1. Incident dementia risk among spouse dementia caregivers
Incident Dementia Risk Among Spousal Dementia Caregivers:
The Cache County Study.
Maria C. Norton, Ken R. Smith, Truls Østbye, JoAnn T. Tschanz, Chris Corcoran,
Sarah Schwartz, Kathleen W. Piercy, Peter V. Rabins, David C. Steffens, Ingmar Skoog,
John C. S. Breitner, Kathleen A. Welsh-Bohmer
For the Cache County Investigators
From the Department of Family Consumer and Human Development, Utah State University
(Drs. Norton & Piercy); the Department of Family and Consumer Studies and the Huntsman
Cancer Institute, University of Utah (Dr. Smith); the Department of Community and Family
Medicine, Duke University Medical Center (Dr. Østbye); the Department of Psychology, Utah
State University (Drs. Norton & Tschanz), the Department of Mathematics and Statistics, Utah
State University (Dr. Corcoran and Ms. Schwartz); Department of Psychiatry and Behavioral
Sciences, The Johns Hopkins University School of Medicine (Dr. Rabins), the Department of
Psychiatry and Behavioral Sciences, Duke University Medical Center (Dr. Steffens); VA Puget
Sound Health Care System and Department of Psychiatry and Behavioral Sciences, University of
Washington School of Medicine, Seattle (Dr. Breitner); the Joseph and Kathleen Bryan
Alzheimer’s Disease Research Center, Duke University (Drs. Østbye & Welsh-Bohmer).
Address correspondence to: Dr. Maria Norton, Cache County Memory Study,
Utah State University, 4440 Old Main Hill, Logan, UT 84322-4440;
Phone: 435-797-1599; Email address firstname.lastname@example.org; Fax: 435-797-2771
Word Count: 4,556
Portions of this paper were presented at the International Conference on Alzheimer’s Disease in
Vienna, Austria, July 11-16, 2009.
This work was supported by NIH grants: AG-031272, AG-011380, and AG-021136
2. Incident dementia risk among spouse dementia caregivers
Chronic psychosocial stressors can increase risk in caregivers for adverse health outcomes
including depression, anxiety, and cognitive decline. Few studies have examined the long-term
effects on incident dementia risk of spousal dementia caregiving. In a population-based sample
of 2,442 subjects (1,221 married couples) aged 65 and older, we tested the hypothesis that
spousal dementia caregiving is associated with increased risk for incident dementia in the
caregiver. Incident dementia was diagnosed in 255 subjects Cox proportional hazards
regression tested the effect of time-dependent exposure to dementia caregiving adjusted for age,
gender, APOE genotype, and (to control for shared socioeconomic status that may impact shared
dementia risk) husband’s education and occupation. A subject (i.e.caregiver) whose spouse
experienced incident dementia onset had a six-fold (HR=6.01, 95% CI: 2.23-16.17, p=.0004)
increase in the hazard for incident dementia in relation to subjects whose spouses were dementia
free. In gender-stratified analyses, husband caregivers were at even higher risk (HR=11.93,
95%CI: 1.67-85.52, p=.0136), compared to wife caregivers (HR=3.66, 95%CI: 1.15-11.61,
p=.0277). These results suggest that the chronic and often severe stress associated with
dementia caregiving exerts substantial risk for the development of dementia in spouse caregivers,
with relative effects greater for husband caregivers.
3. Incident dementia risk among spouse dementia caregivers
Informal (unpaid) dementia caregiving is a difficult job involving great demands on the
caregiver’s time and energy, with typically great physical demands in providing personal and
instrumental assistance to the spouse with dementia (the care recipient). The caregiver burden
often seems to be a cycle repeated every 24 hours which, to many caregivers, may seem like a
“36 hour day” (Mace & Rabins, 2006). A feeling of loss of personal control or loss of individual
self has also been associated with dementia caregiving (Boss, 1990). A growing body of
research on dementia caregiving has demonstrated higher rates of depression (Bergman-Evan,
1994; Schulz et al, 1990; Pinquart & Sorensen, 2003; Sörensen et al, 2006), other physical health
problems (Vitaliano, et al, 2003) and greater mortality (Christakis & Allison, 2006; Schulz &
Beach, 1999) among caregivers. This may be especially true for spouse caregivers who are of
the same age as, or even older than, the person with dementia, with the role of full-time caregiver
potentially jeopardizing their own health.
While spousal dementia caregivers may report a variety of positive aspects to their
caregiving role (Bacon et al, 2009), the unique and extreme challenges of caregiving may result
in physical and mental exhaustion, feelings of anxiety, loneliness and hopelessness. While
providing care to an older person with physical impairments or other illnesses may also include
many of these stressors, a study of over 1,500 caregiving households revealed that dementia
caregivers provided more assistance, reported more personal sacrifices and reported that
caregiving was more stressful, compared to caregivers of physically-impaired older adults (Ory,
4. Incident dementia risk among spouse dementia caregivers
While much has been published on the effect of caregiver stress on depression and
cardiovascular health in the caregiver, very few studies have examined effects on the caregiver’s
cognitive function and subsequent risk for dementia. A plausible biological mechanism has been
suggested by McEwen and Sapolsky through maladaptive neuroendocrine response from chronic
stress (McEwen; 2002; Sapolsky, Krey,& McEwen, 1986). An emerging literature showing
associations between chronic stress from spousal caregiving and cognitive deficits in the
caregivers supports this hypothesis. deVugt et al (2006) found that spouse dementia caregivers
had significantly worse general cognitive functioning, speed of information processing and
verbal memory, compared to age-, sex-, and education-matched controls. Another study found
that caregivers of spouses with dementia performed more poorly on a digit symbol test of
complex attention and cognitive speed (Caswell, et al, 2003). Results from the Nurses Health
Study revealed that female caregivers of ill spouses performed worse than non-caregivers on a
general cognitive screening test, and on a test of immediate and delayed recall of a 10-item word
list (Lee, et al, 2004). Decline in vocabulary over a two-year period was significantly greater in
caregivers of spouses with dementia than in matched controls (Vitaliano, et al, 2005). Finally,
Mackenzie et al (2007) demonstrated that caregivers of relatives in palliative care exhibited
significant impairments in attention and had difficulty monitoring their performance and
regulating attentional resources, compared to a control sample.
While these few studies show consistent evidence that the stress associated with spousal
dementia caregiving may lead to declines in cognitive function for the caregiver, all studies were
limited by small sample sizes. Importantly, no study included longitudinal follow-up of
caregivers combined with in-depth clinical assessment for dementia that confirms whether these
observed cognitive impairments are transitory or may actually lead to dementia in the caregiver.
5. Incident dementia risk among spouse dementia caregivers
We report here results from a population-based study of dementia wherein 2,442 older
married individuals participated in up to 12 years of longitudinal cognitive evaluation. An in-
depth dementia ascertainment protocol was employed to diagnose dementia in both spouses.
Dementia onset ages were determined for each spouse, facilitating time-dependent analyses of
the effects of exposure to spousal dementia caregiving. We hypothesized that after onset of
caregiving for dementia, the caregiver is subsequently at increased risk for incident dementia,
due to the chronic stress associated with spousal dementia caregiving. Further, because of the
higher rates of incident dementia (Miech et al, 2002) and depression (Luijendijk et al, 2008)
among women, and the greater tendency among women caregivers to report poorer well-being as
a function of caregiver burden (Croog, et al, 2001), we hypothesized that the effect of spousal
dementia caregiving on subsequent dementia in the caregiver would be stronger for wives caring
for husbands with dementia, than for husbands caring for wives with dementia.
The Cache County Study on Memory Health and Aging (CCMS) is an on-going
investigation of aging in general, and dementia and cognitive impairment in particular,
among the elderly population of Cache County, Utah. In 1995 we invited all elderly (age 65 or
older) individuals who were permanent residents of Cache County to participate in the first wave
of our study. There were 5,092 participants at the baseline interview in the CCMS, i.e. 90% of
all eligible residents of Cache County. In this cohort, 3299 were married, 1534 were widowed,
167 were divorced or separated, 82 had never married, and 10 had missing marital status, as of
the baseline interview. Of those who were married, 2,738 had a spouse who was also a CCMS
participant. After removing from this sample 296 subjects with prevalent dementia, the final
analysis sample for the present study included 2,442 participants, i.e. 1,221 married couples.
6. Incident dementia risk among spouse dementia caregivers
Each of these 2,442 participants serves as both the index subject whose incident dementia is
being modeled, as well as the spouse whose dementia diagnosis (or lack thereof) is used to
generate the exposure variable for each index subject. Of the 2442 observed participants, 229
were “exposed to” a spouse diagnosed with dementia and thereby to spousal dementia
caregiving. These subjects with known exposure were observed an average of 9.22 (SD =3.05)
years. This time includes observation time before (M = 5.09, SD = 3.20 years) and after (M =
4.14, SD = 3.15 years) spouse’s dementia onset. The remaining 2213 subjects who were never
exposed to a spouse diagnosed with dementia were followed for a slightly longer period, on
average (M = 4.45, SD = 4.24 years).
The CCMS utilized a multi-stage dementia ascertainment protocol described elsewhere
(Breitner et al, 1999), which was repeated in four successive triennial waves (1995-7, 1998-
2000, 2002-4, and 2005-7). Briefly, within each wave, screening began with an in-person
interview that included an adaptation (Tschanz et al, 2002) of the modified Mini-Mental State
examination (3MS; Teng & Chui, 1987) or, for those unable to participate, the Informant
Questionnaire for Cognitive Disorder in the Elderly (IQCODE; Jorm & Jacomb, 1989)
administered to a collateral informant. Participants who screened positive for possible dementia
then completed an in-depth clinical assessment (CA), as did a 19% sub-sample of “designated
controls” who completed a CA in each wave. Specially trained nurses and psychometric
technicians administered the CA, which included a brief physical evaluation, a detailed history of
medical and cognitive symptoms, a structured neurological examination, and a one-hour battery
of neuropsychological tests.
7. Incident dementia risk among spouse dementia caregivers
A geriatric psychiatrist and neuropsychologist reviewed these data and assigned working
diagnoses of dementia (DSM-III-R; 1987) or other cognitive syndromes. Subjects with a
working diagnosis of dementia were then selected for geriatric psychiatric examinations and
laboratory studies including neuroimaging. To substantiate or refine their diagnoses, these
individuals were re-examined 18 months after their initial CA. A multidisciplinary consensus
panel of experts in neurology, geriatric psychiatry, neuropsychology, and cognitive neuroscience
reviewed all available data and assigned final consensus diagnoses. Onset was defined as the
year in which a participant unambiguously met DSM-III-R criteria for dementia.
Imputation of Incident Dementia for Subjects with Incomplete Ascertainment
Incomplete ascertainment was identified in 253 subjects who screened positive for
dementia but failed to complete the CA (due to death, refusal or moving out of area between
screening and CA). Four possible methods for addressing this scenario were considered. One
approach would be to assume that 100% of all such individuals had dementia at the point of the
screen positive result, while a second approach would assume that none of these individuals ever
had dementia. A third approach would discard all observation years between the prior wave
(where subject received a non-dementia diagnosis) to the point of the incomplete ascertainment
with the screen positive result, an option that would discard 1,092 person-years of observation.
A fourth approach (believed to get closest to “truth”) would be to impute dementia onset by
utilization of all available screening data, employing conservative diagnostic criteria for
dementia. It is this fourth approach that was used in the present study, where screening data
were used to infer dementia status for those with incomplete ascertainment at their last wave of
participation. [Note: We also ran models using the third approach (defining right censoring as of
preceding wave with non-dementia diagnosis, i.e. no imputation) but as results were comparable,
8. Incident dementia risk among spouse dementia caregivers
we report herein only the models using the imputation approach which resulted in preservation of
all person-years of observation.].
In Waves 1 and 2, individuals who were screen-positive on either the 3MS or IQCODE
were asked to undergo further evaluation with the Dementia Questionnaire (DQ), a semi-
structured informant interview designed to inventory cognitive or functional difficulties and
medical conditions relevant to dementia (Kawas, et al, 1994). The DQ interviews were rated by
a study neuropsychologist who estimated degree of cognitive impairment as 1=cognitively
normal, 2=mild cognitive impairment, 3=moderate cognitive impairment, 4=severe cognitive
impairment, or 5=dementia. Among subjects with “incomplete ascertainment,” where a DQ
rating was available (Waves 1 and 2), it was used, with dementia imputed if DQ rating = 5, and
non-dementia was imputed if DQ rating <5. When no DQ was available in Waves 1-4, data from
the first screening stage were used, with conservative cut-points. If the IQCODE score was >=
3.6 (Se=.80, and Sp=.82; Jorm, 2004) or if the 3MS was below an age/gender/education-specific
cut-off score (Se=1.00-.61, Sp=.83-.94 for age groups 65-69, 70-74, 75-79, 80-84, 85-89, 90+
years, respectively; Tschanz, et al, 2002), dementia was imputed, and non-dementia was imputed
otherwise. Whenever dementia was imputed with DQ rating, 3MS, or IQCODE data, the onset
age was imputed to be the mid-point between the last wave of complete (non-dementia)
evaluation and the wave with incomplete ascertainment. This strategy “recouped” 1,092 lost
person-years of observation that would otherwise have resulted in shorter observation periods for
subject or spouse (or both).
9. Incident dementia risk among spouse dementia caregivers
Subject’s gender and age (in years) at baseline interview were included as covariates.
Additionally, at baseline interview, we obtained buccal DNA from consenting participants (97%)
for determination of genotype at APOE. Following the method of Richards and colleagues
(1993) APOE genotypes were determined using polymerase chain reaction (PCR) amplification
and a restriction isotyping method described by Saunders and colleagues (1993).
To adjust for shared environmental exposures that might increase risk for dementia in
both spouses, and thereby be confounded with the effect of spousal dementia caregiving, we
adjusted for socioeconomic status (SES), a significant predictor of many health outcomes
including dementia (Fotenos et al, 2008; Goldbourt et al, 2007). SES was measured with
husband’s education and occupation, given that these are much stronger predictors of couple-
level socioeconomic status, on average in this cohort, than wife’s education and occupation.
Husband’s education was recorded in years. Following the Dictionary of Occupational Titles
(U. S. Department of Labor, 1991), husband’s occupation of longest duration was coded into
nominal categories of (1) professional, technical, managerial; (2) clerical and sales; (3) service;
(4) agricultural; (5) processing; (6) machine trades; (7) bench work; (8) structural work, and (9)
miscellaneous. Given the very low frequencies of categories 5-9, these were combined into one
group, labeled hereafter as “machine/misc.”
A series of Cox proportional hazards regression models were computed to test for the
association between spouse dementia caregiving, a time-varying covariate, and subsequent risk
for incident dementia in the caregiver, before and after adjustment for covariates (Cox & Oakes,
10. Incident dementia risk among spouse dementia caregivers
1984). Using expert consensus diagnoses and the imputation protocol described above, each of
the 2,442 subjects was identified as being either an incident dementia case (with corresponding
age of onset—to the nearest integer year), or was considered “right censored” at the last
completed visit. Cox models were fit, stratified by baseline age group (<75, 75-85, 85+ years),
which allows each stratum to have its own baseline hazard rate. Thus, the parameter estimate for
“age” in the model actually reflects the general effect of each additional year of age as applied to
a given age stratum (Cox & Oakes, 1984).
For purposes of modeling time to incident dementia, each subject’s observation time
started at their own baseline screening visit and ended at 1) their own dementia onset, 2) their
own right-censoring, or 3) their spouse’s right-censoring (death or loss to follow-up), whichever
took place first. When spouse’s right-censoring date occurred first, the decision to end subject’s
observation at this earlier date was a conservative but necessary approach due to unknown
exposure to dementia caregiving from the point of spouse’s right-censoring forward. Thus,
spouse dementia onset (i.e. subject’s exposure to dementia caregiving) was ignored if it occurred
after subject’s dementia onset, and subject’s dementia onset was ignored if it occurred after
spouse’s right-censoring (due to unknown exposure status from this point forward).
Dementia Caregiving Exposure Time.
Although many caregiving activities likely commenced previously, for analyses reported
here, exposure to spousal dementia caregiving was presumed to commence at the spouse’s
dementia diagnosis. In epidemiological studies generally, risk factors can have long-lasting
residual effects after initial exposure (e.g. development of AIDS after exposure to the HIV
11. Incident dementia risk among spouse dementia caregivers
virus), or alternatively, risk may return to normal or near-normal levels after exposure has ended
(e.g. risk for development of cancer after exposure to tobacco smoking). Recent studies
demonstrate that caregivers experience continued deterioration of spatial working memory and
delayed episodic memory, even after the death of the care recipient (Mackenzie et al, 2007).
Such associations suggest long-term cumulative effects of psychosocial stress associated with
dementia care-giving. Thus, in the present work, the approach of “once exposed, always
exposed” was implemented with the exposure variable not being reset to “unexposed” after death
of the spouse with dementia.
Adjustments for Widowhood and Dependent Observations within Dyads
The very small number of unexposed subjects (0 men, 4 women) experiencing
widowhood over the entire period of observation time precluded our ability, but obviated the
need, to adjust for the potential confounding effects of widowhood. It is possible that the
association between the risks of dementia shared by spouses may not be the result of dementia
caregiving of one spouse for another but rather because they are exposed to some common but
unobserved factor. This possibility was considered by refitting the Cox proportional hazard
models but with the inclusion of a shared frailty parameter that captures the potential influence
of a shared risk among spouses. A shared frailty model is a random effects model where the
frailties are common among spouses (Cleves, Gould, Gutierrez & Marchenko, 2008). Models
that include a shared frailty generated results (not shown) were comparable to those reported
here. Additionally, analyses were stratified by subject gender, to separately examine effects of
12. Incident dementia risk among spouse dementia caregivers
spousal dementia caregiving on husbands vs. wives (with such models not requiring the inclusion
of the shared frailty parameter).
In this community-based sample of 2,442 older adults (from 1,221 married couples),
husband baseline age ranged from 65.7 to 100.1 (M=75.7, SD=5.9) years and wife baseline age
ranged from 65.4 to 93.7 (M=73.1, SD=5.3) years. Length of marriage before baseline ranged
from 1 to 73 years (M=48.9, SD=11.5). Only four couples divorced some time after baseline
interview, with all other couples remaining married up to index subject’s dementia onset or right
censoring, or up to right-censoring in spouse, whichever occurred first. Table 1 provides
demographic summary statistics to describe the sample, along with statistical tests of bivariate
relationships. Incident dementia was significantly associated with older age (p<.001), exposure
to dementia caregiving (p=.022), and having one or two e4 alleles at APOE (p<.001).
An initial proportional hazards regression model of survival time to incident dementia
revealed a significant hazard ratio for exposure to spousal dementia caregiving (HR=6.399,
95%CI: 2.377-17.226, p=.0002). In the fully-adjusted model (Table 2), older age (p=.0123),
one e4 allele (p=.0068) or two e4 alleles (p<.0001) (versus no e4 allele) at the APOE locus were
associated with significantly higher risk of incident dementia, while higher SES (husband with
professional/ technical/managerial occupation) was associated with significantly lower risk of
incident dementia (p=.0183). In this same fully-adjusted model, exposure to spousal dementia
caregiving was still associated with a six-fold increased risk for incident dementia (HR=6.01,
95%CI: 2.231-16.171, p=.0004). When analyses were stratified by gender of the index subject
13. Incident dementia risk among spouse dementia caregivers
(i.e. dementia caregiver), husband caregivers were at even higher risk (HR=11.93, 95%CI: 1.67-
85.52, p=.0136), compared to wife caregivers (HR=3.66, 95%CI: 1.15-11.61, p=.0277).
To our knowledge, this is the first population-based study to examine risk for incident
dementia among spouse caregivers of persons with dementia. We observed a striking six-fold
increased risk for incident dementia among older adults exposed to spousal dementia caregiving,
compared to married older adults whose spouse never developed dementia. This effect was
robust after adjusting for age, gender, education and APOE genotype, all factors known to be
associated with dementia risk. This effect also remained after further adjustment for
socioeconomic status, a key indicator of shared environmental exposures, which, at least in part,
ruled out confounding with common lifestyle factors that may lead to increased dementia risk in
both spouses (e.g. diet, access to medical care).
Serving as primary care provider to an individual with dementia is stressful regardless of
age (Mace & Rabins, 2006), but the burden for spouse caregivers may be even greater due to
their own advanced age, possible functional limitations, and greater likelihood of fatigue at
physical exertion. As suggested by interdependence theory of dyadic relationships (Kelley &
Thibaut, 1978), spouse caregivers, as with spouses or relationship partners generally, become
interdependent over time through their interactions. As interdependence increases, so does
concern for the partner’s outcomes, and a transformation of motives occurs from motives of self
interest to pro-relationship motives (Lewis, McBride, Pollak, Puleo, Butterfield, & Emmons,
2006). However, cognitive and functional decline in the spouse with dementia forces caregivers
to face the realization that, while there may be occasional “good days” and successes, the
14. Incident dementia risk among spouse dementia caregivers
progressive nature of dementia is relentless. Neuropsychiatric disturbances are particularly
stressful and are the most common reason for institutionalization (Kalapatapu & Neugroschl,
2009). The caregiver must grieve not only the certainty of the spouse’s death in the near future
leaving them with less time together, but must also grieve the loss of the pre-dementia
relationship quality as the spouse with dementia experiences further declines as the disease
advances. The caregiver may or may not have or take the opportunity for respite periodically to
rejuvenate, but if this indulgence is permitted oneself, often there is guilt associated with the
temporary delegation of caregiving duties to others (Albinsson & Strang, 2003). All of these
factors impose great stress burdens on the spouse caregiver and those without strong support
systems or who have more depressed or anxious tendencies are more likely to suffer the
deleterious effects on the brain, increasing their own risk for dementia.
Our findings provide evidence that the cognitive impairments demonstrated by other
studies among spouse dementia caregivers (deVugt et al, 2006; Caswell, et al, 2003; Lee, et al,
2004; Vitaliano, et al, 2005; Mackenzie et al, 2007) are not always transient, and that there is a
significant risk for the chronic stress of caregiving to result in dementia in the care provider.
The magnitude of effect we observed was not borderline in clinical significance but was of
substantial significance—a six-fold increase in risk overall, a finding that can be of vital
importance for healthcare providers. Vigilance in monitoring the well-being and coping of the
spouse dementia caregiver is recommended, while at the same time continuing to treat the person
with dementia. Treatment for depression, stress management interventions, dementia
educational seminars, and referrals to respite services are among the helps that healthcare
providers can offer the spouse caregiver. Further, the caregiver’s health can directly affect the
quality of care provided to care recipients, because preserving cognitive function in the caregiver
15. Incident dementia risk among spouse dementia caregivers
will ensure better attention to a host of factors such as personal hygiene, medication compliance,
and engagement in cognitively stimulating activities for the care recipient. Thus, practitioners
who treat a patient with dementia should also consider the dyad, and provide surveillance,
diagnosis and treatment to two patients, not just one.
Psychosocial stress is often associated with elevated levels of biomarkers of
physiological stress reactivity. Chronic stress, especially in those with maladaptive coping
strategies, may lead to a disregulation of the neuroendocrine response, ultimately resulting in
neuronal damage and higher AD risk (McEwen, 2002; Sapolsky, Krey,& McEwen, 1986). A
chronic stressor such as dementia caregiving may cause excessive activation of the HPA axis and
high levels of glucocorticoids, resulting in cell death in regions of the hippocampus. This damage
is hypothesized to diminish the normal inhibitory feedback of the HPA axis, resulting in ongoing
activation of this system and higher glucocorticoid levels, termed the “glucocorticoid cascade
hypothesis” advanced by Sapolsky et al (1986). This has been demonstrated in Alzheimer’s
caregivers who had significantly higher interleukin-6 (but not C-reactive protein) levels than
controls (von Känel, et al, 2006). Instrumental and emotional social supports may act as
protective, moderating factors to enhance resilience in the face of high caregiver stress burden,
however, access to such supports varies greatly (Wilks & Croom, 2008).
In gender-stratified analyses, both husband and wife dementia caregivers were at
significantly increased risk for incident dementia. However, the effect was even more
pronounced for husband caregivers who experienced an almost 12-fold increased risk, while
wives experienced an almost 4-fold increased risk. While this was contrary to our hypothesis,
there are several possible explanations for this finding. Although many husbands make
satisfactory adjustments to the dementia caregiver role, Calasanti and Bowen (2006) found that
16. Incident dementia risk among spouse dementia caregivers
caregiver husbands were more challenged than wives to adapt to the performance of daily care
tasks, perhaps because dementia caregiving involves tasks with which husbands have been
somewhat less familiar. In addition, husbands’ shrinking social networks and loss of meaningful
civic and work activities may contribute to a profound sense of isolation (Kaye & Applegate,
1994; Russell, 2007). Such isolation may in turn contribute to increased stress levels and a sense
of role captivity, the latter of which has been shown to lead to psychological distress in husband
caregivers (Levesque, Ducharme, Zarit, Lachance, & Giroux, 2008).
Despite such distress, male caregivers may be less likely than women caregivers to seek
support for their feelings of isolation and captivity. Gottman and Carrère (1994) found that men
are less likely than women to use their relationships with others to understand and process their
emotions. Likewise, Julien, Arellano, and Turgeon (1997) reported that men are more likely than
women to use diversion talk or activities as ways to cope with negative affect.
It is also possible that the decline in marital interactions, shared meanings, and overall
marital quality experienced by husband caregivers of wives with dementia (Wright, 1991; 1994)
has a more deleterious impact on husbands because of the centrality of marriage in their support
system. Relational deprivation, defined as the loss of closeness and former levels of
communication in the marriage, is associated with higher levels of psychological distress for
husband caregivers (Levesque et al). The loss of a primary support, coupled with greater
reluctance to reach out to others for emotional support, may increase men’s risk of negative
Finally, self-neglect is not uncommon among family caregivers (Ortiz, Lamdan, Johnson,
& Korbage, 2009; Portnoy, 1993) and is associated with poorer physical functioning (Dong,
17. Incident dementia risk among spouse dementia caregivers
Mendes de Leon, & Evans, 2009). Men are at greater risk for cardiovascular health conditions
(Eggers et al, 2008) and even when not in a caregiving role, are more likely to delay help seeking
when they become ill (Galdas, Cheater, & Marshall, 2005), compared to women. Thus, husband
caregivers may be at increased risk for incident dementia not only from effects of the
psychosocial stress associated with caregiving, but also from their own neglected comorbid
vascular conditions, which are in turn associated with increased risk for dementia (Hayden et al,
2006; Qiu, Ronchi, & Fratiglioni, 2007).
A limitation of this study is that we did not have data on whether these subjects were
providing non-dementia care to spouse or others. To the extent that subjects in the non-exposed
group were actually experiencing chronic stress due to non-dementia caregiving, our results
would be biased toward the null. However, given the strong effect observed for dementia
caregiving, this concern is alleviated. Dementia caregiving was presumed to begin at the onset
of dementia in the spouse and was not defined from direct questioning as to when caregiving
duties actually began, as these data were not available. This is likely a conservative estimate of
exposure to the demands of dementia caregiving, given that caregiving assistance and
compensation for one’s spouse’s declining cognitive abilities likely begins much earlier than
when clinical symptoms progress to unambiguously meet criteria for dementia. In a meta-
analysis of 47 studies, Bäckman and colleagues (2005) suggested that preclinical changes can be
detected for many years prior to AD onset, however, 3 years prior to dementia onset symptoms
of prodromal AD are more reliably seen.
Another limitation of this study was the lack of data on caregiver personality, cognitive
appraisals of the caregiving experience, or perceived stress. These indicators of subjective stress
are potentially important modifiers of the effect of chronic caregiver stress on risk for incident
18. Incident dementia risk among spouse dementia caregivers
dementia in the spouse caregiver. Further, no data were available on social supports, respite, or
whether others shared caregiving duties with the spouse, or general coping strategies used by the
spouse caregiver for dealing with stress generally. Our sample was also not large enough for us
to study the effects of widowhood after commencement of spousal dementia caregiving.
Additionally, while we were able to adjust for the effects of spousal dementia caregiving for
socioeconomic status (husband’s education and occupation), this was an imperfect adjustment
for the entire universe of shared environmental exposures common to both spouses. There may
be residual confounding to the extent that couples may have been exposed to risk factors that
increased risk for dementia in both spouses which were not entirely captured by socioeconomic
status such as a shared active or sedentary lifestyle (Lensegrav-Benson et al, 2004) or shared
habits as to use of neuroprotective agents such as antioxidant vitamin supplements or anti-
inflammatory medications (Fotuhi et al, 2008). To the extent that this is true, the effect we see
here may not be entirely due to cumulative damage due to chronic psychosocial stress in the
caregiver. Finally, it should be noted that while the effect sizes are substantial, the confidence
intervals are large, owing to the relatively small number of spouse pairs where both spouses
received a dementia diagnosis.
Although effect size estimates vary, all results reported here, whether in combined or
stratified analyses, consistently suggest that the chronic stress and work of spouse dementia
caregiving exerts deleterious effects on the caregiver’s own dementia risk. There remains a clear
need for additional work to ascertain important mediators and moderators of the effect of spouse
dementia caregiving on risk for dementia in the caregiver. Future studies of similar design with
larger samples may be able to determine whether incident dementia risk in the spouse caregiver
is reduced after widowhood or institutionalization of the spouse with dementia. More elaborate
19. Incident dementia risk among spouse dementia caregivers
analyses that examine the effect of the spouse’s rate of clinical dementia progression and
behavioral disturbances may identify specific patterns that are more stressful to spouse
caregivers, thereby helping to identify more vulnerable caregivers for interventions.
Consideration of other psychosocial stressors that may compound the effects of dementia
caregiving, whether concurrent or cumulative over the lifespan, will be important to pursue.
Studies are also needed that will compare different coping strategies of caregivers, along with
stress-related biomarkers, to determine which strategies result in less severe neuroendocrine
reactivity and are thus generally more protective against dementia risk in the caregiver. These
will help inform interventions that will not only help the caregiver maintain positive well-being
during the caregiving experience, but may even help preserve cognitive function for their
remaining years, after the work of caregiving is finished.
20. Incident dementia risk among spouse dementia caregivers
Table 1. Bivariate relationships between demographic variables and incident
samples t-test for
square test for
Subject’s Age at baseline interview
74.2 (5.8) 75.9 (5.8) p < .001
Spouse’s Age at baseline interview
74.4 (5.7) 74.7 (6.2)
p = .402
Husband’s education * in years
14.0 (3.4) 13.8 (3.3) p = .508
Exposure to spousal dementia
caregiving (N, %)
195 (8.9%) 34 (13.3%) p=.022
Subject APOE: zero ε4 allele (N, %) 1482 (68.9%) 147 (57.9%)
p < .001
Subject APOE: one ε4 allele (N, %) 625 (29.1%) 85 (33.5%)
Subject APOE: two ε4 alleles (N, %) 43 (2.0%) 22 (8.7%)
Husband Occupation *: Professional,
technical, management (N, %)
938 (42.9%) 100 (39.2%)
Husband Occupation: Clerical, sales
178 (8.1%) 20 (7.8%)
Husband Occupation: Service (N, %) 61 (2.8%) 9 (3.5%)
Husband Occupation: Agriculture
457 (20.9%) 61 (23.9%)
Husband Occupation: Machine/
miscellaneous (N, %)
553 (25.3%) 65 (25.5%)
* Husband’s education (and occupation) refers to subject’s own, if male, and spouse’s, if female
21. Incident dementia risk among spouse dementia caregivers
Table 2. Relative hazard of incident dementia from spousal caregiving, adjusted for
Pr > ChiSq Hazard
1 1.79272 0.50536 12.5842 0.0004 6.006 2.231 16.171
Female gender 1 -0.22712 0.13305 2.9140 0.0878 0.797 0.614 1.034
Age at baseline
1 0.06220 0.02486 6.2601 0.0123 1.064 1.014 1.117
one ε4 allele
1 0.37347 0.13793 7.3318 0.0068 1.453 1.109 1.904
two ε4 alleles
1 1.51376 0.23700 40.7950 <.0001 4.544 2.855 7.230
1 -0.45482 0.19274 5.5685 0.0183 0.635 0.435 0.926
1 -0.41148 0.25922 2.5197 0.1124 0.663 0.399 1.101
1 -0.02075 0.36752 0.0032 0.9550 0.979 0.477 2.013
1 -0.21729 0.18043 1.4503 0.2285 0.805 0.565 1.146
1 0.0003385 0.02394 0.0002 0.9887 1.000 0.954 1.048
* Reference category for APOE genotype was zero ε4 alleles
** Reference category for husband’s occupation was “Machine/miscellaneous”
22. Incident dementia risk among spouse dementia caregivers
ALTERNATIVE TABLE 2
Table 2. Relative hazard of incident dementia from spousal caregiving, adjusted for
covariates (overall model in total sample of 2442 subjects, then stratified by spouse gender).
Total Sample Husband as
1 6.006 2.23 - 16.17 11.79 1.65 - 84.53 3.56 1.12 – 11.29
Female gender 1 0.797 0.61 - 1.03 ------- ------- ------- -------
Age at baseline
1 1.064 1.01 - 1.12 1.02 .98 – 1.07 1.13 1.05 – 1.23
one ε4 allele
1 1.453 1.11 - 1.90 1.41 .99 – 2.01 1.44 .93 – 2.24
two ε4 alleles
1 4.544 2.86 - 7.23 4.61 2.53 – 8.40 3.76 1.65 – 8.58
1 0.635 0.44 - 0.93 .68 .41 – 1.10 .63 .34 – 1.18
1 0.663 0.40 - 1.10 .58 .29 – 1.18 .86 .41 – 1.82
1 0.979 0.48 - 2.01 1.06 .43 – 2.63 .72 .21 – 2.51
1 0.805 0.57 - 1.15 .93 .59 – 1.47 .64 .35 – 1.16
1 1.000 0.95 - 1.05 .99 .94 – 1.06 .99 .92 – 1.07
23. Incident dementia risk among spouse dementia caregivers
Albinsson L, Strang P. Differences in supporting families of dementia patients and cancer
patients: a palliative perspective. Palliative Medicine 2003 Jun;17(4):359-67.
Bäckman, L., S. Jones, A.K. Berger, E.J. Laukka, and B.J. Small. Cognitive impairment in
preclinical Alzheimer's disease: a meta-analysis. Neuropsychology, 2005. 19(4): p. 520-
Bacon E, Milne DL, Sheikh AI, Freeston MH. Positive experiences in caregivers: an exploratory
case series. Behav Cogn Psychother. 2009 Jan;37(1):95-114.
Bergman-Evan B. A health profile of spousal Alzheimer’s caregivers. Journal of Psychosocial
Nursing 1994; 32(9): 25-30.
Boss P, Caron W, Horbal J, Mortimer J. Predictors of depression in caregivers of dementia
patients: Boundary ambiguity and mastery. Family Processes 1990; 29: 245-254.
Breitner JC, Wyse BW, Anthony JC, Welsh-Bohmer KA, Steffens DC, Norton MC, Tschanz JT,
Plassman BL, Meyer MR, Skoog I, Khachaturian A. APOE-epsilon4 count predicts age
when prevalence of AD increases, then declines: the Cache County Study. Neurology.
Calasanti, T., Bowen, M. E. Spousal caregiving and crossing gender boundaries: Maintaining
gendered identities. Journal of Aging Studies. 2006: 20: 253-263.
Caswell LW, Vitaliano PP, Croyle KL, Scanlan JM, Zhang J, Daruwala A. Negative
associations of chronic stress and cognitive performance in older adult spouse caregivers.
Experimental Aging Research 2003; 29: 303-318.
Christakis NA, Allison PD. Mortality after the hospitalization of a spouse. New England Journal
of Medicine 2006 Feb 16;354(7):719-30.
Cleves M, Gould WW, Gutierrez RG, Marchenko Y. An Introduction to Survival Analysis
Using Stata, 2nd Edition. Stata Press. 2008.
Cox DR & D Oakes D. (1984) Analysis of survival data (Chapman & Hall).
Croog SH, Sudilovsky A, Burleson JA, Baume RM. Vulnerability of husband and wife
caregivers of Alzheimer disease patients to caregiving stressors. Alzheimer’s Disease and
Associated Disorders 2001 Oct-Dec;15(4):201-10.
de Vugt, M.E., J. Jolles, L. van Osch, F. Stevens, P. Aalten, R. Lousberg, and F.R. Verhey,
Cognitive functioning in spousal caregivers of dementia patients: findings from the
prospective MAASBED study. Age and Ageing, 2006. 35(2): p. 160-6 16495293.
24. Incident dementia risk among spouse dementia caregivers
Diagnostic and Statistical Manual of Mental Disorders. Third Edition-Revised. 3 ed.
Washington, D.C.: American Psychiatric Association; 1987.
Dong X, Mendes de Leon CF, Evans DA. Is greater self-neglect severity associated with lower
levels of physical function? Journal of Aging and Health; Vol. 21(4) June 2009 596-610.
Eggers KM, Lind L, Ahlström H, Bjerner T, Ebeling Barbier C, Larsson A, Venge P, Lindahl B.
Prevalence and pathophysiological mechanisms of elevated cardiac troponin I levels in a
population-based sample of elderly subjects. European Heart Journal. 2008
Fotenos AF; Mintun MA; Snyder AZ; Morris JC; Buckner RL. Brain volume decline in aging:
evidence for a relation between socioeconomic status, preclinical Alzheimer disease, and
reserve. Archives of Neurology 2008 Jan; Vol. 65 (1), pp. 113-20.
Fotuhi M, Zandi PP, Hayden KM, Khachaturian AS, Szekely CA, Wengreen H, Munger RG,
Norton MC, Tschanz JT, Lyketsos CG, Breitner JC, Welsh-Bohmer K. Better
cognitive performance in elderly taking antioxidant vitamins E and C supplements in
combination with nonsteroidal anti-inflammatory drugs: the Cache County Study.
Alzheimers Dementia 2008 May;4(3):223-7.
Galdas PM, Cheater F, Marshall P. Men and health help-seeking behaviour: literature
review. Journal of Advanced Nursing. 2005 Mar;49(6):616-23.
Goldbourt U; Schnaider-Beeri M; Davidson M. Socioeconomic status in relationship to death of
vascular disease and late-life dementia. Journal of the Neurological Sciences 2007 Jun
15; Vol. 257 (1-2), pp. 177-81.
Gottman, J. M., Carrère, S. Why can’t men and women get along? In D.J. Canary & L. Stafford
(Eds.). Communication and relational maintenance 1994: 203-229. San Diego, CA:
Hayden, K.M., Zandi, P.P., Lyketsos, C.G., Khachaturian, A.S., Bastian, L.A., Charoonruk,
G.*, Tschanz, J.T., Norton, M.C., Pieper, C.F., Munger, R.G., Breitner, J.C.S., & Welsh-
Bohmer, K.A. (2006). Vascular risk factors for incident alzheimer's disease and vascular
dementia: The Cache County Study. Alzheimer Disease & Associated Disorders, 20, 93-
Jorm AF. The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): a
review. International Psychogeriatrics 2004;16: 275-293.
Jorm AF, Jacomb PA. The Informant Questionnaire on Cognitive Decline in the Elderly
(IQCODE): socio-demographic correlates, reliability, validity and some norms.
Psychological Medicine 1989;19:1015-22.
25. Incident dementia risk among spouse dementia caregivers
Julien, D., Arellano, C., & Turgeon, L. Gender issues in heterosexual, gay, and lesbian couples.
In K. Halford & H.J. Markman (Eds.). Clinical handbook of marriage and marital
intervention 1997: 107-127). New York: Wiley.
Kalapatapu RK, Neugroschl JA. Update on neuropsychiatric symptoms of dementia: evaluation
and management. Geriatrics. 2009 Apr;64(4):20-6.
Kawas C, Segal J, Stewart W, Corrada M, Thal L. A validation study of the dementia
questionnaire. Archives of Neurology 1994;51:901-906.
Kaye, L. W., Applegate, J. S. Older men and the family caregiving orientation. In E. H.
Thompson Jr. (Ed.). Older men’s lives. 1994: 218-236. Thousand Oaks, CA: Sage
Kelley, H. H., & Thibaut, J. W (1978). Interpersonal relations: A theory of interdependence.
New York: Wiley.
Lee S, Kawachi I, Grodstein F. Does caregivig stress affect cognitive function in older women?
Journal of Nervous and Mental Disorders 2004; 192: 51-57.
Lensegrav-Benson, T., Tschanz, J., Masters, K.S., Carlson, M.C., Corcoran, C., Lyketsos, C.,
Health, E., Leslie, C., Munger, R.A., Ostbye, T., Welsh-Bohmer, K., Norton, M., &
Hayden, K. Sedentary Lifestyle Increases Dementia Risk: The Cache County Study.
Presentated at the Annual Meeting of the American Psychological Association, Honolulu,
Hawaii, July, 2004.
Levesque, L., Ducharme, F., Zarit, S. H., Lachance, L., Giroux, F. Predicting longitudinal
patterns of psychological distress in older husband caregivers: Further analysis of
existing data. Aging & Mental Health, 2008: 12(3), 333-342.
Lewis, M. A., McBride, C. M., Pollak, K. I., Puleo, E., Butterfield, R. M., & Emmons, K. M
(2006). Understanding health behavior change among couples: an interdependence and
communal coping approach. Social Science & Medicine 62(6), 1369-1380.
Luijendijk HJ, van den Berg JF, Dekker MJ, van Tuijl HR, Otte W, Smit F, Hofman A, Stricker
BH, Tiemeier H. Incidence and recurrence of late-life depression. Arch Gen Psychiatry.
Mace NL, Rabins PV. The 36-Hour Day: A Family Guide to Caring for People with Alzheimer
Disease, Other Dementias, and Memory Loss in Later Life, 4th Edition. 2006. The Johns
Hopkins University Press: Baltimore, MD.
McEwen BS. Sex, stress and the hippocampus: allostasis, allostatic load and the aging process.
Neurobiol Aging. Sep-Oct 2002;23(5):921-939.
26. Incident dementia risk among spouse dementia caregivers
Miech RA, Breitner JC, Zandi PP, Khachaturian AS, Anthony JC, Mayer L. Incidence of AD
may decline in the early 90s for men, later for women: The Cache County study.
Neurology. 2002 Jan 22;58(2):209-18.
Noone JH, Stephens C. Men, masculine identities, and health care utilization. Sociology of
Health and Illness 2008 Jul;30(5):711-25.
Ortiz N, Lamdan R, Johnson S, Korbage A. Caregiver status: a potential risk factor for extreme
self-neglect. Psychosomatics. 2009 Mar-Apr;50(2):166-8.
Ory MG, Hoffman RR, Yee JL, Tennstedt S, Schulz R. Prevalence and impact of caregiving: a
detailed comparison between dementia and non-dementia caregivers. Gerontologist 1999;
Pinquart M, Sörensen S. Differences between caregivers and noncaregivers in psychological
health and physical health: A meta-analysis. Psychology and Aging 2003; 18: 250-267.
Portnoy D. Are you caring or caretaking? American Journal of Hospice and Palliative Care.
Qiu C, Ronchi D, Fratiglioni L. The epidemiology of the dementias : an update. Current Opinions
in Psychiatry. 2007 Jul;20(4):380-5.
Richards B, Skoletsky J, Shuber AP, et al. Multiplex PCR amplification from the CFTR gene
using DNA prepared from buccal brushes/swabs. Human Molecular Genetics 1993;
Russell, R. The work of elderly men caregivers: From public careers to an unseen world. Men
and Masculinities, 2007: 9(3), 298-314.
Sapolsky RM, Krey LC, McEwen BS. The neuroendocrinology of stress and aging: the
glucocorticoid cascade hypothesis. Endocr Rev. Aug 1986;7(3):284-301.
Saunders AM, Strittmatter WJ, Schmechel D, et al. Association of apolipoprotein E allele E4
with late-onset familial and sporadic Alzheimer's disease. Neurology 1993;43:1467-1472.
Schulz R, Beach S. Caregiving as a risk factor for mortality: The Caregiver Health Effects
Study. Journal of the American Medical Association 1999; 282: 2215-2219.
Schulz R, Visintainer P,Williamson GM. Psychiatric and physical morbidity effects of
caregiving. Journal of Gerontology 1990; 45:181-191.
Sörensen S, Duberstein P, Gill D, Pinquart M. Dementia care: mental health effects, intervention
strategies, and clinical implications. Lancet Neuroogy. 2006 Nov;5(11):961-73.
Teng EL, Chui HC. The Modified Mini-Mental State (3MS) examination. Journal of Clinical
Psychiatry 1987 Aug;48(8):314-8.
27. Incident dementia risk among spouse dementia caregivers
Tschanz JT, Welsh-Bohmer KA, Plassman BL, Norton MC, Wyse BW, Breitner JC. An
adaptation of the modified mini-mental state examination: analysis of demographic
influences and normative data: The Cache County Study. Neuropsychiatry
Neuropsychology and Behavioral Neurology 2002;15(1):28-38.
U.S. Department of Labor. (1991). Dictionary of Occupational Titles (4TH
Washington, D.C.: U.S. Government Printing Office.
Vitaliano PP, Echeverria D, Yi J, Phillips PEM, Young J, Siegler IC. Psychophysiological
mediators of caregiver stress and differential cognitive decline. Psychology and Aging
Vitaliano PP, Zhang J, Scanlan JM. Is caregiving hazardous to one’s physical health? A meta-
analysis. Psychological Bulletin 2003; 129:946-972.
von Känel R, Dimsdale JE, Mills PJ, Ancoli-Israel S, Patterson TL, Mausbach BT, Grant I.
Effect of Alzheimer caregiving stress and age on frailty markers interleukin-6, C-reactive
protein, and D-dimer. J Gerontol A Biol Sci Med Sci. 2006 Sep;61(9):963-9.
Wilks SE, Croom B. Perceived stress and resilience in Alzheimer's disease caregivers: testing
moderation and mediation models of social support. Aging Ment Health. 2008
Wright, L. K. The impact of Alzheimer’s disease on the marital relationship. The Gerontologist, 1991:
Wright, L. K. Alzheimer’s disease afflicted spouses who remain at home: Can human dialectics explain
the findings? Social Science & Medicine, 1994: 38(8), 1037-1046.