Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample
(USC Thesis Other)
History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type o f computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely afreet reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back o f the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zed) Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
HISTORY OF DEPRESSION, ANTIDEPRESSANT TREATMENT,
AND OTHER PSYCHIATRIC ILLNESS
AS RISK FACTORS FOR ALZHEIMER'S DISEASE
IN A TWIN SAMPLE
Julie Loebach Wetherell
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
Master of Arts
(Psychology)
December 1996
® 1996 Julie Loebach Wetherell
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 1383550
UMI Microform 1383550
Copyright 1997, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 8 0 0 0 7
This thesis, written by
Julie Loebach Wetherell
under the direction of k.er.— Thesis Committee,
and approved by all its members, has been pre
sented to and accepted by the Dean of The
Graduate School, in partial fulfillm ent of the
requirements for the degree of
Master of Arts
...
THESIS COMMITTEE
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
The Study of Dementia in Swedish Twins, funded by NIA Grant No.
R01 AG08724, Margaret Gatz, Principal Investigator, is an ongoing study
conducted in Stockholm at The Karolinska Institute by the Division of Genetic
Epidemiology, Institute of Environmental Medicine, and the Department of
Geriatric Medicine, and by the Institute of Gerontology in Jonkoping, Sweden,
in collaboration with the University of Southern California. The OCTO-Twin
Study is funded by NIA Grant No. R01 AG08861, Gerald McCleam, Principal
Investigator, and is conducted in Jonkoping, Sweden, at the Institute of
Gerontology, in collaboration with the Center for Developmental and Health
Genetics at The Pennsylvania State University.
The author also appreciates the personal and professional support of
Margaret Gatz, Joseph Loebach Wetherell, and the members of the Section on
Clinical Research in Aging and Psychology.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ..................................................................................ii
LIST OF TABLES................................................................................................iv
ABSTRACT...........................................................................................................v
INTRODUCTION.............................................................................................. 1
METHODS ...................................................................................................... 15
Sample................................................................................................... 15
Instruments and Data Collection ........................................................ 20
Statistical Analyses............................................................................... 21
RESULTS ........................................................................................................ 23
DISCUSSION................................................................................................... 35
REFERENCES................................................................................................ 46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table 1. Discordance Information................................................................. 19
Table 2. Demographic Information............................................................... 25
Table 3. History of Depressive Hlness, Antidepressant or ECT Exposure, and
History of Other Psychiatric Illness as Risk Factors for Alzheimer's
Disease .............................................................................. 27
Table 4. Results by Subgroup ....................................................................... 29
Table 5. Liberal Inclusion Criteria (Partner Alive and Cognitively Intact at
Time of Proband's Dementia O nset).................................................. 32
Table 6. Stringent Inclusion Criteria (Concordant After 7 Years)................ 33
Table 7. Discordant Only Inclusion C rite ria ................................................ 34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
V
ABSTRACT
Although case-control studies have found elevated risk for Alzheimer's
disease (AD) associated with a prior history of depression, none of the previous
research has adequately controlled for familial risk factors. Putative psychiatric
risk factors including history of depression, other psychiatric illness, and
treatment with antidepressant medications or electroconvulsive therapy (ECT)
were evaluated for a registry-based sample of 52 twin pairs discordant for AD.
Odds ratios (OR) were calculated for depressive illness at any time and for
episodes more than 10 years prior to dementia onset. A history of depression
was associated with elevated risk (OR = 4.0, 0.73-21.9). When analysis was
restricted to individuals whose depressive episodes commenced more than 10
years prior to dementia onset, the magnitude of the odds ratio decreased
markedly (OR = 2.0, 0.28-14.4). Exposure to antidepressants or ECT was not
associated with increased risk of AD in this sample (OR = 1.0, 0.13-7.77).
While these findings are not statistically significant due to the small number of
pairs discordant for both disease and exposure in this sample, they suggest that
a history of depression may be associated with an elevated risk for AD. In
particular, these results are consistent with an interpretation of depression as a
prodromal symptom of dementia.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODUCTION
Alzheimer's disease (AD) is the most common form of dementia in older
adults. With no currently known cure, it affects approximately five percent of
those over the age of 65 and 20 percent of those over 80. Applying this
prevalence rate to the current U. S. population, AD currently affects over 1.5
million older Americans; by the year 2040, this number is expected to double
(Office of Technology Assessment, 1987). In addition to the terrible human
costs inflicted on its victims, AD is of great concern to the medical and public
health community. At least half of all nursing home residents are believed to
have AD, and the cost of caring for these patients is estimated at $50 billion
annually (Office of Technology Assessment, 1988). Discovering risk factors
for the disorder could aid in prevention efforts, which are especially important
in the absence of a cure. For example, if particular exposures are associated
with elevated AD risk, public health efforts directed towards limiting these
exposures or reducing the associated risk might reduce the incidence of the
disease (Mortimer, 1995).
The status of depression as a risk factor for AD remains controversial.
It is generally recognized that comorbid depressive symptoms are a common
feature of early stage AD (Migliorelli et al., 1995; Teri & Wagner, 1992).
Additionally, depression is believed to be a prodromal symptom, often noted
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
several years before cognitive impairment becomes significant enough to merit a
dementia diagnosis. Two prospective studies found that depressed,
nondemented older adults were significantly more likely to become demented
within three years than nondepressed respondents (Minami et al., 1995; Reding,
Haycox, & Blass, 1985).
Affective symptoms may represent a reaction to cognitive decline or an
endogenous feature of the dementia process. For example, Liston, Jarvik, and
Gerson (1987) propose that Alzheimer's-related neuronal degeneration in
ascending noradrenergic and cholinergic cortical pathways may lead to the
development of depression. Alternatively, individuals may perceive and react
to cognitive deficits before they are significant enough to be detected by
neuropsychological assessment; personality or other variables may leave certain
groups of people especially vulnerable to reactive depression associated with
dementia.
Many epidemiological investigations, including a number of case-control
studies, have noted a relationship between depression and subsequent onset of
AD. Case-control studies are among the most widely used epidemiological
techniques for eliciting information about risk factors for disease. The
methodology involves comparing the prevalence of exposure to presumed risk
factors in a group of people with the disease under investigation to the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
prevalence of these risk factors in a matched group free from disease. Case-
control studies are more efficient than prospective cohort studies, especially for
relatively rare diseases such as AD, because the investigators do not need to
follow up prohibitively large numbers of unaffected subjects in order to achieve
a sufficiently large sample for analysis.
Recently, the EURODEM team of researchers combined and reanalyzed
data from 11 case-control studies of risk factors for AD conducted in the United
States, Europe, Japan, and Australia. Most were hospital- or clinic-based as
opposed to population-based. Six of the studies included data on history of
depression; the pooled analysis (total N, cases = 743; total N, controls = 818)
found a significantly elevated risk for late onset AD associated with prior
medically treated depression (Jorm et al., 1991).
Because Alzheimer's is a disease of insidious onset, distinguishing risk
factors from early symptoms is especially difficult. In a prospective,
longitudinal study, La Rue and Jarvik (1987) found evidence of lower cognitive
performance in subjects 20 years before dementia diagnosis. They propose that
the disease process may commence decades before an individual meets criteria
for dementia. An alternative explanation is suggested by a study based on
writing samples of young women who joined an order of nuns (Snowdon et al.,
1996). Idea density in the writing samples was strongly predictive of cognitive
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
function, AD diagnoses, and neuropathology an average of 58 years later.
These findings seem to support a threshold model in which dementia is
diagnosed when an individual's cognitive reserve capacity, related to the amount
of functional brain tissue, falls below a threshold at which normal cognitive
function cannot be sustained (Mortimer, 1994).
Cognitive reserve is assumed to be a function of multiple factors,
including age, early cognitive capacity, exposure to environmental toxins and
traumas, and genetic risk factors for AD and other dementias. One implication
of the threshold model is that a number of accumulated exposures over the life
span, each associated with a slightly elevated level of risk, may lower the age of
onset of dementia by reducing brain reserve capacity. A second implication of
the threshold model is that selection of any time point prior to dementia onset is
essentially arbitrary ; yet it remains important to distinguish those factors which
play a role in eroding cognitive reserve from factors which are merely
symptomatic of cognitive decline.
The EURODEM pooled analysis examined the timing of depressive
episodes under the assumption that temporally distant episodes are less likely to
represent an early manifestation of dementia. Four of the studies included in
the analysis (Broe et al., 1990; French et al., 1985; Kokmen et al., 1991;
Shalat, Selzer, Pidcock, & Baker, 1987) provided data on age of onset of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
depression. For those individuals with a history of depression more than ten
years prior to developing late onset AD (after age 65), the relative risk was 2.01
(95% confidence intervals, 1.06-3.80), indicating that individuals with a history
of depression were twice as likely to develop Alzheimer's as individuals who
had never experienced depression. For those individuals with a more recent
history of depression, the relative risk was 4.46 (1.24-16.02). The pooled
analysis found no significantly elevated risk for early onset AD (before age 65)
associated with depression.
These findings suggest that depression is associated with an elevated risk
of developing AD in late life. Furthermore, the fact that higher relative risk
was associated with depressive episodes more proximal to dementia onset
suggests that depression is often an early symptom of dementia, in many cases
appearing before detectable cognitive symptoms are severe enough to meet
criteria for dementia. If the relative risk had been higher for earlier depressive
episodes than for later episodes, evidence would be stronger for a risk factor
model, because the mechanism by which depression leads to dementia would
have the opportunity to work its damage over a longer period.
A subsequent large, population-based case-control study of risk factors
for AD, the Canadian Study of Health and Aging (1994), did not find a
relationship between any prior depressive episodes and increased risk of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dementia (odds ratio estimate of relative risk = 1.05, 0.63-1.77). This study
included only those patients with probable AD (according to NINCDS-ADRDA
criteria) whose symptoms first appeared within three years of study diagnosis
(N, cases = 258; N, controls = 535). Risk factor questionnaires were
completed by proxy informants for both cases and controls. Although this study
did not attempt to analyze early and late onset cases separately, the authors
noted that the average age of the sample was significantly older than samples in
other case-control studies, suggesting that the bulk of their subjects were late
onset.
Recently, Speck et al. (1995) conducted a case-control study in an
HMO-based Alzheimer's registry. Only incident cases meeting NINCDS-
ADRDA criteria for probable AD were included (N, cases = 294; N, controls
= 300); exposure data were collected by interview from proxies for both cases
and controls. The interview elicited information about depression severity and
treatment as well as age of onset and whether the depression began after a
significant loss. After adjusting for age, gender, education, and type of
informant, the researchers found a significantly elevated risk of AD associated
with depressive episodes ten or more years prior to dementia onset (odds ratio
= 1.7, 1.0-2.9). When analysis was restricted to those whose depressions were
not related to loss or grief, the magnitude of association increased (odds ratio =
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
2.4, 1.2-4.5). Interestingly, depression less than ten years before dementia
onset was not associated with statistically greater risk (odds ratio = 1.1, 0.5-
2.3). Furthermore, when analyses were restricted to just those individuals with
treated depression or with depression severe enough to limit their activities, the
results were no longer significant.
Finally, researchers found an association between cognitive impairment
(exclusive of dementia) and current depression only for those without a prior
history of depression in a large, community-based sample (Van Ojen et al.,
1995). Dementia rates did not differ between depressed patients with late and
early depression onset. Although this study did not directly address the issue of
risk factors or early symptoms of Alzheimer's disease, it provides further
support for the notion that late but not early onset of depression may be related
to cognitive decline.
The reasons for these conflicting findings are unclear. The failure of
Speck's team (1995) and the Canadian study (1994) to find a significant
relationship between late onset depression and AD is surprising in light of
empirical evidence and clinical lore. In the EURODEM analysis, unlike the
Canadian study, only medically-treated depression was examined. Thus it is
possible that depressive episodes were more severe in the EURODEM samples.
Although Speck et al. (1995) found no significantly elevated risk for individuals
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
with more severe depression, they measured severity by reported effect on
activity level. This may be an inadequate measure of depression severity;
individuals with relatively mild depressions may still experience a significant
reduction in daily activities.
If depression is not a prodromal symptom of dementia, it is of interest to
identify a mechanism through which it would function as a risk factor. The
most commonly proposed explanation for depression as a risk factor for AD is
an iatrogenic process. Antidepressants alter various neurotransmitter systems,
many of which are involved in AD. In particular, tricyclic antidepressants have
anticholinergic effects, and cholinergic levels are lower in AD patients (Liston
et al., 1987). Electroconvulsive therapy (ECT) may also cause memory deficits
in older adults, although some evidence suggests that these effects are mild and
transient (Lipman, Brown, Silbert, Rains, & Grady, 1993; Rubin, Kinscherf,
Figiel, & Zorumski, 1993). If exposure to anticholinergic agents can have
long-term effects on cholinergic production, release, binding, or reuptake, such
exposure may be a factor in moving an individual closer to the dementia
threshold. Likewise, ECT-related memory impairment may contribute to a
dementing process. Hence treatment, and not depression per se, may increase
risk of subsequent dementia. This hypothesis, however, is not supported by
Speck et al. (1995), who found no significantly elevated risk in those
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
individuals with a history of treated depression. Moreover, the EURODEM
authors examined the impact of antidepressant exposure by combining data from
the two studies that collected data on antidepressant treatment (Broe et al.,
1990; French et al., 1985). They found no elevated risk for those patients who
had been treated with antidepressants (odds ratio = 1.20, 0.52-2.78; Jorm et
al., 1991).
Five case-control studies have explored whether psychiatric illness other
than depression is associated with increased risk of AD. Barclay, Kheyfets,
Zemcov, Blass, and McDowell (1986) found no significantly elevated risk
associated with psychiatric disorders other than depression and alcoholism in a
sample of 259 cases and 36 controls presenting to an outpatient dementia clinic.
Unfortunately, this study had low power to detect significant risks due to the
small number of controls, and the recruitment strategy did not ensure that all
control subjects were free of incipient cognitive impairment. Broe et al. (1990)
found no elevated risk for "nervous breakdown" in a hospital-based study of
170 cases and 170 matched controls. Heyman et al. (1984) calculated an odds
ratio of 2.38 for physician-treated psychiatric disorders, including depression,
in a sample of 40 early onset (before age 70) Alzheimer's patients and 80
matched community controls. This finding was not statistically significant,
presumably due to the small sample size; moreover, the authors were unable to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
determine whether psychiatric symptoms were early manifestations of dementia
or separate clinical entities representing possible risk factors for AD. A brain
bank case-control study compared 407 autopsy-confirmed AD patients to 100
patients with autopsy-confirmed non-Alzheimer dementias and 50 normal
elderly brains. This study did not find an elevated risk of AD associated with
prior psychiatric illness or electroconvulsive therapy (Mendez et al., 1992).
Finally, in a large population-based study conducted through medical records
linkage in Rochester, Minnesota, Kokmen et al. (1991) reported odds ratios for
mania, psychosis, and personality disorder in addition to depression. While
results were not significant for mania or psychosis, personality disorders were
associated with significantly elevated risk for AD. The authors proposed no
explanation for this finding.
In general, failure of the smaller studies to achieve statistically
significant results is probably due to the low base rates of reported psychiatric
disorders. Typically, large samples are required to find significant risk
associated with rare exposures, even when the magnitude of the risk is
substantial (Schlesselman, 1982). The selection criteria for other psychiatric
disorders is unclear in several of the studies; nervous breakdown is not a DSM-
m -R diagnosis, and Kokmen's study appears to have overlooked common
diagnostic categories such as anxiety disorders. Furthermore, since Kokmen’s
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
team relied exclusively on medical records, cases of psychological disorders that
were not medically treated may have been overlooked. Finally, brain bank
samples may not be representative of dementia cases in the population because
individuals who volunteer to donate their remains may differ systematically
from typical patients. Because the findings have varied across studies and
failure to find significant results in some cases may have been influenced by
small sample sizes, the question of whether psychiatric disorders other than
depression are associated with elevated risk for AD merits further investigation.
Establishing the specificity of the depression-dementia link is necessary prior to
isolating the factors associated with depression that may increase risk for
subsequent expression of AD.
Thus the question of whether depression or other psychiatric disorders
are associated with increased risk of AD remains puzzling. Although a further
analysis of EURODEM data by van Duijn and colleagues (1994) suggests that a
history of depression increases the risk for Alzheimer's independent of family
history of dementia, matching for family history is imperfect in community
case-control samples that rely on retrospective reports of dementia in relatives.
Furthermore, the process of choosing controls in community or hospital samples
allows the possibility that the control group differs substantially from the cases
with respect to exposure to other known risk factors, biasing the results.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Because in a twin case-control design, each case is matched with a
healthy twin partner, twin studies of AD control for two well-established risk
factors which are identical in twins, age and family history of dementia
(Breitner et al., 1993). They also control for other nonspecific genetic and
environmental factors better than case-control studies in samples of unrelated
individuals. For diseases like Alzheimer's in which multiple risk factors are
assumed to contribute in reducing brain reserve capacity, greater statistical
power is needed to identify risk factors, each one of which may only slightly
increase risk independent of all others. Twin studies, by virtue of their close
match between cases and controls, provide greater statistical power to detect
such risks. The data presented here represent the first exploration of
depression, antidepressant treatment, and other psychiatric illness as risk factors
for Alzheimer's disease in a twin sample.
The twin sample used in this study is drawn from the Swedish Twin
Registry, the largest and most well-established twin registry in Scandinavia
(Medlund, Cederlof, Floderus-Myrhed, Friberg, & Sorenson, 1977). One
advantage of registry samples is that they are population-based. Hospital- or
clinic-based samples often used in case-control studies, by contrast, are not
representative of all dementia patients because only those individuals brought to
the attention of the health care provider are included as cases. Volunteer twin
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
samples are also typically not representative of their population; they tend to be
disproportionately monozygotic (identical) and female (Breitner et al., 1993).
Population-based samples, by contrast, yield the most representative sampling
of prevalent cases. Furthermore, record-keeping in Scandinavia is systematic
and easily accessible to researchers. Because health care in Sweden is
nationalized, it is easier to track sources of medical care for the sample under
investigation. This is especially important when studying medical-related risk
factors such as history of psychiatric illness or exposure to antidepressant
medications.
Questions about risk associated with depression, other psychiatric
disorders, and their treatment have important clinical implications. Depression
is one of the most common psychological disorders, with an estimated lifetime
prevalence rate of approximately 17% (Kessler et al., 1994). Because it is so
common, uncovering potential long-term sequelae such as elevated risk for late
life dementia is of great clinical significance. If it were demonstrated that the
relationship between depression and increased risk of developing AD is due to
anticholinergic effects of certain antidepressant medications, one primary
prevention strategy could be to discourage the prescription of tricyclic
antidepressants to individuals who are already at elevated risk of Alzheimer's
due to a family history of the disorder or the presence of a known risk factor.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
If depressive illness itself causes structural or biochemical changes that increase
the risk for AD, prevention efforts might be profitably directed toward early
detection and treatment of depression.
Alternatively, the same factors may predispose an individual to mental
disorders generally, with a lower threshold for psychiatric disorders than for
dementia. Thus depression and other psychological disorders could serve as a
marker for an underlying anatomical or biochemical process that may ultimately
lead to dementia. In the event that an effective preventive intervention for AD
is developed, such an intervention might be applied prophylactically to
individuals with psychological disorders in the hopes of delaying the onset of
dementia.
Even if psychiatric disorders do not constitute a risk factor for AD,
further exploration of depression as an early manifestation of dementia would be
valuable. The results from Speck et al. (1995) and the Canadian Study of
Health and Aging (1994) call into question the observation that depression often
heralds dementia onset. Confirmation or disconfirmation of these findings
would provide further insight into the course and progress of early stage AD.
Additional research could then be directed toward identifying subgroups of
patients for whom depression is an early symptom of dementia and exploring
the mechanism for the relationship in these individuals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
METHODS
Sample
The sample is primarily taken from the Study of Dementia in Swedish
Twins (Gatz, 1993; Pedersen, Gatz, Winblad, Pearson, & Berg, 1989), a
subsample of the Swedish Twin Registry, which represents a population-based
sample of more than 25,000 same-sex twin pairs born between 1886 and 1985.
Since 1984, a subsample of the Swedish Twin Registry called the Swedish
Adoption/Twin Study of Aging (SATSA) have been participating in a
longitudinal study of normal aging, including mailed questionnaires and in-
person assessments (Pedersen et al., 1991). The SATSA sample consists of
twin pairs who were reared apart and pairs reared together matched for birth
year and county (Pedersen et al., 1989).
All SATSA participants who were bom before 1935 and alive in 1985
(N = 1,978) were screened for dementia, using either the Mini-Mental State
Examination (Folstein, Folstein, & McHugh, 1975) or a telephone screening
protocol (Gatz et al., 1995). Individuals who screened positively were given a
complete clinical work-up, including physical and neurological evaluation,
neuropsychological assessment, laboratory tests, and neuroimaging (Morris et
al., 1989). Diagnoses were assigned at a multi-disciplinary consensus
conference, following DSM-HI-R criteria for dementia (American Psychiatric
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Association, 1987) and NINCDS/ADRDA criteria for Alzheimer's disease
(McKhann et al., 1984). Partners were administered identical diagnostic
batteries. Cases and partners are being followed longitudinally at 18-month
intervals and diagnoses refined based on clinical course. Autopsy is available
thus far for 10 individuals, with all cases diagnosed as having AD confirmed by
neuropathological findings.
The present study includes only those pairs in which one partner meets
NINCDS/ADRDA criteria for possible or probable Alzheimer's disease and the
other partner is cognitively intact. From 55 twin pairs in the Study of Dementia
in Swedish Twins in which one or both had AD, 37 pairs met these criteria. To
increase the sample size, the sample was supplemented by 15 discordant pairs
from the OCTO-Twin study, another subsample of the Swedish Twin Registry
(Berg et al., 1992). OCTO-Twin includes all intact pairs in the Swedish Twin
Registry who were 80 years of age or older in 1991; OCTO-Twin-dementia
includes those OCTO-Twin pairs in which one or both partners is demented.
The protocol for clinical workup and diagnosis in OCTO-Twin-dementia is
analogous to the Swedish Twin Study of Dementia, except that nonresponders to
the initial questionnaire did not receive follow-up contacts. Since the
prevalence of dementia is typically higher in survey nonresponders, dementia
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
sampling in OCTO-Twin-dementia is probably less complete than in the
Swedish Twin Study of Dementia.
Across both samples, to be counted as discordant, the partner must have
survived past the proband's age of onset by at least three years and died intact
(10 pairs), or the partner must be alive and intact at least three years after the
proband's age of onset (31 pairs), or a pair who have subsequently become
concordant must have remained discordant for over three years (11 pairs). The
inclusion of pairs that eventually became concordant for Alzheimer's is based
on the assumption that the factors leading an individual to develop dementia are
the same factors influencing age of onset. This assumption is consistent with
the threshold model of dementia. The three year criterion was employed in a
previous twin study of AD (Breitner et al., 1994).
To compensate for the difficulty of ascertaining age of dementia onset,
this study uses conservative criteria to calculate length of discordance. The
diagnostic team often assigned an onset range rather than an exact age of
dementia onset. The period of discordance is calculated based on the latest
possible age of onset for the proband and, in pairs which eventually became
concordant, on the earliest possible age of onset for the partner. This
effectively underestimates the length of discordance. For example, if the
proband's age of onset is estimated to range between 65 and 70, the partner
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
must have remained alive and cognitively intact until at least age 73.
Technically, discordance for such a pair would be calculated as three years,
although the actual period could be as long as eight years or more. In fact, only
six of 52 pairs failed to exceed the minimum standard of three years,
conservatively calculated, and only one of those pairs eventually became
concordant for dementia.
Table 1 displays discordance information. For the 31 pairs in which the
control twin is alive and intact, the average period of discordance is 9.7 years;
for the 10 pairs in which the partner died intact, the average period of
discordance is 5.9 years. For the 11 pairs in which the partner became
demented after proband's onset, the average lag is 7.7 years. Arguably, these
periods are clinically significant, especially in a sample with an average age of
84.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 1. Discordance Information
Mean SD
Number of years discordant
Parmer alive and intact (31 pairs) 9.7 5.3
Parmer died intact (10 pairs) 5.9 5.5
Parmer became demented after
proband's onset (11 pairs) 7.7 4.2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Instruments and Data Collection
Information on putative risk factors, including history of depression and
other psychiatric illness and exposure to antidepressants or electroconvulsive
therapy, was elicited via interviews with unaffected twins and family informants
as well as from medical records and linkage to the national psychiatric registry.
Psychotropic medication use for disorders other than depression and exposure to
other anticholinergic agents were not investigated in this study. For both
probands and partners, individuals whose first lifetime episode of psychiatric
illness, ECT, or antidepressant use arose within a year of the proband's earliest
possible dementia onset or thereafter were treated as not having a history of
psychiatric disorder in these analyses. Because sufficient data on psychiatric
illness or treatment were not available for three individuals from two twin pairs,
all calculations were based on a final sample of 50 pairs, 36 from the Study of
Dementia in Swedish Twins and 14 from OCTO-Twin-dementia.
Data were collected in Sweden by a team from the Karolinska Institute's
Department of Environmental Medicine in Stockholm. The data were reviewed,
coded, and entered into computerized databases, copies of which were
distributed to coinvestigators at the University of Southern California. Data
were analyzed using SAS for Windows, release 6.11.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 1
Statistical Analyses
Relative risk is calculated as the prevalence of a disease in a group of
individuals exposed to a risk factor, divided by the prevalence of the disease in
an unexposed group. For rare diseases such as Alzheimer's, the relative risk
can be approximated by the odds ratio, calculated as the product of the number
of individuals with both disease and exposure history and the number with
neither disease nor exposure, divided by the product of those with disease but
no exposure and those with exposure but no disease (Schlesselman, 1982).
The odds ratio estimates of relative risk for a matched pair sample,
based on pairs discordant for both disease and exposure, were calculated to test
the hypotheses that prior episodes of depression or other psychiatric disorders
or antidepressant exposure are associated with elevated risk for AD (Fleiss,
1981). Odds ratios were calculated as the number of discordant pairs in which
the proband has a history of depression and the partner does not, divided by the
number of pairs in which the partner has a history of depression and the
proband does not. Tests of significance and 90 percent confidence intervals
were calculated based on the Mantel-Haenszel method (Schlesselman, 1982).
Ninety percent confidence intervals were chosen because prior research suggests
that a history of depression or other psychiatric illness may be associated with
increased risk of Alzheimer's disease; no previous investigations have suggested
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 2
that such a history is protective (associated with decreased risk). Ninety percent
confidence intervals provide a one-tailed test of significantly elevated risk (odds
ratios greater than 1.0) at the .05 level.
Etiologic fractions were calculated to represent the proportion of AD
cases attributable to a prior history of depression, antidepressant treatment, or
other psychiatric illness (Schlesselman, 1982). The etiologic fraction is used to
assess the public health impact of risk factors by partitioning the proportion of
disease caused by a given exposure, trait, or intervention. It is calculated as the
exposure rate multiplied by the odds ratio minus one, divided by one plus the
numerator. The etiologic fraction is theoretically analogous to an R2 obtained
from linear regression analysis. Because of differences in scaling and in the
underlying assumptions for linear regression and logistic regression modelling,
however, the two measures are not mathematically equivalent; absolute values
of etiologic fractions are much larger than the corresponding R2. For example,
in one study in which R2 for continuous risk factor variables were compared to
the etiologic fractions for those same risk factors expressed as dichotomous
variables, an R2 of .011 corresponded to an etiologic fraction of .28. More
commonly, the etiologic fractions were four to eight times higher than the
corresponding R2 (Wen & Kramer, 1995).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
Analyses were performed for the entire sample as well as by sex,
zygosity, rearing status, and late onset of Alzheimer's disease. In order to
ensure better comparability between cases and controls on one established
genetic risk factor for which data were available in this sample, these analyses
were also performed on a subset of the sample concordant for presence or
absence of the apolipoprotein e-4 allele (ApoE), an established risk factor for
AD (Corder et al., 1993).
To test the robustness of the results, analyses were performed using
different discordance criteria: a liberal condition, in which all pairs were
included in which the partner was alive and cognitively intact at the latest
possible age of onset for the proband (this condition is analogous to a classical
case-control study, in which controls are selected for their status at one point in
time with no longitudinal follow-up); a stringent condition, in which pairs that
eventually became concordant had to remain discordant for seven years; and a
discordant-only condition, in which all pairs that eventually became concordant
were excluded.
RESULTS
Demographics for the 52 pairs discordant for three years appear in Table
2. Twelve pairs are male, 40 female. Approximately two-thirds of the sample
(67.3%) are dizygotic (fraternal) twins. Most (85.1%) have an elementary
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
school education. The average age of the sample as of January 1, 1996 was
the average age of dementia onset among the probands is 74.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2. Demographic Information
Pairs
Sex
Male 12 23.1
Female 40 76.9
Zygosity
Monozygotic 17 32.7
Dizygotic 35 67.3
Individuals
A
Education
Elementary 86 85.1
Vocational 10 9.9
Secondary 3 3.0
University 2 2.0
History of depression 9 8.9
Antidepressant exposure 6 5.9
Exposure to electroconvulsive therapy 1 1.0
History of other psychiatric illness 3 3.0
Mean SD
Current age (n=60) 84.4 6.6
Age at death (n=44) 83.9 7.9
Age of dementia onset (cases only; n==52) 74.3 9.1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
The prevalence of depressive history in this sample was 8.9%. Five of
the nine individuals with a reported history of depression received treatment
with antidepressant medication; one received both antidepressants and ECT.
Three participants had a history of other psychiatric illness. One was diagnosed
with "psychiatric insufficiency syndrome" and anxiety 11 years before dementia
onset, one had a cancer phobia beginning 24 years before onset of dementia,
and one had paranoid psychosis 17 years before dementia onset.
For each putative risk factor, Table 3 displays the number of pairs in
each of four categories (neither twin exposed, case exposed, control exposed,
and both exposed), the odds ratio based on pairs discordant for both disease and
exposure, a 90% confidence interval, and the etiologic fraction. The results for
history of depressive illness, while not statistically significant due to the small
number of pairs discordant for both depression and AD, are suggestive of
elevated risk. The odds ratio for depression at any time prior to dementia onset
is 4.0 (0.73-21.9), and the odds ratio for depression more than ten years prior
to AD onset is 2.0 (0.28-14.4). Etiologic fractions for these exposures are .21
and .03, respectively.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
Table 3. History of Depressive Illness, Antidepressant or ECT Exposure, and
History of Other Psychiatric Illness as Risk Factors for Alzheimer's Disease
Number of Pairs
by Exposure Category
(Case/Control) Etiologic
-/- + /-/+ + /+ OR* (90% Cl) Fraction
History of depressive illness
At any time before AD onset 43 4 1 2 4.0 (0.73-21.9) .21
> 10 years before AD onset 47 2 1 0 2.0 (0.28-14.4) .03
Antidepressant or ECT exposure 45 2 2 1 1.0 (0.13-7.77) .00
History of other psychiatric illness 47 3 0 0 n/a n/a
*Odds ratio estimate of relative risk in a matched sample, based on pairs
discordant for both disease and exposure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
Exposure to antidepressants or electroconvulsive therapy was unrelated
to Alzheimer’s disease in this sample (OR = 1.0, 0.13-7.77). The odds ratio
for a history of other psychiatric disorders could not be calculated (denominator
equal to 0). All three individuals with a history of mental illness other than
depression experienced the onset of psychiatric symptoms more than 10 years
before dementia onset.
Table 4 displays results for the following subgroups: late onset cases
only, men and women, mono- and dizygotic twins, twins reared apart and
reared together, and pairs concordant for the presence or absence of the
apolipoprotein e-4 allele. The sample sizes are too small to draw conclusions
about differential risks between groups, so all comparisons must be considered
purely exploratory. Despite this limitation, some results are suggestive and
may be worthy of further investigation in a larger sample. Restricting analysis
to late onset (after age 65) AD cases eliminates two depression-discordant pairs,
one in which the case had a history of depression, and one in which the control
had a previous depressive episode. In addition, one early onset case had a
history of other psychiatric illness. Thus, in three of seven pairs discordant for
early onset AD, one partner had a prior psychiatric history, corresponding to a
lifetime prevalence rate of psychiatric disorders of 21% (3/14), compared to the
rate of 12% (12/101) in the overall sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Table 4. Results by Subgroup
Number of pairs discordant for both disease and exposure*
Other
Depression Antidepressant Psychiatric
Group (n pairs) Ever >10 Exposure Illnes
Late onset cases (43) 3/0 1/1 2/1 2/0
Women (38) 3/1 0/1 1/2 2/0
Men (12) 1/0 2/0 1/0 1/0
Monozygotic twins (17) 3/1 1/0 2/1 0/0
Dizygotic twins (33) 1/0 1/1 0/1 3/0
Twins reared together (36) 3/1 0/1 1/2 2/0
Twins reared apart (14) 1/0 2/0 1/0 1/0
Concordant for ApoE (20) 3/1 1/0 2/1 1/0
*The numerator is the number of pairs in which the case has a history of the
exposure and the control does not; the denominator is the number of pairs in
which the control has a history of the exposure and the case does not.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
Although the risk associated with any prior depressive episode does not
appear to differ between men and women after considering the predominance of
women in the sample, risks associated with a history of depression more than 10
years before dementia onset and with antidepressant exposure both appear to be
slightly higher in men. Likewise, given the relative size of the subsamples, the
risk associated with early depression onset and antidepressant treatment appears
greater in twins reared apart than in those reared together. Although
monozygotic twins represent only one-third of the sample, raw results suggest
that risk associated with depression and antidepressant treatment may be greater
for identical than for fraternal twin pairs. On the other hand, risk associated
with other psychiatric illness may be higher in dizygotic pairs. Results for pairs
concordant for ApoE are similar to those for monozygotic twins, except that
one ApoE-concordant dizygotic pair was discordant for a history of other
psychiatric illness.
Tables 5 through 7 display the number of pairs discordant for both
disease and exposure using various discordance criteria. The liberal criteria
(Table 5), in which all pairs discordant at any point in time were included,
generated a sample of 55 pairs. Results were identical to results from the three
year discordance criteria for every exposure and every subgroup. Results
excluding pairs in which the partner became demented within seven years of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
proband's dementia onset are displayed in Table 6 (N = 45). Two pairs
discordant for a history of depression at any point and for antidepressant
exposure were excluded based on these criteria; in one pair, the case had
experienced a depressive episode, and in the other the control had been
depressed. Finally, Table 7 displays results excluding any pair that eventually
became discordant (N = 40). Results are identical to the seven year
discordance criteria. This pattern of results provides no support for the notion
that findings are an artifact of inclusion criteria.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
Table 5. Liberal Inclusion Criteria (Partner Alive and Cognitively Intact at
Time of Proband's Dementia Onset)
Number of pairs discordant for both disease and exposure*
Other
Depression Antidepressant Psychiatric
Group (n pairs) Ever >10 Exposure Illnes
All cases (55) 4/1 2/1 2/2 3/0
Late onset cases (48) 3/0 1/1 2/1 2/0
Women (40) 3/1 0/1 1/2 2/0
Men (15) 1/0 2/0 1/0 1/0
Monozygotic twins (19) 3/1 1/0 2/1 0/0
Dizygotic twins (36) 1/0 1/1 0/1 3/0
Twins reared together (38) 3/1 0/1 1/2 2/0
Twins reared apart (17) 1/0 2/0 1/0 1/0
Concordant for ApoE (29) 3/1 1/0 2/1 1/0
*The numerator is the number of pairs in which the case has a history of the
exposure and the control does not; the denominator is the number of pairs in
which the control has a history of the exposure and the case does not.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
Table 6. Stringent Inclusion Criteria (Concordant After 7 Years)
Number of pairs discordant for both disease and exposure*
Other
Depression Antidepressant Psychiatric
Group (n pairs) Ever >10 Exposure Qlnes
All cases (45) 3/0 2/1 1/1 3/0
Late onset cases (39) 2/0 1/1 1/1 2/0
Women (33) 2/0 0/1 0/1 2/0
Men (12) 1/0 2/0 1/0 1/0
Monozygotic twins (14) 2/0 1/0 1/0 0/0
Dizygotic twins (31) 1/0 1/1 0/1 3/0
Twins reared together (32) 2/0 0/1 0/1 2/0
Twins reared apart (13) 1/0 2/0 1/0 1/0
Concordant for ApoE (17) 2/0 1/0 1/0 1/0
*The numerator is the number of pairs in which the case has a history of the
exposure and the control does not; the denominator is the number of pairs in
which the control has a history of the exposure and the case does not.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
Table 7. Discordant Only Inclusion Criteria
Number of pairs discordant for both disease and exposure*
Group (a pairs)
Other
Depression Antidepressant Psychiatr
Ever > 10 Exposure Illness
All cases (40) 3/0 2/1 1/1 3/0
Late onset cases (34) 2/0 1/1 1/1 2/0
Women (31) 2/0 0/1 0/1 2/0
Men (9) 1/0 2/0 1/0 1/0
Monozygotic twins (11) 2/0 1/0 1/0 0/0
Dizygotic twins (29) 1/0 1/1 0/1 3/0.
Twins reared together (30) 2/0 0/1 0/1 2/0
Twins reared apart (10) 1/0 2/0 1/0 1/0
Concordant for ApoE (13) 2/0 1/0 1/0 1/0
♦The numerator is the number of pairs in which the case has a history of the
exposure and the control does not; the denominator is the number of pairs in
which the control has a history of the exposure and the case does not.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
DISCUSSION
These findings, while not statistically significant due to the small number
of pairs discordant for both disease and exposure, generally support previous
research suggesting that a history of depression may be associated with elevated
risk for developing Alzheimer's disease (Barclay et al., 1986; Broe et al., 1990;
French et al., 1985; Kokmen et al., 1991; Shalat et al., 1987). Although the
width of the confidence intervals precludes rejection of the null hypothesis that
depression is associated with statistically significant risk, the magnitude of the
odds ratios is suggestive of elevated risk. Etiologic fractions represent an
estimate of effect size larger than but analogous to the percentage of variance
explained in a linear regression analysis (R2 ). Unlike significance tests or
confidence intervals, they are not affected by sample size. The etiologic
fraction of .21 associated with depression at any point prior to dementia onset
suggests that about one-fifth of AD cases in the population are attributable to a
prior history of depression. By contrast, only three percent of AD cases are
attributable to depressive episodes more than 10 years before dementia onset.
These results are consistent with an interpretation of depression as an early
symptom of AD.
Etiologic fractions calculated from the EURODEM data show a similar
risk pattern. Depressive episodes more than 10 years before late life AD onset
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
are associated with an etiologic fraction of .04, whereas depression at any point
prior to dementia onset is associated with an etiologic fraction of .08. Although
the impact of depression across the lifespan was lower in the EURODEM study
than in the twin case-control design, in part due to the lower lifetime prevalence
of depression in the EURODEM sample (5.7%), the pattern of decreasing risk
with decreasing proximity to dementia onset supports the prodromal depression
model.
Mortimer (1995) recently reported etiologic fractions for three other risk
factors for AD: low education, family history of dementia, and head trauma
with loss of consciousness. He noted that the public health significance of a
risk factor is dependent on both the strength of association between the risk
factor and the disease and the population frequency of exposure to the risk
factor. Thus, because low education is much more common than family history
of dementia, its associated etiologic fraction is .77, whereas the etiologic
fraction associated with positive family history is only .26. The etiologic
fraction associated with head trauma is .04. Results from this study suggest that
the significance of the risk associated with depressive episodes at any point
prior to AD onset may be roughly comparable to that associated with a family
history of dementia, whereas the impact of risk associated with early life
depression may be comparable to that associated with head trauma.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
As some researchers propose, the Alzheimer's disease process may
commence decades before an individual meets criteria for dementia (La Rue &
Jarvik, 1987; Ohm, Muller, Braak, & Bohl, 1995). Thus, a 10 year window
prior to AD onset may be insufficient to distinguish true risk factors from early
symptoms. As the relative magnitude of the etiologic fractions suggests, while
late life depression may herald the onset of AD, the risk associated with earlier
depressive episodes is quite small. Because the risk associated with depression
appears to increase with proximity to dementia onset, these results can be
interpreted as providing further support for the observation that depression may
be an early manifestation of dementia. The methodology of this study does not
permit any conclusions about whether prodromal depression is endogenous or
reactive.
Consistent with previous research, this study found no relationship
between exposure to antidepressant medications or ECT and AD (Broe et al.,
1990; French et al., 1985; Jorm et al., 1991; Speck et al., 1995). Failure to
find elevated risk associated with antidepressant or ECT exposure provides no
support for the supposition that the link between prior episodes of depression
and subsequent onset of AD is iatrogenic in origin. The etiologic fraction of
.00 associated with these exposures suggests that the failure to find significant
results is not merely due to the low statistical power of this study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
Risk associated with a history of other psychiatric illness could not be
evaluated due to the small number of subjects with such a history. All three
subjects had normal partners and became demented more than 10 years after the
onset of their psychiatric symptoms. Psychiatric symptoms could be interpreted
in several ways. On the one hand, retrospective diagnosis of depression may
simply be imprecise, and some of the instances classified as "other psychiatric"
may in reality have been depressive episodes. It is also possible that psychiatric
symptoms other than depression (such as paranoia or anxiety) may arise
secondarily to changes associated with subclinical AD, either as a reaction to
cognitive decline or an endogenous feature of the illness. Alternatively, finding
three individuals with a prior psychiatric history in the sample may represent a
random occurrence unrelated to the general relationship between depression and
dementia. Finally, these results could be interpreted as generally consistent
with a threshold model in which dementia is diagnosed when an individual's
brain reserve capacity, related to the amount of functional brain tissue, falls
below a threshold at which normal cognitive function cannot be sustained
(Mortimer, 1994). If psychiatric illness is related to lower brain reserve or to
processes that serve to reduce reserve, psychiatric illness may represent an early
expression of some underlying vulnerability to AD. This explanation would
better account for the relationship Kokmen et al. (1991) found between a history
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
of personality disorders and AD than a theory that psychiatric symptoms
represent a response to cognitive impairment, as the onset of personality
disorders typically occurs in late adolescence or early adulthood, long before an
affected individual would be expected to notice signs of cognitive decline
associated with dementia.
Results from exploratory comparisons between various subsamples
suggest that pairs in which one partner has early onset AD may have higher
than average rates of prior psychiatric illness than pairs discordant for late onset
AD. This may be an artifact of recall bias; younger subjects and their
informants may be more likely to remember past episodes of depression and
other mental illness than older individuals. Alternatively, factors associated
with early onset AD may be related to psychiatric symptomatology.
Comparing results between subgroups is difficult given the very small
numbers of pairs discordant for both disease and exposure. Hence the
following comparisons are highly speculative and should be used only for the
purpose of generating hypotheses for further study. Risk associated with a
history of depression more than 10 years before dementia onset and with
antidepressant exposure, but not with a history of depression at any time, may
be higher in men than in women. While this trend may be an artifact of the
small number of male pairs in this sample, it merits further investigation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
Although women are at higher risk for depression, Jorm et al. (1991) found no
significant sex differences in risk for Alzheimer's disease associated with prior
depression; they did not look at sex differences in risk associated with early
onset depression or with antidepressant treatment. Additionally, twins reared
apart may experience greater risk for AD associated with early depression onset
and antidepressant treatment. Speculation about the mechanisms for such a
relationship is premature given the very small size of the male and reared apart
subgroups and the highly exploratory nature of these comparisons.
Risk associated with depression and antidepressant treatment appears to
be greater for monozygotic than for dizygotic twins. Monozygotic twins are
genetically identical. Dizygotic (fraternal) twins share, on average, 50 percent
of their segregating genes. The risk associated with depression might be higher
in monozygotic pairs because genetic risk factors other than ApoE may account
for some of the discordance in dizygotic twins. This hypothesis should be
tested in a larger sample.
Because the results were substantially similar using various criteria for
discordant pairs, there is no evidence to suggest that these results are an artifact
of the definition of discordant. Furthermore, findings were not affected by
excluding twins discordant for ApoE. In particular, depression was not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
associated with higher risk among those pairs whose discordance for AD was
not "explainable" by discordance for a major genetic risk factor.
Several important limitations must be considered when interpreting this
study. First, the sample size lacked adequate statistical power. Given the
magnitude of the effects detected in the present analyses, a sample of 15
disease- and exposure-discordant pairs would be needed to test the primary
hypothesis that depression at any point in time is associated with significantly
increased risk of Alzheimer's disease. Higher numbers would be required to
test hypotheses associated with differences between subgroups.
Additionally, depressive episodes may have been underreported in this
sample. The lifetime prevalence of depression in this sample, 8.9%, is about
half of the lifetime prevalence of depression (17.1%) reported by Kessler et al.
(1994) in a U.S. probability sample of younger adults (aged 15-54). The
lifetime prevalence for unipolar depression in the Swedish Twin Registry as
ascertained by self-report questionnaire reflecting DSM-IH-R criteria was
13.2%; this figure included younger as well as older adults (Kendler, Pedersen,
Johnson, Neale, & Mathe, 1993). Although the EURODEM investigators
found a lower prevalence of depressive history, 5.7%, in their AD case-control
sample (Jorm et al., 1991), the Canadian Study of Health and Aging (1994)
reported a lifetime prevalence of depression of 17.4% and Speck et al. (1995)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
found a history of depression in 18.7% of the older adults they surveyed.
Finally, the Lundby study (1957-1972) in rural Sweden estimated the
cumulative probability of contracting any type of depression before 70 years of
age to be 26.9% for men and 45.2% for women (Rorsman et al., 1990). While
cohort differences in actual prevalence or in reporting may make generalization
from younger samples problematic, these comparisons suggest that information
on some depressive episodes may have been missing from our analyses.
Furthermore, 67% of the depressed people in the twin sample received
antidepressant treatment. This figure contrasts with Regier et al.'s (1993)
finding that only half of those with major depression received professional help
of any type, including antidepressant medications, again suggesting that
depression was underreported in this sample.
Underreporting of depressive episodes may have been due to the fact
that data on depression and other psychiatric disorders were not collected
through a structured diagnostic interview. Additionally, psychiatric data were
not consistently collected through informants for both cases and controls.
While data on cases were collected exclusively through informant reports and
medical records, data on controls were gathered by self report. This lack of
comparability may have biased the results towards unity (nonsignificant
findings), since control subjects may have reported more episodes of depression
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
or other psychiatric illness than were listed in medical records or known to
proxies.
Because dizygotic twins share only one half of their segregating genes in
common, and the sample is predominantly dizygotic, the control for genetic
factors in the present study was less than complete. A more desirable sample
for discordant pair analysis would have been based on monozygotic pairs.
Unfortunately, insufficient numbers of such pairs were available. Hence,
conclusions concerning control for sources of similarity within pairs should be
limited to familial causes and not be interpreted as solely genetic. The sample
is too small for an evaluation of the relative importance of genetic and
environmental effects for the depression-dementia association.
The retrospective design of this study may have led to reporting bias.
Informants providing data on behalf of a demented relative may have been more
likely to report a risk factor than those reporting on a control subject; this could
have led to inflated estimates of relative risk. The study attempted to minimize
reporting bias by asking a large number of questions about medical, psychiatric,
and exposure history; exactly which items were being examined as putative risk
factors for Alzheimer's disease was not made known to the informants.
Additionally, informant data were checked against medical records at the time
of the interview and again in preparation for this analysis to confirm historical
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
data, and all cases of reported depression were verified through medical
records.
The sample contains a group of twins reared apart since childhood. This
group is not representative of all twins or all people. Factors associated with
family breakup may limit generalizability to the overall population at risk for
Alzheimer's disease. On the other hand, using a co-twin control design ensures
that both partners in a discordant pair were subjected to the same rearing status.
Thus bias due to the inclusion of pairs reared apart should not have influenced
the results.
Finally, this study included twin pairs from two different samples.
Although both samples were taken from the same population-based source (the
Swedish Twin Registry), they are not identical. Because membership in the
OCTO-Twin-dementia sample required that both partners live past the age of
80, this group is unusually hardy, which may limit generalizability of results.
For purposes of risk factor analysis, however, this should not pose a problem
because both partners in each pair were required to be hardy.
Despite its small sample size, this study includes many advantages over
previous case-control studies. The population-based sampling technique avoids
the issue of volunteer or hospital sample bias. Furthermore, ascertainment of
cases, including longitudinal follow-up through autopsy, is as comprehensive
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and accurate as possible. Many prior studies selected cases and controls at one
point in time; errors in the diagnosis of a progressive disorder such as
Alzheimer's disease are more likely without longitudinal assessment, and some
subjects included as controls in previous investigations may have been
experiencing incipient cognitive impairment. Most importantly, use of a twin
sample ensures comparability of cases and controls, especially with respect to
known risk factors for Alzheimer's disease such as age and family history of
dementia. Since this sample is predominantly dizygotic, genetic influences are
only partially controlled for in these analyses; however, the twin methodology
does control for general familial factors that may influence the depression-
dementia relationship.
Finding no elevated risk associated with depression in this sample would
have suggested that the depression-dementia association previously reported
may reflect some familial mechanism influencing susceptibility to both
depression and AD. The magnitude of these effects after controlling for
familial factors strengthens the idea that the link between depression and
dementia does not reflect a common mechanism, and that depression truly
presages the onset of Alzheimer’s disease.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
REFERENCES
American Psychiatric Association. (1987). Diagnostic and statistical
manual of mental disorders (3rd ed., rev.). Washington, DC: Author.
Barclay, L. L., Kheyfets, S., Zemcov, A., Blass, J. P., & McDowell,
F. H. (1986). Risk factors in Alzheimer's disease. In A. Fisher, I. Hanin, &
C. Lachman (Eds.), Alzheimer's and Parkinson's diseases: Strategies for
research and development (pp. 141-146). New York: Plenum Press.
Berg, S., Johansson, B., Plomin, R., Ahem, F. M., Pedersen, N. L., &
McCleam, G. E. (1992). Origins of variance in the old-old: The first
presentation of the OCTO-Twin study in Sweden [Abstract]. Behavior
Genetics. 22. 708-709.
Breitner, J. C. S., Gatz, M., Bergem, A. L. M., Christian, J. C.,
Mortimer, J. A., McCleam, G. E., Heston, L. L., Welsh, K. A., Anthony, J.
C., Folstein, M. F., & Radebaugh, T. S. (1993). Use of twin cohorts for
research in Alzheimer's disease. Neurology. 43. 261-267.
Breitner, J. C. S., Gau, B. A., Welsh, K. A., Plassman, B. L.,
McDonald, W. M., Helms, M. J., & Anthony, J. C. (1994). Inverse
association of anti-inflammatory treatments and Alzheimer's disease: Initial
results of a co-twin control study. Neurology. 44. 227-232.
Broe, G. A., Henderson, A. S., Creasey, H., McCusker, E., Korten, A.
E., Jorm, A. F., Longley, W., & Anthony, J. C. (1990). A case-control study
of Alzheimer's disease in Australia. Neurology. 40. 1698-1706.
The Canadian Study of Health and Aging. (1994). The Canadian study
of health and aging: Risk factors for Alzheimer’s disease in Canada.
Neurology. 44. 2073-2080.
Corder, E. H., Saunders, A. M., Strittmatter, W. J., Schmechel, D. E.,
Gaskell, P. C., Small, G. W., Roses, A. D., Haines, J. L., & Pericak-Vance,
M. A. (1993). Gene dose of apolipoprotein E type 4 allele and the risk of
Alzheimer's disease in late onset families. Science. 261. 921-923.
Fleiss, J. L. (1981). Statistical methods for rates and proportions (2nd
ed.). New York: John Wiley & Sons.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). "Mini-
Mental State": A practical method for grading the cognitive state of patients for
the clinician. Journal of Psychiatry Research. 12. 189-198.
French, L. R., Schuman, L. M., Mortimer, J. A., Hutton, J. T.,
Boatman, R. A., & Christians, B. (1985). A case-control study of dementia of
the Alzheimer type. American Journal of Epidemiology. 121. 414-421.
Gatz M. (1993). Twin studies of risk factors for Alzheimer's disease
[Abstract]. The Gerontologist. 33. 299.
Gatz, M., Reynolds, C., Nikolic, J., Lowe, B., Karel, M., & Pedersen,
N. (1995). An empirical test of telephone screening to identify potential
dementia cases. International Psychogeriatrics. 7. 429-437.
Heyman, A., Wilkinson, W. E., Stafford, J. A., Helms, M. J., Sigmon,
A. H., & Weinberg, T. (1984). Alzheimer's disease: A study of
epidemiological aspects. Annals of Neurology. 15. 335-341.
Jorm, A. F., Van Duijn, C. M., Chandra, V., Fratiglioni, L., Graves,
A. B., Heyman, A., Kokmen, E., Kondo, K., Mortimer, J. A., Rocca, W. A.,
Shalat, S. L., Soininen, H., & Hofrnan, A. (1991). Psychiatric history and
related exposures as risk factors for Alzheimer's disease: A collaborative re
analysis of case-control studies. International Journal of Epidemiology. 20.
S43-S47.
Kendler, K. S., Pedersen, N., Johnson, L., Neale, M. C., & Mathe, A.
A. (1993). A pilot Swedish twin study of affective illness, including hospital-
and population-ascertained subsamples. Archives of General Psychiatry. 50.
699-706.
Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes,
M., Eshleman, S., Wittchen, H.-U., & Kendler, K. S. (1994). Lifetime and
12-month prevalence of DSM m-R psychiatric disorders in the United States:
Results from the National Comorbidity Study. Archives of General Psychiatry.
Si* 8-19.
Kokmen, E., Beard, C. M., Chandra, V., Offord, K. P., Schoenberg,
B. S., & Ballard, D. J. (1991). Clinical risk factors for Alzheimer's disease:
A population-based case-control study. Neurology. 41. 1393-1397.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
La Rue, A., & Jarvik, L. F. (1987). Cognitive function and prediction
of dementia in old age. International Journal of Aging and Human
Development. 25. 79-89.
Lipman, R. S., Brown, E. A., Silbert, G. A., Rains, D. G., & Grady,
D. A. (1993). Cognitive performance as modified by age and ECT history.
Progress in Neuro-Psychopharmacologv and Biological Psychiatry. 17. 581-
594.
Liston, E. H., Jarvik, L. F., & Gerson, S. (1987). Depression in
Alzheimer's disease: An overview of adrenergic and cholinergic mechanisms.
Comprehensive Psychiatry. 28. 444-457.
McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., &
Stadlan, E. M. (1984). Clinical diagnosis of Alzheimer's disease: Report of
the NINCDS-ADRDA work group under the auspices of the Department of
Health and Human Services Task Force on Alzheimer's disease. Neurology.
34,939-944.
Medlund, P., Cederlof, R., Floderus-Myrhed, B., Friberg, L., &
Sorenson, S. (1977). A new Swedish twin registry. Acta Medica Scandinavica.
1977 (Suppl.l. 1-111.
Mendez, M. F., Underwood, K. L., Zander, B. A., Mastri, A. R.,
Sung, J. H., & Frey, W. H. (1992). Risk factors in Alzheimer's disease: A
clinicopathologic study. Neurology. 42. 770-775.
Migliorelli, R., Teson, A., Sabe, L., Petracchi, M., Leiguarda, R., &
Starkstein, S. E. (1995). Prevalence and correlates of dysthymia and major
depression among patients with Alzheimer's disease. American Journal of
Psychiatry. 152. 37-44.
Minami, Y., Tsuji, I., Fukao, A., Hisamichi, S., Asano, H., Sato, M.
& Shinoda, K. (1995). Physical status and dementia risk: A three-year
prospective study in urban Japan. International Journal of Social Psychiatry.
4L 47-54.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
Morris, J. C., Heyman, A., Mohs, R. C., Hughes, J. P., van Belle, F.,
Fillenbaum, G., Mellits, E. D., & Clark, C. (1989). The consortium to
establish a registry for Alzheimer’s disease (CERAD): I. Clinical and
neuropsychological assessment of Alzheimer's disease. Neurology. 39. 1159-
65.
Mortimer, J. A. (1995). Prospects for prevention of dementia and
associated impairments. In L. A. Bond, S. J. Cutler, & A. Grams (Eds.),
Promoting successful and productive aging (pp. 131-147). Thousand Oaks,
CA: Sage Publications.
Mortimer, J. A. (1994). What are the risk factors for dementia? In F.
A. Huppert, C. Brayne, & D. W. O'Connor (Eds.), Dementia and normal aging
(pp. 208-229). Cambridge, England: Cambridge University Press.
Office of Technology Assessment. (1988). Confronting Alzheimer’s
disease and other dementias. Philadelphia, PA: J. B. Lippincott.
Office of Technology Assessment. (1987). Losing a million minds:
Confronting the tragedy of Alzheimer's disease and other dementias (OTA-BA-
323). Washington, DC: U.S. Government Printing Office.
Ohm, T. G., Muller, H., Braak, H., & Bohl, J. (1995). Close-meshed
prevalence rates of different stages as a tool to uncover the rate of Alzheimer's
disease-related neurofibrillary changes. Neuroscience. 64. 209-217.
Pedersen, N. L., Gatz, M., Winblad, B., Pearson, C. G., & Berg, S.
(1989). Dementias in Swedish twins. In K. Iqbal, H. M. Wisniewski, & B.
Winblad (Eds.), Alzheimer's disease and related disorders (pp. 217-222). New
York: Alan R. Liss.
Pedersen, N. L., McCleam, G. E., Plomin, R., Nesselroade, J. R.,
Berg, S., & DeFaire, U. (1991). The Swedish Adoption/Twin Study of Aging:
An update. Acta Geneticae Medicae et Gemellologiae. 40. 7-20.
Reding, M., Haycox, J., & Blass, J. (1985). Depression in patients
referred to a dementia clinic: A three-year prospective study. Archives of
Neurology. 42. 894-896.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
Regier, D. A., Narrow, W. E., Rae, D. S., Manderscheid, R. W.,
Locke, B. Z., & Goodwin, F. K. (1993). The de facto U.S. mental and
addictive disorders service system: Epidemiologic Catchment Area prospective
1-year prevalence rates of disorders and services. Archives of General
Psychiatry. 50. 85-94.
Rorsman, B., Grasbeck, A., Hagnell, O., Lanke, J., Ohman, R.,
Ojesjo, L., & Otterbeck, L. (1990). A prospective study of first-incidence
depression: The Lundby Study, 1957-1972. British Journal of Psychiatry. 156.
336-342.
Rubin, E. H., Kinscherf, D. A., Figiel, G. S., & Zorumski, C. F.
(1993). The nature and time course of cognitive side effects during
electroconvulsive therapy in the elderly. Journal of Geriatric Psychiatry and
Neurology. 6. 78-83.
Schlesselman, J. J. (1982). Case-control studies: Design, conduct.
analysis. New York: Oxford University Press.
Shalat, S. L., Seltzer, B., Pidcock, C., & Baker, E. L. (1987). Risk
factors for Alzheimer's disease: A case-control study. Neurology. 37. 1630-
1633.
Snowdon, D. A., Kemper, S. J., Mortimer, J. A., Greiner, L. H.,
Wekstein, D. R., & Markesbery, W. R. (1996). Linguistic ability in early life
and cognitive function and Alzheimer's disease in late life: Findings from the
Nun Study. Journal of the American Medical Association. 275. 528-532.
Speck, C. E., Kukull, W. A., Brenner, D. E., Bowen, J. D.,
McCormick, W. C., Teri, L., Pfanschmidt, M. L., Thompson, J. D., &
Larson, E. B. (1995). History of depression as a risk factor for Alzheimer's
disease. Epidemiology. 6. 366-369.
Teri, L., & Wagner, A. (1992). Alzheimer's disease and depression.
Journal of Consulting and Clinical Psychology. 60. 379-391.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
Van Duijn, C. M., Clayton, D. G., Chandra, V., Fratiglioni, L.,
Graves, A. B., Heyman, A., Jorm, A. F., Kokmen, E., Kondo, K., Mortimer,
J. A., Rocca, W. A., Shalat, S. L., Soininen, H., & Hofman, A. (1994).
Interaction between genetic and environmental risk factors for Alzheimer's
disease: A reanalysis of case-control studies. Genetic Epidemiology. 11. 539-
551.
Van Ojen, R., Hooijer, C., Bezemer, D., Jonker, C., Lindeboom, J., &
Van Tilburg, W. (1995). Late-life depressive disorder in the community: I.
The relationship between MMSE score and depression in subjects with and
without psychiatric history. British Journal of Psychiatry. 166. 311-315.
Wen, S. W., & Kramer, M. S. (1995). Etiologic fraction analysis for
continuously distributed outcome variables and empirical analogy with
dichotomized outcome variables. International Journal of Epidemiology. 24.
457-461.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Depression as a risk factor for dementia
PDF
Frequency and risk factors of poststroke dementia
PDF
Alcohol expectancies and consumption: Age and sex differences
PDF
Head injury and dementia: A co-twin control study of Swedish twins
PDF
Cognitive functioning and dementia following cancer: A Swedish twin study
PDF
Experimental evaluation of a new intervention designed to promote dementia caregivers' acceptance and empathic responding towards care -recipients
PDF
Treatment of generalized anxiety disorder in older adults
PDF
Expectancies for alternative behaviors predict drinking problems: A test of a cognitive reformulation of the matching law
PDF
Depression and suicidality in Latino adolescents: A study of acculturation and gender role beliefs
PDF
Hispanic culture, acculturation, and distress among caregivers of dementia patients
PDF
Biopsychosocial factors in major depressive disorder
PDF
Articulated thoughts regarding cognitions toward older adults
PDF
Alcohol treatment entry and refusal in a sample of older veterans
PDF
Effects of cultural values and behaviors in sociocultural stress and coping among Korean-American and Korean caregivers
PDF
Dementia caregiving and ethnicity: African American caregivers and the sociocultural stress and coping model
PDF
Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task
PDF
Mood Induction As An Analog To Depression In Older Adults: Aspects Of Depression And Cognitive Effects
PDF
An examination of affective modulation, psychopathy, and negative schizotypy in college and community samples
PDF
Emotional intelligence and smoking risk factors in early adolescents
PDF
Cognitive dysfunction and occupations with likely significant magnetic field exposure: A cross-sectional study of elderly Mexican Americans
Asset Metadata
Creator
Wetherell, Julie Loebach
(author)
Core Title
History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample
School
Graduate School
Degree
Master of Arts
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Gerontology,health sciences, mental health,OAI-PMH Harvest,Psychology, clinical
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Gatz, Margaret (
committee chair
), [illegible] (
committee member
), Knight, Bob G. (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-9555
Unique identifier
UC11342275
Identifier
1383550.pdf (filename),usctheses-c16-9555 (legacy record id)
Legacy Identifier
1383550.pdf
Dmrecord
9555
Document Type
Thesis
Rights
Wetherell, Julie Loebach
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
health sciences, mental health