Psychology and Aging 2015, Vol. 30, No. 2, 279-285

© 2015 American Psychological Association 0882-7974/15/$ 12.00 http://dx.doi.org/10.1037/pag0000024

Cognitive Aging in Older Black and White Persons Robert S. Wilson, Ana W. Capuano, Joel Sytsma, David A. Bennett, and Lisa L. Barnes Rush University Medical Center During a mean of 5.2 years of annual follow-up, older Black (n = 647) and White (n = 647) persons of equivalent age and education completed a battery of 17 cognitive tests from which composite measures of 5 abilities were derived. Baseline level of each ability was lower in the Black subgroup. Decline in episodic and working memory was not related to race. Decline in semantic memory, perceptual speed, and visuospatial ability was slower in Black persons than White persons, and in semantic memory and perceptual speed this effect was stronger in older than younger participants. Racial differences persisted after adjustment for retest effects. The results suggest subtle cognitive aging differences between Black persons and White persons. Keywords: cognitive aging, racial differences, longitudinal study, retest learning

Most knowledge about cognitive aging conies from research on White people. In the last 15 years, however, several longitudinal studies have compared change in cognitive function in older Black and White people. Some studies have found more rapid cognitive decline in Black people (Lyketsos, Chen, & Anthony, 1999; SachsEricsson & Blazer, 2005; Sawyer, Sachs-Ericsson, Preacher, & Blazer, 2009; Wolinsky et al., 2011), some have found more rapid cognitive decline in White people (Sloan & Wang, 2005; Alley, Suthers, & Crimmins, 2007; Karlamangla et ah, 2009; Early et al., 2013), and some have found no difference (Atkinson et al., 2005; Castora-Binkley, Peronto, Edwards, & Small, 2013; Marsiske et al., 2013; Masel & Peek, 2009). The factors contributing to these inconsistent findings are uncertain. One issue is that some studies

have assessed cognition at only two points (Atkinson et al., 2005; Lyketsos et al., 1999; Sachs-Ericsson & Blazer, 2005; Wolinsky et al., 2011). This makes it difficult to separate level of cognitive function from rate of change which is problematic given BlackWhite differences in cognitive level. With few exceptions (Early et al., 2013; Marsiske et al., 2013), previous studies have assessed cognition with brief global measures such as the Short Portable Mental Status Questionnaire (Sachs-Ericsson & Blazer, 2005; Sawyer et al., 2009), Telephone Interview for Cognitive Status (Alley et al., 2007; Castora-Binkley et al., 2013; Karlamangla et al„ 2009; Masel & Peek, 2009; Sloan & Wang, 2005; Wolinsky et al., 2011), and Mini-Mental State Examination (Atkinson et al., 2005; Lyketsos et al., 1999), which lack measurement precision and the ability to characterize specific domains of cognitive func­ tion. Retest learning (Yang, Reed, & Kuan, 2012), defined as improved cognitive performance due to repeated test administra­ tion, is known to affect estimates of late-life cognitive decline (Wilson, Li, Bienias, & Bennett, 2006; Yang et al., 2012), but it is not known whether there are racial differences in retest learning that are affecting comparisons of cognitive trajectories between racial groups. Finally, it is possible that the comparability of Black and White participants has varied from study to study. The present analyses are based on older Black and White participants in three longitudinal cohort studies. Subgroups of older Black and White people without dementia at enrollment (each n = 647) were selected to be equivalent in age, education, and number of cognitive assessments using propensity matching. At annual intervals for a mean of more than 5 years, they com­ pleted a battery of 17 cognitive tests from which previously established composite measures of five cognitive domains were derived. We used mixed-effects models to test for racial differ­ ences in level of function and rate of change in each cognitive domain.

This article was published Online First May 11, 2015. Robert S. Wilson. Department of Neurological Sciences and Department of Behavioral Sciences, Rush Alzheimer's Disease Center, Rush University Med­ ical Center; Ana W. Capuano, Department of Neurological Sciences, Rush Alzheimer’s Disease Center, Rush University Medical Center; Joel Sytsma, Rush Alzheimer’s Disease Center, Rush University Medical Center; David A. Bennett, Department of Neurological Sciences, Rush Alzheimer’s Disease Center, Rush University Medical Center; Lisa L. Barnes, Department of Neurological Sciences and Department of Behavioral Sciences, Rush Alzhei­ mer’s Disease Center, Rush University Medical Center. This research was supported by National Institutes of Flealth Grants R01AG17917, P30AG10161, and R01AG36042 and the Illinois Depart­ ment of Public Health. The funding organizations had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript. We thank the participants of the Minority Aging Research Study, the Rush Memory and Aging Project, and the Religious Order Study for their invaluable contributions. We thank Charlene Gamboa, Tracy Colvin, Tracey Nowakowski, Barbara Eubeler, Karen Lowe-Graham, and Karen Skish for study recruitment and coordination; John Gibbons and Greg Klein for data management; Alysha Kett for data analysis; and the staff of the Rush Alzheimer’s Disease Center. Correspondence concerning this article should be addressed to Robert S. Wilson, Rush Alzheimer’s Disease Center, Rush University Medical Cen­ ter, 600 South Paulina Street, Suite 1027a, Chicago, IL, 60612. E-mail: [email protected]

Method Participants Participants were drawn from three ongoing longitudinal cohort studies with nearly identical protocols. The Minority Aging Re279

280

WILSON, CAPUANO, SYTSMA, BENNETT, AND BARNES

search Study began in 2004 and involves annual clinical evalua­ tions of older Black people in the Chicago area recruited from the community and the clinical core of the Rush Alzheimer’s Disease Core Center (Arvanitakis, Bennett, Wilson, & Barnes, 2010; Barnes, Shah, Aggarwal, Bennett, & Schneider, 2012). The Rush Memory and Aging Project began in 1997 and involves annual clinical evaluations of older people in the Chicago area; partici­ pants are mostly White but approximately 6% are Black (Bennett et al., 2005; Bennett et ah, 2012). The Religious Orders Study began in 1994 and involves annual clinical evaluations of Catholic nuns, priests, and monks from across the United States; approxi­ mately 7% are Black (Wilson, Bienias, Evans, & Bennett, 2004; Bennett, Schneider, Arvanitakis, & Wilson, 2012). In all three studies, participants signed informed-consent forms after a thor­ ough discussion about the project. Each project was approved by the institutional review board of Rush University Medical Center. Eligibility for these analyses required absence of dementia at baseline and completion of at least one annual follow-up evalua­ tion. At the time of these analyses, 708 Black people and 2,311 White people met these criteria. With the use of propensity scores, we identified subgroups of 647 Black participants (Minority Aging Research Study, 487; Rush Memory and Aging Project, 76; Reli­ gious Orders study, 84) and 647 White participants (Rush Memory and Aging Project, 429; Religious Orders Study, 218) that were balanced in relation to age, education, and number of cognitive assessments because age is related to cognitive decline, education is related to cognitive level, and the duration of observation and number of follow-ups impact the ability to reliably characterize cognitive trajectories. We used a greedy 5-to-l digit algorithm in SAS to match propensity scores and identify the subgroups (Rassen et ah, 2012). As a result of the propensity balancing, the Black subgroup was similar to the White subgroup in age at baseline (73.5 vs. 73.6), r( 1,292) = 0.2, p = .839), years of education (15.2 vs. 15.4), t( 1,292) = 0.9, p = .370, and number of annual cogni­ tive assessments (6.3 vs. 6.1), t( 1.292) = —0.6, p = .548. The subgroups did not differ in gender (percentage women: 77.1 vs. 73.6), x2(l, N = 1,294) = 2.2, p = .138.

Clinical Evaluation At baseline and annually thereafter, participants had a structured clinical evaluation that included a medical history, cognitive test­ ing, and a neurologic examination. On the basis of this evaluation, an experienced clinician diagnosed dementia using 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 which require a history of cognitive decline and evidence of impairment in at least two cognitive domains (McKhann et ah, 1984). Individu­ als who met these criteria at baseline were not included in the present analyses.

Assessment of Cognitive Function Cognitive function was annually assessed with a battery of 17 individual tests in a session lasting approximately 1 hr. To reduce measurement error, particularly floor and ceiling artifacts, we used composite measures of two or more cognitive tests in longitudinal analyses. Supported by previous factor analyses (Krueger, Wilson,

Bennett, & Aggarwal, 2009; Wilson, Aggarwal et ah, 2009; Wil­ son, Barnes, & Bennett, 2003; Wilson et ah, 2005; Wilson et ah, 2002), we constructed measures of episodic memory, which were based on the following seven tests: immediate and delayed recall of Logical Memory Story A (Wechsler, 1987) and the East Boston Story (Albert et ah, 1991; Wilson et ah, 2002) and Word List Memory, Word List Recall, and Word List Recognition (Welsh et ah, 1994), semantic memory (three tests: 15-item Boston Naming Test [Kaplan, Goodglass, & Weintraub, 1983; Welsh et ah, 1994], a 15-item word reading test [Wilson et ah, 2002], and a measure of verbal fluency involving naming examples of animals and vege­ tables in 60-s trials [Welsh et ah, 1994; Wilson et ah, 2002]), working memory (three tests: Digit Span Forward, Digit Span Backward, and Digit Ordering [Wechsler, 1987; Wilson et ah, 2002]), perceptual speed (two tests: oral form of the Symbol Digit Modalities Test [Smith, 1982] and a modified version [Wilson et ah, 2002] of Number Comparison [Ekstrom, French, Harman & Kermen, 1976]), and visuospatial ability (two tests: 16-item ver­ sion of Standard Progressive Matrices [Raven, Court, & Raven, 1992] and a 15-item version of Judgment of Line Orientation [Benton, Si van, Hamsher, Varney, & Spreen. 1994]). We also constructed a composite measure of global cognition based on all 17 tests. Raw scores on individual tests were converted to z scores using the baseline mean and standard deviation in all three parent cohorts combined. The z scores of component tests were averaged to yield composite scores. Additional information on the individual tests and composite scores is contained in earlier publications (Wilson, Barnes, et ah, 2003; Wilson et ah, 2005; Wilson et ah, 2002 ).

Statistical Analysis We analyzed change in cognitive function in a series of mixedeffects models (Laird & Ware, 1982). Time was treated as years since baseline. Models included terms for age at baseline, sex, education, race, and their interactions with time. Random effects accounted for individual differences in the initial level of cognitive function and the rate of cognitive change over time. In separate subsequent models, we tested whether race modified the associa­ tions of age or education with cognitive trajectories. We assessed retest effects in two ways. First, we constructed a mixed-effects model that allowed the slope to change at some variable point. We constrained the change point to occur within 6 years of baseline to avoid the acceleration in cognitive decline known to occur in the last few years of life (Sliwinski et ah, 2006; Wilson, Segawa, Hizel, Boyle, & Bennett, 2012). Because some preliminary models with cognitive domain measures had difficulty converging, we used a previously established composite index of global cognition based on all 17 tests (Wilson et ah, 2002) to maximize measurement precision in the analysis. This approach allowed us to separate cognitive trajectories into components dif­ ferentially influenced by retesting and to test the relation of race to each component. Second, because the change point model sug­ gested that retest effects were strongest during initial follow-up evaluations, we repeated the core analyses, first eliminating the baseline evaluation and then again eliminating baseline plus the first year follow-up.

COGNITIVE AGING

281

Table 1 Relation of Demographic Variables to Trajectories of Change in Cognitive Functions Episodic memory

Semantic memory

Estimate

SE

P

Estimate

-0.048 -0 .0 3 4 -0.185 0.046 -0.177 -0 .0 0 6 -0.005 -0.003 0.010

0.006 0.002 0.036 0.004 0.031 0.001 0.009 0.001 0.008

Cognitive aging in older Black and White persons.

During a mean of 5.2 years of annual follow-up, older Black (n = 647) and White (n = 647) persons of equivalent age and education completed a battery ...
5MB Sizes 0 Downloads 10 Views