of participants with clinically probable AD and about 70% of those with clinically possible AD met the neuropathologic criteria for AD, a rate comparable to the positive predictive value in clinical settings (28).

Composite Frailty Measure

A continuous composite measure of frailty was based on four frailty components: grip strength, timed walk, body composition, and fatigue (1,2,29). This approach, rather than the more standard dichotomous or trichotomous variable (1,2), was used for the primary analyses because our primary interest was to use a sensitive measure to document change in frailty over time and determine its association with incident AD and cognitive decline. Strength was based on grip strength measured with the Jamar hydraulic hand dynamometer (Lafayette Instruments, Lafayette, IN). This sealed hydraulic system features a dual-scale readout that displays isometric grip force from 0 to 200 pounds. Two trials of grip strength were obtained for each hand. The four trials were averaged to yield strength. Gait was based on the time to walk 8 feet. Although almost all participants in our cohort were able to walk at baseline, some were no longer able to ambulate at follow-up but still had other valid frailty components. Therefore, by using a 6-point scale (from 0 to 6) and assigning a 0 to those unable to walk, we were able to include all participants so as to not lose data for longitudinal analyses. For the remaining participants who were able to walk, the walk times were converted to quintiles, with a score of 5 assigned to the fastest times and a score of 1 to the slowest times. Body composition was based on BMI = weight/height (2). As done in previous studies of frailty, we used two questions derived from a modified version of the Center for Epidemiologic Studies-Depression Scale to assess fatigue (questions 7 and 20). These were: a) I felt that everything I did was an effort, and b) I could not get "going" (1,2). Participants answered "yes" or "no" to each question, and the values were summed to yield a three-level variable. The four components used to construct the composite measure of frailty were structured so that higher values would indicate poorer performance (higher frailty) and lower values would indicate better performance (less frailty), to be consistent with prior literature. The composite measure of frailty was constructed by converting the raw score from each of the four component measures to z scores, using the mean and SD values from all participants at baseline.

Categorical Frailty Measure

Frailty, which represents an age-related reduction in physiologic reserve and resistance to stressors, is widely considered to represent a dichotomous or trichotomous state (1–4). Thus, although a continuous composite measure of frailty may provide a useful measure to assess change in frailty over time, such a measure does not represent the state of frailty that has been used in prior research (1–4). Therefore, to provide data similar to previous studies, we dichotomized each of the four components used to construct composite frailty. Similar to previous reports, the lowest quintile of grip, gait, and BMI were defined as frail and any reports of fatigue were considered consistent with frailty. Participants with 0 or 1 frail components were considered not frail and those participants with ≥2 frail components were considered frail. Categorical frailty and composite frailty were highly correlated (ρ = .55; p < .001).

Comorbidities and Other Covariates

Gender and race were recorded at the baseline interview. Race questions and categories were those used by the 1990 US Census. Gender was coded as "1" for men and "0" for women. Age in years was computed from self-reported date of birth and date of the clinical examination at which the strength measures were collected. Education (reported highest grade or years of education) was obtained at the time of the baseline cognitive testing. As previously described (30), to assess the influence of cumulative vascular risk factor and vascular disease burden on motor function, we computed summary scores indicating each individual’s vascular risk factor (i.e., the sum of hypertension, diabetes mellitus, and smoking, resulting in a score ranging from 0 to 3 for each individual (1.15 ± 0.80)) and vascular disease burden (i.e., the sum of heart attack, congestive heart failure, claudication, and stroke), resulting in a score ranging from 0 to 4 for each individual (0.35 ± 0.63). These summary scores were used as covariates in the analyses.

Statistical Analyses

Pearson correlations were used to examine the associations of baseline frailty with age and education. The t tests were used to compare demographic measures among men and women as well as those who did and did not develop AD. We used ordinary least-squares regression to estimate the annual rate of change in frailty for each person (31). Previous studies have used a quadratic term for BMI (BMI x BMI) because both high and low BMI may be associated with adverse health outcomes (31). Because BMI is one of the components of composite frailty, we first examined whether BMI x BMI was associated with the risk of AD. Because BMI x BMI was not associated with incident AD (data not shown), only a term for BMI was considered in subsequent models as part of composite frailty.

A series of Cox proportional hazards models were used to examine the association of frailty with incident AD. All models controlled for age, gender, and education. First, we examined the association of baseline level of frailty with incident AD. Next, to examine the association of baseline level of frailty and annual rate of change in frailty with incident AD, we added a term for the rate of change in frailty as determined by ordinary least squares. Finally, we added terms for vascular risk factors and diseases to determine if these covariates affected the association of the baseline level or change in frailty with incident AD. We then repeated the core Cox models, using the categorical measure for frailty, to see whether it showed similar associations with incident AD.

We next conducted a complementary set of analyses using linear mixed-effects models (32) to examine the association of level of and rate of change in frailty with rate of cognitive decline, which is the clinical hallmark of AD. The outcome for this model was a global measure of cognition and included a term for linear cognitive decline, time (in years from baseline), and a term to account for nonlinear cognitive decline (time x time). The model also included terms for baseline frailty and rate of change in frailty and their interactions with cognitive decline. The terms for frailty and rate of change in frailty indicate the associations of these variables with cognition at baseline, and the interaction terms denote their associations with rate of cognitive decline. In addition, because age, gender, and education are associated with cognitive decline, all models controlled for these demographic variables and their interactions with time. Initial analyses showed that the terms for the interaction of demographic variables with time x time were not significant (results not shown) and were therefore not included in the models presented. Models were examined graphically and analytically and assumptions were judged to be adequately met (33). Programming was done in SAS (34).


RESULTS


Metric Properties of Frailty

Scores on the composite measure of frailty ranged from –1.73 to 1.92 (–0.03 ± 0.56) with higher scores indicating more frailty (poorer performance). Frailty was positively related to age (r = .34; p = < .001), negatively associated with education (r = –0.16; p < .001), and men were less frail than women were (t[821] = 8.71; p = < .001).

Baseline Frailty and Risk of Incident AD

During a mean of about 3 years of follow-up (2.91 ± 1.94 years), there were 98 cases of incident dementia. Eighty-two persons developed probable AD and seven had possible AD including three with cognitive impairment due to stroke, three with Parkinson’s disease dementia or Lewy body disease, and one with depression. Nine persons developed dementia from another cause and were excluded from the following analyses. The mean time to developing AD was about 3 years (2.93 ± 1.85 years). Participants who developed AD were older, had lower cognitive testing scores and higher levels of frailty at baseline than those who did not develop AD (Table 1).

___
TABLE 1. Baseline Characteristics of Study Participants
http://www.psychosomaticmedicine.org/content/69/5/483/T1.expansion.html
___


We first examined the relationship of baseline level of frailty to the risk of AD in a Cox proportional hazards model controlling for the potentially confounding effects of age, gender, and education. In this model, each one tenth unit increase in frailty at baseline was associated with about an 8% increase in the risk of AD (hazard ratio: 1.94; 95% CI: 1.31, 2.87). Figure 1A shows that high frailty at baseline (90th percentile, solid line) was associated with >2.5-fold greater risk of developing AD compared with low frailty at baseline (10th percentile, dotted line). These results were unchanged when we excluded the seven participants with possible AD from the analyses (results not shown). The addition of terms for vascular risk factors or diseases to the previous model did not change the association of baseline frailty with incident AD (results not shown).


___
Figure 1. Frailty and the risk of Alzheimer’s disease (AD). A. Baseline frailty. This figure shows the cumulative hazard of AD during the study for two participants: The first has low baseline frailty, i.e., better performance, measured at baseline (solid line; 10th percentile: –0.75 frailty units) and a second with high baseline frailty, i.e., poorer performance (dotted line; 90th percentile: 0.68 frailty units), adjusted for age, gender, and education. B. Annual rate of change in frailty. This figure shows the cumulative hazard of AD during the study for two participants: The first has a mildly decreasing rate of change in frailty, i.e., improving performance (solid line; 10th percentile: –0.20 frailty units/year) and a second with a rapidly increasing rate of change in frailty, i.e., deteriorating performance (dotted line; 90th percentile: 0.39 frailty units/year), adjusted for age, gender, education, and baseline frailty.
http://www.psychosomaticmedicine.org/content/69/5/483/F1.expansion.html
___


Annual Rate of Change in Frailty and Risk of Incident AD

Like AD, the onset of frailty occurs slowly over time and our measure of frailty was intended to capture the slow progressive nature of the condition. Therefore, we next examined whether the rate of change in frailty was associated with incident AD. Each person’s annual rate of change in frailty was estimated using ordinary least-squares regression. Overall, the rate of change in frailty ranged widely from –1.95 to 1.66 (0.09 ± 0.30). About two thirds (65.4%) of the cohort exhibited increasing frailty. By contrast, the remaining persons (34.6%) exhibited less frailty over time. There was a modest inverse correlation between baseline level of frailty and annual rate of change in frailty (r = –.20; p = < .001).

To examine the association of the rate of change in frailty with the risk of AD, we added a term for the rate of change in frailty, as determined by the least-squares methods, to a Cox proportional hazards model controlling for age, gender, education, and baseline frailty. In this analysis, the rate of change in frailty was associated with an increased risk of AD. Each one tenth of a unit higher level of frailty at baseline was associated with >9% increase in the risk of AD (hazard ratio: 2.24; 95% CI: 1.49, 3.37); each one tenth of a unit annual increase in rate of change in frailty was associated with about a 12% increase in the risk of AD (hazard ratio: 3.30; 95% CI: 1.52, 7.13). The results were unchanged when we excluded seven participants with possible AD (results not shown). Figure 1B shows a person with a rapid increase in frailty (90th percentile, dotted line) who had an almost two-fold increased risk of developing AD compared with a person experiencing a mild decrease in frailty (10th percentile, solid line). The addition of terms for vascular risk factors and diseases to the previous model did not change the association of baseline frailty or rate of change in frailty with incident AD (results not shown).

Categorical Measure of Frailty and Risk of AD

Because many researchers consider frailty to be a state rather than representing a continuum, we examined the association between a categorical measure of frailty and risk of AD (1–4). In a Cox proportional hazards model which controlled for age, gender, and education, dichotomous frailty at baseline was associated with incident AD (hazard ratio: 2.10; 95% CI: 1.27, 3.46). Thus, both the categorical and composite measures of frailty were both associated with incident AD.

Frailty and Cognitive Decline

AD develops slowly over many years and its clinical hallmark is change in cognitive function. Thus, to some extent, a diagnosis of AD consists of placing a cut-point along a continuum of cognition. To ensure that our results were not influenced by diagnostic misclassification, we next examined the association of frailty with change in cognitive function using a previously established composite measure of global cognition (23). At baseline, scores on the composite measure of global cognition ranged from –1.82 to 1.40, with higher scores indicating better cognitive function.

We used a linear mixed-effects model to test the association of baseline frailty and rate of change in frailty with the level of and rate of cognitive decline, adjusting for age, gender, and education. Overall, there was a nonlinear accelerating rate of cognitive decline, as indicated by the negative term for time x time (Table 2). A higher level of baseline frailty was associated with a lower level of cognitive function at baseline, as indicated by the term for baseline frailty (Table 2). In addition, a higher level of baseline frailty was associated with an increased rate of global cognitive decline, as indicated by the interaction of baseline frailty with linear change in cognition (time) but there was only a marginal association with nonlinear change in cognition (time x time) (Table 2). Thus, a higher level of baseline frailty was associated with both baseline level and rate of global cognitive decline. The rate of change in frailty was also associated with cognitive decline as observed by the interaction with both time and time x time (Table 2).

___
TABLE 2. Associations of Frailty With Level of and the Rate of Cognitive Decline
http://www.psychosomaticmedicine.org/content/69/5/483/T2.expansion.html
___



DISCUSSION


In a cohort of >800 well-characterized older persons free of dementia at baseline, we found that a higher baseline level as well as a more rapid increase in frailty were both associated with an increased risk of incident AD. These findings were robust in that they were not due to the influence of other common comorbidities and were observed using both a continuous composite measure of frailty and a categorical measure consistent with previous research. Further, in separate analyses, both the level of and rate of change in frailty were associated with the rate of cognitive decline. These findings suggest that diagnostic misclassification did not strongly influence the association of frailty with incident AD. These findings raise the possibility that frailty and AD may share common etiologies.

Frailty is common among older persons and is associated with adverse health outcomes including increased morbidity and mortality. Previous cross-sectional studies have reported that frailty also is associated with the level of cognition and dementia (1–4). Several of the individual components used to construct the measure of frailty in this study, including slowed gait, slowed movement, and weight loss, have been associated with the development of dementia and incident AD (5–8,35,36). We are also aware of one study that reported the relationship between grip strength and risk of AD (8). However, we are not aware of previous longitudinal studies that have examined the association of frailty or change in frailty with incident AD. The findings in this study were robust in that they were observed with both measures of frailty (1–3,37). Further, we found that both level of and rate of change in frailty were associated with the rate of change in cognition over the course of the study, suggesting that findings with incident AD were not due to misclassification. Thus, the results of the current study suggest that frailty occurs before the onset of clinical AD and is associated with the rate of cognitive decline in older people.

Aging and AD are associated with a wide range of mental health symptoms including agitation, depression, apathy, delusions, hallucinations, and sleep impairment (38). However, the relationship between frailty and these symptoms and other chronic diseases is poorly understood. Our findings suggest that frailty and its components and cognitive decline may share a common etiopathogenesis, i.e., that factors associated with the development of frailty are also associated with development of AD. For example, risk factors for cardiovascular disease (e.g., diabetes) and common vascular diseases (e.g., congestive heart failure, brain infarcts) have been related to both frailty (39) and AD (40). Increased markers of inflammation such as C-reactive protein or proinflammatory interleukins are common and have been implicated in frailty (41), cognitive impairment (42), and AD (43,44). Although controlling for vascular diseases and risk factors had little effect on the association of frailty and AD, we were unable to account for inflammatory markers. Further, both frailty and AD are complex diseases and it is likely that many other factors are also involved.

Another intriguing possibility that could account for the link between frailty and AD relates to the neuropathologic changes of AD (i.e., the development of neuritic plaques and neurofibrillary tangles). Among persons with AD, neurofibrillary tangles in the substantia nigra have been related to motor signs such as gait slowing (45–48). Further, it is now well known that AD pathology can express itself clinically in persons without dementia. Several studies have reported that AD pathology is related to mild cognitive impairment (49) and is associated with subtle memory deficits in persons without cognitive impairment (50). Second, some of the individual signs of frailty such as grip strength (8), body composition (31), and gait disturbance (7,35) have been related to mild cognitive impairment and risk of AD. Third, AD pathology is known to accumulate in neural systems that subserve motor function such as the substantia nigra, primary and supplementary motor cortices, and striatum (46–48). Recent evidence suggests that neurofibrillary tangles in the substantia nigra are related to gait disturbance even in persons without dementia (51), and widespread neuritic plaques and tangles are related to BMI in persons without dementia (52). Overall, these findings suggest that frailty may be related to the accumulation of neuritic plaques and neurofibrillary tangles in neural systems that underlie motor function, body composition, and fatigue. Because the biology of frailty is poorly understood, further studies may be helpful to examine the link between AD pathology and frailty in older persons.

Importantly, frailty is a multidimensional construct that includes a psychological dimension, fatigue. Studies examining how the psychological dimension of frailty is linked with the risk of AD and cognitive decline are needed, as this remains unclear. For example, although AD is accompanied by a wide range of mental health problems in the elderly (38), current data show that psychological factors such as depression and loneliness are associated with AD and cognitive decline but not with AD pathology (53,54). Such findings suggest that the psychological dimension of frailty may be linked with dementia and cognitive decline through a mechanism other than AD neuropathology. Thus, although the current study is a necessary first step toward examining the relationship of frailty, AD, and cognitive decline, the biology underlying this association remains unclear. Future research will need to examine other mechanisms that may link frailty with dementia and cognitive decline, and assess if the specific components of frailty such as fatigue mediate or modify the association of the other dimensions of frailty with AD and cognitive decline.

This study has some limitations. Our study is restricted to older persons willing to provide organ donation at the time of their death, which could have introduced bias. Although we were able to measure change in frailty, some of the individual components such as fatigue were assessed crudely, which may have resulted in an underestimation of change in frailty. Follow-up in this study was short and findings could differ over longer periods of study. Persons with dementia at baseline, those who did not survive to the time of their first follow-up examination, and those who were not yet eligible for follow-up were excluded from analyses. The inclusion of these persons could have affected our associations. However, several factors increase confidence in the findings from this study. These findings are based on a large number of well-characterized older persons free of dementia at baseline, and both composite measure of frailty as well as a dichotomous measure of frailty, cognitive function, and AD were based on uniform structured procedures with high rates of follow-up increasing internal validity.

We thank the many Illinois residents for their participating in the Rush Memory and Aging Project; Traci Colvin, MPH, and Tracy Hagman for coordinating the study; Liping Gu, MS, and Woojeong Bang, MS, for statistical programming; George Dombrowski, MS, and Greg Klein for data management.

NOTES

Received for publication May 16, 2006; revision received March 1, 2007.

This research was supported by Grants R01AG17917 (D.A.B.) and R01AG24480 (A.S.B.) from the National Institute on Aging, the Illinois Department of Public Health, and the Robert C. Borwell Endowment Fund.

DOI:10.1097/psy.0b013e318068de1d

REFERENCES
1. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, Seeman T, Tracy R, Kop WJ, Burke G, McBurnie MA. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 2001;56:M146–56.[Abstract/Free Full Text]
2. Ferrucci L, Guralnik JM, Studenski S, Fried LP, Cutler GB Jr, Walston JD. Designing randomized, controlled trials aimed at preventing or delaying functional decline and disability in frail, older persons: a consensus report. J Am Geriatr Soc 2004;52:625–34.[CrossRef][Medline]
3. Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci 2004;59:255–63.[Medline]
4. Mitnitski AB, Song X, Rockwood K. The estimation of relative fitness and frailty in community-dwelling older adults using self-report data. J Gerontol A Biol Sci Med Sci 2004;59:M627–32.[Abstract/Free Full Text]
5. Cronin-Stubbs D, Beckett LA, Scherr PA, Field TS, Chown MJ, Pilgrim DM, Bennett DA, Evans DA. Weight loss in people with Alzheimer's disease: a prospective population based analysis. BMJ 1997;314:178–9.[Free Full Text]
6. Scarmeas N, Albert M, Brandt J, Blacker D, Hadjigeorgiou G, Papadimitriou A, Dubois B, Sarazin M, Wegesin D, Marder K, Bell K, Honig L, Stern Y. Motor signs predict poor outcomes in Alzheimer disease. Neurology 2005;64:1696–703.[Abstract/Free Full Text]
7. Mitchell SL, Rockwood K. The association between parkinsonism, Alzheimer's disease, and mortality: a comprehensive approach. J Am Geriatr Soc 2000;48:422–5.[Medline]
8. Wang L, Larson EB, Bowen JD, van Belle G. Performance-based physical function and future dementia in older people. Arch Intern Med 2006;166:1115–20.[Abstract/Free Full Text]
9. Bennett DA, Schneider JA, Buchman AS, Mendes de Leon C, Bienias JL, Wilson RS. The Rush memory and aging project: study design and baseline characteristics of the study cohort. Neuroepidemiology 2005;25:163–75.[CrossRef][Medline]
10. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state." A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–98.[CrossRef][Medline]
11. Wechsler D. Wechsler Memory Scale—Revised Manual. San Antonio, Texas: Psychological Corporation; 1987.
12. Albert M, Smith LA, Scherr PA, Taylor JO, Evans DA, Funkenstein HH. Use of brief cognitive tests to identify individuals in the community with clinically diagnosed Alzheimer's disease. Int J Neurosci 1991;57:167–78.[Medline]
13. Wilson RS, Beckett LA, Barnes LL, Schneider JA, Bach J, Evans DA, Bennett DA. Individual differences in rates of change in cognitive abilities of older persons. Psychol Aging 2002;17:179–93.[CrossRef][Medline]
14. Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum G, Mellits ED, Clark C. The Consortium to Establish a Registry for Alzheimer's Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer's disease. Neurology 1989;39:1159–65.[Abstract/Free Full Text]
15. Kaplan E, Goodglass H, Weintraub S. The Boston Naming Test. Philadelphia: Lea & Febiger; 1983.
16. Cooper JA, Sagar HJ. Incidental and intentional recall in Parkinson's disease: an account based on diminished attentional resources. J Clin Exp Neuropsychol 1993;15:713–31.[Medline]
17. Smith A. Symbol Digit Modalities Test Manual—Revised. Los Angeles: Western Psychological Services; 1992.
18. Ekstrom R, French J, Harman H, Kermen D. Manual for Kit of Factor-Referenced Cognitive Tests. Princeton, NJ: Educational Testing Service; 1976.
19. Trenerry M, Crosson B, DeBoe J, Leber W. The Stroop Neuropsychological Screening Test. Odessa, FL: Psychological Assessment Resources; 1989.
20. Benton A, Sivan A, Hamsher K, Varney N, Spreen O. Contributions to Neuropsychological Assessment. 2nd ed. New York: Oxford University Press; 1994.
21. Raven J, Court J, Raven J. Manual for Raven's Progressive Matrices and Vocabulary: Standard Progressive Matrices. Oxford: Oxford Psychologists Press; 1992.
22. Goodglass H, Kaplan E. The Assessment of Aphasia and Related Disorders. Philadelphia: Lea & Febiger; 1972.
23. Wilson RS, Barnes LL, Krueger KR, Hoganson G, Bienias JL, Bennett DA. Early and late life cognitive activity and cognitive systems in old age. J Int Neuropsychol Soc 2005;11:400–7.[Medline]
24. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA work group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease. Neurology 1984;34:939–44.[Abstract/Free Full Text]
25. Langston JW, Widner H, Goetz CG, Brooks D, Fahn S, Freeman T, Watts R. Core assessment program for intracerebral transplantations (CAPIT). Mov Disord 1992;7:2–13.[Medline]
26. McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, Salmon DP, Lowe J, Mirra SS, Byrne EJ, Lennox G, Quinn NP, Edwardson JA, Ince PG, Bergeron C, Burns A, Miller BL, Lovestone S, Collerton D, Jansen EN, Ballard C, de Vos RA, Wilcock GK, Jellinger KA, Perry RH. Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology 1996;47:1113–24.[Abstract/Free Full Text]
27. Robins LN, Helzer JE, Ratcliff KS, Seyfried W. Validity of the diagnostic interview schedule, version II: DSM-III diagnoses. Psychol Med 1982;12:855–70.[Medline]
28. Bennett DA, Schneider JA, Aggarwal NT, Arvanitakis Z, Shah RC, Kelly JF, Fox JH, Cochran EJ, Arends D, Treinkman AD, Wilson RS. Decision rules guiding the clinical diagnosis of Alzheimer's disease in two community-based cohort studies compared to standard practice in a clinic-based cohort study. Neuroepidemiology 2006;27:169–76.[CrossRef][Medline]
29. Wilson RS, Buchman AS, Arnold SE, Shah RC, Tang Y, Bennett DA. Harm avoidance and disability in old age. Exp Aging Res 2006;32:243–61.[CrossRef][Medline]
30. Boyle PA, Wilson RS, Aggarwal NT, Arvanitakis Z, Kelly J, Bienias JL, Bennett DA. Parkinsonian signs in subjects with mild cognitive impairment. Neurology 2005;65:1901–6.[Abstract/Free Full Text]
31. Buchman AS, Wilson RS, Bienias JL, Shah RC, Evans DA, Bennett DA. Change in body mass index and risk of incident Alzheimer disease. Neurology 2005;65:892–7.[Abstract/Free Full Text]
32. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics 1982;38:963–74.[CrossRef][Medline]
33. Collett D. Modeling Survival Data in Medical Research 2nd ed. Boca Raton, Florida: Chapman & Hall; 2003.
34. SAS. SAS/STAT User's Guide, Version 8. Cary, NC: SAS Institute Inc; 2000.
35. Waite LM, Grayson DA, Piguet O, Creasey H, Bennett HP, Broe GA. Gait slowing as a predictor of incident dementia: 6-year longitudinal data from the Sydney older persons study. J Neurol Sci 2005;229–230:89–93.
36. Wilson RS, Schneider JA, Bienias JL, Evans DA, Bennett DA. Parkinsonianlike signs and risk of incident Alzheimer disease in older persons. Arch Neurol 2003;60:539–44.[Abstract/Free Full Text]
37. Rockwood K, Howlett SE, MacKnight C, Beattie BL, Bergman H, Hebert R, Hogan DB, Wolfson C, McDowell I. Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian study of health and aging. J Gerontol A Biol Sci Med Sci 2004;59:1310–7.[Abstract/Free Full Text]
38. Lyketsos CG, Lopez O, Jones B, Fitzpatrick AL, Breitner J, DeKosky S. Prevalence of neuropsychiatric symptoms in dementia and mild cognitive impairment: results from the cardiovascular health study. JAMA 2002;288:1475–83.[Abstract/Free Full Text]
39. Newman AB, Gottdiener JS, McBurnie MA, Hirsch CH, Kop WJ, Tracy R, Walston JD, Fried LP. Associations of subclinical cardiovascular disease with frailty. J Gerontol A Biol Sci Med Sci 2001;56:M158–66.[Abstract/Free Full Text]
40. Arvanitakis Z, Wilson RS, Bienias JL, Evans DA, Bennett DA. Diabetes mellitus and risk of Alzheimer disease and decline in cognitive function. Arch Neurol 2004;61:661–6.[Abstract/Free Full Text]
41. Puts MT, Visser M, Twisk JW, Deeg DJ, Lips P. Endocrine and inflammatory markers as predictors of frailty. Clin Endocrinol (Oxf) 2005;63:403–11.[CrossRef][Medline]
42. Weaver JD, Huang MH, Albert M, Harris T, Rowe JW, Seeman TE. Interleukin-6 and risk of cognitive decline: MacArthur studies of successful aging. Neurology 2002;59:371–8.[Abstract/Free Full Text]
43. Ehl C, Kolsch H, Ptok U, Jessen F, Schmitz S, Frahnert C, Schlosser R, Rao ML, Maier W, Heun R. Association of an interleukin-1beta gene polymorphism at position –511 with Alzheimer's disease. Int J Mol Med 2003;11:235–8.[Medline]
44. Ma SL, Tang NL, Lam LC, Chiu HF. The association between promoter polymorphism of the interleukin-10 gene and Alzheimer's disease. Neurobiol Aging 2005;26:1005–10.[CrossRef][Medline]
45. Liu Y, Stern Y, Chun MR, Jacobs DM, Yau P, Goldman JE. Pathological correlates of extrapyramidal signs in Alzheimer's disease. Ann Neurol 1997;41:368–74.[CrossRef][Medline]
46. Burns JM, Galvin JE, Roe CM, Morris JC, McKeel DW. The pathology of the substantia nigra in Alzheimer disease with extrapyramidal signs. Neurology 2005;64:1397–403.[Abstract/Free Full Text]
47. Gearing M, Levey AI, Mirra SS. Diffuse plaques in the striatum in Alzheimer disease (AD): relationship to the striatal mosaic and selected neuropeptide markers. J Neuropathol Exp Neurol 1997;56:1363–70.[Medline]
48. Wolf DS, Gearing M, Snowdon DA, Mori H, Markesbery WR, Mirra SS. Progression of regional neuropathology in Alzheimer disease and normal elderly: findings from the Nun study. Alzheimer Dis Assoc Disord 1999;13:226–31.[CrossRef][Medline]
49. Bennett DA, Schneider JA, Bienias JL, Evans DA, Wilson RS. Mild cognitive impairment is related to Alzheimer disease pathology and cerebral infarctions. Neurology 2005;64:834–41.[Abstract/Free Full Text]
50. Bennett DA, Schneider JA, Arvanitakis Z, Kelly JF, Aggarwal NT, Shah RC, Wilson RS. Neuropathology of older persons without cognitive impairment from two community-based studies. Neurology 2006;66:1837–44.[Abstract/Free Full Text]
51. Schneider JA, Li JL, Li Y, Wilson RS, Kordower JH, Bennett DA. Substantia nigra tangles are related to gait impairment in older persons. Ann Neurol 2006;59:166–73.[CrossRef][Medline]
52. Buchman AS, Schneider JA, Wilson RS, Bienias JL, Bennett DA. Body mass index in older persons is associated with Alzheimer disease pathology. Neurology 2006;67:1949–54.[Abstract/Free Full Text]
53. Wilson RS, Krueger KR, Arnold SE, Schneider JA, Kelly JF, Barnes LL, Tang Y, Bennett DA. Loneliness and risk of Alzheimer disease. Arch Gen Psychiatry 2007;64:234–40.[Abstract/Free Full Text]
54. Wilson RS, Schneider JA, Bienias JL, Arnold SE, Evans DA, Bennett DA. Depressive symptoms, clinical AD, and cortical plaques and tangles in older persons. Neurology 2003;61:1102–7.[Abstract/Free Full Text]