University of Southern California Dissertations and Theses

Are life events differentially associated with dementia risk by gender? A twin study

(USC Thesis Other)

Are life events differentially associated with dementia risk by gender? A twin study

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Are Life Events Differentially Associated with Dementia Risk by Gender?
A Twin Study
by
Alice Jinna Kim, B.A.
A Thesis Presented to the
FACULTY OF THE USC DORNSLIFE COLLEGE OF LETTERS, ARTS AND SCIENCES
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements of the Degree
MASTER OF ARTS
PSYCHOLOGY
December 2020
Copyright 2020 Alice Jinna Kim

ii
Acknowledgements

I would first like to express my sincere gratitude to my thesis advisor, Dr. Christopher Beam, for
his constant guidance and encouragement. I could not have asked for a better advisor and
mentor on this journey.

I would also like to thank Drs. Margaret Gatz, David Walsh, and Jennifer Ailshire, to whom I am
indebted for their valuable comments on this thesis.

Finally, I am grateful for the love and support of my family – my parents, Mi J. and David K. S.
Kim; my sister, Hannah S. Kim; my grandparents, Woonam and John E. Hahn and Chun O. and
Sang M. Kim; and my partner, Jordan Limperis. This accomplishment could not have been
possible without them. Thank you.

iii
TABLE OF CONTENTS
Acknowledgements……………………………………………………………………………………………………………………..….ii

List of Tables……………………………………………………………………………………………………………….………………..iv

List of Figures………………………………………………………………………………………………………………………………..v

Abstract…………………………………………………………………………………………………………………………………….....…vi

Introduction………………………………………………………………………………………………………………………………….....1
Life Events, Stress, and Health: Social Causation…………………………………..……………...2
Life Events and Dementia: Social Causation or Social Selection?.......................................4
Life Events and Dementia: The Role of Gender?..............................................................................6
Twin Studies: Social Selection vs. Social Causation……………………………………..…….12
Biological Mechanisms for Social Causation………………………………………………………14
Proposed Study……………………………………………………………………………………………………..16

Methods………………………………………………………………………………………………………………………………………...17
Sample………………………………………………………………………………………………………………..…17
Measures…………………………………………………………………………………………………………….…18
Data Analysis………………………………………………………………………………………………………..20

Results…………………………………………………………………………………………………………………………..………………23
Confirmatory Factor Analyses……………………………………………………………………………...23
Phenotypic Analyses……………………………………………………………………………………...……..27
Cross-Twin Cross-Trait Logistic Regressions……………………………………………………...28
Bivariate Twin Models……………………………………………………………………………………..…..30

Discussion………………………………………………………………………………………………………………………………...…..34

References………………………………………………………………………………………………………………………...…………..39

Appendices……………………………………………………………………………………………………………………………..…….55
Appendix A: Items in the modified version of the Social Readjustment Rating
Scale in SATSA……………………………………………………………………………55
Appendix B: Results of exploratory factor analysis on life event items………………56

iv
List of Tables

Table 1. Goodness-of-Fit Indicators of Models for Measurement Invariance by Gender………......24
Table 2. Confirmatory Factor Analysis for Men – Factors and Loadings……………………...……………25
Table 3. Confirmatory Factor Analysis for Women – Factors and Loadings…………………...................26
Table 4. Phenotypic Effect of Life Events on Dementia by Gender………………………......................…….27
Table 5. Model-Fitting Results for Bivariate Twin Models – Men…………………........................….…….....31
Table 6. Parameter Estimates from Best-Fitting Models – Men……………………….……...............................32
Table 7. Model-Fitting Results for Bivariate Twin Models – Women………………….....................……….33
Table 8. Parameter Estimates from Best-Fitting Models – Women………………………….......................….34

v
List of Figures

Figure 1. Flowchart of Participants Included in Present Study………………………………..........................….18
Figure 2. Biometric Regression Model for Life Event Factor and Dementia……...............……………...23
Figure 3. Forest Plot of Cross-Twin Cross-Trait Odds Ratios by Zygosity and Life Event Factor
for Men…………………………………………………………….............................................................……………….…..28
Figure 4. Forest Plot of Cross-Twin Cross-Trait Odds Ratios by Zygosity and Life Event Factor
for Women………………………………………………………...........................................................……………….……29

vi
Abstract
Stressful life events are associated with cognitive aging and dementia risk. It remains unclear
whether negative and positive life events are causally related to dementia risk and if existing
associations are accounted for by genetic and shared environmental confounds. Further, it is
unclear whether these life events are differentially associated with dementia risk by gender.
Gender may affect the pathways through which different life events correlate with dementia risk.
This study examined whether there is support for a causal association between six different life
event domains and dementia risk by gender. The study tested and controlled for the confounding
influence of genetics and shared environmental factors in explaining the association between
these six life event domains and dementia risk for men and women. Data from the first three
waves of the Swedish Adoption/Twin Study of Aging (SATSA) were used (MZM = 199, MZF =
269, DZM = 345, DZF = 556). Life events were assessed using standard measures modified for
older adults. Dementia diagnoses were made based on in-person interviews or Swedish national
registry linkage. Structural equation modeling was used to evaluate the hypothesis that there is a
causal relationship between six life event domains and dementia risk by gender, as well as
quantify and account for shared genetic and environmental influences that may confound the
association. We hypothesized that life events would differentially associate with dementia risk
by gender via both familial confounding and causal processes. There was little evidence for
causal processes for men and women across life event domains. Familial confounding (genes
and common environment) appeared to drive the observed phenotypic associations. Gender
differences were found for phenotypic associations as well as between-family effects.

1

Are Life Events Differentially Associated with Dementia Risk by Gender?: A Twin Study
Introduction
Lifetime stress exposure has been shown to be associated with increased risk of
psychological and physical health outcomes including depression and acute coronary heart
disease (Kendler, Karkowski, & Prescott, 1999; Rafanelli et al., 2005; Slavich, 2016). Life
events – particularly negative or undesirable life events – are a commonly used measurement of
lifetime stress exposure (Holmes & Rahe, 1967; Brugha & Cragg, 1990). In using life events,
researchers often are interested in causal interpretation of the association with health outcomes,
such that stressful life events themselves cause disease.
Increases in stressful life events may also correlate with heightened dementia risk.
However, the relation between life events and dementia has not been consistently observed.
Some studies show that specific negative life events, like widowhood and late-life disability, are
related to dementia risk (Persson & Skoog, 1996; Norton et al., 2009) while another reported no
association between specific negative life events and dementia (Sundstrom et al., 2014). One
study reported no association with aggregate negative life events and dementia, yet another team
found two or more negative life events predict increased risk (Sundstrom et al., 2014; Gerritsen
et al., 2017). Studies on positive life events are fewer and similarly inconclusive (Sundstrom et
al., 2014; Sundstrom et al., 2016).
In addition, gender differences in the association between life events and dementia have
rarely been explored. A large body of literature exists on the role of gender in stress exposure,
reactivity, and coping. The genetic risk conferred by dementia-related alleles may also differ by
men and women. As such, different life events may correlate with dementia risk through
different pathways for men and women.

2

One of the primary difficulties in research on life events and dementia is establishing the
causal relation between life events and dementia risk, as random assignment to life events and
dementia status is not feasible. Though a causal interpretation of correlations between life events
and dementia is of interest to most researchers, social selection factors may contribute towards or
completely account for the correlations. In social selection, the association between life events
and dementia reflects the origin of life event risk rather than the effects of life event risk as in
social causation (Johnson, Turkheimer, Irving, & Bouchard, 2009).
Genetically informed research designs, such as twin studies, are one solution
(Turkheimer & Harden, 2014), as they test and adjust for unmeasured genetic and shared
environmental confounds that may obscure observed associations between life events and
dementia risk. Thus, these designs are a particularly rigorous way of strengthening causal
arguments. To my knowledge, there are no published studies examining the causal relation
between life events and dementia using genetically informed designs.
The proposed study will examine whether six different life event domains predict risk of
dementia in men and women, accounting for genetic and shared environmental confounds.
Life Events, Stress, and Health: Social Causation

Much of the research on life events has developed from the stress literature. Stress is
state of arousal that stems from a discrepancy between environmental circumstances and an
individual’s ability to adapt to them. External demands requiring adaptive responses are called
stressors; some stressors easily elicit adaptive responses in most people, while adjustment to
others are more difficult (Aneshensel, 1992). Researchers discovered highly similar levels of
adaptation were required for life events such as marriage, death of a child, and retirement across
populations (Holmes & Rahe, 1967; Rahe, Lundberg, Theorell, & Bennett, 1971). Thus, life

3

events were conceptualized as stressors when they evoked significant changes in most people’s
normal routines (Holmes & Rahe, 1967; Kobasa, 1979). Negative or undesirable events such as
personal illness and positive or desirable events such as marriage could both qualify as stressors
by causing significant changes to a normal routine (Holmes & Rahe, 1967). Further research
suggested the most stressful life events were undesirable life events (Ross & Mirowsky, 1979).
Negative life events correlate with a multitude of outcomes including depression,
rheumatoid arthritis, and variability in HIV progression (Hammen, 2005; Cutolo & Straub, 2006;
Leserman et al., 2002). As in earlier studies, researchers have examined both the effect of
individual life events as well as the aggregate effect of total life events (Lillberg et al., 2003;
Rafanelli et al., 2005). Researchers taking a stress perspective on life events are often interested
in how they impact health outcomes.
Methodological advances in meta-analyses and an increase of life event studies have also
allowed researchers to make broader statements regarding the general effect of life events on
different health outcomes. One meta-analysis examining life events and breast cancer risk did
not reveal an association for overall life events but did find modest support for a positive
relationship between death of spouse and risk of breast cancer (Duijts, Zeegers, & Borne, 2003).
Other meta-analyses revealed support for an association between total life event exposure and
increased risk of psychotic disorder (n = 13, OR = 3.19), exacerbation in multiple sclerosis (n =
14, d = .53), and depression (n = 24, r = .15; Beards et al., 2013; Mohr et al., 2004; Kraaij,
Arensman, & Spinhoven, 2002). The possibility that life events might be reliably associated
with health outcomes has become more salient and testable.
For many of these studies on life events and health, researchers are interested in causal
interpretation of the correlations. Often, the hypothesized causal direction points from life

4

events towards health outcomes: do stressful life events cause disease? Does cumulative stress
from total life events causes one to experience a higher risk of dementia? Thus, the stress
perspective frequently incorporates a social causation hypothesis, which suggests that stressful
life events themselves might cause undesirable health outcomes. Though the social causation
hypothesis is preferred by most researchers, an alternative hypothesis, social selection, may also
contribute towards or account for observed associations between life events and health outcomes.
Life Events and Dementia: Social Causation or Social Selection?

As the second leading cause of disability for adults over 70 years and the seventh leading
cause of death, dementia is a major health outcome of interest. The syndrome entails
deterioration of cognitive function and significant impairment of one’s ability to perform daily
activities. There are around 50 million people living with dementia worldwide and as the older
adult population increases, the number of people with dementia is expected to reach 152 million
people by 2050 (“Towards a dementia plan,” 2018). In order to address the public health impact
of dementia, it is important to identify modifiable risk factors for the prevention or delay of
disease onset.
Life events measures encompass the intersection of modifiable risk factors related to
social engagement, socioeconomic conditions, and physical health conditions considered to raise
risk of dementia (Livingston et al., 2017). Yet, compared to other health outcomes, the literature
on whether and how individual and aggregate negative and positive life events affect dementia
risk remain understudied and report inconsistent associations.
Specific negative life events inconsistently correlate with risk of dementia. Retirement, a
life event that has both desirable and undesirable aspects, has been shown to confer a higher risk
of developing dementia if it occurs at earlier ages compared to later ages (Dufouil et al., 2014).

5

Late-life disability in adults without dementia also predicts dementia incidence (Fauth et al.,
2012). Having a spouse or child suffering from severe illness in late-life has also been
implicated in higher risk of dementia (Person & Skoog, 1996; Norton et al., 2009) and mild
cognitive impairment (MCI; Pilleron et al., 2015). Widowhood increases dementia risk
(Hakansson et al., 2009; Sundstrom, Westerlund, & Kotrylo, 2016; Sommerlad et al., 2018) as
does divorce (Hakansson et al., 2009; Sundstrom et al., 2016). However, one meta-analysis did
not find any association with dementia for divorce (Sommerlad et al., 2018).
Specific positive life events predict decreased risk of dementia. Entering marriage at
mid-life and late-life correlates with lower risk of dementia (Sundstrom et al., 2016; Sommerlad
et al., 2018; Hakansson et al., 2009). Having more frequent social engagements in late-life is
also correlated with reduced risk of dementia (Sorman et al., 2015; Wang et al., 2002).
Yet, null findings have been reported for many specific life events examined in several
studies. Thirteen individual negative events occurring in mid-life or late-life (Persson & Skoog,
1996; Pilleron et al., 2015) and 56 individual negative and positive life events occurring in mid-
life and late-life (Perris, 1984; Sundstrom, Ronnlund, Adolfsson, & Nilsson, 2014) were not
associated with dementia incidence. Thus, one conclusion might be that confounds unaccounted
for in these studies obscure the observed associations and lead to inconsistent findings.
Life events have also been inconsistently shown to have an aggregate risk for dementia.
Persson & Skoog (1996) found a relationship between all examined negative life events and
dementia incidence such that of 3% of those reporting no events developed dementia, 8% of
those reporting one to two events developed dementia, and 20% of those reporting three or more
events developed dementia. Gerritsen et al. (2017) also found that having experienced two or
more negative life events at baseline put subjects at higher risk of incident all-cause dementia

6

and particularly vascular dementia. Johansson et al. (2013) reported that greater total negative
psychosocial stressors in midlife predicted higher incidence of late-life dementia and
Alzheimer’s disease. Peary et al. (2012) found that experiencing a relatively greater number of
highly stressful life events was associated with transition from MCI to dementia. Greenberg et
al. (2014) conducted a literature review on the role of cumulative stressful life event exposure on
negative long-term cognitive outcomes. The authors found reasonable support for the role of life
stress, with several caveats: existing studies show only modest effects and future studies need to
better control for confounding factors that complicate causal inferences. Null results have also
been reported for any association between aggregate negative life events and incident dementia
(Fountoulakis, Pavlidis, & Tsolaki, 2011; Sundstrom et al., 2014; Pilleron et al., 2015), although
total negative life events correlated with MCI (Pilleron et al., 2015).
There is currently no evidence for a protective effect of aggregate positive life events on
dementia risk (Sundstrom et al., 2014). Sundstrom et al. (2014) and Pilleron et al. (2015)
reported that people who eventually developed dementia in their studies reported fewer positive
life events than those who did not develop dementia. A limitation of these studies is that life
event measures were based on self-report, so persons with dementia may have had limited ability
to recall life events. Again, these inconsistent findings for aggregate life event risk as with
individual life event risk may reflect the presence of uncontrolled confounds such as social
selection effects.
Life Events and Dementia: The Role of Gender?
Gender may complicate the association between stressful life events and dementia. First,
gender may determine differential exposure to life events; in objective terms, certain types of life
events may be experienced more commonly for one gender relative to another. In the context of

7

longer life expectancies, larger social networks, and sociocultural norms toward intimacy,
nurturance, and connection, women are more likely to experience the loss of spouses, siblings,
and friends (Haug & Folmar, 1986; McLaughlin et al., 2010; Felmlee & Muraco, 2009; Matud,
2004). Related to those same factors, women are more likely to serve as caregivers (Stoller,
1990); in reporting life events over a two-year span, women were more likely to report network
events (Matud, 2004). For older cohorts, men may be more likely to experience chronic work-
related stress and retirement after employment (Matud, 2004). The cumulative effect of certain
stressors may then disproportionately affect one gender over another.
Second, if the frequency of certain life events tends to differ by gender, the subjective
meaning, related responsibilities, and implications of the event may differ by gender as well. For
example, women tend to experience a greater financial cost of widowhood, particularly for older
cohorts in which women forwent careers due to familial responsibilities (Halleröd, 2013).
Caregiving for a spouse is generally a more normative task for women than for men; as such,
when men are placed in caregiving roles, there may be additional stressors related to lack of role
familiarity. Swinkels et al. (2019) found that when more hours of care were required by men,
their sense of caregiving burden increased; this was not true for women. The study also reported
that while women reported greater burden than men overall, the gap was mediated by more
secondary stressors experienced by women (i.e., relational and financial problems). A meta-
analysis of 229 studies reported that women performing care work reported more hours of care
provision, types of caregiving tasks, care recipient behavioral issues, and burden (Pinquart &
Sörensen, 2006). These differences in meaning, responsibilities, and implications can then lead
to differential associations with dementia risk by gender.

8

Several studies have reported a higher subjective experience of stress for women in
response to undesirable network life events (Kressler & McLeod, 1984; Oman & King, 2000;
Kendler et al., 2001). This heightened vulnerability to network stress may put women at greater
risk of health consequences. Kressler & McLeod (1984) proposed that women might incur
vulnerability to the impact of stressful network life events through the cost of emotional caring;
if women are more emotionally involved in the lives of others, they are likely more liable to the
stressful effects of that emotional involvement. Additionally, if women are more likely to be
called upon as caregivers or supports, they are more liable to being overwhelmed by the overall
number of emotional demands. Kressler & McLeod (1984) noted that serious life crises
involving spouses and children were rated as similarly stressful for men and women in their
study; it was in the distal network events that women experienced significant stress while men
did not.
Some studies also suggest that women report higher overall appraisals of stress – not just
for network events (Miller & Kirsch, 1987; Kroenke & Spitzer, 1998; Troisi, 2001). The
experience of gender is enshrined within social, historical, and political context. Women have
generally been disadvantaged by unequal access to and control over resources and decision-
making power (Ravindran, 2002). Men tend to be overrepresented in prestigious occupational
positions while women continue to be more responsible for household labor, even as they take on
increasing occupational responsibility (Roos & Reskin, 1984; U. S. Bureau of Labor Statistics,
2019). As such, a general tendency for women to experience events as more stressful and less
controllable is not surprising and may put women are disproportionately higher risk for
dementia.

9

Third, certain life events may elicit different types of coping by gender, which may have
differential downstream effects on health outcomes; this would make sense if the life event in
question is experienced more frequently or holds different implications by gender. Such a result
would be somewhat consistent with the role-constraint hypothesis, which asserts that gender
differences in coping are due to the differential likelihood of men and women to occupy certain
types of roles and access role-related resources and opportunities; this corresponds to differential
frequency of life event exposure (e.g., the role of caregiver and the life event of caregiving).
However, the role-constraint hypothesis also asserts that men and women would cope in the
same way given similar social roles (e.g., if women and men were both caregivers, they would
cope in the same way); this ignores the possibility of different meanings, responsibilities, and
implications of the role for men and women (Rosario et al., 1998).
Men and women may also exhibit general preferences for different coping strategies
across life stressors; for example, several studies suggest that women are more likely than men to
cope by seeking social support (Folkman & Lazarus, 1980; Hamilton & Fagot, 1988). This
pattern would be consistent with the socialization hypothesis of coping, which asserts that these
broader coping patterns are informed by conditioning of traditional gender roles (Ptacek et al.,
1992). If women are socialized to be communal and emotionally expressive, it would be no
surprise for women to more frequently cope by utilizing social ties and eliciting network support
(Helgeson, 1994; Almeida & Kessler, 1998). If men are socialized to be independent, confident,
and control- and goal-oriented, they would be expected to more frequently take active and
problem-solving approaches to coping (Stone & Neale, 1984). In situations where there are no
clear or possible solutions, men may then favor avoidance or denial strategies to save face in line
with traditional gender roles favoring control (Felsten, 1998).

10

In a review of the literature on coping strategies of men and women, Tamres et al. (2002)
found that women are likelier than men to employ more types of coping strategies; women also
engage in more overall coping compared to men. This may be suggestive of the broader
structural disadvantages and increased pressures faced by women. The most robust gender
difference across studies was the finding that across different stressors, women are more likely to
seek emotional support compared to men. Women were overall more likely to use coping
involving verbal expression to self (e.g., rumination, positive self-talk) and others (e.g.,
emotional support). This result partially supports the socialization hypothesis of coping.
Women are more likely to use emotional support, which is in line with traditional gender roles;
however, men are not likelier to use any coping strategy more often than women, including
active and problem-solving strategies.
In contrast to the above results, Tamres et al. (2002) noted some findings in which the
nature of the stressor was associated with differential coping strategies by gender. For stressors
related to personal and others’ health, women were likelier than men to use more coping
strategies. This result might be related to the higher likelihood of women to serve as caretakers
and experience care-related distress. Men were more likely to respond to others’ health stressors
by using avoidance and withdrawal strategies; this is possibly consistent with Kressler &
MeLeod’s (1984) finding of heightened stress in response to network events in women but not
men. For relationship stressors, men were also more likely to use avoidance and venting than
women; women were more likely to use isolation, rumination, active coping, and problem-
focused coping. This is consistent with the gender differences found in lab-based studies of
older married couples, such that women are more likely to acknowledge and respond to negative
interactions while men avoid or ignore conflict (Carstensen et al., 1995). These results also

11

provide partial support for the socialization hypothesis of coping, as the differences between men
and women by stressor are generally consistent with traditional gender roles (except for women
utilizing active and problem-focused coping). The results also support the notion that the same
life events can hold different meanings and responsibilities for men and women, such that men
and women experience different associations with dementia risk.
Regarding biological influences, it is possible that the strength of the association between
life event- and dementia-related alleles differ for men and women. For example, the E4 allele of
the APOE gene is the strongest known genetic risk factor for late-life Alzheimer’s disease at
present (Mielke et al., 2014). There is some evidence for the finding that APOE4 confers greater
risk for Alzheimer’s disease for women than men. While several studies found such results
(Altmann et al., 2014; Farrer et al., 1997), another study reported that APOE4 may confer greater
risk for verbal memory and learning decline but not dementia for women than men (Beydoun et
al., 2012). APOE4 may also confer a unique risk for coronary events in men but not in women
(Scuteri et al., 2001). At a more micro-level of analysis, SNPs (single nucleotide
polymorphisms) also may pose differential risk for Alzheimer’s disease for men and women
(Mielke et al., 2014). Thus, differences in biological risk underlying life events and dementia for
men and women may drive different associations between life events and dementia. It is
important to note that while sociocultural and biological influences for men and women were
discussed distinctly in this section, the presence of one does not preclude the other. Gene-
environment correlation and interaction are possible and likely, but this is beyond the scope of
the present study.

12

Twin Studies: Social Selection vs. Social Causation

Given random assignment to stressful life events and dementia is unethical and
impractical, one of the primary difficulties in the life events and dementia literature has been
establishing their causal association. As with other health outcomes, most researchers argue a
social causation perspective: stressful life events themselves increase risk of dementia. Stressful
life events, for example, may increase dementia risk through increased glucocorticoid levels,
which in turn may lead to dysregulation of physiological functioning necessary for adequate
brain health, subsequently resulting in increased risk of dementia.
However, social selection factors, like nonrandom exposure to negative life events due to
genetic predisposition (e.g., impulsivity) and environmental exposures (e.g., poverty) may also
underlie correlation between life events and dementia, possibly explaining all or part of their
associations. In social selection, the inception of life event risk accounts for the association
between life events and dementia rather than the effects of life event risk as in social causation
(Johnson et al., 2009). An individual may have a genotype that predisposes her to
disagreeableness, which may increase both the risk of experiencing stressful life events and
dementia. Though social selection and social causation are not mutually exclusive, without
accounting for social selection confounds, studies that attempt to draw conclusions about the
causal relation between life events and dementia are at a disadvantage.
Genetically informed research designs, such as twin studies, are one solution as they
adjust for unmeasured familial confounds that may obscure observed associations between life
events and dementia risk (Beam et al., 2017; Turkheimer & Harden, 2014). Familial confounds
can be genetic and environmental in origin and are shared entirely by monozygotic (MZ) twins.
Commonly referred to as gene-environment correlation (i.e., nonrandom exposure to

13

environments based on genotype), people with the greatest genetic risk of dementia might be
exposed to more stressful life events. Stressful life events are heritable (Plomin et al., 1990;
Foley, Neale, & Kendler, 1996; Kendler, Karkowski, & Prescott, 1999) as is dementia (Gatz et
al., 1997; Gatz et al., 2006). Yet, shared environmental components might also predispose
people to both stressful life events and dementia. Shared environments refer to any
environmental factor that affects family members in the same way, like socioeconomic status,
parental education, or even neighborhood quality. Each of these has been implicated in the risk
of experiencing stressful life events and dementia (McLeod & Kessler, 1990; Hatch &
Dohrenwend, 2007; Mortimer & Graves, 1993; Fischer et al., 2008; King & Ogle, 2014; Kind et
al., 2017). For MZ twins, once familial confounds are accounted for, differences in effects of
life events on dementia risk must be attributed to differences in their unique environmental
experiences, as both unmeasured genetic and environmental confounds shared by twins are fully
adjusted.
Twin studies use pairs of twins of varying degrees of genetic relatedness to decompose
observed variance into additive genetic (A), shared environmental (C), and unique (nonshared)
environmental components (E). Additive genetic effects refer to the cumulative effect of
genotype that makes two twins similar to one another. In MZ twins, genotype is perfectly
correlated while for fraternal or dizygotic (DZ) twins genotype is correlated 0.5 as they share
half their genes, on average. Shared environments are perfectly correlated for identical and
fraternal twins, as zygosity status has no bearing on twins’ exposure to these environments.
Nonshared environmental factors refer to any factor that makes twins phenotypically differ from
one another, including measurement error. Thus, any phenotypic difference in dementia within
identical twin pairs could then be attributable to non-shared environmental effects i.e. divergent

14

social experiences such as adulthood life events as well as any measurement error; E effectively
denotes the causal effect of life events on dementia (Johnson et al., 2009; Beam et al., 2017;
Turkheimer & Harden, 2014).
Twin studies are quasi-causal in the sense that they are natural experiments yet lack
random assignment. Thus, they are still liable to third variable confounds that are not captured
by A and C. However, by controlling for genetic and shared environmental selection, twin
studies account for two significant sources of social selection confounds, allowing for stronger
inferences of causality stemming from unique environmental sources. To my knowledge, there
are no published studies examining the causal relationship between stressful life events and
dementia using genetically informed designs.
Biological Mechanisms for Social Causation

If this project produces evidence pointing towards causal relations between life event
domains and dementia using a twin design, several biological mechanisms might explain how
life events eventually cause dementia. The main systems involved are the endocrine, central
nervous, immune, and cardiovascular systems. Under psychological stress, the hypothalamic-
pituitary-adrenal (HPA) axis is activated and produces an increase in glucocorticoid hormone
levels (Lupien et al., 1999; Peavy et al., 2007; Ouanes & Popp, 2019). Glucocorticoids have
been found to promote oxidative stress and have been implicated in increases in amyloid beta
and tau pathology which are hallmark biomarkers of dementia (Toledo et al., 2012; Goodman et
al., 1996; Green et al., 2006). Increases in glucocorticoids may further decrease clearance of
amyloid beta from the brain (Harris-White et al., 2001). Cortisol, the main glucocorticoid in
humans, has been found in higher levels in patients with Alzheimer’s dementia; it has also been

15

shown to be positively correlated with severity of cognitive impairment (Dong and Csernansky,
2009; Pedersen et al., 2001; Zverova et al., 2013).
Immune system dysregulation may also be a mechanism through which stressful life
events and dementia are associated (Graham, Christian, & Kiecolt-Glaser, 2006; McEwen &
Stellar, 1993). Brief psychological stress such as stress stemming from periods of academic
examination can slow wound healing and dysregulate production of inflammatory cytokines,
suggesting common stressors can impact the immune system (Kiecolt-Glaser et al., 1998). More
chronic types of stress, such as stress stemming from conflict in close relationships, the kind
implicated in the types of stressful life events under examination, are associated with heightened
and lasting immune dysregulation. Couples who exhibit chronically unsupportive and negative
interaction patterns show more susceptibility to illness and slower wound healing (Cohen, 2005;
Kiecolt-Glaser et al., 2005). Family caregivers and bereaved individuals have also shown poorer
immune function and greater dysregulation (Gerra et al., 2003; Glaser et al., 2000; Esterling et
al., 1994). Even if chronic stress does not trigger an inflammatory response, it can lead to
chronic low-grade inflammation that may contribute over time to disease risk (Leonard, 2007).
Research implicates cytokine-mediated neuroinflammation in the development of dementia
(Takeda, Sato, & Morishita, 2014; Chang, Yee, & Sumbria, 2017).
Cardiovascular systems are also implicated in the development of dementia and can be
linked to stress. Studies have found that intense emotional stress is associated with heart disease
(Sharkey et al., 2005; Wittstein et al., 2005; Dimsdale, 2008). Chronic work stress and marital
stress are associated with increased risk of heart disease (Orth-Gomer, 2000; Matthews & Gump,
2002). Heart disease is a known risk factor for dementia due to reduced cerebral blood flow,
which causes a buildup of amyloid beta and tau proteins (Justin, Turek, & Hakim, 2013). Given

16

evidence for a causal association between life event domains and dementia, there are thus several
convincing mechanisms through which life events may cause dementia risk.
Proposed Study

Because there are no twin studies that have investigated the genetic and environmental
effects underlying the observed association between life events and dementia by gender, this
project has three specific aims.
Aim 1: Examine whether genetic and shared environmental influences involved in
experiencing stressful life events also account for variability in experiencing dementia for men
and women.
Hypothesis 1: I expect that there will be common genetic and shared
environmental influences involved in the experience of six domains of stressful
life events and dementia for men and women.
Aim 2: Adjusting for any common genetic and shared environmental influences between
dementia risk and life events, examine whether within-family differences (i.e., nonshared
environmental) in life events increase dementia risk for men and women, which would be
consistent with causal arguments.
Hypothesis 2: I expect that there will be evidence of within-family differences in
the associations between the six life event domains and dementia risk for men and
women. I anticipate that within-family differences in only the domain of positive
life events will decrease dementia risk; within-family differences in the other five
domains will increase dementia risk.
Aim 3: Examine if and how gender affects the types of between-family factors and quasi-
causal effects implicated in the association between life event domains and dementia risk.

17

Hypothesis 3: I expect that the strength of any quasi-causal effects will differ by
gender. Women may show an overall pattern of stronger quasi-causal effects.
To sum, this study contributes new perspective with which to examine the relationship
between life events and dementia prevalence for men and women. By using a genetically
informed design and latent variable modeling, stronger inferences about the presence of a causal
association between life event domains and risk of dementia can be made.

Methods

Sample

This project uses data from the first three waves of the Swedish Adoption/Twin Study of
Aging (SATSA; MZM = 199, MZF = 269, DZM = 345, DZF = 556) to examine whether
different life events are causally associated with dementia risk (Finkel & Pedersen, 2004).
SATSA is an ongoing project that began in 1984 after mailing questionnaires to identical and
same-sex fraternal twins identified as having been reared apart starting from before 10 years of
age in the population-based Swedish Twin Registry. A matched sample of twins by sex, country
of birth, and birth year who were reared together was also contacted. After the initial contact in
1984 (Q1), twins were again sent questionnaires in 1987 (Q2) and 1990 (Q3). All questionnaires
included the same measures of life events. In-person testing was also performed on a subset of
twins starting in 1986 (IPT 1). One follow-up was conducted in 1989. A health examination,
structured interviews, and tests on cognition, memory, and functional capacity were performed
(Pedersen, 2015). For the purposes of the study, twins under the age of 50 years at baseline were
excluded (Figure 1).

18

Figure 1
Flowchart of Participants Included in Present Study

Measures
Life events. A modified version of the Social Readjustment Rating Scale, a multiple life
events measure, was administered to all twins (Holmes & Rahe, 1967; Persson, 1980). This
measure tailored to older adults contains 25 life event items spanning domains of illnesses,
finances, deaths, relationship changes for self, family, and friends (see Appendix A). Both
negative and positive events are included in the measure. Examples include divorce, nursing
home care of spouse, home care, of self, major improvement in financial status, retirement after
employment, and death of siblings or friends. Occurrence of life events were recorded across
waves as occurring or not (1 = not occurred, 2 = occurred).
Gender. Participant sex was collected (1 = male, 2 = female) from the Swedish Twin
Registry. In this study, this variable was conceptualized as measuring gender considering the
cohort and sociohistorical context; several studies examining differences between men and
women using SATSA have done the same (Read et al., 2006, Kato & Pedersen, 2005; Finkel et
al., 2015). As the scope and proposed explanations of the study are rooted in structural

19

inequities and differential sociocultural norms and roles, gender was chosen for this study.
However, the risk of conflating gender and sex remains (Johnson et al., 2009; Rich-Edwards et
al., 2018).
Dementia diagnosis. Dementia diagnoses were determined using multiple approaches
for participants who had completed both a SATSA in-person testing visit and questionnaire, just
the questionnaire, or had not responded to the questionnaire. Screening began during the second
wave of SATSA. Those with in-person cognitive testing results were screened according to a
cut-off of 24.5 on the Mini Mental State Exam and overall clinical impressions from nurses.
Those with only questionnaire data and non-respondents were screened through use of a
telephone interview called TELE that included the Mental Status Questionnaire and items
designed to tap health and functional status. Relatives were asked questions about cognitive
function and functional ability if respondents were unable to answer. Anyone left in SATSA was
followed through registry linkage with the Swedish National Patient Register (NPR) and Cause
of Death Register (CDR). These registries contained dementia diagnoses from International
Classification of Disease (ICD) codes. Those who did not screen positive through any of these
means were due for follow-up in another three years (Gatz et al., 1997; Beam et al., 2018).
Participants who screened positive were then referred for in-person clinical work-ups
which included physical and neuropsychological assessment, lab tests, and neuroimaging. A
preliminary assessment of dementia for one twin meant the co-twin participated in the clinical
work-up, too. Final clinical diagnoses were determined by a diagnostic consensus board using
collected data from the in-person clinical work-up and participant medical records. Dementia
diagnoses were determined using DSM-III-R and DSM-IV criteria, and NINCDS-ADRDA
criteria for Alzheimer’s dementia (Gatz et al., 1997; Beam et al., 2018). Dementia status is

20

dichotomously coded (0 = no, 1 = yes). Only dementia cases occurring after Q3 (1990) were
included.
Data Analysis

The 25 life event variables were recoded so that 1 = occurred and 0 = did not occur in
each wave using R (R Core Team, 2013). New scores were then created for endorsement of each
item over the three waves Q1 - Q3 to measure if any item was ever endorsed across waves. For
example, a resulting score of 0 for the life event “divorce” means no endorsement of this event
across the lifespan. A score of 1 would mean endorsing divorced status during at least one wave.
Exploratory factor analysis (EFA) was then conducted on these scores to identify a
possible underlying factor structure of the 25 life event items. We examined up to 8 factors.
Previous research on life events tend to operationalize life event domains based on theory or
common sense (Perris, 1984; Myers, Lindenthal, Pepper, & Ostrander, 1972; Dohrenwend, 1973;
Kendler, Karkowski, & Prescott, 1999). Such an approach has led to the use of sum scores,
which is problematic because sum scores introduce measurement error into the life event
outcomes. EFA extracts the maximum amount of shared variance in a set of variables and
identifies underlying variables called factors using a pre-specified rotation; for current purposes,
we used an oblique rotation, as it was sensible that life event factors would correlate. The
chosen factor solution is one that best fits theoretically and structurally (Kline, 2015). Latent
variable modeling facilitates an empirical test of theoretical life event domains. These resulting
participants’ life event domain factor scores will be unbiased by measurement error.
The resulting structure was validated using confirmatory factor analysis (CFA);
diagonally weighted least squares (WLSMV) estimation was used due to binary observed data.
WLSMV does not assume normality for observed variables but does for the latent distribution

21

underlying the observed variables (Li, 2015; Muthén & Muthén, 1998-2012). The within-family
factor correlations from the CFA were checked for the assumption that twins are
indistinguishable. Measurement invariance across gender was tested; configural invariance best
fit the observed data, meaning that factor loadings and thresholds could not be set equal.
Following creation of life event factors, within-family correlations by zygosity were
examined for each of the life event factors to determine the unique estimates of genetic (A),
shared environment (C), and non-shared environment influences (E) by gender. For the
dementia variable, within-family correlations by zygosity were examined to determine estimates
of A, C, and E by gender. Genetic confounds are implicated if MZ twin correlations are greater
than DZ twin correlations. Shared environmental effects are implicated if DZ twin correlations
are greater than half the MZ correlations (Rijsdijk & Sham, 2002). Nonshared environmental
effects are implicated if the MZ twin correlation is less than 1.0.
To determine if there were any significant phenotypic associations between life event
domains and dementia by gender, logistic regressions were conducted. These regressions did not
account for the possibility of A and/or C confounds and are most directly comparable to
estimates of the observed association between life events and dementia from other studies that do
not use twin methods.
A cross-twin, cross-trait logistic regression analysis was then run for each life event
factor and dementia by gender using Mplus 8.2 (Muthén & Muthén, 2018). These analyses
allowed for inferences regarding familial confounding in the observed associations between life
event domains and dementia. They are interpreted in the same manner as twin correlations of
single variables. This logistic regression examined the association between twin 1’s life event
experience and twin 2’s binary dementia status. To the extent that twin 1’s life event is equally

22

or more predictive of twin 2’s dementia status in DZ twins than MZ twins, a CE model is
implied. If twin 1’s life event is more predictive of twin 2’s dementia status in MZ twins than
DZ twins, there is evidence of A and C. Odds ratios were calculated for each life event factor
and dementia by zygosity and gender.
Bivariate twin models were then fit to the data. These models account for any genetic
and environmental confounding, such that any remaining associations could be inferred as
support for a quasi-causal effect of life events on dementia status. The following model-fitting
sequence was followed. Model 1 was the baseline biometric ACE model (Figure 2). Model 2 fit
an A=C model, in which the common factors generated from the observed data are
indistinguishable; a good fit for this model indicates a general between-family effect. Model 3
fit an AE model in which there is no C effect. Model 4 fit a CE model in which there is no A
effect. The purpose of Models 3 and 4 was to explore which selection process – genetic or
environmental – was the most plausible confounding mechanism if Model 2 was the best-fitting
model.
Aim 1 is addressed by examining whether A and/or C factors account for variability in
life event domains and dementia risk. Aims 2 and 3 are addressed by examining the effect of life
event domains on dementia risk while simultaneously adjusting for any genetic and shared
environmental correlation between life events and dementia risk. All analyses were conducted
using Mplus 8.2 (Muthén & Muthén, 2018).

23

Figure 2
Biometric Regression Model for Life Event Factor and Dementia

Note. A significant bE estimate indicates evidence for a quasi-causal relationship between the
life event domain and dementia. Significant σ2A and σ2C estimates indicate evidence for the
presence of A and C effects, respectively.
Results

Data was analyzed from 885 twin families (MZM = 199, MZF = 269, DZM = 345, DZF
= 556). Twins were over the age of 50 years at baseline and 60.26% women. 15.05% of the
sample had a diagnosis of dementia occurring after Q3 (1990). The results of the confirmatory
factor analysis, phenotypic analyses, cross-twin cross-trait logistic regressions, and bivariate twin
models are reported below.
Confirmatory Factor Analyses
In the full sample, the confirmatory factor analysis confirmed the 6-factor solution from
the exploratory factor analysis (see Appendix B). In support of this conclusion, the RMSEA was
.03, below the .05 threshold of “good” fit, suggesting good fit between the model and the
observed data. The 6-factor solution was examined for measurement invariance by gender. The
configural model provided a good fit to the data while the scalar model provided a worse fit than
the configural model (Table 1). This means that while the factor structure is equal across men
and women, the loadings and thresholds are not, indicating that each life event item may not

24

measure the same thing in men and women. The items most representative of factors are
different across gender for five out of six life event domains. As such, we proceeded with
separate life event factors for men and women (Table 2; Table 3).
Table 1

Goodness-of-Fit Indicators of Models for Measurement Invariance by Gender

Model
Description
Model
Comparison
χ2 df Δ χ2 Δ df p RMSEA TLI SRMR

Model 1 -
Configural -
959.65 388 - - - 0.045 0.899 0.099

Model 2 -
Scalar 2 vs. 1
1013.41 398 62.52 10 0 0.046 0.894 0.1

Note. Italicized text indicates the model that provided the best fit to the data. The configural
model sets the factor structures across gender to be equal. The scalar model additionally sets the
factor loadings and thresholds as equal across gender.

25

Table 2

Confirmatory Factor Analysis for Men – Factors and Loadings

Life Events Multidomain
Loss*
NLE
Children
Self-Illness* Family
Strife*
NLE
Spouse*
Positive
LE*
Retirement after
employment
0.84
Major deterioration in
financial status
0.75
Death of siblings or
friends
0.74
Loss of sexual ability
of interest
0.59
Serious illness in child 0.87
Death of child 0.73
Forced change in
residence b/c one can’t
manage to look after
oneself
0.88
Home care, self 0.81
Mental illness, self 0.78
Forced change in
residence with
reduced contact
0.65
Serious conflicts with
child
0.81
Deterioration in
married life
0.81
Divorce 0.62
Nursing home care,
spouse
0.91
Death of spouse 0.79
Home care of spouse
by proband
0.78
Somatic illness,
spouse
0.76
Mental illness, spouse 0.62
Making an
acquaintance
0.72
Getting married 0.65
Major improvement in
financial status
0.65
Improvement in
married life
0.64

26

Note. An asterisk (*) in the life event factor column indicates that the most representative item
differs by gender.
Table 3

Confirmatory Factor Analysis for Women – Factors and Loadings

Life Events Multidomain
Loss*
NLE
Children
Self-Illness* Family
Strife*
NLE
Spouse*
Positive
LE*
Loss of sexual ability
or interest
0.77
Major deterioration in
financial status
0.74
Retirement after
employment
0.71
Death of siblings or
friends
0.70
Serious illness in child 0.99
Death of child 0.58
Home care, self 0.78
Forced change in
residence b/c one can’t
manage to look after
oneself
0.69
Mental illness, self 0.66
Forced change in
residence with reduced
contact
0.65
Deterioration in
married life
0.83
Serious conflicts with
child
0.78
Divorce 0.50
Somatic illness,
spouse
0.92
Home care of spouse
by proband
0.82
Nursing home care,
spouse
0.75
Death of spouse 0.70
Mental illness, spouse 0.59
Getting married 0.76
Making an
acquaintance
0.67
Improvement in
married life
0.65
Major improvement in
financial status
0.60

27

Note. An asterisk (*) in the life event factor column indicates that the most representative item
differs by gender.
Phenotypic Analyses

Only one life event domain each for men and women showed a significant phenotypic
association with dementia; these domains differed by gender (Table 4). For women, the total
effect of life events on dementia was detected for the domain of multidomain loss; the odds of
dementia indicate that women had 21% higher risk of dementia for every unit increase in
multidomain loss (OR = 1.21, 95% CI [1.08, 1.36], p = .001). For men, family strife was
associated with dementia; the men had a 24% higher risk of dementia for every unit increase in
family strife (OR = 1.24, 95% CI [1.03, 1.51], p = .03). For women, the domain of negative life
events for spouse trended toward statistical significance, as risk of dementia was 9% (with an
upper estimate of 23% increased risk and a lower estimate of 1% decreased risk) for every unit
increase in negative life events for spouse. OR = 1.09, 95% CI [0.99, 1.23], p = .09). For men,
the effect of negative life events for children, a second index of negative life events surrounding
family, trended toward statistical significance; men had 26% increased dementia risk (with an
upper estimate of 65% increased risk and a lower estimate of 3% decreased risk) for every unit
increase in negative life events for children(OR = 1.26, 95% CI [0.97, 1.65], p = .09).
Table 4
Phenotypic Effect of Life Events on Dementia by Gender
Men Women
Phenotype
OR 95% CI p OR 95% CI p
Multidomain Loss
1.12 [0.96, 1.29] 0.16 1.21 [1.08, 1.36] 0.001
NLE Children
1.26 [0.97, 1.65] 0.09 0.99 [0.75, 1.31] 0.95
Self-Illness
1.06 [0.85, 1.34] 0.60 0.97 [0.83, 1.12] 0.66
Family Strife
1.24 [1.03, 1.51] 0.03 0.92 [0.79, 1.07] 0.26
NLE Spouse
1.12 [0.94, 1.32] 0.21 1.09 [0.99, 1.23] 0.09
Positive LE
1.03 [0.77, 1.38] 0.86 1.19 [0.96, 1.46] 0.12

Note. Bolded text indicates p < .05. Italicized text indicates p < .10.

28

Cross-Twin Cross-Trait Logistic Regressions
Cross-twin cross-trait logistic regressions suggested the presence of familial confounding.
Of note, no significant differences were found by zygosity across any of the life event domains
for men and women given the reduced sample size per zygosity group. Point estimates of the
odds ratios are thus, interpreted only. For men (Figure 3), multidomain loss, self-illness, and
family strife suggest confounding of A and C, as effects in MZ twins were greater than in DZ
twins. Negative life events of children, negative life events for spouse, and positive life events
suggest confounding of C, as effects in DZ twins were greater than in DZ twins.
Figure 3
Forest Plot of Cross-Twin Cross-Trait Odds Ratios by Zygosity and Life Event Factor for Men

29

For women (Figure 4), multidomain loss, negative life events of children, self-illness,
negative life events of spouse, and positive life events suggest the presence of C confounds. The
domain of family strife suggests confounding of A and C.
Figure 4
Forest Plot of Cross-Twin Cross-Trait Odds Ratios by Zygosity and Life Event Factor for
Women

30

Bivariate Twin Models
Model-fitting results for men (Table 5) reveal best-fitting models by factor. For men, a
model with environmental effects only (CE model) best fit the life event domain of multidomain
loss. Models with indistinguishable family level confounds (A=C model) best fit the life event
domains of negative life events of children and self-illness. Shared environmental factors likely
could be dropped (AE model) from domains of family strife, negative life events of spouse, and
positive life events. Parameter estimates from the best-fitting models for men (Table 6) reveal
significant variances underlying each life event in line with the hypothesized paths (e.g., AE
model shows significant A variance component) except in the case of family strife. No
regression coefficients were significant (p > .05), suggesting that observed associations between
life event domains and dementia risk are entirely mediated by genetic confounds, shared
environmental confounds, or both.

31

Table 5
Model-Fitting Results for Bivariate Twin Models – Men
Model Description χ2 df Δ χ2 Δ df p TLI SRMR RMSEA
LE 1 - Multidomain Loss
Baseline (ACE)
106.733 94 - - - 0.977 0.129 0.027
A=C
105.116 96 0.773 2 .68 0.984 0.131 0.023
CE
103.902 96 0.239 2 .89 0.986 0.13 0.021
LE 2 - NLE Children
Baseline (ACE)
37.817 31 - - - 0.925 0.146 0.035
A=C
37.642 33 0.147 2 .93 0.952 0.147 0.028
AE
37.616 32 0.182 1 .67 0.941 0.147 0.031
CE
37.74 32 0.235 1 .63 0.939 0.148 0.032
LE 3 - Self-Illness
Baseline (ACE)
90.852 94 - - - 1.01 0.17 0
A=C
91.189 96 0.526 2 .77 1.015 0.173 0
AE
90.926 95 0.194 1 .66 1.013 0.171 0
CE
91.326 95 0.481 1 .49 1.012 0.171 0
LE 4 - Family Strife
Baseline (ACE)
73.044 58 - - - 0.923 0.204 0.038
A=C
72.954 60 0.579 2 .75 0.936 0.204 0.035
AE
72.77 59 0.003 1 .95 0.931 0.204 0.036
CE
72.57 59 0.024 1 .88 0.932 0.204 0.036
LE 5 - NLE Spouse
Baseline (ACE)
139.757 138 - - - 0.996 0.174 0.008
AE
138.607 140 0.143 2 .93 1.003 0.174 0
LE 6 - Positive LE
Baseline (ACE)
118.811 94 - - - 0.921 0.147 0.038
A=C
118.349 96 1.432 2 .49 0.93 0.148 0.036
AE
117.552 95 0 1 .99 0.929 0.147 0.036

Note. Italicized text indicates the model that provided the best fit to the data for each life event
domain. All models compared to Baseline (ACE) model. MZM = 199; DZM = 345.

32

Table 6
Parameter Estimates from Best-Fitting Models – Men

Estimates
Model Description σ2A σ2C σ2E A REG C REG E REG
LE 1 - Multidomain
Loss

CE - .54 (.08) .78 (.07) - 0.23 (0.45) 0.06 (0.21)
LE 2 - NLE Children*

A=C .24 (.12) .54 (.13) 1.90 (1.01) -0.73 (0.53)
LE 3 - Self-Illness*

A=C .42 (.04) .46 (.08) 0.29 (0.80) -0.27 (1.10)
LE 4 - Family Strife*

AE .50 (.51) - .78 (.18) 0.86 (1.95) - -0.07 (0.38)
LE 5 - NLE Spouse

AE .47 (.10) - .67 (.08) 1.02 (0.78) - -0.41 (0.45)
LE 6 - Positive LE*

AE .40 (.18) - .43 (.07) 0.82 (1.11) - -1.01 (1.15)
Note. An asterisk (*) in the life event row indicates that the best-fitting model differs by gender.
Italics indicate an unstandardized regression coefficient that is significant at p < .10.

Model-fitting results for women (Table 7) reveal best-fitting models by factor. For
women, a model with environmental effects only (CE model) best fit the life event domain of
multidomain loss, negative life events of children, and positive life events. Shared
environmental factors could be eliminated (AE model) for the domains of negative life events of
spouse, and also self-illness. A model that could not distinguish between familial confounds
(A=C model) best fit the life event domain of family strife. Women and men shared the same
best-fitting models for two out of six life event domains. Parameter estimates from the best-
fitting models for women (Table 8) reveal significant variances in line with the hypothesized
paths (e.g., AE model shows significant A variance component) except in the case of positive life

33

events. No regression coefficients were significant (p > .05), indicating that observed
associations between life event domains and dementia are entirely mediated by genetic
confounds, shared environmental confounds, or both.
Table 7
Model-Fitting Results for Bivariate Twin Models – Women
Model Description χ2 df Δ χ2 Δ df p TLI SRMR RMSEA
LE 1 - Multidomain Loss
Baseline (ACE) 111.473 94 - - - 0.979 0.097 0.027
A=C 108.455 96 0.56 2 .76 0.985 0.099 0.022
AE 113.186 95 1.599 1 .21 0.985 0.1 0.027
CE 107.624 96 0.293 2 .86 0.985 0.097 0.021
LE 2 - NLE Children
Baseline (ACE) 23.564 31 - - - 1.107 0.106 0
A=C 24.503 33 0.788 2 .67 1.115 0.109 0
AE 23.753 32 0.076 1 .78 1.115 0.107 0
CE 24.085 33 0.267 2 .88 1.12 0.106 0
LE 3 - Self-Illness
Baseline (ACE) 91.357 94 - - - 1.013 0.145 0
AE 91.169 96 0.126 2 .94 1.023 0.145 0
LE 4 - Family Strife
Baseline (ACE) 48.791 58 - - - 1.055 0.127 0
A=C 50.249 60 1.333 2 .51 1.056 0.127 0
AE 49.625 59 0.763 1 .38 1.055 0.127 0
CE 50.741 59 1.653 1 .20 1.048 0.127 0
LE 5 - NLE Spouse
Baseline (ACE) 137.223 138 - - - 1.001 0.127 0
AE 134.241 140 0.151 2 .93 1.006 0.127 0
LE 6 - Positive LE
Baseline (ACE) 101.63 94 - - - 0.982 0.096 0.018
A=C 100.108 96 0.487 2 .78 0.99 0.097 0.013
AE 100.976 95 0.016 1 .90 0.986 0.096 0.015
CE 101.052 95 0.006 1 .94 0.986 0.096 0.016

Note. Italicized text indicates the model that provided the best fit to the data for each life event
domain. All models compared to Baseline (ACE) model. MZF = 269; DZF = 556.

34

Table 8
Parameter Estimates from Best-Fitting Models – Women

Estimates
Model Description σ2A σ2C σ2E A REG C REG E REG
LE 1 - Multidomain
Loss

CE - .57 (.06) .68 (.06) - 0.51 (0.31) -0.03 (0.21)
LE 2 - NLE Children*

CE - .13 (.24) .37 (.28) - -1.04 (2.59) 0.39 (0.71)
LE 3 - Self-Illness*

AE .72 (.08) - .44 (.13) -0.33 (0.71) - 0.49 (1.24)
LE 4 - Family Strife*

A=C .23 (.21) .95 (.14) 0.79 (0.84) -0.27 (0.16)
LE 5 - NLE Spouse

AE .50 (.10) - .78 (.07) 0.27 (0.56) - 0.03 (0.22)
LE 6 - Positive LE*

CE - .15(.37) .47 (.08) - -1.20 (6.46) 0.57 (0.46)
Note. An asterisk (*) in the life event row indicates that the best-fitting model differs by gender.
Italics indicate an unstandardized regression coefficient that is significant at p < .10.

Discussion
The present study found that life event domains mean different things for men and
women. The most representative items in five out of six life event domains were different by
gender. In the example domain of negative spousal events, a spouse’s somatic illness best
represented this domain of experiences for women, while for men, it was a spouse’s nursing
home care. Even for negative life events of children, the factor for which the most representative
item was the same for men and women, the pattern of loadings suggested that women may
experience this factor as more driven by a child’s serious illness while men may experience it as
a more even split between a child’s serious illness and death.

35

Two significant phenotypic associations were detected between life events and dementia,
one each for men and women. The significant associations were different by gender. For
women, the odds of dementia were 21% higher for every unit increase in multidomain loss. For
men, the odds of a dementia diagnosis were 24% higher for every unit increase in family strife.
Examination of the cross-twin cross-trait odds ratios showed that overall, between-family
effects mattered for men and women; there was possible confounding of A and C (e.g., between-
family effects). A and C confounds were implicated in the observed phenotypic associations,
given the significant odds ratios found for multidomain loss and dementia in women as well as
family strife and dementia in men were no longer significant when accounting for zygosity.
Further, these cross-twin cross-trait analyses indicated a possible suppression effect of A and C;
although there were very few significant phenotypic associations (within-twin, no accounting for
zygosity), MZ and DZ differences emerged, albeit nonsignificant ones. In other words,
unaccounted A and C confounds in the phenotypic model may suppress otherwise observable
population differences between life events and dementia.
The bivariate twin models allowed for testing of common genetic and shared
environmental influences as well as inference of quasi-causal associations between life event
domains and dementia. In best-fitting models, significant common genetic and shared
environmental influences for life events and dementia were detected for nearly all best-fitting
models (Aim 1). However, after adjustment for familial confounding, there was no evidence for
any associations between the six life event domains and dementia for men and women; there was
no evidence for quasi-causal effects of life events on dementia (Aims 2 and 3). Taking all the
analyses together, the results suggest that any observed relationships between life events and

36

dementia are explained by family-level processes (between-family effect) rather than individual
differences (within-family effect).
The idea that life events are not individual but familial experiences in the context of
dementia risk is an intriguing one. But perhaps it is not inexplicable. In the example of family
strife, the kinds of shared influences and experiences that make twins similar for deterioration in
married life, serious conflicts with child, and divorce may also heighten selection into situations
that raise dementia risk. The notion that one twin’s family strife may be related to the other
twin’s dementia risk is plausible because marital discord, parent-child conflict, and divorce can
have family-wide implications. Similarly, for the domain of negative spousal events, one twin’s
experience with spousal stress may affect the other twin’s dementia risk by way of caregiving
effects and strain through the family network.
Of note, the best-fitting models by men and women for life event domains were not
consistent; only 2 out of 6 domains shared the same best-fitting model (multidomain loss and
negative spousal events). To the extent this pattern holds, an accompanying question is what
might explain the differences in fit by gender for the remaining life event domains. A likely
contributing factor is that many of the domains represented different things for men and women,
as evidenced by the most representative items. For example, the most representative item for
positive life events is getting married for women while it is making an acquaintance for men; the
CE model provides best fit for women while the AE model provides best fit for men. It is also
possible that the same life events can denote different things by way of meaning, related
responsibilities, and implications, such that even if the relative item loadings for a domain were
the same, the representative experience for men and women is distinct. For example, the domain
of negative child events is best represented by serious illness in child for both men and women;

37

however, experience of the domain for women may involve more caregiving. The best fitting
model for women is environmental effects only while for men the genetic and shared
environmental pathways are indistinguishable. Another consideration is that although the best-
fitting models differ for men and women in 4 out of 6 domains, the general conclusion is that
familial confounds could not be distinguished. Further study of life events and dementia in other
twin samples could provide clarity on the family processes that drive observed associations.
When genetic sources of variance account for the association between dementia and
negative life events, this opens a new avenue for investigating whether specific genetic effects
contribute to life event experiences and dementia. A next step would be to identify these genetic
influences using genome-wide association studies and polygenic scores. Additionally, it permits
further testing of environmental factors that may moderate genetic influences common to both
negative life events and dementia. When shared environmental sources of variance account for
some of the association, social mechanisms underlying the association should be examined
further. For example, social environments inducing hostility may engender risk of both
multidomain loss and dementia.
This study has several limitations. One regards the interpretation of the life events in
this study. The life events under examination were measured for occurring during the entire life
course, so comparisons about events occurring during childhood, adulthood, and late-life cannot
be made. The second regards the interpretation of the bivariate twin model, as it does not specify
which genetic or shared environmental influences common to both life events and dementia are
involved. In addition, it is possible that cognitive decline related to dementia diagnosis drives
selection into life events in a way that requires additional model specification (e.g., higher-order
common pathways model) Further research is needed to understand the specific genetic or

38

shared environmental confounds implicated in the association between life event domains and
dementia risk.
The results of this study suggest that instead of attempting to minimize dementia risk by
mitigating impacts of life events, future research should examine the kinds of genotypes and
environments that increase the risk of experiencing both negative life events and dementia by
gender. By preventing or intervening in the impacts of these mechanisms, it may be possible to
decrease the incidence of both negative life events and dementia.

39

References

Almeida, D. M., & Kessler, R. C. (1998). Everyday stressors and gender differences in daily
distress. Journal of Personality and Social Psychology, 73(3), 670-80.
Altmann, A., Tian, L., Henderson, V., & Grecius, M. (2014). Sex modifies the APOE-related risk
of developing Alzheimer disease. Annals of Neurology, 75(4), 563-73.
Alzheimer’s Association. (2018). Alzheimer’s Disease Facts and Figures. Alzheimer’s Dement,
14(3), 367-429.
Aneshensel, C. (1992). Social stress: theory and research. Annual Review of Sociology, 18, 15-
38.
Beam, C., Dinescu, D., Emery, R., & Turkheimer, E. (2017). A twin study on perceived stress,
depressive symptoms, and marriage. Journal of Health and Social Behavior, 58(1), 37-
53.
Beam, C., Kaneshiro, C., Jang, J. Y., Reynolds, C., Pedersen, N., & Gatz, M. (2018). Differences
between men and women in incidence rates of dementia and Alzheimer’s disease. J
Alzheimer’s Disease, 64(4), 1077-83.
Beards, S., Gayer-Anderson, C., Borges, S., Dewey, M. E., Fisher, H. L., & Morgan, C. (2013).
Life events and psychosis: a review and meta-analysis. Schizophrenia Bulletin, 39(4),
740-7.
Beydoun, M., Boueiz, A., Abougergi, M., Kitner-Tiolo, M., Beydoun, H., Resnick, S., O’Brien,
R., & Zonderman, A. (2012). Sex differences in the association of the apolipoprotein E
epsilon 4 allele with incidence of dementia, cognitive impairment and decline.
Neurobiology of Aging, 33(4), 720-31.

40

Brugha, T. S., & Cragg, D. (1990). The list of threatening experiences: the reliability and validity
of a brief life events questionnaire. Acta Psychiatrica Scandinavica, 82, 77–81.
http://dx.doi.org/10.1111/j.1600-0447.1990. tb01360.x
Carstensen, L., Gottman, J., & Levenson, R. (1995). Emotional behavior in long-term marriage.
Psychology and Aging, 10, 140-49.
Chang, R., Yee, K. L., & Sumbria, R. K. (2017). Tumor necrosis factor α Inhibition for
Alzheimer's Disease. Journal of Central Nervous System Disease, 9, 1179573517709278.
doi:10.1177/1179573517709278
Cohen S. (2005). The Pittsburgh common cold studies: psychosocial predictors of susceptibility
to respiratory infectious illness. Int. J. Behav. Med., 12, 123–131.
Cutolo, M., & Straub, R. H. (2006). Stress as a risk factor in the pathogenesis of rheumatoid
arthritis. Neuroimmunomodulation, 13(5-6), 277-82.
Dimsdale, J. E. (2008). Psychological stress and cardiovascular disease. Journal of the American
College of Cardiology, 51(13), 1237-46.
Dohrenwend, B. (1973). Social status and stressful life events. Journal of Personality and Social
Psychology, 28(2), 225-35.
Dohrenwend, B. S., & Dohrenwend, B. P. (Eds.). (1974). Stressful life events: Their nature and
effects. Oxford, England: John Wiley & Sons.
Dong, H., & Csernansky, J. G. (2009). Effects of stress and stress hormones on amyloid-beta
protein and plaque deposition. Journal of Alzheimer's Disease, 18(2), 459-69.
Dufouil, C., Pereira, E., Chêne, G. et al. (2014). Older age at retirement is associated with
decreased risk of dementia. Eur J Epidemiol, 29: 353-61.
https://doi.org/10.1007/s10654-014-9906-3

41

Duijts, S. F., Zeegers, M. P. & Borne, B. V. (2003). The association between stressful life events
and breast cancer risk: a meta‐analysis. Int. J. Cancer, 107, 1023-1029.
doi:10.1002/ijc.11504
Escott-Price, V., Shoai, M., Pither, R., Williams, J., & Hardy, J. (2017). Polygenic score
prediction captures nearly all common genetic risk for Alzheimer's disease. Neurobiology
of Aging, 49, 214-211. https://doi.org/10.1016/j.neurobiolaging.2016.07.018.
Esterling, B. A., Kiecolt-Glaser, J. K., Bodnar, J., & Glaser, R. (1994). Chronic stress, social
support, and persistent alterations in the natural killer cell response to cytokines in older
adults. Health Psychology, 13, 291-99.
Farrer, L., Cupples, L., Haines, J., Hyman, B., Kukull, W., Mayeux, R., et al. (1997). Effects of
age, sex, and ethnicity on the association between apolipoprotein E genotype and
Alzheimer disease. A meta analysis. APOE and Alzheimer disease meta analysis
consortium. JAMA, 278, 1349-56.
Fauth, E. B., Schwartz, S., Tschanz, J. T., Østbye, T., Corcoran, C., & Norton, M. C. (2012).
Baseline disability in activities of daily living predicts dementia risk even after
controlling for baseline global cognitive ability and depressive symptoms. International
Journal of Geriatric Psychiatry, 28(6), 597-606.
Felmlee, D., & Muraco, A. (2009). Gender and friendship norms among older adults. Research
on Aging, 31(3), 318-44.
Felsten, G. (1998). Gender and coping: use of distinct strategies and associations with stress and
depression. Anxiety, Stress, and Coping, 11(4), 289-309.

42

Finkel, D., & Pedersen, N. L. (2004). Processing speed and longitudinal trajectories of change
for cognitive abilities: The Swedish Adoption/Twin Study of Aging. Aging,
Neuropsychology, and Cognition, 11(2-3), 325-345.
http://dx.doi.org/10.1080/13825580490511152
Finkel, D., Ernsth-Bravell, M., & Pedersen, N. (2015). Sex differences in genetic and
environmental influences on longitudinal change in functional ability in late adulthood.
The Journals of Gerontology Series B: Psychology Sciences and Social Sciences, 70(5),
709-17.
Fischer, P., Zehetmayer, S., Jungwirth, S., et al. (2008). Risk factors for Alzheimer dementia in a
community-based birth cohort at the age of 75 years. Dement Geriatr Cogn Disorder, 25,
501-507. doi: 10.1159/000128577
Foley, D. L., Neale, M. C., & Kendler, K. S. (1996). A longitudinal study of stressful life events
assessed at interview with an epidemiological sample of adult twins: the basis of
individual variation in event exposure. Psychological Medicine, 26(6), 1239-1252.
http://dx.doi.org/10.1017/S0033291700035960
Folkman, S., & Lazarus, R. (1980). An analysis of coping in a middle-aged community sample.
Journal of Health and Social Behavior, 21(3), 219-39.
Fountoulakis, K., Pavlidis, I., & Tsolaki, M. (2011). Life events and dementia: what is the nature
of their relationship? Psychiatry Research, 190(1), 156-58.
Gatz, M., Pedersen, N., Berg, S., Johansson, B., Johansson, K., Mortimer, J., Posner, S.,
Viitanen, M., Winblad, B., & Ahlbom, A. (1997). Heritability for Alzheimer's Disease:
the study of dementia in Swedish twins. The Journals of Gerontology: Series A, 52(2),
117–125. https://doi.org/10.1093/gerona/52A.2.M117

43

Gatz, M., Reynolds, C. A., Fratiglioni L, et al. (2006). Role of genes and environments for
explaining Alzheimer disease. Arch Gen Psychiatry, 62(2), 168–174.
doi:10.1001/archpsyc.63.2.168
Gerra, G., Monti, D., Panerai, A. E., Sacerdote, P., Anderlini, R., Avanzini, P., Zaimovic, A.,
Brambilla. F., & Franceschi, C. (2003). Long-term immune-endocrine effects of
bereavement: relationships with anxiety levels and mood. Psychiatry Res., 121, 145-58.
Gerritsen, L., Wang, H. X., Reynolds, C. A., Fratiglioni, L., Gatz, M., & Pedersen, N. L. (2017).
Influence of negative life events and widowhood on risk for dementia. The American
Journal of Geriatric Psychiatry, 25(7), 766-778.
Glaser, R., Sheridan, J. F., Malarkey, W. B., MacCallum, R. C. & Kiecolt-Glaser, J. K. (2000).
Chronic stress modulates the immune response to a pneumococcal pneumonia vaccine.
Psychosom. Med., 62, 804-7.
Goodman, Y., Bruce, A. J., Cheng, B., & Mattson, M. P. (1996). Estrogens attenuate and
corticosterone exacerbates excitotoxicity, oxidative injury, and amyloid beta-peptide
toxicity in hippocampal neurons. J. Neurochem, 66, 1836–1844. 10.1046/j.1471-
4159.1996.66051836.x
Graham, J. E., Christian, L. M., & Kiecolt-Glaser, J. K. (2006). Stress, age, and immune
function: toward a lifespan approach. Journal of Behavioral Medicine, 29(4), 389-400.
Green, K. N., Billings, L. M., Roozendaal, B., McGaugh, J. L., & LaFerla, F. M. (2006).
Glucocorticoids increase amyloid-beta and tau pathology in a mouse model of
Alzheimer’s disease. J. Neurosci, 26, 9047–9056. 10.1523/JNEUROSCI.2797-06.2006

44

Håkansson, K., Rovio, S., Helkala, E. L., Vilska A. R., Winblad, B., Soininen, H., et al. (2009).
Association between mid-life marital status and cognitive function in later life:
population based cohort study. BMJ, 339:b2462.
Halleröd, B. (2013). Gender inequality from beyond the grave: intra-household distribution and
wellbeing after spousal loss. Ageing and Society, 33(5), 783-803.
Hamilton, S., & Fagot, B. (1988). Chronic stress and coping styles: a comparison of male and
female undergraduates. Journal of Personality and Social Psychology, 55(5), 819-23.
Hammen, C. (2005). Stress and depression. Annual Review of Clinical Psychology, 1, 293-319.
Harris-White, M., Chu, T., Miller, S., Simmons, M., Teter, B., Nash, D., Cole, G., & Frautschy,
S. (2001). Estrogen (E2) and glucocorticoid (Gc) effects on microglia and Aβ clearance
in vitro and in vivo. Neurochemistry International, 39, 435-48. 10.1016/S0197-
0186(01)00051-1.
Hatch, S. L., & Dohrenwend, B. P. (2007), Distribution of traumatic and other stressful life
events by race/ethnicity, gender, SES and age: a review of the research. American
Journal of Community Psychology, 40: 313-332. doi:10.1007/s10464-007-9134-z
Haug, M. R., & Folmar, S. J. (1986). Longevity, gender, and life quality. Journal of Health and
Social Behavior, 27(4), 323-45.
Helgeson, V. S. (1994). Relation of agency and communion to well-being: evidence and
potential explanations. Psychological Bulletin, 116(3), 412-28.
Holmes, T. H., & Rahe, R. H. (1967). The social readjustment rating scale. Journal of
Psychosomatic Research, 11, 213–218. http://dx.doi.org/10.1016/0022-3999(67)90010-4

45

Johansson, L., Guo, X., Hällström, T., et al. (2013). Common psychosocial stressors in middle-
aged women related to longstanding distress and increased risk of Alzheimer’s disease: a
38 year longitudinal population study. BMJ Open, 3: e003142. doi:10.1136/ bmjopen-
2013-003142
Johnson, J., Greaves, L., & Repta, R. (2009). Better science with sex and gender: facilitating the
use of a sex and gender-based analysis in health research. International Journal for
Equity in Health, 8(14).
Johnson, W., Turkheimer, E., Gottesman, I. I., & Bouchard, T. J. (2010). Beyond heritability:
twin studies in behavioral research. Current Directions in Psychological Science, 18(4),
217-220.
Justin, B. N., Turek, M., & Hakim, A. M. (2013). Heart disease as a risk factor for dementia.
Clinical Epidemiology, 5, 135-45. doi:10.2147/CLEP.S30621
Kato, K., & Pedersen, N. (2005). Personality and coping: a study of twins reared apart and twins
reared together. Behavior Genetics, 32(2), 147-58.
Kendler, S., Karkowski, L. M., & Prescott, C. (1999). Causal relationship between stressful life
events and the onset of major depression. The American Journal of Psychiatry, 156, 837-
41. 10.1176/ajp.156.6.837.
Kendler, K., Karkowski, L., & Prescott, C. (1999). The assessment of dependence in the study of
stressful life events: validation using a twin design. Psychological Medicine, 29(6), 1455-
1460.
Kendler, K. S., Thornton, L. M., & Prescott, C. A. (2001). Gender differences in the rates of
exposure to stressful life events and sensitivity to their depressogenic effects. The
American Journal of Psychiatry, 158(4), 587-93.

46

Kiecolt-Glaser, J. K. (1998). Stress, personal relationships, and immune function: health
implications. Brain. Behav. Immun., 13, 61–72.
Kiecolt-Glaser, J. K., Loving, T. J., Stowell, J. R., Malarkey, W. V., Lemeshow, S., Dickinson,
S. L., & Glaser, R. (2005). Hostile marital interactions, proinflammatory cytokine
production, and wound healing. Arch. Gen. Psychiatry, 62, 1377-84.
Kind, A. J., Bendlin, B., Kim, A. J., Koscik, R. L., Buckingham, W. R., Gleason, C., Blennow,
K., Zetterberg, H., Carlsson, C., & Johnson, S. C. (2017). Neighborhood socioeconomic
contextual disadvantage, baseline cognition, and Alzheimer’s disease biomarkers in the
Wisconsin Registry for Alzheimer’s Prevention (WRAP) study. Alzheimer’s & Dementia,
13(7), P195-6.
King, K., & Ogle, C. (2014). Negative life events vary by neighborhood and mediate the relation
between neighborhood context and psychological well-being. PLoS ONE, 9(4), e93539.
https://doi.org/10.1371/journal.pone.0093539
Kline, B. (2015). Principles and practice of structural equation modeling. New York, NY, US:
The Guilford Press.
Kobasa, S. C. (1979). Stressful life events, personality, and health. Journal of Personality and
Social Psychology, 3(1), 1-11.
Kraaij, V., Arensman, E., & Spinhoven, P. (2002). Negative life events and depression in elderly
persons: a meta-analysis. The Journals of Gerontology: Series B, 57(1), 87-94.
https://doi.org/10.1093/geronb/57.1.P87
Kressler, R., & McLeod, J. D. (1984). Sex differences in vulnerability to undesirable life events.
American Sociological Review, 49(5), 620-31.

47

Kroenke, K., & Spitzer, R. L. (1998). Gender differences in the reporting of physical and
somatoform symptoms. Psychosomatic Medicine, 60(2), 150-55.
Leonard, B. E. (2007). Inflammation, depression, and dementia: are they connected?
Neurochemical Research, 32(10), 1749-56.
Leserman, J., Petitto, J. M., Gu, H., Gaynes, B. N., Barroso, J., Golden, R. N., . . . Evans, D. L.
(2002). Progression to AIDS, a clinical AIDS condition and mortality: psychosocial and
physiological predictors. Psychological Medicine, 32(6), 1059-
1073.http://dx.doi.org/10.1017/S0033291702005949
Li, C. (2015). Confirmatory factor analysis with ordinal data: comparing robust maximum
likelihood and diagonally weighted least squares. Behavior Research Methods, 48, 436-
49.
Lillberg, K., Verkasalo, P., Kaprio, J., Teppo, L., Helenius, H., Koskenvuo, M. (2003). Stressful
life events and risk of breast cancer in 10,808 women: a cohort study. American Journal
of Epidemiology, 157(5), 415–423. https://doi.org/10.1093/aje/kwg002
Livingston, G., Sommerlad, A., Orgeta, V., Costafreda, S. G., Huntley, J., Ames, D. , Ballard, C.,
Banerjee, S., Burns, A., Cohen-Mansfield, J., Cooper, C., Fox, N., Gitlin, L.N., Howard,
R., Kales, H.C., Larson, E.B., Ritchie, K., Rockwood, K., Sampson, E.L., Samus, Q.,
Schneider, L.S., Selbæk, G., Teri, L., & Mukadam, N. (2017). Dementia prevention,
intervention, and care. Lancet, 390(10113), 2673-2732.
Lupien S. J., Maheu F., Tu M., Fiocco A., & Schramek T. E. (2007). The effects of stress and
stress hormones on human cognition: implications for the field of brain and cognition.
Brain Cogn, 65, 209–237. 10.1016/j.bandc.2007.02.007

48

Matthews, K., & Gump, B. (2002). Chronic work stress and marital dissolution increase risk of
posttrial mortality in men from the Multiple Risk Factor Intervention trial. Arch Intern
Med, 162, 309–15.
Matud, M. P. (2004). Gender differences in stress and coping styles. Personality and Individual
Differences, 37, 1401-15.
McEwen, B. S., & Stellar, E. (1993). Stress and the individual: mechanisms leading to disease.
Arch Intern Med, 153(18), 2093–2101. doi:10.1001/archinte.1993.00410180039004
McLaughlin, D., Vagenas, D., Pachana, N. A., Begum, N., & Dobson, A. (2010). Gender
differences in social network size and satisfaction in adults in their 70s. Journal of Health
Psychology, 15(5), 671-79.
McLeod, J. D., & Kessler, R. C. (1990). Socioeconomic status differences in vulnerability to
undesirable life events. Journal of Health and Social Behavior, 31(2), 162-172.
http://dx.doi.org/10.2307/2137170
Mielke, M., Vemuri, P., & Rocca, W. (2014). Clinical epidemiology of Alzheimer’s disease:
assessing sex and gender differences. Clinical Epidemiology, 6, 37-48.
Miller, S., & Kirsch, N. (1987). Sex differences in cognitive coping with stress. In R. C. Barnett,
L. Biener, & G. K. Baruch (Eds.), Gender and stress (p. 278–307). Free Press.
Mohr, D. C., Hart, S. L., Julian, L., Cox, D., & Pelletier, D. (2004). Association between
stressful life events and exacerbation in multiple sclerosis: a meta-analysis. BMJ (Clinical
research ed.), 328(7442), 731.
Mortimer, J., & Graves, A. (1993). Education and socioeconomic determinants of dementia and
Alzheimer’s disease. Neurology, 43(4), S39-S44.

49

Muthén, LK.; Muthén, BO. MPlus (8.2) [Computer software]. Los Angeles: Muthén & Muthén;
2018.
Muthén, L.K., & Muthén, B.O. (1998-2012). Mplus User’s Guide. Seventh Edition. Los
Angeles, CA: Muthén & Muthén
Myers, J., Lindenthal, J., Pepper, M., & Ostrander, D. (1972). Life events and mental status: a
longitudinal study. Journal of Health and Social Behavior, 13(4), 398-406.
Norton, M. C., Østbye, T., Smith, K. R., Munger, R. G. & Tschanz, J. T. (2009). Early parental
death and late-life dementia risk: findings from the Cache County study. Age and Ageing,
38, 340–343. doi:10.1093/ageing/ afp023.
Oman, R., & King, A. (2000). The effect of life events and exercise program format on the
adoption and maintenance of exercise behavior. Health Psychology, 19(6), 605-12.
Orth-Gomer, K., Wamala, S., Horsten, M., Schenck-Gustafsson, K., Schneiderman, N., &
Mittleman, M. (2000). Marital stress worsens prognosis in women with coronary heart
disease: the Stockholm female coronary risk study. JAMA, 284, 3008–14.
Ouanes, S., & Popp, J. (2019). High Cortisol and the Risk of Dementia and Alzheimer's Disease:
A Review of the Literature. Frontiers in Aging Neuroscience, 11, 43.
doi:10.3389/fnagi.2019.00043
Peavy, G. M., Lange, K. L., Salmon, D. P., Patterson, T. L., Goldman, S., Gamst, A. C., Mills, P.
J., Khandrika, S., … Galasko, D. (2007). The effects of prolonged stress and APOE
genotype on memory and cortisol in older adults. Biological Psychiatry, 62(5), 472-8.
Pedersen, W. A., Wan, R., & Mattson M. P. (2001). Impact of aging on stress-responsive
neuroendocrine systems. Mech. Ageing Dev, 122, 963–983. 10.1016/S0047-
6374(01)00250-0

50

Pedersen, N. L. Swedish Adoption/Twin Study on Aging (SATSA), 1984, 1987, 1990, 1993,
2004, 2007, and 2010. Ann Arbor, MI: Inter-university Consortium for Political and
Social Research [distributor], 2015-05-13. https://doi.org/10.3886/ICPSR03843.v2
Perris, H. (1984). Life events and depression: part 1. effect of sex, age, and civil status. Journal
of Affective Disorders, 7(1), 11-24.
Persson, G. (1980). Life event ratings in relation to sex and marital status in a 70‐year‐old urban
population. Acta Psychiatrica Scandinavica, 62, 112-118. doi: 10.1111/j.1600-
0447.1980.tb00599.x
Persson, G. & Skoog, I. (1996). A prospective population study of psychosocial risk factors for
late onset dementia. International Journal of Geriatric Psychiatry, 11, 15–22.
doi:10.1002/(SICI)1099-1166(199601)11:1<15::AIDGPS262>3.0.CO;2-5.
Pilleron, S., Guerchet, M., Ndamba-Bandzouzi, B., Mbelesso, P., Dartigues, J. F, Preux P. M., &
Clément J. P. (2015). Association between stressful life events and cognitive disorders in
Central Africa: results from the EPIDEMCA program. Neuroepidemiology, 44, 99-107.
doi: 10.1159/000375462
Pinquart, M., & Sörensen, S. (2006). Gender differences in caregiver stressors, social resources,
and health: an updated meta-analysis. The Journals of Gerontology: Series B, 61(1), 33-
45.
Plomin, R., Lichtenstein, P., Pedersen, N. L., McClearn, G. E., & Nesselroade, J. R. (1990).
Genetic influence on life events during the last half of the life span. Psychology and
Aging, 5(1), 25-30. http://dx.doi.org/10.1037/0882-7974.5.1.25
Ptacek, J. T., Smith, R. E., & Zanas, J. (1992). Gender, appraisal, and coping: a longitudinal
analysis. The Journal of Personality, 60(4), 747-70.

51

Ravindran, T. K. (2002). Editorial. Journal of Health Management 4(2), v-vii.
R Core Team (2013). R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. http://www.R-project.org/.
Rafanelli, C., Roncuzzi, R., Milaneschi, Y., Tomba, E., Colistro, M. C., Pancaldi, L. G., & Di
Pasquale, G. (2005). Stressful life events, depression and demoralization as risk factors
for acute coronary heart disease. Psychother Psychosom, 74, 179-184. doi:
10.1159/000084003
Rahe, R. H., Lundberg, U., Theorell, T., & Bennett, L. K. (1971). The social readjustment rating
scale: a comparative study of Swedes and Americans. Journal of Psychosomatic
Research, 51, 241-249.
Read, S., Pedersen, N., Gatz, M., Berg, S., Vuoksimaa, E., Malmberg, B., Johansson, B., &
McClearn, G. (2006). Sex differences after all those years? Heritability of cognitive
abilities in old age. The Journals of Gerontology: Series B, 61(3), 137-43.
Rich-Edwards, J. W., Kaiser, U., Chen, G., Manson, J., & Goldstein, J. (2018). Sex and gender
differences research design for basic, clinical, and population studies: essentials for
investigators. Endocrine Reviews, 39(4), 424-39.
Rijsdijk, F. V., & Sham, P. C. (2002). Analytic approaches to twin data using structural equation
models. Brief Bioinform, 3(2), 119-33.
Roos, P., & Reskin, B. (1984). Institutional factors contributing to sex segregation in the
workplace. Sex segregation in the workplace: trends, explanations, remedies, 235-60.
Rosario, M., Shinn, M., Morch, H., & Huckabee, C. (1988). Gender differences in coping and
social supports: testing socialization and role constraint theories. Journal of Community
Psychology, 16, 55-69.

52

Ross, C., & Mirowsky, J. (1979). A comparison of life-event-weighting schemes: change,
undesirability, and effect-proportional indices. Journal of Health and Social Behavior,
20(2), 166-77.
Scuteri, A., Bos, A., Zonderman, A., Brant, L., Lakatta, E., & Fleg, J. (2001). Is the apoE4 allele
an independent predictor of coronary events? The American Journal of Medicine, 110(1),
28-32.
Sharkey, S., Lesser, J., Zenovich, A., et al. (2005). Acute and reversible cardiomyopathy
provoked by stress in women from the United States. Circulation, 111, 472–9.
Slavich, G. M. (2016). Life stress and health: a review of conceptual issues and recent findings.
Teaching of Psychology, 43(4), 346–355. https://doi.org/10.1177/0098628316662768.
Sommerlad, A., Ruegger, J., & Singh-Manoux, A., et al. (2018). Marriage and risk of dementia:
systematic review and meta-analysis of observational studies. J Neurol Neurosurg
Psychiatry, 89, 231-38.
Sörman, D., Rönnlund, M., Sundström, A., Adolfsson, R., & Nilsson, L. (2015). Social
relationships and risk of dementia: A population-based study. International
Psychogeriatrics, 27(8), 1391-1399. doi:10.1017/S1041610215000319
Stoller, E. (1990). Males as helpers: the role of sons, relatives, and friends. The Gerontologist,
30(2), 228-35.
Stone, A., & Neale, J. (1984). New measure of coping: development and preliminary results.
Journal of Personality and Social Psychology, 46(4), 892-906.
Sundström, A., Rönnlund, M., Adolfsson, R., & Nilsson, L. G. (2014). Stressful life events are
not associated with the development of dementia. International Psychogeriatrics, 26(1),
147-54.

53

Sundström, A, Westerlund, O., & Kotyrlo, E. (2016). Marital status and risk of dementia: a
nationwide population-based prospective study from Sweden. BMJ Open, 6:e008565.
doi: 10.1136/bmjopen-2015-008565.
Swinkels, J., van Tilburg, T., Verbakel, E., & van Groenou, M. (2019). Explaining the gender
gap in the caregiving burden of partner caregivers. The Journals of Gerontology: Series
B, 74(2), 309-17.
Takeda, S., Sato, N., & Morishita, R. (2014). Systemic inflammation, blood-brain barrier
vulnerability and cognitive/non-cognitive symptoms in Alzheimer’s disease: relevance to
pathogenesis and therapy. Frontiers in Aging Neuroscience, 6, 1-8.
Tamres, L., Janicki, D., & Helgeson, V. (2002). Sex differences in coping behavior: a meta-
analytic review and an examination of relative coping. Personality and Social Psychology
review, 6(1), 2-30.
Toledo, J. B., Toledo, E., Weiner, M. W., Jack, C. R., Jr., Jagust, W., Lee, V. M., et al. (2012).
Cardiovascular risk factors, cortisol, and amyloid-beta deposition in Alzheimer’s disease
neuroimaging initiative. Alzheimer’s Dement, 8, 483–489. 10.1016/j.jalz.2011.08.008
Towards a dementia plan: a WHO guide. Geneva: World Health Organization; 2018. Licence:
CC BY-NC-SA 3.0 IGO.
Troisi, A. (2001). Gender differences in vulnerability to social stress: a Darwinian perspective.
Physiology and Behavior, 73(3), 443-49.
Turkheimer, E., & Harden, K. P. (2014). Behavior genetic research methods: testing quasi-causal
hypotheses using multivariate twin data. In H. T. Reis & C. M. Judd (Eds.), Handbook of
research methods in social and personality psychology (pp. 159-187). New York, NY,
US: Cambridge University Press.

54

U. S. Bureau of Labor Statistics. (2019). American time use survey – 2018 results.
Wang, H., Karp, A., Winblad, B., & Fratiglioni, L. (2002). Late-life engagement in social and
leisure activities is associated with a decreased risk of dementia: a longitudinal study
from the Kungsholmen Project. American Journal of Epidemiology, 155(12), 1081-87.
https://doi.org/10.1093/aje/155.12.1081
Wimo, A., Jonsson, L., Bond, J., Prince, M., Winblad, B., & Alzheimer Disease International.
(2013). The worldwide impact of dementia 2010. Alzheimer’s & Dementia, 9(1), 1-11.
Wittstein, I., Thiemann, D., Lima, J., et al. (2005). Neurohumoral features of myocardial
stunning due to sudden emotional stress. N Engl J Med, 352, 539–48.
Zverova, M., Fisar, Z., Jirak, R., Kitzlerova, E., Hroudova, J., & Raboch, J. (2013). Plasma
cortisol in Alzheimer’s disease with or without depressive symptoms. Med. Sci. Monit,
19, 681–689. 10.12659/MSM.889110

55

Appendix A
Items in the modified version of the Social Readjustment Rating Scale in SATSA

LE # LE Meaning
1 Retirement after employment
2 Major deterioration in financial status
3 Serious illness in child
4 Death of child
5 Serious conflicts with child
6 Somatic illness, self
7 Forced change in residence b/c one can’t manage to look after oneself
8 Divorce
9 Home care of spouse by proband
10 Getting married (NOT IN IPT3)
11 Deterioration in married life
12 Somatic illness, spouse
13 Death of spouse
14 Nursing home care, spouse
15 Mental illness, spouse
16 Improvement in married life (NOT IN IPT3)
17 Home care, self
18 Forced change in residence with reduced contact
19 Mental illness, self
20 Death of siblings or friends
22 Loss of sexual ability or interest
24 Major improvement in financial status (NOT IN IPT3)
25 Making an acquaintance (NOT IN IPT3)

56

Abstract (if available)

Abstract Stressful life events are associated with cognitive aging and dementia risk. It remains unclear whether negative and positive life events are causally related to dementia risk and if existing associations are accounted for by genetic and shared environmental confounds. Further, it is unclear whether these life events are differentially associated with dementia risk by gender. Gender may affect the pathways through which different life events correlate with dementia risk. This study examined whether there is support for a causal association between six different life event domains and dementia risk by gender. The study tested and controlled for the confounding influence of genetics and shared environmental factors in explaining the association between these six life event domains and dementia risk for men and women. Data from the first three waves of the Swedish Adoption/Twin Study of Aging (SATSA) were used (MZM = 199, MZF = 269, DZM = 345, DZF = 556). Life events were assessed using standard measures modified for older adults. Dementia diagnoses were made based on in-person interviews or Swedish national registry linkage. Structural equation modeling was used to evaluate the hypothesis that there is a causal relationship between six life event domains and dementia risk by gender, as well as quantify and account for shared genetic and environmental influences that may confound the association. We hypothesized that life events would differentially associate with dementia risk by gender via both familial confounding and causal processes. There was little evidence for causal processes for men and women across life event domains. Familial confounding (genes and common environment) appeared to drive the observed phenotypic associations. Gender differences were found for phenotypic associations as well as between-family effects.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Dementia, co-morbid depressive symptoms, and mortality in older adults: a twin study

PDF

Associations between longitudinal loneliness, epigenetic age, and dementia risk

PDF

Underlying neural mechanisms of depression and dementia

PDF

Examining the effect of fertility on lifespan using data from the Swedish Twin Registry

PDF

Affective neuropsychiatric symptoms and neural connectivity in the early stages of Alzheimer’s disease

PDF

Untreated hyperglycemia associated with tau pathology and worse cognitive performance in older adults

PDF

Internet communication use, psychological functioning and social connectedness at older ages

PDF

Evaluating a cultural process model of depression and suicidality in Latino adolescents

PDF

The association of cerebrovascular disease risk factors with brain structure and its modification by genetic variation

PDF

A neuropsychological exploration of low-SES adolescents’ life goals and their motives

PDF

Renin-angiotensin system antihypertensive medicines with and without blood-brain-barrier crossing potential: effects on cognitive outcomes in the elderly

PDF

Family aggression exposure and community violence exposure associated with brain volume in late adolescence: a comparison of automated versus manual segmentation

PDF

Is affluence a developmental risk for Hong Kong Chinese adolescents?

PDF

A genome wide association study of multiple sclerosis (MS) in Hispanics

PDF

Developmental etiology of reactive and proactive aggression, electrodermal reactivity and their relationships: a twin study

PDF

Adolescent life stress and the cortisol awakening response: the moderating roles of emotion regulation, attachment, and gender

PDF

Personalized normative feedback applied to undergraduates with problem drinking: a comparison with psychoeducation and an examination of cognitive-affective change mechanisms via the articulated ...

PDF

Metabolic consequences of obesity-associated inflammation during puberty and perinatal development

PDF

Genetic epidemiological approaches in the study of risk factors for hematologic malignancies

PDF

Quality of life of patients with cleft lip and palate undergoing orthodontic treatment during early vs. late adolescence

Asset Metadata

Creator Kim, Alice Jinna (author)

Core Title Are life events differentially associated with dementia risk by gender? A twin study

School College of Letters, Arts and Sciences

Degree Master of Arts

Degree Program Psychology

Publication Date 10/12/2020

Defense Date 04/28/2020

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

Tag dementia,gender,life events,OAI-PMH Harvest,twin study

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Beam, Christopher R. (committee chair), Ailshire, Jennifer (committee member), Gatz, Margaret (committee member), Walsh, David (committee member)

Creator Email ajk258@cornell.edu,alicejki@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-380569

Unique identifier UC11666483

Identifier etd-KimAliceJi-9035.pdf (filename),usctheses-c89-380569 (legacy record id)

Legacy Identifier etd-KimAliceJi-9035.pdf

Dmrecord 380569

Document Type Thesis

Rights Kim, Alice Jinna

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA