University of Southern California Dissertations and Theses

Frequency and risk factors of poststroke dementia

(USC Thesis Other)

Frequency and risk factors of poststroke dementia

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content FREQUENCY AND RISK FACTORS OF POSTSTROKE DEMENTIA
By
Kecia Fumi Watari
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
December 2003
Copyright 2003 Kecia Fumi Watari
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
UMI Number: 3133351
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3133351
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This dissertation, written by
kecioi kMi'an'
under the direction o f h Q T dissertation committee, and
approved by all its members, has been presented to and
accepted by the Director o f Graduate and Professional
Programs, in partial fulfillment of the requirements for the
degree of
DOCTOR OF PHILOSOPHY
Director
Date Qppemher 1 7 , 2003
Dissertation Committee
Chair
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
A cknowledgements
I would like to acknowledge my advisor, Margy Gatz, who is my mentor,
teacher and role model. I feel very privileged to be one of her students. I would also
like to acknowledge my dissertation committee: Nancy Pedersen, Bob Knight,
Sandra Howell and Andy Johnson, for their feedback and wisdom. I am grateful for
the support of my peers at U.S.C., including the SCRAP lab, my dissertation support
group, and my very special friend, Krista Barbour. I would like to acknowledge my
family. I would not have had the confidence to even pursue a graduate degree
without their encouragement. Finally, I thank my husband for his patience. So there,
Margy!
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table of Contents
Acknowledgements...........................................................................................................ii
List of Tables..................................................................................................................... v
Abstract............................................................................................................................vii
1. Introduction.................................................................................................................1
Vascular Dementia............................................................................................... 2
Poststroke Dementia............................................................................................ 7
Vascular Factors...................................................................................................8
Hypertension.......................................................................................... 9
Diabetes...................................................................................................9
Behavioral Factors..............................................................................................11
Smoking.................................................................................................12
Alcohol................................................................................................... 13
Physical Activity................................................................................... 14
D ie t......................................................................................................... 16
2. Purpose and Hypotheses............................................................................................19
3. M ethod........................................................................................................................21
Population........................................................................................................... 21
Participants.........................................................................................................21
Base Sample.......................................................................................... 21
Old Cohort............................................................................................ 25
D esign................................................................................................................ 27
Matched Case Control.......................................................................... 27
Co-Twin Control...................................................................................31
M easures............................................................................................................ 33
Cognitive Screening..............................................................................33
Dementia............................................................................................... 35
Stroke.....................................................................................................36
Diabetes..................................................................................................41
Hypertension......................................................................................... 43
Demographics.......................................................................................44
Behavioral Factors............................................................................... 44
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Analyses............................................................................................................. 52
Stroke......................................................................................................52
Specific Aim I ........................................................................................ 52
Specific Aim I I .......................................................................................53
4. Results........................................................................................................................55
Stroke.................................................................................................................. 55
Prevalence R ates.....................................................................................55
Probandwise Concordance R ates.......................................................... 56
Specific Aim I: Matched Case Control Analyses............................................58
Descriptives............................................................................................. 58
Stroke as a Risk for Dementia................................................................58
Specific Aim I: Co-Twin Control Analyses..................................................... 62
Descriptives............................................................................................. 62
Stroke as a Risk for Dementia................................................................62
Specific Aim II: Matched Case Control Analyses...........................................64
Descriptives............................................................................................. 64
Stroke Characteristics..............................................................................65
Behavioral Factors...................................................................................66
5. Discussion...................................................................................................................76
Stroke.................................................................................................................. 76
Stroke as a Risk for Dementia.......................................................................... 76
Behavioral Factors Related to Poststroke Dementia.......................................84
Limitations and Strengths..................................................................................89
Conclusions and Implications........................................................................... 91
References........................................................................................................................94
Appendices
International Classification of Diseases (ICD) - Cerebrovascular Disease 107
The Swedish Twin Registry M easures................................................................. 108
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
V
List o f Tables and Figures
Table 1. Summary of Respondents and Non-respondents.................................. 23
Table 2. Characteristics of the Base Sample Participants (1998-2000),
N= 14,446................................................................................................ 24
Table 3. Comparison of Telephone Screening Respondents and Non-
Respondents, N=20,130 ........................................................................ 25
Table 4. Characteristics of the Old Cohort Participants with Stroke,
N=610......................................................................................................26
Table 5. Matched Case-Control Characteristics (Dementia): Means and
Standard Deviations for Cases (with Dementia, N=492) and
Controls (Intact, N=l,544)..................................................................... 29
Table 6. Matched Case-Control Characteristics (Dementia): Frequencies for
Cases (with Dementia, N=492) and Controls (Intact, N=1,544)........29
Table 7. Matched Case-Control Characteristics (Poststroke Dementia):
Means and Standard Deviations for Cases (with Poststroke
Dementia, N=51) and Controls (Poststroke, Intact, N=138).............. 31
Table 8. Matched Case-Control Characteristics (Poststroke Dementia):
Frequencies for Cases (with Poststroke Dementia, N=51) and
Controls (Poststroke, Intact, N =138)...................................................31
Table 9. Co-Twin Control Characteristics (Dementia): Means and Standard
Deviations for Twins Discordant for Dementia (N=T30 Like-Sex
P airs).......................................................................................................33
Table 10. Stroke Characteristics of Base Sample Participants with Stroke,
N= 1,290...................................................................................................39
Table 11. Stroke Characteristics of Old Cohort Participants, N=610................. 40
Table 12. Frequencies of Behavioral Risk Factors - Old Cohort (N=610).......45
Figure 1. Twin Pair Concordances for Stroke: Ascertainment of Twin Pairs ... 57
Table 13. Crude Odds Ratios (Matched Case-Control Design): Stroke as a
Risk for Dementia.................................................................................. 59
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
v i
Table 14. Adjusted Odds Ratios (Matched Case-Control Design): Stroke as
a Risk for Dementia................................................................................ 59
Table 15. Adjusted Odds Ratios (Matched Case-Control Design): Stroke as
a Risk for Dementia; Exclude Transient Ischemic Attacks (TIA).....62
Table 16. Crude Odds Ratios (Co-Twin Control Design): Stroke as a Risk
for Dementia............................................................................................63
Table 17. Adjusted Odds Ratios (Co-Twin Control Design): Stroke as a Risk
for Dementia; Exclude Transient Ischemic Attacks (TIA)................. 64
Table 18. Stroke Characteristics among Poststroke Participants........................65
Table 19. Univariate Conditional Logistic Regression Analyses (Matched
Case-Control Design): Stroke Risks for Poststroke Dementia...........66
Table 20. Univariate Conditional Logistic Regression Analyses (Matched
Case-Control Design): Behavioral Risks for Poststroke Dementia... 67
Table 21. Frequencies of Behavioral Factors for Cases (Poststroke Dementia,
N=51) and Controls (Poststroke, Intact, N=138)................................ 68
Table 22. Univariate Conditional Logistic Regression Analyses (Matched
Case-Control Design): Behavioral Risks for Poststroke
Dementia - Women................................................................................70
Table 23. Univariate Conditional Logistic Regression Analyses (Matched
Case-Control Design): Behavioral Risks for Poststroke
Dementia - Men.................................................................................... 71
Table 24. Adjusted Odds Ratios (Matched Case-Control Design): Behavioral
Risks for Poststroke Dementia.............................................................73
Table 25. Intercorrelations of Risk Factors and Confounding Variables
(N=189, Combined Poststroke Dementia and Intact)..........................75
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Abstract
The first purpose of this study was to investigate the risk of dementia
following a stroke, using a population-based twin sample and both classic case-
control and co-twin control designs. Second, mid-life behaviors were examined as
predictors of poststroke dementia. Participants included members of the Swedish
Twin Registry, aged 65 years and older, who participated in a comprehensive
telephone screening between 1998 and 2000. Those who screened positively for
cognitive impairment received comprehensive clinical evaluations for dementia.
Information about stroke was obtained from the Swedish Hospital Discharge
Register. Information about mid-life behaviors came from surveys of the twins in
1961 to 1970. In the case-control design (n=492 cases, n=l,544 controls),
conditional logistic regressions indicated that stroke (excluding transient ischemic
attacks) increased risk of dementia by 1.65 times, after adjusting for covariates. In
the co-twin control design («=130 like-sexed twin pairs), stroke was not related to
dementia. Although not significant, this finding was probably due to the limited
number of twin pairs where both partners had a stroke, as the odds ratios were
similar in the case-control and co-twin control designs. Among participants who had
a stroke, heavier smoking was related to greater risk of poststroke dementia, while
low to moderate amounts o f alcohol were related to a decreased risk o f poststroke
dementia. Overall, these findings suggest that for those who have suffered a stroke,
behaviors during mid-life may be influential in the development of dementia. Early
family environment and genetics appeared to have little effect in explaining which
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
individuals who had suffered strokes would subsequently become demented. Future
research further investigating behavioral determinants of poststroke dementia may be
useful in establishing preventive interventions, because once behavioral risks are
known, interventions to change behaviors can be designed.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1
1. Introduction
Vascular dementia is an increasingly relevant area of research, as it is one of
the only preventable types of dementia at the present time (Gorelick, 1997). Vascular
dementia is the second most common type of dementia in the United States and
Europe (Skoog, 1998) and the most common type in China and Japan (Jorm, 1991;
Li et al., 1989; Ueda, Kawano, Hasuo, & Fujishima, 1992; Yoshitake et al., 1995).
Surprisingly, the research base is much smaller than that of Alzheimer’s disease.
Vascular dementia is based on the recognition that cerebrovascular types of diseases
can cause brain injury and lead to cognitive impairment (Chui et al., 1992). There
are multiple etiologic pathways that can lead to vascular types of dementia, of which
the most commonly studied is stroke (i.e., poststroke dementia). However, the
understanding and diagnostic criteria of vascular dementia have changed over time,
resulting in inconsistent frequency rates and risk factors. Consequently, researchers
suggest focusing on a specific vascular mechanism leading to brain injury, such as
stroke. Much of the risk factor research for vascular and poststroke dementia has not
focused on behavioral factors. These factors may be more amenable to preventive
interventions, e.g., modifying diet or changing smoking habits, which suggests a role
for psychologists.
In the present study, stroke as a risk for dementia was examined using a
sample of older Swedish twins from a population-based twin registry, which allowed
both classic case-control and co-twin control designs. The co-twin design benefits by
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
controlling for some genetic and early environmental factors. In addition, the present
study benefited from longitudinal data, so that behaviors during mid-life, prior to
stroke onset, could be examined prospectively. The main purposes of the present
study were to: (1) examine the likelihood of dementia following a stroke using more
conservative research designs, and (2) examine mid-life behavioral risk factors for
poststroke dementia.
Vascular Dementia
Vascular dementia has become an increasingly relevant area of research
primarily because of the potential benefit from preventive interventions. Generally,
vascular dementia is based on the understanding that cardiovascular types of diseases
can cause brain injury leading to cognitive impairment (Chui et al., 1992). The
advancement of brain-imaging techniques revealed that several types of vascular
events result in cognitive impairment. The National Institute of Neurological
Disorders and Stroke (NINDS) and the Association Internationale pour la Recherche
et l’Enseignement en Neurosciences (AIREN) proposed a diagnostic criteria for use
with research studies (Roman et al., 1993). A diagnosis of vascular dementia requires
the following: (1) the presence of dementia, (2) evidence of cerebrovascular disease,
and (3) a temporal relationship between the cerebrovascular event and the onset of
dementia. Dementia is defined as a decline in memory and other cognitive or
intellectual abilities by which the decline interferes with daily function. The course
can either be progressive (e.g., leukoaraiosis) or stable (Roman et al., 1993). Other
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
3
conditions that impair ability, including physical impairment due to stroke are
excluded, as well as instances of delirium and altered consciousness. Decline should
be demonstrated through patient history and neuropsychological testing.
Cerebrovascular disease is defined “by the presence of focal neurologic signs
consistent with stroke” (Roman et al., 1993, p. 253).
There is more than one set of diagnostic criteria for vascular dementia,
however, which poses difficulty when comparing findings across studies. Other
criteria include: the Diagnostic and Statistical Manual of Mental Disorders (DSM-in,
DSMIII-R, DSMIV), the International Classification of Diseases, 10th revision
(ICD-10), and the State of California Alzheimer’s Disease Diagnostic and Treatment
Centers (ADDTC) (Chui et al., 1992). Although the Hachinski Ischemic Score
(Hachinski et al., 1975) is also used for diagnostic purposes, it was originally
designed to distinguish between degenerative types of dementia (e.g., Alzheimer’s
disease) and multi-infarct dementia. Comparisons of the various criteria resulted in
classifying different groups of patients with vascular dementia, suggesting that
comparing frequencies across criteria may not be appropriate (Chui et al., 2000;
Verhey, Lodder, Rozendaal, & Jolles, 1996; Wetterling, Kanitz, & Borgis, 1996).
The overall prevalence of dementia among adults aged 65 and older is about
4.5% (Chui, 1998), and the prevalence doubles every five years (Jorm, Korten, &
Henderson, 1987). In the United States and Europe, vascular dementia tends be the
second most common type of dementia after Alzheimer’s disease (Skoog, 1998) and
accounts for 10% to 50% of dementia cases (Skoog, 1994). The range reflects issues
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
4
of diagnostic criteria as much as genuine population differences. One study of 85-
year-olds living in Sweden reported higher rates of vascular dementia (47%) than
Alzheimer’s disease (30%) (Skoog, Nilsson, Palmertz, Andreasson, & Svanborg,
1993). In Japan and China, frequency rates are higher for vascular dementia than
Alzheimer’s disease (Jorm, 1991; Li et al., 1989; Ueda et al., 1992; Yoshitake et al.,
1995). Given its high prevalence, it is surprising that research has been lagging far
behind that of Alzheimer’s disease. For example, Hachinski (1992) called attention
to vascular dementia and questioned why more research has not been conducted,
especially when vascular dementia risks, such as stroke, are preventable.
One of the most widely studied vascular risk factors for dementia is stroke.
According to the National Institute of Neurological and Communicative Disorders
and Stroke, stroke is defined as “a sudden, nonconvulsive, focal neurologic deficit
persisting for >24 hours” (Foulkes, Wolf, Price, Mohr, & Hier, 1988, p. 548). More
generally, stroke has been described as “death of tissue in the central nervous system
owing to infarction or hemorrhage” (Gilroy, 1995, p. 3) and often referred to as a
“cerebrovascular accident” (Wiebe-Velazquez & Hachinski, 1991, p. 111). Stroke is
one of the leading causes of death and disability among older adults in the United
States (American Heart Association, 2000). The incidence of stroke over age 55
doubles each decade (American Heart Association, 2000). The age-adjusted
incidence rates of stroke (per 1,000 person years) were 4.44 among Black men, 3.10
among Black women, 1.78 among White men, and 1.24 among White women
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
5
(Rosamond et al., 1999). A U.S. study based on interviews from the National Health
and Nutrition Examination Surveys (NHANES) reported that the prevalence of
stroke in the U.S. among those aged 75 years and older was 11.3% (Muntner,
Garrett, Klag, & Coresh, 2002). In Sweden, the prevalence of stroke among adults
aged 75 years and older is 10% for men and 8% for women (Zhu et al., 1998).
The consequences of stroke depend primarily on severity of brain damage
(Weibe-Velazquez & Hachinski, 1991). Aside from a range of physical impairments
(e.g., paralysis, aphasia), stroke can cause cognitive (e.g., memory impairment) and
psychological impairments (e.g., depression). The estimated prevalence of vascular
dementia in community-based studies over the age of 60 is 2.6% for men and 2.1%
for women, which increases with age to about 16.3% among men in their 80’s and
9.2% among women in their 80’s (Leys, Parnetti, & Pasquier, 1999).
As mentioned above, a diagnosis of vascular dementia requires evidence of
cerebrovascular disease, so risk factors for vascular dementia overlap with risk
factors for stroke. Gorelick (1997) reviewed the risk factor research for dementia
associated with stroke, graded the quality of the studies, and concluded that
increasing age was the only well-documented risk for dementia associated with
stroke. Gorelick (1997) also rated other factors as strong contributors to dementia
associated with stroke, including ethnicity (i.e., Asian and African Americans), low
education, hypertension, smoking, myocardial infarction, diabetes, high cholesterol,
volume of cerebral tissue loss, number of infarctions, and ventricular size. Risk
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
factors that were rated as having poor evidence, due to lack of research or lack of
well-designed studies, included male sex, genetic factors, infarct location, white
matter lesions, and silent cerebral infarcts (Gorelick, 1997). A few other factors
have also been examined but were not mentioned in Gorelick’s review, including
heavy alcohol use, aspirin, stress early in life, blue collar employment, exposure to
pesticides or herbicides, and personality traits that increase vulnerability to stress
(Skoog, 1998). Aside from alcohol, the other factors have not been widely studied.
Few studies have examined the role of genetic influences on vascular
dementia, which is likely related to small numbers of twin pairs with vascular
dementia. However, one twin study reported no significant differences in pairwise
or probandwise concordance rates of vascular dementia between monozygotic and
dizygotic twin pairs, suggesting environmental factors may be more influential rather
than genetic factors (Bergem, Engedal, & Kringlen, 1997). However, a review of the
literature on the heritability of stroke, including family and twin studies, suggested
that there may be a genetic component (aside from autosomal dominant disorders)
(Plassman & Breitner, 1996).
The lack of consensus regarding risk factors for vascular dementia is likely
the result of methodological and diagnostic differences between studies and an
inability to control various confounding factors, such as education. Due to the
multiple etiologic pathways leading to vascular types of dementia and the present
limitations in conceptualization and diagnosis, researchers have suggested focusing
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
7
on a specific vascular mechanism leading to brain injury. This study focused
specifically on stroke as a risk for dementia or poststroke dementia.
Poststroke Dementia
Poststroke dementia refers to a subtype of vascular dementia in which
dementia is associated with stroke. There is no official diagnostic category for
poststroke dementia, but it is the most commonly studied type of vascular dementia.
The frequency of dementia among people with a history of stroke ranges from 14%
to 35% of people with stroke, and stroke survivors have an increased risk of
dementia of 4 to 12 times that of the general aging population (Andersen,
Vestergaard, Riis, & Ingeman-Nielsan, 1996; Barba et al., 2000; Censori et al., 1996;
Kokmen, Whisnant, O’Fallon, Chu, & Beard, 1996; Loeb, Gandolfo, Croce, &
Conti, 1992; Tatemichi et al., 1992). Among hospitalized stroke cohorts, dementia
occurs in about one-fourth to one-third of cases (Chui, 1998). These numbers differ
from those mentioned above for vascular dementia, in that these studies are limited
to stroke patients. There are several factors that contribute to the wide range of
frequency rates, of which varying diagnostic criteria and assessment protocols for
diagnosing poststroke dementia (e.g., time from stroke to dementia was not
consistent) and sample limitations (e.g., not population-based, small sample size)
should be considered. As with earlier studies of poststroke dementia, the present
study hypothesized that stroke would be related to an increased risk of dementia,
using a sample of older Swedish twins from a population-based twin registry.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
8
Although stroke increases the risk of dementia, not every person who suffers
from a stroke becomes demented. An examination risk and protective factors will
help improve our understanding of preventing this disease. Much of the risk factor
research for poststroke dementia (i.e., risks incurred above and beyond those of
stroke) has been focused in the biological and medical sciences. For example,
research on stroke-related features, such as severity (e.g., size of infarction, bilateral
lesions, etc.) and location of brain injury have been examined, as well as vascular-
related factors, such as hypertension and diabetes. In addition, demographic
variables, including advancing age and low education have been tested, with age
fairly well established as a risk for poststroke dementia. However, none of these
factors can independently explain why some people with stroke develop dementia
while others do not (Tatemichi et al., 1993). It is likely that multiple factors
contribute, which may include behaviors (e.g., smoking, diet); however, few have
specifically examined the role of behaviors.
Vascular Factors
In much of the research, many vascular factors (e.g., hypertension) were
examined as potential risks for poststroke dementia, primarily because of its
relationship with stroke. However, there has been little agreement between studies.
For example, Tatemichi et al. (1993) reported that diabetes increased risk o f
poststroke dementia, whereas hypertension and cardiac disease did not. In contrast,
Gorelick et al. (1993) reported that myocardial infarction increased risk of dementia,
hypertension decreased risk of dementia, and diabetes was not related to dementia.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
9
Regardless of discrepancies, it is likely that vascular factors may influence risk of
poststroke dementia, and controlling for their effect would decrease risk of
confounding results. In the present study hypertension and diabetes were included, as
they have been the most widely discussed in relation to poststroke dementia.
Hypertension. Many studies have examined the relationships between blood
pressure and cognitive decline or dementia (see Skoog, 1997 for a review), and blood
pressure and stroke (Gorelick et al., 1999). It is fairly well-established that reducing
blood pressure reduces risk of stroke (Collins et al., 1990). Hypertension during mid
life seems to be related to cognitive decline and dementia in older adulthood,
whereas the relationship between concurrent measures of blood pressure and
dementia has not been consistent. Among older adults who are already demented,
research suggests that blood pressure is actually low (Guo, Viitanen, Fratiglioni, &
Winblad, 1996). Hypertension has not been consistently supported as a risk for
poststroke dementia (Barba et al., 2000; Censori et al., 1996; Gorelick et al., 1993;
Inzitari et al., 1998; Kokmen et al., 1996). However, all of these studies considered a
previous diagnosis of hypertension or elevated current blood pressure as meeting the
diagnosis for hypertension, so distinction between past and current blood pressure
was not clarified. It appears relevant to not only account for hypertension, but also
time o f onset.
Diabetes. Diabetes, in particular non-insulin dependent or type 2 diabetes, is
a risk for stroke (Stegmayr & Asplund, 1995) and has been associated with dementia
(Hassing et al., 2002; Leibson et al., 1997; Ott et al., 1999). Ott et al. (1999)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
10
reported that older diabetic adults were twice as likely to become demented than
those who were not diabetic. A longitudinal study indicated that diabetes during mid
life was predictive of cognitive decline six years later (Knopman et al., 2001).
Diabetes may be linked to dementia via vascular and nonvascular means. For
example, Kumari, Brunner, & Fuhrer (2000) summarized some of the research on
diabetes and memory impairment, and reported that type 2 diabetes, which is
characterized by hyperinsulinemia — a metabolic syndrome, was related to cognitive
decline. Diabetes may also accelerate atherosclerosis - a vascular process (Gorelick
et al., 1999).
When compared to Alzheimer’s disease, diabetes has been associated with
vascular dementia more often than with Alzheimer’s disease (Boston, Dennis, &
Jagger, 1999; Hassing et al., 2002; Katzman et al., 1989; Nielson et al., 1996). One
study reported a threefold increased risk of vascular dementia among older Swedish
twins with type 2 diabetes (Hassing et al., 2002). Another study found an association
between abnormal glucose metabolism and vascular dementia 25 years later among
Japanese American men; this association was not significant for Alzheimer’s disease
(Curb et al., 1999). When considering treatment of type 2 diabetes, those at the
highest risk of vascular dementia have been insulin treated as opposed to oral
medication or no treatment (Ott et al., 1996; Ott et al., 1999). It was suggested that
insulin-treated diabetics are more severe and have been exposed to diabetes for a
longer period of time (Ott et al., 1999).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
11
Although there has been empirical support for diabetes as a risk for cognitive
decline, dementia, and vascular dementia, it is unclear whether stroke could account
for these observations. Ott et al. (1996) reported a significant effect for diabetes
after adjusting for stroke. In contrast, the relationship between diabetes and
poststroke dementia has been less consistent. A few studies provided support for
diabetes as a risk for poststroke dementia (Censori et al., 1996; Tatemichi et al.,
1993; Yoshitake et al., 1995), but many did not (Barba et al., 2000; Gorelick et al.,
1993; Inzitari et al., 1998; Kokmen et al., 1996; Loeb et al., 1992; Pohjasvaara et al.,
1998). However, identification of diabetes in these studies ranged from self-report,
to history of diabetes treatment, fasting glucose level prior to stroke, and fasting
glucose level after the stroke. If diabetes is both a risk for stroke and dementia,
including identification and onset of diabetes might help clarify some of these
discrepancies.
Behavioral Factors
There has been remarkably little study of behavioral factors, which represent
potentially modifiable risks. Psychology emphasizes behavioral factors in health
because once behavioral risks are known, interventions to change behaviors can be
designed. The present study examined the following behavioral factors as risks or
protective against poststroke dementia: Smoking, alcohol intake, diet, and physical
activity. These factors have in some capacity been related to both stroke and
cognitive impairment, which suggest that they may relate to dementia beyond risks
related to stroke alone.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
12
Smoking. Smoking has been related to an increased risk of stroke (Gorelick
et a., 1999; Shinton & Beevers, 1989) and cognitive impairment (Galanis et al.,
1997; Launer, Feskens, Kalmijn & Kromhout, 1996). Thus, it is plausible that it may
also increase risk of poststroke dementia. However, the relationship between
smoking and cognitive impairment has not been consistent. For example, a three-
year longitudinal study in Boston reported that smoking was not significantly related
to changes in cognitive functioning (as measured by a brief cognitive exam) among
current or former smokers (Hebert et al., 1993). In contrast, the Zutphen Elderly
Study of men also followed participants for three years and reported that smokers
made 20% more errors on a brief mental status exam than never smokers at baseline,
and the greatest cognitive decline was among those with cardiovascular disease or
diabetes (Launer et al., 1996). Further, smokers who were cardiovascular disease
free experienced almost no decline in cognitive functioning, whereas, never smokers
with cardiovascular disease experienced about the same degree of cognitive decline
as smokers. This suggests that those who are susceptible to or have a compromised
vascular system may be most affected by smoking (Laurner et al., 1996).
Several studies reported an association between smoking and vascular
dementia (Boston et al., 1999; Knopman et al., 2001), but it was not consistent across
all studies (Ott et al., 1998). One study reported that when normotensive older adults
diagnosed with multi-infarct dementia stopped smoking, cognition improved
compared to those who continued to smoke (Meyer, Judd, Tawakina, Rogers, &
Mortel, 1986). Current smoking has been related to an increased risk of poststroke
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
13
dementia (Gorelick et al., 1993; Pohjasvaara et al., 1998), but overall, smoking has
gained little empirical support as a risk for poststroke dementia (Barba et al., 2000;
Censori et al., 1996; Inzitari et al., 1998; Kokmen et al., 1996; Loeb et al., 1992;
Tatemichi et al., 1993; Yoshitake et al., 1995). Aside from difficulties associated
with varying measures of smoking, including onset, duration and quantity, it appears
that smoking should increase risk of dementia, especially among those poststroke.
Alcohol. Heavy amounts of alcohol have been associated with elevated blood
pressure, an increased risk of stroke, and even death (Smith, 1995). Alcohol may
also have negative effects on cognition, and can cause alcohol-induced dementia or
Wernicke-Korsakoff syndrome with prolonged heavy drinking (Smith & Atkinson,
1995). However, research examining the relationship between alcohol with vascular
and poststroke dementia has been subject to mixed findings. One study of Japanese
older adults, particularly among men, reported that moderate amounts of alcohol
were related to an increased risk of vascular dementia (Yoshitake et al., 1995).
Another study reported that Canadian older adults with a history of alcohol abuse
were at a greater risk for vascular dementia (Lindsay, Hebert, & Rockwood, 1997).
In contrast, a literature review by Gorelick et al. (1999) found several studies
supporting low to moderate amounts of alcohol as protective against dementia and
related with higher cognitive scores, compared to drinking excessive amounts or no
alcohol. Alcohol as a risk for poststroke dementia has gained little support (Barba et
al., 2000; Gorelick et al., 1993; Henon et al., 1997; Pohjasvaara et al., 1998;
Tatemichi et al., 1993).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
14
Although this research seems mixed, it was hypothesized that smoking and
heavy alcohol intake during mid-life would be related to an increased risk of
poststroke dementia, while moderate levels of alcohol would be protective.
Poststroke dementia studies may find different results if more specific measures of
alcohol intake and smoking were documented. The present study benefited from
prospective longitudinal data and multiple measures of smoking and drinking.
Physical Activity. It is generally accepted that exercise and physical activity
promote health benefits for older adults, but their effect on neuropsychological and
cognitive performance has been mixed (Bonner & Cousins, 1996; Okumiya et al.,
1996). Case-control studies tend to report that physically active older adults have
higher cognitive scores than sedentary older adults (Clarkson-Smith & Hartley,
1989; Dustman et al., 1990; Rogers, Meyer, & Mortel, 1990). In a longitudinal
community-based study in Canada, physical activity was protective against cognitive
impairment and dementia, especially Alzheimer’s disease and among women
(Laurin, Verreault, Lindsay, MacPherson, & Rockwood, 2001). In addition, a dose-
response relationship was observed in which increasing level of physical activity was
related to a decreasing risk of dementia. Similarly, Yoshitake et al. (1995) observed
that daily leisure activity and moderate physical activity during work was protective
against Alzheimer’s disease.
In contrast, few exercise intervention studies report cognitive gains
(Blumenthal et al., 1991; Blumenthal & Madden, 1988; Emery & Gatz, 1990; Hill,
Storandt, & Malley, 1993; Madden, Blumenthal, Allen, & Emery, 1989; McMurdo
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
15
& Burnett, 1992; Okumiya et al., 1996). However, they do tend to report
physiological improvements, such as increased aerobic capacity or oxygen intake,
relative to a non-aerobic or waitlist control group (Blumenthal et al., 1991; Hill et al.,
1993; Madden et al., 1989; McMurdo & Burnett, 1992; Okumiya et al., 1996). One
explanation for this discrepancy is that exercise programs need to be longer in
duration, lasting several years instead of months (Dustman, Emmerson, & Shearer,
1994). Most studies implement aerobic exercise groups that range from one exercise
trial to 12 weeks, and even up to about one year. Cognitive benefits from exercise
may require longer exercise training than is needed for physiological benefits.
Another explanation entails speculation that differences observed in cross-sectional
and correlational studies comparing active to sedentary adults might be attributed to
preexisting cognitive abilities associated with people who have adopted an overall
physically active lifestyle (Etnier et al., 1997).
Another explanation is that most intervention studies recruit healthy
community samples. Netz and Jacob (1994) suggest that exercise may produce less
of an effect in older adults who are healthy and active, whereas inactive or ill older
adults may receive greater benefit. Stones and Dawe (1993), for example, utilized
nursing home residents and found cognitive benefits after only one trial of exercise.
However, few have examined the effects o f exercise among nursing home or
institutionalized older adults. Physical activity or exercise may also be most
beneficial for older adults with cognitive impairment or dementia. Improvements in
memory were reported among cognitively impaired Japanese older adults who
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
16
engaged in walking exercises (Satoh, Sakurai, Miyagi, & Hohshaku, 1995). Gains in
attention, verbal ability, and mental status were also reported among Alzheimer’s
disease patients after 3 months of aerobic exercise (Palleschi et al., 1996). However,
both Satoh et al. (1995) and Palleschi et al. (1996) had small sample sizes and no
control groups.
Most of the epidemiological research has been conducted with Alzheimer’s
disease, yet physical activity may protect against vascular types of dementia given
that it is potentially protective against stroke (Hu et al., 2000). For example, regular
exercise was protective against vascular dementia among women in the Canadian
Study of Aging (Hebert et al., 2000). Physical activity may help sustain cerebral
circulation by decreasing blood pressure and lipid levels, thus slowing cognitive
decline (Yoshitake et al., 1995). The present study benefited from data about general
physical activity earlier in life and it is a population-based sample, which reduces the
likelihood of a self-selected sample (e.g., healthy volunteers interested in exercise).
Given the evidence supporting the relationship of physical activity to cognitive
impairment, it was predicted that a high level of physical activity during mid-life
would be protective against poststroke dementia. To our knowledge, no studies have
examined the relationship between physical activity and poststroke dementia.
Diet. Research on the relationship between diet and vascular dementia has
produced competing results. It seems logical to hypothesize that high amounts of
saturated fats and cholesterol should be related to an increased risk of vascular
dementia, as they have been related to cardiovascular disease (Kalmijn et al., 1997).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
17
In turn, polyunsaturated fatty acids have an antithrombotic effect, which may
decrease risk of vascular dementia. Kalmijn et al. (1997) examined the relationship
between fat intake and incident dementia using a prospective population-based
sample (the Rotterdam study). High levels of total fat and saturated fat were related
to an increased risk of dementia, specifically vascular dementia. Fish, a marker of
omega-3 polyunsaturated fatty acid, was related with a decreased risk of dementia,
especially Alzheimer’s disease. These relationships remained significant after
controlling for low education, smoking, and alcohol intake. Similarly, eating
shellfish was protective against vascular dementia in the Canadian Study of Health
and Aging (Hebert et al., 2000). Finally, among a sample of non-demented Swedish
men, obesity (based on body mass index) was related to poorer performance on
cognitive tests (Kilander, Nyman, Bobert, & Lithell, 1997).
In contrast, several studies suggested that a high fat diet is protective against
dementia. One study reported that preference for a Westernized diet, which
traditionally tends to be high in animal fat and protein while low in carbohydrates
was protective against vascular dementia among Japanese American men (Ross et
al., 1999). A traditional Japanese diet tends to be the opposite, low in animal fat and
protein, and high in carbohydrates. This study did not specify type of fat, but
saturated fatty acids are typically found in animal products. The Japanese
historically had high rates of stroke mortality, which has consistently declined over
the years, possibly due to the decreased rates of hypertension or Westernized
influence on diet, which may include increased protein and decreased salt intake
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
18
(Ross et al., 1999; Ueda et al., 1988). Moreover, rates of vascular dementia have
decreased with the rates of stroke among Japanese men (Kiyohara et al., 1994).
Another study reported that obesity was protective against vascular dementia among
a predominantly African American sample (Gorelick et al., 1993).
The relationship between a diet and vascular dementia may be related to the
relationship between diet and stroke. One study reported that a sample of middle-
aged men living in the United States (from the Framingham Heart Study) who
ingested high levels of fat, saturated fat and monounsaturated fats based on a 24-hour
dietary recall had a lower of risk of stroke (Gillman, Cupples, Millen, Ellison, &
Wolf, 1997).
Even though these findings appear contradictory and indicate a need for more
research, it is likely that modifying diet can possibly affect the development of
poststroke dementia. Studies that tend support a higher fat diet as protective against
vascular dementia are among primarily Japanese American and African American
samples. The sample in the present study was comprised of Swedish twins, so it was
predicted that a higher fat diet would be related to an increased risk of poststroke
dementia, while fish consumption would protect against poststroke dementia.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
19
2. Purpose and Hypotheses
There were two primary aims to the present study. First, the risk of dementia
following a stroke was examined, using a population-based twin sample and both
classic case-control and co-twin control designs. For both designs, it was predicted
that individuals with a history of stroke would be more likely to develop dementia
than those with no history of stroke. As summarized above, this hypothesis has been
well-established. However, in past studies the risk of dementia varied from 4 to 12
times those with no history of stroke. The present study examined the risk of
dementia using more stringent research designs, specifically the co-twin design.
Twin data are a unique and valuable resource for research that allows for
investigation of some genetic and environmental influences (Lichtenstein et al.,
2002). The co-twin control method benefits by providing implicit control over
potential confounding effects of early rearing and family environments (e.g.,
socioeconomic status). Comparison of risks using the classic case-control and co
twin control designs helps clarify whether an exposure effect is related to early life
environmental factors or the actual exposure.
Second, given a stroke, mid-life behaviors were examined as predictors of
dementia. These behaviors included smoking, alcohol use, diet, and physical activity.
It was predicted that after controlling for potential covariates, including age,
education, hypertension and diabetes, heavy smoking and a high fat diet during mid
life would increase risk of poststroke dementia. Excessive alcohol intake would also
increase risk of poststroke dementia, while moderate levels would be protective.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 0
Finally, high physical activity and consumption of greater amounts of fish were
predicted as protective against poststroke dementia. Targeting behaviors can provide
effective ways of reducing overall risks of poststroke dementia. Before behaviors can
be targeted, good evidence is needed as to which behaviors are most likely to have
protective effects.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
21
3. Method
Population
For the first aim (which examines stroke as a risk for dementia), the present
study used data from the Swedish Twin Registry (STR), including all members who
were 65 years and older and alive at the time of the comprehensive telephone
screening in 1998 to 2000. This is referred to as the base sample. For the second aim
of the study (which examines mid-life behaviors as predictors of dementia),
members from the base sample who were same sexed pairs, born before 1926, and
who were positive for stroke were included.
Participants
Base Sample. The STR was first established in 1959 to study the effects of
smoking and alcohol on cancer and cardiovascular disease (Lichtenstein et al., 2002).
It is a population-based registry of all twins residing in Sweden and consists of
several cohorts. For this study, of relevance are those bom between 1886 and 1925
(“old cohort”) and those born between 1926 and 1958 (“middle cohort”).
In 1961, self-report questionnaires were mailed to all living like-sexed twin
pairs in the old cohort. Questionnaires were again sent in 1963 and 1967, and to a
select group in 1970 (those who did not respond to the 1967 questionnaire and their
partners). Broadly, the questionnaires inquired about twin similarity, demographics
(e.g., marital status, education), smoking and drinking habits, diet, health and illness.
Similar questionnaires were mailed to the middle cohort in 1972-1973, with
additional questions on personality and environmental exposures.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 2
From March 1998 to May 2001, twins from the STR (regardless of the
gender composition of the pair) who were still alive and living in Sweden were
contacted to participate in a telephone screening. Among like-sexed twins, this
follow up occurred 27 to 32 years after their initial contact either in the old or middle
cohort (Lichtenstein et al., 2002). The 1998-2001 telephone screening consisted of a
computer assisted telephone interview, which was conducted by trained interviewers
with a medical background. Every month, a random sample of 1,000 twin pairs was
screened by telephone. Telephone numbers were obtained by linking their name and
address to the telephone company’s files. The interview consisted of detailed
questions regarding birth and twin information, occupation, education, medical and
medication history, psychiatric history, and lifestyle behaviors (see Lichtenstein et
al., 2002 for a detailed summary). Cognitive status was also evaluated. If one of the
twins in a pair committed a series of errors (determined by a scoring algorithm) in
the cognitive screening, the proband (i.e., individual who failed the screening) and
his or her twin partner were referred for a dementia workup.
The base sample of the present study included 20,210 individuals aged 65
years and older when available for the telephone screening and screened between the
years of 1998 and 2000. Data on stroke history were obtained from the Swedish
Hospital Discharge Register, which is one of the national health care registries
matched to the STR. The Hospital Discharge Register included discharges through
December 31, 2000, so twins interviewed after that date were excluded («=76
individuals screened in 2 0 0 1 ).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
Individuals may not have participated in the telephone screening for a variety
of reasons, including: Refusals to participate (22%, «=4,317), they were not
reachable (6 %, n= 1,171), or they died between selection and screening ( 1 %, n=216).
In addition, a number of participants were reached by telephone but could not be
screened, because they were too confused or cognitively impaired (1%, «=183), too
sick (1%, «=T91) or unable to hear (1%, n=258) and there was no proxy available to
complete the proxy form of the screening interview. Overall response rate was 71%,
including those who were screened and those who were reached by telephone but
could not complete the interview (Table 1).
Table 1
Summary o f Respondents and Non-respondents
Respondents N Non-respondents n
Base
Sample
TOTAL
"Finished" 13,734
NIT Cognitively Impaired 183 Refused 4,317
NIT too sick 191 Not Reachable 1,171
NIT impaired hearing
Previous Dementia Study
258
80
Died 276
TOTAL 14,446 TOTAL 5,764 2 0 , 2 1 0
Note. NIT = twins who were reached, not interviewable.
Additional participants from the base sample of the present study were not
contacted for the telephone screening because they had already been screened in an
earlier dementia study (w=80). Their screening results were included in the present
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 4
study, but were not counted toward the response rate, resulting in 14,446
participants. Overall, average age of base sample participants was 73.84 (SD=6.83),
and 56% were women. There were 24% monozygotic twins, 42% dizygotic twins,
32% twins from unlike sex pairs, and 2% whose zygosity was unknown (Table 2).
Table 2
Characteristics o f the Base Sample Participants (1998-2000), N= 14,446
% n
Gender Men 44 6,315
Women 56 8,131
Age 65-69 33 4,738
70-74 26 3,789
75-79 2 1 2,935
80-84 1 2 1,773
85-89 6 8 8 6
90+ 2 325
Zygosity Monozygotic 24 3,337
Dizygotic 42 6,018
Unlike Sex 32 4,679
Unknown 2 312
Note. Age = age at the time of the telephone screening.
There were significant differences in characteristics between respondents and
non-respondents (Table 3). Response rates excluded those screened in the earlier
dementia study. Response rates were significantly higher for men than women.
Among men, 74% responded; among women, 69% responded. Responders were
significantly younger than non-responders; however, this difference was less than
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
half a year. Average age of responders was 73.80 (*SZ)=6.81), and average age of
non-responders was 74.22 (SD=6.75). There were also significant differences in
zygosity, in particular, only 27% of twins of unknown zygosity participated in the
study, likely reflecting their lack of involvement in any twin registry data collection.
Among the others, 76% of monozygotic, 75% of dizygotic and 71% of unlike sex
pairs were participants.
Table 3
Comparison o f Telephone Screening Respondents and Non-Respondents, N =20,130
Respondents
(n = 14,366)*
%
Non-Respondents
(n = 5,764)
%
x2
Gender Men 44 38 53.13***
Women 56 62
l l
Zygosity Monozygotic 24 19 1170.56***
Dizygotic 41 35 df= 3
Unlike Sex 33 32
Unknown 2 14
t
Age Mean 73.80 74.22
2 9 7 ***
SD 6.81 6.75
Note. Age = age at the time of the telephone screening. * For response rates, n
excluded participants screened in an earlier dementia study («=80). ***/?< .0 0 0 1 .
Old Cohort. For the second aim of the study, participants included members
from the base sample who were same sexed pairs born before 1926, responded to the
STR questionnaires about relevant risk factors, participated in the telephone
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 6
screening, and positive for stroke. Information about early to midlife cardiovascular
disease history, alcohol use, smoking, diet, and physical activity were obtained from
the 1961-1970 STR questionnaires (Old Cohort). Of the 14,446 participants in the
base sample, 5,050 (35%) were bom before 1926 and responded to 1961-1970 STR
questionnaires; of which 610 participants had a stroke. Thus, 610 participants were
included in the second specific aim. Average age was 80.09 years (SD=5.04), and
56% were women. There were 34% monozygotic twins, 64% dizygotic twins and
2% whose zygosity was unknown (Table 4).
Table 4
Characteristics o f the Old Cohort Participants with Stroke, N — 610
% n
Gender Men 44 267
Women 56 343
Age 70-74 8 49
75-79 40 246
80-84 29 174
85-89 17 106
90+ 6 35
Zygosity Monozygotic 34 209
Dizygotic 64 392
Unknown 2 9
Note. Age = age at the time o f the telephone screening.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 7
Design
Matched Case-Control Design. A classic matched case-control design was
used for the first and second aims of the study, which compares twins with a disease
to external controls (unrelated twins without the disease) and estimates risk of the
disease given a particular exposure (Lichtenstein et al., 2002). In the present study,
twins with dementia (cases) were compared to unrelated cognitively intact twins
(controls), who were matched on gender and age at the time of the telephone
screening. Among complete twin pairs, one member of the pair was randomly
selected, to ensure that cases were not related, controls were not related, and cases
and controls were not related to each other. Controls were randomly chosen from the
pool of twins who screened intact and were matched four controls to one case, by
age (using two-year age bands) and gender.
For the first specific aim, which examined stroke as a risk for dementia, 585
participants in the base sample were diagnosed with dementia (4%), of which 493
were singletons and 92 were pairs. One participant from every pair was randomly
selected, resulting in 539 individuals with dementia. Among the cognitively intact
participants (u=10,327), 4,441 were singletons and 5,886 were pairs. After
randomly selecting one participant from each pair, 7,384 intact individuals remained.
When cases and controls were merged, cognitively intact twins among pairs
discordant for dementia were excluded (»=134), resulting in 539 individuals with
dementia and 7,250 cognitively intact individuals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 8
Participants were organized into two-year age bands from age 65 through 90.
Individuals over the age of 90 were excluded from analyses, as cases (n=47)
outnumbered controls (n=46). Within each gender, four controls for every case
within an age band were randomly chosen from the remaining pool of cognitively
intact twins. However, among older age groups, there were more cases and fewer
controls, resulting in a fewer number of controls matched to cases. Among women,
in the age 83 to 84 age band, three controls were matched to one case; in the 85 to 8 6
and 89 to 90 age bands, one control was matched to one case; and in the 87 to 8 8 age
band, two controls were matched to one case. Among men, in the 85 to 8 6 and 87 to
8 8 age bands, two controls were matched to one case; and in the 89 to 90 age band,
three controls were matched to one case.
The final sample included 492 dementia cases and 1,544 cognitively intact
controls (Table 5 and 6 ), matched on age and gender. Among cases, average age was
80.66 (SD=5.92), and 63% were women. There were 27% monozygotic twins, 46%
dizygotic twins, 23% opposite sex twins, and 4% whose zygosity was unknown.
Among controls, average age was 78.95 (SD=5.61), and 59% were women. There
were 27% monozygotic twins, 47% dizygotic twins, 25% opposite sex twins, and 1%
whose zygosity was unknown.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Table 5
Matched Case-Control Characteristics (Dementia): Means and Standard Deviations
for Cases (with Dementia, N=492) and Controls (Intact, N=1,544)_______________
Dementia Cases Intact Controls
M SD n M SD n
Age (years) 80.66 5.92 492 78.95 5.61 1,544
Education (years) 7.48 2.48 373 8.38 2.91 1,529
Note. Age = age at the time of the telephone screening.
Table 6
Matched Case-Control Characteristics (Dementia): Frequencies fo r Cases (with
Dementia, N=492) and Controls (Intact, N=1,544)_______________________
Dementia Cases Intact Controls
% N % n
Gender Women 63 312 59 904
Men 37 180 41 640
Zygosity Monozygotic 27 133 27 412
Dizygotic 46 224 47 726
Unlike Sex 23 114 25 391
Unknown 4 2 1 1 15
For the second specific aim, which examined mid-life behaviors as predictors
of poststroke dementia, 610 participants were included, who were in the old cohort
and positive for a history of stroke. Of these, 55 were diagnosed with dementia (9%);
all were singletons. Among the other 555 participants, 310 screened as cognitively
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 0
intact, including 282 singletons and 14 pairs (28 individuals). After randomly
selecting one participant from each pair, 296 intact individuals remained. When
cases and controls were merged, there were 55 demented and 296 cognitively intact
unrelated individuals. Aside from those aged 73 to 75, participants were organized
into two-year age bands from age 76 through 90. As with the first aim, participants
over age 90 were excluded (n=16). Within each gender, four controls for every case
within an age band were randomly chosen from the remaining pool of intact twins,
with the exception of older age groups. Among women, in the 85 to 8 6 , 87 to 8 8 , and
89 to 90 age bands, one control was matched to one case. Among men, in the 85 to
8 6 age band, two controls were matched to one case; in the 87 to 8 8 and 89 to 90 age
bands, one control was matched to one case.
The final sample included 51 dementia cases and 138 controls, matched on
age and gender (Table 7 and 8 ). Among cases, average age was 82.82 (*SD=4.59),
and 59% were women. There were 39% monozygotic twins, 59% dizygotic twins,
and 2% whose zygosity was unknown. Among controls, average age was 80.58
(579=4.05), and 50% were women. There were 38% monozygotic twins, 61%
dizygotic twins, and 1 % whose zygosity was unknown.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
Table 7
Matched Case-Control Characteristics (Poststroke Dementia): Means and Standard
Deviations fo r Cases (with Poststroke Dementia, N=51) and Controls (Poststroke,
Intact, N=138)__________________________________________________________
Poststroke
Dementia Cases
Poststroke
Intact Controls
M SD n M SD N
Age (years) 82.82 4.59 51 80.58 4.05 138
Education (years) 6.76 1.79 45 9.02 3.26 136
Note. Age = age at the time of the telephone screening.
Table 8
Matched Case-Control Characteristics (Poststroke Dementia): Frequencies fo r
Cases (with Poststroke Dementia, N=51) and Controls (Poststroke, Intact, N=138)
Poststroke Poststroke
Dementia Cases Intact Controls
% n % n
Gender Women 59 30 50 69
Men 41 2 1 50 69
Zygosity Monozygotic 39 2 0 38 52
Dizygotic 59 30 61 85
Unknown 2 1 1 1
Co-Twin Control Design. A co-twin control design was used in the first
specific aim only, which involved comparisons between twins discordant for a
disease (use of internal controls) to estimate risk of that disease given a particular
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 2
exposure (Lichtenstein et al., 2002). In the first specific aim, twins who were
discordant for dementia were examined. Analyses included complete twin pairs from
the base sample («=4,534 pairs), of which, 286 individuals were diagnosed with
dementia. Twins who were concordant for dementia were excluded (n= 92
individuals, 46 pairs), resulting in 194 pairs who were discordant for dementia. Pairs
were considered as discordant if both members of the pair participated in the clinical
diagnosis procedure and one member of the pair was diagnosed as demented while
the other member of the pair was diagnosed as not demented. Pairs were also
included if one member of the pair was diagnosed as demented and the twin partner
did not participate in clinical diagnosis but was cognitively intact when screened
over the phone. Fifty-one partners were ascertained to be cognitively intact when
screened and did not participate in complete clinical diagnosis. Nine partners of
demented twins did not participate in the clinical diagnosis nor were ascertained to
be cognitively intact during the screening. Therefore, these pairs were excluded from
analyses, resulting in 185 twin pairs who were discordant for dementia. Finally,
unlike sex pairs were excluded, which resulted in 130 same sex twin pairs who were
discordant for dementia (n=42 male pairs, n = 8 8 female pairs). Average age was 79
years (SD=6.5) for both groups (Table 9). There were 25% monozygotic twins
(n=46), 43% dizygotic twins («— 80), and 2% whose zygosity was unknown {n -4).
For the second specific aim, a co-twin design was not performed because there were
no twin pairs who were concordant for stroke and discordant for dementia.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
Table 9
Co-Twin Control Characteristics (Dementia): Means and Standard Deviations for
Twins Discordant for Dementia (N=130 Like-Sex Pairs)______________________
Dementia Cases Intact Controls
M SD n M SD N
Age (years)
Education (years)
79.05
8.30
6.47 130
2.65 103
79.22
7.93
6.50
2.71
130
115
Note. Age = age at the time of the telephone screening.
Measures
Cognitive Screening. Cognitive screening was based on a combination of
information from the twin's performance on the telephone cognitive screening
instrument (Gatz, Reynolds, Nikolic, Lowe, Karel & Pedersen, 1995; Gatz,
Reynolds, John, Johansson, Mortimer & Pedersen, 2002) and/or the informant's
report on the Blessed Dementia Rating Scale (Blessed, Tomlinson, & Roth, 1968) of
how much the twin's cognitive status interfered with daily functioning. The
cognitive screening included the 10-item Mental Status Exam (MSQ; Kahn, Pollack,
& Goldfarb, 1961) which assessed gross cognitive functioning and orientation. There
were three additional questions from the Similarities subtest on the Swedish
Wechsler Adult Intelligence Scale (Jonsson & Molander, 1964), three-item recall,
and serial three subtractions. The informant interview assessed health and daily
functioning, as well as the Blessed Dementia Rating Scale.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 4
A scoring algorithm determined whether a participant should be referred for a
dementia workup. Scores ranged from 0 to 3, indicating increasing level of
impairment. A score of 0 indicated that the twin completed the cognitive screening
with 5 or fewer errors and no domain impaired. A score of 1 was assigned when the
twin completed the cognitive screening with impairment in one domain (MSQ,
similarities, or 3-word recall). Scoring 2 indicated that the twin was unable to
complete the interview or performed poorly (failed a third of the items or was
impaired in more than one domain), but number of errors on MSQ was less than 3,
and either there was no informant available or the informant did not judge the twin to
be impaired. A score of 3 was assigned if the twin was unable to complete the
interview or performed poorly (failed a third of the items or was impaired in more
than one domain) and the Blessed Dementia Rating Scale score was 1.5 or higher or
informant directly reported cognitive impairment affecting the twin's daily
functioning or the informant specifically mentioned dementia. A score of 3 could be
obtained if the number of errors on MSQ equaled or exceeded 3 or if the twin
admitted to memory problems requiring help with daily activities, even without an
informant's confirmation.
Twins scoring 0 or 1 were considered cognitively intact for purposes of
identifying normal controls, and twins scoring 3 were considered cognitively
impaired for purposes of referring for a dementia workup. The telephone screening is
a validated tool for identifying cases of dementia (Gatz et al., 1995; 2002).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 5
Dementia. When a twin was referred for the dementia workup, a trained
doctor and nurse team assessed the twin in his or her residence. Nurses administered
the cognitive battery and conducted interviews with an informant regarding a twin’s
current functioning and observed behavior changes. In addition, the doctors
performed a neurological exam and complete physical workup. The nurses drew the
blood, which was sent to a lab to do the blood panel. Medical records were obtained
and reviewed by the doctor.
The neuropsychological assessment included the Mini-Mental State Exam
standardized from the Canadian Study of Health and Aging (Molloy, & Standish,
1997) and a cognitive battery. The cognitive battery included sections from the
Consortium to Establish a Registry of Alzheimer’s Disease neuropsychological
battery (CERAD) (Morris et al., 1989), subtests from the Swedish Wechsler Adult
Intelligence Scale (Jonsson & Molander, 1964), a prospective memory test (Huppert,
Johnson & Nickson, 2000), and the Memory in Reality test (Swedish “apartment
test”) (Johansson, 1988/89).
A knowledgeable informant was administered a structured interview
regarding the twin’s cognitive ability and current functioning. Parts of the interview
were based on the Clinical Dementia Rating scale (Morris et al., 1997). Onset and
pattern o f cognitive decline were also established (Gatz et al., 2002). Finally,
behavioral changes associated with dementia were assessed using the
Neuropsychiatric Inventory (Cummings et al., 1994).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 6
Clinical diagnoses of dementia were based on the Diagnostic and Statistical
Manual of Mental Disorders, Fourth Edition (American Psychiatric Association,
1994), and participants were categorized as demented, questionable, or intact.
Diagnoses were determined by a diagnostic review board, consisting of a
psychologist and neurologist who specialize in dementia. Each reviewed the
assessment results and made diagnoses independently. Agreement of dementia
diagnoses was evaluated for 204 twins. Weighted kappa was .87, suggesting good
agreement between them. Therefore, the remaining cases were diagnosed by only
one judge, with the option to consult with the other if there were any disagreements.
For the case-control design, if a participant was referred for the dementia
workup and diagnosed with dementia, they were coded as cases. If a participant
scored a 0 or 1 in the cognitive screening they were coded as controls. For the co
twin design, if a partner was diagnosed with dementia, they were coded as cases.
Since twin partners were systematically referred for a dementia workup when one of
the twins failed the cognitive screening, partners who were diagnosed as not
demented or who scored a 0 or 1 on the cognitive screening were coded as controls.
Stroke. The Swedish Hospital Discharge Register was used to determine
stroke status. The Discharge Register records all public and private hospital
admissions and discharges, with 94% accuracy (Lindblad et al., 1993). Data include
demographic (e.g., gender, age of admission), administrative (e.g., date of admission
and discharge, hospital), and medical (e.g., diagnoses) information. The present
study included diagnostic data beginning in 1967, with complete coverage of all
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 7
public hospitalizations from 1987 through the year 2000. Within our sample, there
were relatively few strokes from the first years of the registry. While it is possible
that few strokes were reported due to limited coverage during the 1970’s, the pattern
that we observed would also be expected based on survival effects. Twins now in
the sample who had strokes longer ago would be likely to have their strokes at a
younger age. Those who did have strokes at an older age are not likely to have
survived to be included in the population. For example, among twins in our base
sample who were diagnosed with stroke between 1969 and 1986, the average age of
stroke was 58 years. Among twins diagnosed between 1987 and 1996, the average
age of stroke was 71 years; and between 1997 and 2000, the average age was 76.
Since the Discharge Register spans across 40 years, various revisions of the
International Classification of Disease (ICD) were used for diagnoses, including the
ICD-7 (prior to 1969), ICD- 8 (1969-1986), ICD-9 (1987-1996), and ICD-10 (1997-
2000). For the ICD7, 8 , and 9 versions, codes for cerebrovascular diseases were
between 430 through 438, and included hemorrhages, occlusions, transient cerebral
ischemia, and ill-defined cerebrovascular diseases. For the ICD-10, codes for
cerebrovascular disease were between 160 to 169, and codes for transient ischemic
attacks were classified as G45. Appendix A lists ICD codes for stroke and
corresponding diagnoses.
Stroke was defined as any hospital discharge with a primary diagnosis of
cerebrovascular disease or transient ischemic attack. The primary diagnosis was
generally defined as the reason a person came to the hospital. The data were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 8
organized by hospital discharge, i.e., each hospital discharge was a new entry or row
in the data file. The data were transposed so that each row represented a twin and
each hospital admission was represented in a column. Data from the most recent
stroke diagnosis was retained (e.g., date of discharge), and if multiple strokes
occurred, dates of those events were noted. Data were limited to twins who
participated in the telephone screening prior to December 31, 2000, which totaled
1,332 twins with one or more strokes. If the first stroke occurred after the onset of
dementia, participants were coded as negative for stroke (n=42). If a first stroke
occurred during the same year of dementia onset, participants were coded as positive
for stroke («=18).
This resulted in a total of 1,290 twins (9% of the base sample participants)
who were discharged with a primary diagnosis of one or more strokes, and the stroke
occurred prior to dementia onset. Average age at the time of stroke was 76.34
(«SD=6.77) and 50% were women (n=649). There were 23% monozygotic twins,
42% dizygotic twins, 33% twins from unlike sex pairs, and 2% whose zygosity was
unknown. Sixty-one percent were discharged once with a primary diagnosis of
stroke, 21% had two strokes, and 10% had three strokes. The remaining 8 %
experienced a range of 4 to 13 discharges with stroke. Type of the most recent stroke
included occlusions (53%), transient ischemic attacks (23%), “not otherwise
specified” strokes (13%), and hemorrhages (11%). The “not otherwise specified”
category included those who were diagnosed with a “cerebrovascular accident not
otherwise specified,” “other cerebrovascular disease” or “sequelae of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3 9
cerebrovascular disease.” Of those with stroke, 1 0 % were diagnosed with dementia.
Table 10 summarizes stroke characteristics.
Table 10
Stroke Characteristics o f Base Sample Participants with Stroke, N= 1,290
% N
Stroke Type Occlusions 53 680
Transient Ischemic Attacks 23 301
Hemorrhages 1 1 141
Not Otherwise Specified 13 168
Number of Strokes 1 61 781
2 2 1 272
3 1 0 136
4-13 8 1 0 1
Zygosity Monozygotic 23 292
Dizygotic 42 549
Unlike Sex 33 422
Unknown 2 27
Note. Number of strokes is defined as the number of hospital discharges with a
primary diagnosis of stroke.
For the second aim, which examined behavioral predictors of poststroke
dementia, all participants were positive for stroke. Of the 610 old cohort participants,
61% (n -371) were discharged once with a diagnosis of stroke, 21% (w=129) had two
strokes, 10% (»=59) had three strokes and the remaining 8% («=51) had a range o f 4
to 13 strokes. The most recent stroke type included 54% occlusions (n=332), 23%
transient ischemic attacks (n=138), 13% “not otherwise specified” strokes (n=80),
and 10% hemorrhages (n=6 0). An estimated 15% of the sample was diagnosed with
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 0
dementia. Table 11 summarizes stroke characteristics of participants from the old
cohort.
Table 11
Stroke Characteristics o f Old Cohort Participants, N=610_____________
% N
Stroke Type Occlusions 54 332
Transient Ischemic Attacks 23 138
Hemorrhages 13 80
Not Otherwise Specified 1 0 60
Number of Strokes 1 61 371
2 2 1 129
3 1 0 59
4-13 8 51
Note. Number of strokes is defined as the number of hospital discharges with a
primary diagnosis of stroke
The Discharge Register appears to be a valid measure of the occurrence of a
stroke, as over 90% of Swedish people who have had a stroke sought help from a
hospital (Alfredsson et al., 1986), and the Hospital Register diagnoses are estimated
to be 94% accurate (Lindblad et al., 1993). In a previous study, self report of stroke
was compared to the Swedish Hospital Discharge Register (see Watari, Gatz, &
Pedersen, 2002). There was moderate agreement between the Register and self
report {kappa = .62), so the Hospital Register alone was used to measure the
occurrence of a stroke. Among participants of whom the Hospital Register indicated
there was no history of stroke, only 2 % of participants reported one.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
Diabetes. History of diabetes was obtained because it is commonly
researched in relation to poststroke dementia. Twins were asked about diabetes
(including old age diabetes, excluding pregnancy diabetes) in the 1998-2000
telephone screening, specifically, status and date or age of onset. Disease status was
based on self-report and supplemented with medical records or the proxy interview
obtained through the telephone screening. Telephone screening questions are
illustrated in Appendix B.
Participants were coded as positive for diabetes if they reported having
diabetes in the telephone screening and onset was before the first stroke event. If
negative for stroke, onset of diabetes was before the telephone screening. Anyone
who developed diabetes after their first stroke was coded as negative for diabetes
(«=42 of 1221 who were positive for diabetes in the telephone screening). Among
cognitively intact participants, diabetes status was largely determined by their
telephone screening responses. Medical records were used to supplement missing
data in less than 1% of the sample. However, fewer participants with dementia had
self report data (through the telephone screening), resulting in utilizing proxy
information and medical records in approximately 25% of those participants. The
proxy was asked about diabetes status (i.e., whether a participant was diabetic), but
not time of onset. Medical records, which were available only among those who
participated in the dementia workup, were used to establish onset.
An additional analysis was conducted to examine the validity of self report of
diabetes status. Overall, there was substantial agreement between self report during
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 2
the telephone screening and medical records, with 96% agreement. Of the 879 twins
who responded to the telephone screening and provided medical records, 842 agreed
(kappas.84). Among those who disagreed, 3% (n=29) denied having diabetes in the
telephone screening but their medical records indicated they that were diabetic; only
1 % (n=8 ) reported being diabetic, whereas their medical records did not support this.
Self-reported answers could not be compared to a proxy report, because a proxy was
only used when participants were too cognitively impaired to respond.
One methodological limitation was the inability to reliably determine who
had Type I vs. Type II diabetes, as Type II is usually associated with both stroke and
dementia. Although diabetes type was asked of the participants, they frequently
answered that they did not know what type of diabetes they had or refused to answer.
However, it is fairly rare to have onset of Type I diabetes after age 20. In addition,
insulin was not available when these participants were under 20 years. Less than
0.1% (n= 18) of our sample reported age of onset younger than 20 years. Thus,
participants who indicated they were diabetic, regardless of type, were coded as
positive for diabetes in the present study.
Overall, 9% of the base sample participants were diagnosed with diabetes
(n=1,242). Among those with dementia, 16% («=6 6 ) were diabetic; among those
who were cognitively intact, 9% (n=937) were diabetic. For the second specific aim
(i.e., examination of behavioral factors related to poststroke dementia), 1 1 % of the
sample were diagnosed with diabetes («=61). Among those with dementia, 8 %
(n=3) had diabetes; among those who were intact, 11% (»=34) had diabetes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Hypertension. Blood pressure history is also commonly researched in
relation to poststroke dementia. Twins were asked about blood pressure in the
telephone screening and the 1963 STR questionnaire; however, onset information
was not available. For the first aim, the STR questionnaire could not be relied on
alone, as it was only asked of the old cohort resulting in 32% of the base sample with
data on blood pressure (n= 4,564). STR responses could not be supplemented with
telephone screening responses because time of hypertension onset could not be
established. Thus, hypertension was not included in the analyses for the first specific
aim examining stroke as a risk for dementia. Hypertension was included in the
second specific aim, using only the 1963 STR questionnaire responses (old cohort).
Of the second specific aim participants, 91% responded to the hypertension question.
Three percent («=17) of the old cohort participants with stroke were hypertensive in
mid-life. Among those with dementia, 4% (n=2) had hypertension; among those
who were intact, 2% (n=9) had hypertension. These data undoubtedly underestimate
hypertension because it could have developed between 1963 and the occurrence of
stroke. Average year of stroke onset was 1995.
Nonetheless, self-report of hypertension without onset posed a problem.
Examination of self-report of hypertension during the telephone screening (which
asks about current or past blood pressure levels) indicated that a greater proportion of
non-demented participants were hypertensive than demented participants. It is likely
that our respondents reported current blood pressure status, lending support to earlier
observations that older adults with dementia are typically hypotensive (Guo et al.,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 4
1996). Additionally, Pohjasvaara et al. (1998) examined determinants of poststroke
dementia, and reported no relationship between history of hypertension and
poststroke dementia, but lower blood pressure levels at the time of evaluation were
related to poststroke dementia.
Demographics. Demographic data included age at the time of the telephone
screening, gender, and education level. Education was defined as the number of
years of education. In the present sample, years of education ranged from 0 to 20.
For the analyses of risk factors, education was reversed, so that higher numbers
reflected lower education levels.
Behavioral Factors. Information on midlife behavioral risk factors was
obtained from the 1961-1970 STR questionnaires (old cohort), which included
alcohol use, physical activity, smoking, and diet. Risk factors were analyzed in the
second specific aim only (n=610). Sample sizes differed depending upon each risk
factor (Table 12).
Alcohol intake was assessed in 1961, 1967 and 1970. For the present study,
alcohol intake was measured as the quantity of various types of alcohol consumed
(i.e., beer, wine, or hard liquor), drinking status and lifetime use. Quantity of alcohol
was defined as grams of pure alcohol per month, which was based on an algorithm
comprised o f frequency o f alcohol consumption for each type o f alcoholic beverage
and the amount consumed per sitting.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
Table 12
Frequencies o f Behavioral Risk Factors - Old Cohort (N=610)
% n
Alcohol Status Abstainer 2 0 108
Former Drinker 2 8
Current: <600 g/mos 76 410
Current: 600 + g/mos 3 14
Lifetime Abstainer 15 58
Former 28 106
Drinker 57 216
Physical Activity Status Hardly Any 1 0 54
(leisure) Easy 71 402
Regular 1 0 59
Demanding 9 53
Physical Activity Status Sedentary 18 75
(work) Active 56 238
Physically Strenuous 26 1 1 1
Smoking Quantity Nonsmoker 54 326
Former 15 92
Current: Cigar/pipe only 9 51
Current: 2-10 1 1 67
cigarettes/day
Current: 11+ 6 37
cigarettes/day
Current: 5 28
Cigarettes+pipe/cigar
Lifetime Nonsmoker 51 2 0 2
Former 41 164
Smoker 8 33
Diet Fish & Seafood No part 1 7
Small Part 39 198
Medium Part 51 254
Great Part 9 43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
Table 12 (continued)
% n
Fruits & Vegetables No part 1 5
Small part 24 124
Medium part 56 288
Great part 19 96
Beef No part 1 7
Small part 31 155
Medium part 63 311
Great part 5 2
Pork No part 2 8
Small part 32 165
Medium part 61 313
Great part 5 25
Sausage No part < 1 3
Small part 24 119
Medium part 69 349
Great part 7 35
Cakes & Pastries More seldom 33 173
Once/day 47 251
Several times/day 2 0 106
Candy More seldom 85 445
Once/day 13 6 8
Several times/day 2 13
Drinking status included the categories of: (1) abstainer, (2) former drinker,
(3) current drinker (light, less than 600 grams per month) and (4) current drinker
(intermediate to heavy, 600+ grams per month). Drinking status was first determined
through a twin’s response to the 1967 and 1970 questionnaires, which included:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 7
abstainer, presently a drinker or previously a drinker. If a twin responded only once,
their status was based on their response for that one year. If the twin responded
multiple times, their status was based on the combination of their responses. When a
twin responded that they were currently a drinker, quantity of alcohol (grams/month)
was used to create the categories of “light” or “intermediate to heavy.” Divisions of
drinking status were based on a previous study which examined alcohol and
myocardial infarction using the STR (Hammar, Romeljso & Alfredsson, 1997).
However, in the previous study, alcohol quantity was reported as grams per day as
opposed to grams per month in the present study. Hammar et al. (1997) did not find
differences between heavy (>900 grams of alcohol per month) and intermediate
drinkers (600 to 899 grams per month) due to very few cases of heavy drinkers.
Similar proportions were noted in the present study, so intermediate and heavy
drinkers were combined into one category.
Lifetime alcohol use was measured using STR (1961-1970) responses and the
telephone screening. If a twin reported abstaining from drinking alcohol in the STR
and in the recent telephone screening, they were coded as abstainers. Twins were
coded as former drinkers if they reported drinking alcohol at any point of contact, but
denied drinking alcohol in the past year (telephone screening). Finally, twins were
coded as “lifetime” drinkers if they reported drinking in the STR and again in the
telephone screening. An additional six twins were also coded as “lifetime” drinkers,
which included twins who responded as former drinkers or nondrinkers in the STR,
but reportedly drank alcohol within one year of the telephone screening.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 8
Average quantity of alcohol consumed per month during middle age (STR)
was 125.93 (SD= 197.78, n=526). Most participants were classified as light drinkers
(76%) or abstainers (20%). The remaining participants consisted of former drinkers
(2%) or intermediate to heavy drinkers (3%). Lifetime alcohol use consisted of 15%
abstainers, 28% former drinkers and 57% drinkers.
Due to a fairly positively skewed distribution, responses were compared to
telephone screening responses for consistency. First, the telephone screening
included a measure of alcohol abuse history, based on a scoring algorithm applied to
a number of questions that asked about a time when participants were drinking
heavily or the most they had in their lives. Only 1% of our sample («=4) were
classified as alcohol abusers, and 48% endorsed some of the questions but were not
classified as abusers (n=295). The remaining 311 participants did not receive an
alcohol abuse score, which suggests that they did not endorse having a period in their
life when they drank too much, someone else did not indicate they drank too much,
or that they drank instead of working. Based upon the observation that very few
participants in our sample were classified as alcohol abusers, it appeared that the
observed skewed distribution in the STR was a reliable indicator of alcohol use
during mid-life.
Physical activity during leisure and at work was assessed in 1967 and 1970.
Leisure physical activity was defined as “How much exercise have you basically had
between the ages of 25 and 50?” Answer options ranged from 1 to 4, which
included: ( 1 ) hardly any, (2 ) light exercise (example: regular walks or easy yard
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 9
work), (3) regular sports-related exercise, and (4) demanding physical workouts. If a
twin answered both questionnaires in 1967 and 1970, their two ratings were averaged
(31%, n=\12). Within this sample, 71% engaged in light exercise, 10% in hardly
any exercise, 10% in regular sports-related exercise, and 9% in demanding physical
workouts.
Physical activity at work was measured as one of three categories, which
included: (1) sedentary, (2) active, and (3) physically strenuous. If a twin answered
both questionnaires in 1967 and 1970, their two ratings were averaged. Within this
sample, 56% were physically active at work, 26% engaged in physically strenuous
work, and 18% were sedentary at work.
Smoking was assessed in 1961, 1967 and 1970, and was defined as the
average quantity of tobacco intake and lifetime use of tobacco. Smoking status was
first determined, which included (1) nonsmokers, (2) former smokers, or (3) current
smokers. If a twin responded only once, their status was based on their response for
that one year. If the twin responded multiple times, their status was based on the
combination of their responses. For example, if a twin consistently responded all
three years that they did not smoke or did smoke, they were categorized as
nonsmokers or current smokers, respectively. If a twin responded that they smoked
in 1961, but not in 1967 or 1970, they were categorized as former smokers. Quantity
of smoking combined current smoking status, type (e.g., cigars, cigarettes), and
quantity smoked. Categories included: (1) nonsmoker, (2 ) former smoker, (3) current
smoker, only cigars and/or pipe, (4) current smoker, 2 through 10 cigarettes per day,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 0
(5) current smoker, 11+ cigarettes per day, (6 ) current smoker, cigarettes plus cigar
and/or pipe.
Lifetime use of tobacco was measured using STR (1961-1970) responses and
the telephone screening. Twins who reported not smoking during the STR and in the
telephone screening were coded as “lifetime” nonsmokers. If a twin reported being a
former smoker in the STR and a nonsmoker now or if they had reportedly smoked at
any point of contact but denied smoking in the telephone screening, they were coded
as former smokers. Finally, if a twin reported being a smoker in the STR and during
the telephone screening they were coded as “lifetime” smokers. Of note, the
“lifetime” smoker category included one twin who was categorized as a previous
smoker and another twin who was categorized as a nonsmoker in the STR, but a
current smoker in the telephone screening.
Overall, most participants were nonsmokers (54%) or former smokers (15%)
during middle age. Among the smokers, 11% smoked 2-10 cigarettes per day, 9%
smoked only cigars and/or a pipe, 6 % smoked 11 or more cigarettes per day, and 5%
smoked cigarettes plus cigars/pipe. Regarding lifetime tobacco use, 51% reported
never smoking, 41% were formed smokers and 8 % were current smokers. There
were fewer responses for lifetime tobacco use relative to smoking quantity because
the lifetime tobacco use category required twins to answer in both the STR 1961-
1970 questionnaire and the telephone screening.
As with alcohol, STR responses were compared to telephone screening
responses for consistency. The telephone screening asked participants whether they
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
had ever smoked or used snuff. If a participant indicated that they never “even tried
it,” or “only tried it,” they were coded as “nonsmokers” (n=227). If a participant
indicated that they “smoked now and then” or “smoked regularly,” they were coded
as “former smokers” (n=139). If a participant indicated that they “smoke now and
then” or “smoke regularly,” they were coded as “current smokers” («=33). When
responses were compared to smoking status in the STR, 70% were in complete
agreement, 24% were plausible scenarios (e.g., nonsmoker in the STR, but former or
current smoker in the telephone screening; current smoker in the STR but former
smoker in the telephone screening), but 6% disagreed (e.g., current smoker in the
STR and nonsmoker in the telephone screening). Thus, in most cases STR responses
were consistent with the telephone screening, in that most participants were either
nonsmokers or former smokers. However, it is possible that participants may have
changed their smoking habits prior to their stroke and would be misclassified based
on their STR responses.
Diet was assessed in 1967 and defined as the proportion of consumption of
various food types, including pork, sausage, beef, fish and seafood, and fruits and
vegetables. Answer options ranged from 1 to 4: (1) no part, (2) small part, (3)
medium part, and (4) great part of one’s diet. In addition, frequency of
cakes/pastries and candy consumption were assessed, with answer options: ( 1) more
seldom, (2) once a day, and (3) several times a day. Animal products (i.e., pork,
sausage, beef) and foods high in refined sugars (i.e., cakes and pastries, candy) tend
to be comprised of saturated fatty acids, whereas fish are higher in polyunsaturated
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
fatty acids. In the present study, information on how food was prepared or cooked
was unavailable, which can affect how fatty or lean a food might be considered. So,
food types were used in the analyses, instead of creating categories of “saturated
fats” or “polyunsaturated fats.” Frequencies for each individual food type are
reported in Table 12 (above).
Analyses
Analyses for descriptive information on stroke and for each specific aim are
outlined below. All statistical analyses were run using the SAS statistical software
Version 8.02.
Stroke. Stroke was examined as a descriptive variable. Given the present
sample is derived from a population-based twin registry, prevalence rates of stroke
for monozygotic and dizygotic twins were calculated, as well as the frequency of
stroke in our specific samples. Second, probandwise concordance rates were
calculated for monozygotic and dizygotic pairs separately, which is the probability
that a partner of an affected twin (i.e., positive for stroke) will also be positive for
stroke. When concordance rates are higher among monozygotic pairs relative to
dizygotic pairs, it is an indicator that genetic influences may contribute to stroke
occurrence. Finally, characteristics of stroke among twins with poststroke dementia
were reported, which included number o f years from stroke to dementia onset, stroke
type and number of strokes.
Specific Aim I. To examine stroke as a risk for dementia, both classic
matched case-control and co-twin control designs were implemented using the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
PROC PHREG procedure in SAS. For the matched case-control design, cases (with
dementia) were compared to controls (intact), stratified for age at the time of
screening and gender. The first set of analyses was primarily descriptive, consisting
of univariate conditional logistic regressions for age at the time of screening, low
education and diabetes status. Next, conditional logistic regression analyses (crude
and adjusted) examined whether stroke exposure differed between cases and
controls. Additional analyses were conducted to further investigate the findings,
specifically risk of dementia excluding TIA’s and for each gender. Odds ratios with
95% confidence intervals (Cl) were calculated. Covariates included education and
history of diabetes.
For the co-twin control design, twin pairs were discordant for dementia,
which is comparable to a matched case-control design, with implicit matching for
unmeasured early life environment and some genetics. Conditional logistic
regression analyses were used to calculate odds ratios for stroke as a risk factor with
95% confidence intervals (both crude and adjusted for covariates), as well as
separate analyses that exclude TIA’s and for each gender.
Specific Aim II. To examine behaviors related to poststroke dementia, a
classic matched case-control design was implemented using the PROC PHREG
procedure in SAS. Cases, twins with a history o f stroke who were diagnosed with
dementia (i.e., poststroke dementia) were compared to unrelated twins with a history
of stroke who screened intact (controls). These analyses were divided into three
parts. The first set of analyses was primarily descriptive and consisted of univariate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
conditional logistic regressions, stratified on age and gender. Predictors included
demographic variables, history of hypertension and diabetes, stroke-related factors
(i.e., stroke type, number of strokes) and behaviors. The second set of analyses
consisted of multivariate conditional logistic regression models, using only
significant variables identified in the univariate analyses to reduce the likelihood of
obtaining spurious results. The best fitting model was determined by the -2 Log
Likelihood estimates, and the significance of the model was evaluated using a chi-
square statistic. Analyses were also run separately for each gender. Odds ratios were
obtained from conditional logistic regression analyses, with 95% confidence
intervals that matched for age at the time of screening and gender, as well as adjusted
for education, history of diabetes and history of hypertension. Third, correlation
coefficients of the primary variables of interest and potential confounds were
reported to provide information on interrelationships between variables and to help
further explain the results. The PROC CORR procedure was used to calculate
correlations.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
4. Results
Stroke
Prevalence Rates. Within the base sample, history of stroke (including TIA)
was present in 1,290 twins (9%) of the base sample participants. Stroke was
prevalent among 10% of men and 8% of women. For the matched case-control
analyses examining stroke as a risk for dementia, 16% (n=79) of cases and 12%
(«=177) of controls were positive for stroke. These percents are higher than the
percent in the base sample because the case control sample is older than the total
base sample. When participants who had a transient ischemic attack (TIA) were
excluded as positive for stroke, 13% (n=64) of cases and 8% («=127) of controls
were positive for stroke. For the co-twin control analyses, 13% (n= 17) of cases and
8% (/?=10) of controls were positive for stroke. After excluding TIA’s, 11% (n=14)
of cases and 4% (n=5) of controls were positive for stroke.
For the second aim (which examines behaviors as predictors of poststroke
dementia), a co-twin design could not be tested because there were no pairs who
were concordant for stroke and discordant for dementia, resulting in further
investigation of these limited numbers. Overall prevalence rate of stroke within the
present sample of Swedish twins was estimated at 8% for monozygotic and 8% for
dizygotic twins. These percents are lower than the base sample o f 9% because a
greater number of complete monozygotic and dizygotic twin pairs were women than
men, whose prevalence rate was 8% in the base sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
Probandwise Concordance Rates. Probandwise concordance rates were
calculated for monozygotic and dizygotic twin pairs. Dizygotic twin pairs included
both same-sex and opposite-sex twin pairs. Figure 1 outlines the process of
ascertaining twin pairs and the calculation of probandwise concordance rates. For
each (monozygotic or dizygotic), only twin pairs were included and classified as
either concordant or discordant for stroke. Among monozygotic twins, only 16 pairs
were concordant for stroke; among dizygotic pairs, only 17 pairs were concordant for
stroke. In this sample, the frequency of stroke was lower among individuals in
complete pairs (8%) than individuals in singletons (11%), suggesting that stroke
might have been underestimated because partners to a twin with stroke may have
died resulting in fewer complete pairs.
Probandwise concordance rate of stroke for monozygotic twin pairs was
18.9% and probandwise concordance rate for dizygotic pairs was 11.9%. Overall,
both concordance rates were low; however, a higher concordance rate among
monozygotic twins relative to dizygotic twins indicates that stroke events may be
influenced in part by genes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
Figure 1
Twin Pair Concordances fo r Stroke: Ascertainment o f Twin Pairs
N = 14,446
(BASE SAMPLE)
Included
Excluded
MZ
n = 3,437
individuals
DZ
n = 6,018
individuals
PAIRS
n = 1,120
(2,240
individuals)
PAIRS
n = 1,803
(3,606
individuals)
MZ PAIRS
(TOTAL PAIRS=1,120)
Twin A
Twin B
No
Stroke Stroke
No Stroke 967 60
Stroke 77 16
UNK
n = 312
individuals
OPP
n = 4,679
individuals
SINGLES
MZ
n = 1,197
individuals
SINGLES
DZ
n = 2,412
individuals
DZ PAIRS
(TOTAL PAIRS=1,803)
Twin A
Twin B
No
Stroke Stroke
No
Stroke
1,535 134
Stroke 117 17
Probandwise Concordance Rate = (2 X pairs concordant for stroke) / (2 X pairs
concordant for stroke) + (pairs discordant for stroke)
MZ : (16 X 2) / (2 X 16) + 77 + 60 = 18.9%
DZ: (17 X 2 )/(2 X 17)+ 127+ 134 = 11.9%
Note. MZ = Monozygotic, DZ = Dizygotic, OPP = Opposite sex, UNK = Unknown
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Specific Aim I: Matched Case Control Analyses
Descriptives. Comparisons were made between the dementia cases and
unrelated intact controls on demographics and diabetes status. Case-control analyses
included 492 participants with dementia (cases) and 1,544 intact participants
(controls), who were matched on age and gender and excluded any participant older
than age 90. Univariate conditional logistic regression analyses were applied to age
at the time of screening, low education and diabetes status. Demented participants
were more likely to be older (OR = 1.40, 95% CI= 1.13-1.72), have fewer years of
education (OR = 1.13, 95% Cl = 1.08-1.19) and have a history of diabetes (OR =
1.88, 95% C /= 1.36-2.60), relative to intact participants. On average, cases were
80.7 years old and had 7.5 years of education; while controls were 79.0 years old and
had 8.4 years of education. Even though they were matched on age, the differences
in age were due to fewer available controls in the older age bands. Among cases,
17% (n=62) were diabetic relative to 10% (n= 154) of the controls.
Stroke as a Risk fo r Dementia. Conditional logistic regressions were applied
to estimate odds ratios. The crude or unadjusted risk of developing dementia after
stroke was 1.49 (95% Cl = 1.11-2.00) relative to participants with no history of
stroke. Table 13 lists the unadjusted odds ratios for stroke and other descriptive
variables. After adjusting for age, education, and diabetes, the risk o f developing
dementia after stroke was 1.57 times relative to participants with no history of stroke
(95% Cl — 1.12-2.42). All other covariates also remained significant risks for
dementia, which included older age (OR = 1.33), low education (OR = 1.12) and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
history of diabetes (OR = 1.76). Table 14 lists the adjusted odds ratios and 95% Cl’s
for each variable.
Table 13
Crude Odds Ratios (Matched Case-Control Design): Stroke as a Riskfor Dementia
N= 492 cases, 1,544 controls_____________________________________________
Odds Ratio 95% Confidence Interval
Stroke 1.49 1.11-2.00
Low Education 1.13 1.08-1.19
Age 1.40 1.33 - 1.72
Diabetes 1.88 1.36-2.60
Note. Education is reverse coded.
Table 14
Adjusted Odds Ratios (Matched Case-Control Design): Stroke as a Riskfor
Dementia
N = 492 cases, 1,544 controls
Odds Ratio 95% Confidence Interval
Stroke 1.57 1.12-2.20
Low Education 1.12 1.06-1.18
Age 1.33 1.04-1.69
Diabetes 1.76 1.25-2.48
Note. Education is reverse coded.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 0
It was unexpected that the risk of dementia after stroke would be higher after
adjusting for covariates. Separate conditional logistic regressions were performed
with stroke and each covariate (i.e., age, education and diabetes). When stroke and
diabetes were entered into a regression model alone, the risk of dementia increased
to 1.87 (95% Cl = 1.60-2.23). Further logistic regressions examining stroke as a risk
for dementia were performed for diabetics and non-diabetics separately. Among
diabetics, the risk of dementia following a stroke dropped below significance
(OR=Q.77, 95% CI= 0.30-2.03). However, among non-diabetics, the risk of
dementia following stroke was increased to 1.79 (95% CI= 1.26-2.54), suggesting
that diabetics may have died before they could develop dementia.
This finding prompted an exploratory analysis of whether diabetes posed a
risk for stroke (with unconditional logistic regression using PROC LOGISTIC). For
these analyses, twins with a history of stroke were compared to twins with no history
of stroke (from the base sample). After adjusting for age, education and gender,
diabetes was related to an increased risk of stroke (OR= 1.30, 95% C/=1.07-1.58).
This appears to further support the logic that diabetes is a risk factor for stroke, and
that stroke is a risk for dementia, but diabetics who suffer a stroke may die before
developing dementia. Therefore, diabetes as a risk factor for dementia may be
underestimated. These results also explain why the odds ratio for stroke as a risk for
dementia increases when diabetes is controlled for.
To further test stroke as a risk for dementia, conditional logistic regressions
were applied separately for men and women. Among men, the unadjusted risk of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
dementia after stroke was 1.66 (95% C7=1.09-2.54). Among women, the unadjusted
risk of dementia after stroke was 1.34 (95% CI=0.89-2.03). Twenty-two percent
(n=39) of men with dementia had a history of stroke, relative to 15% (n=93) of men
who were intact. For women, 13% («=40) of those with dementia and 9% (n=84) of
those who were intact had a history of stroke. Similar to the findings mentioned
above, stroke was not related to an increased risk of dementia among diabetic men
(OR=.95, 95% CI=.24-3.69) or diabetic women (OR=.64, 95% CI=. 16-2.53), but
was related to an increased risk of dementia among non-diabetic men (OR= 1.91,
95% C/= 1.52-3.14) and women (OR=1.68, 95% C /= l.03-2.75).
Conditional logistic regressions excluding those who had a transient ischemic
attack (TIA) (w=65 excluded) were applied to estimate odds ratios. The results were
similar. The unadjusted risk of developing dementia after a stroke was 1.67 (95% Cl
= 1.21-2.32), which remained significant (OR=1.65, 95% Cl = 1.13-2.42) after
adjusting for covariates (Table 15). All other covariates also remained significant
risks for dementia, which included older age (OR = 1.33), low education (OR = 1.12)
and history of diabetes (OR = 1.77).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
Table 15
Adjusted Odds Ratios (Matched Case-Control Design): Stroke as a Risk fo r
Dementia; Exclude Transient Ischemic Attacks (TIA)
N= 492 cases, 1,544 controls_______________________________________
Odds Ratio 95% Confidence Interval
Stroke 1.65 1.29-2.42
Low Education 1.12 1.06-1.18
Age 1.32 1.04-1.69
Diabetes 1.77 1.26-2.50
Note. Education is reverse coded.
Specific Aim I: Co-Twin Control Analyses
Descriptives. Comparisons on demographic and vascular risk factors were
conducted between 130 like sexed twin pairs discordant for dementia. Univariate
conditional logistic regressions were applied to education and history of diabetes.
There were no significant differences in education (OR = 0.98, 95% CI= 0.86-1.12)
or diabetes status (OR = 1.38, 95% Cl = 0.55-4.42), when cases (with dementia)
were compared to controls (intact). Overall, average age at the time of screening was
79 years, and average education level was about 8 years. Among cases, 16% (n=\6)
had diabetes relative to 12% (n= 14) of controls.
Stroke as a risk fo r dementia. Conditional logistic regressions were applied to
estimate odds ratios. The unadjusted risk of developing dementia after stroke was
1.70 (95% C /= 0.78-3.71). Among men, the unadjusted risk of dementia after stroke
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
was 4.50 (95% C/=.97-20.83); among women, the unadjusted risk of dementia after
stroke was 1.00 (95% C/=.38-2.66). Twenty-one percent (n=9) of men with
dementia had a history of stroke relative to only 5% (n=2) of men who were intact.
In contrast, 9% («=8) of women with dementia and 9% («=8) of women who were
intact had a history of stroke. Unadjusted odds ratios for stroke and each covariate
are reported in Table 16. Analyses adjusting for covariates were not run, because
stroke alone was not related to a greater risk for dementia using the co-twin control
design.
Table 16
Crude Odds Ratios (Co-Twin Control Design): Stroke as a Risk fo r Dementia
T V = 130 Pairs Discordant for Dementia
Odds Ratio 95% Confidence Interval
Stroke 1.70 0.78-3.71
Low Education 0.98 0.86-1.12
Diabetes 1.38 0.55-3.42
Note. Education is reverse coded.
Conditional logistic regressions excluding those who had a TIA (n= 14
excluded) were performed on the same model. The unadjusted risk of developing
dementia after a stroke was 2.80 (95% Cl — 1.01-7.77). After adjusting for
covariates, the risk of dementia after stroke was not significant relative to twins with
no history of stroke (OR = 2.25, 95% CI = 0.69-7.33) (Table 17). Among men, the
unadjusted risk of dementia after stroke was 8.00 (95% C/= 1.00-63.96); among
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
women, the unadjusted risk of dementia after stroke was 1.50 (95% C/=0.42-5.33).
The differential influence of diabetes on stroke as a risk for dementia could not be
evaluated, because there were no twins who had a stroke and were also diabetic.
Table 17
Adjusted Odds Ratios (Co-Twin Control Design): Stroke as a Riskfor Dementia;
Exclude Transient Ischemic Attacks (TIA)
N= 116 discordant pairs________________________________________________
Odds Ratio 95% Confidence Interval
Stroke 2.25 0.69-7.33
Low Education 0.97 0.85-1.11
Diabetes 1.18 0.46-3.01
Note. Education is reverse coded.
Specific Aim II: Matched Case-Control Analyses
Descriptives. Matched case-control analyses included 51 participants with
poststroke dementia (cases) and 138 participants with stroke who screened intact
(controls), matched on age and gender. Univariate conditional logistic regressions
were applied to age at the time of screening, low education, history of diabetes,
history of hypertension, stroke-related factors and behaviors. Participants with
poststroke dementia were more likely to have fewer years o f education (O R = \A \,
95% CI= 1. 15-1.73) relative to participants who were poststroke and intact, but there
were no significant differences in age or history of diabetes and hypertension.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
Stroke Characteristics. Stroke characteristics were compared between
poststroke dementia and poststroke intact participants. Descriptively, average age of
the most recent stroke among twins with poststroke dementia was 77.82 (SD = 6.21)
and 45% had one stroke. Among those with poststroke dementia, most (76%)
developed dementia within 5 years of their stroke, of which 48% developed dementia
within one year. The average number of years from the first stroke to dementia onset
was 3.39 years (51)=3.94). Among twins who were poststroke and intact, average
age of the most recent stroke was 77.24 (SD = 6.81) and 37% had more than one
stroke. Average number of years from the first stroke to time of the telephone
screening was 4.41 (SD=5.20). Table 18 summarizes stroke characteristics.
Table 18
Stroke Characteristics among Poststroke Participants
N = 51 cases, 138 controls______________________________________________
Poststroke Poststroke
Dementia Cases Intact Controls
% n % N
Years from Stroke to
Dementia Onset 0 - 1 48 24
1 .5 -3 18 9
3 .5 -5 10 5
5 .5 -7 6 3
7 .5 -9 6 3
9.5-11 6 3
11.5-13 4 2
13.5-14 2 1
Number of Strokes 1 stroke 45 23 37 51
More than 1 55 28 63 87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 6
Stroke-related factors were not related to an increased risk of dementia
following a stroke, which included occlusions relative to other stroke types,
hemorrhagic strokes relative to other stroke types, and having more than one stroke
relative to only one stroke (Table 19).
Table 19
Univariate Conditional Logistic Regression Analyses (Matched Case-Control
Design): Stroke Risks for Poststroke Dementia
N= 51 cases, 138 controls________________________________________________
Odds Ratio 95% Confidence Interval
Hemorrhagic stroke 1.64 0.59-4.59
Occlusion stroke 1.22 0.63-2.38
More than one stroke 1.67 0.83-3.34
Behavioral Factors. There were no differences between cases and controls in
physical activity (i.e., leisure and work) or consumption of various food types (i.e.,
meat, fish and seafood, cakes and pastries, and fruits and vegetables). However,
participants with poststroke dementia were more likely to be heavier smokers during
their middle ages (O R= \29, 95% CI= 1.02-1.63) and lifetime smokers (OR=4.77,
95% Cl 1.37-16.60) than poststroke non-demented participants. In contrast, lifetime
alcohol drinkers were less likely to have poststroke dementia than abstainers
(OR=0.37, 95% C/=0.15-0.89). Odds ratios and frequencies for each behavioral
factor are in Tables 20 and 21.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
Table 20
Univariate Conditional Logistic Regression Analyses (Matched Case-Control
Design): Behavioral Risks for Poststroke Dementia
N = 51 cases, 138 controls___________________________________________
Odds Ratio 95% Confidence Interval
Low Education 1.41 1.15-1.73
Age 1.78 0.98-3.24
Diabetes 0.54 0.15 - 1.98
Hypertension 3.56 0.30-42.87
Smoking - Quantity 1.29 1.02-1.63
Smoking - Lifetime 4.77 1.37-16.60
Physical Activity (leisure) 1.21 0.78-1.90
Physical Activity (work) 1.40 0.75-2.62
Alcohol - Status 1.08 0.68-1.74
Alcohol - Lifetime 0.37 0.15-0.89
Alcohol - grams/month 1.00 0.62-1.60
Fish & Seafood 1.44 0.80-2.59
Fruits & Vegetables 0.89 0.48-1.66
Candy 0.88 0.38-2.04
Cakes & Pastries 0.89 0.53 -1.49
Beef 1.23 0.63-2.49
Pork 1.64 0.81-3.30
Sausage 1.00 0.49-2.03
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
Table 21
Frequencies o f Behavioral Factors fo r Cases (Poststroke Dementia, N=51) and
Controls (Poststroke, Intact, N=138)____________________________________
Dementia Cases Intact Controls
% n % n
Alcohol Abstainer 20 9 15 20
Former Drinker 0 0 2 2
Current: <500 g/mos 76 34 82 106
Current: 501-900 g/mos 2 1 1 1
Current: 901+ g/mos 2 1 0 0
Smoking Nonsmoker 53 27 55 76
Former 12 6 17 23
Current: Cigar/pipe only 8 4 7 10
Current: 2-10 cigarettes/day 14 7 12 17
Current: 11+ cigarettes/day 8 4 3 4
Current: Cigarettes+pipe/cigar 6 3 5 7
Physical Hardly Any 11 5 10 13
Activity Easy 64 30 66 86
(Leisure) Regular 8 4 17 22
Demanding 17 8 7 10
Physical Sedentary 26 10 21 22
Activity Active 38 15 67 70
(Work) Physically Strenuous 36 14 12 13
Fruits & No part 0 0 0 0
Vegetables Small part 29 12 20 23
Medium part 46 19 63 74
Great part 25 10 17 20
Fish & No part 0 0 3 3
Seafood Small part 37 14 39 45
Medium part 47 18 51 59
Great part 16 6 7 8
Beef No part 0 0 2 2
Small part 29 11 31 36
Medium part 63 24 64 73
Great part 8 3 3 3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
Table 21 (continued)
% n % N
Pork No part 0 0 1 1
Small part 26 10 38 46
Medium part 68 26 56 67
Great part 5 2 5 6
Sausage No part 0 0 1 1
Small part 31 12 25 29
Medium part 64 25 70 82
Great part 5 2 4 5
Candy More seldom 90 37 82 98
Once/day 3 1 16 19
Several times/day 7 3 3 3
Cakes & More seldom 39 16 32 39
Pastries Once/day 37 15 50 60
Several times/day 24 10 18 22
Conditional logistic regressions were also applied separately for men and
women. Among men who were poststroke, low education (OR= 1.47, 95% Cl = 1.07-
2.02) was related to an increased risk of dementia relative to those who were intact.
In addition, heavier smoking was related to an increased risk of dementia, however,
only approached significance (OR=\32, 95% C7=0.97-1.79). Among women who
were poststroke, low education (OR= 1.37, 95% CI= 1.05-1.78) and greater physical
activity (OR=231, 95% C/=l. 14-6.72) were related to an increased risk of dementia,
while lifetime drinking was related to a decreased risk of poststroke dementia
(OR=0.24, 95% C/=0.07-0.80). No other behaviors were significant predictors of
poststroke dementia. Odds ratios for each gender are in Tables 22 and 23.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
Table 22
Univariate Conditional Logistic Regression Analyses (Matched Case-Control
Design): Behavioral Risks fo r Poststroke Dementia - Women
N = 30 cases, 69 controls____________________________________________
Odds Ratio 95% Confidence Interval
Low Education 1.37 1.05-1.78
Diabetes 0.56 0.11-2.84
Hypertension 1.16 0.07-18.60
Smoking - Quantity 1.24 0.86-1.79
Smoking - Lifetime 2.59 0.66-10.20
Physical Activity (leisure) 2.77 1.14-6.72
Physical Activity (work) 2.12 0.77-5.81
Alcohol - Status 0.76 0.45-1.30
Alcohol - Lifetime 0.24 0.07-0.80
Alcohol - grams/month 1.00 0.99-1.01
Fish & Seafood 1.89 0.90-3.97
Fruits & Vegetables 1.24 0.51-2.99
Candy 1.32 0.40-4.41
Cakes & Pastries 1.38 0.71 -2.67
Beef 0.79 0.30-2.08
Pork 1.04 0.40-2.68
Sausage 0.84 0.33-2.13
Note. Education is reverse coded.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Table 23
Univariate Conditional Logistic Regression Analyses (Matched Case-Control
Design): Behavioral Risks for Poststroke Dementia - Men
N = 2 \ cases, 69 controls____________________________________________
Odds Ratio 95% Confidence Interval
Low Education 1.47 1.07-2.02
Diabetes 0.50 0.06-4.45
Smoking - Quantity 1.32 0.97-1.79
Physical Activity (leisure) 0.89 0.51 - 1.55
Physical Activity (work) 1.06 0.47-2.37
Alcohol - Lifetime 0.90 0.15-5.39
Alcohol - grams/month 1.00 1.0 0 - 1.00
Fish & Seafood 0.84 0.30-2.34
Fruits & Vegetables 0.63 0.25 - 1.59
Candy 0.60 0.17-2.17
Cakes & Pastries 0.45 0.19-1.08
Beef 2.06 0.71-6.01
Pork 2.95 0.94 - 9.26
Sausage 1.27 0.42-3.88
Multivariate conditional logistic models were calculated to estimate the
overall fit of the model and to identify independent predictors of poststroke
dementia. Univariate conditional logistic regression models identified the following
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 2
independent predictors: low education, smoking quantity, lifetime tobacco use, and
lifetime alcohol use. Both lifetime smoking and smoking quantity were related to an
increased risk of poststroke dementia; however, only smoking quantity was used in
the multivariate analysis because it was a significant predictor of poststroke dementia
among men and when gender groups were combined. In addition both smoking
variables were highly correlated (r- 0.75,/?< 0001, n=135), but smoking quantity
had a larger number of respondents (n=188) relative to lifetime smoking (n=136).
The multivariate conditional logistic regression model included the following
independent predictors: low education, smoking (quantity), and lifetime alcohol
consumption. The overall fit of the model was significant, x2 (3) = 12.49,/? =.006,
with a -2 Log Likelihood of 21.03 (with covariates). The model identified greater
amounts of smoking (OR=2.21, 95% (77=1.14-4.27) as a significant risk for
poststroke dementia, and lifetime alcohol consumption as protective against
poststroke dementia (OR=0.3l, 95% CI=0.10-0.95). Low education was not
significantly related to poststroke dementia.
Separate regression models were performed for each gender, which included
those independent variables identified in their respective univariate analyses.
Among men, the independent predictors included low education and smoking
(quantity). The overall fit o f the model was significant, y2 (2) = 12.2 9, p =.002, with
a -2 Log Likelihood of 51.94 (with covariates). The model identified low education
as a significant risk for poststroke dementia (OR=1.47, 95% C/= 1.06-2.04), but
heavier smoking was no longer significant. For women, education, physical activity
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
(leisure) and lifetime alcohol consumption was included in the regression model. The
overall fit of the model was significant, y l (3) = 8.13, p =.04, with a -2 Log
Likelihood of 12.71 (with covariates). None of the variables predicted poststroke
dementia in the multivariate model. Table 24 summarizes the analyses for all
participants and for each gender.
Table 24
Adjusted Odds Ratios (Matched Case-Control Design): Behavioral Risks fo r
Poststroke Dementia
N = 51 cases, 138 controls__________________________________________
All Men Women
OR 95% Cl OR 95% Cl OR 95% Cl
Low Education 1.37 0.68-2.78 1.48 1.06-2.04 0.94 0.56-1.56
Smoking -
Quantity
2,21 1.14-4.27 1.31 0.92-1.86
Physical Activity 2.93 0.41-20.80
Alcohol -
Lifetime
0.31 0.10-0.95 0.27 0.06-1.28
Correlational analyses were performed to help explain the findings from the
conditional logistic regressions (Table 25). In particular, it was unexpected that
greater amounts of physical activity (leisure) were related to an increased risk of
poststroke dementia among women. In the correlational analyses, greater amount of
physical activity at work was positively related to physical activity during leisure,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
lower education and greater amounts of smoking. Additionally, greater amounts of
cake and pastry consumption were positively related to greater amounts of physical
activity at work and at leisure, as well as negatively related to greater smoking.
Finally, drinking alcohol was positively related to greater physical activity at leisure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 25
Intercorrelations of Risk Factors and Confounding Variables (N=189, Combined Poststroke Dementia and Intact)
Measure i 2 3 4 5 6 7 8 9
10
11 12
1. Low Education — .01
n=180
-.02
n=162
.02
n=170
.24**
n=140
.09
n=152
.09
n=150
-.05
n=146
-.10
n=147
-.06
n=152
-.02
n=155
-.03
n=156
2. Smoking — .20*
n=169
-.00
n=178
-.18*
n=144
-.04
n=158
.00
n=156
.06
n=152
.03
n=153
-.04
n=158
-.10
n=161
-.18*
n=162
3. Alcohol — .18*
n=162
-.08
n=135
.13
n=146
-.04
n=T44
.02
n=141
.01
n=142
.04
n=146
.16
n=149
.03
n=150
4. Physical Activity
(Leisure)
—
29**
n=139
-.08
n=155
.03
n=152
-.02
n=T49
-.03
n=149
-.06
n=154
-.09
n=157
.20*
n=158
5. Physical Activity
(W ork)
— -.05
n=124
.04
n=122
-.17
n=118
.01
n=119
-.09
n=124
.07
n=127
.20*
n=128
6. Pork —
32***
n=154
.28**
n=151
.18*
n=151
-.12
n=155
.12
n=157
.09
n=158
7. Sausage — .23*
n=149
.10
n=150
-.05
n=154
.10
n=155
.14
n=156
8. Beef —
27**
n=147
.09
n=151
.06
n=151
.09
n=152
9. Fish/Seafood —
29**
n=152
.06
n=153
.13
n=153
10. Fruits/Vegies .14
n=157
-.02
n=158
11. Candy — .21*
n=161
12. Cakes/Pastries —
Note. Education is reverse code. *** p < .000 , **/><.001, *p< .05.
o
C/1
7 6
5. Discussion
The present study examined the risk of dementia following a stroke, using a
population-based twin sample and both classic case-control and co-twin control
designs. To my knowledge, no other studies have examined stroke as risk for
dementia using both designs within the same study. In addition, mid-life behaviors
were examined prospectively as predictors of dementia, after controlling for the risk
posed by stroke. Most studies tend to focus on lesion features (e.g., size and location
of a lesion), which are important determinants of poststroke dementia; however,
alone cannot completely account for why stroke some people develop dementia after
stroke while others do not (Tatemichi et al., 1993). It is likely that multiple factors,
acting in particular combinations lead to dementia (Tatemichi et al., 1993). Few
studies have focused on the role of behaviors in relationship to poststroke dementia.
Stroke
In the present study, stroke cases were identified through a Hospital Register
in Sweden. It was estimated that 9% of the base sample had a history of stroke (10%
of men and 8% of women). As expected, these estimates were very similar to Zhu et
al. (1998), who also utilized the Swedish Hospital Discharge Register to examine
stroke.
Stroke as a Risk fo r Dementia
For the first aim, it was predicted that individuals with a history of stroke
would be more likely to develop dementia than those with no history of stroke.
Results from the case-control analyses suggested that individuals with a history of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
stroke were 1.49 times more likely to develop dementia than those with no history of
stroke. After adjusting for the effects of age, education and diabetes, the risk of
dementia remained about the same at 1.57 times higher those with no stroke history.
When participants who had a TIA were excluded from the analyses, risk of dementia
was similar at 1.65 times participants who no stroke history.
It was, however, unexpected that the risk of dementia following a stroke
would increase after controlling for covariates (using the sample that included
TIA’s), resulting in further investigation of this finding. Diabetes appeared to
moderate the risk of stroke. Surprisingly, among diabetics, stroke did not increase the
risk of dementia. Among non-diabetics, stroke was a significant risk for dementia.
However, a post-hoc exploratory analysis suggested that twins with a history of
diabetes were at an increased risk of stroke. One possible explanation is that
participants with both diabetes and stroke died early, before having the chance to
develop dementia. Moreover, the risk appeared greater among men than women, in
that, men with history of stroke had an increased risk of dementia by 1.7 times those
with no history of stroke. Stroke did not increase risk of dementia among women.
Results from the co-twin control design analyses indicated that individuals
with a history of stroke were not significantly more likely to develop dementia than
those with no stroke history. Given greater control of early environmental and
familial factors in the co-twin design, the lack of statistical significance in the co
twin design relative to the matched case-control design might suggest that early
environmental or familial factors influenced the development of dementia after a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
stroke. However, the odds ratio in the co-twin design (1.70) was actually higher than
in the matched case-control design (1.49). This lack of significance is likely related
to low power rather than the influence of genetics or early familial environment.
Separate analyses for men and women revealed a possible trend, in that the
risk of dementia among men with stroke may be greater than women with stroke.
Descriptively, 21% of men with dementia had a history of stroke as compared to 5%
of their twin partners who were intact. Among women, 9% of both the demented and
non-demented co-twin pairs had a history of stroke. Previous research has reported
higher rates of stroke among Swedish men than women, but no differences in gender
in relation to dementia after stroke (Zhu et al., 1998). Overall, these results suggested
that stroke increases risk of dementia, particularly among men; however, it seems
unlikely that genetics or early rearing environmental factors shared by twins
influence risk.
The present study was unable to directly examine genetic influences of
poststroke dementia due to the lack of twin pairs concordant for stroke and
discordant for dementia, but genetic influences on the occurrence of stroke were
examined. Probandwise concordance rates of stroke were calculated. Higher
probandwise concordance rates among monozygotic (18.9%) relative to dizygotic
twin pairs (11.9%) were found, suggesting a genetic component to stroke, but a role
for non-genetic influences. This pattern of concordance rates was also found among
a population based sample of Danish twins, as probandwise concordance rates were
19% for monozygotic and 13% for dizygotic twin pairs (Bak et al., 2002). Among a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 9
sample of U.S. veteran male twins, probandwise concordance rates for stroke were
17.7% for monozygotic and 3.6% for dizygotic pairs (Brass et al., 1992).
In comparison to previous research, the risk afforded by stroke in the present
study was actually lower than earlier estimates. It is difficult to speculate why the
risk of poststroke dementia may differ, especially given different study populations,
data collection procedures and diagnostic criteria. However, the present study used a
population-based sample, in contrast to other studies which may have utilized
hospital-based or clinical samples. Stroke risk may be overestimated among
hospitalized samples, as they may be in poorer health overall. In addition, biases may
be introduced as patients with stroke and dementia may be more likely to be referred
to certain hospitals or neurology clinics. Within hospital cohorts, the frequency of
dementia among stroke patients ranged from about 25% to 32% (e.g., Barba et al.,
2000; Censori et al., 1996; Tatemichi et al., 1992). This differs from a community-
based sample, which reported a 12% frequency of dementia (Kase et al., 1998) or a
population-based sample in Rochester, Minnesota which reported a 20% frequency
of dementia (Kokmen et al., 1996). Prencipe et al. (1997) differed from the present
study and the two mentioned above (Kase et al., 1998; Kokmen et al., 1996) by
examining stroke as a risk for vascular dementia only. Prencipe et al. (1997) reported
a 30% frequency o f vascular dementia among a population-based sample o f stroke
survivors in Italy. However, other factors may also be influential.
The length of time from stroke onset to dementia onset may relate to
differential rates of poststroke dementia. The highest risk of poststroke dementia
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
tends to be within one year of the stroke. For example, odds ratios of dementia
within the first year of a stroke was 8.8 times that of non-stroke participants, 4.2
times within 3 years, 3.5 times within 5 years, and 2.5 times within 10 years
(Kokmen et al., 1996). In the present study, risk of dementia following a stroke was
1.49 times higher than non-stroke participants, with an average number of years
between the first stroke and dementia onset of 3.4 years. In addition, Kokmen et al.
(1996) reported 7% incidence of dementia within the first year of stroke. Similarly,
among all stroke participants in the present sample (n= 1,290), 5.2% developed
dementia within the first year of stroke. As expected, the percentage in the present
study was likely underestimated because we were unable to determine who might
have died before developing dementia. The present study also could not evaluate risk
of dementia onset within one year of stroke because a comparison group of
nondemented stroke participants were not followed during the same time period.
It is also important to establish prestroke cognitive functioning level.
Although the present study did not have access to longitudinal cognitive functioning,
anyone whose first stroke event followed dementia onset was not considered positive
for stroke. This provides a gross estimate of prestroke cognitive functioning, even
though it is possible that a participant may have had some type of cognitive
impairment prior to stroke The Framingham Study, a prospective longitudinal,
community-based study evaluated participants’ cognitive functioning over a 13-year
period, enabling them to compare pre- and poststroke cognitive functioning (Kase et
al., 1998). They estimated that 12% of their sample developed dementia following a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
stroke, which is similar to the present study. Inzitari et al. (1998) utilized an
informant interview method to determine dementia at stroke onset and estimated that
16.8% of their patients had poststroke dementia. Further, Zhu et al. (1998) reported
higher relative risks for dementia when all strokes were included (OR= 3.0)
compared to relative risks when those who demonstrated dementia symptoms prior
to stroke were excluded (OR= 2.6). It is important to compare the findings of the
present study to the Zhu et al. (1998) study, who also utilized the same Hospital
Discharge Register. Their slightly higher estimate is likely attributable to their older
age limit of 75 years and older, whereas the present study included participants 65
years and older when evaluated for cognitive impairment and excluded anyone over
90 years.
Covariates were also significant independent predictors of poststroke
dementia. As found in previous research, increased age has been a consistent
predictor of vascular dementia and poststroke dementia (Barba et al., 2000; Gorelick,
1997; Pohjasvaara et al., 1998). Low education has also been associated with
poststroke dementia (Gorelick et al., 1993; Pohjasvaara et al., 1998; Prencipe et al.,
1997), thus, it is not surprising that older age and low education increased risk of
dementia by 1.3 and 1.1 times those who are younger or with higher education,
respectively. Additionally, history o f diabetes (with onset before a stroke) increased
the risk of dementia by 1.76 times those with who were not diabetic in the matched
case-control design, even after controlling for the effects of stroke. However, in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
present study, an unexpected finding led us to speculate that twins with a history of
diabetes and stroke may have died before developing dementia.
There has been past disagreement regarding whether diabetes poses an
independent risk for dementia, outside of stroke. For example, Tatemichi et al.
(1993) found that diabetes increased risk of dementia after a stroke, but questioned
whether the risk of dementia was mediated through the relationship of diabetes and
stroke. In some studies, diabetes was defined as either a past diagnosis or current
treatment of diabetes (e.g., Barba et al., 2000; Inzitari et al., 1998). In the present
study, diabetes was defined as positive only if onset was prior to stroke, likely
leading to longer term diabetics. It was suggested that those who have been exposed
to diabetes for a longer period of time or are more severe (i.e., receiving insulin-
dependent treatments) had a higher risk of dementia (Ott et al., 1999), and in the
present study, perhaps, a higher risk of mortality.
Finally, stroke features were examined as predictors of dementia after stroke.
The present studied combined all stroke types in one “stroke” category, which
included occlusions, hemorrhages and TIA’s. Stroke type did not affect risk of
poststroke dementia. Most studies tended to exclude hemorrhagic strokes and TIA’s,
but it was included in the present study to be comparable to Zhu et al. (1998).
Additionally, a small percentage o f strokes were comprised o f hemorrhages (11%),
and neither type of stroke influenced risk of poststroke dementia in the present
sample. The occurrence of more than one stroke was not related to an increased risk
of poststroke dementia. Of the population-based samples examining poststroke
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
dementia, there has been inconsistency. Like the present study, Prencipe et al.
(1997) found that multiple vs. single infarcts were not related to poststroke dementia;
however, Kokmen et al. (1996) reported that having a second stroke was related to
an increased risk of poststroke dementia. The present study did not examine specific
anatomical features of stroke; however, the findings have also been inconsistent.
Volume of cerebral tissue loss, number of infarcts, and cerebral atrophy have been
supported as risks for dementia associated with stroke among one or more well
designed clinical studies (Gorelick et al., 1997). The influences of infarct location,
white matter lesions and silent cerebral infarcts on dementia associated with stroke
have been examined, but results may have been inconclusive or the studies were not
methodologically sound (Gorelick et al., 1997).
In summary, stroke was related to an increased risk of dementia by about
some 1.5 times those with no history of stroke. This risk of stroke may be influenced
by in part by genetics; however, genetics may play less of a role in the development
of dementia after stroke. In addition, risk may be different among men than women.
It is clear that multiple factors influence prevalence rates of dementia after stroke,
including (but not limited to) sample selection (i.e., population or hospital), time
span from stroke to assessment of dementia, and prestroke cognitive decline. It
appears the population-based samples and those that account for prestroke cognitive
decline report lower estimates of dementia after stroke, whereas those that estimate
risk within 3 months to a year of stroke onset report higher rates. The role of stroke
features and cardiovascular diseases (e.g., diabetes) remain unclear.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
Behavioral Factors Related to Poststroke Dementia
For the second aim, various mid-life behaviors were investigated as
predictors of poststroke dementia, beyond the risk posed by stroke. Several studies
have asserted that stroke is largely preventable (Gorelick et al., 1999; Marmot &
Poulter, 1992). One literature review paper provided recommendations for
preventing a first stroke, including treatment of hypertension, myocardial infarction,
atrial fibrillation, diabetes, high cholesterol, carotid artery disease, and modifying
lifestyle factors - lifestyle factors included smoking, alcohol, physical activity and
diet (Gorelick et al., 1999). It was predicted that these behaviors influence risk of
poststroke dementia.
Results from the case-control analyses revealed that fewer years of education,
heavier smoking during mid-life, and lifetime smoking (i.e., reported being a smoker
during mid-life and again in the 1998-2000 telephone screening) were related to a
greater risk of poststroke dementia, while lifetime alcohol consumption (i.e.,
reported drinking alcohol during mid-life and again in the 1998-2000 telephone
screening) was related to a decreased risk of poststroke dementia. Physical activity
(work or leisure) and diet (meats, sweets, fish and seafood, or fruits and vegetables)
were not related to risk of poststroke dementia.
The results revealed different risks when analyses were run separately for
men and women. Among men, univariate results showed greater amounts of smoking
approached significance as a risk for poststroke dementia. Among women, a greater
amount of physical activity during leisure was related to an increased risk of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
poststroke dementia, and lifetime alcohol consumption was related to a decreased
risk of poststroke dementia. Fewer years of education was related to an increased
risk of poststroke dementia among both genders. However, no behavioral risk factor
was significant in the within gender multivariate analyses.
As predicted, heavier smoking during mid-life increased risk of poststroke
dementia, particularly among men. However, it is important to note that 79% of
women in this sample were nonsmokers. To my knowledge, no other study has
specifically examined mid-life smoking habits and risk of poststroke dementia.
Although several studies found that current smoking was related to an increased risk
of poststroke dementia (Gorelick et al., 1993; Pohjasvaara et al., 1998), research on
the relationship of smoking to poststroke or vascular dementia has been equivocal.
For example, other studies that examined lifetime tobacco exposure (Tatemichi et al.,
1993), smoking greater than 10 cigarettes per day (Censori et al., 1996) or current
smoking status (Barba et al., 2000) did not find a relationship between smoking and
poststroke dementia. One obvious difficulty is related to different measures of
tobacco use. Nonetheless, the present study is consistent with a Finnish study that
also reported a higher frequency of current smoking among poststroke dementia
participants relative to nondemented poststroke participants (Pohjasvaara et al.,
1998).
In contrast, lifetime alcohol consumption was related to a decreased risk of
poststroke dementia. When examining quantity of alcohol, the distribution of alcohol
intake was bimodal, in that over 95% of the participants were either nondrinkers or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
light drinkers; thus, the present study could not examine whether moderate levels of
alcohol were protective or whether heavy alcohol consumption was a risk for
poststroke dementia. The limited number of heavier alcohol drinkers was likely
related to Sweden’s early system of alcohol rationing for spirits, which was active
until 1955 (Hammar, Romelsjo, & Alffedsson, 1997). In addition, married women
were not allowed to buy alcoholic beverages. Regarding the present sample, women
from the old cohort were at least age 29 years or older at the time rationing was
abolished and may have already established very low levels of alcohol consumption.
Another possible factor may be related to an unwillingness to admit to higher
consumption of alcohol.
However, given the low quantities of alcohol consumption, our measure of
lifetime alcohol use (i.e., abstainer, former drinker, or current drinker) provided an
indirect means of examining whether moderate drinking during mid- and late-life
modified the risk of poststroke dementia relative to lifetime abstainers or former
drinkers. The results indicated that lifetime drinking was protective against
poststroke dementia. Average grams of alcohol per day among lifetime drinkers in
the present sample (old cohort) were 3.11 grams per day for women and 6.13 grams
per day for men. These results were consistent with a recent meta-analysis on alcohol
consumption and stroke risk, in that consumption of less than 12 grams of alcohol
per day was associated with a lower risk of stroke (Reynolds et al., 2003), which
equates to almost one drink per day (by European standards). Further, drinking less
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
than one drink per week was associated with a reduced risk of dementia, with an
increasing reduction in risk through 13 drinks per week (Mukamal et al., 2003).
A surprising finding was that strenuous physical activity (leisure) during mid
life increased risk of poststroke dementia among women, although this finding did
not hold up in multivariate analyses. This contradicts earlier research supporting
regular exercise as protective against vascular dementia (Hebert et al., 2000;
Yoshitake et al., 1995). In the present sample, heavier physical activity at leisure
was positively correlated with alcohol consumption and eating cakes or pastries, as
well as with low education. One possibility is that more strenuous physical activity
was related to poorer lifestyle habits, which may increase risk of both stroke and
dementia. Alternatively, extreme levels of physical activity (too much or too little)
may be the critical factor. In the present sample, 24% of participants with poststroke
dementia engaged in strenuous physical activity, and 19% rarely engaged in physical
activity. In contrast, the non-demented stroke group had only 9% and 12% of
participants who rarely engaged or engaged in strenuous physical activity,
respectively. Over 70% of the non-demented stroke participated in easy or regular
physical activity. Moreover, physical activity during leisure was positively
correlated with physical activity at work, possibly suggesting that those who were
very active at work were also active during their leisure time. The results o f the
present study illustrate the complexity of disentangling behavioral risk factors.
Of all the behaviors examined in the present study, diet is probably the least
studied in relation to poststroke dementia. It was predicted that a high fat diet during
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
mid-life would relate to an increased risk of poststroke dementia. In the present
study, specific measures of fats or saturated fatty acids were not used. However,
estimates of fat may be derived from certain food types. For example, pastries and
cakes tend to be fairly high in fat, particularly fried desserts like pie crusts, danishes,
croissants or doughnuts (according to a USDA National Nutrient Database). Also,
meats, such as beef, pork or sausages tend to have higher levels of fats. In the present
study, none of the “high fat” types of food were related to post stroke dementia. Also,
the prediction that eating greater amounts of fish would be protective against
poststroke dementia was not supported. Overall, participants consumed the various
foods in moderation, as there tended to be very few people who endorsed any one
food as a “great part” or “no part” of their diet. One study which examined diet
variables in the Swedish Twin Registry in relation to prostate cancer risk postulated
that the distinction between categories, such as “medium part” or “great part” of
one’s diet may be too subtle to detect differences (Henrik, Damber, L., & Damber, J,
1996). Due to the paucity of research in this area, the purpose of examining diet was
primarily exploratory, to identify possible influences on poststroke dementia.
In sum, heavier smoking was positively related to risk of poststroke
dementia, while lifetime alcohol consumption in low to moderate amounts was
related lower risks o f poststroke dementia. There was no relationship between
poststroke dementia and diet.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
Limitations and Strengths
One limitation of the present study was that the data were prevalence based,
in that dementia diagnoses were determined at one time point as opposed to being
followed longitudinally. For example, in a cohort design of poststroke dementia,
participants with stroke would be followed from stroke onset, over a period of time,
and numbers of incident dementia cases could be compared to others who were
followed during that same time period. Given the increased mortality rate following
a stroke, our frequencies of dementia may be underestimates due to early mortality.
However, this limitation actually suggests that the relative risks in the present study
may be underestimating the strength of the relationship between stroke and
dementia.
Second, this study was limited to examining stroke in relation to dementia,
not a specific diagnosis of vascular dementia. The determination of a vascular
dementia diagnosis was beyond the scope of this paper. On the other hand, there are
difficulties in agreeing on vascular dementia diagnoses. The greatest difficulty in
determining a vascular dementia diagnosis is accurately judging whether the stroke is
directly related to the onset of dementia (Ekinjuntti, 1997). For example, the NINDS-
AIREN requires the onset of dementia within three months of the recognized stroke
(Roman et al., 1993). This time frame may seem arbitrary, not to mention, strokes
may not always present with obvious symptoms (Roman et al., 1993). The ADDTC
criterion requires demonstrated brain injury through brain imaging (e.g., CT or MRI
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9 0
scans) (Chui et al., 1992); however, imaging cannot always establish timing.
Consequently, the ADDTC criterion accepts either a clear temporal relationship
between a single stroke and dementia or evidence of two or more strokes (no
temporal relationship required). In the present study, the occurrence of stroke was
fairly well established through the Swedish Hospital Discharge Register. All strokes
occurred prior to dementia onset. In the present sample, 48% of those who had
poststroke dementia, onset of dementia was within one year, and 55% experienced
two or more strokes. The remaining participants developed dementia anywhere from
2 years on, which is an obvious limitation of the study. Although causality cannot be
established, conclusions regarding the association of stroke with dementia may be
reasonably supported.
Third, behaviors in relation to poststroke dementia were assessed in the
1960’s and 1970’s, with little information on changes in lifestyle behaviors until the
telephone screening (1998-2000). This limitation also benefits the present study, as
data were collected prospectively. Further, the present study focused on mid-life
behaviors prior to stroke onset as risks for poststroke dementia. It is suspected that
strategies for preventing poststroke dementia prior to stroke versus after a stroke are
different, thus, examining risk factors separately would be relevant.
Fourth, there was limited variability among some behavioral factors,
particularly with alcohol consumption and diet. However, accuracy of coding was
checked with telephone screening information, where possible. In addition, our study
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
was consistent with other studies who have used the Swedish Twin Registry and also
reported low rates of alcohol use and a moderate diet (e.g., Terry et al., 1999).
There were several strengths of the present study. In specific, the Swedish
Twin Registry presented a unique opportunity to study both environmental and
genetic influences, within a large, population-based sample. Given a population-
based sample, biases related to self-selection were not a concern. Second, data
collection spanned over an approximate 30-year time period, which enabled the
study of predictors prospectively, reducing biases related to retrospective reporting.
Third, due to multiple data collection points and sources of data (e.g., medical
records, national data sources), information was available to either supplement or
complement variables of interest. The multiple sources of data were also useful in
establishing reliable estimates of disease onset, including diabetes, stroke, and
dementia. Finally, the present study utilized both matched case-control and co-twin
control designs, providing greater statistically efficiency relative to statistically
adjusting for covariates, which theoretically should lead to more accurate estimates
of risk.
Conclusions and Implications
The purpose of the present study was to examine stroke as a risk for dementia
using both case-control and co-twin control designs, and to identify mid-life
behaviors that might influence risk of dementia given the presence of stroke. As
expected, stroke increased risk of dementia. The risk may be modified by gender or
behaviors, and less likely by genetics or early rearing environment. Specifically,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9 2
heavier smoking was a predictor of poststroke dementia, while lifetime drinking of
low to moderate levels of alcohol was protective.
Overall, it appears that the path to poststroke dementia is multifactorial,
which may include influence of behaviors. Further, behaviors during mid-life may
have consequences during late life. The present study focused on behaviors that
occurred before the stroke onset. Future research that investigates predictors of
poststroke dementia prior to and post- stroke would be useful in establishing
preventive interventions. The role of genetics in poststroke dementia should also be
further examined, as well as a more detailed look at relevant behaviors, e.g.,
exploring the dose-response relationship between alcohol use and risk of poststroke
dementia.
Most of the research in poststroke dementia has been with neurologists or
epidemiologists, but the results of the study uncover a critical role for psychologists.
Based on the results of the present study, preventive interventions that target
improving health behaviors, such as smoking cessation, minimal or moderate alcohol
consumption and moderate physical activity are supported. O f course, one is not
encouraged to start drinking alcohol or stop exercising; however, this study provides
evidence that other factors aside from stroke alone, demographics and stroke-related
features matter. Specifically, behaviors, even during mid-life may influence risk o f
poststroke dementia. Yet, it is well known that motivation to change and lack of
adherence to treatment or preventive interventions are significant barriers. For
example, in the present sample 70% of participants who reported smoking during
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
mid-life were also smoking in late-life. Research suggests that psychologists may
not only be helpful in directly working with a patient to assist in behavior change,
but they are also useful in training physicians to improve treatment adherence
(Mullet et al., 1997; Samsa et al., 1997). Prevention of stroke would most likely be
an immediate and effective means to reduce poststroke dementia, and ideally during
middle age or sooner (Nyenhuis & Gorelick, 1998). Secondary prevention, after a
first stroke, would also be a critical time to intervene as risk of dementia and patient
motivation to adhere to treatment tends to be high.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
References
Alffesson, L., Von Arbin, M., & De Faire, U. (1986). Mortality from and incidence
of stroke in Stockholm. British Medical Journal, 292, 1299-1303.
American Heart Association. (2000). 2001 Heart and stroke statistical update.
Dallas, TX: American Heart Association.
American Psychiatric Association. (1994). Diagnostic and statistical manual fo r
mental disorders (4th Rev., pp. 133-55). Washington, DC: Author.
Andersen, G., Vestergaard, K., Riis, J. O., Inteman-Nielsen, M. (1996). Intellectual
impairment in the first year following stroke, compared to an age-matched
population sample. Cerebrovascular Disease, 6, 363-369.
Barba, R., Martmez-Espinosa, S., Rodriguez-Garcia, E., Pondal, M., Vivancos, J., &
Ser, T. D. (2000). Poststroke dementia: Clinical features and risk factors.
Stroke, 31, 1494-1501.
Bergem, A. L. M., Engedal, K., & Kringlen, E. (1997). The role of heredity in late-
onset Alzheimer’s disease and vascular dementia: A twin study. Archives o f
General Psychiatry, 54, 264-270.
Blessed, G., Tomlinson, B. E., & Roth, M. (1968). The association between
quantitative measures of dementia and of senile change in the cerebral grey
matter of elderly subjects. British Journal o f Psychiatry, 114, 797-811.
Blumenthal, J. A., Emery, C. F., Madden, D. J., Schniebolk, S., Walsh-Riddle, M.,
George, L. K., McKee, D. C., Higginbotham, M. B., Cobb, F. R., &
Coleman, R. E. (1991). Long-term effects of exercise on psychological
functioning in older men and women. Journal o f Gerontology:
Psychological Sciences, 46, P352-P361.
Blumenthal, J. A., & Madden, D. J. (1988). Effects of aerobic exercise training,
age, and physical fitness on memory-search performance. Psychology and
Aging, 3, 280-285.
Bonner, A. P., & Cousins, S. O. (1996). Exercise and Alzheimer’s disease: Benefits
and barriers. Activities, Adaptation & Aging, 20, 21-34.
Boston, P. F., Dennis, M. S., & Jagger, C. (1999). Factors associated with vascular
dementia in an elderly community population. International Journal o f
Geriatric Psychiatry, 14, 761-766.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
Brass, L. M., Isaacsohn, J. L., Merikangas, K .R., & Robinette, C. D. (1992). A
study of twins and stroke. Stroke, 23, 221-223.
Censori, B., Manara, O., Agostinis, C., Camerlingo, M., Casto, L., Galavotti, B.,
Partziguian, T., Servalli, M. C., Cesana, B., Belloni, G., & Mamoli, A.
(1996). Dementia after first stroke. Stroke, 27, 1205-1210.
Chui, H. C. (1998). Rethinking vascular dementia: Moving from myth to
mechanism. In J. H. Growdon & M. N. Rossor (Eds.), Dementias (pp. 377-
401). Woburn, MA: Butterworth-Heinemann.
Chui, H. C., Mack, W., Jackson, J. E., Mungas, D., Reed, B. R , Tinklenberg, J.,
Chang, F., Skinner, K., Tasaki, C., & Jagust, W. J. (2000). Clinical criteria
for the diagnosis of vascular dementia: A multicenter study of comparability
and interrater reliability. Archives o f Neurology, 57, 191-196.
Chui, H. C., Victoroff, J. I., Margolin, D., Jagust, W., Shankle, R., & Katzman, R.
(1992). Criteria for the diagnosis of ischemic vascular dementia proposed by
the State of California Alzheimer’s Disease Diagnostic and Treatment
Centers. Neurology, 42, 473-480.
Clarkson-Smith, L., & Hartley, A. A. (1989). Relationships between physical
exercise and cognitive abilities in older adults. Psychology and Aging, 4,
183-189.
Collins, R , Peto, R., MacMahon, S., Hebert, P., Fiebach, N.H., Eberlein, K. A.,
Godwin, J., et al. (1990). Blood pressure, stroke, and coronary heart
diseases. Part 2, short term reductions in blood pressure: Overview of
randomized drug trials in their epidemiological context. Lancet, 335, 827-
838.
Cummings, J. L., Mega, M., Gray, K., Rosenberg-Thompson, S., Carusi, D. A., &
Gornbein, J. (1994). The Neuropsychiatric Inventory: comprehensive
assessment of psychopathology in dementia. Neurology, 44, 2308-2314.
Curb, J. D., Rodriguez, B. L., Abbott, R. D., Petrovitch, H., Ross, G. W., Masaki, K.
H., Foley, D., Blanchette, P. L., Harris, T., Chen, R., & White, L. R. (1999).
Longitudinal association of vascular and Alzheimer’s dementias, diabetes,
and glucose tolerance. Neurology, 52*971-975.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9 6
Dustman, R. E., Emmerson, R. Y., Ruhling, R. O., Shearer, D. E., Steinhaus, L. A.,
Johnson, S. C., Bonekat, H. W., & Shigeoka, J. W. (1990). Age and fitness
effects on EEG, ERP’s, visual sensitivity, and cognition. Neurobiology o f
Aging, 11, 193-200.
Dustman, R. E., Emmerson, R., & Shearer, D. (1994). Physical activity, age, and
cognitive-neuropsychological function. Journal o f Aging and Physical
Activity, 2, 143-181.
Emery, C. F., & Gatz, M. (1990). Psychological and cognitive effects of an exercise
program for community-residing older adults. Gerontologist, 30, 184-188.
Erkinjuntti, T. (1997). Vascular dementia: Challenge of clinical diagnosis.
International Psychogeriatrics, 9(Suppl. 1), 51-58.
Etnier, J. L., Salazar, W., Landers, D. M., Petruzzello, S. J., Han, M., & Nowell, P.
(1997). The influence of physical fitness and exercise upon cognitive
functioning: A meta-analysis. Journal o f Sport & Exercise Psychology, 19,
249-277.
Foulkes, M. A., Wolf, P.A., Price, T. R., Mohr, J. P., & Hier, D. B. (1988), The
Stroke Data Bank: Design, methods, and baseline characteristics. Stroke, 19,
547-554.
Galanis, D. J., Petrovitch, H., Launer, L. J., Harris, T. B., Foley, D. J., & White, L.
R. (1997). Smoking history in middle age and subsequent cognitive
performance in elderly Japanese-American men: The Honolulu-Asia Aging
Study. American Journal o f Epidemiology, 145, 507-515.
Gatz, M., Pedersen, N. L., Berg, S., Fratiglioni, L., Johansson, B., Mortimer, J. A.,
Reynolds, C. A., & Fiske, A. (2002). Study of dementia in the Swedish twin
registry [Abstract], Neurobiology o f Aging, 23, S309.
Gatz, M., Pedersen, N. L., Berg, S., Johansson, B., Johansson, K., Mortimer, J. A.,
Posner, S. F., Viitanen, M., Winblad, B., & Ahlbom, A. (1997). Heritability
for Alzheimer’s disease: The study of dementia in Swedish twins. Journals
o f Gerontology: Medical Sciences, 52A, M l 17-M125.
Gatz, M., Reynolds, C., Nikolic, J., Lowe, B., Karel, M., & Pedersen, N. (1995). An
empirical test of telephone screening to identify potential cases.
International Psychogeriatrics, 7 , 429-438.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9 7
Gillman, M. W., Cupples, L. A., Millen, B. E., Ellison, R. C , & Wolf, P. A. (1997).
Inverse association of dietary fat with development of ischemic stroke in
men. JAMA, 278, 2145-2150.
Gilroy, J. (1995). Clinical applications and introduction to cerebrovascular
anatomy, physiology, and pathology. In C. L. Rumbaugh, A. Want, & F. Y.
Tsai (Eds.), Cerebrovascular Disease: Imaging and Interventional
Treatment Options (pp. 3-9). New York: Igaku-Shoin Medical Publishers,
Inc.
Gorelick, P. B. (1997). Status of risk factors for dementia associated with stroke.
Stroke, 28, 459-463.
Gorelick, P. B., Brody, J., Cohen, D., Freels, S., Levy, P., Dollear, W., Forman, H.,
Harris, Y. (1993). Risk factors for dementia associated with multiple
cerebral infarcts: A case-control analysis in predominantly African-
American hospital-based patients. Archives o f Neurology, 50, 714-120.
Gorelick, P. B., Erkinjuntti, T., Hofman, A., Rocca, W. A., Skoog, I., & Winblad, B.
(1999). Prevention of vascular dementia. Alzheimer’ s Disease and
Associated Disorders, 13 (Suppl. 3), S131-S139.
Gorelick, P.B., Sacco, R. L., Smith, D. B., Alberts, M., Mustone-Alexander, L.,
Rader, D. Ross, J. L., et al. (1999). Prevention of a first stroke: A review of
guidelines and a multidisciplinary consensus statement from the National
Stroke Association. JAMA, 281, 1112-1120.
Guo, Z., Viitanen, M., Fratiglioni, L., & Winblad, B. (1996). Low blood pressure
and dementia in elderly people: the Kungsholmen project. British Medical
Journal, 312, 805-808.
Hachinski, V., Iliff, L. D., Zilhka, E., Du Boulay, G. H., McAllister, V. L., Marshall,
J., Russell, R. W. R., & Symon, L. (1975). Cerebral blood flow in dementia.
Archives o f Neurology, 32, 632-637.
Hachinski, V. (1992). Preventable senility: A call for action against the vascular
dementias. Lancet, 340, 645-648.
Hammar, N., Romelsjo, A. & Alfredsson, L. (1997). Alcohol consumption, drinking
pattern and acute myocardial infarction. A case referent study based on the
Swedish Twin Register. Journal o f Internal Medicine, 241, 125-131.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
Hassing, L.B., Johansson, B., Nilsson, S.E., Berg, S., Pedersen, N.L., Gatz, M., &
McClearn, G. (2002). Diabetes mellitus is a risk factor for vascular dementia,
but not for Alzheimer’s disease: A population-based study of the oldest old.
International Psychogeriatrics, 14, 239-248
Hebert, L. E., Scherr, P. A., Beckett, L. A., Albert, M. S., Rosner, B., Taylor, J. O.,
& Evans, D. A. (1993). Relation of smoking and low-to-moderate alcohol
consumption to change in cognitive function: A longitudinal study in a
defined community of older persons. American Journal o f Epidemiology,
137, 881-891.
Hebert, R., Lindsay, J., Verreault, R , Rockwood, K., Hill, G., & Dubois, M. (2000).
Vascular dementia: Incidence and risk factors in the Canadian study of health
and aging. Stroke, 31, 1487-1493.
Henon, H., Pasquier, F., Durieu, I., Godeffoy, O., Lucas, C., Lebert, F., & Leys, D.
(1997). Preexisting dementia in stroke patients: Baseline frequency,
associated factors, and outcome. Stroke, 28, 2429-2436.
Henrik, G., Damber, L., & Damber, J. (1996). Total food consumption and body
mass index in relation to prostate cancer risk: A case-control study in Sweden
with prospectively collected exposure data. Journal o f Urology, 155, 969-
974.
Hill, R. D., Storandt, M., & Malley, M. (1993). The impact of long-term exercise
training on psychological function in older adults. Journal o f Gerontology:
Psychological Sciences, 48, P12-P17.
Hu, F. B., Stampfer, M. J., Colditz, G. A., Ascherio, A., Rexrode, K. M., Willett, W.
C., & Manson, J. E. (2000). Physical activity and risk of stroke in women.
Journal o f the American Medical Association, 283, 2961-2967.
Huppert, F. A., Johnson, T., & Nickson, J. (2000). High prevalence of prospective
memory impairment in the elderly and in early-stage dementia: Findings
from a population-based study. Applied Cognitive Psychology, 14, S63-S81.
Inzitari, D., Di Carlo, A., Pracucci, G., Lamassa, M., Vanni, P., Romanelli, M.,
Spolveri, S., Adriani, P., Meucci, I., Landini, G., & Ghetti, A. (1998).
Incidence and determinants of poststroke dementia as defined by an
informant interview method in a hospital-based stroke registry. Stroke, 29,
2087-2093.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
Johansson, B. (1988/89). The MIR-Memory-in-Reality Test. Psykologif rlaget,
AB, Stockholm, Sweden.
Jonsson, C.-O., & Molander, L. (1964). Manual till CVB-skalan [Manual of the
CVB Scales], Stockholm: PsykologiF rlaget.
Jorm, A. F. (1991). Cross-national comparisons of the occurrence of Alzheimer’s
and vascular dementias. European Archives o f Clinical Neurosciences, 240,
218-222.
Jorm, A. F., Korten, A. E., & Henderson, A. S. (1987). The prevalence of dementia:
A quantitative integration of the literature. Acta Psychiatrica Scandinavica,
76, 465-479.
Kahn, R. L., Pollack, M., & Goldfarb, A. I. (1961). Factors related to individual
differences in mental status of institutionalized aged. In P. Hoch & J. Zubin
(Eds.), Psychopathology o f Aging, (pp. 104-113). New York: Grune &
Stratton.
Kalmijn, S., Laurner, L. J., Ott, A., Witteman, J. C. M., Hofman, A., & Breteler, M.
M. B. (1997). Dietary fat intake and the risk of incident dementia in the
Rotterdam study. Annals o f Neurology, 42, 776-782.
Kase, C. S., Wolf, P. A., Kelly-Hayes, M., Kannel, W. B., Beiser, A., & D’Agostino,
R. B. (1998). Intellectual decline after stroke: The Framingham study.
Stroke, 29, 805-812.
Katzman, R., Aronson, M., Fuld, P., Kawas, C., Brown, T., Morgenstern, H.,
Frishman, W., Gidez, L., Eder, H., & Ooi, W. L. (1989). Development of
dementing illnesses in an 80-year-old volunteer cohort. Annals o f Neurology,
25, 317-324.
Kilander, L., Nyman, H., Boberg, M., & Lithell, H. (1997). Cognitive function,
vascular risk factors and education. A cross-sectional study based on a cohort
of 70-year-old men. Journal o f Internal Medicine, 242, 313-321.
Kiyohara, Y., Yoshitake, T., Kato, I., Ohmura, T., Kawano, H., Ueda, K., &
Fujishima, M. (1994). Changing patterns in the prevalence of dementia in a
Japanese community: The Hisayama study. Gerontology, 40(Suppl. 2), 29-
35.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
Knopman, D., Boland, L. L., Mosley, T., Howard, G., Liao, D., Szklo, M.,
McGovern, P., & Folsom, A. R. (2001). Cardiovascular risk factors and
cognitive decline in middle-aged adults. Neurology, 56, 42-48.
Kokmen, E., Whisnant, J. P., O’Fallon, W. M., Chu, C.-P., & Beard, C. M. (1996).
Dementia after first ischemic stroke: A population-based study in Rochester,
Minnesota (1960-1984). Neurology, 19, 154-159.
Kumari, M., Brunner, E., Fuhrer, R. (2000). Minireview: Mechanisms by which
the metabolic syndrome and diabetes impair memory. Journal o f
Gerontology: Biological Sciences, 55A, B228-B232.
Launer, L. J., Feskens, E. J. M., Kalmijn, S., & Kromhout, D. (1996). Smoking,
drinking, and thinking: The Zutphen Elderly Study. American Journal o f
Epidemiology, 143, 219-227.
Laurin, D., Verreault, R., Lindsay, J., MacPherson, K., & Rockwood, K. (2001).
Physical activity and risk of cognitive impairment and dementia in elderly
persons. Archives o f Neurology, 58^ 498-504.
Leibson, C. L., Rocca, W. A., Hanson, V. A., Cha, R , Kokmen, E., O’Brien, P., C.,
& Palumbo, P. J. (1997). Risk of dementia among persons with diabetes
mellitus: A population-based cohort study. American Journal o f
Epidemiology, 145, 301-308.
Leys, D., Parnetti, L., & Pasquier, F. (1999). Vascular dementia. In M. Fisher & J.
Bogousslavsky (Eds.), Current Review o f Cerebrovascular Disease (3rd ed.)
(pp. 137-147). Philadelphia, PA: Current Medicine, Inc.
Li, G., Shen, Y. C., Chen, C. H., Zhao, Y. W., Li, S. R , & Lu, M. (1989). An
epidemiological survey of age-related dementia in an urban area of Beijing.
Acta Psychiatrica Scandinavica, 79, 557-563.
Lichtenstein, P., De Faire, U., Floderus, B., Svartengren, M., Svedberg, P., &
Pedersen, N. L. (2002). The Swedish Twin Registry: A unique resource for
clinical, epidemiological and genetic studies. Jounral o f Internal Medicine,
252, 184-205.
Lindblad, U., Rastam, L., Ranstam, J., & Peterson, M. (1993). Validity of register
data on acute myocardial infarction and acute stroke: the Skaraborg
Hypertension Project. Scandinavian Journal o f Social Medicine, 21, 3-9.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
Lindsay, J., Hebert, R., & Rockwood, K. (1997). The Canadian study of health and
aging: Risk factors for vascular dementia. Stroke, 28, 526-530.
Loeb, C., Gandolfo, C., Croce, R., & Conti, M. (1992). Dementia associated with
lacunar infarction. Stroke, 23, 1225-1229.
Madden, D. J., Blumenthal, J. A., Allen, P. A., & Emery, C. F. (1989). Improving
aerobic capacity in healthy older adults does not necessarily lead to improved
cognitive performance. Psychology and Aging, < 307-320.
Marmot, M. G., & Poulter, N. R. (1992). Primary prevention of stroke. Lancet,
339, 344-347.
McMurdo, M. E. T., & Burnett, L. (1992). Randomised controlled trial of exercise
in the elderly. Gerontology, 38, 292-298.
Meyer, J. S., Judd, B. W., Tawakina, T., Rogers, R. L., & Mortel, K. F. (1986).
Improved cognition after control of risk factors for multi-infarct dementia.
Journal o f the American Medical Association, 256, 2203-2209.
Molloy, D. W., & Standish, T. I. M. (1997). Mental status and neuropsychological
assessment: A guide to the standardized Mini-Mental State Examination.
International Psychogeriatrics, 9(Suppl. 1), 87-94.
Morris, J. C., Ernesto, C., Schafer, K., Coats, M., Leon, S., Sano, M., Thai, L. J.,
Woodbury, P., & the Alzheimer’s Disease Cooperative Study. (1997).
Clinical Dementia Rating training and reliability in multicenter studies: The
Alzheimer’s disease cooperative study experience. Neurology, 48, 1508-
1510.
Morris, J. C., Heyman, A., Mohs, R. C., Hughes, J. P., van Belle, C., Fillenbaum, G.,
Melits, E. D., & Clark, C. (1989). The consortium to establish a registry for
Alzheimer’s disease (CERAD): I. Clinical and neuropsychological
assessment of Alzheimer’s disease. Neurology, 39, 1159-1165.
Mukamal, K. J., Kuller, L. H., Fitzpatrick, A. L., Longstreth, W. T., Mittleman, M.
A., & Siscovick, D. S. (2002). Prospective study o f alcohol consumption and
risk of dementia in older adults. JAMA, 289, 1405-1413.
Mullen, P. D., Simons-Morton, D. G., Ramirez, G., Frankowski, R. F., Green, L. W.,
& Mains, D. A. (1997). A meta-analysis of trials evaluating patient
education and counseling for three groups of preventive health behaviors.
Patient Education and Counseling, 32, 157-173.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
Muntner, P., Garrett, E., Klag, M. J., & Coresh, J. (2002). Trends in stroke
prevalence between 1973 and 1991 in the US population 25 to 74 years of
age. Stroke, 33, 1209-1213.
Netz, Y., & Jacob, T. (1994). Exercise and the psychological state of
institutionalized elderly: A review. Perceptual and Motor Skills, 79, 1107-
1118.
Nielson, K. A., Nolan, J. H., Berchtold, N. C., Sandman, C. A., Mulnard, R. A., &
Cotman, C. W. (1996). Apolipoprotein-E genotyping of diabetic dementia
patients: Is diabetes rare in Alzheimer’s disease? Journal o f the American
Geriatrics Society, 44, 897-904.
Nyenhuis, D. L., & Gorelick, P. B. (1998). Vascular dementia: A contemporary
review of epidemiology, diagnosis, prevention and treatment. Journal o f the
American Geriatrics Society, 46, 1437-1448.
Okumiya, K., Matsubayashi, K., Wada, T., Kimura, S., Doi, Y., & Ozawa, T.
(1996). Effects of exercise on neurobehavioral function in community-
dwelling older people more than 75 years of age. Journal o f the American
Geriatrics Society, 44, 569-572.
Ott, A., Slooter, A. J. C., Hofman, A., van Harskamp, F., Witteman, J. C. M., Van
Broeckhoven, C., van Duijn, C. M., & Breteler, M. M. B. (1998). Smoking
and risk of dementai and Alzheimer’s disease in a population-based cohort
study: The Rotterdam Study. Lancet, 351, 1840-1843.
Ott, A., Stolk, R. P., Hofman, A., van Harskamp, E., Grobbee, D. E., & Breteler, M.
M. B. (1996). Association of diabetes mellitus and dementia: The
Rotterdam study. Diabetologia, 39, 1392-1397.
Ott, A., Stolk, R. P., van Harskamp, F., Pols, H. A. P., Hofman, A., & Breteler, M.
M. B. (1999). Diabetes mellitus and the risk of dementia: A Rotterdam
study. Neurology, 53, 1937-1942.
Palleschi, L., Vetta, F., De Gennaro, E., Idone, G., Sottosanti, G., Gianni, W., &
Marigliano, V. (1996). Effect of aerobic training on the cognitive
performance of elderly patients with senile dementia of Alzheimer’s type.
Archives o f Gerontology and Geriatrics^ Suppl. 5), 47-50.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
Plassman, B. L., & Breitner, J. C. S. (1996). Recent advances in the genetics of
Alzheimer’s disease and vascular dementia with an emphasis on gene-
environment interactions. Journal o f the American Geriatrics Society, 44,
1242-1250.
Pohjasvaara, T., Erkinjuntti, T., Ylikoski, R., Hietanen, M., Vataja, R., & Kaste, M.
(1998). Clinical determinants of poststroke dementia. Stroke, 29, 75-81.
Prencipe, M., Ferretti, C., Casini, A. R , Santini, M., Giubilei, F., & Culasso, F.
(1997). Stroke, disability, and dementia: Results of a population survey.
Stroke, 28, 531-536.
Reynolds, K., Lewis, L. B., Nolen, J. D. L., Kinney, G. L., Sathya, B., & He, J.
(2003). Alcohol consumption and risk of stroke: A meta-analysis. JAMA,
289, 579-588.
Rogers, R. L., Meyer, J. S., & Mortel, K. F. (1990). After reaching retirement age
physical activity sustains cerebral perfusion and cognition. Journal o f the
American Geriatrics Society, 38, 123-128.
Roman, G. C., Tatemichi, T. K., Erkinjuntti, T., Cummings, J. L., Masdeu, J. C.,
Garcia, J. H., Amaducci, L., Orgogozo, J. -M., Brun, A., Horman, A.,
Moody, D. M., O’Brien, M., D., Yamaguchi, T., Grafman, J., Drayer, B. P.,
Bennertt, D. A., Fisher, M. A., Ogata, J., Kokmen, E., Bermejo, F., Wolf, P.
A., Gorelick, P. B., Bick , K. L., Paueau, A. K., Bett, M. A., DeCarli, C.,
Culebras, A., Korczyn, A. D., Bogousslavsky, J., Hartmann, A., &
Scheinberg. (1993). Vascular dementia: Diagnostic criteria for research
studies: Report of the NINCS-AIREN Internal workshops. Neurology, 43,
250-260.
Rosamond, W. D., Folsom, A. R., Chambless, L. E., Wang, C., McGovern, P.G.,
Howard, G., Copper, L. S., & Shahar, E. (1999). Stroke incidence and
survival among middle-aged adults: 9-year follow-up of the atherosclerosis
risk in communities (ARIC) cohort. Stroke, 30, 736-743.
Ross, G. W., Petrovitch, H., White, L. R , Masaki, K. H., Li, C. Y., Curb, J. D.,
Yano, K., Rodriguez, B. L., Foley, D. J., Blanchette, P. L., & Havlik, R.
(1999). Characterization of risk factors for vascular dementia: The Honolulu-
Asia Aging study. Neurology, 53, 337-343.
Samsa, G. P., Cohen, S. J., Goldstein, L. B., Bonito, A. J., Duncan, P. W., Enarson,
C., DeFriese, G. H., Horner, R. D., & Matchar, D. B. (1997). Knowledge of
risk among patients at increased risk of stroke. Stroke, 28, 916-921.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
Satoh, T., Sakurai, I., Miyagi, K., & Hohshaku, Y. (1995). Walking exercise and
improved neuropsychological functioning in elderly patients with cardiac
disease. Journal o f Internal Medicine, 238, 423-428.
Shinton, R., & Beevers, G. (1989). Meta-analysis of relation between cigarette
smoking and stroke. British Medical Journal, 295 789-794.
Skoog, I. (1994). Risk factors for vascular dementia: A review. Dementia, 5, 137-
144.
Skoog, I. (1997). The relationship between blood pressure and dementia: A review.
Biomedicine & Pharmacotherapy, 51, 367-375.
Skoog, I. (1998). Status of risk factors for vascular dementia. Neuroepidemiology,
17, 2-9.
Skoog, I., Nilsson, L., Palmertz, B., Andreasson, L., & Svanborg, A. (1993). A
population-based study of dementia in 85-year-olds. New England Journal o f
Medicine, 328, 153-158.
Smith, J. W. (1995). Medical manifestations of alcoholism in the elderly.
International Journal o f the Addictions, 30, 1749-1798.
Smith, J. W., & Atkinson, R. M. (1995). Alcoholism and dementia. International
Journal o f the Addictions, 30x 1843-1869.
Stegmayr, B., & Asplund, K. (1995). Diabetes as a risk factor for stroke: A
population perspective. Diabetologia, 38, 1061-1068.
Stone, M. J., & Dawe, D. (1993). Acute exercise facilitates semantically cued
memory in nursing home residents. Journal o f the American Geriatrics
Society, 41, 531-534.
Tatemichi, T. K., Desmond, D. W., Mayeux, R., Paik, M., Stern, Y., Sano, M.,
Remien, R. H., Williams, J. B. W., Mohr, J. P., Hauser, W. A., & Figueroa,
M. (1992). Dementia after stroke: Baseline frequency, risks, and clinical
features in a hospitalized cohort. Neurology, 42, 1185-1193.
Tatemichi, T. K., Desmond, D. W., Paik, M., Figueroa, M., Gropen, T. I., Stern, Y.,
Sano, M., Remien, R., Williams, J. B. W., Mohr, J. P., & Mayeux, R. (1993).
Clinical determinants of dementia related to stroke. Annals o f Neurology, 33,
568-575.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
Terry, P., Baron, J. A., Weiderpass, E., Yuen, J., Lichtenstein, P. & Nyren, O.
(1999). Lifestyle and endometrial cancer risk: A cohort study from the
Swedish Twin Registry. International Journal o f Cancer, 82, 38-42.
Ueda, K., Hasuo, Y., Kiyohara, Y., Wada, J., Kawano, H., Kato, I., Fujii, I., Yanai,
T., Omae, T., & Fujishima, M. (1988). Intracerebral hemorrhage in a
Japanese community, Hisayama: Incidence, changing pattern during long
term follow-up, and related factors. Stroke, 19, 48-52.
Ueda, K., Kawano, H., Hasuo, Y., & Fujishima, M. (1992). Prevalence and etiology
of dementia in a Japanese community. Stroke, 23, 798-803.
U.S. Department of Agriculture, Agricultural Research Service. (2002). USDA
National Nutrient Database fo r Standard Reference, Release 15. Retrieved
March 3, 2003, from http://www.nal.usda.gov/fnic/foodcomp.
Verhey, F. R. J., Lodder, J., Rozendaal, N. & Jolles, J. (1996). Comparison of seven
sets of criteria used for the diagnosis of vascular dementia.
Neuroepidemiology, 15, 166-172.
Watari, K., Gatz, M., & Pedersen, N. L. (2002, November). Agreement of self-
reported stroke with a hospital registry in Sweden [Abstract], Gerontologist,
42 (Special Issue I), 209.
Weibe-Velazquez, S., & Hachinski, V. (1991). Overview of clinical issues in
stroke. In R. A. Bomstein & G. G. Brown (Eds.), Neurobehavioral aspects o f
cerebrovascular disease, (pp. 110-130). New York: Oxford University
Press.
Wetterling, T., Kanitz, R , & Borgis, K. (1996). Comparison of different diagnostic
criteria for vascular dementia (ADDTC, DSM-IV, ICD-10, NINDS-AIREN).
Stroke, 27, 30-36.
World Health Organization. (1967). Manual o f the International Statistical
Classification o f Diseases, Injuries, and Causes o f Death (8th Rev.). Geneva:
Author.
World Health Organization. (1977). Manual o f the International Statistical
Classification o f Diseases, Injuries, and Causes o f Death (9th Rev.). Geneva:
Author.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
World Health Organization. (1992-1993). International Statistical Classification o f
Diseases and Related Health Problems (10th Rev.). Geneva: Author.
Yoshitake, T., Kiyohara, Y., Kato, I., Ohmura, T., Iwamoto, H., Nakayama, K.,
Ohmori, S., Nomiyama, K., Kawano, H., Ueda, K., Sueishi, K., Tsuneyoshi,
M., & Fujishima, M. (1995). Incidence and risk factors of vascular dementia
and Alzheimer’s disease in a defined elderly Japanese poulation: The
Hisayama study. Neurology, 45, 1161-1168.
Zhu, L., Fratiglioni, L., Zhenchao, G., Agrjero-Torres, H., Winblad, B., & Viitanen,
M. (1998). Association of stroke with dementia, cognitive impairment, and
functional disability in the very old: A population-based study. Stroke, 29,
2094-2099
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
Appendix A
International Classification of Diseases (ICD) - Cerebrovascular Disease
(Includes major headings and ICD codes)
ICD - 8th Revision
Cerebrovascular disease (430-438)
430 Subarachnoid haemorrhage
431 Cerebral haemorrhage
432 Occlusion of pre-cerebral arteries
433 Cerebral thrombosis
434 Cerebral embolism
435 Transient cerebral ischaemia
436 Acute but ill-defined cerebrovascular disease
437 Generalized ischaemic cerebrovascular disease
438 Other and ill-defined cerebrovascular disease
ICD - 9th Revision
Cerebrovascular Disease (430-438)
430 Subarachnoid hemorrhage
431 Intracerebral hemorrhage
432 Other and unspecified intracranial hemorrhage
433 Occlusion and stenosis of precerebral arteries
434 Occulsion of cerebral arteries
435 Transient cerebral ischemia
436 Acute, but ill-defined, cerebrovascular disease
437 Other and ill-defined cerebrovascular disease
438 Late effects of cerebrovascular disease
ICD - 10th Revision
Cerebrovascular Disease (160-169)
160 Subarachnoid haemorrhage
161 Intracerebral haemorrhage
162 Other nontraumatic intracranial haemorrhage
163 Cerebral infarction
164 Stroke, not specified as haemorrhage or infarction
165 Occlusion and stenosis of precerebral arteries, not resulting in cerebral infarction
166 Occlusion and stenosis of cerebral arteries, not resulting in cerebral infarction
167 Other cerebrovascular diseases
168 Cerebrovascular disorders in diseases classified elsewhere
169 Sequelae of cerebrovascular disease
G45 Transient cerebral ischaemic attacks and related syndromes
Note: ICD - 7t h Revision codes not listed; only relevant to one participant.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
Appendix B
Swedish Twin Registry Measures
OLD COHORT
1961
Cardiovascular History
Question____________________ Answer Options
High blood pressure l=ill
0=not ill
Cerebral haemorrhage
Diabetes
1967
Smoking
Question____________________ Answer Options
Blue Smoking status 0=non-smoker
l=presently a smoker
2=previously a smoker
9=unknown
Blue Smoking group 0=non-smoker
l=cigar and pipe
2=2-10 cgt/d
3=11-20 cgt/d
4=21- cgt/d
5=2-10 cgt/d + cgr/pipe
6=11-20 cgt/d + cgr/pipe
7=21- cgt/d + cgr/pipe
8=1 cgt/d
9=unknown
Alcohol
Question_________________________________________________ Answer Options
Have you at any time during the past year consumed beer, wine, or 0=no
l=yes
other alcoholic drinks (liquor)?
l=almost daily
2=once or twice a week
3=a few times a month
How often do you usually drink beer?
4=once or twice a month
5=once or twice a year
6=less often
7=never
l=one glass or two
2=one bottle
3=two bottles
On a day when you drink beer, how much do you usually drink?
4=three bottles or more
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
How often do you usually drink wine?
On a day when you drink wine, how much do you usually drink?
How often do you usually drink liquor?
How much do you usually drink when you drink hard liquor?
l=almost daily
2=once or twice a week
3=a few times a month
4=once or twice a month
5=once or twice a year
6=less often
7=never
l=a wine-glass or two
2=a half bottle
3=half a bottle-one
bottle
4=one bottle or more
l=almost daily
2=once or twice a week
3=a few times a month
4=once or twice a month
5=once or twice a year
6=less often
7=never
l=less than 1 shot
2=1-3 shots
3=3-7 shots
4=more than 7 shots
Blue Alcohol status
0=never drinks
l=presently a drinker
2=previously a drinker
9=unknown
Diet
Question Answer Options
What proportions of the cooked foods you eat?
Pork
Sausage, chopped meat, other
Beef or similar meat
Fish and seafood
Fruit and vegetables
How often do you eat cakes and pastries?
How often do you eat candies?
l=great part
2=medium part
3=small part
4=no part
l-2=several times a day
3-4=once a day
5-6=more seldom
l-2=several times a day
3-4=once a day
5-6=more seldom
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
Physical Activity
Question___________________________________ Answer Options
How much physical exercise have you had from age 25-
l=hardly any physical exercise
50?
2=light exercise e.g. regular walks,
light gardening
3=regular exercise
4=hard physical training
1970
Smoking
Question_______________________ Answer Options
Mahogany Smoking status 0=non-smoker
l=presently a smoker
2=previously a smoker
9=unknown
Mahogany Smoking group 0=non-smoker
l=cigar and pipe
2=2-10 cgt/d
3=11-20 cgt/d
4=21- cgt/d
5=2-10 cgt/d + cgr/pipe
6=11-20 cgt/d + cgr/pipe
7=21- cgt/d + cgr/pipe
8=1 cgt/d
9=unknown
Alcohol
Question Answer Options
Have you at any time during the past year consumed beer, wine, or
other alcoholic drinks (liquor)?
How often do you usually drink beer?
On a day when you drink beer, how much do you usually drink?
How often do you usually drink wine?
0=no
l=yes
l=almost daily
2=once or twice a week
3=a few times a month
4=once or twice a month
5=once or twice a year
6=less often
7=never
l=one glass or two
2=one bottle
3=two bottles
4=three bottles or more
l=almost daily
2=once or twice a week
3=a few times a month
4=once or twice a month
5=once or twice a year
6=less often
7=never
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I l l
On a day when you drink wine, how much do you usually drink?
How often do you drink liquor?
How much do you usually drink when you drink hard liquor?
l=a wine-glass or two
2=a half bottle
3=half a bottle - one bottle
4=one bottle or more
l=almost daily
2=once or twice a week
3=a few times a month
4=once or twice a month
5=once or twice a year
6=less often
7=never
l=less than 1 shot
2=1-3 shots
3=3-7 shots
4=more than 7 shots
Mahogany Alcohol status 0=never drinks
l=presently a drinker
2=previously a drinker
9=unknown
Physical Activity
Question Answer Options
How much physical exercise have you had from age l=hardly any physical exercise
25-50? 2=light exercise e.g. regular walks,
light gardening
3=regular exercise
4=hard physical training
TELEPHONE SCREENING (1998 - 2000)
Diabetes
Questions______________________________________________________ Answer options
Do you have or have you had diabetes (including old age diabetes)
(excluding pregnancy diabetes)?
Yes
No
Don’t know
Refuse
You mentioned earlier that you have or have had diabetes. At what age did
you get diabetes?
Age
Year
Don’t know
Refuse
Blood Pressure
Questions_______________________________________________________________ Answer options
Do you have or have you had high blood pressure? Yes
No
Don’t know
Refuse
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
Smoking
Questions______________________________________________ Answer options
Have you ever smoked or used snuff? No, not even tried it
Yes, but only tried it
Smoked now and then (e.g., at parties)
Smoked regularly
Smoked at parties
Used snuff now and then (e.g., at parties)
Snuffed regularly
Smoke now and then
Smoke regularly
Smoke at parties
Snuff now and then
Snuff regularly
Don’t know
Refuse
Alcohol
Questions Answer
options____________________________________________________ __________________
Think about your use of alcohol over your entire life. Has there ever
been a period in your life when you drank too much?
Yes
No
Don’t know
Refuse
Has there ever been a period in your life when someone else objected to
your drinking, i.e., one person objected more than once or several people
objected?
Yes
No
Don’t know
Refuse
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Depression as a risk factor for dementia

PDF

Experimental evaluation of a new intervention designed to promote dementia caregivers' acceptance and empathic responding towards care -recipients

PDF

Dementia caregiving and ethnicity: African American caregivers and the sociocultural stress and coping model

PDF

History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample

PDF

Effects of cultural values and behaviors in sociocultural stress and coping among Korean-American and Korean caregivers

PDF

Hispanic culture, acculturation, and distress among caregivers of dementia patients

PDF

An evaluation of a brief HIV /AIDS prevention intervention using normative feedback to promote risk reduction among sexually active college students

PDF

Associations and mechanisms among attention deficit hyperactivity disorder symptoms, cognitive functioning, and drinking habits

PDF

Cross -cultural examination of mental health measures in dementia caregivers: Assessment of the Center for Epidemiologic Studies -Depression Scale and the Zarit Burden Inventory

PDF

Education and intelligence test scores: Predictors of dementia?

PDF

Childhood videotaped neuromotor and social precursors of schizophrenia: A prospective investigation

PDF

Determining factors that predict physical activity behavior and compliance with an intervention program in Latino women

PDF

Anorexia nervosa: A developmental perspective

PDF

Impact of language and culture on a neuropsychological screening battery for Hispanics

PDF

Emotional intelligence and smoking risk factors in early adolescents

PDF

FGF8, isoform b, is an etiological factor in prostate tumorigenesis

PDF

Alcohol expectancies and consumption: Age and sex differences

PDF

Hostile family environments and children's perceptions of social support

PDF

Emotion regulation as a mediator between family-of-origin aggression and marital aggression

PDF

Alcohol treatment entry and refusal in a sample of older veterans

Asset Metadata

Creator Watari, Kecia Fumi (author)

Core Title Frequency and risk factors of poststroke dementia

School Graduate School

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Gerontology,health sciences, pathology,OAI-PMH Harvest,Psychology, clinical

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Gatz, Margaret (committee chair), Howell, Sandra (committee member), Johnson, C. Anderson (committee member), Knight, Bob G. (committee member), Pedersen. Nancy (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-498683

Unique identifier UC11335685

Identifier 3133351.pdf (filename),usctheses-c16-498683 (legacy record id)

Legacy Identifier 3133351.pdf

Dmrecord 498683

Document Type Dissertation

Rights Watari, Kecia Fumi

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA