Article

Hasselbalch BJ, Knorr U, Kessing LV. Cognitive impairment in the remitted state of unipolar depressive disorder: a systematic review. J Affect Disorder 134: 20-31

Department of Psychiatry, University Hospital of Copenhagen, Rigshospitalet, Denmark.
Journal of Affective Disorders (Impact Factor: 3.38). 12/2010; 134(1-3):20-31. DOI: 10.1016/j.jad.2010.11.011
Source: PubMed

ABSTRACT

It is unclear whether cognitive impairment is prevalent in the remitted state of unipolar disorder.
To evaluate whether cognitive function is impaired in the remitted state in patients with unipolar depression compared with healthy control individuals, and to investigate the association to prior course of illness, i.e. the number, duration and severity of prior depressive episodes.
Systematic search on existing on-line databases and hand-search of original published papers.
A total of 11 studies fulfilled the selection criteria and were included in the review, including a total of 500 patients remitted from unipolar depression and 471 healthy control individuals. In nine of the eleven studies performance on neuropsychological tests was found to be decreased in patients compared to healthy control individuals in at least one of the tests. Methodological drawbacks were prevalent including non-stringent definition of remission and non-correction for multiple testing. Only few studies investigated the association between cognition and prior course of illness and the results were divergent.
Stringent criteria were used in the assessment of eligibility of studies. The studies were first and foremost selected according to the criteria for remission used.
Cognitive dysfunction seems to be present in individuals suffering from unipolar disorder in the remitted state. We recommend that future studies should focus on disentangling the state and trait characteristics of cognitive dysfunction in unipolar disorder and further clarify the associations with clinical phenotype, course of illness and subsyndromal psychopathology. Furthermore, there is a need to identify the cognitive difficulties in individuals suffering from unipolar disorder in relation to psychosocial function, quality of life and risk of recurrence and to assess the effect of treatment intervention on cognitive function.

Full-text

Available from: Lars Kessing, Apr 19, 2016
Review
Cognitive impairment in the remitted state of unipolar depressive disorder:
A systematic review
Bo Jacob Hasselbalch
, Ulla Knorr, Lars Vedel Kessing
Department of Psychiatry, University Hospital of Copenhagen, Rigshospitalet, Denmark
article info abstract
Article history:
Received 3 June 2010
Received in revised form 7 September 2010
Accepted 7 November 2010
Available online 15 December 2010
Background: It is unclear whether cognitive impairment is prevalent in the remitted state of
unipolar disorder.
Aim: To evaluate whether cognitive function is impaired in the remitted state in patients with
unipolar depression compared with healthy control individuals, and to investigate the association
to prior course of illness, i.e. the number, duration and severity of prior depressive episodes.
Method: Systematic search on existing on-line databases and hand-search of original published
papers.
Results: A total of 11 studies fulfilled the selection criteria and were included in the review,
including a total of 500 patients remitted from unipolar depression and 471 healthy control
individuals. In nine of the eleven studies performance on neuropsychological tests was found to
be decreased in patients compared to healthy control individuals in at least one of the tests.
Methodological drawbacks were prevalent including non-stringent definition of remission and
non-correction for multiple testing. Only few studies investigated the association between
cognition and prior course of illness and the results were divergent.
Limitations: Stringent criteria were used in the assessment of eligibility of studies. The studies
were first and foremost selected according to the criteria for remission used.
Conclusion: Cognitive dysfunction seems to be present in individuals suffering from unipolar
disorder in the remitted state. We recommend that future studies should focus on disentangling
the state and trait characteristics of cognitive dysfunction in unipolar disorder and further clarify
the associations with clinical phenotype, course of illness and subsyndromal psychopathology.
Furthermore, there is a need to identify the cognitive difficulties in individuals suffering from
unipolar disorder in relation to psychosocial function, quality of life and risk of recurrence and to
assess the effect of treatment intervention on cognitive function.
© 2010 Elsevier B.V. All rights reserved.
Keywords:
Unipolar Disorder
Cognitive impairment
Cognitive dysfunction
Remission
Euthymia
State
Trait
Scar
Neuropsychology
Psychometrics
Contents
1. Introduction ....................................................... 21
1.1. Objective ..................................................... 21
2. Method ......................................................... 21
2.1. Protocol ...................................................... 21
2.2. Eligibility criteria ................................................. 21
2.3. Strategy of data collection ............................................. 21
2.4. Study selection process. .............................................. 22
Journal of Affective Disorders 134 (2011) 2031
Corresponding author. Tel.: +45 35456147; fax: +45 35456238.
E-mail address: jacob.hasselbalch@regionh.dk (B.J. Hasselbalch).
0165-0327/$ see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jad.2010.11.011
Contents lists available at ScienceDirect
Journal of Affective Disorders
journal homepage: www.elsevier.com/locate/jad
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 1
3. Results ......................................................... 22
3.1. Study selection and data collection process ..................................... 22
3.2. Main ndings ................................................... 22
4. Discussion........................................................ 25
4.1. Main ndings ................................................... 25
4.2. Rationale for the selection criteria used in this review................................ 26
4.2.1. Criteria for the assessment of state of remission .............................. 26
4.2.2. Neuropsychological tests ......................................... 26
4.3. The association between cognitive outcome and the number, duration and subtype/severity of previous episodes . . . 26
4.3.1. Clinical manifestation, subtype/severity .................................. 26
4.3.2. Episodes, number and duration ...................................... 26
4.4. The associations between cognitive outcome and other clinical factors ........................ 27
4.4.1. Age at onset ............................................... 27
4.4.2. The impact of subclinical psychopathology on cognitive test performance .................. 27
4.4.3. The time period elapsed since the patients experienced the last episode of depression ............ 27
4.4.4. Medication and other treatment interventions ............................... 27
4.4.5. Age and gender ............................................. 28
4.4.6. Number of previous hospitalisations .................................... 28
4.4.7. Co-morbid psychiatric disorders ...................................... 28
4.5. General conclusion on the specicity and nature of the neuropsychological ndings ................. 28
5. Estimation of risk of bias in the included studies ..................................... 29
6. Conclusions and recommendations for future research ................................... 29
Role of funding source .................................................... 30
Conict of interest ...................................................... 30
Acknowledgement ...................................................... 30
References .......................................................... 30
1. Introduction
Cognitive impairment has been widely reported in patients
during episodes of major depression e.g. (Burt et al., 1995;
Ravnkilde et al., 2002), in the euthymic phase of bipolar
disorder e.g. (Quraishi and Frangou, 2002), and in rst degree
relatives of patients with unipolar e.g. (Christensen et al., 2006)
and bipolar disorder e.g. (Bora et al., 2009). In bipolar disorder,
attention and executive function seem to be impaired across all
phases of the illness, but in respect to learning memory and
visuospatial skills, there is little evidence that decits reect
more than attentional disturbance (Goodwin and Jamison,
2007). Cognitive impairment may not only be a signicant
factor affecting the individual's ability to function socially and
occupationally in everyday life during episodes of illness and in
remission (Jaeger et al., 2006), but may have a long term impact
on cognition associated with a more deteriorating course of
illness than previously assumed.
A number of reviews and meta-analyses have been
published on cognitive function in the remitted state of
bipolar disorder. Furthermore a number of studies seem to
suggest that people suffering from unipolar disorder may
also present cognitive dysfunctions during remitted states,
but no systematic reviews or meta-analysis has so far been
pre sented on this subjec t.
1.1. Objective
The aim of this study was to conduct a systematic review
and meta-analysis to evaluate if cognitive functi on is
impaired during the remitted state in patients with unipolar
affective disorder compared to healthy control individuals.
Our hypothesis was that development of cognitive dysfunc-
tion is dependent on the number, duration and severity of
previous depressive episodes. Furthermore we aimed to
investigate whether the following factors could contribute
to the observed ndings: 1) age at onset 2) subsyndromal
depressive symptoms 3) current-/lifetime treatment with
psychoactive medication 4) the time elapsed since the
participants suffered the last depressive episode 5) age and
gender 6) co-morbid psychiatric disorder and 7) number of
episodes requiring hospitalizations. Additionally we aimed to
assess the possible sources of bias.
2. Method
2.1. Protocol
Methods of the analysis were specied in advance and
documented in a protocol.
2.2. Eligibility criteria
We assessed the eligibility of originally reported studies
providing data on cognitive function in the remitted state of
unipolar depressive disorder compared to control partici-
pants as measured in performance on neuropsychological
tests.
2.3. Strategy of data collection
Studies were i dentied by searching the MEDLINE-,
EMBASE- and PsychInfo databases of original papers published
between 1980- November 2009 using the specied search
terms: 1) Pub Med Mesh search terms; Construct A: Mood
disorder or unipolar disorder or depressive disorder. Construct
B: Cognition disorders or cognition or neuropsychological tests.
Construct C: Remit* or recover* or euthyme*. 2) PsychInfo
21B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 2
search terms; Construct A: Affective disorder or unipolar
disorder or depressive disorder. Construct B: Cognitive assess-
ment and cognitive ability. Construct C: Remit* or recover*or
euthyme*. 3) EMBASE search terms: Construct A: Major
depression. Construct B: Cognition or cognitive defect or
neuropsychological. Construct C: Remit* or recover* or
euthyme*. Furthermore, database free keyword search of
papers and reviews on affective disorder and cognitive function
in general was performed, and references of retrieved articles
were hand-searched for not previously identied studies.
Papers identied by searching the individual databases, and
by hand-search, were registered in separate ReferenceManager
12.0.2 databases that were nally merged into one single
database, from which duplicates were removed. In the case of
multiple publications including an update on partly or
completely the same patient sample the last published study
was included.
2.4. Study selection process
Papers were included, if they met the following criteria:
1. Were published in English, German or French.
2. Included adult patient samples (age 18).
3. Included a sample of N N 10 individuals remitted from
unipolar depression that was compared to a control group
of people who did not suffer from affective or any other
major psychiatric disorder.
4. Cognitive function was assessed by neuropsychological
tests which provided an estimate of global cognitive
function or specic estimates of cognitive function in the
domains of attention, executive functions, and learning or
memory.
5. The diagnosis of unipolar depressive disorder was estab-
lished according to ICD or DSM diagnostic criteria.
6. Participants were judged to be in a remitted state when
neuropsychological testing was performed, and this was
conrmed by a clinical evaluation according to specied
criteria for remission and using a cut-off score of 8 on the
17-item Hamilton Rating Scale for Depression (HDRS) or
12 on the Montgomery-Asberg Depression Rating Scale
(MADRS).
Using the specied criteria relevant articles were identi-
ed by screening of abstracts, and when found potentially
eligible for inclusion, full articles were retrieved for further
assessment. Final inclusion was done by consensus between
the authors. Data extraction was done using a pre-dened
data extraction sheet.
3. Results
3.1. Study selection and data collection process
Using the above-specied database search strategy, a total
of one hundred and eleven articles were identied. Of these
eighty two were excluded on the initial screening of abstracts
and twenty nine were retrieved for further assessment. By
database free keyword search and hand-searches an addi-
tional forty one articles of interest were identied. From a
total of seventy articles retrieved for further assessment 11
articles were found to meet the criteria for inclusion. Of the
fty seven articles which were not included, two articles were
not included since they did not full criteria one, two since
they did not full criteria two, eleven since they did not full
criteria three, four since they did not full criteria four
and nally thirty eight since they did not ful
l criteria six. Of
the papers which were not included according to criteria
six, twenty two were not included since a strategy using
specied criteria and/or using a cut-off score on the HDRS or
MADRS to assess whether participants were remitted was
not employed, and sixteen since higher cut-off scores (HDRS
range 916, MADRS none) were used.
3.2. Main ndings
Using the above-specied criteria a total of 11 studies
were included in the review, including a total of 500 patients
who, according to our selection criteria, were dened as
remitted from episodes of unipolar depression, and 471
healthy control individuals.
Characteristics of the patients and of the methodology and
diagnostic instruments used in the included studies are
presented in Table 1. Most of the studies included middle-
aged or older patients and healthy control individuals. One
study aimed only to include young adults (Smith et al., 2006)
and two studies aimed only to include older adults (Beats
et al., 1996; Yuan et al., 2008). Two studies used ICD-10 to
establish the diagnosis (Kessing, 1998; Preiss et al., 2009),
eight studies used DSM-VI (Weiland-Fiedler et al., 2004; Clark
et al., 2005a, 2005b; Neu et al., 2005; Paelecke-Habermann
et al., 2005; Smith et al., 2006; Biringer et al., 2007; Nakano
et al., 2008; Yuan et al., 2008) and one study used DSM-III-R
(Beats et al., 1996). The cut-off scores on the HDRS used were
8 in three studies (Kessing, 1998; Clark et al., 2005a, 2005b;
Smith et al., 2006), 7 in two studies (Biringer et al., 2007;
Yuan et al., 2008) and 6 in two studies (Neu et al., 2005;
Nakano et al., 2008). The cut-off scores on the MADRS used
were 12 in one study (Preiss et al., 2009), 9 in one study
(Beats et al., 1996), and 5 in one study (Weiland-Fiedler
et al., 2004). The strategies employed to assess whether
participants were in a remitted state were diverse. Five
studies relied solely on a cut-off score to assess if participants
were in a remitted state (Beats et al., 1996; Kessing, 1998;
Clark et al., 2005a, 2005b; Biringer et al., 2007; Nakano et al.,
2008), whereas in the remaining studies, participants were
deemed clinically remitted for a period of minimum of one
month in one study (Smith et al., 2006), two months in one
study (Preiss et al., 2009), three months in two studies
(Weiland-Fiedler et al., 2004; Paelecke-Habermann et al.,
2005) and six months in two studies (Neu et al., 2005; Yuan
et al., 2008). It was unclear in most of the studies if a strategy
was employed t o asses s whethe r the heal thy c ontrol
individuals suffered from affective disorder or other major
or minor psychiatric disorder.
Results of neuropsychological test in patients compared to
healthy control individuals and the possible predictors of
outcome assessed in the patient sample are summarised in
Table 2.
A large number of different neuropsychological tests were
employed. In nine out of the eleven studies included in this
review, a signicantly decreased performance on neuropsy-
chological tests were found in at least one test assessing
22 B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 3
Table 1
Characteristics of the patients and of the methodology and diagnostic instruments used in the included studies.
Study Design Sample characteristics Strategy employed to establish diagnosis and to asses if patients
were in a remitted state
Neuropsychological tests used
N total
unipolar
remitted/control
Age,
patients:
Mean (SD)
Gender,
patients:
Female/male
Diagnostic
instrument
Cut-off scores on
the HDRS-17 or
MADRS rating
scales.
General strategy for
assessment of euthymia
Beats et al. (1996) Longitudinal 19/15 73.6 (5.4) 10/9 DSM-III-R MADRS 9Dened by cut-off score Cambridge Automated Neuropsychological Test Battery
Biringer et al. (2007) Longitudinal 17/50 N.a. N.a. DSM-IV HDRS 7Dened by cut-off score Stroop Colour, Stroop Word, Stroop Colour Word,
Wisconsin Card Sorting Test perseverative responses,
and California Computerized Assessment Package
simple reaction time
Clark et al.
(2005a, 2005b)
Crosssectional Uni-
and bipolar
patients and 1st
degree relatives
15/46 45.2 (10.9) 11/4 DSM-IV HDRS 8Dened by cut-off score Rapid visual information processing task, California
Verbal Learning Test and the intradimensional/
extradimensional shift task of the Cambridge Cognitive
Examination test battery.
Kessing (1998) Crosssectional Uni-
and bipolar
patients
118/58 69.3 (12.7) 82/36 ICD-10 HDRS 8Dened by cut-off score Cambridge Cognitive Examination, Mattis Dementia
Rating Scale (MATTIS), Gottfries-Bråne-Steen, Dementia
Rating Scale, Mini-Mental State Examination, Global
Deterioration Scale
Nakano et al. (2008) Crosssectional age
a
79/85 45.1 (8.3) 41/38 DSM-V HDRS 6Dened by cut-off score Wisconsin Card Sorting Test, Stroop and Verbal Fluency
Test
Neu et al. (2005) Longitudinal 27/30 53.43
(10.84)
19/8 DSM-IV HDRS 6Dened as HDRS 6
continuously for at least
6 months by monthly evaluation
Rey auditory learning test, Reitan trail-making test (part
A), Verbal Fluency Test (animal naming), Wechsler
memory subscale (visual memory)
Paelecke-Habermann
et al. (2005)
Crosssectional No.
of episodes
a
40/20 44.40
(10.4)
b
48.15
(8.81)
N.a. DSM-IV MADRS 12 Dened by cut-off score and a
period of N 3 months where
criteria of a depressive episode
were not fullled
Visual orienting task (computerized test), Sustained
attention task (card sorting task), Executive attention,
Behavioural Assessment of the Dysexecutive Syndrome
test.
Preiss et al. (2009) Crosssectional 97/97 46.3 (12.1) 51/46 ICD-10 MADRS 11 Dened by cut-off score
andN 2 months where the
patients functioned well
Auditory Verbal Learning test Trail Making Test part A
Trail Making Test part B
Smith et al. (2006) Crosssectional Uni-
and bipolar
patients
42/33 21.3 (1.87) 29/13 DSM-IV HDRS 8Dened by cut-off score and
deemed clinically remitted for
N 1 month
California Verbal Learning Test Brixton spatial
anticipation Test Trail Making Test part A Trail Making
Test part B Stroop Colour Word
Weiland-Fiedler et al.
(2004)
Crosssectional 28/23 37 (12.2) 18/10 DSM-IV MADRS 5Dened by cut-off score and a
period of N 3 months where
patients were off antidepressant
medication
Cambridge Automated Neuropsychological Test Battery
Yuan et al. (2008) Crosssectional 18/14 67.4 (7.3) 10/9 DSM-IV HDRS 7Dened by cut-off score and
deemed clinically remitted for
N 6 months
Rey Auditory Verbal Learning Test (RAVLT), Trailmaking
A and B, Clock drawing test and Digit Span test
Only data on remitted unipolar patients included. N.a. = data not available/unclear/not assessed.
a
Sample divided into groups by the specied indication.
b
Mean and S.D. according to subgroup ( 3 episodes/N3 episodes).
23B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 4
Table 2
Results of neuropsychological tests in patients compared to healthy control individuals and possible predictors of outcome of the tests.
Study Possible predictors of cognitive function
a
Results of cognitive tests
Subsyndromal
depressive symptoms
Age of
onset
Severity of
previous episodes
Medication status Number of episodes
Beats et al. (1996) Signicant
correlation with
HDRS-score and
thinking times
Early and
late-onset/
n.a
N.a./n.a. Medicated and un-medicated/n.a. N.a./signicant correlation with
response latency
No signicant difference between patients
and healthy control individuals in any tests.
Biringer et al. (2007) N.a. N.a./n.a Moderate
(minimum)/n.a
Medicated and un-medicated/n.a. Recurrent episodes/n.a. No signicant difference between patients
and healthy control individuals in any tests.
Clark et al.
(2005a, 2005b)
No signicant
correlation with
HDRS-scores on any
tests used
N.a./n.a N.a./n.a. Medicated and un-medicated/n.a. Recurrent episodes/n.a. Paper b): Patients made signicantly more
errors on the extradimensional shift task
of the CAMCOG compared to healthy control
individuals.
Paper a):No signicant difference between
patients and healthy control individuals in
any tests.
Kessing (1998) Signicant
correlation between
BDI-scores and on
some of the tests used
N.a./n.a N.a./n.a Medicated and un-medicated/no
signicant correlation with current
psychoactive treatment. Signicant
correlations with lifetime treatment with
Lithium
Single and recurrent episodes/signicant
correlation with performance in all tests
Patients signicantly impaired on all tests.
b
Nakano et al. (2008) N.a N.a./n.a. N.a/n.a All medicated/n.a. Single and recurrent episodes/n.a. Patients signicantly impaired compared to
healthy control individuals on the Stroop test.
Neu et al. (2005) N.a N.a./no
signicant
correlation
Melancholic
subtype/n.a
Medicated and un-medicated/no
signicant correlation with current or
lifetime-treatment
No signicant correlation Patients signicantly impaired compared to
healthy control individuals on the Verbal
uency- and the Rey auditory learning test.
Paelecke-Habermann
et al. (2005)
N.a. N.a./n.a. N.a./n.a Medicated and un- medicated/n.a. Single and recurrent episodes/signicant
correlation
Patients signicantly impaired compared to
healthy control individuals.
Preiss et al. (2009) Signicant
correlation between
MADRS-scores and
some of the tests used
Early
onset/n.a.
N.a./n.a Medicated and un-medicated/no
correlation with current psychoactive
treatment
N.a/no signi cant correlation Patients signicantly impaired compared to
healthy control individuals on all measures
on the Auditory
Verbal Learning test and the Trail Making
Test part A.
Smith et al. (2006) N.a. Early
onset/n.a.
N.a./n.a Un-medicated N.a./n.a. Patients signicantly impaired compared to
healthy control individuals on trials 4 and
5 of the California Verbal Learning Test
and on the Trailmaking part A and B.
Weiland-Fiedler et al.
(2004)
Signicant
correlation between
HDRS-scores and
some of the tests used
N.a./n.a. N.a./n.a Un-medicated Recurrent episodes/n.a. Patients signicantly impaired compared to
healthy control individuals on two of the
subtests of the Cambridge Automated
Neuropsychological Test Battery.
Yuan et al. (2008) N.a. Late-onset N.a. Un-medicated Single episode/n.a. Patients signicantly impaired compared to
healthy control individuals on the delayed
recall of RAVLT and the Trailmaking A and B.
N.a. = data not available/unclear/not assessed.
a
Characteristics of included patients reported/statistical analysis performed to asses a possible association with outcome.
b
Separate analysis of patients with unipolar disorder compared to healthy control individuals performed but not presented in the paper.
24 B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 5
global cognitive function, or on specic tests assessing
cognitive function in the domains of attention, memory or
executive function in patients compared to the healthy
control individuals (Kessing, 1998; Clark et al., 2005b; Clark
et al., 2005a; Paelecke-Habermann et al., 2005; Neu et al.,
2005; Smith et al., 2006; Nakano et al., 2008; Yuan et al.,
2008; Preiss et al., 2009). In two of the included studies no
signicant difference was found in any of the employed tests
(Beats et al., 1996; Biringer et al., 2007).
The associations between cognitive function and subsyn-
dromal depressive symptoms were investigated in three of
the studies. In two of the studies (Weiland-Fiedler et al.,
2004; Preiss et al., 2009) signicant correlations were found
between MADRS-scores and performance in some of the
tests, and in the study by Weiland-Fiedler et al. (2004) some
of the results became non-signicant when the statistical
analysis was adjusted for subsyndromal depressive symp-
toms. In one study no signicant correlation was found in any
of the tests used (Clark et al., 2005a, 2005b). Furthermore, in
two of the studies (Kessing, 1998; Preiss et al., 2009)
signicant correlations were found between Beck Depression
Inventory (BDI) scores and decreased performance in some of
the tests used.
Estimates of the time period since the participants experi-
enced their last depressive episode were presented in two of
the cross-sectional studies (Kessing, 1998; Weiland-Fiedler
et al., 2004), and the statistical analyses of performance on
neuropsychological tests were not adjusted for the time period
elapsed since the patients experienced their last depressive
episode in any of the studies. None of the studies aimed to
investigate the possible effect of co-morbid psychiatric
disorders on cognitive performance. In one study only,
associations between age of onset and test performance was
investigated and no signicant correlation was found (Neu
et al., 2005).
None of the studies aimed to investigate the possible
associations between cognitive function and the clinical
manifestation of the previous episodes of depression. In some
of the studies a strategy was employed to select patients
according to the clinical subtype of previous episodes; thus in
two of the studies only patients with a history of melancholic
episodes (Neu et al., 2005), moderate severity (according to
HDRS-score at initialassessment) were included (Biringer et al.,
2007), and in one study, patients with a history of psychotic
episodes were excluded (Nakano et al., 2008).
In three of the studies, associations between current
medication status and test performance was investigated but
no signicant associations were found (Kessing, 1998; Neu
et al., 2005; Preiss et al., 2009). In three of the studies a strategy
was employed only to include currently un-medicated patients
(Weiland-Fiedler et al., 2004; Smith et al., 2006; Yuan et al.,
2008) and in one study only medicated patients were included
(Nakano et al., 2008), whereas in the remaining studies a mixed
sample of medicated and un-medicated participants were
included. In two of the studies, associations between lifetime
treatment with psychotropic medication and test performance
were investigated. In one of the studies (Kessing, 1998)the
number of Dened Daily Doses (DDD) of prior lithium
treatment was found to be associated with decreased perfor-
mance on the CAMCOG (Cambridge Cognitive Examination)
test battery, whereas in the other study (Neu et al., 2005)no
association between lifetime treatment with psychotropic
medication and performance in any tests was found.
In ve of the studies assoc iations between cognitive
function and the number of depressive episodes were inves-
tigated (Beats et al., 1996; Kessing et al., 1998; Neu et al., 2005;
Paelecke-Habermann et al., 2005; Preiss et al., 2009). In one
study, an association between the total number of episodes and
severity of cognitive impairment as measured by CAMCOG-
score (Kessing, 1998) was found, and in one study it was found
that patients with N 2 episodes were more impaired than
patients with 12 episodes on test which can be attributed to
the domain of executive function (Paelecke-Habermann et al.,
2005). In three studies, no association was found (Beats et al.,
1996; Neu et al., 2005; Preiss et al., 2009).
The possible association between cognitive function and
the cumulated duration of depressive episodes was not
investigated in any of the studies.
Very few studies accounted for the selection of participants
and the participation rate, and in only one study comparisons
between participants and non-participants were performed
(Kessing, 1998).
In one of the studies an attempt was made to blind the
investigator to the history of affective illness of the participants
when cognitive testing was performed (Kessing, 1998). In the
remaining studies it was not reported whether the investiga-
tors who performed the neuropsychological tests were blinded
towards whether the participant was a patient remitted from
depression or a healthy control individual and on the history of
affective illness of the patients. In none of the studies statistics
on inter-rater agreement were reported.
Matching of participants was done on various variables
such as age, gender, measures of pre-morbid IQ and
education. In one of the studies no exclusion criteria were
used (Smith et al., 2006).
4. Discussion
4.1. Main ndings
A total of 11 studies were included in the present review,
including a total of 500 patients remitted from unipolar
depression and 471 healthy control individuals. In nine of the
eleven studies it was found that performance on neuropsycho-
logical tests in domains of sustained and selective attention,
memory and executive function or in tests providing an
estimate on global cognitive function was decreased in patients
compared to the healthy control individuals in at least one of
the tests used. Two studies found no signicant difference
(Beats et al., 1996; Biringer et al., 2007) in any of the applied
tests, but these results may be due to low sample-sizes (N =17
and 19, respectively) and thus low statistical power to detect
discrete differences.
Multiple signicance testing increases the risk of a type I
error i.e. to erroneously conclude the presence of a signicant
association when none exist. Less than half of the studies took
this into account and performed modied t-tests to compare
the groups of interest using e.g. the Bonferroni method to
adjust the p-values. (Beats et al., 1996; Weiland-Fiedler et al.,
2004; Nakano et al., 2008; Preiss et al., 2009). Due to the
diversity in the neuropsychological tests used and in the
25B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 6
clinical spectrum of the included participants, we did not nd
it meaningful to do a meta-analysis.
4.2. Rationale for the selection criteria used in this review
4.2.1. Criteria for the assessment of state of remission
In 1988 the MacArthur Foundation Research Network on
the Psychobiology of Depression task force described how
inconsistencies in descriptions of change points in the course of
illness in affective disorder might impede research in this eld
(Frank et al., 1991). Furthermore, it was emphasised that
employment of internally consistent, empirically dened
conceptual schemes for the terms remission, recovery, relapse
and recurrence are crucial in dening change points in the
course of illness. Tentative operational criteria for each of these
terms were initially put forward by Frank et al. (1991).
However the employment of stringent operational criteria
seems not to be fully adopted by researchers in this eld yet.
According to Frank et al. (1991), cut-off scores of 7on
the17-item HDRS or 10 on the MADRS should be used to
dene remission. Some ndings support the use of even lower
cut-off scores on these scales (Zimmerman et al., 2005). The
HDRS rating scale (HAMILTON, 1960) remains the most widely
used psychometric instrument to assess the s everity of
depressive symptoms. A score of 10 on the MADRS has
been shown to be equivalent to the HRSD cut-off score of 7
(Zimmerman et al., 2004).
In the present review our main aim was to investigate
whether cognitive function is impaired during the remitted
state of unipolar affective disorder. To serve this purpose we
rst and foremost employ the strategy only to include studies
using a cut-off score 8 on the17-item HDRS, or a score of 12
on the MADRS. Using a cut-off score 7 on the HDRS or 10 on
the MADRS, might have been a more rigorous approach to
explore the aim of our investigation, but such an approach
would have resulted in very few studies (N=6 (Beats et al.,
1996; Neu et al., 2005; Biringer et al., 2007; Nakano et al., 2008;
Weiland-Fiedler et al., 2004; Yuan et al., 2008)). We did not
include studies using other rating scales (one study excluded),
since results are not comparable to the main body of research.
Secondly, in accordance with the recommendations by
Frank et al., 1991 we would ideally have aimed only to include
studies which employed a duration criterion according to the
recommended N 8 weeks in which no depressive episodes did
occur in order to dene full remission, but we felt that this
approach would be too restrictive. Studies solely relyingon self-
reported rating scale scores, such as the BDI, to assess if the
participant was in a remitted state were not included, since a
diagnoses of depression can only be established by clinical
evaluation.
4.2.2. Neuropsychological tests
Studies on the neuropsychology of affective disorder have
traditionally focused on the major cognitive domains of
attention, executive function and memory. Neuropsychological
impairment may also be present in other modalities of
cognition such as sensory perception, processing of emotionally
laden information and expression of thought disorders. A
new eld of research in these areas has recently developed with
afocusonhot cognition (Goodwin and Jamison, 2007), in
which the possible signicance of affective valence on
information processing is being studied. However, since the
processes involved in the regulation of mood could be entirely
independent of the processing of emotional valent information
(Goodwin and Jamison, 2007), these concepts need further
clarication. Thus, since the body of research in these areas are
not sufciently developed yet, we chose not to include studies
that relied on cognitive tests, which have not yet been
validated.
4.3. The association between cognitive outcome and the
number, duration and subtype/severity of previous episodes
4.3.1. Clinical manifestation, subtype/severity
The clinical spectrum of depressive disorder is hetero-
geneous in nature. Accordingly, attempts to characterise
specic cognitive prol es in accordance with clinical
manifestation have so fa r been inconclusi ve. However, the
severity of depression and in part icular the melancholic
subtype appear to contribu te to the differen ces in cognitive
decit s expressed in depr ess ion (Austin et al., 2001).
Furthermore, it has been suggested that n europsychological
decits d uring psych otic depressive ep isodes migh t be
similar to, but less severe, than those observed in patients
with sch izophrenia (Hill et al., 2004). However, due to the
fact that studies so far have utilized different methods of
ass essment, the nature and severity of cognitive impair-
ment in affective psychosis is yet to be established (Fleming
et al., 2004). These associations were not directly investi-
gated in any of the included studies. However, in some of
the studies this was taken into account by limiting selection
into the study accord ing to the subtype/se verity of previous
episodes or by excl uding participan ts with a histo ry of
episodes with psychotic features, but i t is not poss ible to
make comparisons across studies.
4.3.2. Episodes, number and duration
It has been suggested that repeated episodes of depression
may induce a decline in global cognitive function (Kessing,
1998). In the study by Kessing (1998),thenumberof
depressive episodes (mean 5.2, SD 3.7) was found to be
associated with global cognitive decline, whereas the number
of manic or mixed episodes did not seem to contribute to
cognitive decline. Patients with one episode did not differ in
cognitive function from healthy control individuals in
any measure. However, the results were not presented
separately for the groups of patients according to diagnostic
entity in relation to the control individuals. In the study by
Paelecke-Habermann et al. (2005) patients with N 3episodes
(mean 5.25 SD 2.53) were more impaired than patients with
12 episodes (mean 1.55 SD 0.51). Preiss et al. (2009) and
Biringer et al. (2007) did not replicate these ndings. However,
in the study by Preiss et al. (2009) the number of episodes the
participants had suffered were on average smaller (mean2.7 SD
1.6) and in the study by
Biringer et al. (2007) the analysis was
not sufciently powered to establish whether cognitive
function returned to normal at follow-up. Furthermore, as
noted in the paper, practise effects, which tend to show a
marked improvement on retesting, and lack of retesting of the
healthy control individuals, might have led to an overestima-
tion of the improvement of the performance of the patients. In
the study by Beats et al. (1996) reduced response latencies
26 B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 7
correlated with the number of depressive episodes. However it
is unclear whether this effect remained signicant after
correction for HDRS-scores.
Observations from brain imaging studies have suggested
that the duration of depressive episodes is associated with
atrophy in the CNS (Sheline et al., 1999). As stated by Biringer
et al. (2007), associations between longer duration of illness
and lower cogn itive function have not previously been
detected. In this study an association between the duration
of depression and decline in cognitive function was not found
and it was concluded that these ndings speak against the
presence of a scar effect in recurrent unipolar disorder.
However, this conclusion is somehow unclear since it seems
that it was the effect of the duration of illness and not the
effect of depressive episodes on cognition which was
investigated.
4.4. The associations between cognitive outcome and other
clinical factors
4.4.1. Age at onset
Patients with early and late-onset depression may show
important differences in neuropsychological test perfor-
mance, in that pat ients wi th earl y onset may suff er
predominantly from impaired episodic memory, and those
with late-onset mainly from reductions of executive function
and processing speed (Herrmann et al., 2007). Late-
onset depression may be associated with an increased
presence of vascular disease and more severe cognitive
impairments (Salloway et al., 1996). Consequently it has
been proposed that late-onset depression should be consid-
ered a distinct disorder in terms of aetiology and symptom-
atology (Alexopoulos et al., 1997). Furthermore, it has been
reported that patients with late-onset depression are less
likely to have a positive family history for affective disorders
compared to patients with early onset (HOPKINSON, 1964).
However this is a subject of ongoing discussion. The possible
association between age at onset and differences in test
performance was only investigated in one of the included
studies. In the study by Neu et al. (2005) correlations
between age at onset and test performance were investigated
but no signicant effect was found. However a signicant
correlation with the presence of signal hyperintensities on
MRI magnetic resonance imaging (MRI) presumably of
ischemic origin and cognitive performance was also
observed. In conclusion, age of onset may be associated
with different risk factors including genetic predispostition to
developing recurrent depression, somatic disease and vascu-
lar pathology. These distinctions appear to have importance
in relation to the neuropsychological and neurobiological
abnormalities observed in unipolar disorder.
4.4.2. The impact of subclinical psychopathology on cognitive
test performance
Performance on neuropsychological test are not readily
dissociable from psychopathology (Austin et al., 2001), and
some tests seem to be highly dependent on current mood
severity, while others are not (Austin et al., 2001). Some of
the studies intended to investigate the impact of subsyn-
dromal depressive symptoms on cognitive outcome. In the
study by Biringer et al. (2007) the patients' cognitive
performance was assessed longitudinally during episodes of
depression and in remission. In this study, remission of
depression was followed by improvement in verbal memory
function up to the level of healthy controls individuals. No
associations between improvement in depression and im-
provement in other dimensions of cognitive function were
found. However the sample-size of the remitted patients was
small (N=17). Clark et al. (2005a, 2005b) did not
nd any
signicant correlations on any test measures, but this could
also be attributed to the inability to detect small differences
due to the low sample size. In the much larger study by Preiss
et al. (2009) a signicant correlation between MADRS-score
and performance in some of the tests used was found and in
the study by Weiland-Fiedler et al. (2004), some of the results
became non-signicant when the statistical analysis was
adjusted for subsyndromal depressive symptoms (MADRS-
score). Additionally in the study by Kessing (1998) and by
Preiss et al. (2009) associations were found between BDI-
scores and performance on some of the tests.
It is a well known clinical experience that a subgroup of
patients with depressive disorder does not exhibit cognitive
decits neither in the acute state nor after remission or
recovery, but it is unclear how large this fraction of patients is.
It is unclear how this is so, but for some patients depressive
psychopathology does not seem to play an important role. It is
an important focus area for future research to investigate why
some patients with depressive disorder present with cogni-
tive dysfunction and others not.
4.4.3. The time period elapsed since the patients experienced the
last episode of depression
Cognitive decits might persist longer than the depressive
episode and this seems to be true for all depressive sub-
diagnoses (Neu et al., 2001). This suggest that if studies wish
to assess whether depressive episodes are associated with
cognitive decits after recovery, the time period that elapsed
since patients experienced the last episode of depression
should be taken into consideration. However this was not
systematically investigated in any of the studies.
4.4.4. Medication and other treatment interventions
The impact of medication on cognitive function has not
been clearly established. A common perception is that
tricyclic antidepressants and antidepressants with anticho-
linergic effects and mainly lithium should have a deteriorat-
ing effect on cognition (Lund et al., 1982; Shaw et al., 1987).
Conversely, some recent ndings suggest that antidepressant
agents (Duman, 2009; Kessing et al., 2009) and Lithium
(Kessing et al., 2008; Kessing et al., 2010) could have a
neuroprotective effect. Some evidence from discontinuation
and reinstatement studies suggests that a possible negative
effect of lithium on cognition is expressed immediately and
remits partially with discontinuation (Goodwin and Jamison,
2007). However given the nature that the choice of treatment
regimes and dosage levels is inherently connected to the
severity of the disorder and the course of illness of the patient,
this association is not easily assessed. In the studies by Neu
et al. (2005), Kessing (1998) and Preiss et al. (2009) no
signicant association between current psychotropic treat-
ment and test performance was found. Moreover in the study
by Kessing (1998) and Neu et al. (2005) no association
27B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 8
between lifetime treatment with psychotropic medication
and test performance was found. In the study by Kessing
(1998) lifetime exposure to Lithium was associated with a
non-signicant decrease in performance on two out of ve
tests of cognitive function. However, this effect did not
remain signicant when the analysis was adjusted for the
effect of factors associated with the course of illness. In
conclusion, the impact of medication on cognitive function
has not been clearly established. Observations in the included
studies do not suggest that psychotropic medication should
have a negative impact on cognitive function. However none
of the studies were designed to assess this association. These
ndings are not suggestive of causality and should be
interpreted with the greatest caution. Future studies should
take into consideration the association between medication
and cognitive function. Furthermore, the number of observa-
tions and the possibility of multiplicative interactions should
be considered when assessing these associations statistically,
as insufcient sample size could inate the p-values and
consequently increase the risk of type I e rror. These
considerations also apply to treatment interventions such as
Electro Convulsive Therapy.
4.4.5. Age and gender
Presently there are no indications of a gender effect on
cognition in depression and this was not investigated in any
of the included studies.
It is generally accepted that elderly patients may show
more severe cognitive decits compared to young patients
during episodes of depression. As mentioned above, a
possible role of vascular lesions has been established, but
this seems to apply in particular to patients with late-onset
depressive disorder. In the study by Nakano et al. (2008), the
presence of vascular lesions was shown to affect performance
in tests of executive function independent of age. In the study
done by Beats et al. (1996), adjusting the statistical analysis
for age increased the signicance level on the Tower of
London test independent of the presence of vascular
pathology. However since these changes seem not to be
unique to depression in the elderly conclusions on causality
or on to what extent specic residual cognitive decits are
unique to elderly u nipolar patients cannot be directly
inferred. Furthermore, in a large Danish prospective cohort
study it was shown that the association between white-
matter hyperintensities (presumed to be of ischemic origin)
and cognitive impairment was signicant only for cognitive
decline in the decade 7080 years (Garde et al., 2000).
In conclusion, ischemic lesions may play a role in both
normal and pathological ageing and it is only one among
many factors contributing to cognitive decline in ageing.
Moreover, these changes seem not always to be associated
with cognitive dysfunction and the association to unipolar
disorder remains unclear.
4.4.6. Number of previous hospitalisations
Patients with a history of hospitalisation may perform
worse on some measures of neuropsychological performance
than those without (Purcell et al., 1997). In the study by Preiss
et al. (2009), patients with the most ho spitalisation s
performed worse in test of attention and executive function
compared to healthy control individuals, but not in verbal
memory tests. However, it is unclear whether cognitive
impairment is a factor leading to hospitalisation, or whether
factors related to hospitalisation exacerbate cognitive decits.
Additionally, hospitalisation reects the complete breakdown
in psychosocial function of the patient, and could thus be a
subsidiary indicator of the severity of episodes.
4.4.7. Co-morbid psychiatric disorders
The presence of co-morbid disorders has been shown to
be a predictor of poor cognitive performance in depression
(Baune et al., 2009), and could possibly to some extent
explain the diversity in ndings. Especially anxiety disorders
may in some aspect share common traits with depression.
However, in none of the included studies the authors aimed
to assess the possible impact of co-morbid psychiatric
disorders on cognitive function.
4.5. General conclusion on the specicity and nature of the
neuropsychological ndings
Presently there is no general con sensus on to which
extent patients with affective disorders manifest gl obal
impairment of brain functions (Veiel, 1997; Kessing, 1998)
or impairment within specic cognitive domains (Austin
et al., 2001). In the studies included in the pr esent review,
cognitive impairment was found within various cognitive
domains. However , it is li kely that the diversity of
methodological strategies u sed in the included studies,
heterogeneity in the clinical characteristics of the included
samples and the possible s ources of bias in the studies, are
contributing to t he observed ndings. Furthermore, cogni-
tive processes are mutually i nterdependent, and abnormal-
ities of the greatest magnitude reected in a test score do
not necessarily reect main core decits. Neuropsycholog-
ical tests diverge on a number of cognit ive domains thereby
making it difcult to assess any primary functional
impairment with a specic test. In addition to this, the
lack of adjus tment for confounding variables may inuence
estimates, e.g. in the study by Nakano et al. (2008),ageand
education, and in the study by Preiss et al., 2009,subclinical
psychopathology, were shown to have a differential effect on
different measures of cognition, and consequently effect sizes
varied across the individual neuropsychological tests. Thus,
since a large number of different neuropsychological tests and
statistical analyses were used across studies with lack of
adjustment for a number of possible confounders, it is difcult
to validly estimate effect sizes. In general cognitive decline was
found within a number of neuropsychological tests that may be
related to sustained and selective attention, memory and
executive functions, and it is not possible to assess whether
impairments in some cognitive domains are more prevalent
than in others.
In conclusion, from the studies included in our review, it is
difcult to assess whether the neuropsychological decits
observed in the patients are signs of multiple independent
impairments or of one or more domain specic cognitive
impairments. We highlight the need for future research to
take into account covariates which may act as confounders in
the context of the selected patient sample and hypothesis,
and that the selection of neuropsychological tests should be
hypothesis-driven.
28 B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 9
5. Estimation of risk of bias in the included studies
A major source of bias in studies using multiple tests is the
possibility of type I error introduced by multiple signicance
testing and only half of the studies included in this review
corrected for multiple testing when relevant.
Very few studies accounted for the selection of partici-
pants and controls and the participation rate. Therefore, the
results from most of the included studies could be affected by
selection bias. In one of the studies the investigator tried to do
comparisons between participa nts and non-participants.
Thus, in the study by Kessing (1998) non-participants were
rated with the Global Deteriorating Scale Score (GDS) and the
scores were found to be lower in participants compared to
non-participants. However, this possible trend toward selec-
tion of patients with poorer cognitive performance could be
attributed to the different techniques applied in face-to-face
interview vs. telephone interview the latter enforcing more
conservative estimates in favour of higher GDS scores.
In one study (Kessing, 1998), an attempt was made to
blind the investigator to the history of affective illness of the
participants. In the remaining studies it was not reported
whether the investigators who performed the neuropsycho-
logical tests were blinded toward whether the participant
was a person remitted from depression or a healthy control
individuals, and the history of affective illness in the
participants. Moreover, none of the studies reported statistics
on inter-rater agreement when relevant. Lack of blinding
would give a trend toward a bias in the direction of
overestimating the difference in test performance between
the patient and control sample and in the assessment of
variables related to the course of illness such as number of
episodes. Matching of participants was done on various
variables su ch as a ge, gender, measures of p re-morbid IQ/
pre -morbid capacity and education. However, only a few
studies seemed to have adjusted their analysis for these
importa nt factors.
6. Conclusions and recommendations for future research
In conclusion, due to limitations in the design of the
included studies, the plausibility of the hypotheses explored
in this review, are not easily assessed. One of the major
drawbacks in the eld of research of cognitive function in the
remitted state of unipolar affective disorder is that very few
studies have employed strict denitions to establish whether
patients were fully remitted when neuropsychological testing
was performed (see Section 4.2.1). Furthermore, due to the
diversity of methodological strategies employed, the hetero-
geneity in the clinical characteristics of the included samples
and the possible sources of bias, assessment of the relative
contribution of the effect of depressive episodes and the
impact of factors, such as medication, on cognitive function
cannot be readily made.
In the present review cognitive impairment was found in
various cognitive domains. Howeve r, from the present
ndings it cannot directly be inferred on to what extent
these impairments are signs of multiple independent impair-
ments, or a sign of domain specic cognitive impairments.
Furthermore, based on the neuropsychological ndings from
the studies included in this review, it can be concluded, that
although a number of studies is suggestive of an effect of
episodes per se, it is still unclear whether episodes of
depression induces cognitive decline.
It is an ongoing debate whether cognitive dysfunction is a
state or a trait related phenomenon and whether repeated
episodes of depression may induce a scar. It has been
hypothesised that each depressive episode increases the risk
of cognitive impairment during remission, which further
increase the risk of future recurrence (Kessing, 2001). It is
however not possible based on the current review to qualify
this hypothesis further mainly since high quality prospective
longitudinal studies are lacking. Findings from neuroimaging
studies suggest that repeated episodes of depression and the
duration of illness may be associated with atrophy of structures
of the CNS vita l to c ognitive function such as e.g. the
hippocampus (McKinnon et al., 2009). However, it has been
found that hippocampal volume reduction may be present in
healthy individuals with a genetic predisposition to unipolar
depression (Baare et al., 2010
). Furthermore, depressive
disorder may be associated with an increased risk of developing
dementia (Kessing and Nilsson, 2003; Kessing and Andersen,
2004) and studies have suggested, that rather than a prodrome,
depression may be a risk factor for Alzheimer's disease (Ownby
et al., 2006). No prospective clinical study, however, has
investigated patients with early signs of possible dementia
such as Mild Cognitive Impairment (MCI) before the onset of
the rst depressive episode. However, a study from our group
found that discrete cognitive impairment may be present in the
pre-morbid state (Christensen et al., 2006). Thus, it is not
known whether cognitive impairment in unipolar disorder is
neurodegenerative or neurodevelopmental in nature and the
association with course of illness is unclear. In relation to this,
disentangling persistent cognitive dysfunction in the context of
depression and co-existence of MCI is an important focus for
future research.
In theory, biological and cognitive trait markers of
unipolar disorder could produce a pre-episode effect
which, in the pathway through epigenetic modication and
environmental interactions, could act by increasing the rate of
episodes and thus the sensitivity to cognitive decline. In this
theoretical model the end-state, wherein cognitive decline
could occur, could be reached by two pathwa ys: high
frequency and duration of episodes and high genetic risk to
develop episodes. This association would optimally be
investigated in a prospective study with neuropsychological
testing of individuals in genetic risk of depression before
onset of illness, during consecutive periods of remission and
following successive depressive episodes. Such a study is time
and research consuming and hard to conduct with a sufcient
participation rate over many years of follow-up.
Combining neuropsychological evidence with evidence
from brain imaging and biomarker studies, could contribute
further insight into the main core of neuropsychological
decits in unipolar affective disorder. We recommend that
future studies should focus on disentangling the state and
trait chara cteristics of cognitive dysfunction in unipolar
disorder and further clarify the associations with clinical
phenotype, course of illness and subsyndromal psychopa-
thology. From a clinical perspective, there is a need to identify
the cognitive difculties in individuals suffering from uni-
polar disorder in relation to psychosocial function, quality of
29B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 10
life and risk of recurrence. Studies should seek to identify
potential prophylaxis and treatment interventions for cogni-
tive dysfunction, and it should be investigated whether
pharmacological treatment, cognitive remediation and early
and sustained intervention for depressive episodes decreases
the risk of cognitive decline.
Role of funding source
No funding was provided for this study.
Conict of interest
Lars Vedel Kessing has been a consultant for Bristol-Myers Squibb, Eli
Lilly, Lundbeck, AstraZeneca, Pzer,Wyeth, Servier and Janssen-Cilag. All
other authors declare that they have no conicts of interest.
Acknowledgement
None.
References
Alexopoulos, G.S., Meyers, B.S., Young, R.C., Campbell, S., Silbersweig, D.,
Charlson, M., 1997. Vascular depression hypothesis. Arch. Gen.
Psychiatry 54 (10), 915922.
Austin, M.P., Mitchell, P., Goodwin, G.M., 2001. Cognitive decits in
depression: possible implications for functional neuropathology. Br. J.
Psychiatry 178, 200206.
Baare, W.F., Vinberg, M., Knudsen, G.M., Paulson, O.B., Langkilde, A.R.,
Jernigan, T.L., Kessing, L.V., 2010. Hippocampal volume changes in
healthy subjects at risk of unipolar depression. J. Psychiatr. Res. 44,
655662. doi:10.1016/j.jpsychires.2009 .12.009 doi: S0022-3956(09)
00284-2 [pii].
Baune, B.T., McAfoose, J., Leach, G., Quirk, F., Mitchell, D., 2009. Impact of
psychiatric and medical comorbidity on cognitive function in depression.
Psychiatry Clin. Neurosci. 63 (3), 392400. doi:10.1111/j.1440-
1819.2009.01971.x doi: PCN1971 [pii].
Beats, B.C., Sahakian, B.J., Levy, R., 1996. Cognitive performance in tests
sensitive to frontal lobe dysfunction in the elderly depressed. Psychol.
Med. 26 (3), 591603.
Biringer, E., Mykletun, A., Sundet, K., Kroken, R., Stordal, K.I., Lund, A., 2007. A
longitudinal analysis of neurocognitive function in unipolar depression.
J. Clin. Exp. Neuropsychol. 29 (8), 879891.
Bora, E., Yucel, M., Pantelis, C., 2009. Cognitive endophenotypes of bipolar
disorder: a meta-analysis of neuropsychological decits in euthymic
patients and their rst-degree relatives. J. Affect. Disord. 113 (12), 120.
doi:10.1016/j.jad.2008.06.009 doi: S0165-0327(08)00246-2 [pii].
Burt, D.B., Zembar, M.J., Niederehe, G., 1995. Depression and memory
impairment: a meta-analysis of the ass ociation, its pattern, and
specicity. Psychol. Bull. 117 (2), 285305.
Christensen, M.V., Kyvik, K.O., Kessing, L.V., 2006. Cognitive function in unaffected
twins discordant for affective disorder. Psychol. Med. 36 (8), 11191129.
doi:10.1017/S0033291706007896 doi: S0033291706007896 [pii].
Clark, L., Kempton, M.J., Scarna, A., Grasby, P.M., Goodwin, G.M., 2005a.
Sustained attention-decit conrmed in euthymic bipolar disorder but
not in rst-degree relatives of bipolar patients or euthymic unipolar
depression. Biol. Psychiatry 57 (2), 183187.
Clark, L., Sarna, A., Goodwin, G.M., 2005b. Impairment of executive function
but not memory in rst-degree relatives of patients with bipolar I
disorder and in euthymic patients with unipolar depression. Am. J.
Psychiatry 162 (10), 19801982 -1982.
Duman, R.S., 2009. Neuronal damage and protection in the pathophysiology
and treatment of psychiatric illness: stress and depression. Dialogues
Clin. Neurosci. 11 (3), 239255.
Fleming, S.K., Blasey, C., Schatzberg, A.F., 2004. Neuropsychological correlates
of psychotic features in major depressive disorders: a review and meta-
analysis. J. Psychiatr. Res. 38 (1), 2735 doi: S0022395603001006 [pii].
Frank, E., Prien, R.F., Jarrett, R.B., Keller, M.B., Kupfer, D.J., Lavori, P.W., Rush,
A.J ., Weissman, M.M., 1991. Conceptualization and rationale for consensus
denitions of terms in major depressive disorder. Remission, recovery,
relapse, and recurrence. Arch. Gen. Psychiatry 48 (9), 851855.
Garde, E., Mortensen, E.L., Krabbe, K., Rostrup, E., Larsson, H.B., 2000. Relation
between age-related decline in intelligence and cerebral white-matter
hyperintensities in healthy octogenarians: a longitudinal study. Lancet
356 (9230), 628634. doi:10.1016/S0140-6736(00)02604-0 doi: S0140-
6736(00)02604-0 [pii].
Goodwin, F.K., Jamison, K.R., 2007. Manic-Depressive Illness. Oxford
University Press.
Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg.
Psychiatry 23, 5662.
Herrmann, L.L., Goodwin, G.M., Ebmeier, K.P., 2007. The cognitive neuropsy-
chology of depression in the elderly. Psychol. Med. 37 (12), 16931702.
doi:10.1017/S0033291707001134 doi: S0033291707001134 [pii].
Hill, S.K., Keshavan, M.S., Thase, M.E., Sweeney, J.A., 2004. Neuropsycholog-
ical dysfunction in antipsychotic-naive rst-episode unipolar psychotic
depression. Am. J. Psychiatry 161 (6), 9961003.
Hopkinson, G., 1964. A genetic study of affective illness in patients over 50.
Br. J. Psychiatry 110, 244254.
Jaeger, J., Berns, S., Uzelac, S., Davis, C.S., 2006. Neurocognitive decits and
disability in major depressive disorder. Psychiatry Res. 145 (1), 3948 -
48.
Kessing, L.V., 1998. Cognitive impairment in the euthymic phase of affective
disorder. Psychol. Med. 28 (5), 10271038.
Kessing, L.V. 2001. Course and cognitive outcome in major a ffective
disorders: Doctoral Thesis. Lægeforeningens Forlag, Copenhagen.
Kessing, L.V., Andersen, P.K., 2004. Does the risk of developing dementia
increase with the number of episodes in patients with depressive
disorder and in patients with bipolar disorder? J. Neurol. Neurosurg.
Psychiatry 75 (12), 16621666. doi:10.1136/jnnp.2003.031773 doi: 75/
12/1662 [pii].
Kessing, L.V., Nilsson, F.M., 2003. Increased risk of developing dementia
in patients with major affective disorders compared to patients with
other medical illnesses. J. Affect. Disord. 73 (3), 26126 9 doi:
S0165032702000046 [pii].
Kessing, L.V., Andersen, P.K., Mortensen, P.B., 1998. Predictors of recurrence
in affective disorder. A case register study. J. Affect. Disord. 49 (2),
101108 doi: S0165-0327(97)00163-8 [pii].
Kessing, L.V., Sondergard, L., Forman, J.L., Andersen, P.K., 2008. Lithium
treatment and risk of dementia. Arch. Gen. Psychiatry 65 (11),
13311335. doi:10.1001/archpsyc.65.11.1331 doi: 65/11/1331 [pii].
Kessing, L.V., Sondergard, L., Forman, J.L., Andersen, P.K., 2009. Antidepres-
santsanddementia.J.Affect.Disord.117(12), 2429. doi :10.1016/
j.jad.2008.11.020 doi: S0165-0327(08)00473-4 [pii].
Kessing, L.V., Forman, J.L., Andersen, P.K., 2010. Does lithium protect against
dementia? Bipolar Disord. 12 (1), 87 94. doi:10.1111/j.1 399-
5618.2009.00788.x doi: BDI788 [pii].
Lund, Y., Nissen, M., Rafaelsen, O.J., 1982. Long-term lithium treatment and
psychological functions. Acta Psychiatr. Scand. 65 (3), 233244.
McKinnon, M.C., Yucel, K., Nazarov, A., MacQueen, G.M., 2009. A meta-
analysis examining clinical predictors of hippocampal volume in patients
with major depressive disorder. J. Psychiatry Neurosci. 34 (1), 4154.
Nakano, Y., Baba, H., Maeshima, H., Kitajima, A., Sakai, Y., Baba, K., Suzuki, T.,
Mimura, M., Arai, H., 2008. Executive dysfunction in medicated, remitted
state of major depression. J. Af fec t. Disord. 111 (1), 4 651. doi :10.1016/
j.jad.2008.01.027 doi : S0165-0327(08)00066-9 [pii].
Neu, P., Kiesslinger, U., Schlattmann, P., Reischies, F.M., 2001. Time-related
cognitive deciency in four different types of depression. Psychiatry Res.
103 (23), 237247 doi: S0165-1781(01)00286-4 [pii].
Neu, P., Bajbouj, M., Schilling, A., Godemann, F., Berman, R.M., Schlattmann,
P., 2005. Cognitive function over the treatment course of depression in
middle-aged patients: correlation with brain MRI signal hyperintensi-
ties. J. Psychiatr. Res. 39 (2), 129135.
Ownby, R.L., Crocco, E., Acevedo, A., John, V., Loewenstein, D., 2006.
Depression and risk for Alzheimer disease: systematic review, meta-
analysis, and metaregression analysis. Arch. Gen. Psychiatry 63 (5),
530538. doi:10.1001/archpsyc.63.5.530 doi: 63/5/530 [pii].
Paelecke-Habermann, Y., Pohl, J., Leplow, B., 2005. Attention and executive
functions in remitted major depression patients. J. Affect. Disord. 89 (13),
125135. doi:10.1016/j.jad.2005.09.006 doi: S0165-0327(05)00290-9
[pii].
Preiss, M., Kucerova, H., Lukavsky, J., Stepankova, H., Sos, P., Kawaciukova, R.,
2009. Cognitive decits in the euthymic phase of unipolar depression.
Psychiatry Res. 169 (3), 235239. doi:10.1016/j.psychres.2008.06.042
doi: S0165-1781(08)00220-5 [pii].
Purcell, R., Maruff, P., Kyrios, M., Pantelis, C., 1997. Neuropsychological
function in young patients with unipolar major depression. Psychol.
Med. 27 (6), 12771285.
Quraishi, S., Frangou, S., 2002. Neuropsychology of bipolar disorder: a review.
J. Affect. Disord. 72 (3), 209226 -226.
Ravnkilde, B., Videbech, P., Clemmensen, K., Egander, A., Rasmussen, N.A.,
Rosenberg, R., 2002. Cognitive decits in major depression. Scand. J.
Psychol. 43 (3), 239251.
Salloway, S., Malloy, P., Kohn, R., Gillard, E., Duffy, J., Rogg, J., Tung, G.,
Richardson, E., Thomas, C., Westlake, R., 1996. MRI and neuropsycho-
logical differences in early- and late-life-onset geriatric depression.
Neurology 46 (6), 15671574.
30 B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 11
Shaw, E.D., Stokes, P.E., Mann, J.J., Manevitz, A.Z., 1987. Effects of lithium
carbonate on the memory and motor speed of bipolar outpatients.
J. Abnorm. Psychol. 96 (1), 6469.
Sheline, Y.I., Sanghavi, M., Mintun, M.A., Gado, M.H., 1999. Depression
duration but not age predicts hippocampal volume loss in medically
healthy women with recurrent major depression. J. Neurosci. 19 (12),
50345043.
Smith, D.J., Muir, W.J., Blackwood, D.H.R., 2006. Neurocognitive impairment
in euthymic young adults with bipolar spectrum disorder and recurrent
major depressive disorder. Bipolar Disord. 8 (1), 4046 -46.
Veiel, H.O., 1997. A preliminary prole of neuropsychological decits
associated with major depression. J. Clin. Exp. Neuropsychol. 19 (4),
587603.
Weiland-Fiedler, P., Erickson, K., Waldeck, T., Luckenbaugh, D.A., Pike, D.,
Bonne, O., Charney, D.S., Neumei ster, A., 2004. Evidence for continuing
neuropsychological impairments in depression. J. Affect. Disord. 82 (2),
253258. doi:10.1016/j.jad.2003.10.009 doi: S0165032703003185 [pii].
Yuan,Y.,Zhang,Z.,Bai,F.,Yu,H.,Shi,Y.,Qian,Y.,Liu,W.,You,J.,Zhang,
X., Liu, Z., 2008. Abnormal neural activity in the patients with
remitted geriatric depression: a resting-state functi onal magnetic
resonance imagingstudy. J. Affect. Disord. 111 (23), 145152. doi:10.1016/
j.jad.2008.02.016 doi: S0165-0327(08)00091-8 [pii].
Zimmerman, M., Posternak, M.A., Chelminski, I., 2004. Derivation of a
denition of remission on the Montgomery-Asberg depression rating
scale corresponding to the denition of remission on the Hamilton rating
scale for depression. J. Psychiatr. Res. 38 (6), 577582. doi:10.1016/j.
jpsychires.2004.03.007 S0022395604000433 [pii].
Zimmerman, M., Posternak, M.A., Chelminski, I., 2005. Is the cutoff to dene
remission on the Hamilton Rating Scale for Depression too high? J. Nerv.
Ment. Dis. 193 (3), 170175 doi: 00005053-200503000-00003 [pii].
31B.J. Hasselbalch et al. / Journal of Affective Disorders 134 (2011) 2031
Downloaded from ClinicalKey.com at BS - University of Copenhagen April 19, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.
Page 12
  • Source
    • "For example, a systematic review of 11 studies involving 500 adult patients in remission from MDD documented persistent deficits in sustained and selective attention, memory, and executive function. However, the heterogeneity across studies did not allow for the estimation of precise effect sizes [37]. In a subsequent study of individuals with unipolar depressive disorder in remission, Hasselbalch et al. [38] reported that the severity of cognitive impairment increases as a function of the cumulative duration of MDEs and, additionally, indicated that the occurrence of psychotic features is associated with worse cognitive performance. "
    [Show abstract] [Hide abstract] ABSTRACT: Background: Cognitive dysfunction in major depressive disorder (MDD) encompasses several domains, including but not limited to executive function, verbal memory, and attention. Furthermore, cognitive dysfunction is a frequent residual manifestation in depression and may persist during the remitted phase. Cognitive deficits may also impede functional recovery, including workforce performance, in patients with MDD. The overarching aims of this opinion article are to critically evaluate the effects of available antidepressants as well as novel therapeutic targets on neurocognitive dysfunction in MDD. Discussion: Conventional antidepressant drugs mitigate cognitive dysfunction in some people with MDD. However, a significant proportion of MDD patients continue to experience significant cognitive impairment. Two multicenter randomized controlled trials (RCTs) reported that vortioxetine, a multimodal antidepressant, has significant precognitive effects in MDD unrelated to mood improvement. Lisdexamfetamine dimesylate was shown to alleviate executive dysfunction in an RCT of adults after full or partial remission of MDD. Preliminary evidence also indicates that erythropoietin may alleviate cognitive dysfunction in MDD. Several other novel agents may be repurposed as cognitive enhancers for MDD treatment, including minocycline, insulin, antidiabetic agents, angiotensin-converting enzyme inhibitors, S-adenosyl methionine, acetyl-L-carnitine, alpha lipoic acid, omega-3 fatty acids, melatonin, modafinil, galantamine, scopolamine, N-acetylcysteine, curcumin, statins, and coenzyme Q10. The management of cognitive dysfunction remains an unmet need in the treatment of MDD. However, it is hoped that the development of novel therapeutic targets will contribute to 'cognitive remission', which may aid functional recovery in MDD.
    Full-text · Article · Dec 2016 · BMC Medicine
    • "[13] However, it is not clear whether deficits in attentional shifting in depression are simply statedependent correlates, scars of previous episodes, or vulnerability factors that may confer risk for the onset of depressive episodes. [11, 17] Although distinct constructs, attentional shifting, and rumination may influence one another. An inability to disengage attention from mood-congruent stimuli in depression may facilitate the prolonged processing of such material via rumination, thereby impairing the ability to use adaptive regulatory strategies such as cognitive reappraisal. "
    [Show abstract] [Hide abstract] ABSTRACT: Background: Major depressive disorder often is characterized by a lack of cognitive and emotional flexibility, resulting in an impaired ability to adapt to situational demands. Adolescence is an important period of risk for the first onset of depression, yet relatively little is known about whether aspects of inflexibility, such as rumination and deficits in attentional shifting, could confer risk for the development of the disorder during this time. Method: In the present study, a sample of 285 never-depressed adolescents completed self-report and behavioral measures of rumination and attentional shifting at a baseline visit, followed by up to 4 years of annual prospective follow-up diagnostic assessments. Results: Survival analyses indicated that adolescents with greater levels of rumination or poorer attentional shifting experienced a shorter time until the first onset of major depressive episodes, even after accounting for baseline symptoms and demographic characteristics. Although girls were twice as likely as boys to experience the first onset of depression, rumination predicted a shorter time until depression onset only for boys. Rumination and attentional shifting were not correlated and predicted time until onset of major depression independently of one another. Conclusions: These results provide evidence that components of cognition that are characterized by rigidity and perseveration confer risk for the first onset of major depression during adolescence. Evaluating rumination and attentional shifting in adolescence may be useful in identifying individuals who are at risk for depression and who may benefit from interventions that target or alter the development of these characteristics.
    No preview · Article · Apr 2016 · Depression and Anxiety
  • Source
    • "Others have noted significantly improved verbal memory abilities following a successful course of antidepressant treatment in depressed adults (Vythilingam et al., 2004; Wroolie et al., 2006). Although these findings provide support for the notion that verbal memory improves when depressive symptoms diminish or remit, several studies have found remitted patients to display impaired auditory verbal episodic memory (Hasselbalch, Knorr, & Kessing, 2011; Preiss et al., 2009; Smith, Muir, & Blackwood, 2006; Weiland Fiedler et al., 2004; Yuan et al., 2008). The findings on verbal memory of remitted depressed patients have been varied and the factors contributing to this variability is unclear. "
    [Show abstract] [Hide abstract] ABSTRACT: The substantial functional and financial costs of Major Depression, particularly Refractory Depression, have led to the systematic investigation of effective treatment. As such, understanding the neuropsychological profile of the syndrome has been an important line of scientific inquiry, although the negative effect Major Depression has on neuropsychological function is not clearly understood. To date, there are no studies that have examined the neuropsychological profiles of Refractory Depression when defined as two failed antidepressant medication trials, from two different classes of adequate dose and duration. This study examined the neuropsychological profile differences between Refractory (RD, n=26) and Non-refractory Depression (NRD, n=46). It was hypothesized that (1) RDs would perform worse than NRDs on a battery of cognitive tests measuring abilities in attention, executive function, and verbal memory; and (2) depression chronicity (the total length of time since the onset of initial depressive episode) and severity would significantly contribute to the variance of cognitive function. Lastly, an exploratory hypothesis aimed to examine the effects three medication classes (SSRI, SNRI, and Wellbutrin) had on cognitive function. Results revealed that (1) the RDs and NRDs of this sample do not differ in cognitive function; (2) depression chronicity and severity did not significantly contribute additional variance to cognitive performance; and (3) that medication classes (SSRI, SNRI, and Wellbutrin) do not significantly differ from one another in their effect across the domains of cognition tested. Considering the strong need for a better understanding of cognitive abilities in RD, this study represents a respectable initial effort in this area. Although no significant differences were found between groups, RDs generally performed better than NRDs across cognitive domains tested. Because of the similarities on certain aspects between the sample groups (depression chronicity, severity, length of current depressive episode), future investigation is warranted to draw meaning from these results. The several limitations of this study include similarities and relatively small size across the two sample groups, likely medication confounds, unknown reliability of chronicity variable (e.g., length of time in months since the initial depressive episode), and limited generalizability (participants solely consisted of outpatients). Future studies would benefit from an agreed upon classification system and a unified approach to investigating aspects of Major Depression. Classification and the unified methodological approach should be widely agreed upon by a large group of international experts and should be used to ensure consistency of in future studies. Results of the present study provide a framework that may assist clinicians better understand the cognitive capabilities of Refractory Depression and help clinicians tailor more individualized treatment.
    Full-text · Thesis · Nov 2015
Show more