Bipolar Disorder Programme, Clinical Institute of Neuroscience, University Hospital Clinic, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona
Bipolar Disorder Programme, Clinical Institute of Neuroscience, University Hospital Clinic, IDIBAPS, Barcelona and Psychiatry Department, Universidad Autonoma de Madrid
Bipolar Disorder Programme, Clinical Institute of Neuroscience, University Hospital Clinic, IDIBAPS, Barcelona, Spain
Correspondence: Dr Eduard Vieta, Clinical Institute of Neuroscience, University Hospital Clinic of Barcelona, Villarroel 170, 08036 Barcelona, Spain. Tel: +34 93 2275401; fax: +34932275477; email: evieta{at}clinic.ub.es
Funding detailed in Acknowledgements.
|
|
|---|
Aims To compare the cognitive performance of patients with bipolar II disorder with that of patients with bipolar I disorder and a healthy control group.
Method The study included 71 euthymic patients with bipolar disorder (38 bipolar I, 33 bipolar II), who were compared on clinical and neuropsychological variables (e.g. executive function, attention, verbal and visual memory) and contrasted with 35 healthy controls on cognitive performance.
Results Compared with controls, both bipolar groups showed significant deficits in most cognitive tasks including working memory (DigitSpan Backwards, P=0.002) and attention (DigitSpan Forwards, P=0.005; Trail Making Test, P=0.001). Those with type II disorders had an intermediate level of performance between the bipolar I group and the control group in verbal memory (P<0.005) and executive functions (Stroop interference task, P=0.020).
Conclusions Cognitive impairment exists in both subtypes of bipolar disorder, although more so in the bipolar I group. The best predictors of poor psychosocial functioning in bipolar II disorder were subclinical depressive symptoms, early onset of illness and poor performance on a measure related to executive function.
|
|
|---|
The main aim of our study was to identify the cognitive performance in patients with bipolar II disorder in comparison with those with bipolar I disorder and a healthy control group. We predicted that the bipolar II group would exhibit an intermediate profile between the bipolar I group and the healthy controls with an emphasis on domains of verbal memory, attention and executive functions, which are the most common cognitive deficits in bipolar illness in general. A further hypothesis was that neuropsychological performance would also influence psychosocial functioning in patients with bipolar II disorder. As far as we know, this is the first study to evaluate specifically cognitive dysfunctions in bipolar II disorder, employing a rigorous definition of euthymia, with a design involving two control groups: one comprising patients with bipolar I disorder and the other healthy participants.
|
|
|---|
Clinical variables were collected as part of the Bipolar Disorders Programme protocol of the University Hospital Clinic of Barcelona. The clinical variables included in this study were number and type of episodes, duration of illness (chronicity); age at onset of illness; number of hospitalisations; suicide attempts; family history of affective disorders; history of psychotic symptoms; and diagnostic type I or II.
Psychosocial functioning was assessed using the Global Assessment of Functioning scale (GAF; American Psychiatric Association, 1994) as a measure of functional outcome. The original GAF instructions call for rating symptoms or functioning. As many other measures of mood symptoms were obtained as part of the evaluation, the rater was instructed to use the GAF to measure psychosocial functioning in the month prior to rating. Occupational adaptation, as an additional measure of functional outcome, was established as `good' when patients were working at a good or acceptable level of functioning or `poor' if they did not work at all or had poor occupational functioning during the 3 years prior to the evaluation. This information was provided by the patient and confirmed by a first-degree relative or a partner. The clinical interview, including psychosocial functioning, was conducted by a trained psychiatrist, and the neuropsychological evaluation was carried out by a trained neuropsychologist, masked to the results of the clinical and psychosocial assessments.
Neuropsychological measures
An extensive review of previous literature on this issue guided our choice
of neuropsychological tests. To enhance replication, only tests frequently
documented in the neuropsychological literature were used
(Lezak, 1995). Participants
completed a comprehensive battery of tests spanning 4 broad cognitive domains.
Tests were administered according to standard instructions and took about 90
min to complete. The tasks were given in the same order to the whole sample.
The instruments administered for each domain are described elsewhere
(Martinez-Aran et al,
2004a):
Statistical analyses
The three groups (bipolar I, bipolar II and healthy controls) were compared
on clinical and socio-demographic variables using analysis of variance (ANOVA)
and chi-squared tests. Multivariate analysis of variance was performed to show
overall differences in neuropsychological tests between groups. Since multiple
dependent variables were used, a prior protective analysis of covariance was
performed with age as covariate and group as a main factor. The differences
shown between the scores on the YMRS and HRSD, when controlled for, did not
significantly alter the results, so these variables were not finally included
as covariates. Since neuropsychological tests are naturally correlated, this
procedure was considered better than Bonferroni inequality correction, which
would have increased type II error. Group differences between the bipolar I,
bipolar II and control samples were tested in one-way ANOVA, followed by Tukey
post hoc comparison procedure when significant main effects were
present. The effects sizes have been calculated to find the difference between
the groups in terms of standard deviation. Pearson correlations were used to
analyse which clinical and neurocognitive measures were related to
psychosocial functioning, as measured by the GAF, taking into account
variables that showed group differences (P0.1). In patients with
bipolar II disorder, we used a multiple linear regression model to identify
the variables that would be good predictors of psychosocial functioning. The
clinical and neuropsychological variables that correlated with the GAF were
introduced in the model using a hierarchical stepwise method: clinical
variables were introduced in block 1 and neuropsychological variables in block
2. A logistical regression test was also performed to identify predictive
variables of occupational adaptation, as defined above. The variables included
in the analysis were the same as in the multiple linear regression model. Data
analyses were performed using the Statistical Package for the Social Sciences,
version 10.0 for Windows.
|
|
|---|
|
View this table: [in a new window] |
Table 1 Demographic and clinical characteristics of the study sample
|
With regard to neuropsychological variables, results are shown in Table 2. Multivariate analysis of covariance yielded Pillai's F=1.952, d.f.=30, 170 (P=0.004) for the main effect, indicating that there were overall differences in neuropsychological performance between groups. For 12 of 15 comparisons the differences reached statistical significance (P<0.05). In general, patients with type II disorder performed poorly on most neuropsychological measures compared with healthy controls, especially on measures related to semantic verbal fluency (animal naming) and verbal learning and memory (CVLT learning task, cued short-delay and long-delay-recall and recognition hits). Both bipolar disorder groups performed worse than the control group on attention (TMT part A and Digit-Span Forwards) and working memory measures (DigitSpan Backwards). In another measure of working memory (TMT part B) only a trend towards a poorer performance was detected in patients compared with controls. Patients with type II disorder, as well as the bipolar I group, showed a trend towards a higher number of WCST perseverative errors compared with healthy controls (F=2.90. P=0.06). Tukey post hoc analysis showed that the bipolar I group performed worse on most measures than the bipolar II group, who in turn performed worse than the control group, so patients with bipolar II disorder showed an intermediate cognitive profile between patients with type I disorder and healthy participants.
|
View this table: [in a new window] |
Table 2 Performance on neuropsychological tests
|
The bipolar II group showed an intermediate level of performance, between the bipolar I and control groups, on the Stroop interference task and on all measures of verbal memory (CVLT). In this regard medium effect sizes were observed, as shown in Table 2 (Cohen's d values; Cohen, 1988). Analysis of the effect sizes pointed to small differences between the patient groups, suggesting that cognitive deficits are present in both groups but these dysfunctions are quantitatively more marked in bipolar I disorder. Cognitive dysfunction was present in the bipolar II group relative to the control group but differences were medium in terms of effect size. Pearson correlations were also used in order to establish which clinical variables correlated with the neuropsychological measures in the patient groups. In the bipolar II group we found a correlation between psychosocial functioning as measured by the GAF and the age at illness onset (R=-0.42, P=0.026), the HRSD (R=-0.48, P=0.004) and the Trail Making Test part B (R=-0.45, P=0.009). Patients with longer illness duration showed more slowness or diminished attention (TMT part A), more working memory dysfunctions (DigitSpan Backwards sub-test) and more deficits in executive functions (animal naming, and higher perseverative errors from the WCST).
In the bipolar I group psychosocial functioning was related to some frontal executive functions such as the FAS (R=0.41, P=0.009), the DigitSpan Backwards sub-test (R=0.39, P=0.013) and the TMT part B (R=-0.36, P=0.025), as well as the learning (R=0.37, P=0.019), short-delay recall (R=0.35, P=0.027), free and cued long-delay recall (R=0.39, P=0.013); (R=0.37, P=0.021) and recognition (R=0.32; P=0.045) measures from the CVLT.
In the bipolar II group, after selecting all the variables that were correlated with the GAF, stepwise multiple linear regression analysis showed that the variables that best predicted psychosocial functioning, as measured through the GAF, were higher HRSD score, TMT part B score and the age at illness onset. This model accounted for nearly half (49.7%) of the variance (F=9.55, P<0.001). The TMT part B accounted for nearly 18% of the variance after controlling for the effect of the clinical variables (ß=-0.41, t=-2.93, P=0.007). On the other hand, 14 of 33 patients showed poor occupational adaptation. Consistently with these results, logistical regression analysis also showed that higher TMT part B scores appear to be nearly significant as an indicator of poor occupational adaptation (Exp(B)=1.021, P=0.058).
|
|
|---|
Cognitive performance in bipolar II disorder
Patients with bipolar II disorder had many verbal memory deficits compared
with healthy controls. When compared with bipolar I patients, the bipolar I
group showed quantitatively more dysfunctions than the bipolar II. This is
consistent with a growing body of evidence that people with bipolar disorder
experience impairment in verbal learning and memory which persists during the
euthymic state (Cavanagh et al,
2002; Glahn et al,
2004; Martinez-Aran et al,
2004a,b;
Balanza-Martinez et al,
2005; Kieseppa et al,
2005). A longitudinal study would better address the differences
in cognitive performance in hypomania and mania, but all studies so far have
been cross-sectional.
Regarding executive functions, patients with type II disorder seem to make more perseverative errors in the Wisconsin Card Sorting Test. Perseverative errors may also be related to greater impulsivity, so this could be related to a higher comorbidity related to the impulsivity spectrum in type II disorder (Goldberg & Harrow, 1999; Vieta et al, 2000).
After controlling for age, the bipolar I and II groups had a worse performance than the control group on working memory measures (DigitSpan Backwards and TMT part B) and attention (TMT part A). Patients in the bipolar II group showed an intermediate level of performance between the bipolar I and control groups in verbal memory and executive functions (Stroop interference task). This suggests that working memory may be correlated with illness severity. However, bipolar II disorder has been reported to be not just a milder form of bipolar illness, but a particularly malignant subtype with regard to frequency of episodes (Vieta et al, 1997). In fact, participants with bipolar II disorder in this study had on average three more episodes than those with bipolar I disorder, but differences did not reach statistical significance owing to the higher standard deviation of the bipolar II sample.
Role of clinical and social factors
A severe illness course probably has a negative impact on social and
occupational functioning as well as on cognition. The correlations found
between psychosocial outcome and verbal memory in the bipolar I group are
consistent with other findings by our research group (Martinez-Aran et
al,
2004a,b;
2006). Patients with type II
disorder initially showing a better clinical profile than those with type I
disorder may have a worse illness course because of the greater number of
episodes, with significantly more major and minor depressive episodes and
shorter inter-episode intervals (Vieta
et al, 1997; Judd
et al, 2003). In bipolar II disorder, patients experience
more severe and longer depressions than in bipolar I disorder
(Ayuso-Gutierrez & Ramos-Brieva,
1982) and have more persistent residual depressive symptoms
(Cassano & Savino, 1997;
Benazzi, 2001). Partial
remission as well as cognitive dysfunctions may lead to impaired psychosocial
functioning in bipolar disorder. These subtle depressive symptoms might
explain why patients with bipolar II disorder have more cognitive complaints
and cognitive dysfunctions than healthy individuals even when the effect of
subtle affective symptoms is controlled for. Rapid-cycling might carry higher
risk of cognitive impairment, but as these patients were equally split between
the two groups, there is a little chance that this factor could explain the
differences between type I and II disorder in our study. Other possible
factors involved when comparing executive function between the two types of
bipolar disorder are prior psychotic symptoms and lithium treatment, which
were both more frequent in participants with bipolar I disorder. However,
looking at the effect sizes we cannot conclude that taking or not taking
lithium would explain the differences in cognitive performance between the two
groups (P=0.023). In one study
(Stip et al, 2000) it
was observed that medium-term lithium administration did not impair explicit
memory and attention in healthy participants.
Regarding psychotic symptoms, the important reduction of the effect size (approximately 50%) may mean that the higher prevalence of psychotic symptoms in bipolar I disorder would partially explain the differences in performance v. type II disorder. The presence of psychotic symptoms is a baseline diagnostic difference between the two diagnostic categories (Vieta et al, 1997) and the specific effect of psychotic features on cognitive function in bipolar disorder has not been adequately examined. A recent study did not reveal any correlation between prior history of psychotic symptoms and cognitive impairment (Selva et al, 2006). Frontal executive dysfunctions, specifically related to working memory impairment, may be related to a poorer psychosocial functioning in bipolar II disorder. Working memory dysfunctions have been found to be present in euthymic patients with bipolar disorder, even when residual depressive symptoms were covaried for (Ferrier et al, 1999). Therefore, executive dysfunctions are likely to constitute good predictors of social and occupational difficulties in patients with type II disorder, whereas problems in retaining and recovering information may be more relevant in type I disorder. These results suggest that perhaps different neurocognitive processes are involved in the psychosocial difficulties of the two bipolar subtypes. However, further research would be required to clarify our findings.
Limitations of the study
Our study was cross-sectional, whereas a longitudinal follow-up might
provide more information about the progression of cognitive dysfunctions. It
remains unclear whether cognitive dysfunction is a premorbid issue or actually
progressive in the course of the illness. A larger sample size would have
allowed more sophisticated analyses and might have shown clearer differences
between the groups, for instance with respect to the executive functions.
Another relevant issue is the baseline difference between patients and
controls in terms of medication and history of psychotic symptoms. In the
bipolar I group there was a significantly higher percentage of patients with a
previous history of psychotic symptoms compared with the bipolar II group, so
the potential impact of this variable on cognition deserves specific attention
in further research.
Clinical implications
Persistent cognitive dysfunctions, including deficits in attention,
executive function and verbal memory, exist in bipolar II disorder as in type
I disorder, so cognitive functioning should be routinely examined in patients
with either subtype. In patients with bipolar II disorder, working memory
dysfunction seems to be a good predictor of functional impairment, after
controlling for the effect of sub-syndromal symptoms. Rehabilitation
interventions should take into account potential cognitive differences between
the two subtypes, especially regarding their impact on functioning. An early
diagnosis of type II disorder is important to prevent or remediate as much as
possible the cognitive problems of these patients.
|
|
|---|
|
|
|---|
Related articles in BJP:
This article has been cited by other articles:
![]() |
C. Simonsen, K. Sundet, A. Vaskinn, A. B. Birkenaes, J. A. Engh, A. Faerden, H. Jonsdottir, P. A. Ringen, S. Opjordsmoen, I. Melle, et al. Neurocognitive Dysfunction in Bipolar and Schizophrenia Spectrum Disorders Depends on History of Psychosis Rather Than Diagnostic Group Schizophr Bull, May 14, 2009; (2009) sbp034v1. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Savitz, L. van der Merwe, D. J. Stein, M. Solms, and R. Ramesar Neuropsychological status of bipolar I disorder: impact of psychosis The British Journal of Psychiatry, March 1, 2009; 194(3): 243 - 251. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. L. Phillips and E. Vieta Identifying Functional Neuroimaging Biomarkers of Bipolar Disorder: Toward DSM-V Schizophr Bull, July 1, 2007; 33(4): 893 - 904. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||