Journal of Alzheimer’s Disease 54 (2016) 185–199 185

DOI 10.3233/JAD-160209
IOS Press

Late-Life Depressive Symptoms

and Lifetime History of Major Depression:
Cognitive Deficits are Largely Due
to Incipient Dementia rather than
Depression
Kathrin Hesera,∗ , Markus Bleckwennb , Birgitt Wiesec , Silke Mamonec , Steffi G. Riedel-Hellerd ,
Janine Steind , Dagmar Lühmanne , Tina Posselte , Angela Fuchsf , Michael Pentzekf , Siegfried
Weyererg , Jochen Werleg , Dagmar Weegh , Horst Bickelh , Christian Brettschneideri , Hans-Helmut
Königi , Wolfgang Maiera,j , Martin Scherere,1 and Michael Wagnera,j,1 for the AgeCoDe Study
Group
a Department of Psychiatry and Psychotherapy, University of Bonn, Bonn, Germany
b Department of General Practice and Family Medicine, University of Bonn, Bonn, Germany
c Working Group Medical Statistics and IT Infrastructure, Institute of General Practice, Hannover Medical

School, Hannover, Germany

d Institute of Social Medicine, Occupational Health and Public Health, University of Leipzig, Leipzig, Germany
e Department of Primary Medical Care, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
f Institute of General Practice, Medical Faculty, Heinrich-Heine University Düsseldorf, Düsseldorf, Germany
g Central Institute of Mental Health, Medical Faculty Mannheim/Heidelberg University, Mannheim, Germany
h Department of Psychiatry, Technical University Munich, Munich, Germany
i Department of Health Economics and Health Services Research, Hamburg Center for Health Economics,

University Medical Center Hamburg-Eppendorf, Hamburg, Germany

j DZNE, Center for Neurodegenerative Diseases, Bonn, Germany

Handling Associate Editor: Montse Alegret

Accepted 24 May 2016

Abstract.
Background: Late-life depression is frequently accompanied by cognitive impairments.
Objective: Whether these impairments indicate a prodromal state of dementia, or are a symptomatic expression of depression
per se is not well-studied.

1 shared
last authorship. Freud-Str. 25, 53105 Bonn, Germany. Tel.: +49 228 287 19827;
∗ Correspondence to: Kathrin Heser, PhD, Department of Fax: +49 228 287 90 19827; E-mail: Kathrin.Heser@ukb.uni-
Psychiatry and Psychotherapy, University of Bonn, Sigmund- bonn.de.

ISSN 1387-2877/16/$35.00 © 2016 – IOS Press and the authors. All rights reserved
186 K. Heser et al. / Depression, Cognition, Subsequent Dementia

Methods: In a cohort of very old initially non-demented primary care patients (n = 2,709, mean age = 81.1 y), cognitive
performance was compared between groups of participants with or without elevated depressive symptoms and with or
without subsequent dementia using ANCOVA (adjusted for age, sex, and education). Logistic regression analyses were
computed to predict subsequent dementia over up to six years of follow-up. The same analytical approach was performed
for lifetime major depression.
Results: Participants with elevated depressive symptoms without subsequent dementia showed only small to medium cog-
nitive deficits. In contrast, participants with depressive symptoms with subsequent dementia showed medium to very large
cognitive deficits. In adjusted logistic regression models, learning and memory deficits predicted the risk for subsequent
dementia in participants with depressive symptoms. Participants with a lifetime history of major depression without sub-
sequent dementia showed no cognitive deficits. However, in adjusted logistic regression models, learning and orientation
deficits predicted the risk for subsequent dementia also in participants with lifetime major depression.
Conclusion: Marked cognitive impairments in old age depression should not be dismissed as “depressive pseudodementia”,
but require clinical attention as a possible sign of incipient dementia. Non-depressed elderly with a lifetime history of major
depression, who remained free of dementia during follow-up, had largely normal cognitive performance.

Keywords: Cognition, dementia, depression, depressive symptoms, executive function, memory

INTRODUCTION were associated with poorer cognitive function on all

Depression and dementia are frequent disorders in
the elderly [1]. Meta-analyses showed that depression
increases the risk for dementia [2]. Late-life depres-
sion could be of particular relevance [3]. Dementia
is defined by cognitive impairment and depression
can be accompanied by cognitive deficits as well
[4]. Late-life depression was hallmarked by deficits
in executive function [5], although deficits in other
cognitive domains such as processing speed or visu-
ospatial abilities were found as well [6]. Marked
cognitive deficits in late-life depression that are not
caused by neurodegeneration due to dementia and
that show improvement after the depression symp-
tomatology remitted are often considered as a sign of
“depressive pseudodementia” [7].
Longitudinal studies are essential to examine
whether cognitive deficits in depressed participants
are a sign of incipient dementia or a rather benign
symptomatic expression of depression per se. If
late-life depression with cognitive deficits is a pro-
dromal state of dementia, then the degree, and
possibly also the profile of cognitive impairments
should predict future dementia, because dementia
(and particularly Alzheimer’s dementia (AD)) is pre-
ceded by cognitive impairment (and particularly by
memory impairments). Furthermore, “pure” late-life
depression without subsequent dementia might be
accompanied by rather minor cognitive deficits. How-
ever, to date this topic is understudied and only a few
studies on the prediction of dementia by using the per-
formance in different cognitive domains in depressed
subjects exist. The available results are ambiguous.
Ganguli et al. [8] found that depressive symptoms
tests cross-sectionally, but only subsequent demen-
tia and not depression was associated with cognitive
decline longitudinally. Whereas Vilalta-Franch et al.
[9] found that late-onset depression with executive
dysfunction increased the risk for dementia, Potter et
al. [10] found that deficits in recognition and execu-
tive function in older depressed participants predicted
subsequent dementia over about 5.5 y. Rushing et al.
[11] found that long-term memory impairment pre-
dicted the development of AD over 7 y of follow-up.
von Gunten et al. [12] found that immediate prose
free recall and visual organization were associated
with later dementia in depressed patients with subjec-
tive memory complaints. However, Vilalta-Franch et
al. [9] did not include different cognitive domains in
their analyses. Potter et al. [10], Rushing et al. [11],
and von Gunten et al. [12] used a rather small sample
without a non-depressed comparison group.
For clinicians it would be important to know which
depressed patients are at an increased risk to develop
subsequent dementia. We hypothesized that execu-
tive dysfunction in depressed participants would be
a symptom of late-life depression independent of
incipient dementia, whereas learning and memory
deficits in depressed participants would be associ-
ated with an increased risk for subsequent dementia.
In accordance with the concept of cognitive reserve
[13], participants with better global cognition, higher
cortical functions, and higher intellectual function
should have a reduced risk for subsequent dementia,
and this assumption should also apply to depressed
participants. We therefore examined the cognitive
performance of participants with elevated depres-
sive symptoms or with a history of lifetime major
 K. Heser et al. / Depression, Cognition, Subsequent Dementia
Fig. 1. Sample selection flowchart.
depression and compared the predictive power of dif-
ferent cognitive domains for subsequent dementia in
depressed and non-depressed participants.
MATERIALS AND METHODS
Sample
We present data from a prospective multi-center
study (the German Study on Aging, Cognition, and
Dementia in Primary Care Patients (AgeCoDe)).
At baseline, participants (≥75 y) were recruited via
general practitioners (GPs) in six German cities
(Bonn, Düsseldorf, Hamburg, Leipzig, Mannheim,
and Munich). The local ethics committees approved
the study. The interview was conducted in-person
by a trained research assistant. Follow-up assess-
ments after baseline in 2003/2004 were conducted
every 18 months. History of major depression was
first assessed at follow-up 1. Therefore, the present
study referred to data of covariates and predictors
assessed at follow-up 1 and included information of
dementia status until follow-up 5. Follow-up 5 was
conducted approximately 6 y after follow-up 1. The
sample initially consisted of 3,327 baseline partici-
pants. Participants with a diagnosis of dementia at
baseline or follow-up 1 were excluded (n = 173). Data
was not available due to other reasons for 420 partic-
ipants (see Fig. 1).
Sample characteristics are given in Table 1.
Covariates
In the adjusted analyses, we controlled for age
at follow-up 1, sex (female, male), and education
187
according to CASMIN [14], as these demographi-
cal covariates are often associated with depression,
dementia, and neuropsychological performance.
Depressive symptoms
Depressive symptoms at follow-up 1 were assessed
using the short version of the Geriatric Depression
Scale [15]. The scale contains 15 items. A cut-off
score of at least 6 was supposed to be indicative of
clinically relevant depression and was used by several
authors [16]. Therefore, this cut-off score was used
to create two depressive symptom groups (with and
without elevated depressive symptoms).
Major depression
Lifetime prevalence of major depression according
to DSM-IV criteria [17] was assessed at follow-up 1
using a shortened and modified version of section E
of the Composite International Diagnostic Interview
(CIDI; [18]; for further details on modification see
[19]).
Cognitive performance
The Consortium to Establish a Registry for
Alzheimer’s Disease (CERAD) verbal fluency and
memory subtest scores and the domain scores of
the Structured Interview for Diagnosis of Demen-
tia of Alzheimer type, Multi-infarct Dementia and
Dementia of other Aetiology according to DSM-
IV and ICD-10 criteria (SIDAM) cognitive section
(SISCO) were used to assess different cognitive func-
tions. The Mini-Mental State Examination (MMSE;
[20]), which is part of the SISCO, was used to assess
global cognitive functioning. Semantic verbal fluency
(animals named within 60 seconds) was used as a
measure of executive function. From the CERAD
ten-word list, the scores of immediate recall, delayed
recall, and recognition were considered. The sum
of three consecutive learning trials of the CERAD
word list was computed to obtain the immediate recall
score. The number of words recalled after ten min-
utes represented the delayed recall score. The word
list recognition score was computed by the sum of
the ten correctly recognized words and ten new words
that were correctly rejected (minus 10). With respect
to SISCO, the cognitive domains of orientation (10
items), memory (20 items), higher cortical functions
(20 items), and intellectual function (5 items) were
computed.
188 K. Heser et al. / Depression, Cognition, Subsequent Dementia

Table 1
Sample characteristics with status of dementia diagnosis until follow-up 5 and status of predictors at follow-up 1
No dementia until FU5 Dementia until FU5
n = 2,305 n = 404
Age, M (SD) 80.88 (3.47) 82.30 (3.41)
Sex, n (%)
Female 1484 (64.4) 283 (70.0)
Male 821 (35.6) 121 (30.0)
Education, n (%)
Low 1385 (60.1) 267 (66.1)
Medium 651 (28.2) 95 (23.5)
High 269 (11.7) 42 (10.4)
GDS, M (SD) 2.34 (2.36) 2.85 (2.62)
GDS Cutoff 6, n (%)
No 2057 (89.2) 334 (82.7)
Yes 242 (10.5) 69 (17.1)
Missing, n (%) 6 (0.3) 1 (0.2)
Major Depression (lifetime), n (%)
No 1789 (77.6) 317 (78.5)
Yes 212 (9.2) 32 (7.9)
Missing, n (%) 304 (13.2) 55 (13.6)
Classification of education according to CASMIN [14]: low:< 9 years of education, medium: 9 to 13 years of
education, high: >13 years of education; FU5, follow-up 5; M, mean; SD, standard deviation; GDS, Geriatric
Depression Scale [15].

Dementia diagnoses was tested by (planned) contrasts using the group

The dementia assessment was based on the SIDAM
[21] implemented by a trained research assistant.
When SIDAM could not be assessed, a Global Dete-
rioration Scale [22] score of at least 4 and/or a
Blessed Dementia Rating Scale [23] score were used.
Each case of incident dementia and its aetiologi-
cal subtype was validated by a geriatric expert. AD
was diagnosed according to the DSM-IV criteria
[17]. Vascular dementia was diagnosed according to
the NINDS-AIREN criteria [24]. All-cause dementia
(including AD, vascular dementia, mixed forms, spe-
cific types of dementia such as Lewy body dementia,
or dementia caused by substance abuse) was used as
dementia outcome variable. The presence or absence
of dementia was diagnosed at every follow-up assess-
ment. Incident dementia cases from follow-up 2 until
follow-up 5 were considered, whereas prevalent cases
until follow-up 1 were excluded.
Statistical analyses
The cognitive test performance mean scores of
four groups of participants with and without depres-
sive symptoms at follow-up 1 according to the GDS
cut-off score of 6 and with and without subse-
quent dementia until follow-up 5 were compared
using ANCOVA analyses controlling for age, sex,
and education. The significance of group differences
without depressive symptoms without subsequent
dementia as reference group. The effect sizes (ES)
of group differences between the three depressive
symptoms and/or dementia groups compared to the
“healthy” reference group without depressive symp-
toms without subsequent dementia were computed
by using the spreadsheet of Thalheimer and Cook
[25]. Logistic regression analyses were computed to
investigate which cognitive performance tests pre-
dicted the incidence of subsequent dementia when
all cognitive domains except the MMSE were con-
sidered; age, sex, and education were entered at
first step using the “enter” procedure and cognitive
test performance scores were entered in a sec-
ond step using the “backwards” selection procedure
(likelihood-ratio) in which all variables are included
altogether in the regression model and are tested
for exclusion until each selected predictor signifi-
cantly improves the model. Apart from the MMSE,
all cognitive test performance scores were entered
(i.e., verbal fluency, immediate recall, delayed recall,
recognition, orientation, memory, higher cortical
functions, intellectual function). The same analyti-
cal approach using ANCOVA analyses and logistic
regression models was also applied to participants
with and without a history of lifetime depression.
The level of significance was set to ␣ = 0.05. Sta-
tistical analyses were conducted with SPSS Version
22.
 K. Heser et al. / Depression, Cognition, Subsequent Dementia
RESULTS
Depressive symptoms
In ANCOVA analyses adjusting for age, sex, and
education, the cognitive performance at follow-up 1
of four groups of participants without/with depressive
symptoms and without/with subsequent demen-
tia were compared. Participants without depressive
symptoms at follow-up 1 without subsequent demen-
tia until follow-up 5 were used as “healthy” reference
group in planned contrasts and effect size esti-
mation (see Table 2). Tests of between-subject
effects in ANCOVA were significant for all cogni-
tive test performances (all ps < 0.001). The groups
with depressive symptoms and/or subsequent demen-
tia performed significantly worse on all cognitive
tests (all ps < 0.05).
The effect sizes of test performance differences
between participants without depressive symptoms
without subsequent dementia (“control group”) and
the other three groups (“treatment groups”) were
estimated (see last column in Table 2). Results are
given in Figure 2. The differences between the
control group and depressed participants without
subsequent dementia were predominantly small (ES
range: 0.19 – 0.45). Medium effect sizes were found
for verbal fluency (ES = 0.40) and intellectual func-
tion (ES = 0.45). CERAD immediate (ES = 0.39), and
delayed recall (ES = 0.37) effect size differences were
small, but approached the cut-off to medium effect
sizes. These results indicated that late-life depressive
symptoms not due to subsequent dementia are asso-
ciated with significant but rather minor deficits only.
In contrast, predominantly very large effect sizes
were found for the cognitive performance differences
between participants with depressive symptoms with
subsequent dementia and the control group (ES
range: 0.44 – 1.41). Pronounced deficits of the group
with depressive symptoms with subsequent dementia
were indicated by very large to large effect size dif-
ferences compared to the reference group regarding
CERAD recognition (ES = 1.41), CERAD immedi-
ate recall (ES = 1.41), SISCO memory (ES = 1.34),
CERAD delayed recall (ES = 1.27), SISCO orien-
tation (ES = 1.17), MMSE (ES = 1.13), and verbal
fluency (ES = 1.02). Medium deficits of participants
with elevated depressive symptoms with subsequent
dementia were found for SISCO higher cortical
functions (ES = 0.63) and SISCO intellectual func-
tion (ES = 0.44). The effect sizes for differences
between participants without depressive symptoms
189
with subsequent dementia and the control group
were intermediate in relation to both other dis-
ease groups (see Fig. 2); the greatest effect sizes
were found for learning and memory parameters
in this group (SISCO memory: ES = 0.77; CERAD
immediate recall: ES = 0.72; CERAD delayed recall:
ES = 0.70).
In a next step, we examined whether subsequent
dementia risk was predicted by different cognitive
test performances in participants with and without
depressive symptoms. Logistic regression analy-
ses were used to predict incident dementia until
follow-up 5 by covariates and cognitive performances
at follow-up 1 separately for participants without
depressive symptoms and with depressive symptoms
at follow-up 1 (see Table 3). The covariates age, sex,
and education were entered at step 1. All cognitive
test performances (except MMSE) were entered at
step 2 using the stepwise inclusion (backward elimi-
nation). In participants without depressive symptoms
at follow-up 1, worse performances on verbal flu-
ency (OR = 0.94, 95% CI = 0.92 – 0.97, p < 0.001)
and three measures of memory and learning (i.e.,
CERAD immediate recall: OR = 0.93, 95% CI = 0.89
– 0.97, p = 0.001; CERAD delayed recall: OR = 0.90,
95% CI = 0.83 – 0.98, p = 0.014; SISCO memory:
OR = 0.83, 95% CI = 0.78 – 0.89, p < 0.001) predicted
subsequent dementia. In participants with depressive
symptoms, CERAD immediate recall (OR = 0.83,
95% CI = 0.76 – 0.90, p < 0.001) and SISCO mem-
ory (OR = 0.75, 95% CI = 0.66 – 0.86, p < 0.001)
were also retained in the logistic regression model,
whereas CERAD delayed recall and verbal fluency
were eliminated. Unexpectedly, higher intellectual
function (OR = 1.55, 95% CI = 1.00 – 2.40, p = 0.05)
was associated with a marginally increased risk for
subsequent dementia in participants with depressive
symptoms. This association was not found in for-
mer analyses when only covariates but not cognitive
performance on other tests were included.
History of lifetime major depression
Tests of between-subjects effects in age-, sex-,
and education-adjusted ANCOVA were significant
for all cognitive test performances (all ps < 0.01; see
Table 4). Planned contrasts showed that participants
with a history of lifetime major depression with-
out subsequent dementia did not significantly differ
from the “healthy” reference group without lifetime
major depression without subsequent dementia on
any cognitive test. Participants without lifetime major
190 K. Heser et al. / Depression, Cognition, Subsequent Dementia

Table 2
Comparison of cognitive test performances adjusted for age, sex, and education (ANCOVA) between four groups of participants with and
without elevated depressive symptoms at follow-up 1 and with and without subsequent dementia until follow-up 5
Groups Tests of Effect sizes
between-
subjects-effects
1a 2b 3c 4d p d
Dep- Dem- Dep+Dem- Dep- Dem+ Dep+Dem+
MMSE, M (S.E.) 28.03 (0.04) 27.57 (0.10) 27.02 (0.09) 25.98 (0.20) <0.001 2 versus 1 : 0.26*
(p < 0.001)
3 versus 1 : 0.56*
(p < 0.001)
4 versus 1 : 1.13*
(p < 0.001)
Verbal fluency, M (S.E.) 20.57 (0.12) 18.41 (0.34) 17.40 (0.30) 15.05 (0.65) <0.001 2 versus 1 : 0.40*
(p < 0.001)
3 versus 1 : 0.58*
(p < 0.001)
4 versus 1 : 1.02*
(p < 0.001)
CERAD immediate recall, 19.83 (0.09) 18.24 (0.25) 16.92 (0.21) 14.12 (0.47) <0.001 2 versus 1 : 0.39*
M (S.E.) (p < 0.001)
3 versus 1 : 0.72*
(p < 0.001)
4 versus 1 : 1.41*
(p < 0.001)
CERAD delayed recall, 6.02 (0.05) 5.18 (0.14) 4.44 (0.12) 3.15 (0.26) <0.001 2 versus 1 : 0.37*
M (S.E.) (p < 0.001)
3 versus 1 : 0.70*
(p < 0.001)
4 versus 1 : 1.27*
(p < 0.001)
CERAD recognition, 9.08 (0.03) 8.81 (0.10) 8.30 (0.08) 7.17 (0.18) <0.001 2 versus 1 : 0.20*
M (S.E.) (p = 0.008)
3 versus 1 : 0.57*
(p < 0.001)
4 versus 1 : 1.41*
(p < 0.001)
SISCO orientation, M (S.E.) 9.82 (0.01) 9.70 (0.03) 9.67 (0.03) 9.29 (0.06) <0.001 2 versus 1 : 0.26*
(p < 0.001)
3 versus 1 : 0.32*
(p < 0.001)
4 versus 1 : 1.17*
(p < 0.001)
SISCO memory, M (S.E.) 17.25 (0.05) 16.75 (0.14) 15.52 (0.12) 14.21 (0.26) <0.001 2 versus 1 : 0.22*
(p < 0.001)
3 versus 1 : 0.77*
(p < 0.001)
4 versus 1 : 1.34*
(p < 0.001)
SISCO higher cortical 17.61 (0.04) 17.27 (0.12) 17.09 (0.10) 16.47 (0.22) <0.001 2 versus 1 : 0.19*
functions, M (S.E.) (p = 0.007)
3 versus 1 : 0.29*
(p < 0.001)
4 versus 1 : 0.63*
(p < 0.001)

(Continued)
 K. Heser et al. / Depression, Cognition, Subsequent Dementia 191

Table 2
(Continued)
Groups Tests of Effect sizes
between-
subjects-effects
1a 2b 3c 4d p d
Dep- Dem- Dep+Dem- Dep- Dem+ Dep+Dem+
SISCO intellectual function, 4.76 (0.01) 4.55 (0.04) 4.61 (0.03) 4.56 (0.08) <0.001 2 versus 1 : 0.45*
M (S.E.) (p < 0.001)
3 versus 1 : 0.32*
(p < 0.001)
4 versus 1 : 0.44*
(p = 0.012)
Dep, depression; Dem, dementia; M, mean; S.E., standard error; MMSE, Mini-Mental State Examination [20]; CERAD, Consortium to
Establish a Registry for Alzheimer’s Disease; SISCO, score of SIDAM (Structured Interview for Diagnosis of Dementia of Alzheimer
type, Multi-infarct Dementia and Dementia of other Aetiology according to DSM-IV and ICD-10 criteria; [21]). Effect sizes: negligible:
–0.15–0.14, small: 0.15–0.39, medium: 0.40–0.74, large: 0.75–1.09, very large: 1.10–1.44, huge: ≥1.45. *significant difference of cognitive
performance between the groups (p values are given in parentheses). a n ranges from 2,040 to 2,057; b n ranges from 237 to 242; c n ranges
from 327 to 334; d n ranges from 64 to 69.

Fig. 2. Inverted effect sizes of cognitive test performance adjusted for age, sex, and education comparing three groups of participants with
elevated depressive symptoms at follow-up 1 and/or subsequent dementia until follow-up 5 with the reference group of participants without
elevated depressive symptoms at follow-up 1 without subsequent dementia until follow-up 5 are depicted.

depression with subsequent dementia performed sig-
nificantly worse on all cognitive tests compared to
the reference group (all ps < 0.01). Participants with
a history of lifetime major depression with subse-
quent dementia performed significantly worse on all
cognitive tests compared to the reference group (all
ps < 0.01), except no significant differences between
verbal fluency and intellectual function performances
(both ps < 0.10).
Although participants without lifetime major
depression without subsequent dementia showed the
best performances on most but not all cognitive
measures, we decided to use these participants as
control group as planned contrasts showed that differ-
ences between participants without major depression
without dementia and participants with lifetime
major depression without subsequent dementia, who
descriptively showed higher scores on SISCO mem-
ory and SISCO intellectual function, were not
significant (see Table 4).
The magnitude of effect sizes of the cognitive
performance differences between groups with life-
time major depression and/or subsequent dementia
(“treatment groups”) and the reference group with-
out lifetime major depression without later incident
dementia (“control group”) is displayed in Figure 3.
The effect sizes of differences between the “healthy”
reference group and the group of participants with
192 K. Heser et al. / Depression, Cognition, Subsequent Dementia

Table 3
Prediction of subsequent dementia until follow-up 5 in participants without elevated depressive symptoms (GDS score <6) and in participants
with elevated depressive symptoms (GDS score ≥6) by cognitive test performances at follow-up 1 in logistic regression analyses (covariates:
enter, cognitive test scores: backward elimination; likelihood ratio)
Participants without elevated depressive
symptoms at follow-up 1 (n = 2,355)
p value Odds Ratio 95% Confidence
Interval
Age <0.001 1.07 1.03–1.11
Sex 0.022 0.71 0.53–0.95
Education 0.133
Education (1) 0.752 1.05 0.77–1.43
Education (2) 0.045 1.58 1.01–2.46
Verbal fluency <0.001 0.94 0.92–0.97
CERAD immediate recall 0.001 0.93 0.89–0.97
CERAD delayed recall 0.014 0.90 0.83–0.98
SISCO memory <0.001 0.83 0.78–0.89
Rate of correct classification: no incident dementia = 98.6%, incident dementia = 11.1%, overall = 86.5%
-2 Log-Likelihood = 1603.054; Cox & Snell R2 = 0.115; Nagelkerke’s R2 = 0.208.

Participants with elevated depressive

symptoms at follow-up 1 (n = 298)
p value Odds Ratio 95% Confidence
Interval
Age 0.594 1.02 0.94–1.12
Sex 0.259 0.64 0.30–1.39
Education 0.395
Education (1) 0.675 1.19 0.54–2.62
Education (2) 0.176 2.27 0.69–7.44
CERAD immediate recall <0.001 0.83 0.76–0.90
SISCO memory <0.001 0.75 0.66–0.86
SISCO intellectual functions 0.050 1.55 1.00–2.40
Rate of correct classification: no incident dementia = 94.5%, incident dementia = 31.7%, overall = 81.2%
-2 Log-Likelihood=237.127; Cox & Snell R2 = 0.210; Nagelkerke’s R2 = 0.327.
GDS, Global Deterioration Scale; classification of education according to CASMIN [14]; education, low, medium, and high level, low
education was used as reference category; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; SISCO, score of SIDAM
(Structured Interview for Diagnosis of Dementia of Alzheimer type, Multi-infarct Dementia and Dementia of other Aetiology according to
DSM-IV and ICD-10 criteria; [21]).

lifetime major depression without incident demen-
tia were all negligible (ES range: 0.00 – 0.14). The
effect sizes of differences between the “healthy”
reference group and the group of participants
without lifetime major depression with subsequent
dementia were medium (MMSE: ES = 0.64; verbal
fluency: ES = 0.69; CERAD recognition: ES = 0.56;
SISCO orientation: ES = 0.41) to large in the cog-
nitive domains of learning and memory (CERAD
delayed recall: ES = 0.83; CERAD immediate recall:
ES = 0.84; SISCO memory: ES = 0.92), but also
small effect sizes were found (SISCO higher corti-
cal function: ES = 0.31; SISCO intellectual function:
ES = 0.15). The effect sizes of differences between
the “healthy” reference group and the group of
participants with lifetime major depression with sub-
sequent dementia until follow-up 5 ranged from
small (verbal fluency: ES = 0.30; SISCO intellec-
tual function: ES = 0.21), medium (SISCO higher
cortical functions: ES = 0.59; CERAD recognition:
ES = 0.70; CERAD immediate recall: ES = 0.71), and
large (CERAD delayed recall: ES = 0.75; SISCO
memory: ES = 0.85; MMSE: ES = 1.07) to even huge
(SISCO orientation: ES = 1.48).
In a next step, we examined whether subsequent
dementia risk was predicted by different cognitive
test performances in participants without and in par-
ticipants with lifetime major depression in logistic
regression analyses (see Table 5). In participants
without lifetime major depression, worse perfor-
mances on verbal fluency (OR = 0.93, 95% CI = 0.90
– 0.96, p < 0.001) and three measures of memory
and learning (CERAD immediate recall: OR = 0.93,
95% CI = 0.89 – 0.98, p = 0.003; CERAD delayed
recall: OR = 0.90, 95% CI = 0.83 – 0.99, p = 0.02;
SISCO memory: OR = 0.82, 95% CI = 0.77 – 0.88,
p < 0.001) predicted subsequent dementia (as in the
group of participants without depressive symptoms,
 K. Heser et al. / Depression, Cognition, Subsequent Dementia 193

Table 4
Comparison of cognitive test performances adjusted for age, sex, and education (ANCOVA) between four groups of participants with and
without history of lifetime major depression and with and without subsequent dementia until follow-up 5
Groups Tests of Effect sizes
between-
subjects-effects
1a 2b 3c 4d p d
Dep-Dem- Dep+Dem- Dep- Dem+ Dep+Dem+
MMSE, M (S.E.) 28.08 (0.04) 27.88 (0.11) 27.00 (0.09) 26.27 (0.28) <0.001 2 versus 1 : 0.12
(p = 0.085)
3 versus 1 : 0.64*
(p < 0.001)
4 versus 1 : 1.07*
(p < 0.001)
Verbal fluency, M 20.40 (0.13) 20.00 (0.38) 16.63 (0.31) 18.77 (0.96) <0.001 2 versus 1 : 0.07
(S.E.) (p = 0.316)
3 versus 1 : 0.69*
(p < 0.001)
4 versus 1 : 0.30
(p = 0.094)
CERAD immediate 19.66 (0.09) 19.46 (0.27) 16.45 (0.22) 16.96 (0.68) <0.001 2 versus 1 : 0.05
recall, M (S.E.) (p = 0.486)
3 versus 1 : 0.84*
(p < 0.001)
4 versus 1 : 0.71*
(p < 0.001)
CERAD delayed 5.97 (0.05) 5.70 (0.15) 4.22 (0.12) 4.39 (0.38) <0.001 2 versus 1 : 0.13
recall, M (S.E.) (p = 0.084)
3 versus 1 : 0.83*
(p < 0.001)
4 versus 1 : 0.75*
(p < 0.001)
CERAD recognition, 9.10 (0.04) 8.93 (0.10) 8.18 (0.08) 7.92 (0.26) <0.001 2 versus 1 : 0.10
M (S.E.) (p = 0.116)
3 versus 1 : 0.56*
(p < 0.001)
4 versus 1 : 0.70*
(p < 0.001)
SISCO orientation, 9.82 (0.01) 9.82 (0.03) 9.64 (0.03) 9.19 (0.09) <0.001 2 versus 1 : 0.00
M (S.E.) (p = 0.996)
3 versus 1 : 0.41*
(p < 0.001)
4 versus 1 : 1.48*
(p < 0.001)
SISCO memory, 17.27 (0.05) 17.31 (0.14) 15.32 (0.12) 15.48 (0.37) <0.001 2 versus 1 : 0.02
M (S.E.) (p = 0.768)
3 versus 1 : 0.92*
(p < 0.001)
4 versus 1 : 0.85*
(p < 0.001)
SISCO higher cortical 17.70 (0.04) 17.47 (0.12) 17.18 (0.10) 16.71 (0.31) <0.001 2 versus 1 : 0.14
functions, M (S.E.) (p = 0.079)
3 versus 1 : 0.31*
(p < 0.001)
4 versus 1 : 0.59*
(p = 0.002)

(Continued)
194 K. Heser et al. / Depression, Cognition, Subsequent Dementia

Table 4
(Continued)
Groups Tests of Effect sizes
between-
subjects-effects
1a 2b 3c 4d p d
Dep-Dem- Dep+Dem- Dep- Dem+ Dep+Dem+
SISCO intellectual 4.74 (0.02) 4.75 (0.04) 4.62 (0.04) 4.56 (0.11) 0.004 2 versus 1 : 0.01
function, M (S.E.) (p = 0.930)
3 versus 1 : 0.15*
(p = 0.001)
4 versus 1 : 0.21
(p = 0.095)
Dep, depression; Dem, dementia; M, mean; S.E., standard error; MMSE, Mini-Mental State Examination [20]; CERAD, Consortium to
Establish a Registry for Alzheimer’s Disease; SISCO, score of SIDAM (Structured Interview for Diagnosis of Dementia of Alzheimer
type, Multi-infarct Dementia and Dementia of other Aetiology according to DSM-IV and ICD-10 criteria; [21]). Effect sizes: negligible:
–0.15–0.14, small: 0.15–0.39, medium: 0.40–0.74, large: 0.75–1.09, very large: 1.10–1.44, huge: ≥1.45. *significant difference of cognitive
performance between the groups (p values are given in parentheses). a n ranges from 1,774 to 1,789; b n ranges from 207 to 212; c n ranges
from 306 to 317; d n = 32.

Fig. 3. Inverted effect sizes of cognitive test performance adjusted for age, sex, and education comparing three groups of participants with
history of lifetime major depression and/or subsequent dementia until follow-up 5 with the reference group of participants without history
of lifetime major depression without subsequent dementia until follow-up 5 are depicted.

cf. Table 3). In participants with lifetime major
depression, CERAD immediate recall (OR = 0.88,
95% CI = 0.79 – 0.97, p = 0.013) and SISCO orien-
tation (OR = 0.43, 95% CI = 0.22 – 0.83, p = 0.011)
were retained in the logistic regression model to pre-
dict subsequent dementia.
DISCUSSION
As late-life depression can be a discrete disorder
or a prodrome of dementia, we studied the cognitive
performance of participants with and without late-
life depressive symptoms and regarded whether or
not dementia was developed subsequently. Compared
to the reference group without depressive symp-
toms and without subsequent dementia, participants
with depressive symptoms and/or subsequent demen-
tia performed significantly worse in age-, sex-, and
education-adjusted analyses. However, participants
with depressive symptoms in late-life that did not
subsequently develop dementia within the period
of observation predominantly showed rather small
deficits on most cognitive tests according to effect
sizes; medium deficits of participants with late-life
depressive symptoms without subsequent dementia
compared to the reference group were only found
for verbal fluency and intellectual function perfor-
mance. Verbal fluency is a measure of executive
function, and several studies found that executive
function deficits are a cognitive component of late-
life depression [5]. In contrast, pronounced cognitive
deficits indicated by large to very large effect sizes
were found for participants with late-life depressive
symptoms and subsequent dementia especially in
 K. Heser et al. / Depression, Cognition, Subsequent Dementia 195

Table 5
Prediction of subsequent dementia until follow-up 5 in participants without history of lifetime major depression and in participants with
history of lifetime major depression by cognitive test performances at follow-up 1 in logistic regression analyses (covariates: enter, cognitive
test scores: backward elimination; likelihood ratio)
Participants without history of major depression
at follow-up 1 (n = 2067)
p value Odds Ratio 95% Confidence Interval
Age 0.001 1.06 1.02–1.10
Sex 0.031 0.71 0.53–0.97
Education 0.140
Education (1) 0.693 1.07 0.77–1.48
Education (2) 0.047 1.60 1.01–2.53
Verbal fluency <0.001 0.93 0.90–0.96
CERAD immediate recall 0.003 0.93 0.89–0.98
CERAD delayed recall 0.020 0.90 0.83–0.99
SISCO memory <0.001 0.82 0.77–0.88
Rate of correct classification: no incident dementia = 97.8%, incident dementia = 17.8%, overall = 86.1%
-2 Log-Likelihood = 1419.951; Cox & Snell R2 = 0.136; Nagelkerke’s R2 = 0.241.

Participants with history of major depression

at follow-up 1 (n = 239)
p value Odds Ratio 95% Confidence Interval
Age 0.370 1.06 (0.93–1.21)
Sex 0.616 1.28 (0.49–3.33)
Education 0.191
Education (1) 0.088 0.39 (0.13–1.15)
Education (2) 0.387 0.46 (0.08–2.65)
CERAD immediate recall 0.013 0.88 (0.79–0.97)
SISCO orientation 0.011 0.43 (0.22–0.83)
Rate of correct classification: no incident dementia = 99.5%, incident dementia = 12.5%, overall = 87.9%
-2 Log-Likelihood = 158.947; Cox & Snell R2 = 0.115; Nagelkerke´s R2 = 0.211.
Classification of education according to CASMIN [14]; education, low, medium, and high level, low education was used as reference
category; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; SISCO, score of SIDAM (Structured Interview for Diagnosis
of Dementia of Alzheimer type, Multi-infarct Dementia and Dementia of other Aetiology according to DSM-IV and ICD-10 criteria; [21]).

the domains of learning and memory. Interestingly,
the strongest effect was found for recognition per-
formance, which replicates the neuropsychological
finding that restricted recognition performance is a
sign of cognitive impairment and not a consequence
of depression in late-life [26]. As pronounced impair-
ments in depressed participants were confined to the
group with incipient dementia, the term “pseudode-
mentia” [7] will in most cases be inappropriate for the
co-occurrence of depression and cognitive impair-
ment in old age. Due to the reviewed findings on
its reversibility and potential biological mechanisms,
Kennedy [27] concluded that depressive pseudo-
dementia is more closely related to true dementia
than previously assumed. In accordance with our
results, Sáez-Fonseca et al. [28] showed that depres-
sive pseudodementia in the elderly was associated
with an increased risk to develop dementia over seven
years. The effect sizes for differences between non-
depressed participants with subsequent dementia and
the control group were intermediate in relation to both
other disease groups in general; the highest effect
sizes were found for learning and memory parameters
in this group.
When cognitive subdomains were entered in
covariate-adjusted logistic regression analyses,
higher scores of verbal fluency and of three measures
of learning and memory (CERAD immediate recall,
CERAD delayed recall, SISCO memory) predicted
a decreased risk for subsequent dementia in the
group without depressive symptoms, whereas the
risk for subsequent dementia in the group with
depressive symptoms was not predicted by verbal
fluency, but significantly decreased when higher
scores of two measures of learning and memory
(CERAD immediate recall, SISCO memory) were
attained. These results were in accordance with our
assumptions that executive dysfunction (represented
by verbal fluency) in participants with depressive
symptoms should preliminary be a consequence of
late-life depression, whereas it indicates an incipient
dementia process in nondepressed participants. As
expected, learning and memory deficits were an
indicator of the subsequent risk for dementia in both
196 K. Heser et al. / Depression, Cognition, Subsequent Dementia
groups. The effects of recognition performance on
subsequent dementia in participants with depressive
symptoms that were found in ANCOVA analyses
were not found in logistic regression analyses prob-
ably due to intercorrelations with the other cognitive
performance measures. Unexpectedly, we found that
higher scores on SISCO intellectual function were
associated with an increased risk for subsequent
dementia in the depressed participant group in the
logistic regression analyses when the performance
on other cognitive domains was controlled. As this
association was not expected, similar results of other
studies should be presented.
Depressive symptoms were associated with sub-
sequent AD only in the higher-educated group in
a Dutch sample; the authors hypothesized that a
greater cognitive reserve of the participants with
higher education has buffered the early cognitive
decline processes of AD, whereas the noncognitive
symptoms such as depression were not positively
influenced by cognitive reserve [29, 30]. These
authors also found that the association between
depressive symptoms and AD was stronger when
memory performance was controlled. It remains to be
investigated if meta-cognitive processes or awareness
of subjective cognitive impairment might be involved
in this context. In accordance with this post hoc inter-
pretation, Spitznagel et al. [31] found that depressive
symptoms were associated with the awareness of
cognitive impairment in participants with higher cog-
nitive reserve, whereas no association was found in
participants with lower cognitive reserve in a sample
of participants with questionable or mild dementia.
In addition to late-life depression assessed by
depressive symptoms, the lifetime history of major
depression was also considered. The results for life-
time major depression were less consistent than the
results for depressive symptoms regarding the cogni-
tive deficits, probably due to the fact that the former
mentioned group was more heterogeneous than the
last-mentioned. Interestingly, participants with life-
time major depression without subsequent dementia
did not perform significantly worse on any cognitive
domain compared to the healthy participant group
without lifetime major depression without subse-
quent dementia. As expected, participants without
lifetime major depression with subsequent demen-
tia performed significantly worse than the healthy
comparison group with the most pronounced deficits
on learning and memory measures. Participants with
lifetime major depression and subsequent dementia
did not significantly perform worse on all, but on
several cognitive domains compared to the reference
group. Despite medium to large deficits on learn-
ing and memory domains, this group showed large
deficits on MMSE and huge deficits on SISCO orien-
tation. Verbal fluency and intellectual functions were
not significantly impaired in participants with life-
time major depression and subsequent dementia, but
deficits approached the significance level; the small
group sample size should be considered in that con-
text. In the logistic regression model for participants
without lifetime major depression, the cognitive per-
formance tests that predicted subsequent dementia
were the same as those that predicted subsequent
dementia in participants without depressive symp-
toms (i.e., verbal fluency, CERAD immediate and
delayed recall, SISCO memory). In participants with
lifetime major depression, only CERAD immediate
recall and SISCO orientation deficits were associ-
ated with an increased risk for subsequent dementia,
whereas the other cognitive domains did not pre-
dict the risk for subsequent dementia. In another
study, orientation was used as an indicator of long-
term memory, and long-term memory was a better
predictor of AD than executive function in this sam-
ple of older depressed participants with a mean
age of about 68 y [12]. The importance of orienta-
tion in the prediction of subsequent dementia risk
in participants with lifetime major depression is
interesting, as it might imply associations with the
hippocampus, which is involved in depression and
dementia [32], beside orientation functions. Deficits
in the orientation domain in participants with lifetime
major depression independent of current depres-
sive symptoms might be interpreted as a warning
sign by clinicians for a possibly upcoming dementia
process.
The results of our study match with the results
of another longitudinal study with younger elderly
participants which found that although depressive
symptoms were associated with cognitive deficits
cross-sectionally, they did not predict longitudinal
cognitive decline over up to 12 y of observation, but
instead incipient dementia might have caused these
impairments [8]. Our study has several strengths such
as its longitudinal design and an extensive assess-
ment of neuropsychological tests and psychosocial
variables in a large elderly sample. Although our
sample is not representative of the general popu-
lation, using primary care patients was an attempt
to approach representativity as a high percentage of
very old individuals are regularly treated by gen-
eral practitioners. It should be noted that our sample
 K. Heser et al. / Depression, Cognition, Subsequent Dementia
consisted of initially non-hospitalized, community-
dwelling elderly aged 75 and older. Whether our
findings and conclusions hold true for severe, clin-
ically treated cases of old age depression will need
to be shown. Depressive symptoms were assessed
with an instrument that was designed for geriatric
samples. Additionally, a clinical questionnaire of his-
tory of lifetime major depression was assessed. In
contrast to other studies that also examined cogni-
tion in late-life depression and the risk of subsequent
dementia, we had a non-depressed comparison group.
Besides these strengths, there are also limitations
of our study. Although we assessed major depres-
sion following CIDI, our standardized interview was
partly abbreviated and modified. Although our sam-
ple of participants with depressive symptoms was
larger compared to other studies, we only had 311
participants with elevated depressive symptoms and
244 participants with lifetime major depression in
our sample. Data on history of major depression was
missing for about 13% of the sample, but the reported
lifetime prevalence of about 10% was comparable to
the prevalence found by others in an older cohort [33].
Unfortunately, verbal fluency was the only executive
function measure assessed, but learning and mem-
ory were assessed with different subtests. Compared
to the commonly long-lasting process of dementia
development, the period of follow-up assessment of
about 6 y in this study was rather short. Conversion
to dementia beyond the observation period could not
be considered.
In sum, we provided new evidence suggesting that
elevated late-life depressive symptoms without sub-
sequent dementia are associated with only minor
cognitive deficits, whereas participants with ele-
vated depressive symptoms and subsequent dementia
showed pronounced cognitive deficits. The result that
higher intellectual functions increased the risk for
subsequent dementia in participants with depressive
symptoms, after the other cognitive test scores were
controlled, was not expected, but effects of cognitive
reserve or awareness of personal cognitive deficits
might provide post hoc explanations. Learning and
memory deficits in participants with late-life depres-
sive symptoms were most indicative of subsequent
dementia. Verbal fluency deficits did not predict the
risk for subsequent dementia in depressed partic-
ipants, although verbal fluency was predictive in
nondepressed participants of our study. Executive
function deficits might go back to late-life depression
in that subsample and might not be caused by incipi-
ent dementia. Orientation deficits might be a warning
197
sign for an incident dementia process in participants
with a lifetime history of major depression. Thus,
older depressive individuals with objectively veri-
fiable and substantial learning and memory deficits
had an increased risk for subsequent dementia as
these deficits largely did not result from depressive
symptoms and lifetime major depression. Therefore,
clinicians should take these deficits seriously, apply
further cognitive and dementia diagnostic evaluation,
and closely monitor these patients with neuropsycho-
logical tests. This provides the opportunity for early
dementia detection and preventive intervention such
as health care or medical treatment in the future.
ACKNOWLEDGMENTS
We want to thank both all participating patients
and their general practitioners for their good collab-
oration.
MEMBERS OF THE AGECODE STUDY
GROUP
Principal Investigators*: Wolfgang Maier, Martin
Scherer
Heinz-Harald Abholz, Christian Brettschneider,
Cadja Bachmann, Horst Bickel, Wolfgang Blank,
Hendrik van den Bussche, Sandra Eifflaender-Gorfer,
Marion Eisele, Annette Ernst, Angela Fuchs, Kathrin
Heser, Frank Jessen, Hanna Kaduszkiewicz, Teresa
Kaufeler, Mirjam Köhler, Hans-Helmut König,
Alexander Koppara, Carolin Lange, Diana Lubisch,
Tobias Luck, Dagmar Lühmann, Melanie Luppa,
Manfred Mayer, Edelgard Mösch, Michael Pentzek,
Tina Posselt, Jana Prokein, Steffi Riedel-Heller,
Susanne Röhr, Anna Schumacher, Janine Stein,
Susanne Steinmann, Franziska Tebarth, Michael
Wagner, Klaus Weckbecker, Dagmar Weeg, Jochen
Werle, Siegfried Weyerer, Birgitt Wiese, Steffen
Wolfsgruber, Thomas Zimmermann.
*Hendrik van den Bussche (2002-2011)
This publication is part of the German Research
Network on Dementia (KND), the German Research
Network on Degenerative Dementia (KNDD), and
the Study on Needs, Health Service Use, Costs
and Health-related Quality of Life in a large Sam-
ple of Oldest-old Primary Care Patients (85+)
(AgeQualiDe) and was funded by the German Fed-
eral Ministry of Education and Research (grants
KND 01GI0102, 01GI0420, 01GI0422, 01GI0423,
01GI0429, 01GI0431, 01GI0433 and 01GI0434;
198 K. Heser et al. / Depression, Cognition, Subsequent Dementia
grants KNDD 01GI0710, 01GI0711, 01GI0712,
01GI0713, 01GI0714, 01GI0715 and 01GI0716;
and grants AgeQualiDe 01GY1322A, 01GY1322B,
01GY1322 C, 01GY1322D, 01GY1322E, 01GY
1322F, 01GY1322 G).
Authors’ disclosures available online (http://j-
alz.com/manuscript-disclosures/16-0209r2).
REFERENCES
[1] Steffens DC, Otey E, Alexopoulos GS, Butters MA, Cuth-
bert B, Ganguli M, Geda YE, Hendrie HC, Krishnan RR,
Kumar A, Lopez OL, Lyketsos CG, Mast BT, Morris
JC, Norton MC, Peavy GM, Petersen RC, Reynolds CF,
Salloway S, Welsh-Bohmer KA, Yesavage J (2006) Per-
spectives on depression, mild cognitive impairment, and
cognitive decline. Arch Gen Psychiatry 63, 130-138.
[2] Ownby RL, Crocco E, Acevedo A, John V, Loewenstein D
(2006) Depression and the risk for Alzheimer disease: Sys-
tematic review, meta-analysis, and metaregression analysis.
Arch Gen Psychiatry 63, 530-538.
[3] Diniz BS, Butters MA, Albert SM, Dew MA, Reynolds
CF 3rd (2013) Late-life depression and risk of vascular
dementia and Alzheimer’s disease: Systematic review and
meta-analysis of community-based cohort studies. Br J Psy-
chiatry 202, 329-335.
[4] Zihl J, Reppermund S, Thum S, Unger K (2010) Neuropsy-
chological profiles in MCI and in depression: Differential
cognitive dysfunction patterns or similar final common path-
way disorder? J Psychiatr Res 44, 647-654.
[5] Herrmann LL, Le Masurier M, Ebmeier KP (2008) White
matter hyperintensities in late life depression: A systematic
review. J Neurol Neurosurg Psychiatry 79, 619-624.
[6] Butters MA, Whyte EM, Nebes RD, Begley AE, Dew MA,
Mulsant BH, Zmuda MD, Bhalla R, Meltzer CC, Pollock
BG, Reynolds CF 3rd, Becker JT (2004) The nature and
determinants of neuropsychological functioning in late-life
depression. Arch Gen Psychiatry 61, 587-595.
[7] Kang H, Zhao F, You L, Giorgetta C, D V, Sarkhel S, Prakash
R (2014) Pseudo-dementia: A neuropsychological review.
Ann Indian Acad Neurol 17, 147-154.
[8] Ganguli M, Du Y, Dodge HH, Ratcliff GG, Chang CC
(2006) Depressive symptoms and cognitive decline in late
life: A prospective epidemiological study. Arch Gen Psychi-
atry 63, 153-160.
[9] Vilalta-Franch J, López-Pousa S, Llinàs-Reglà J, Calvó-
Perxas L, Merino-Aguado J, Garre-Olmo J (2013)
Depression subtypes and 5-year risk of dementia and
Alzheimer disease in patients aged 70 years. Int J Geriatr
Psychiatry 28, 341-350.
[10] Potter GG, Wagner HR, Burke JR, Plassman BL, Welsh-
Bohmer KA, Steffens DC (2013) Neuropsychological
predictors of dementia in late-life major depressive disorder.
Am J Geriatr Psychiatry 21, 297-306.
[11] Rushing NC, Sachs-Ericsson N, Steffens DC (2014)
Neuropsychological indicators of preclinical Alzheimer’s
disease among depressed older adults. Neuropsychol Dev
Cogn B Aging Neuropsychol Cogn 21, 99-128.
[12] von Gunten A, Giannakopoulos P, Duc R (2005) Cognitive
and demographic determinants of dementia in depressed
patients with subjective memory complaints. Eur Neurol
54, 154-158.
[13] Stern Y (2006) Cognitive reserve and Alzheimer disease.
Alzheimer Dis Assoc Disord 20, 112-117.
[14] König W, Lüttinger P, Müller W (1988) A comparative
analysis of the development and structure of educational
systems: Methodological foundations and the construction
of a comparative education scale, CASMIN Working Paper
12, University Mannheim, Mannheim, Germany.
[15] Sheikh JI, Yesavage JA (1986) Geriatric Depression Scale
(GDS): Recent evidence and development of a shorter ver-
sion. In Clinical Gerontology: A Guide to Assessment and
Intervention, TL Brink, ed. Haworth, New York, pp. 165-
173.
[16] Wancata J, Alexandrowicz R, Marquart B, Weiss M,
Friedrich F (2006) The criterion validity of the Geriatric
Depression Scale: A systematic review. Acta Psychiatr
Scand 114, 398-410.
[17] APA (1994) DSM-IV: Diagnostic and Statistical Manual
of Mental Disorders, 4th edition, American Psychological
Association, Washington DC.
[18] Wittchen HU, Pfister H (1997) DIA-X-Interviews [German
version], Swets & Zeitlinger, Frankfurt/Main.
[19] Heser K, Tebarth F, Wiese B, Eisele M, Bickel H, Köhler
M, Mösch E, Weyerer S, Werle J, König HH, Leicht H,
Pentzek M, Fuchs A, Riedel-Heller SG, Luppa M, Prokein
J, Scherer M, Maier W, Wagner M, the AgeCoDe Study
Group (2013) Age of major depression onset, depressive
symptoms, and risk for subsequent dementia: Results of
the German Study on Ageing, Cognition, and Dementia in
Primary Care Patients (AgeCoDe). Psychol Med 43, 1597-
1610.
[20] Folstein MF, Folstein SE, McHugh PR (1975) “Mini-Mental
State”: A practical method for grading the cognitive state of
patients for the clinician. J Psychiatr Res 12, 189-198.
[21] Zaudig M, Hiller W (1996) SIDAM Handbook. Structured
Interview for Diagnosis of Dementia of Alzheimer type,
Multi-infarct Dementia and Dementia of other Aetiology
according to DSM-III-R, DSM-IV, and ICD-10 [German
version], Huber, Bern.
[22] Reisberg B, Ferris SH, de Leon MJ, Crook T (1982)
The Global Deterioration Scale for assessment of primary
degenerative dementia. Am J Psychiatry 139, 1136-1139.
[23] Blessed G, Tomlinson BE, Roth M (1968) The association
between quantitative measures of dementia and of senile
changes in the cerebral grey matter of elderly subjects. Br J
Psychiatry 114, 797-811.
[24] Roman RC, Tatechimi TK, Erkinjuntii T, Cummings JL,
Masdeu JC, Garcia JH, Amaducci L, Orgogozo JM, Brun
A, Hofman A, Moody DM, O’Brien MD, Yamaguchi T,
Grafman J, Drayer BP, Bennett DA, Fisher M, Ogata J,
Kokmen E, Bermejo F, Wolf PA, Gorelick PB, Bick KL,
Pajeau AK, Bell MA, DeCarli C, Culebras A, Korczyn AD,
Bogousslavsky J, Hartmann A, Scheinberg P (1993) Vascu-
lar dementia: Diagnostic criteria for research studies. Report
of the NINDS-AIREN International Workshop. Neurology
43, 250-260.
[25] Thalheimer W, Cook S (2002) How to calculate effect sizes
from published research articles: A simplified methodology.
http://www.bwgriffin.com/gsu/courses/edur9131/content/
Effect Sizes pdf5.pdf, Accessed on January 20, 2016.
[26] Dierckx E, Engelborghs S, De Raedt R, De Deyn PP,
Ponjaert-Kristoffersen I (2007) Differentiation between
mild cognitive impairment, Alzheimer’s disease and depres-
sion by means of cued recall. Psychol Med 37, 747-755.
[27] Kennedy J (2015) Depressive pseudodementia – how
“pseudo” is it really? Old Age Psychiatrist 62, 30-37.
 K. Heser et al. / Depression, Cognition, Subsequent Dementia
[28] Sáez-Fonseca JA, Lee L, Walker Z (2007) Long-term out-
come of depressive pseudodementia in the elderly. J Affect
Disord 101, 123-129.
[29] Geerlings MI, Schmand B, Braam AW, Jonker C, Bouter
LM, van Tilburg W (2000) Depressive symptoms and risk of
Alzheimer’s disease in more highly educated older people.
J Am Geriatr Soc 48, 1092-1097.
[30] Geerlings MI, Schoevers RA, Beekman AT, Jonker C, Deeg
DJ, Schmand B, Adèr HJ, Bouter LM, van Tilburg W (2000)
Depression and risk of cognitive decline and Alzheimer’s
disease: Results of two prospective community-based
studies in The Netherlands. Br J Psychiatry 176,
568-575.
199
[31] Spitznagel MB, Tremont G, Brown LB, Gunstad J (2006)
Cognitive reserve and the relationship between depressive
symptoms and awareness of deficits in dementia. J Neu-
ropsychiatry Clin Neurosci 18, 186-190.
[32] Rodrigues R, Petersen RB, Perry G (2014) Parallels between
major depressive disorder and Alzheimer’s disease: Role of
oxidative stress and genetic vulnerability. Cell Mol Neuro-
biol 34, 925-949.
[33] Kessler RC, Berglund P, Demler O, Jin R, Merikangas
KR, Walters EE (2005) Lifetime prevalence and age-of-
onset distributions of DSM-IV disorders in the National
Comorbidity Survey Replication. Arch Gen Psychiatry 62,
593-602.