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
/
Underlying neural mechanisms of depression and dementia
(USC Thesis Other)
Underlying neural mechanisms of depression and dementia
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
UNDERLYING NEURAL MECHANISMS OF DEPRESSION AND DEMENTIA
by
Jessica Anne Brommelhoff
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
December 2010
Copyright 2010 Jessica Anne Brommelhoff
ii
Dedication
In loving memory of my grandfather, Dr. Carl E. Ekberg, Jr., the “Ekberg Expert.”
1920-2007
iii
Acknowledgements
I would like to gratefully acknowledge Margaret Gatz, Ph.D. for her dedicated
support and supervision throughout every step of this project. In addition to Dr. Gatz, I
would like to also thank Nancy Pedersen, Ph.D. for allowing me access to the Swedish
Twin Registry data, and to Drs. Bryan Spann and John Go for lending their expertise in
neurology and neuroradiology and for spending countless hours rating the CT scans. I
would also like to thank Wendy Mack, Ph.D. and Carol Prescott, Ph.D. for providing
highly-skilled guidance, especially in regards to the study design and analyses.
Thank you as well to all past and present Gatz Lab members (Cynthia Pearson,
Ph.D., Barbara Yuen, Maggi Mackintosh, Ph.D., Poorni Otilingam, Ph.D., Lewina Lee,
M.A., Emily Schoenhofen, M.A., Trish George, M.A., and Carlos Rodriguez), for their
support (moral and otherwise), comments, and suggestions over the past six years, and to
Kristen Haut, Ph.D. (my “dissertation buddy”), Linda Ercoli, Ph.D., Paul Cernin, Ph.D.,
and David Crawford for keeping me sane during my internship year. Additionally, I
would like to thank my friends and classmates, especially Lara Heflin, Ph.D., Kent
Holmes, Jennifer Stevens, Anne Hong, Jacob Lentz, Carrie Willingham, and Jesse
Eidsness for keeping me “well-rounded” and helping me maintain a sense of humor.
This was all made possible by the love and constant support of my family
members, and I would like to particularly acknowledge my favorite aunt, Janet Ekberg
for all the weekend getaways, shopping/spa sprees, and essentially ensuring the little joys
in life would not lay by the wayside. Finally, I would like to thank my grandparents, Carl
and Dorothy Ekberg for providing the inspiration, and my parents, Jurgen and Gretchen
Brommelhoff, for decades of believing in me, and in turn, helping me believe in myself.
iv
Table of Contents
Dedication ii
Acknowledgements iii
List of Tables v
Abstract viii
Chapter 1: Introduction 1
Neurobiological Mechanisms of Depression and Dementia 3
Specific Aim and Study Hypotheses 10
Chapter 2: Methods 12
Participants 12
Figure 1: Study Sample 13
Measures 14
Analyses 30
Chapter 3: Results 35
Depression Status—Narrow Definition 44
Hypothesis 1 49
Hypothesis 2 55
Hypothesis 3 57
Hypothesis 4 59
Depression Status—Broad Definition 61
Hypothesis 1 64
Hypothesis 2 70
Hypothesis 3 71
Hypothesis 4 73
Dementia Severity 75
Source of Depression Information 78
Chapter 4: Discussion 79
Limitations 90
Chapter 5: Conclusion 94
References 96
Appendix A: Visual Rating Worksheet and Corresponding Variable Names 104
Appendix B: Comparisons between Logistic Regression and GEE Method 105
Results
v
List of Tables
Table 1: Sample Demographics and Covariates by Dementia Type 36
Table 2: Correlation Coefficients between Demographic Characteristics 37
of Sample
Table 3: Association between Demographic Variables and CT Measures 39
Table 4: Phi Correlation Coefficients between Coronary Artery Disease 42
Indicators
Table 5: Phi Correlation Coefficients between Risk Factors for 42
Cerebrovascular Disease
Table 6: Presence of Regional Hypodensities by CAD Status 43
Table 7: Presence of Regional Hypodensities by History of TIA 45
Table 8: Sample Demographics by Depression Group 47
Table 9: Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset 50
Depression and Late-Life Depression compared with No Depression
for All Dementia, Alzheimer’s Disease, and Vascular Dementia
Table 10: Risk of Confluent Frontal Lobe Deep Matter Hypodensities in 51
Late-Onset Depression and Late-Life Depression compared with No
Depression for All Dementia, Alzheimer’s Disease, and Vascular
Dementia
Table 11: Risk of Subcortical White Matter Hypodensities in Late-Onset 53
Depression and Late-Life Depression compared with No Depression
for All Dementia, Alzheimer’s Disease, and Vascular Dementia
Table 12: Number of Striatal Hypodensities in Late-Onset Depression and 54
Late-Life Depression compared with No Depression for All Dementia,
Alzheimer’s Disease, and Vascular Dementia
Table 13: Right and Left Frontal Ventricular Width (in mm) for All Dementia, 60
Alzheimer’s Disease, and Vascular Dementia
Table 14: Sample Demographics by Depression Group (Broad) 62
vi
Table 15: Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset 66
Depression and Late-Life Depression compared with No Depression
for All Dementia, Alzheimer’s Disease, and Vascular Dementia
Table 16: Risk of Confluent Frontal Lobe Deep Matter Hypodensities in 67
Late-Onset Depression and Late-Life Depression compared with No
Depression for All Dementia, Alzheimer’s Disease, and Vascular
Dementia
Table 17: Risk of Subcortical White Matter Hypodensities in Late-Onset 68
Depression and Late-Life Depression compared with No Depression
for All Dementia, Alzheimer’s Disease, and Vascular Dementia
Table 18: Number of Striatal Hypodensities in Late-Onset Depression and 69
Late-Life Depression compared with No Depression for All Dementia,
Alzheimer’s Disease, and Vascular Dementia
Table 19: Right and Left Frontal Ventricular Width (in mm) for All Dementia, 74
Alzheimer’s Disease, and Vascular Dementia
Table 20: Numbers of Individuals by Dementia Severity Rating and Level of 76
Impairment in Each CDR Domain
Table 21: Mean Global CDR Score and CDR Domain Scores by Depression 77
Group
Table 22: Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset 107
Depression (Narrow) compared with No Depression for All Dementia
and Alzheimer’s Disease Using Logistic Regression and Generalized
Estimating Equations
Table 23: Risk of Subcortical White Matter Hypodensities in Late-Onset 108
Depression (Narrow) compared with No Depression for All Dementia
and Alzheimer’s Disease Using Logistic Regression and Generalized
Estimating Equations
Table 24: Number of Striatal Hypodensities in Late-Onset Depression (Narrow) 109
compared with No Depression for All Dementia and Alzheimer’s
Disease Using Logistic Regression and Generalized Estimating Equations
Table 25: Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset 110
Depression (Broad) compared with No Depression for All Dementia and
Alzheimer’s Disease Using Logistic Regression and Generalized
Estimating Equations
vii
Table 26: Risk of Subcortical White Matter Hypodensities in Late-Onset 111
Depression (Broad) compared with No Depression for All Dementia
and Alzheimer’s Disease Using Logistic Regression and Generalized
Estimating Equations
Table 27: Number of Striatal Hypodensities in Late-Onset Depression (Broad) 112
compared with No Depression for All Dementia and Alzheimer’s
Disease Using Logistic Regression and Generalized Estimating
Equations
viii
Abstract
Studies have shown that white matter changes and other neuropathology
frequently found in individuals with dementia, also may be related to late-life depression,
or “vascular dementia.” As these two conditions frequently coexist, the question of
whether there is a relationship between these neuropathological changes and depression
among individuals with manifest dementia has not been established. The aim of the
present study was to examine whether there were neuroanatomical differences evident on
the CT scans of individuals with dementia based upon depression onset (no depression
versus early-onset versus late-onset) and history of late-life depression (any episode of
depression after age 60). We hypothesized that individuals with dementia and late-onset
depression and/or late-life depression would be more likely than non-depressed
individuals with dementia to exhibit frontal lobe deep white matter, subcortical white
matter, and subcortical gray matter hypodensities. We found that compared to
individuals with Alzheimer’s disease and no depression, individuals with Alzheimer’s
disease and late-onset depression had a greater number of striatal hypodensities (gray
matter hypodensities in the caudate nucleus and lentiform nucleus, which includes the
putamen and globis pallidus). In addition, we found that although there were no
differences between the non-depressed and late-onset and late-life depression groups with
respect to white matter hypodensities, the late-onset depression and late-life depression
groups in comparison to the non-depressed group displayed a significantly higher degree
of global functional impairment, as well as impairment within the domains of memory,
orientation, and in the context of their home activities and hobbies. These findings
ix
suggest that late-onset depression may be a process that is distinct from the
neurodegenerative changes caused by Alzheimer’s disease.
1
Chapter 1: Introduction
Studies have shown that dementia and a history of depression frequently coexist
(Jorm, 2001). However, the mechanism behind this association remains controversial. A
meta-analysis by Jorm (2001) suggested that a history of depression nearly doubles the
risk of dementia, as found by both case-control studies (RR = 2.01, CI = 1.16-3.50) and
prospective studies (RR = 1.87, 1.09-3.20). While Jorm concluded that depression was a
risk factor for dementia, many studies have found that depression acts as a risk factor
predominantly when the depression has developed for the first time closer to dementia
onset (Brommelhoff et al., 2009; Chen, Ganguli, Mulasnt, & DeKosky, 1999; Wetherell,
Gatz, Johansson, & Pedersen, 1999; Yaffe et al., 1999). In these studies the authors
concluded that depression is a prodromal feature of dementia, suggesting that depression
manifests itself as an early symptom of dementia.
Depression is also a common co-morbid condition in older adults with dementia.
Though prevalence estimates vary, approximately 40-50% of Alzheimer’s Disease (AD)
patients experience a depressed mood and 10-20% meet clinical criteria for a depression
diagnosis (Wragg & Jeste, 1989). A study by Bassuk, Berkman, and Wypij (1998) found
that depression did not precede cognitive impairment, but instead was present in the
participants who had already experienced a reduction in cognitive abilities. They
concluded that depression was likely to be a reaction to an individual’s insight about
decline in cognitive functioning, which may precede an actual diagnosis of dementia by
several years.
Research into the phenomenology of late-onset depression (LOD) has suggested
that while LOD and early-onset depression (EOD) symptom profiles are similar, etiology
2
of these disorders are different (Brodaty et al., 2001; Gallagher et al., 2009). In a review
of research on depression in AD and other dementias, Boland (2000) concluded that the
co-morbid depression found in dementia differs from classic depression. Rather than co-
morbid depression being a reaction to cognitive decline, he suggested that the depression
arises from anatomical damage to the brain caused by the neuropathological course of
dementia. In other words, the depression is caused by the pathology in the brain that
leads to cognitive impairment and eventually dementia. Suggesting that depression is not
a reaction to the actual manifestation or diagnosis of dementia per se, Carpenter et al.
(2008) measured depressive symptoms both before and after an individual was diagnosed
with dementia and found that there was no increase in depressive symptoms post-
diagnosis.
Boland’s conclusion is further supported by studies looking at the
neuroanatomical features of depression and dementia. Compared to non-depressed
patients with dementia, patients with dementia and depression have different
histopathological and neurochemical characteristics, suggesting a specific pathogenesis
rather than a reactive phenomenon (Liao et al., 2003). In a review of the clinical and
biological differences between late-onset depression and early-onset depression,
Schweitzer, Tuckwell, O’Brien, and Ames (2002) concluded that the white matter
changes that are common in individuals with late-onset depression were associated with
cognitive impairment and that these changes were indicative of a prodrome to dementia,
most likely of the vascular type.
A prospective study by Heiden et al. (2005) investigated the associations between
the extent of white matter brain changes and clinical outcomes in older adult depressed
3
patients. They found that WMH were associated with a poorer outcome both cognitively
and with respect to the course of the depressive illness. Thus it is possible that the brain
changes that accompany late-onset depression may have effects beyond their contribution
in the development of a depressive disorder. In fact, in a recent review, Thomas and
O’Brien (2008) concluded that there is ample evidence that the neurocognitive
impairment seen in individuals with late-onset depression persists even after the
depression has remitted.
Neurobiological Mechanisms of Depression and Dementia
Definitions of Neurobiological Terms
It is important to note that the terms “white matter lesions,” “white matter
hypodensities,” and “white matter hyperintensities” are used rather interchangeably
throughout much of the neuroimaging literature. This is due to the fact that on CT scans
white matter lesions show up as hypodense (or darker) regions and on MRI as
hyperintense (or brighter) regions. We will encompass all terms referring to these white
matter lesions with the acronym WMH, although in the present study we used CT scans,
so WMH with respect to our study methodology and results will refer to white matter
hypodensities. This study also differentiates between deep WMH and subcortical WMH.
Deep white matter hypodensities (DWMH) were defined as cortical WMH that were
located medially from the sulci. The term “subcortical” will refer to the basal ganglia and
thalamic regions of the brain, with subcortical white matter referring to the internal
capsule and subinsular region.
4
White Matter Changes
Many studies have suggested an association between an increased number of
white matter lesions and depression, specifically late-onset depression (Lesser et al.,
1996) both with (O’Brien et al., 1996, Thomas & O’Brien, 2008) and without cognitive
impairment or dementia (Thomas et al., 2002, Firbank et al., 2005, Jorm et al., 2005,
Krishnan et al., 2006). Other studies have suggested that white matter lesions are often
evident among individuals lacking any form of pathology, with the estimated prevalence
varying considerably, though increasing with age (see Coffey et al., 1993).
As WMH have been shown to be indicative of cerebrovascular disease (CVD;
Jagust et al., 2008), it has been suggested that there could be a vascular component to
late-onset depression. To test the theory of vascular depression (Alexopoulos et al.,
1997)—the idea that the sequelae of cerebrovascular disease, such as cerebral white
matter lesions, may cause or exacerbate late-life depression—Thomas et al. (2001)
compared the postmortem brain tissue of 20 non-demented patients with a history of
depression to 20 non-demented controls. They found that individuals with a history of
depression had significantly greater levels of vascular disease than the controls. In a
subsequent study, Thomas et al. (2002) carried out in vitro neuroimaging on the slices of
brain tissue from the individuals in the prior study. They found that in the samples from
the depressed patients, WMH were more frequently found in the dorsolateral prefrontal
cortex and more frequently were ischemic in origin compared to samples from those
without a history of depression. The patients from whom the tissue had been taken had
not been scanned prior to death, however, so correlations with antemortem findings were
not possible.
5
A greater amount of total brain WMH are also thought to be indicative of more
severe depression or a greater number of depressive symptoms. For example, among
cognitively intact older adults, both Jorm et al. (2005) and Heiden et al. (2005) found a
direct association between total brain WMH and severity of depression and number of
depressive symptoms. Additionally, in a population of older adults with and without
cognitive impairment or dementia, Lavretsky et al. (2008) found an association between
higher white matter lacunar volume and symptoms of depressed mood, anhedonia,
anergia, and apathy at the time of assessment, even after controlling for cognitive status.
The direction of causation from these studies was not established due to their cross-
sectional study designs. One longitudinal study, however, found that white matter
changes pre-dated and independently predicted depressive symptoms in older adult
participants (Teodorczuk et al., 2007), providing some evidence that WMH are an
antecedant to depression.
There is some evidence that depression with a vascular etiology is unique to late-
onset depression. One study found that a greater amount of total brain WMH was only
more prevalent in participants with late-onset depression (over age 50) compared with the
early-onset and no depression groups (Lesser et al., 1996). Additionally, the late-onset
group showed deficits in executive functions, which suggests frontal lobe dysfunction.
This observation supports the findings from studies showing that it is specifically frontal
lobe WMH that are associated with a higher rate of depressive symptoms among those
without dementia (Clark et al., 1998, Firbank et al., 2005). In addition, for non-depressed
older adults, more frontal lobe WMH are shown to be associated with more severe levels
of dementia due to AD or vascular dementia (Capizzano et al., 2004).
6
While periventricular white matter lesions are often associated with AD, lesions
in the subcortical white matter are more often associated with late-onset depression
(Schweitzer, Tuckwell, O’Brien, & Ames, 2002). Hickie et al. (1995) evaluated 39
hospital inpatients over age 60 with severe depression and found an association between
later age of first depressive episode and increased subcortical deep white matter
hyperintensities. After a mean follow-up time of 14 months, they found that 27% of the
original 39 inpatients developed a probable dementia syndrome, which was predicted by
a later age of depression onset and subcortical deep WMH (Hickie, Scott, Wilhelm, &
Brodaty, 1996). This suggests that individuals with late-onset depression may experience
a worse cognitive trajectory than those with early-onset depression or no depression.
These findings of an association between frontal lobe and subcortical deep WMH
and depression suggest that damage to frontal-subcortical circutry may be responsible for
late-onset depression. Such damage would be consistent with the fronto-striatal
hypothesis of depression (Shah, Glabus, Goodwin, & Ebmeier, 2002), which suggests
that damage to the frontal lobe and striatum (input nuclei for the basal ganglia comprised
of the caudate nucleus and the putamen) are associated with depression, though the exact
nature of the association remains unclear. In her review focusing on subcortical ischemic
vascular dementia, Chui (2007) posits that deep white matter lesions can disrupt frontal-
subcortical loops and the white matter tracts therein, which are important for cognition
and emotion.
Basal Ganglia-Thalamic Lesions
In their review of studies that investigate structural brain abnormalities in
affective disorders, Taylor and Krishnan (2003) posit that for late-onset depression
7
“[e]vidence is strongest for a contributory effect of [subcortical hyperintensities],
particularly as to hyperintensities of the basal ganglia” (p. 61). In fact some of the
earliest studies in this area dating back nearly 20 years noted the prevalence and severity
of subcortical hyperintensities among depressed older adults. For example, Coffey,
Figiel, Djang, and Weiner (1990) compared subcortical hyperintensities for both normal
and depressed older adults (60 years of age and older) and found lesions in the
subcortical gray matter nuclei (basal ganglia and thalamus) to be significantly more
common in the group of depressed older adults than in non-depressed older adults.
Furthermore depressed older adults with a history of neurologic illness, such as mild
dementia or stroke, had the highest prevalance of moderate to severe subcortical
hyperintensities.
Another more recent study also found differences in the subcortical gray matter in
depressed adults. Kim, Hamilton, and Gotlib (2008) found that adults with depression
have reduced caudate nucleus (part of the lentiform nucleus) volumes compared to their
non-depressed counterparts. In addition, Greenwald et al. (1998) sought further
localization of the subcortical lesions found to be common specifically in late-life
depression, by comparing the subcortical gray matter hyperintensities in a group of older
adults receiving treatment for depression to a group of older adult community controls.
They found that left-hemisphere putaminal hyperintensities significantly predicted
assignment into the depressed group. As the putamen makes up part of the lenticular
nucleus, which is a component of the striatum, this finding also supports the fronto-
striatal hypothesis.
8
Cortical Atrophy and Ventricular Enlargement
In contrast to the suggestion that neuroanatomical damage may lead to depression,
some reasearch has suggested that chronic depression may lead to dementia (Jorm, 2001).
A proposed mechanism for depression increasing risk of dementia is that depression leads
to damage to the hippocampus and temporal lobe through a glucocorticoid cascade (e.g.,
Jorm, 2001). In an article summarizing findings of structural changes in the brain due to
depression, Kanner (2004) reported that decreases in hippocampal volume are directly
associated with chronicity, severity, and duration of depression. Additionally, in their
study of 38 women with major depressive disorder (MDD), Sheline, Gado, and Kraemer
(2003) found that there was a significant positive correlation between amount of
hippocampal volume reduction and duration of untreated depression.
Atrophy of structures in the temporal lobe, especially the hippocampus, is also
often found in patients with AD (Schweitzer, Tuckwell, Ames, & O’Brien, 2001). A
study comparing CT scans of 12 twin pairs from a prior wave of the Study of Dementia
in Swedish Twins found that twins with AD exhibited more dilation of ventricles and
temporal horns, as well as more atrophy of the temporal lobes, when compared to their
healthy cotwins (Pedersen et al., 1999). Furthermore, dementia in subcortical ischemic
vascular disease has been found to correlate with hippocampal atrophy (Fein et al., 2000).
Frontal lobe atrophy has also been found in studies of individuals with late-onset
depression and in studies of individuals with dementia. In their study regarding the
neuroanatomical aspects of depression in older adults, Almeida, Burton, Ferrier,
McKeith, and O’Brien (2003) compared 27 older adults experiencing late-onset
depression with 24 older adults with early-onset depression and 37 controls. In this
9
study, late-onset depression was defined as cases where onset of depression occurred at
or after age 60, while early-onset depression occurred before age 60. Almeida et al.
(2003) found that participants with late-onset depression had greater reduction in frontal
lobe volume than participants with early-onset depression and controls. Almeida et al.
(2003) also suggested that changes in brain volume or structure have a role in late-onset
depression. In addition to atrophy of structures in the temporal lobe in patients with
Alzheimer’s disease, van der Flier et al. (2002) found that atrophy also was pronounced
in the frontal lobe.
Finally, while many neuroimaging studies have found that whole brain atrophy
and ventricular dilation are among some of the structural changes that accompany
dementia, these changes also have been seen in several studies of depression (Schweitzer,
Tuckwell, O’Brien, & Ames, 2002). Alexopoulos, Young, and Shindledecker (1992)
compared individuals with late-onset depression (onset after age 60) and individuals with
early-onset depression using CT. They found that not only did the individuals with late-
onset depression have overall larger ventricles than individuals with early-onset
depression, but their ventricle size was comparable to the ventricle size of individuals
with Alzheimer’s disease. Similarly, Rabins, Pearlson, Aylward, Kumar, and Dowell
(1991) found greater sulcal atrophy among individuals with depression onset after age 60.
Some of the most recent research into the neuropathology of depression and
dementia, however, has indicated that LOD shows a stronger neuroanatomical correlation
with white matter changes rather than with cortical atrophy (Kohler et al., 2010; Mueller
et al., 2010; Starkstein et al., 2009). In fact, a study by Mueller et al. (2010) concluded
that depressed mood and cognitive impairment have different pathological correlates,
10
specifically that depressed mood is associated with WMH, whereas cognitive function is
associated with gray matter atrophy.
Cerebrovasular Disease
There is increasing evidence that the pathenogenesis of white matter lesions may
be largely due to cerebrovascular disease (CVD). One study (Jefferson et al., 2007)
indicated that cardiac insufficiency, one of the sequelae of cardiovascular disease, results
in a hypoperfusion (decreased blood flow) of blood in subcortical regions of the brain.
Fronto-subcortical loops and the associated white matter tracts seem to be especially
vulnerable to hypoperfusion and ischemia (Chui, 2007). Thus, it would follow that even
in the absence of a stroke, brain damage seen as white matter changes could occur
secondary to the systemic hypoperfusion and ischemia caused by CVD risk factors. In
conjunction with the vascular depression hypothesis (Alexopoulos et al., 1997), it would
stand to reason that CVD be considered as a potentially important covariate in the
relationship between late-onset/late-life depression and white matter pathology.
Specific Aim and Study Hypotheses
The aim of the present study was to examine whether there were neuroanatomical
differences evident on the CT scans of individuals with dementia based upon depression
onset (no depression versus early-onset versus late-onset), and whether there was a
history of late-life depression (any episode of depression after age 60). We hypothesized
that individuals with dementia and LOD and/or LLD would be more likely than non-
depressed individuals with dementia to exhibit hypodensities in the frontal lobe deep
white matter and subcortical white matter (subinsular region and internal capsule), as
11
well as the subcortical striatal gray matter (caudate nucleus and lentiform nucleus, which
includes the putamen and globis pallidus), and display a greater degree of frontal lobe
cerebral atrophy (as measured by frontal ventricular width) on CT scans. We
hypothesized that these patterns would hold for total dementia and when Alzheimer’s
disease or vascular disease were considered separately. CVD comorbidities have been
shown to strongly correlate with neurovascular pathology and have a presumed role in
vascular depression. Therefore, we considered CVD as a potentially important covariate.
In addition, because of the cross-sectional study design, particular attention was paid to
the severity and duration of dementia at the time of the CT scan.
12
Chapter 2: Methods
Participants
Participants in the study were from the Study of Dementia in Swedish Twins
(HARMONY; Gatz et al., 2005). The HARMONY study population included all the
twins in the Swedish Twin Registry (Lichtenstein et al., 2002) who were aged 65 and
older and alive during the telephone-screening phase. The sample for the present study
(N = 238) consisted of all twins from HARMONY who 1) were diagnosed as having
dementia using DSM-IV criteria, 2) had a hard copy of their CT scan (performed as part
of the clinical phase of dementia assessment) that could be assessed by the CT raters in
the present study, and 3) did not meet any of the exclusion criteria. (See Figure 1.)
Exclusion criteria included missing information on timing of depressive episodes (N=4)
or had first depression onset more than six months after the scan date (N = 2). It is
important to note that although this study uses data from Swedish twins, it is not a twin
study per se.
13
Figure 1. Study Sample
244 CT Scans Available
4 Missing Age of
Depression Onset
2 Depression Onset After Scan
238 CT Scans Rated
23 Minor Stroke
(not excluded)
Data Analyzed:
182 Rated for WMH
176 Atrophy Measurements
45 MCA/PCA Stroke
2 Hydrocephalus
4 Poor Quality Scan
5 Other Severe Brain
Pathology
14
Dementia Diagnosis
Dementia was ascertained in two phases, a screening phase and a clinical phase.
In the screening phase, cognitive status was assessed over a 2.5-year period starting
March 1998. The TELE cognitive screening instrument (Gatz, et al., 2002; Gatz,
Reynolds, Nikolic, Lowe, & Karel, 1995) was used to screen for cognitive dysfunction.
If an individual performed poorly on the TELE, a relative was subsequently interviewed
with the Blessed Dementia Rating Scale (BDRS; Blessed, Tomlinson, & Roth, 1968) to
determine if the twin’s cognitive status affected functioning on a daily basis. The scores
from the TELE and BDRS were then combined to create a cognitive status score. Those
with cognitive impairment sufficient to affect dailyfunctioning were invited to the clinical
phase, wherein a clinical work up was performed to determine a clinical dementia
diagnosis. Of the 1569 participants who screened positive, 1101 proceeded to the clinical
phase. The remaining 468 did not proceed either by design because the twin’s partner
would not be informative due to having died before age 65 or because they refused. After
the assessment team made a preliminary assessment that a twin was demented, their co-
twin was referred for clinical evaluation, regardless of whether the co-twin had screened
positive or negative for cognitive impairment.
Measures
Clinical evaluations involved a visit by an assessment team, comprised of a nurse
and a physician, typically occurring in the twin’s home. To ensure that the assessment
team was blind to twin zygosity, the same team did not visit both twins in a pair. The
evaluation protocol generally followed the Consortium to Establish a Registry for
Alzheimer’s Disease (CERAD; Morris et al., 1989), including a physical and
15
neurological examination, a complete medical history based on medical record review
and informant interview, a neuropsychological evaluation, a collection of blood for
laboratory tests and DNA extraction for genotyping, and a referral for neuroimaging.
Dementia was clinically diagnosed using the DSM-IV criteria. Differential
diagnoses were made using the NINCDS/ADRDA criteria for Alzheimer’s disease
(McKhann et al., 1984), NINDS-AIREN criteria for vascular dementia (Roman et al.,
1993), Lund and Manchester criteria for frontal temporal dementia (Lund and Manchester
Groups, 1994; Neary et al., 1998), and consensus criteria for dementia with Lewy bodies
(McKeith et al., 1996).
Diagnostic procedure followed three steps. First, based on workup and review of
medical records, the assessment team completed an initial diagnosis. Next, the
assessment information was given to a diagnostic review board, comprised of a
psychologist and neurologist, who independently constructed a clinical dementia
diagnosis, including differential diagnosis of type of dementia. Finally, disagreements
between board members as well as disagreements between assessment teams and board
members were submitted back to the diagnostic review board for resolution, which was
determined by consensus. The diagnosis was made initially without taking neuroimaging
into account, as neuroimaging was not available for everyone. Thereafter, the board
reconsidered the diagnosis in light of neuroimaging (for details, see Johansson,
Fratiglioni, Pedersen, and Gatz, 2005). The initial clinical diagnosis is used in this study.
The HARMONY study used two sources of information to estimate the age of
dementia onset: informant reporting and medical records (Fiske, Gatz, Aadnoy, &
Pedersen, 2005). During the clinical phase, to assess age of onset and course of
16
dementia, an in-depth, semi-structured interview, designed specifically for the
HARMONY study was administered to an informant by trained nurses. Informants were
usually either adult children or spouses of the proband, had a long history with the
proband, and were in at least weekly contact with the proband. Information was also
extracted from medical records, including the dates of any prior workup conducted to
assess dementia, in which there was an indication of some level of cognitive impairment.
Based on the informant interview and the information from medical records, the
assessment team determined the age of onset for dementia.
Dementia severity. Data pertaining to dementia severity came from the Clinical
Dementia Rating (CDR; Morris et al., 1997) scale, which was completed during the
clinical workup phase based on an informant interview. The CDR includes an overall
dementia severity score (0.5 = very mild dementia, 1.0 = mild dementia, 2.0 = moderate
dementia, 3 = severe dementia), as well as specific scores for the level of impairment in
each of the six domains of cognitive and functional performance, including memory,
orientation, judgment and problem-solving, community affairs, home and hobbies, and
personal care (0 = no impairment, 0.5 = questionable impairment, 1.0 = mild impairment,
2.0 = moderate impairment, 3 = severe impairment). The interviewers were trained to
criterion according to the CDR protocol defined by Morris et al. (1997) using videotaped
assessments of subjects in various stages of Alzheimer’s disease. The CDR was tested
for inter-rater reliability using cases that are the most difficult to distinguish and rate
(mild cognitive impairment and mild dementia). Agreement between assessment team
members was 76% (Gatz et al., 2005). CDR data were missing for six individuals from
the no depression group and one individual from the LOD group.
17
Neuroimaging
The referral for neuroimaging was specifically for CT scans because at the time of
the clinical diagnostic assessment not enough participants lived close enough to MRI
centers. A small subset of pairs (n=37) who lived near an academic medical center with
MRI facilities were invited to participate in an MRI study, and did not also provide a CT
scan. Otherwise, since the purpose of the neuroimaging was to aid in diagnosis, only
probands were included in neuroimaging, and not their non-demented co-twin. Twins
were instructed to have their scan done at whichever clinic was the most convenient. To
ensure that the CT scans would be comparable, the CT technicians were all given
standardized instructions. They were to perform a CT of the brain with standard slices
and non-contrast enhancement. The slices were to be four to five millimeters from the
base of the skull and eight to ten millimeters from the pars petrosa ossis temporalis.
Contrast enhancement was performed if it was needed for clinical reasons. Finally, the
diagnostic protocol allowed individuals who had a CT within six months of the clinical
work-up to provide a copy of that scan instead.
Scans were read clinically so that participants could be provided with any
medically relevant findings, such as evidence of a brain tumor. Thereafter, scans were
deidentified and made available for research purposes.
Visual Rating of CT Scans
Drs. Bryan Spann (USC; Department of Neurology) and John Go (USC;
Department of Neuroradiology) served as the CT raters. Raters were blind to clinical
diagnosis and any demographic information, including which scans were from twin pairs,
18
as well as the age, gender, and zygosity of the scanned individual. Visual rating
worksheet is presented in Appendix A.
Since these scans were acquired throughout Sweden in different hospitals, there is
some inconsistency in image quality. To account for this, the raters were asked to
provide an image quality score for each CT, which took into consideration factors such as
image clarity, the individual’s location in the scanner, whether the participant’s head was
tilted during the scanning process, and whether the images are either too dark or
overexposed. If either rater determined that an image’s quality was unacceptable, it was
excluded from the analyses.
The raters assessed each CT scan for the number and location of white matter
hypodensities (WMH), the number and location of hypodensities in the basal ganglia and
thalamic region, and the level of ventricular atrophy in the frontal, temporal, parietal, and
occipital lobes. Additionally, the raters were asked to determine whether there is
evidence of either stroke or hydrocephalus. If there was evidence of a major stroke, such
as middle cerebral artery or posterior cerebral artery stroke, the CT was not included in
the analyses. Individuals with small lacunar infarcts, however, were not excluded.
Hydrocephalus was considered to be present when the ratio of the anterior third to the
posterior third of the ventricles exceeds one; ratios less than one are typically indicative
of atrophy as opposed to hydrocephalus. CT scans that indicated hydrocephalus were
also excluded from the analyses. Finally, scans that indicated other severe
neuropathology, such as evidence of a brain tumor, major brain surgery, or traumatic
brain injury, were also excluded from the analyses.
19
Frontal deep white matter hypodensities. As recommended by Hachinski et al.
(2006), cortical and subcortical WMH were rated and counted separately for the left and
right hemispheres. Peri-ventricular, peri-atrial, and peri-occipital WMH were also rated
separately for each hemisphere. While WMH or capping around these regions may be
indicative of age-related changes, to be consistent with the recommendations of
Hachinski et al. (2006), the raters were not asked to distinguish between changes related
to age versus changes of a more pathological nature.
WMH that were not located in the peri-ventricular, peri-atrial, or peri-occipital
regions were rated on a modified version of the Age-Related White Matter Changes
(ARWMC) Scale (Wahlund et al., 2001), where 0 = an absence of hypodensities, 1 = one
focal hypodensity (≥5 mm), 2 = more than one focal hypodensity, 3 = confluent
hypodensities, and 4 = confluent hypodensities with additional discrete focal
hypodensities. Focal hypodensities were defined as discrete hypodensities greater than
five millimeters in size. Confluent hypodensities were present when discrete
hypodensities could not be separately defined. Location of WMH were recorded as being
either subcortical (beneath the gyri) or in the deep white matter (located medially from
the sulci). Thus, deep white matter hypodensities (DWMH) were defined as WMH that
were located medially from the sulci. (During the rating process, the term “subcortical”
referred to cortical WMH immediately beneath the gyri. However, for the purposes of
the present study, the term “subcortical” will refer to the basal ganglia and thalamic
regions of the brain.) Lobe (frontal, temporal, parietal, occipital, or global) and
hemisphere (right, left, or equal) where WMH were dominant was also indicated.
20
Spearman’s rank correlation indicated that the ratings of frontal DWMH in the
right and left hemisphere were very similar among the participants in the present study
(Spearman's rho = 0.92, p < 0.0001). Therefore, scores for frontal DWMH were created
by comparing the scores for the right and left hemispheres and utilizing whichever score
was greater.
Basal ganglia-thalamic hypodensities. The raters also examined each scan for
white matter and gray matter hypodensities in the basal ganglia and thalamic region.
During the rating process the homogeny of the gray matter in the striatum (caudate
nucleus and lentiform nucleus), substantia nigra, and thalamus were compared to one
another in order to determine if there was evidence of hypodensities within these regions.
Assessment included separate counts of the number of hypodensities present in the
caudate nucleus and the lentiform nucleus for both right and left hemispheres. By adding
the number of hypodensities in these two regions bilaterally we were able to calculate the
total number of hypodensities in the striatum.
Subcortical white matter hypodensities. The density of the white matter in the
internal capsule and subinsular region (comprised of the external capsule, claustrum, and
extreme capsule) was compared to the density of homogenous areas of white matter in
the frontal lobe to determine if there were relative hypodensities evident in these two
components of the basal ganglia white matter. The location, side (left vs. right
hemisphere), and rating of the hypodensities within these regions were indicated, with
hypodensities rated on the modified version of the ARWMC Scale.
Spearman’s rank correlation indicated that the ratings of subcortical WMH in the
right and left hemisphere were significantly similar among the participants in the present
21
study (Spearman's rho = 0.79, p < 0.0001). Therefore, scores for subcortical WMH were
created by comparing the scores for the right and left hemispheres and utilizing
whichever score was greater.
Frontal ventricular width. Ventricular width was assessed as an approximation of
cortical atrophy. Ventricular width in the frontal lobe was calculated by measuring the
maximum width of the most anterior portion of the frontal horn for both the right and left
hemispheres. Each measurement was made on the slice which displayed the frontal horn
at the maximum level. Then, this measurement was converted into millimeters using the
scale included on the scan.
Inter-rater reliability. Each CT scan was scored separately by both raters using
the measurement guidelines described above. In cases of disagreement, the scans were
re-rated conjointly, and a consensus rating was established by the two raters. We
evaluated inter-rater reliability for each of the three major CT variables (WMH, basal
ganglia-thalamic hypodensities, and frontal ventricular width) using their independent
ratings. Inter-rater reliability for degree of bilateral WMH was very substantial
(weighted kappa, right WMH = 0.92; weighted kappa, left WMH = 0.89) and slightly
lower though adequate for bilateral basal ganglia-thalamic hypodensity ratings (weighted
kappa, right = 0.73; weighted kappa, left = 0.62). Intraclass correlation (ICC) indicated
high inter-rater concordance in measurement of both right and left frontal lobe ventricular
width (right, ICC coefficient = 0.81; left, ICC coefficient = 0.82).
22
Depression
History of depression, co-morbid depression, and estimated age of onset of
depression were determined by combining information from four sources: 1) the national
computerized Inpatient Discharge Registry (IDR), 2) the national registry of inpatient
psychiatric hospital services, 3) medical history given by an informant, and 4) medical
records.
Inpatient Discharge Registry. The Swedish Twin Registry is linked to the
national computerized Inpatient Discharge Registry that records inpatient discharges
every time any individual leaves a hospital in Sweden. The Inpatient Discharge Registry
was begun in 1964, but 100% coverage throughout Sweden was not achieved until 1987,
three years after reporting to the Inpatient Discharge Registry became mandatory.
Because of the Swedish universal health care system, almost all hospital care is public,
and Sweden’s National Board of Health and Welfare estimates that about 99% of all
public hospitalizations are included in the IDR registry (Centre for Epidemiology, 1997).
The discharge diagnosis is given in terms of an International Classification of Disease
(ICD) code. If the discharge date was prior to 1969, the ICD-7 coded depression
diagnoses as 302 (involutional melancholia), 314 (depressive neurosis), and 790.2 (other
recurrent depressive disorder). If the discharge date was 1969-1986, the ICD-8 coded
depression diagnoses as 296.0 (involutional melancholia), 298 (reactive depressive
psychosis), 300.4 (depressive neurosis), and 790.2 (other recurrent depressive disorder).
If the date of discharge was 1987-1996, the ICD-9 coded depression diagnoses similar to
the ICD-8, with 296.2 (depressive psychosis), 296.3 (recurrent depressive psychosis),
296.82 (atypical depression), 300.4 (dysthymia), and 311 (depression, NOS). Finally,
23
beginning in 1997, the ICD-10 depression diagnoses included F32 (depressive episode),
F33 (recurrent depressive disorder), and F34.1 (dysthymia). There were 11 twins with
discharge diagnoses of depression in the Inpatient Discharge Registry.
Inpatient psychiatric hospital services registry. The Swedish Twin Registry has
also been linked to a national registry of inpatient psychiatric hospital services that was
maintained between 1967 and 1983 and then discontinued. For each person entered in
this registry, there is a record of the discharge diagnosis and the date of hospitalization.
All diagnoses are given in terms of an ICD-8 diagnosis (see above). There were two
twins with diagnoses of depression in the inpatient psychiatric hospital services registry,
both of whom also had a diagnosis of depression in the Inpatient Discharge Registry.
Thus, a total of 11 participants in this study had at least one depression-related
discharge diagnosis in the IDR between the years of 1964 and 2004, a span of 40 years,
or between 1967 and 1984 in the inpatient psychiatric hospital services registry. This
indicates that in this sample there was a 6.04% morbid risk of being hospitalized for
depression, which is more than twice the morbid risk estimation (2.96%) for the entire
STR population (Kendler, Pedersen, Johnson, Neale, & Mathe, 1993). The most
common depression-related discharge diagnosis was depression, not otherwise specified,
followed by dysthymia, recurrent depressive disorders, and depressive episodes of mild,
severe, or other characteristics. Discharge diagnoses in both the IDR and the psychiatric
hospital discharge registry that were regarded as not depression-related included bipolar
affective disorder and manic-depressive reaction, manic or unspecified type,
schizoaffective disorder, and unspecified mood disorders.
24
Medical history. Medical history reported by an informant was collected during
the clinical evaluation. The history includes whether the individual had any history of
“major depressive disorder” or “reactive depression,” and if so, the date or dates of onset.
Also included in medical history is information regarding the use of an antidepressant
medication. If the medical history indicated any depressive disorder or the use of
antidepressants, the participant was considered to have a history of depression. The first
depressive episode according to the medical history was determined by taking the earliest
date recorded for either the onset of a depressive disorder or the start date for initiating
the use of an antidepressant medication. The most recent episode of depression was
determined by taking the latest date recorded for either a depressive episode or the use of
antidepressant medication. There were 37 individuals for whom an informant indicated a
history of depression. Of these individuals, 14 were reported to have both a diagnosis of
depression and a history of using antidepressants, 7 had a diagnosis of depression, but no
history of antidepressant use, and 16 had a history of antidepressant use, but were not
reported as having a history of a depression diagnosis.
Medical records. Medical records, ordered during the clinical evaluation phase,
were coded by the assessment team to reflect whether the twin had been diagnosed with
depression. Records typically go back approximately ten years or until the general time
of onset of symptoms of cognitive dysfunction. Thus, these records were most helpful in
knowing whether there was depression just prior to the dementia or concurrently with the
dementia. Data include the onset and dates of the illnesses. There is also information
regarding the use of antidepressive medication and the date of the first recorded use, as
well as information on later, and in most cases, most recent use of an antidepressant.
25
According to medical records, 30 twins had a diagnosis of depression and were also on
antidepressants, 11 twins had a diagnosis of depression and were not on antidepressants,
and 32 twins did not have a record of depression diagnosis, but were recorded as being on
antidepressants.
Combining Sources of Depression
The assessment of depression was based upon clinician report (IDR, psychiatric
IDR, and medical records) and informant-reported medical history. Prior studies
examining comparability of information on depression history in individuals with
dementia have reported adequate concordance between these sources of information. For
example, Teri and Wagner (1991) showed that for depressed individuals with AD there
were no differences between informant and clinician report of dementia and that accuracy
did not differ by dementia severity. In addition, across various neuropsychiatric
problems in dementia, Victoroff, Nielson, and Mungas (1997) showed that caregiver and
clinician inter-rater reliability was highest for depression.
Of the 81 participants with a history of depression, only eleven had ever been
hospitalized for depression. All eleven of these individuals also had evidence of a history
of depression in their medical records and ten individuals had informant-reported
depression history. Among the remaining 70 participants who had never been
hospitalized for depression, eight had evidence of a history of depression only in their
informant-reported medical history, 40 had evidence of depression only in their medical
records, and 22 had evidence of depression history in both their medical records and
informant-reported medical history. In three of the cases where an informant indicated a
history of depression, but the medical records did not, it was because the depression
26
occurred prior to earliest available medical record. In sum, these numbers are consistent
with the fact that most people who have a history of depression are not hospitalized for
the disorder, and therefore will only have evidence of a history of depression in a source
that is not limited to information from hospitalization records. Additionally, it is possible
that depression is underreported in informant-reported medical history.
Definition of history of depression diagnosis. There were 54 people who had data
in their medical records and/or medical history indicating that they had taken
antidepressants (20 in their medical records only, 22 informant report only, and 12
according to both medical records and informant report). Of these individuals, 23 had at
least one other source corroborating a diagnosis of depression (six were in the IDR, 17
had a depression diagnosis in their medical records, and 19 had an informant-reported
diagnosis of depression). Of the 31 remaining individuals, 19 had taken antidepressants
only according to the medical records, four had taken antidepressants only according to
an informant report, and eight had taken antidepressants according to both medical
records and an informant report.
There were 31 individuals who were coded as having a history of depression
based solely upon having taken antidepressants. Thus, it is possible that these individuals
took these medications for reasons other than depression. Therefore, we defined a history
of depression using both a narrow and a broader definition. The broad definition
included the individuals coded as depressed based solely upon having taken
antidepressants and the narrow definition did not include these individuals. All analyses
were run using both definitions of depression history.
27
Timing of Depression
Early-onset and late-onset depression. The age of the first episode of depression
was determined by discerning the age at the earliest reported occurrence of depression
across all sources. Individuals with a first episode of depression prior to age 60 were
considered to have early-onset depression (EOD). Individuals with a first episode of
depression at age 60 or older were considered to have late-onset depression (LOD). The
narrow definition of depression classified 36 people as having late-onset depression and
14 people as having early-onset depression. The more inclusive, broad definition coded
66 individuals as having LOD and 15 individuals having EOD.
Late-life depression. Across all sources of depression history data, we were able
to assess for the presence of later episodes of depression, including the most recent
episode of depression. Individuals who had at least one episode of depression after age
60 were considered to have a history of late-life depression (LLD). This number
represents a combination of those with LOD and those with EOD who experienced a late-
life recurrence of depression. Narrowly defined depression coded 45 individuals with
LLD. The broad definition of depression identified 75 individuals as having LLD. Nine
individuals with EOD had a recurrence of depression after age 60 for both narrowly and
broadly defined depression. Individuals with EOD but without an identified episode of
depression after age 60 were not considered to have a history of late-life depression
(narrow definition: N = 5; broad definition: N = 6) and were not included in analyses
involving history of LLD.
28
Cerebrovascular Disease
CVD has been shown to correlate with a greater amount of white matter and
subcortical gray matter lesions and is the proposed pathological mechanism behind the
theory of vascular depression. Furthermore, it has been shown that AD and CVD
pathology often coexist (Jagust et al., 2008). Not surprisingly, white matter pathology
has been shown to be frequently associated with cerebrovascular disease risk factors,
such as hypertension, diabetes, atrial fibrillation, peripheral artery disease, transient
ischemic attack, and coronary artery disease (CAD) indicators, which include myocardial
infarction, angina, and heart failure. Thus cerebrovascular disease risk factors, as
indicators of CVD, were examined as potential confounders or effect modifiers. Data on
hypertension, diabetes, atrial fibrillation, peripheral artery disease, transient ischemic
attack, and coronary artery disease indicators came from the participant’s medical records
(see above section entitled “Medical Records”). Coronary artery disease was considered
present if the individual had a history of myocardial infarction, angina, or heart failure.
Other Covariates
In addition to CVD, covariates included age at the time of the scan, dementia
duration (i.e., age at the time of the scan compared with age of dementia onset), gender,
zygosity, and number of years of education completed. Zygosity was determined by a
developed algorithm that was based on the answers to questions such as, “During
childhood, were you and your twin partner as like as ‘two peas in a pod’ or not more
alike than siblings in general” and “How often did strangers have difficulty in
distinguishing between you and your twin partner when you were children.” The
algorithm was validated by testing 13 DNA markers on 199 adult pairs, proving to be
29
correct in 99% of the pairs. Only one pair classified as MZ was misclassified
(Lichtenstein et al., 2002).
Pre-rating Exclusions
All participants whose CT scans were rated had a clinical dementia diagnosis. Six
scans were not rated because either the depression information was incomplete or
because the onset of depression occurred more than six months after the date of the scan.
(See Figure 1.) Of these scans, there were four cases where an informant reported that
the proband was depressed, but did not know when the depression occurred. As there
were no other data endorsing depression and an onset date for these four individuals, they
were excluded from being rated. Also, we determined the time differences between when
the scan was performed and the onset of depression. There were two scans excluded
where depression onset was more than eight months after the scan date.
Post-rating Exclusions
Of the 238 CT scans scored by the raters, 56 were excluded during the rating
process. (See Figure 1.) Four scans were excluded due to the poor quality of the scan, 45
because there was evidence of major stroke, 2 because of hydrocephalus, and 5 because
of other major brain problems (e.g. evidence of major brain surgery or traumatic brain
injury). Those with evidence of a minor stroke were not excluded. The final sample
included white matter hypodensity ratings for 182 individuals. Of these, there were six
scans that did not receive measurements for ventricular width either because the
measurement scale on the scan was not sufficiently precise or because the scale was
missing. Thus, there were 176 individuals with ventricular atrophy ratings.
30
Analyses
For the first steps in the initial model building, the associations between
depression onset (no depression versus LOD versus EOD) and history of LLD (no
depression versus LLD) and all potential demographic and medical confounders were
examined in order to determine which were appropriate factors to include as covariates in
the final multivariate model. Because the design of the study was cross-sectional,
particular attention was paid to duration of dementia at the time of the CT scan
(calcluated by subtracting age of dementia onset from the age at the time of the CT scan).
Chi-square tests (PROC FREQ) were used to examine whether gender, zygosity,
or risk factors for CVD (hypertension, diabetes, atrial fibrillation, peripheral artery
disease, transient ischemic attack, and coronary artery disease indicators) differed by 1)
depression onset (no depression versus LOD versus EOD), or 2) history of LLD (no
depression versus LLD) using first the narrow and then the broad definitions of history of
depression. One-way ANOVA (PROC GLM) also was used to determine whether age at
CT scan, age of dementia onset, duration of dementia at the time of the CT scan, or level
of education differed by depression onset, and pooled-variance independent group
Student’s t-tests (PROC TTEST) examined whether they differed by history of LLD.
The results of the bivariate analyses including the EOD group are reported for
descriptive purposes only. Due to the small size and probable under-ascertainment of
EOD, multivariate analyses did not examine differences between the EOD and non-
depressed or EOD and LOD groups. Thus for the multivariate analyses, depression onset
only refers to the non-depressed and LOD groups.
31
Kruskal-Wallis chi-square was used to examine whether frontal deep white matter
or subcortical white matter score varied by age at CT scan or dementia duration. Fisher’s
Exact Test was used to determine whether frontal deep white matter or subcortical white
matter score differed by gender. Simple linear regression was used to test whether
number of striatal hypodensities or total frontal ventricular width varied by age, gender,
or dementia duration. We took a conservative approach to determining which potential
confounders should be included in the final model, including covariates with an
association of p < 0.15.
Logistic regression (PROC LOGISTIC) controlling for gender and history of TIA
was used to examine the association between the both the presence and confluence of
frontal deep WMH and LOD and the presence and confluence of frontal deep WMH and
history of LLD. To analyze the relationship between subcortical WMH and LOD and
subcortical WMH and history of LLD, ordinal logistic regression (PROC LOGISTIC)
was used, controlling for age at CT scan and zygosity. The association between striatial
hypodensities and LOD and striatal hypodensities and histoy of LLD was examined using
one-way ANOVA (PROC GLM). Finally, one-way ANOVA (PROC GLM) controlling
for age and gender were used to determine whether ventricular width varied by
depression onset or history of LLD. Analyses were run using both the narrow and
broader definition of depression. Only results for the no depression vs. LOD and no
depression vs. LLD comparisons are reported, as no meaningful differences were
expected to be found with the comparisons using the EOD group.
32
Dementia Severity
We also took into consideration the possibility that differences in dementia
severity between the LOD, EOD, LLD, and non-depressed groups could obscure the
relationship between WMH, striatal hypodensities, frontal lobe atrophy, and LOD or
LLD. Due to the cross-sectional design, it was likely that the participants differed a great
deal in severity at their point of entry into the study. One-way ANOVA (PROC GLM)
were used to determine whether dementia severity (as measured by the CDR)
significantly differed by depression onset or history of late-life depression. Logistic
regression (PROC LOGISTIC) was used to determine whether presence of frontal lobe
deep white matter hypodensities differed by global CDR score. Ordinal logistic
regression was used to examine whether severity of subcortical WMH differed by global
CDR score. Finally, one-way ANOVA (PROC GLM) were used to investigate whether
total number of hypodensities in the striatum differed by global CDR score. In addition,
as dementia duration is likely to be associated with greater dementia severity, we also
used a one-way ANOVA (PROC GLM) to characterize the relationship between
dementia duration and dementia severity in our sample. The final multivariate regression
model included the significant terms shown to be related to dementia severity in the
bivariate analyses.
Correlation Between Twin Pairs
There were nine complete twin pairs in the study sample. There were seven
complete twin pairs included in the analyses comparing the no depression to late-onset
depression groups using both the narrow and the broad definition. Therefore we also
used generalized estimating equations (GEE; PROC GENMOD) to account for the
33
possibility that these seven pairs had correlated data. GEE controlling for gender and
history of TIA were used to examine the association between the presence of frontal deep
WMH and LOD for both the total sample and a sample only including individuals with
Alzheimer’s disease. GEE controlling for age at CT were used to examine the
association between the amount of subcortical WMH and LOD for both the total sample
and a sample only including individuals with Alzheimer’s disease. Finally, GEE analysis
was used to examine the association between number of striatal hypodensities and LOD
for both the total sample the the Alzheimer’s disease only sample. Tables providing
comparisons of the outcomes using logistic regression and GEE analyses are provided in
Appendix B.
CVD Risk Factors and Coronary Artery Disease Indicators
Phi coefficients of correlation (PROC FREQ) were calculated to determine
intercorrelation between CVD risk factors and between the CAD indicators (angina,
myocardial infarction, and heart failure). Assessment of CVD risk factors revealed that
nearly one-third of the sample had at least one indicator of CAD, the onset of which has
previously been shown to be related to a greater risk of subsequent depression onset
among a larger sample of individuals in the STR (Kendler, Gardner, Fiske, & Gatz,
2009). To explore this relationship further in our sample, chi-square (PROC FREQ) tests
were used to determine whether frequency of CAD differed between the EOD, LOD,
LLD, and non-depressed groups. Logistic regression (PROC LOGISTIC) was used to
determine whether presence of frontal lobe deep WMH differed by history of CAD.
Ordinal logistic regression was used to examine whether severity of subcortical WMH
34
differed by history of CAD. Finally, one-way ANOVA (PROC GLM) was used to
investigate whether total number of hypodensities in the striatum differed by CAD.
Statistical Software
All analyses were conducted using Statistical Analysis System version 9.2 (SAS
Institute, Cary, NC). For multivariate analyses, a p-value less than 0.05 was considered
to be statistically significant.
35
Chapter 3: Results
The most frequent type of dementia for the 182 individuals included in the present
study was Alzheimer’s disease (N = 127, 69.8%), followed by vascular dementia (N =
29, 15.9%) and dementia not otherwise specified (N = 16, 8.8%). An additional five
participants (2.7%) were diagnosed with dementia of a mixed type and another five
(2.7%) individuals had other forms of dementia (e.g. frontal temporal dementia).
Average duration of dementia (from dementia onset to date of CT scan) was 5.2 years
(SD = 4.25, range 0 – 23 years). The majority of participants were female (n = 116,
63.7%) and the average age at the time of the CT scan was 80.6 years (SD = 6.7 years,
range 56 – 96 years). Participants had 7.4 years (SD = 2.4) of education on average, and
51 participants (28.0%) were from a monozygotic twin pair.
Demographic characteristics are presented by dementia group (all dementia,
Alzheimer’s disease, and vascular dementia) in Table 1. Compared to individuals with
vascular dementia, individuals with Alzheimer’s disease were more likely to be a
member of a monozygotic twin pair (RR = 4.57, 95% CI = 1.17, 17.82, p = 0.0072).
Individuals with vascular dementia were more likely than individuals with Alzhiemer’s
disease to have a history of a TIA (RR = 3.65, 95% CI = 1.19, 11.15, p = 0.0179).
Correlation coefficients between the demographic characteristics of the sample are
presented in Table 2. Gender and age at CT scan were significantly associated (r = 0.22,
p = 0.0028), with a greater proportion of women as age increased. Dementia duration
and dementia severity were also significantlhy assoicated (r = 0.22, p = 0.0046), with a
longer duration of dementia associated with a higher level of severity. However, age at
36
Table 1
Sample Demographics and Covariates by Dementia Type
CDR = Clinical Dementia Rating Scale; PAD = peripheral artery disease; CAD = coronary artery
disease; TIA = transient ischemic attack; CVD = cerebrovascular disease; AD = Alzheimer’s
disease; VaD = vascular dementia.
All Dementia
Alzheimer’s
Disease
Vascular
Dementia
AD vs.
VaD
N 182 127 29
Age at CT Scan (SD)
80.6 (6.7) 80.8 (6.9) 81.7 (6.6) n.s.
Duration of Dementia (SD)
5.2 (4.2) 5.1 (4.2) 5.0 (4.9) n.s.
Global CDR, mean (SD) 1.87 (0.87) 1.91 (0.86) 1.82 (0.92) n.s.
Female, n (%) 116 (63.7) 86 (67.7) 19 (65.5) n.s.
Education, years (SD) 7.4 (2.4) 7.5 (2.6) 7.0 (1.2) n.s.
Monozygotic, n (%) 51 (28.0) 40 (31.5) 2 (6.9) p = 0.0070
Hypertension, n (%) 153 (84.1) 104 (81.9) 26 (89.7) n.s.
Diabetes, n (%) 43 (23.6) 24 (18.9) 9 (31.0) n.s.
Arrhythmia, n (%) 36 (19.8) 25 (19.7) 5 (17.2) n.s.
PAD, n (%) 4 (2.2) 2 (1.6) 1 (3.4) n.s.
CAD, n (%) 59 (32.4) 38 (30.0) 10 (34.5) n.s.
TIA, n (%) 14 (7.7) 6 (4.7) 5 (17.2) p = 0.0175
Any CVD, n (%) 166 (91.2) 113 (89.0) 27 (93.1) n.s.
37
Table 2
Correlation Coefficients between Demographic Characteristics of Sample (N = 176)*
Variables 1 2 3 4 5 6
1. Age at CT Scan
−
2. Dementia Duration .09
−
3. Dementia Severity .10 .21**
−
4. Gender .22** -.01 .09
−
5. Years of Education .00 .11 .04 -.03
−
6. Monozygosity .03 -.08 .09 .10 .14
−
*Data regarding dementia severity were only available for 176 individuals. ** p < .01.
38
the time of the CT scan was not correlated with either dementia duration or dementia
severity.
Associations Between Imaging Data and Sample Demographic Characteristics
Frontal Lobe and Subcortical White Matter Hypodensities. Across all study
participants, 39.6% (N = 72) had at least one hypodensity in the frontal lobe deep white
matter and 51.6% (N = 94) had at least one subcortical white matter hypodensity.
Among the 94 participants with subcortical white matter hypodensities, the majority (N =
88, 93.6%) had at least one hypodensity in the subinsular region, with only 6 individuals
having hypodensities solely within the internal capsule, and 30.8% (N = 29) of these
individuals having hypodensities both in the subinsular region and the internal capsule.
Kruskal-Wallis chi-square analyses indicated that age at the time of the scan,
duration of dementia, also at the time of the scan, and education were not associated with
frontal deep white matter hypodensities. (See Table 3.) A Fisher’s Exact Test indicated
that male gender was associated with more frontal deep white matter hypodensities, and
as the association indicated that p < 0.15, gender was therefore included as a covariate in
the final logistic regression analysis. Zygosity was not associated with presence of
frontal lobe deep WMH. A Kruskal-Wallis chi-square indicated that age at the time of
the CT scan was significantly associated with subcortical white matter hypodensities,
such that older individuals had more subcortical WMH (χ
2
= 1.59, df = 3, p = 0.0251). A
chi-square test indicated that individuals who were monozygotic twins were less likely
than non-monozygotic twins to have subcortical WMH (χ
2
= 10.43, df = 3, p = 0.0153).
Gender, education, and duration of dementia were not associated with subcortical white
39
Table 3
Association between Demographic Variables and CT Measures
DWMH = deep white matter hypodensities; WMH = white matter hypodensities
*p-value from Fisher’s exact test. Associations significant at p < 0.15 are bolded.
Dependent Variable Independent Variable Χ
2
df p
Frontal DWMH Age at CT Scan 1.59 3 0.6606
Male Gender - - 0.0539*
Dementia Duration 1.23 3 0.7447
Subcortical WMH Age at CT Scan 9.34 3 0.0251
Male Gender - - 0.9000*
Dementia Duration 0.32 3 0.9548
Monozygosity 10.43 3 0.0153
β t(181) p
Striatal Hypodensities Age at CT Scan 0.01 1.09 0.2785
Gender 0.06 0.37 0.7138
Dementia Duration -0.01 -0.27 0.7912
β t(175) P
Frontal Ventricular Width Age at CT Scan 0.18 2.22 0.0274
Gender -2.23 -2.03 0.0442
Dementia Duration 0.17 1.38 0.1696
40
matter hypodensities. Thus, age at the time of the CT scan and zygosity were included as
covariates in the multivariate models.
Striatal Hypodensities. Across all study participants, 30.8% (N = 56) had at least
one hypodensity in the striatum. A simple linear regression analysis indicated that age at
the time of the scan, education, and the duration of dementia were not associated with the
number of striatal hypodensities. One-way ANOVA indicated that gender and zygosity
were also not associated with striatial hypodensities. (See Table 3.)
Frontal Ventricular Width. T-test results indicated that ventricular width was
significantly greater in men compared with women for the left frontal lobe (male width =
14.5 mm, SD = 4.13; female width = 13.2 mm, SD = 3.30; t(174) = 2.27, p = 0.0242) but
not for the right frontal lobe (male width = 13.8 mm, SD = 3.95; female width = 12.9
mm, SD = 3.56; t(174) = 1.63, p = 0.1051). Greater total frontal ventricular width was
also significantly associated with an older age at the time of the CT scan (r = 0.1663, p =
0.0274). When fit to a multiple regression model gender (β = -3.01, F = 7.33, p = 0.0075)
and age at CT scan significantly predicted bilateral frontal ventricular width (β = 0.23, F
= 8.18, p = 0.0048). Duration of dementia, education, and zygosity were not associated
with frontal ventricular width. Therefore, gender and age at CT scan were included as
covariates in all analyses with frontal ventricular width as the outcome of interest.
Cerebrovascular Disease
One hundred fifty-three (84.1%) individuals had a history of hypertension, 43
(23.6%) had a history of type II diabetes, 36 (19.8%) had a history of arrhythmia, four
(2.2%) had a history of peripheral artery disease, 59 (32.4%) had at least one coronary
artery disease indicator, and 14 (7.7%) had a history of transient ischemic attack. Only
41
16 (8.8%) individuals had no evidence in their medical records of any of these risk
factors for cerebrovascular disease.
As expected, indicators of CAD (myocardial infarction, angina, and heart failure)
were all significantly intercorrelated. (See Table 4.) Risk factors for cerebrovascular
disease, in general, were not significantly intercorrelated. (See Table 5.) The significant
association between CAD and arrhythmia was likely due to the relationship between
arrhythmia and heart failure (r
φ
= 0.34, p < 0.0001).
Coronary Artery Disease. Using the narrow definition of history of depression,
prevalence of CAD was 1.43 times greater among individuals with LOD compared to
non-depressed individuals (95% CI = 0.92, 22.33, p = 0.1327) and 3.10 times more
common in the LOD group compared to the EOD group (p = 0.0461). Using the broader
definition of history of depression, prevalence of coronary artery disease was 1.83 times
greater among individuals with LOD compared to non-depressed individuals (95% CI =
1.20, 2.77, p = 0.0047) and 3.53 times greater compared to EOD individuals (p = 0.0167).
A comparison of the presence of frontal lobe deep WMH and subcortical WMH, and total
number of striatal hypodensities by coronary artery disease status is presented in Table 6.
A pooled variance t-test indicated that total number of striatal hypodensities did not vary
by history of CAD. Presence of frontal deep WMH and level of subcortical WMH
severity also did not vary by history of CAD.
Transient Ischemic Attack. Using the narrow definition of depression, history of a
TIA was present for 14 individuals in the non-depressed group and no individuals with a
history of depression. When the broader definition of depression was used, history of a
TIA was present for 12 individuals in the non-depressed group and two individuals in the
42
Table 4
Phi Correlation Coefficients between Coronary Artery Disease Indicators (N = 182)
*p < .05. **p < .01. ***p < .001.
Table 5
Phi Correlation Coefficients between Risk Factors for Cerebrovascular Disease (N =
182)
Variables 1 2 3 4 5 6
1. Diabetes
−
2. High Blood Pressure .14
−
3. Arrhythmia .02 .10
−
4. TIA -.01 -.04 .01
−
5. PAD .09 -.04 -.07 -.04
−
6. CAD -.03 -.02 .25*** .02 .06
−
TIA = transient ischemic attack; PAD = peripheral artery disease; CAD = coronary artery
disease.
***p < .001.
Variables 1 2 3
1. Myocardial infarction
−
2. Angina .37***
−
3. Heart failure .16* .20**
−
43
Table 6
Presence of Regional Hypodensities by CAD Status
WMH = white matter hypodensities; CAD = coronary artery disease; OR = odds ratio; CI =
confidence interval
No CAD CAD OR 95% CI p-value
Frontal Deep WMH, n (%) 53 (43.1) 22 (29.3) 0.78 0.41, 1.50 0.4641
Subcortical WMH, n (%) 59 (48.0) 35 (59.3) 1.41 0.79, 2.51 0.2491
df t-value p-value
Striatum, mean (SD) 0.53 (1.07) 0.61 (1.16) 180 -0.47 0.6403
44
LOD group. Controlling for gender, a comparison of the presence of frontal lobe deep
WMH by history of TIA indicated that individuals with a history of TIA were 3.92 (95%
CI = 1.16, 13.24, p = 0.0278) times more likely that individuals without a history of TIA
to have hypodensities in the frontal lobe deep white matter. (See Table 7.) Amount of
subcortical WMH and striatal hypodensities did not vary by history of TIA.
Associations between Imaging Data and Dementia Type
An ordinal logistic regression model controlling for gender indicated that
individuals with vascular dementia were 2.67 times more likely to have higher ranked
amount of frontal deep WMH than individuals without vascular dementia (95% CI =
1.24, 5.76, p = 0.0122). Gender also remained significant and men were 1.85 times more
likely to have a higher ranked amount of frontal deep WMH than women (95% CI = 1.08,
3.61, p = 0.0253). Although individuals with vascular dementia were more likely to have
a higher ranked amount of subcortical WMH, this association was not statistically
significant. Presence of striatal hypodensities and average frontal ventricular width did
not differ according to dementia type.
Depression Status—Narrow Definition
There were 132 (72.5%) individuals who did not have a history of a depression
diagnosis and 50 (28.5%) individuals who had a history of a depression diagnosis. Of
these individuals, 36 had their first depressive episode after age 60, with a mean age of
first depressive episode of 74.2 years (SD = 7.0). Fourteen participants had their first
depressive episode before age 60, with a mean age of first depressive episode of 43.2
years (SD = 13.3). Of these 14 participants, nine also had an episode of depression
45
Table 7
Presence of Regional Hypodensities by History of TIA
WMH = white matter hypodensities; TIA = transient ischemic attack; OR = odds ratio; CI =
confidence interval.
No TIA TIA OR 95% CI p-value
Frontal Deep WMH, n (%) 53 (43.1) 22 (29.3) 3.92 1.16,13.24 0.0278
Subcortical WMH, n (%) 85 (50.6) 9 (64.3) 1.57 0.49, 5.02 0.4442
df t-value p-value
Striatum, mean (SD) 0.55 (1.10) 0.57 (1.16) 180 -0.06 0.9536
46
occurring after age 60 (mean age = 74.3 years, SD = 9.6). Thus, 45 of the 182
participants were considered to have late-life depression.
Table 8 presents the results of the bivariate analyses for each of the demographic
characteristics by depression onset group. Associations significant at the p < 0.15 level
are bolded in the table. Group comparisons by ANOVA indicated that 1) the EOD group
was on average significantly younger at the time of the CT scan compared with the non-
depressed group (β = 6.18, t(146) = 3.36, p = 0.0009) and the LOD group (β = 6.06, t(50)
= 3.36, p = 0.0037), and 2) the average age of dementia onset in the EOD group was
significantly earlier compared with the non-depressed group (β = 8.76, t(146) = 4.30, p <
0.0001) and LOD group (β = 9.68, t(50) = 4.24, p < 0.0001). Average age of dementia
onset was significantly earlier in the EOD group compared to both the non-depressed
group (β = 8.76, t(146) = 4.30, p < 0.0001) and the LOD group (β = 9.68, t(50) = 4.24, p
< 0.0001). These differences were neither observed between the non-depressed and LOD
groups nor between the non-depressed and the LLD groups (with the latter results not
shown in the table). A chi-square analysis indicated that participants in the LOD group
were more likely to be a member of a monozygotic twin pair compared to participants in
the non-depressed or EOD groups. Chi-square analyses (for categorical variables) and
one-way ANOVA (for continuous variables) indicated that gender and years of education
did not significantly differ by either depression onset (non-depressed versus LOD) or
history of LLD (latter results not shown in table).
47
Table 8
Sample Demographics by Depression Group
No
Depression
LOD EOD p-value
*
N 132 36 14
Age at CT Scan (SD)
**
81.1 (6.1) 81.0 (7.0) 75.0 (9.3) 0.0039
Age at Dementia Onset (SD)
***
75.9 (6.9) 76.9 (6.7) 67.2 (10.8) 0.0001
Dementia Duration, yrs (SD)
****
5.2 (4.2) 4.2 (3.5) 7.8 (5.5) 0.0249
% Female 63.6 63.9 64.3 0.9986
Education, years (SD) 7.5 (2.4) 7.4 (2.8) 6.9 (1.6) 0.6390
Monozygotic, n (%) 33 (25.0) 15 (41.7) 3 (21.4) 0.1211
Hypertension, n (%) 113 (85.6) 30 (83.3) 10 (71.4) 0.3834
Diabetes, n (%) 32 (24.2) 6 (16.7) 5 (35.7) 0.3451
Arrhythmia, n (%) 25 (18.9) 8 (22.2) 3 (21.4) 0.1842
PAD, n (%) 3 (2.3) 1 (2.8) 0 (0.0) 0.8293
CAD, n (%) 41 (31.1) 16 (44.4) 2 (14.3) 0.1008
TIA, n (%) 14 (10.7) 0 (0.0) 0 (0.0) 0.0566
Associations significant at p < 0.15 are bolded.
LOD = late-onset depression; EOD = early-onset depression; PAD = peripheral artery disease; CAD =
coronary artery disease; TIA = transient ischemic attack.
*Group comparisons by ANOVA for continuous variables and by chi-square for categorical variables.
**EOD group significantly younger at age of CT scan compared with non-depressed group (p = 0.0009)
and LOD group (p = 0.0037). No significant difference between non-depressed and LOD groups. ***EOD
group had significantly earlier dementia onset than non-depressed group (p < 0.0001) and LOD group (p <
0.0001). No significant difference between non-depressed group and LOD groups. ****EOD group had
significantly longer dementia duration than non-depressed group (p = 0.0291) and LOD group (p = 0.0066)
but not significantly longer than non-depressed group. No significant difference between non-depressed
group and LOD groups.
48
Duration of Dementia
Across study participants, the mean time since dementia onset at the time of the
CT scan was 5.2 years (SD = 4.2). A one-way ANOVA indicated significant differences
between the no depression, LOD group, and EOD group with respect to duration of
dementia at the time of the CT scan. At the time of the CT scan those with LOD had the
shortest duration of dementia at 4.2 years (SD = 3.5), followed by individuals with no
history of depression at 5.2 years (SD = 4.2). Individuals with EOD, on average, had 7.8
years (SD = 5.5) between the time of dementia onset and when they were scanned. When
compared separately, there was a significant difference in dementia duration both
between the EOD and non-depressed group (β = 2.59, t(146) = 2.20, p = 0.0291) and
between the EOD and LOD group (β = 3.62, t(50) = 2.75, p = 0.0066), but not between
LOD and non-depressed (β = -1.03, t(168) = 1.31, p = 0.1909).
Cerebrovascular Disease Risk Factors
Chi-square analyses of cerebrovascular disease risk factors by depression group
indicated that CAD and presence of transient ischemic attack differed by depression
group. (See Table 8.) Individuals in the non-depressed group were significantly more
likely to have had a transient ischemic attack than individuals in the LOD group (p =
0.0419) but not when compared to individuals in the EOD group. Individuals with LOD
were more likely than individuals with EOD (p = 0.0563) and non-depressed individuals
(RR = 1.43, 95% CI = 0.92, 22.33, p = 0.1327) to have a history of CAD. CAD was not
more prevalent, however, in the LLD group compared to the no depression group.
49
Hypothesis 1
Compared to individuals with dementia and no history of depression, individuals
with dementia and LOD or dementia and LLD will have a greater amount of frontal lobe
deep WMH, subcortical WMH, and striatal hypodensities.
White Matter Hypodensities and Timing of Depression (Narrow Definition)—All
Dementia
Among individuals with LOD, if there were hypodensities, they were generally
rated as confluent. There was only one individual rated as having one focal hypodensity
in the frontal deep white matter and no one in the LOD group was rated as having more
than one focal hypodensity. Therefore, we analyzed frontal DWMH as a dichotomous
variable—presence (one, more than one, or confluent) versus absence of frontal DWMH.
Controlling for gender and history of TIA, no association was observed either between
frontal DWMH and LOD (non-depressed vs. LOD—OR = 0.83, 95% CI = 0.38, 1.79, p =
0.5161) or frontal DWMH and LLD (non-depressed vs. LLD—OR = 0.78, 95% CI =
0.39, 1.59, p = 0.9237). (See Table 9.)
We also tested to see whether LOD was associated with confluent hypodensities.
We again analyzed frontal DWMH as a dichotomous variable—confluent frontal lobe
DWMH versus non-confluent (none, one focal, more than one focal) frontal lobe
DWMH. Controlling for gender, no significant association was observed either between
confluent frontal DWMH and LOD (non-depressed vs. LOD—OR = 1.58, 95% CI =
0.72, 3.76, p = 0.2572) or confluent frontal DWMH and LLD (non-depressed vs. LLD—
OR = 1.38, 95% CI = 0.66, 2.88, p = 0.3882). (See Table 10.)
Table 9
Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset Depression and Late-Life Depression compared with
No Depression for All Dementia, Alzheimer’s Disease, and Vascular Dementia
-WMH = Frontal deep white matter hypodensities absent; +WMH = Frontal deep white matter hypodensities present; OR =
Binary logit odds ratio estimate, controlling for gender; CI = confidence interval; AD = Alzheimer’s disease; VaD = vascular
dementia.
*Late-onset depression group compared to no depression group. **Late-life depression group compared to no depression
group.
No Depression Late-Onset Depression Late-Life Depression
-WMH
N (%)
+WMH
N (%)
-WMH
N (%)
+WMH
N (%)
OR
(95% CI)
*
-WMH
N (%)
+WMH
N (%)
OR
(95% CI)
**
All Dementia 75 (56.8) 57 (43.2) 22 (61.1) 14 (38.9)
0.96
(0.44, 2.08)
28 (62.2) 17 (37.8)
0.90
(0.44, 1.84)
AD Only 59 (62.1) 36 (37.9) 16 (64.0) 9 (36.0)
0.88
(0.35, 2.25)
17 (60.7) 11 (39.3)
1.02
(0.42, 2.48)
VaD Only 7 (30.4) 16 (69.6) 2 (66.7) 1 (33.3)
0.22
(0.02, 2.83)
4 (80.0) 1 (20.0)
0.11
(0.01, 1.16)
50
Table 10
Risk of Confluent Frontal Lobe Deep White Matter Hypodensities in Late-Onset Depression and Late-Life Depression
compared with No Depression for All Dementia, Alzheimer’s Disease, and Vascular Dementia
-ConWMH = Confluent frontal deep white matter hypodensities absent; +ConWMH = Confluent frontal deep white matter hypodensities
present; OR = Binary logit odds ratio estimate, controlling for gender; CI = confidence interval; AD = Alzheimer’s disease; VaD =
vascular dementia.
*Late-onset depression group compared to no depression group. **Late-life depression group compared to no depression group
No Depression Late-Onset Depression Late-Life Depression
-Con
WMH
N (%)
+Con
WMH
N (%)
-Con
WMH
N (%)
+Con
WMH
N (%)
OR*
(95% CI)
-Con
WMH
N (%)
+Con
WMH
N (%)
OR**
(95% CI)
All Dementia 97 (73.5) 35 (26.5) 23 (63.9) 13 (36.1)
1.58
(0.72, 3.46)
30 (66.7) 15 (33.3)
1.38
(0.66, 2.88)
AD Only 73 (76.8) 22 (23.2) 16 (64.0) 9 (36.0)
1.83
(0.70, 4.79)
17 (60.7) 11 (39.3)
2.12
(0.85, 5.28)
VaD Only 12 (52.2) 11 (47.8) 2 (66.7) 1 (33.3)
0.58
(0.04, 8.15)
4 (80.0) 1 (20.0)
0.29
(0.02, 3.37)
51
52
Controlling for age at CT scan and zygosity, no association was observed either
between subcortical WMH and LOD (non-depressed vs. LOD—OR = 0.97, 95% CI =
0.48, 1.92, p = 0.9188) or subcortical WMH and LLD (non-depressed vs. LLD—OR =
0.97, 95% CI = 0.51, 1.83, p = 0.9315). (See Table 11.) The association between greater
age at the time of the CT scan and a greater amount of subcortical WMH remained
significant (OR = 1.08, 95% CI = 1.01, 1.12, p = 0.0058).
Striatal Hypodensities and Timing of Depression (Narrow Definition)—All Dementia
One-way ANOVA indicated a relationship between total number of striatal
hypodensities and late-onset depression, wherein individuals with late-onset depression
had a greater number of striatal hypodensities compared with non-depressed individuals
(β = 0.38, t(168) = 1.83, p = 0.0693) and individuals with late-life depression had a
greater number of striatal hypodensities compared with non-depressed individuals (β =
0.31, t(177) = 1.67, p = 0.0975). (See Table 12.)
Table 11
Risk of Subcortical White Matter Hypodensities in Late-Onset Depression and Late-Life Depression compared with No
Depression for All Dementia, Alzheimer’s Disease, and Vascular Dementia
OR = Ordinal cumulative logit odds ratio estimate, controlling for age at time of CT scan; CI = confidence interval; AD = Alzheimer’s
disease; VaD = vascular dementia.
Level of WMH: 0 = none, 1 = one discrete hypodensity, 2 = more than one discrete hypodensity, 3 = confluent hypodensities.
* All numbers for Level of WMH given in terms of N (%). **Late-onset depression group compared to no depression group. ***Late-
life depression group compared to no depression group.
No Depression Late-Onset Depression Late-Life Depression
Level of
WMH
*
0 1 2 3 0 1 2 3
OR
(95%
CI)
**
0 1 2 3
OR
(95%
CI)
***
All
Dementia
62
(47.0)
16
(12.1)
25
(18.9)
29
(22.0)
17
(47.2)
3
(8.33)
9
(25.0)
7
(19.4)
0.97
(0.48,
1.92)
21
(46.7)
6
(13.3)
10
(22.2)
8
(17.8)
0.97
(0.51,
1.83)
AD Only
48
(50.5)
11
(11.6)
18
(18.9)
18
(18.9)
12
(48.0)
2
(8.0)
5
(20.0)
6
(24.0)
1.16
(0.48,
2.58)
13
(46.4)
3
(10.7)
6
(21.4)
6
(21.4)
1.12
(0.50,
2.48)
VaD
Only
7
(30.4)
4
(17.4)
4
(17.4)
8
(34.8)
1
(33.3)
-
2
(66.7)
-
0.67
(0.08,
5.93)
2
(40.0)
1
(20.0)
2
(40.0)
-
0.46
(0.08,
2.72)
53
Table 12
Number of Striatal Hypodensities in Late-Onset Depression and Late-Life Depression compared with No Depression for All
Dementia, Alzheimer’s Disease, and Vascular Dementia
LOD = late-onset depression; LLD = late-life depression
*p-value
No Depression LOD No Depression vs. LOD
*
LLD No Depression vs. LLD
*
All Dementia
N 132 36 45
Striatum, mean (SD) 0.48 (0.96) 0.86 (1.53) 0.0693 0.80 (1.45) 0.0975
Alzheimer’s Disease Only
N 95 25 28
Striatum, mean (SD) 0.42 (0.92) 0.96 (1.72) 0.0317 0.86 (1.65) 0.0706
Vascular Dementia Only
N 23 3 5
Striatum, mean (SD) 0.52 (0.79) 0.67 (0.58) 0.7731 0.80 (0.84) 0.4858
54
55
Hypothesis 2
Compared to individuals with Alzheimer’s disease and no depression, individuals
with Alzheimer’s disease and LOD or individuals with Alzheimer’s disease and LLD will
have a greater amount of frontal lobe deep WMH, subcortical WMH, and striatal
hypodensities.
White Matter Hypodensities and Timing of Depression (Narrow Definition)—Alzheimer’s
Disease Only
Among individuals with Alzheimer’s disease, there was no one in the LOD group
who was rated as having one focal hypodensity or more than one focal hypodensity in the
frontal deep white matter. Therefore, we analyzed frontal DWMH as a dichotomous
variable—presence (confluent) versus absence of frontal DWMH. Controlling for gender
and history of TIA, no association was observed either between frontal DWMH and LOD
(non-depressed vs. LOD—OR = 0.88, 95% CI = 0.35, 2.25, p = 0.7935) or frontal
DWMH and LLD (non-depressed vs. LLD—OR = 1.02, 95% CI = 0.42, 2.48, p =
0.7935). (See Table 9.) We also tested to see whether LOD was associated with
confluent hypodensities among individuals with Alzheimer’s disease. We again analyzed
frontal DWMH as a dichotomous variable—confluent frontal lobe DWMH versus non-
confluent (none, one focal, more than one focal) frontal lobe DWMH. Controlling for
gender, among individuals with Alzheimer’s disease, no significant association was
observed either between confluent frontal DWMH and LOD (non-depressed vs. LOD—
OR = 1.83, 95% CI = 0.70, 4.79, p = 0.2151) or confluent frontal DWMH and LLD (non-
depressed vs. LLD—OR = 2.12, 95% CI = 0.85, 5.28, p = 0.1045). (See Table 10.)
56
Controlling for age at CT scan and zygosity, no association was observed either
between subcortical WMH and LOD (non-depressed vs. LOD—OR = 1.16, 95% CI =
0.48, 2.58, p = 0.7977) or subcortical WMH and LLD (non-depressed vs. LLD—OR =
1.12, 95% CI = 0.50, 2.48, p = 0.7847). (See Table 11.) The association between greater
age at the time of the CT scan and a greater amount of subcortical WMH remained
significant (OR = 1.10, 95% CI = 1.04, 1.17, p = 0.0012).
Striatal Hypodensities and Timing of Depression (Narrow Definition)—Alzheimer’s
Disease Only
One-way ANOVA among individuals with Alzheimer’s disease indicated that
individuals with late-onset depression had a significantly greater number of hypodensities
in the striatum compared with non-depressed individuals (β = 0.54, t(120) = 2.17, p =
0.0317; See Table 12). GEE was used to account for the possible correlation between
twin pairs. The relationship between a greater number of striatal hypodensities and LOD
remained significant (OR = 2.27, 95% CI = 1.02, 5.05, p = 0.0441; See Appendix B). A
similar trend was evident when individuals with late-life depression were compared to
non-depressed individuals, although this relationship was not statistically significant (β =
0.44, t(123) = 1.82, p = 0.0706).
According to the bivariate level analyses (see Table 2), there were no associations
between striatal hypodensities and any of the demographic or medical history
confounders. Only CAD differed between the LOD and non-depressed group in the
individuals with AD (χ
2
= 3.89, df = 1, p = 0.0485), with the LOD group more likely to
have CAD than the non-depressed group (48.0% vs. 27.4%, respectively). However,
CAD was not associated with number of total striatal hypodensities. Thus there were no
57
other tested covariates which could explain the association we found between the LOD
group and striatal hypodensities among the AD group.
Hypothesis 3
Compared to individuals with vascular dementia and no depression, individuals
with vascular dementia and LOD or individuals with vascular dementia and LLD will
have a greater amount of frontal lobe deep WMH, subcortical WMH, and striatal
hypodensities.
White Matter Hypodensities and Timing of Depression (Narrow Definition)—Vascular
Dementia Only
Among individuals with vascular dementia, there was no one in the LOD group
who was rated as having one focal hypodensity or more than one focal hypodensity in the
frontal deep white matter. Therefore, we analyzed frontal DWMH as a dichotomous
variable—presence (confluent) versus absence of frontal DWMH. Controlling for
gender, no association was observed either between frontal DWMH and LOD (non-
depressed vs. LOD—OR = 0.22, 95% CI = 0.02, 2.83, p = 0.2445) or frontal DWMH and
LLD (non-depressed vs. LLD—OR = 0.11, 95% CI = 0.01, 1.16, p = 0.2445). (See Table
9.) We also tested to see whether LOD was associated with confluent hypodensities
among individuals with vascular dementia. We again analyzed frontal DWMH as a
dichotomous variable—confluent frontal lobe DWMH versus non-confluent (none, one
focal, more than one focal) frontal lobe DWMH. No significant association was
observed either between confluent frontal DWMH and LOD (non-depressed vs. LOD—
58
OR = 0.58, 95% CI = 0.04, 8.15, p = 0.6887) or confluent frontal DWMH and LLD (non-
depressed vs. LLD—OR = 0.29, 95% CI = 0.02, 3.37, p = 0.3238). (See Table 10.)
No association was observed either between subcortical WMH and LOD (non-
depressed vs. LOD—OR = 0.67, 95% CI = 0.08, 5.93, p = 0.7191) or subcortical WMH
and LLD (non-depressed vs. LLD—OR = 0.46, 95% CI = 0.08, 2.72, p = 0.3886). (See
Table 11.) There was no association between age at the time of the CT scan and amount
of subcortical WMH among individuals with vascular dementia.
Striatal Hypodensities and Timing of Depression (Narrow Definition)—Vascular
Dementia Only
Among individuals with vascular dementia, one-way ANOVA indicated that there
was no relationship between the total number of striatal hypodensities and LOD (β =
0.14, t(26) = 0.29, p = 0.7731), nor was there a relationship between LLD and number of
striatal hypodensities (β = 0.28, t(28) = 0.71, p = 0.4858). (See Table 12.)
59
Hypothesis 4
For all dementia types, Alzheimer’s disease only, and vascular dementia only,
compared to individuals without a history of depression, individuals with LOD or LLD
will have greater frontal ventricular width.
Frontal Ventricular Width and Timing of Depression (Narrow Definition)—All Dementia,
Alzheimer’s Disease Only, and Vascular Dementia Only
One-way ANOVA controlling for age at scan and gender, indicated that across all
types of dementia, right and left frontal ventricular width did not vary by either LOD (p =
0.9024 and p = 0.9348, respectively) or LLD (p = 0.7882 and p = 0.9012, respectively).
(See Table 13.) Among individuals with Alzheimer’s disease there was also no
difference in right and left frontal ventricular width by either LOD (p = 0.8022 and p =
0.7109, respectively) or LLD (p = 0.8903 and p = 0.7971, respectively). Finally among
individuals with vascular dementia there was no difference in right and left ventricular
frontal width by either LOD (p = 0.1359 and p = 0.1072, respectively) or LLD (p =
0.2539 and p = 0.2683, respectively).
Table 13
Right and Left Frontal Ventricular Width (in mm) for All Dementia, Alzheimer’s Disease Only, and Vascular Dementia
LOD = late-onset depression; LLD = late-life depression.
*p-value derived from ANOVA controlling for gender and age at time of CT scan.
No Depression LOD No Dep vs. LOD
*
LLD No Dep vs. LLD
*
All Dementia
N 127 36 45
Right Frontal, mean (SD) 13.25 (3.79) 13.33 (3.57) 0.9024 13.02 (3.66) 0.7882
Left Frontal, mean (SD) 13.70 (3.80) 13.75 (3.39) 0.9348 13.56 (3.36) 0.9012
Alzheimer’s Disease Only
N 93 25 28
Right Frontal, mean (SD) 12.99 (3.79) 13.24 (3.50) 0.8022 13.18 (3.31) 0.8903
Left Frontal, mean (SD) 13.56 (4.14) 13.92 (3.33) 0.7109 13.86 (3.17) 0.7971
Vascular Dementia Only
N 21 3 5
Right Frontal, mean (SD) 14.14 (3.74) 10.33 (3.21) 0.1359 11.80 (5.21) 0.2539
Left Frontal, mean (SD) 14.52 (3.53) 10.67 (3.05) 0.1072 12.40 (4.77) 0.2683
60
61
Depression Status—Broad Definition
There were 101 (55.5%) individuals who did not have a history of depression and
81 (44.5%) individuals who had a history of depression. Of these individuals, 66 had
their first depressive episode after age 60, with a mean age of first depressive episode of
76.3 years (SD = 6.7). Fifteen participants had their first depressive episode before age
60, with a mean age of first depressive episode of 43.0 years (SD = 12.9). Of these 15
participants, nine also had an episode of depression occurring after age 60. Thus, 75 of
the 176 participants were considered to have late-life depression.
Table 14 presents the results of the bivariate analyses for each of the demographic
characteristics by depression group. Associations significant at the p < 0.15 level are
bolded in the table. Group comparisons by ANOVA indicated that 1) the EOD group
was on average significantly younger at the time of the CT scan compared with the non-
depressed group (β = 5.98, t(116) = 3.30, p = 0.0011) and the LOD group (β = 5.98, t(81)
= 2.93, p = 0.0038), and 2) the average age of dementia onset in the EOD group was
significantly earlier compared with the non-depressed group (β = 7.92, t(116) = 3.92, p =
0.0001) and the LOD group (β = 8.34, t(81) = 3.99, p < 0.0001). The average age of
dementia onset in the EOD group was significantly earlier compared to the non-depressed
group (β = 7.92, t(116) = 3.92, p = 0.0001) and compared to the LOD group (β = 8.34,
t(81) = 3.99, p < 0.0001). These differences were neither observed between the non-
depressed and LOD groups nor between the non-depressed and the LLD groups (with the
latter not being shown in the table). A chi-square analysis indicated that the LOD group
had a greater number of years of education compared to the non-depressed group
62
Table 14
Sample Demographics by Depression Group
No Depression LOD EOD p-value
N 101 66 15
Scan Age(SD)
*
81.3 (6.2) 80.8 (6.4) 75.3 (9.0) p = 0.0048
Dementia Onset Age(SD)
**
75.9 (7.2) 76.4 (6.4) 68.0 (10.9) p = 0.0003
Dementia Duration, yrs (SD)
***
5.4 (4.2) 4.4 (3.8) 7.3 (5.6) p = 0.0516
% Female 60.4 68.2 66.7 p = 0.5748
Education, years (SD) 7.2 (2.0) 7.9 (3.0) 6.8 (1.5) p = 0.1239
Monozygotic, n (%) 26 (25.7) 21 (31.8) 4 (2.7) p = 0.5059
Hypertension, n (%) 83 (82.2) 59 (89.4) 11 (73.3) p = 0.2279
Diabetes, n (%) 25 (24.7) 12 (18.1) 5 (33.3) p = 0.4115
Arrhythmia, n (%) 17 (16.5) 14 (21.9) 6 (40.0) p = 0.1842
PAD, n (%) 2 (2.0) 2 (3.0) 0 (0.0) p = 0.7512
CAD, n (%) 26 (25.7) 31 (47.0) 2 (13.3) p = 0.0042
TIA, n (%) 12 (11.9) 2 (3.0) 0 (0.0) p = 0.0436
Associations significant at p < 0.15 are bolded.
LOD = late-onset depression; EOD = early-onset depression; PAD = peripheral artery disease;
CAD = coronary artery disease; TIA = transient ischemic attack.
*EOD group significantly younger at age of CT scan compared with non-depressed group (p =
0.0011) and LOD group (p = 0.0038). No significant difference between non-depressed and LOD
groups. **EOD group had significantly earlier dementia onset than non-depressed group (p =
0.0001) and LOD group (p < 0.0001). No significant difference between non-depressed group
and LOD groups. ***EOD group had significantly longer dementia duration than LOD group (p
= 0.0186) but not significantly longer than non-depressed group. No significant difference
between non-depressed and LOD groups or non-depressed and LLD groups.
63
(β = 0.67, t(167) = 1.76, p = 0.0807) and compared to individuals with EOD (β = 1.09,
t(81) = 1.59, p = 0.1127). Chi-square analyses (for categorical variables) and one-way
ANOVA (for continuous variables) indicated that gender and zygosity did not
significantly differ by either depression onset or history of LLD (with the latter not being
shown in the tables).
Duration of Dementia
Across study participants, the mean time since dementia onset at the time of the
CT scan was 5.2 years (SD = 4.2). A one-way ANOVA indicated that differences
between the no depression, LOD group, and EOD group with respect to duration of
dementia at the time of the CT scan met the p < 0.15 significance criterion for covariates
(F = 3.01, df = 2, p = 0.0516). At the time of the CT scan those with LOD had the
shortest duration of dementia at 4.45 years (SD = 3.79), followed by individuals with no
history of depression at 5.37 years (SD = 4.22). Individuals with EOD, on average, had
7.30 years (SD = 5.63) between the time of dementia onset and when they were scanned.
When compared separately, there was a significant difference between the EOD and
LOD groups (β = 2.85, t(79) = 2.37, p = 0.0186) and EOD and non-depressed groups (β =
1.93, t(116) = 1.66, p = 0.0981) in dementia duration at the time of the scan. The
differences in dementia duration between the LOD and non-depressed groups and LLD
and non-depressed groups were not significant. Thus, with respect to duration of
dementia: EOD > non-depressed > LOD.
64
Cerebrovascular Disease Risk Factors
Chi-square analyses of cerebrovascular disease risk factors by depression group
indicated that only presence of coronary artery disease and transient ischemic attack
differed by depression group. (See Table 14.) Frequency of coronary artery disease was
significantly different by depression group—1.83 times greater in the LOD group than in
the non-depressed group (p = 0.0021) and 3.53 times greater than in the EOD group (p =
0.0131). There was no statistically significant difference between prevalence of coronary
artery disease in the EOD versus the non-depressed group (p = 0.3111). Individuals in
the non-depressed group were significantly more likely to have had a transient ischemic
attack than individuals in both the LOD and EOD groups (p = 0.0359).
Hypothesis 1
Compared to individuals with dementia and no history of depression, individuals
with dementia and LOD or dementia and LLD will have a greater amount of frontal lobe
deep WMH, subcortical WMH, and striatal hypodensities.
White Matter Hypodensities and Timing of Depression (Broad Definition)—All Dementia
Among individuals with LOD, there was only one individual rated as having one
focal hypodensity in the frontal deep white matter and only one individual in the LOD
group was rated as having more than one focal hypodensity. Therefore, we analyzed
frontal DWMH as a dichotomous variable—presence (one, more than one, and confluent)
versus absence of frontal DWMH. Controlling for gender and history of TIA, no
association was observed either between frontal DWMH and LOD (non-depressed vs.
LOD—OR = 0.55, 95% CI = 0.29, 1.06, p = 0.0743) or frontal DWMH and LLD (non-
65
depressed vs. LLD—OR = 0.55, 95% CI = 0.29, 1.02, p = 0.0589). (See Table 15.) We
also tested to see whether LOD was associated with confluent hypodensities. We again
analyzed frontal DWMH as a dichotomous variable—confluent frontal lobe DWMH
versus non-confluent (none, one focal, more than one focal) frontal lobe DWMH.
Controlling for gender, no significant association was observed either between confluent
frontal DWMH and LOD (non-depressed vs. LOD—OR = 1.18, 95% CI = 0.59, 2.34, p =
0.6444) or confluent frontal DWMH and LLD (non-depressed vs. LLD—OR = 1.11,
95% CI = 0.57, 2.16, p = 0.7527). (See Table 16.)
Controlling for age at CT scan and zygosity, no association was observed either
between subcortical WMH and LOD (non-depressed vs. LOD—OR = 0.91, 95% CI =
0.51, 1.62, p = 0.7457) or subcortical WMH and LLD (non-depressed vs. LLD—OR =
0.91, 95% CI = 0.52, 1.60, p = 0.7509). (See Table 17.) The association between greater
age at the time of the CT scan and a greater amount of subcortical WMH remained
significant (OR = 1.07, 95% CI = 1.02, 1.12, p = 0.0089).
Striatal Hypodensities and Timing of Depression—All Dementia
One-way ANOVA indicated no relationship between total number of striatal
hypodensities and late-onset depression (β = 0.09, t(167) = 0.50, p = 0.6210) and no
relationship between late-life depression and number of striatal hypodensities (β = 0.08,
t(176) = 0.47, p = 0.6404). (See Table 18.)
Table 15
Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset Depression and Late-Life Depression compared with
No Depression for All Dementia, Alzheimer’s Disease, and Vascular Dementia
-WMH = Frontal deep white matter hypodensities absent; +WMH = Frontal deep white matter hypodensities present; OR = Binary logit
odds ratio estimate, controlling for gender and history of TIA; CI = confidence interval; AD = Alzheimer’s disease; VaD = vascular
dementia.
*Late-onset depression group compared to no depression group. **Late-life depression group compared to no depression group.
No Depression Late-Onset Depression Late-Life Depression
-WMH
N (%)
+WMH
N (%)
-WMH
N (%)
+WMH
N (%)
OR*
(95% CI)
-WMH
N (%)
+WMH
N (%)
OR**
(95% CI)
All Dementia 52 (51.5) 49 (48.5) 44 (66.7) 22 (33.3)
0.61
(0.31, 1.18)
50 (66.7) 25 (33.3)
0.60
(0.32, 1.14)
AD Only 39 (56.5) 30 (43.5) 35 (70.0) 15 (30.0)
0.60
(0.27, 1.32)
36 (67.9) 17 (32.1)
0.66
(0.31, 1.43)
VaD Only 5 (26.3) 14 (73.7) 4 (57.1) 3 (42.9)
0.27
(0.04, 1.64)
6 (66.7) 3 (33.3)
0.18
(0.03, 0.99)
66
Table 16
Risk of Confluent Frontal Lobe Deep White Matter Hypodensities in Late-Onset Depression and Late-Life Depression
Compared with No Depression for All Dementia, Alzheimer’s Disease, and Vascular Dementia
-ConWMH = Confluent frontal deep white matter hypodensities absent; +ConWMH = Confluent frontal deep white matter hypodensities
present; OR = Binary logit odds ratio estimate, controlling for gender; CI = confidence interval; AD = Alzheimer’s disease; VaD =
vascular dementia.
*Late-onset depression group compared to no depression group. **Late-life depression group compared to no depression group.
No Depression Late-Onset Depression Late-Life Depression
-Con
WMH
N (%)
+Con
WMH
N (%)
-Con
WMH
N (%)
+Con
WMH
N (%)
OR*
(95% CI)
-Con
WMH
N (%)
+Con
WMH
N (%)
OR**
(95% CI)
All Dementia 73 (72.3) 28 (27.7) 46 (69.7) 22 (30.3)
1.18
(0.59, 2.34)
53 (70.7) 22 (29.3)
1.11
(0.57, 2.16)
AD Only 52 (75.4) 17 (24.6) 36 (72.0) 14 (28.0)
1.32
(0.56, 3.07)
37 (69.8) 16 (30.2)
1.47
(0.64, 3.36)
VaD Only 10 (52.6) 9 (47.4) 4 (57.1) 3 (42.9)
0.82
(0.14, 4.73)
6 (66.7) 3 (33.3)
0.56
(0.11, 2.95)
67
Table 17
Risk of Subcortical White Matter Hypodensities in Late-Onset Depression and Late-Life Depression compared with No
Depression for All Dementia, Alzheimer’s Disease, and Vascular Dementia
OR = Ordinal cumulative logit odds ratio estimate, controlling for age at time of CT scan; CI = confidence interval; AD = Alzheimer’s
disease; VaD = vascular dementia.
Level of WMH: 0 = none, 1 = one discrete hypodensity, 2 = more than one discrete hypodensity, 3 = confluent hypodensities.
* All numbers given in terms of N (%). **Late-onset depression group compared to no depression group. ***Late-life depression group
compared to no depression group.
No Depression Late-Onset Depression Late-Life Depression
Level of
WMH
*
0 1 2 3 0 1 2 3
OR
(95%
CI)
**
0 1 2 3
OR
(95%
CI)
***
All
Dementia
46
(45.5)
14
(13.9)
17
(16.8)
24
(23.8)
32
(48.5)
5
(7.6)
17
(25.8)
12
(18.2)
0.91
(0.51,
1.62)
36
(48.0)
8
(10.7)
18
(24.0)
13
(17.3)
0.91
(0.52,
1.60)
AD Only
34
(49.3)
9
(13.0)
12
(17.4)
14
(20.3)
25
(50.0)
4
(8.0)
11
(22.0)
10
(20.0)
0.97
(0.49,
1.94)
26
(49.1)
5
(9.4)
12
(22.6)
10
(18.9)
0.98
(0.50,
1.93)
VaD
Only
5
(26.3)
4
(21.1)
3
(15.8)
7
(36.8)
3
(42.9)
-
3
(42.9)
1
(14.3)
0.56
(0.12,
2.72)
4
(44.4)
1
(11.1)
3
(33.3)
1
(11.1)
0.45
(0.10,
1.92)
68
Table 18
Number of Striatal Hypodensities in Late-Onset Depression and Late-Life Depression compared with No Depression for All
Dementia, Alzheimer’s Disease, and Vascular Dementia
LOD = late-onset depression; LLD = late-life depression
*p-value
No Depression LOD No Depression vs. LOD
*
LLD No Depression vs. LLD
*
All Dementia
N 101 66 75
Striatum, mean (SD) 0.53 (1.02) 0.62 (1.25) 0.6210 0.61 (1.23) 0.6404
Alzheimer’s Disease Only
N 69 50 53
Striatum, mean (SD) 0.48 (0.99) 0.62 (1.34) 0.4973 0.58 (1.31) 0.6046
Vascular Dementia Only
N 19 7 9
Striatum, mean (SD) 0.53 (0.77) 0.57 (0.79) 0.9009 0.67 (0.87) 0.6691
69
70
Hypothesis 2
Compared to individuals with Alzheimer’s disease and no history of depression,
individuals with Alzheimer’s disease and LOD or Alzheimer’s disease and LLD will
have a greater amount of frontal lobe deep WMH, subcortical WMH, and striatal
hypodensities.
White Matter Hypodensities and Timing of Depression (Broad Definition)—Alzheimer’s
Disease Only
Among individuals with Alzheimer’s disease, there was no one in the LOD group
who was rated as having one focal hypodensity or and only one person who was rated as
having more than one focal hypodensity in the frontal deep white matter. Therefore, we
analyzed frontal DWMH as a dichotomous variable—presence versus absence of frontal
DWMH. Controlling for gender and history of TIA, no association was observed either
between frontal DWMH and LOD (non-depressed vs. LOD—OR = 0.60, 95% CI = 0.27,
1.32, p = 0.2049) or frontal DWMH and LLD (non-depressed vs. LLD—OR = 0.66, 95%
CI = 0.31, 1.43, p = 0.2953). (See Table 15.) We also tested to see whether LOD was
associated with confluent hypodensities among individuals with Alzheimer’s disease.
We again analyzed frontal DWMH as a dichotomous variable—confluent frontal lobe
DWMH versus non-confluent (none, one focal, more than one focal) frontal lobe
DWMH. Controlling for gender, no significant association was observed either between
frontal confluent DWMH and LOD (non-depressed vs. LOD—OR = 1.32, 95% CI =
0.56, 3.07, p = 0.5244) or frontal confluent DWMH and LLD (non-depressed vs. LLD—
OR = 1.47, 95% CI = 0.64, 3.36, p = 0.3578). (See Table 16.)
71
Controlling for age at CT scan and zygosity, no association was observed either
between subcortical WMH and LOD (non-depressed vs. LOD—OR = 0.97, 95% CI =
0.49, 1.94, p = 0.9355) or subcortical WMH and LLD (non-depressed vs. LLD—OR =
0.98, 95% CI = 0.50, 1.93, p = 0.9511). (See Table 17.) The association between greater
age at the time of the CT scan and a greater amount of subcortical WMH remained
significant (OR = 1.10, 95% CI = 1.04, 1.17, p = 0.0012).
Striatal Hypodensities and Timing of Depression—Alzheimer’s Disease Only
One-way ANOVA among individuals with Alzheimer’s disease indicated no
relationship between late-onset depression and number of hypodensities in the striatum (β
= 0.14, t(119) = 0.68, p = 0.4973) and no relationship between late-life depression and
number of hypodensities in the striatum (β = 0.11, t(122) = 0.52, p = 0.6046). (See Table
18.)
Hypothesis 3
Compared to individuals with vascular dementia and no history of depression,
individuals with vascular dementia and LOD or vascular dementia and LLD will have a
greater amount of frontal lobe deep WMH, subcortical WMH, and striatal hypodensities.
White Matter Hypodensities and Timing of Depression (Broad Definition)—Vascular
Dementia Only
Among individuals with vascular dementia, there was no one in the LOD group
who was rated as having one focal hypodensity or more than one focal hypodensity in the
frontal deep white matter. Therefore, we analyzed frontal DWMH as a dichotomous
variable—presence versus absence of frontal DWMH. Among those with vascular
72
dementia gender did not vary by frontal DWMH. No association was observed between
frontal DWMH and LOD (non-depressed vs. LOD—OR = 0.27, 95% CI = 0.04, 1.64, p =
0.1542). However, contrary to our hypothesis, individuals in the LLD group were
significantly less likely to have frontal DWMH than non-depressed individuals (non-
depressed vs. LLD—OR = 0.18, 95% CI = 0.03, 0.99, p = 0.0498). (See Table 15.) We
also tested to see whether LOD was associated with confluent hypodensities among
individuals with vascular dementia. We again analyzed frontal DWMH as a dichotomous
variable—confluent frontal lobe DWMH versus non-confluent (none, one focal, more
than one focal) frontal lobe DWMH. No significant association was observed either
between frontal confluent DWMH and LOD (non-depressed vs. LOD—OR = 0.82, 95%
CI = 0.14, 4.73, p = 0.8241) or frontal confluent DWMH and LLD (non-depressed vs.
LLD—OR = 0.56, 95% CI = 0.11, 2.95, p = 0.4951). (See Table 16.)
No association was observed either between subcortical WMH and LOD (non-
depressed vs. LOD—OR = 0.56, 95% CI = 0.12, 2.72, p = 0.4729) or subcortical WMH
and LLD (non-depressed vs. LLD—OR = 0.45, 95% CI = 0.10, 1.92, p = 0.2785). (See
Table 17.) There was no association between age at the time of the CT scan and amount
of subcortical WMH among individuals with vascular dementia.
Striatal Hypodensities and Timing of Depression—Vascular Dementia Only
Among individuals with vascular dementia, one-way ANOVA indicated that there
was no relationship between the total number of striatal hypodensities and LOD (β =
0.04, t(26) = 0.13, p = 0.9009), nor was there a relationship among LLD and number of
striatal hypodensities (β = 0.14, t(29) = 0.43, p = 0.6691). (See Table 18.)
73
Hypothesis 4
For all dementia types, Alzheimer’s disease only, and vascular dementia only,
compared to individuals without a history of depression, individuals with LOD or LLD
will have greater frontal ventricular width.
Frontal Ventricular Width and Timing of Depression (Broad Definition)—All Dementia,
Alzheimer’s Disease Only, and Vascular Dementia
One-way ANOVA controlling for gender and age indicated that across all types of
dementia, right and left frontal ventricular width did not vary between the no depression
and late-onset depression groups (R: β = 0.57, t(162) = 0.96, p = 0.3361; L: β = 0.62,
t(162) = 1.09, p = 0.2757) or between the no depression and late-life depression groups
(R: β = 0.38, t(171) = 0.92, p = 0.5066; L: β = 0.51, t(171) = 0.92, p = 0.3590). (See
Table 19.) Among individuals with Alzheimer’s disease right and left frontal ventricular
width also did not vary by either late-onset depression (R: β = 0.40, t(117) = 0.59, p =
0.5580; L: β = 0.48, t(117) = 0.71, p = 0.4807) or late-life depression (R: β = 0.34, t(120)
= 0.52, p = 0.6053; L: β = 0.42, t(120) = 0.64, p = 0.5228). Finally among individuals
with vascular dementia right and left frontal ventricular width did not vary between the
no depression and late-onset depression groups (R: β = 1.19, t(24) = 0.61, p =0.5482; L: β
= 0.93, t(24) = 0.51, p = 0.6159) or between the no depression and late-life depression
groups (R: β = 0.92, t(30) = 0.52, p = 0.6102; L: β = 1.00, t(30) = 0.61, p = 0.5488).
Table 19
Right and Left Frontal Ventricular Width (in mm) for All Dementia, Alzheimer’s Disease Only, and Vascular Dementia
LOD = late-onset depression; LLD = late-life depression. *p-value derived from ANOVA controlling for gender and age at time of CT scan.
No Depression LOD No Dep vs. LOD
*
LLD No Dep vs. LLD
*
All Dementia
N 97 65 74
Right Frontal, mean (SD) 13.13 (3.82) 13.55 (3.57) 0.3361 13.34 (3.64) 0.5066
Left Frontal, mean (SD) 13.56 (3.91) 13.98 (3.41) 0.2757 13.84 (3.40) 0.3590
Alzheimer’s Disease Only
N 68 49 52
Right Frontal, mean (SD) 13.00 (3.91) 13.20 (3.42) 0.5580 13.17 (3.32) 0.6053
Left Frontal, mean (SD) 13.56 (4.14) 13.80 (3.29) 0.4807 13.77 (3.21) 0.5228
Vascular Dementia Only
N 17 7 9
Right Frontal, mean (SD) 13.33 (3.46) 14.43 (4.86) 0.5706 14.33 (5.17) 0.6102
Left Frontal, mean (SD) 13.76 (3.29) 14.71 (4.64) 0.5948 14.78 (4.74) 0.5488
74
75
Dementia Severity
Dementia severity was assessed by the CDR. All participants were determined to
have at least very mild dementia (global CDR = 0.5), with 63.1% of the sample having
moderate to severe dementia. (See Table 20.) To determine whether clinical severity of
dementia differed by depression onset or history of LLD, we compared global CDR
scores and CDR domain scores for the non-depressed and LOD, and non-depressed and
LLD groups. (See Table 21.) As the means indicate, compared to the non-depressed
group both the LOD group and LLD group had significantly higher global scores as well
as significantly higher scores within the domains of memory, orientation,
judgment/problem solving, and home/hobbies.
One-way ANOVA indicated that severity of dementia on the CDR was also significantly
related to a longer length of dementia duration (F = 2.72, df = 3, p = 0.0459), with each
one-point increase in severity associated with a duration increase of one year. (See Table
20.) Controlling for dementia duration, a history of depression (either LOD or EOD)
significantly predicted dementia severity on global CDR (β = 0.36, t(175) = 2.84, p =
0.0050). Excluding the EOD group, a multiple regression model indicated that LOD (β
= 0.38, t(175) = 2.81, p = 0.0056) and longer duration of dementia (β = 0.05, t(175) =
2.80, p = 0.0057) were each significantly associated with a higher level of dementia
severity. This suggests that the degree of dementia was clinically more severe in the
LOD group, and that this association remains significant even when controlling for
duration of dementia.
76
Table 20
Numbers of Individuals by Dementia Severity Rating and Level of Impairment in Each
CDR Domain
*Average years of dementia duration for each level of dementia severity.
Dementia Severity/Level of Impairment
0 0.5 1 2 3
Global CDR, n (%) 0
(0)
15
(8.5)
50
(28.4)
60
(34.1)
51
(29.0)
Dementia Duration, years (SD)
*
-- 3.9
(4.5)
4.3
(4.1)
5.3
(3.5)
6.5
(4.9)
Memory, n (%) 2
(1.1)
7
(4.0)
38
(21.6)
85
(48.3)
44
(25.0)
Orientation, n (%) 5
(2.8)
22
(12.5)
54
(30.7)
54
(30.7)
41
(23.3)
Judgment, n (%) 1
(0.6)
15
(8.6)
43
(24.6)
57
(32.6)
59
(33.7)
Community, n (%) 7
(4.0)
17
(9.7)
45
(25.6)
57
(32.4)
50
(28.4)
Home/Hobby, n (%) 5
(2.8)
12
(6.8)
40
(22.7)
52
(29.6)
67
(38.1)
Personal Care, n (%) 36
(20.6)
5
(2.7)
56
(32.0)
39
(22.3)
39
(22.3)
Table 21
Mean Global CDR Score and CDR Domain Scores by Depression Group
CDR = Clinical Dementia Rating Scale; LOD = late-onset depression; EOD = early-onset depression; LLD = late-life depression
*p-value. Associations significant at p < 0.05 are bolded for Global CDR score. Associations significant at p < 0.06 are bolded for
CDR domain scores.
No
Depression
LOD EOD
No
Depression
vs. LOD*
LLD
No
Depression
vs. LLD*
N 96 65 14 74
Global CDR, mean (SD) 1.7 (0.87) 2.1 (0.85) 2.1 (0.80) 0.0152 2.0 (0.84) 0.0178
Memory, mean (SD) 1.8 (0.78) 2.2 (0.73) 1.8 (0.77) 0.0059 2.1 (0.75) 0.0344
Orientation, mean (SD) 1.5 (0.91) 1.9 (0.88) 1.6 (0.97) 0.0075 1.8 (0.88) 0.0264
Judgment, mean (SD) 1.8 (0.91) 2.1 (0.86) 2.1 (0.83) 0.0341 2.1 (0.85) 0.0445
Community, mean (SD) 1.7 (0.92) 1.9 (0.97) 1.9 (0.96) 0.2001 1.9 (0.96) 0.2209
Home/Hobbies, mean (SD) 1.9 (0.94) 2.2 (0.92) 2.1 (0.91) 0.0450 2.1 (0.92) 0.0514
Personal Care, mean (SD) 1.3 (1.02) 1.5 (1.12) 1.6 (1.11) 0.2448 1.6 (1.10) 0.2114
77
78
Next we assessed whether total number of striatal hypodensities or white matter
hypodensity ratings in the frontal lobe and subcortical white matter differed by global
CDR score. Total number of striatal hypodensities, presence of frontal lobe DWMH
(controlling for gender), and higher rating of subcortical WMH (controlling for age at CT
scan and zygosity) were not associated with higher levels of dementia severity, or higher
scores in any of the functional domains. This did not vary by depression onset or history
of late-life depression. Thus, higher levels of dementia severity were not associated with
a greater severity of WMH or severity of subcortical gray matter hypodensities.
Source of Depression Information
Despite the demonstrated concordance between medical records and informant-
reported depression, we examined in a post-hoc analysis whether the relationship differed
based upon the source of information about depression. There was no significant
difference in the relationship between frontal DWMH, subcortical WMH, striatal
hypodensities and depression when informant-reported medical history information was
disregarded. There were also eight individuals missing either an informant report (N = 6)
or coded medical records (N = 2). Dropping these individuals from our analyses did not
affect the outcomes of our analyses.
79
Chapter 4: Discussion
The aim of the present study was to investigate the underlying neural mechanisms
linking depression and dementia. This goal stems from the idea that there could be a
pathophysiological basis in the brain for the comorbidity of depression and dementia
(Jorm, 2001). According to the vascular depression hypothesis, white and gray matter
lesions resulting from cerebrovascular disease (as well as normal aging) may cause late-
onset depression. It has been further posited that lesions occurring in the fronto-striatal
pathway specifically lead to late-onset depression. Thus, in line with the vascular
depression and fronto-striatal hypotheses, we predicted that demented individuals with a
LOD or a history of LLD would have a higher prevalence of WMH in the frontal lobe
deep white matter and subcortical white matter, and a greater amount of gray matter
hypodensities in the striatum, compared to individuals with dementia and no depression.
We also predicted that we would find a greater degree of frontal lobe atrophy in the LOD
and LLD groups compared to the no depression group. While similar hypotheses have
been tested in the past among individuals with late-onset and/or late life depression, they
have rarely been considered among individuals with both dementia and late-onset and/or
late life depression.
Late-onset depression was defined in two ways, using both a narrow and a
broader definition. The broad definition included 31 individuals coded as depressed
based solely upon having taken antidepressants and the narrow definition did not include
these individuals. All analyses were run using both definitions of depression history. In
addition, we made an alternative late life depression variable comprised of those with
80
late-onset depression plus those with early-onset depression who experienced a
recurrence in old age.
Although we looked descriptively at data from those with early-onset depression,
we ultimately dropped this group from the comparison. The group was very small, and it
was disproportionately comprised of individuals whose depression had been sufficiently
severe to result in hospitalization (narrow definition: 50.0% EOD vs. 11.3% LOD, p =
0.0060; broad definition: 46.7% EOD vs. 6.1% LOD, p < 0.0001). Medical records could
not be abstracted back as far as the individual’s early and middle adulthood. Therefore,
the only means for assessing a history of less severe EOD (not requiring hospitalization)
would have been through the interview with the proband’s informant. Many of the
informants were the proband’s adult children, and understandably less aware of whether
the older adult with dementia had been depressed earlier in their lives. Therefore, the
degree of underascertainment was likely substantially greater for those whose depression
had an early life onset and did not recur compared to those with depression as older
adults, which may have affected comparisons between the LOD/LLD and EOD groups in
unknown ways.
Narrow Definition of Depression
In support of one of the main hypotheses of the study, individuals specifically
with Alzheimer’s disease and late-onset depression had a significantly greater total
amount of striatal hypodensities compared to individuals with Alzheimer’s disease and
no history of depression. Further analysis revealed that this association was not
confounded by any other medical or demographic covariate. Though not statistically
significant (p < 0.10), analyses run using individuals with any type of dementia (not
81
specifically Alzheimer’s disease) revealed a similar trend in that the total number of
striatal hypodensities was greater in the LOD group. In addition, there was a similar,
although non-significant, association between a greater number of striatal hypodensities
in the LLD group compared to the no depression group among both individuals with all
types of dementia as well as among individuals with Alzheimer’s disease only. Thus,
there was a consistent pattern of findings for striatal hypodensities, but the finding was
only statistically significant among those with Alzheimer’s disease, in particular, and
with late onset depression with no earlier history of depression.
With respect to the white matter components of the fronto-striatal pathway, our
findings did not confirm an association between greater amount of WMH and LOD or
LLD. Though there were a number of people in the nondepressed group with one or
more focal hypodensities in the frontal lobe deep white matter, among individuals in the
late-onset depression group, frontal lobe deep WMH were either completely absent or
confluent. Thus, analyses tested two associations: 1. Whether presence versus absence
of frontal lobe deep WMH varied between the nondepressed and late-onset depression
groups or between the nondepressed and late-life depression groups, and 2. Whether
confluence versus non-confluence of frontal deep WMH varied between the
nondepressed and LOD or nondepressed and LLD groups.
For all dementia as well as Alzheimer’s disease only, the presence (versus
absence) of frontal lobe deep WMH did not differ between the non-depressed and late-
onset depression or late-life depression groups. Male gender and a history of TIA were
the only two significant predictors of presence of frontal lobe deep WMH. Male gender
was a significant predictor of frontal lobe deep WMH, with a higher prevalence of frontal
82
lobe deep WMH in men compared with women. Controlling for gender, a comparison of
the presence of frontal lobe deep WMH by history of TIA indicated that individuals with
a history of TIA were 3.92 (95% CI = 1.16, 13.24, p = 0.0278) times more likely that
individuals without a history of TIA to have hypodensities in the frontal lobe deep white
matter. While the prevalence of ischemic neuropathology on CT in patients with TIAs
has been estimated at approximately 40%, most abnormalities are diffuse throughout the
cerebral white matter or basal ganglia (Stevens, Barber, Keslake, Broz, & Barker, 1991).
Therefore, our finding that hypodensities were specific to the frontal lobe deep white
matter is somewhat surprising.
In the vascular dementia group, results indicated that the presence of frontal lobe
deep WMH was significantly greater in the non-depressed group than in the late-life
depression group. However, the small size of the late-onset and late-life groups makes
such an association spurious. Such a relationship would be strikingly contrary to the
established understanding of vascular dementia pathology. When we tested whether
confluence of frontal deep WMH was associated with LOD or LLD, the odds ratios,
though nonsignificant, were consistent with the pattern predicted by our original
hypotheses among the total sample as well as for the Alzheimer’s disease only group,
particularly among individuals with LLD and Alzheimer’s disease. History of TIA was
not significantly associated with confluence of frontal lobe deep WMH.
Severity of subcortical white matter hypodensities did not differ between the
nondepressed and LOD groups or the nondepressed and LLD groups among in any of the
groups (all dementia, Alzheimer’s disease only, or vascular dementia only). Severity of
subcortical WMH was only associated with age at CT scan and monozygosity. Finally
83
frontal ventricular width did not vary by onset of depression or history of late-life
depression. Greater ventricular width was associated with male gender and older age at
the time of the CT scan.
Broad Definition of Depression
Using the broad definition of depression history, in which those who were
prescribed anti-depressants were scored as depressed even when a history of depression
was not confirmed by any other collateral source, we were unable to demonstrate
relationship between any of the imaging variables and onset of depression or history of
late-life depression. Instead we found that among individuals with dementia in general
and among individuals specifically with Alzheimer’s disease, hypodensities in the frontal
lobe deep white matter, subcortical white matter, and striatal gray matter did not differ
between the LOD/LLD groups and the non-depressed group. We also found that frontal
lobe ventricular width did not vary by either depression onset or history of late-life
depression. As with the analyses using the narrower definition of depression history,
covariates included gender and history of TIA for analyses examining presence of frontal
lobe deep WMH, age at CT scan and zygosity for analyses examining level of subcortical
WMH, and age and gender for analyses investigating frontal lobe ventricular width.
White Matter Hypodensities and LOD/LLD
Our lack of findings with respect to the white matter components of the fronto-
striatal pathway was surprising in light of the solid base of literature supporting its
association with late-onset and late-life depression. Several differences between the
present study and past studies may explain this discrepancy. Most of the study
populations that found the association between WMH and LOD and LLD were
84
comprised of cognitively intact older adults or older adults with mild cognitive
impairment. All of the individuals in the present study had dementia, with approximately
70% diagnosed with Alzheimer’s disease. Therefore this study at least partially
“controls” for the neuropathology related to dementia. However, it could be that by the
time cognitive and functional impairment has progressed to the point of dementia, the
neurological damage caused by dementia is widespread enough to subsume the
neuropathology of late-onset depression such that the two neuropathologies become
indistinct. In fact, this may even happen in the early stages of dementia, as suggested by
a study of mildly demented patients both with and without a subsyndromal level of
depressive symptoms, which failed to find any significant relationship between white
matter changes and depressive symptoms (Lind et al., 2006). At variance with this,
however, Lavretsky et al. (2008) found that higher lacunar volume in the white matter
was associated with neuropsychiatric symptoms (including depressed mood, anhedonia,
anergia, and apathy) in a group of both cognitively intact and cognitively impaired
individuals, even after controlling for cognitive status.
In addition, the current study utilizes CT imaging, which, compared to MRI, is
much less sensitive to white matter changes. In a study comparing the effectiveness of
CT and MRI with respect to detecting white matter changes (Erkinjuntii et al., 1987),
MRI detected white matter changes in 36.4% of individuals with Alzheimer’s disease,
whereas the CT images of the same individuals detected white matter changes in only
4.5% of the scans. Our study detected frontal deep WMH in 37.0% and subcortical
WMH in 48.8% of individuals with Alzheimer’s disease. However, with respect to the
frontal deep WMH, 70.2% of the WMH identified were rated as confluent, which would
85
correspond to relatively severe white matter changes. Therefore, in our study it is
possible that the less severe, discrete focal WMH were not always detected by the CT
images in our study.
Subcortical Gray Matter Lesions and LOD in Alzheimer’s Disease
The finding that individuals with LOD and Alzheimer’s disease had a greater
number of striatal hypodensities compared to individuals without depression and
Alzheimer’s disease is consistent with several prior imaging studies of late-life
depression in nondemented individuals demonstrating a greater prevalence of lesions in
the putamen among individuals with geriatric depression (Greenwald et al., 1998) and
reduced caudate nucleus volume among adults with depression (Kim, Hamilton, &
Gotlib, 2008). The striatum represents the subcortical gray matter components of the
fronto-striatal pathway. Thus, the finding that individuals with AD and LOD have a
greater amount of striatal hypodensities compared to individuals with AD and no
depression provide some support for the fronto-striatal hypothesis of depression among
individuals with dementia.
Although Alzheimer’s disease is a “cortical” dementia, the individuals with LOD
had a greater prevalence of hypodensities in the subcortical gray matter compared to the
individuals with Alzheimer’s disease and no depression. As subcortical changes are not
typically associated with Alzheimer’s disease, the greater amount of striatal
hypodensities evident in the individuals with Alzheimer’s disease and LOD may
represent a neuropathological process that is distinct from the neurodegenerative changes
caused by Alzheimer’s disease. In fact, dementia characterized by the neurodegeneration
of subcortical structures, or subcortical dementia, is often associated with depression-like
86
symptoms, which may provide further support to the idea that the subcortical gray matter
lesions may represent a process specific to LOD. Therefore, the individuals in the
present study with Alzheimer’s disease and LOD have a “double dose” of
neuropathology—one that is primarily cortical (Alzheimer’s disease) and the other
subcortical (LOD), which, as discussed in the following section, may explain why
individuals with LOD were more cognitively and functionally impaired than their non-
depressed counterparts.
Dementia Severity
Individuals in the LOD/LLD groups on average had the shortest duration of time
between dementia onset and date of the CT scan. Therefore, we considered that one
reason we may not have found a relationship between LOD/LLD and white matter
hypodensities was that the degree of dementia in the LOD and LLD samples was simply
not as extreme as the degree in the non-depressed sample. This, however, was not the
case. Quite to the contrary, when the LOD and LLD groups were compared with the
non-depressed group, both the LOD and LLD groups were more impaired globally and
more impaired within the domains of memory, orientation, judgment/problem-solving,
and home/hobbies compared to the nondepressed group. Consistent with the lack of
association between LOD/LLD and WMH and frontal ventricular width, global CDR
scores and CDR domain scores were not related to total frontal and subcortical white
matter hypodensities or ventricular width. The degree of dementia in the LOD group was
higher, and this greater level of dementia severity could not be explained by duration of
dementia.
87
There could be several reasons as to why the LOD group was more impaired. As
there is much overlap in the symptomatology of depression and early-stage dementia, one
possibility is that the depression hid the diagnosis of dementia. In other words it could be
that individuals in the LOD group were diagnosed with depression when, in actuality,
they may have been better diagnosed as demented. Therefore, the duration of dementia at
the time of the CT scan in reality might have been longer than the data would have
reflected. Another reason for the greater impairment in the LOD group could be that the
behavioral sequelae of depression cause certain aspects of the functional impairment
associated with dementia to appear more severe. Although some studies have found that
the neurocognitive impairment seen in individuals with late-onset depression persists
even after the depression has remitted (Thomas & O’Brien, 2008), this could suggest that
dementia severity may be somewhat attenuated if the depression were to remit. Finally,
in conjunction with the finding that individuals with AD and LOD had more striatal
hypodensities, it could be that individuals with co-morbid dementia and depression have
two different neurodegenerative processes occurring at the same time, which together
may lead to greater functional impairment.
Coronary Artery Disease
Coronary artery disease was almost twice as common among individuals with
late-onset depression compared to individuals without depression and 3.53 times more
common in late-onset depression than in early-onset depression. This relationship
between late-onset depression and coronary artery disease is unsurprising given the
Swedish Twin Registry study indicating the modest (OR = 1.3) lifetime association
between major depression and CAD, and that CAD onset was associated with a three
88
times greater risk of depressive onset in the same year and a two times greater risk of
depressive onset in successive years (Kendler, Gardner, Fiske, and Gatz, 2009).
However, our current analysis is based on individuals with dementia; whereas, the prior
report required individuals to be sufficiently cognitively intact to complete a depression
screening interview. We did not find any relationship between CAD and WMH, striatal
hypodensities, and frontal ventricular width. Therefore, the relationship between CAD
and late-onset depression does not seem to be related to any of the neuroanatomical
regions examined in this study. This finding is in contrast to the relationship between
CAD and neuroanatomical features reported by others (e.g., Hoshide et al., 2001; Uehara
et al., 1999) that suggest an association between CAD and subcortical gray matter
lesions.
Heterogeneity of Demographic Characteristics and Depression-Related Neuropathology
by Onset
Across several analyses the EOD group was markedly different from either the
LOD/LLD or no depression groups. Though not statistically significant, individuals in
the EOD group had fewer WMH and striatal hypodensities compared to individuals in the
LOD or LLD groups. This is consistent with prior studies showing that depression in
young adults and middle-aged adults is not associated with WMH (Savitz & Drevets,
2009). In addition, those with no recurrence of depression later in life were
comparatively free of risk factors for cerebrovascular disease, although given such a
small number of individuals in the EOD-only group, this observation should be
interpreted with caution. These differences support the theoretical consideration that
LOD has a different pathophysiology from depression earlier in life.
89
It is important to acknowledge, however, that the EOD group in this study may
have been very select for methodological and substantive reasons. Methodologically,
other than the retrospective informant report, the only other depression measures that
would have indicated an early-life episode of depression were the inpatient discharge
registries. Medical records for the study participants went back an average of 10 years
and therefore, would have only indicated later-life episodes of depression. Therefore, we
had fewer sources of information regarding early-life depression, and these sources
would have been more likely to pick up depression that was severe and required
hospitalization. Another reason for possible selection bias in the EOD group could be
that EOD is not a risk factor for dementia, and therefore, EOD was less prevalent than
LOD in this population with dementia. Findings from a recent study of the population
from which the present study sample was drawn, indicated that depression closer in time
to dementia onset, and not earlier life depression, was associated with an increased risk of
dementia (Brommelhoff et al., 2009).
Demographically the individuals in the EOD group were on average younger at
the time of the scan and had an earlier onset of dementia. There are several possible
explanations for this finding. One possibility is that for the individuals with EOD who
later developed dementia, the early history of depression may have “sped up” the
dementia onset. Two features of the EOD group, however, suggest that this finding could
be an artifact. First, the sample size of the EOD group was quite small and therefore
more susceptible to being affected by a small number of individuals having a very early
onset of dementia (i.e., four individuals in the EOD group had a dementia age of onset
prior to age 60, whereas there were no individuals in the LOD group with an age of onset
90
prior to age 60). Also, as discussed above, it is likely that the early depressive episodes
represented in the EOD group were more severe. Although individuals with a history of
a psychotic disorder, schizophrenia, and bipolar disorder were excluded from this study,
it is possible that some of the early-onset depressive episodes could have been
representative of a more severe psychiatric or neurological process that led to an earlier
onset of cognitive decline.
Limitations
Technological. This study has many notable limitations. First is the use of CT
instead of the more technologically advanced MRI. CT has better specificity than MRI
for white matter changes of clinical significance—95.5% versus 68.2%, respectively for
frontal white matter changes, and 95.5% versus 63.3% for whole brain white matter
changes (Erkinjuntti et al., 1987). However, CT is known for being less sensitive in
detecting subtle microvascular ischemic changes, especially in individuals with
Alzheimer’s disease, the diagnosis assigned to the majority of this study’s participants.
In one study comparing MRI to CT neuroimaging of individuals with Alzheimer’s
disease, white matter changes in the frontal lobe were evident in 7 out of the 22
participants using MRI, but only 1 out of the same 22 participants using CT.
Additionally, it was determined that sensitivity to mild frontal lobe WMH was only
31.0% for CT imaging versus 75.9% for MRI (Erkinjuntii et al., 1987). Thus, it is
possible that the white matter changes found by other studies are too subtle to be detected
by CT. Nonetheless, white matter changes in the present study were found to increase
with age, as would be expected, and to be more common in vascular dementia than in
other dementias.
91
Cross-Sectional Study Design. A major limitation inherent in the study design is
that CT data were available for only one point in time, and on average five years after the
age of dementia onset. Due to our lack of premorbid CT data, we can make no causal
statements regarding depression and the presence of white or gray matter hypodensities
or ventricle size, nor can we make any statements about change over time relative to
progression of dementia. At the same time, there was not a significant relationship
between duration of dementia and amount of white matter hypodensities in either the
frontal lobe or subcortical white matter. For striatal hypodensities there was no
relationship when the entire sample was examined as a whole. However, when the non-
depressed and LOD groups were examined separately, there appeared to be a slightly
inverse trend between dementia duration and number of striatal hypodensities for the
nondepressed group and the approximated regression line for the LOD group suggested
the opposite trend. That is, for the LOD group, total number of striatal hypodensities
appeared to be associated with a longer duration of dementia. Taken together with the
greater amount of striatal hypodensities in the LOD/Alzheimer’s disease group compared
to nondepressed Alzheimer’s disease group, this association, although cross-sectional,
lends further support to the idea that there may be two separate neurodegenerative
processes occurring among the individuals with co-morbid Alzheimer’s disease and
LOD.
Antidepressants. The association between striatal hypodensities and late-onset
depression varied by whether antidepressant use was included as an indicator of
depression history. When use of antidepressants was not included as an indicator of
depression, 45% of the individuals coded as LOD using the broader definition of
92
depression were re-coded as non-depressed. The rationale for dropping antidepressants
as an indicator of depression history was to ensure that people taking the medication for
reasons other than depression were not included in the depressed group, especially as
antidepressants may be over-prescribed in geriatric populations. It is unlikely, however,
that all 31 of these individuals were prescribed these medications for reasons other than
depression. Thus, when using the narrower definition of depression, we may have been
misclassifying some individuals with depression as non-depressed.
Depression assessment. As discussed previously, another limitation was that the
methods for ascertaining a history of depression were more sensitive to detecting recent
episodes. Indeed, in our sample only 8.2% had a depressive episode prior to age 60,
which is well below the estimated lifetime prevalence rate of 19.5% for major depression
in Swedish twins (Kendler, Gatz, Gardner, & Pedersen, 2006). Thus individuals in the
late-onset group may have been misclassified due to unreported early episodes of
depression, although they would have been correctly classified in analyses of late-life
depression. In addition to less sensitive ascertainment of EOD, it is also possible that
there was differential survival in individuals with early-onset depression, such that they
were not as likely as individuals with LOD to develop dementia, and thus were under-
represented as a whole in this sample.
Another limitation stems from the fact that there is no consensus as to what age
constitutes early-onset versus late-onset and late-life depression. A recent meta-analysis
(Herrmann, Masurier, & Ebmeier, 2008) of WMH in late-life depression noted that there
was substantial variability across studies in the age cut-offs used to define EOD and
LOD, with ages ranging from 45 to 65 years of age. Whereas Alexopoulos et al. (1997)
93
proposed that vascular depression be defined by an age of onset after 65 years, Krishnan
et al. (1995) suggest a definition using an age cutoff of 50 years. Devandand et al (2004)
recommended an age cutoff of 60 years because differences in cardiovascular risk factors
between EOD and LOD are greatest at this point.
Another reason for the differences in findings between our study and the findings
of some previous studies of depression in those without dementia could be the nature of
our depression measures. Our depression measures were dichotomous and there was no
way to distinguish between levels of depression severity. As suggested by previous
studies of late life depression (e.g. Coffey, Figiel, Djang, & Weiner, 1990; Hickie, Scott,
Wilhelm, & Brodaty, 1996) white matter changes are related to the severity of late-onset
depression, with differences between non-depressed and late-onset depression groups
being evident only when the level of depression is relatively severe. Therefore, as the
late-onset depression group in our study likely represented a wide range of depression
severity (and because there were not enough individuals in our sample for a separate
analysis with depression severe enough to require hospitalization), differences between
the non-depressed and LOD groups could have been obscured due to the number of
individuals with milder levels of depression.
94
Chapter 5: Conclusions
In summary, the aim of the present study was to examine whether there were
neuroanatomical differences evident on the CT scans of individuals with dementia based
upon history of depression, depression onset (early-onset versus late-onset), and whether
there was a history of late-life depression (any episode of depression after age 60). In
support of the fronto-striatal hypothesis of depression, the results indicated that among
individuals with Alzheimer’s disease, there was a significant relationship between late-
onset depression and a greater amount of striatal hypodensities (gray matter
hypodensities in the caudate nucleus and lentiform nucleus) compared to people with no
depression. We also found that compared to the nondepressed group, individuals with
LOD displayed a significantly higher degree of global functional impairment, as well as
impairment within the domains of memory, orientation, judgment/problem-solving, and
in the context of their home activities and hobbies. Thus, it could be that there are two
separate and additive neurodegenerative processes occurring among the individuals with
co-morbid Alzheimer’s disease and LOD. The association between LOD and striatal
hypodensities in individuals with Alzheimer’s disease, however, did not hold when
individuals with a history of depression defined solely by antidepressant use were
included in the depressed group.
We did not find an association between LOD/LLD and white matter
hypodensities in the frontal deep white matter or subcortical white matter. Other studies
in individuals without dementia finding such an association have typically used MRI.
Thus, associated white matter changes may have been too subtle to be detected by CT
imaging. However, as prior studies were among individuals without dementia, it is
95
equally likely that the neurological damage in the white matter caused by dementia
subsumes the neuropathology of late-onset depression, such that the two
neuropathologies become indistinct with respect to white matter. Finally, this study did
not demonstrate a significant relationship between late-onset depression and increased
frontal ventricular width. However, some of the most recent studies in this area have
suggested that frontal lobe atrophy is associated with cognitive impairment rather than
late-onset depression.
Depression is a heterogeneous disorder, and an ever-growing body of literature
suggests that late-life depression is unique in that it is a result of age-associated vascular
lesions. In older adults, depression is associated with chronic medical conditions,
decreases in subjectively reported quality of life, and functional disability, independent of
the type or severity of the medical illness (Patten, 1999; Steffens et al., 1999). Thus, for
older adults with dementia and late-life depression, there are important public health
implications for identifying the pathophysiologic mechanisms linking these two
conditions. Given that many treatments for depression are mediated by neurobiological
factors, the efficacy of these treatments may be compromised in individuals with
neuroanatomical abnormalities. An understanding of the unique brain changes associated
with dementia versus those associated with LOD may elucidate avenues for treatments
with better efficacy in older adults.
96
References
Alexopoulos, G., S., Young, R. C., & Shindledecker, R. D. (1992). Brain computed
tomography findings in geriatric depression and primary degenerative dementia.
Biological Psychiatry, 31, 591-599.
Alexopoulos, G. S. S., Meyers, B., Young, R. C., Campbell, S., Silbersweig, D., &
Charlson, M. (1997). ‘Vascular depression’ hypothesis. Archives of General
Psychiatry, 54, 915-922.
Almeida, O.P., Burton, E.J., Ferrier, N., McKeith, G., & O’Brien, J.T. (2003).
Depression with late onset is associated with right frontal lobe atrophy.
Psychological Medicine, 33, 675-681.
Bassuck, S. S., Berkman, L. F., & Wypij, D. (1998). Depressive symptomatology and
incident cognitive decline in an elderly community sample. Archives of General
Psychiatry, 55, 1073-1081.
Blessed, G., Tomlinson, B. E., & Roth, M. (1968). The association between
quantitative measures of dementia and of senile change in the cerebral grey matter
of elderly subjects. British Journal of Psychiatry, 114, 797-811.
Boland, R. J. (2000). Depression in Alzheimer’s disease and other dementias.
Current Psychiatry Reports, 2, 427-433.
Brodaty, H., Luscombe, G., Parker, G., Wilhelm, K., Hickie, I., Austin, M.-P., &
Mitchell, P. (2001). Early and late onset depression in old age: Different
aetiologies, same phenomenology. Journal of Affective Disorders, 66, 225-
236.
Brommelhoff, J.A., Gatz, M., Johansson, B., McArdle, J. J., Fratiglioni, L., & Pedersen,
N. L. (2009). Depression as a Risk Factor or Prodomal Feature for Dementia?
Findings in a Population-Based Sample of Swedish Twins. Psychology and
Aging, 24, 373-384.
Capizzano, A. A., Acion, L., Bekinschtein, T., Furman, M., Gomila, H., Martinez, A.,
Mizrahi, R., & Starkstein, S. E. (2004). White matter hyperintensities are
significantly associated with cortical atrophy in Alzheimer’s disease. Journal of
Neurology, Neurosurgery & Psychiatry, 75, 822-827.
Carpenter, B. D., Xiong, C., Porensky, E. K., Lee, M., Brown, P. J., Coats, M., Johnson,
D., & Morris, J. C. (2008). Reaction to a dementia diagnosis in individuals with
Alzheimer’s disease and mild cognitive impairment. Journal of the American
Geriatrics Society, 56, 405-412.
97
Centre for Epidemiology, National Board of Health and Welfare. (1997). The Swedish
Hospital Discharge Register 1987-1995. Stockholm: Author.
Chen, P., Ganguli, M., Mulsant, B. H., & DeKosky, S. T. (1999). The temporal
relationship between depressive symptoms and dementia. Archives of General
Psychiatry, 56, 261-266.
Chui, H. (2007). Subcortical ischemic vascular dementia. Neurologic Clinics, 25, 717-
740.
Clark, L. M., McDonald, W. M., Welsh-Bohmer, K. A., Siegler, I. C., Dawson, D. V.,
Tupler, L. A., & Krishnan, K. R. R. (1998). Magnetic resonance imaging
correlates of depression in early- and late-onset Alzheimer’s disease. Biological
Psychiatry, 44, 592-599.
Coffey, C. E., Figiel, G. S., Djang, W. T., & Weiner, R. D. (1990). Subcortical
hyperintensity on magnetic resonance imaging: A comparison of normal and
depressed elderly subjects. American Journal of Psychiatry, 147, 187-189.
Coffey, C. E., Wilkinson, W. E., Weiner, R. D., Parashos, I. A., Djang, W. T., Webb, M.
C., Figiel, G. S., & Spritzer, C. E. (1993). Quantitative cerebral anatomy in
depression. Archives of General Psychiatry, 50, 7-16.
Devanand, D. P., Adorno, E., Cheng, J., Burt, T., Pelton, G. H., Roose, S. P., &
Sackheim, H. A. (2004). Late onset dysthymic disorder and major depression
differ from early onset dysthymic disorder and major depression in elderly
outpatients. Journal of Affective Disorders, 78, 259-267.
Erkinjuntii, T., Ketonen, L., Sulkava, R., Sipponen, J., Vuorialho, M., & Iivanainen, M.
(1987). Do white matter changes on MRI and CT differentiate vascular dementia
from Alzheimer’s disease. Journal of Neurology, Neurosurgery, and Psychiatry,
50, 37-42.
Fein, G., DiSclafani, V., Tanabe, J., Cardenas, V., Weiner, M. W., Jagust, W. J., Reed,
B.R., Norman, D., Schuff, N., Kusdra, L., Greenfield, T., & Chui, H. (2000).
Hippocampal and cortical atrophy predict dementia in subcortical ischemic
vascular disease. Neurology, 55, 1626-1635.
Firbank, M. J., O’Brien, J. T., Pakrassi, S., Pantoni, L., Simoni, M., Erkinjuntti, T.,
Wallin, A., Wahlund, L. O., van Straaten, I., & Inzitari, D. (2005). White matter
hyperintensities and depression—Preliminary results from the LADIS study.
International Journal of Geriatric Psychiatry, 20, 674-679.
98
Fiske, A., Gatz, M., Aadnoy, B., & Pedersen, N. L. (2005). Assessing age of dementia
onset: Validity of informant reports. Alzheimer’s Disease and Associated
Disorders, 19, 128-134.
Gatz, M., Fratiglioni, L., Johansson, B., Berg, S., Mortimer, J. A., Reynolds, C. A., et al.
(2005). Complete ascertainment in the Swedish Twin Registry: The HARMONY
study. Neurobiology of Aging, 26, 439-447.
Gatz, M., Reynolds, C., Nikolic, J., Lowe, B., & Karel, M. (1995). An empirical test of
telephone screening to identify potential dementia cases. International
Psychogeriatrics, 9, 429-437.
Gatz, M., Reynolds, C. A., John, R., Johansson, B., Mortimer, J. A., & Pedersen, N. L.
(2002). Telephone screening to identify potential dementia cases in a population-
based sample of older adults. International Psychogeriatrics, 14, 273-289.
Greenwald, B. S., Kramer-Ginsberg, E., Krishnan, R. R., Ashtari, M., Auerbach, C., &
Patel, M. (1998). Neuroanatomic localization of magnetization of magnetic
resonance imaging signal hyperintensities in geriatric depression. Stroke, 29,
613-617.
Hachinski, V., Iadecola, C., Petersen, R. C., Breteler, M. M., Nyenhuis, D. L., Black, S.
E., Powers, W. J., DeCarli, C., Merino, J. G., Kalaria, R. N., Vinters, H. V.,
Holtzman, D. M., Rosenberg, G. A., Dichgans, M., Marler, J. R., & Leblanc, G.
C. (2006). National Institute of Neurological Disorders and Stroke—Canadian
Stroke Network Vascular Cognitive Impairment Harmonization Standards.
Stroke, 37, 2220-2241.
Heiden, A., Kettenback, J., Fischer, P., Schein, B., Ba-Ssalamah, A., Frey, R., Naderi, M.
M., Gulesserian, T., Schmid, D., Trattnig, S., Imhof, H., & Kasper, S. (2005).
White matter hyperintensities and chronicity of depression. Journal of
Psychiatric Research, 39, 285-293.
Herrmann, L. L., Le Masurier, M., & Ebmeier, K. P. (2008). White matter
hyperintensities in late life depression: A systematic review. Journal of
Neurology, Neurosurgery, and Psychiatry, 79, 619-624.
Hickie, I., Scott, E., Mitchell, P., Wilhelm, K., Austin, M.-P., & Bennett, B. (1995).
Subcortical hyperintensities on magnetic resonance imaging: Clinical correlates
and prognostic signficance in patients with severe depression. Biological
Psychiatry, 37, 151-160.
Hickie, I., Scott, E., Wilhelm, K., & Brodaty, H. (1996). Subcortical hyperintensities on
magnetic resonance imaging in patients with severe depression—A longitudinal
evaluation. Biological Psychiatry, 42, 367-374.
99
Hoshide, S., Kario, K., Mitsuhashi, T., Sato, Y., Umeda, Y., Katsuki, T., & Shimada, K.
(2001). Different patterns of silent cerebral infarct in patients with coronary
artery disease or hypertension. American Journal of Hypertension, 14, 509-515.
Jagust, W. J., Zheng, L., Harvey, D. J., Mack, W. J., Vinters, H. V., Weiner, M. W., Ellis,
W. G., Zarow, C., Mungas, D., Reed, B. R., Kramer, J. H., Schuff, N., DeCarli,
C., & Chui, H. C. (2008). Neuropathological basis of magnetic resonance
imaging in aging and dementia. Annals of Neurology, 63, 72-80.
Jefferson, A. L., Tate, D. F., Poppas, A., Brickman, A. M., Paul, R. H., Gunstad, J., &
Cohen, R. A. (2007). Lower cardiac output is associated with greater white
matter hyperintensities in older adults with cardiovascular disease. Journal of the
American Geriatrics Society, 55, 1044-1048.
Jorm, A. F. (2001). History of depression as a risk factor for dementia: An updated
review. Australian and New Zealand Journal of Psychiatry, 35, 776-781.
Jorm, A. F., Anstey, K. J., Christensen, H., dePlater, G., Kumar, R., Wen, W., &
Sachdev, P. (2005). MRI hyperintensities and depressive symptoms in a
community sample of individual 60-64 year olds. American Journal of
Psychiatry, 162, 699-705.
Kanner, A. M. (2004). Is major depression a neurologic disorder with psychiatric
symptoms? Epilepsy and Behavior, 5, 636-644.
Kendler, K. S., Gardner, C. O., Fiske, A., & Gatz, M. (2009). Major depression and
coronary artery disease in the Swedish Twin Registry. Archives in General
Psychiatry, 66, 857-863.
Kendler, K. S., Gatz, M., Gardner, C. O., & Pedersen, N. L. (2006). A Swedish national
twin study of lifetime major depression. American Journal of Psychiatry, 163,
109-114.
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, 699-700.
Kim, M. J., Hamilton, J. P., & Gotlib, I. H. (2008). Reduced caudate gray matter volume
in women with major depressive disorder. Psychiatry Research: Neuroimaging,
164, 114-122.
Kohler, S., Thomas, A. J., Lloyd, A., Barber, R., Almeida, O.P., & O’Brien, J. T. (2010).
White matter hyperintensities, cortisol levels, brain atrophy, and continuing
cognitive deficits in late-life depression. The British Journal of Psychiatry, 196,
143-149.
100
Krishnan, K. R. R., Hays, J. C., Tupler, L. A., George, L. K., & Blazer, D. G. (1995).
Clinical and phenomenological comparisons of late-onset and early-onset
depression. American Journal of Psychiatry, 152, 785-788.
Krishnan, M. S., O’Brien, J. T., Firbank, M. J., Pantoni, L., Carlucci, G., Erkinjuntti, T.,
Wallin, A., Wahlund, L.- O., Scheltens, P., van Straaten, E. C. W., & Inzitari, D.
(2006). Relationship between periventricular and deep white matter lesions and
depressive symptoms in older people. The LADIS study. International Journal
of Geriatric Psychiatry, 21, 983-989.
Lavretsky, H., Zheng, L., Weiner, M. W., Mungas, D., Reed, B., Kramer, J. H., Jagust,
W., Chui, H., & Mack, W. J. (2008). The MRI brain correlates of depressed
mood, anhedonia, apathy, and anergia in older adults with and without cognitive
impairment or dementia.
Lesser, I. M., Boone, K. B., Mehringer, C. M., Wohl, M. A., Miller, B. L., & Berman, N.
G. (1996). Cognition and white matter hyperintensities in older adult patients.
American Journal of Psychiatry, 153, 1280-1287.
Liao, Y.-C., Liu, R.-S., Lee, Y.-C., Sun, C.-M., Liu, C.-Y., Wang, P.-S., Wang, P.-H., &
Liu, H.-C. (2003). Selective hyperperfusion of anterior cingulated gyrus in
depressed AD patients: A brain SPECT finding by statistical parametric mapping.
Dementia and Geriatric Cognitive Disorders, 16, 238-244.
Lichtenstein, P., DeFaire, U., Floderus, B., Svartengren, M., Svedberg, P., & Pedersen,
N. L. (2002). The Swedish Twin Registry: A unique resource for clinical,
epidemiological and genetic studies. Journal of Internal Medicine, 252, 184-205.
Lind, K., Jonsson, M., Karlsson, I., Sjogren, M., Wallin, A., & Edman, A. (2006).
Depressive symptoms and white matter changes in patients with dementia.
International Journal of Geriatric Psychiatry, 21, 119-125.
Lund and Manchester Groups. (1994). Clinical and neuropathological criteria for
frontotemporal dementia. Journal of Neurology and Neurosurgical Psychiatry,
57, 416-418.
McKeith, I. G., Galasko, D., Kosaka, K., Perry, E. K., Dickson, D. W., Hansen, L. A., et
al. (1996). Consensus guidelines for the clinical and pathological diagnosis of
dementia with Lewy bodies (DLB): Report of the consortium on DLB
international workshop. Neurology, 47, 1113-1124.
101
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 Department of Health and Human
Services task force on Alzheimer’s disease. Neurology, 34, 939-944.
Morris, J.C., Ernesto, C., Schafer, K., Coats, M., Leon, S., Sano, M., et al. (1997).
Clinical Dementia Rating training and reliability in multicenter studies: The
Alzheimer’s disease cooperative study experience. Neurology, 48, 1508-1510.
Morris, J. C., Heyman, A., Mohs, R. C., Hughes, J. P., van Belle, G., Fillenbaum, G., et
al. (1989). The Consortium to Establish a Registry for Alzheimer’s Disease
(CERAD): 1. Clinical and neuropsychological assessment of Alzheimer’s
disease. Neurology, 39, 1159-1165.
Mueller, S. G., Mack, W. J., Mungas, D., Kramer, J. H., Cardenas-Nicolson, V.,
Lavretsky, H., Greene, M., Schuff, N., Chui, H. C., & Weiner, M. W. (2010).
Influences of lobar gray matter and white matter lesion load on cognition and
mood. Psychiatry Research, 181, 90-96.
Neary D., Snowden, J. S., Gustafson, L., Passant, U., Stuss, D., Black, S., et al. (1998).
Frontotemporal lobar degeneration: A consensus on clinical diagnostic criteria.
Neurology, 51, 1546-1554.
O’Brien, J., Desmond, P., Ames, D., Schweitzer, I., Harrigan, S., & Tress, B. (1996). A
magnetic resonance imaging study of white matter lesions in depression and
Alzheimer’s disease. British Journal of Psychiatry, 168, 477-485.
Patten, SB. (1999). Long-term medical conditions and major depression in the Canadian
population. Canadian Journal of Psychiatry, 44, 151-157.
Pedersen, N. L., Miller, B. L., Wetherell, J. L., Vallo, J., Toga, A. W., Knutson, N.,
Mehringer, C. M., Small, G. W., & Gatz, M. (1999). Neuroimaging findings in
twins discordant for Alzheimer’s disease. Dementia and Geriatric Cognitive
Disorders, 10, 51-58.
Rabins, P. V., Pearlson, G. D., Aylward, E., Kumar, A. J., Dowell, K. (1991). Cortical
magnetic resonance imaging changes in elderly inpatients with major depression.
American Journal of Psychiatry, 148, 617-620.
Roman, G. C., Tatemichi, T. K., Erkinjuntti, T., Cummings, J. L., Masdeu, J. C., Garcia,
J. H., et al. (1993). Vascular dementia: Diagnostic criteria for research studies.
Report of the NINDS-AIREN international workshop. Neurology, 43, 256-260.
Savitz, J. B., & Drevets, W. C. (2009). Imaging phenotypes of major depressive
disorder: Genetic correlates. Neuroscience, 164, 300-330.
102
Schweitzer, I., Tuckwell, V., Ames, D. & O’Brien, J. (2001). Structural neuroimaging
studies in late-life depression. World Journal of Biological Psychiatry, 2, 83-88.
Schweitzer, I., Tuckwell, V., O’Brien, J., & Ames, D. (2002). Is late onset depression a
prodrome to dementia? International Journal of Geriatric Psychiatry, 17, 997-
1005.
Shah, P. J., Glabus, M. F., Goodwin, G. M., & Ebmeier, K. P. (2002). Chronic,
treatment-resistant depression and right fronto-striatal atrophy. British Journal of
Psychiatry, 180, 434-440.
Sheline, Y. I., Gado, M.H., & Kraemer, H.C. (2003). Untreated depression and
hippocampal volume loss. American Journal of Psychiatry, 160, 1516-1518.
Starkstein, S. E., Mizrahi, P., Capizzano, A. A., Acion, L., Brockman, S., & Power, B. D.
(2009). Neuroimaging correlates of apathy and depression in Alzheimer’s
disease. The Journal of Neuropsychiatry and Clinical Neurosciences, 21, 259-
265.
Steffens, D.C., O’Connor, C.M., Jiang, W. J., Pieper, C. F., Kuchibhatla, M. N., Arias,
R. , M., Look, A., Davenport, C., Gonzalez, M. B., Krishnan, K. R. R. (1999).
The effect of major depression on functional status in patients with coronary
artery disease. Journal of the American Geriatrics Society, 47, 319-322.
Stevens, J. M., Barber, C. J., Keslake, R., Broz, M., & Barker, S. (1991). Extended use
of cranial CT in the evaluation of patients with stroke and transient ischaemic
attacks. Neuroradiology, 33, 200-206.
Taylor, W. D., & Krishnan, R. R. (2003). Structural brain investigations in affective
disorders. In J. C. Soares (Ed.), Brain Imaging in Affective Disorders (pp.53-78).
New York: Marcel Dekker, Inc.
Teodorczuk, A., O’Brien, J. T., Firbank, M. J., Pantoni, L., Poggesi, A., Erkinjunnti, T.,
Wallin, A., Wahlund, L.-O., Gouw, A., Waldemar, G., Schmidt, R., Ferro, J. M.,
Chabriat, H., Bazner, H., & Inzitari, D. (2007). White matter changes and late-
life depressive symptoms. British Journal of Psychiatry, 191, 212-217.
Teri, L., & Wagner, A. W. (1991). Assessment of depression in patients with
Alzheimer’s disease: Concordance between informants. Psychology of Aging, 6,
280-285.
Thomas, A. J., Ferrier, I. N., Kalaria, R. N., Perry, R. H., Brown, A., & O’Brien, J. T.
(2001). A neuropathological study of vascular factors in late-life depression.
Journal of Neurology, Neurosurgery and Psychiatry, 70, 83-87.
103
Thomas, A. J., & O’Brien, J. T. (2008). Depression and cognition in older adults.
Current Opinion in Psychiatry, 21, 8-13.
Thomas, A. J., O’Brien, J. T., Davis, S., Ballard, C., Barber, R., Kalaria, R. N., & Perry,
R. H. (2002). Ischemic basis for deep white matter hyperintensities in major
depression. A neuropathological study. Archives of General Psychiatry, 59, 785-
792.
Uehara, T., Tabuchi, M., & Mori, E. (1999). Risk factors for silent cerebral infarcts in
subcortical white matter and basal ganglia. Stroke, 30, 378-382.
van der Flier, W.M., van den Heuval, D.M.J., Weverling-Rijnsburger, A.W.E., Bollen,
E.L.E.M, Westendorp, R.G.J., van Buchem, M.A., et al. (2002). Magnetization
transfer imaging in normal aging, mild cognitive impairment, and Alzheimer’s
disease. Annals of Neurology, 52, 62-67.
Victoroff, J., Nielson, K., & Mungas, D. (1997). Caregiver and clinical assessment of
behavioral disturbances: The California Dementia Behavior Questionnaire.
International Psychogeriatrics, 9, 155-174.
Wahlund, L. O., Barkhof, F., Fazekas, F., Bronge, L., Augustin, M., Sjogren, M., Wallin,
A., Ader, H., Leys, D., Pantoni, L., Pasquier, F., Erkinjuntti, T., & Scheltens, P.
(2001). A new rating scale for age-related white matter changes applicable to
MRI and CT. Stroke, 32, 1318-1322.
Wetherell, J. L., Gatz, M., Johansson, B., & Pedersen, N. L. (1999). History of
depression and other psychiatric illness as risk factors for Alzheimer disease in a
twin sample. Alzheimer Disease and Associated Disorders, 13, 47-52.
Wragg, R. E., & Jeste, D. V. (1989). Overview of depression and psychosis in
Alzheimer’s disease. American Journal of Psychiatry, 146, 577-687.
Yaffe, K., Blackwell, T., Gore, R., Sands, L., Reus, V., & Browner, W. S. (1999).
Depressive symptoms and cognitive decline in nondemented elderly women.
Archives of General Psychiatry, 56, 425-430.
104
Appendix A
Visual Rating Worksheet and Corresponding Variable Names
105
Appendix B
Comparisons between Logistic Regression and GEE Method Results
1. Table 22—Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset
Depression (Narrow) compared with No Depression for All Dementia and Alzheimer’s
Disease Using Logistic Regression and Generalized Estimating Equations
2. Table 23—Risk of Subcortical White Matter Hypodensities in Late-Onset Depression
(Narrow) compared with No Depression for All Dementia and Alzheimer’s Disease
Using Logistic Regression and Generalized Estimating Equations
3. Table 24—Number of Striatal Hypodensities in Late-Onset Depression (Narrow)
compared with No Depression for All Dementia and Alzheimer’s Disease Using Logistic
Regression and Generalized Estimating Equations
4. Table 25—Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset
Depression (Broad) compared with No Depression for All Dementia and Alzheimer’s
Disease Using Logistic Regression and Generalized Estimating Equations
5. Table 26—Risk of Subcortical White Matter Hypodensities in Late-Onset Depression
(Broad) compared with No Depression for All Dementia and Alzheimer’s Disease Using
Logistic Regression and Generalized Estimating Equations
106
6. Table 27—Number of Striatal Hypodensities in Late-Onset Depression (Broad)
compared with No Depression for All Dementia and Alzheimer’s Disease Using Logistic
Regression and Generalized Estimating Equations
Table 22
Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset Depression (Narrow) compared with No Depression for
All Dementia and Alzheimer’s Disease Using Logistic Regression and Generalized Estimating Equations
-WMH = Frontal deep white matter hypodensities absent; +WMH = Frontal deep white matter hypodensities present; OR =
Binary logit odds ratio estimate; CI = confidence interval; AD = Alzheimer’s disease
*Late-onset depression group compared to no depression group, controlling for gender and history of transient ischemic attack.
**Late-onset depression group compared to no depression group, controlling for gender.
No Depression Late-Onset Depression
-WMH
N (%)
+WMH
N (%)
-WMH
N (%)
+WMH
N (%)
Logistic Regression
OR (95% CI)
*
GEE
OR (95% CI)
**
All Dementia 75 (56.8) 57 (43.2) 22 (61.1) 14 (38.9) 0.96 (0.44, 2.08)
1.02 (0.46, 2.25)
AD Only
59 (62.1) 36 (37.9) 16 (64.0) 9 (36.0) 0.88 (0.35, 2.25) 0.94 (0.37, 2.38)
107
Table 23
Risk of Subcortical White Matter Hypodensities in Late-Onset Depression (Narrow) compared with No Depression for All
Dementia and Alzheimer’s Disease Using Logistic Regression and Generalized Estimating Equations
OR = Ordinal cumulative logit odds ratio estimate, controlling for age at time of CT scan; CI = confidence interval; AD = Alzheimer’s
disease
Level of WMH: 0 = none, 1 = one discrete hypodensity, 2 = more than one discrete hypodensity, 3 = confluent hypodensities.
* All numbers for Level of WMH given in terms of N (%). **Late-onset depression group compared to no depression group.
No Depression Late-Onset Depression
Level of
WMH
*
0 1 2 3 0 1 2 3
Logistic Regression
OR (95% CI)
**
GEE
OR (95% CI)
**
All Dementia
62
(47.0)
16
(12.1)
25
(18.9)
29
(22.0)
17
(47.2)
3
(8.33)
9
(25.0)
7
(19.4)
0.97
(0.48, 1.92)
0.96
(0.48, 1.92)
AD Only
48
(50.5)
11
(11.6)
18
(18.9)
18
(18.9)
12
(48.0)
2
(8.0)
5
(20.0)
6
(24.0)
1.16
(0.48, 2.58)
1.12
(0.48, 2.60)
108
Table 24
Number of Striatal Hypodensities in Late-Onset Depression (Narrow) compared with No Depression for All Dementia and
Alzheimer’s Disease Using Logistic Regression and Generalized Estimating Equations
LOD = late-onset depression
*p-values less than 0.05 are bolded.
No Depression LOD
ANOVA
(β, SE, p)
GEE
(β, SE, p)
All Dementia
N 132 36
Striatum, mean (SD) 0.48 (0.96) 0.86 (1.53) 0.38, 0.21, 0.0693 0.57, 0.34, 0.0916
Alzheimer’s Disease Only
N 95 25
Striatum, mean (SD) 0.42 (0.92) 0.96 (1.72) 0.54, 0.25, 0.0317 0.82, 0.41, 0.0441
109
Table 25
Risk of Frontal Lobe Deep White Matter Hypodensities in Late-Onset Depression (Broad) compared with No Depression for
All Dementia and Alzheimer’s Disease Using Logistic Regression and Generalized Estimating Equations
-WMH = Frontal deep white matter hypodensities absent; +WMH = Frontal deep white matter hypodensities present; OR = Binary logit
odds ratio estimate, controlling for gender and history of TIA; CI = confidence interval; AD = Alzheimer’s disease
*Late-onset depression group compared to no depression group, controlling for gender and history of transient ischemic attack.
**Late-onset depression group compared to no depression group, controlling for gender.
No Depression Late-Onset Depression
-WMH
N (%)
+WMH
N (%)
-WMH
N (%)
+WMH
N (%)
Logistic Regression
OR (95% CI)
*
GEE
OR (95% CI)
**
All Dementia
52 (51.5) 49 (48.5) 44 (66.7) 22 (33.3) 0.55 (0.29, 1.06) 0.62 (0.32, 1.20)
AD Only 39 (56.5) 30 (43.5) 35 (70.0) 15 (30.0) 0.60 (0.27, 1.32) 0.62 (0.28, 1.36)
110
Table 26
Risk of Subcortical White Matter Hypodensities in Late-Onset Depression (Broad) compared with No Depression for All
Dementia and Alzheimer’s Disease Using Logistic Regression and Generalized Estimating Equations
OR = Ordinal cumulative logit odds ratio estimate, controlling for age at time of CT scan; CI = confidence interval; AD = Alzheimer’s
disease.
Level of WMH: 0 = none, 1 = one discrete hypodensity, 2 = more than one discrete hypodensity, 3 = confluent hypodensities.
* All numbers given in terms of N (%). **Late-onset depression group compared to no depression group.
No Depression Late-Onset Depression
Level of
WMH
*
0 1 2 3 0 1 2 3
Logistic Regression
OR (95% CI)
**
GEE
OR (95% CI)
**
All
Dementia
46
(45.5)
14
(13.9)
17
(16.8)
24
(23.8)
32
(48.5)
5
(7.6)
17
(25.8)
12
(18.2)
0.91 (0.51, 1.62)
0.91 (0.51, 1.62)
AD Only
34
(49.3)
9
(13.0)
12
(17.4)
14
(20.3)
25
(50.0)
4
(8.0)
11
(22.0)
10
(20.0)
0.97 (0.49, 1.94)
0.97 (0.48, 1.95)
111
Table 27
Number of Striatal Hypodensities in Late-Onset Depression (Broad) compared with No Depression for All Dementia and
Alzheimer’s Disease Using Logistic Regression and Generalized Estimating Equations
LOD = late-onset depression
No Depression LOD
ANOVA
(β, SE, p)
GEE
(β, SE, p)
All Dementia
N 101 66
Striatum, mean (SD) 0.53 (1.02) 0.62 (1.25) 0.09, 0.17, 0.6210 0.14, 0.31, 0.6537
Alzheimer’s Disease Only
N 69 50
Striatum, mean (SD) 0.48 (0.99) 0.62 (1.34) 0.14, 0.21, 0.4973 0.25, 0.39, 0.5138
112
Abstract (if available)
Abstract
Studies have shown that white matter changes and other neuropathology frequently found in individuals with dementia, also may be related to late-life depression, or "vascular dementia." As these two conditions frequently coexist, the question of whether there is a relationship between these neuropathological changes and depression among individuals with manifest dementia has not been established. The aim of the present study was to examine whether there were neuroanatomical differences evident on the CT scans of individuals with dementia based upon depression onset (no depression versus early-onset versus late-onset) and history of late-life depression (any episode of depression after age 60). We hypothesized that individuals with dementia and late-onset depression and/or late-life depression would be more likely than non-depressed individuals with dementia to exhibit frontal lobe deep white matter, subcortical white matter, and subcortical gray matter hypodensities. We found that compared to individuals with Alzheimer's disease and no depression, individuals with Alzheimer's disease and late-onset depression had a greater number of striatal hypodensities (gray matter hypodensities in the caudate nucleus and lentiform nucleus, which includes the putamen and globis pallidus). In addition, we found that although there were no differences between the non-depressed and late-onset and late-life depression groups with respect to white matter hypodensities, the late-onset depression and late-life depression groups in comparison to the non-depressed group displayed a significantly higher degree of global functional impairment, as well as impairment within the domains of memory, orientation, and in the context of their home activities and hobbies. These findings suggest that late-onset depression may be a process that is distinct from the neurodegenerative changes caused by Alzheimer's disease.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Accuray of subjective cognitive complaints in a longitudinal context: the effect of depression and dementia status
PDF
Affective neuropsychiatric symptoms and neural connectivity in the early stages of Alzheimer’s disease
PDF
Dementia, co-morbid depressive symptoms, and mortality in older adults: a twin study
PDF
Communication and home-based interventions in Alzheimer’s disease -- a review
PDF
Crosscultural differences in dementia diagnosis and care-seeking in Hispanic and non-Hispanic white outpatients
PDF
Predicting cognitive decline and dementia in elderly twins from indicators of early life oral health
PDF
Untreated hyperglycemia associated with tau pathology and worse cognitive performance in older adults
PDF
Are life events differentially associated with dementia risk by gender? A twin study
Pdf
Alzheimer's disease: risk factors, value, and alternative payment models
PDF
Neural crest-derived cranial pericytes and their dysfunction in disease
PDF
The role of the locus coeruleus in Alzheimer’s disease and cerebrovascular function: insights from neuroimaging, neuropsychology, and biofluid markers
PDF
Age-related androgen depletion and the development of Alzheimer's disease
PDF
Public awareness campaign for dementia in Taiwan
PDF
Cardiovascular disease risk factors and cognitive function
PDF
Comparison of participant and study partner predictions of cognitive impairment in the Alzheimer's disease neuroimaging initiative 3 study
PDF
Neuroimaging markers of risk & resilience to brain aging and dementia
PDF
Alzheimer’s disease: dysregulated genes, ethno-racial disparities, and environmental pollution
PDF
Blood-brain barrier pathophysiology in cognitive impairment and injury
PDF
Vascular contributions to brain aging along the Alzheimer's disease continuum
PDF
Longitudinal relationships between bereavement and physical health‐related quality of life in middle‐ to older‐aged women
Asset Metadata
Creator
Brommelhoff, Jessica Anne
(author)
Core Title
Underlying neural mechanisms of depression and dementia
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
11/24/2010
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Alzheimer's disease,computed tomography,dementia,late-life depression,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Gatz, Margaret (
committee chair
), Mack, Wendy J. (
committee member
), Pedersen, Nancy L. (
committee member
), Prescott, Carol A. (
committee member
), Spann, Bryan M. (
committee member
)
Creator Email
brommelh@usc.edu,jbrommelhoff@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m3563
Unique identifier
UC1407730
Identifier
etd-Brommelhoff-4148 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-409934 (legacy record id),usctheses-m3563 (legacy record id)
Legacy Identifier
etd-Brommelhoff-4148.pdf
Dmrecord
409934
Document Type
Dissertation
Rights
Brommelhoff, Jessica Anne
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
Alzheimer's disease
computed tomography
dementia
late-life depression