494 Adult Development and Aging
job stress (Muir, 1998; Williams et al., 1997) represent social
psychological risk factors for cardiovascular disease, and
that the lethal mechanisms include increases in BP, blood
cholesterol (Muir, 1998), and sympathetic and cardiovascular
responses (Williams, 1994). We need a systematic series
of studies that examine the cumulative impact of both
biological-cardiovascular and psychosocial-cardiovascular
risk factors on cognitive performance, and the impact of
aging on cognitive functioning.
METHODOLOGICAL CONSIDERATIONS
WHEN STUDYING AGING
The distinction between a risk factor for cardiovascular
disease and disease itself is dif“cult and often academic
(M. Elias et al., 2001). Cardiovascular diseases are risk fac-
tors for other diseases. Clinically de“ned hypertension is a
good example. Several overlapping de“nitions of the term
risk factor emerged early in the course of the Framingham
Heart Study (Kannel, Dawber, Kagan, Revortskie, & Stokes,
1961; Kannel & Sytkowski, 1987): (a) a correlate of cardio-
vascular disease, (b) a characteristic of an individual that pre-
disposes that individual to cardiovascular disease; and (c) a
factor that emerges as a cause of a cardiovascular disease.
Because associations between risk and cardiovascular dis-
ease are more easily demonstrated than causal relationships,
the “rst and second de“nitions have been employed more fre-
quently in the literature dealing with vascular risk factors for
cognitive decline. There is general agreement that variables
such as blood pressure, hypertension, diabetes, obesity, ciga-
rette smoking, and total cholesterol, among others, are risk
factors for disease.
However, a major conceptual problem is created because
age is itself a risk factor for cardiovascular disease. This has
implications for three of the most frequently employed analy-
ses in the health psychology of aging: (a) examine interac-
tions of age cohort status (or change over time, aging), and a
cardiovascular risk factor; (b) via regression or covariance
analyses, subtract the effects of aging on cognitive function-
ing from effects of CVD risk or disease; (c) subtract the
effects of CVD risk or disease from the effects of aging.
M. Elias et al. (1990) note that failure to render age a non-
signi“cant predictor of cognitive functioning by adjusting out
the in”uence of one or several risk factors is frequently cited
as evidence that risk for disease is unimportant with respect
to relations between aging and cognitive functioning. Such
conclusions re”ect a naive assumption that age or aging vari-
ables re”ect little more than primary aging (nondisease)
processes. The variable age in any study re”ects both primary
aging processes and all diseases and risk factors that are cor-
related with age.
The relative importance of age, versus Type II diabetes,
diastolic BP, and cigarette smoking as risk factors for poor
cognitive functioning is illustrated by data from the Framing-
ham Heart Study (M. Elias et al., 1998b, 2001). Beginning in
1950, all participants were screened for cardiovascular risk
factors and events every two years. All subjects were free of
history of de“nite completed stroke and Type I diabetes. All
were ages 55 to 85 at the time of neuropsychological testing.
A summary of the level of independent risk of lowered cog-
nitive functioning is shown in Table 21.1.
Thus, every “ve years of age produces an increased risk of
67% of declines in learning and memory, 61% in the com-
posite score, 44% on similarities, and 19% on attention/
concentration. This age risk, controlling for the very well-
measured disease and risk information, is the largest factor.
Whether one considers diabetes and hypertension as risk fac-
tors or comorbid conditions, they do have increasing effects
on cognitive decline.
In another set of analyses based on this same archival data
set, M. Elias et al. (2001) employed a risk factor scale that re-
”ected the cumulative impact of multiple risk factors on cog-
nitive functioning. These investigators designed a simple risk
factor scale that was used to determine the combined impact of
multiple cardiovascular risk factors, excluding aging. Subjects
were given a score (e.g., 0, 1, 2), depending on the number ofTABLE 21.1 Adjusted Odds Ratios of Performing At or Below the 25th Percentile on the Framingham Neu-
ropsychological Test Measurements (covariates included education, occupation, gender, alcohol consumption,
previous history of cardiovascular disease, and antihypertensive treatment)
Neuropsychological Type II Diabetes Diastolic BP Cigarettes/day Age in Years
Test (per 5 years) (per 10 mm HG) (per 5 cigs.) (per 5 years)
Composite score 1.21 1.30 1.04 1.61
Learning and memory 1.22 1.25 1.03 1.67
Similarities 1.19 1.01 1.09 1.44
Attention/concentration 1.00 1.15 0.98 1.19
BP exams 4...15; Diabetes exams 1...15; cigarettes/day at time of neuropsychological assessment; age ranged from 55 to 88
at time of neuropsychological testing.