Share this article on:

Cognitive impairment and dementia in late-life bipolar disorder

Aprahamian, Ivan; Nunes, Paula V.; Forlenza, Orestes V.

doi: 10.1097/YCO.0b013e32835ac5f6
PERSONALITY DISORDERS: Edited by Charles B. Pull and Aleksandar Janca

Purpose of review This work aims to review the most recent publications on cognitive impairment and dementia associated with bipolar disorder (BPD), especially in the elderly.

Recent findings In the last years, a growing number of studies aiming at better understanding of cognitive impairment in BPD were found. Impairments found in BPD were compared with other psychiatric disorders and primary cognitive diseases. The impact of cognitive impairment on functionality was also recently highlighted. With respect to neurobiology, studies that explored inflammatory, neurotrophic and pathological cascades possibly associated with BPD and cognition were published. Finally, the first study covering treatment of cognitive impairment was carried out with pramipexole, and it raised important methodological issues for future research in BPD.

Summary Cognitive impairment and dementia in BPD should be better explored with cognitive and functionality protocols along with biological and neuroimaging markers.

Laboratory of Neuroscience, Department and Institute of Psychiatry, Faculty of Medicine, University of São Paulo, São Paulo, Brazil

Correspondence to Ivan Aprahamian, Rua dr. Ovidio Pires de Campos, 785, 3rd floor, Pinheiros, São Paulo, Brazil 05403-010. Tel: +55 11 3069 7267; e-mail:

Back to Top | Article Outline


Many affective disorders such as bipolar disorder (BPD) and depression are associated with increased risk for cognitive impairment and dementia, mostly after long-time follow-up. Initially, cognitive deficits seen on these disorders were associated with continuous or intermittent burden of mood expression, treatment adverse effects from medications, comorbidities such as alcohol and drug dependence, anxiety and personality disorders. They were also associated with degenerative disorders’ neuropathological changes, the most severe cases or the psychosocial consequences of having a chronic and recurrent disease such as BPD. In fact, none of the latter theories fell apart with the current evidence. Cognitive impairment and dementia are progressively suggested to be part of the clinical expression of BPD and closely connect with its own pathological substrate.

Back to Top | Article Outline


Most studies of cognition on BPD are conducted in samples of young to middle-aged adults. Deficits have been described affecting verbal memory, executive function, attention, visual memory, mental speed and, to a lesser extent, language and visuospatial function. The most frequently reported impairments involve executive function and verbal memory, which seem to occur regardless of age. These deficits are not usually severe in nondemented BPD patients, ranging between 0.2 and 1.0 SDs below the respective norms [1]. Nevertheless, cognitive impairment negatively affects the functional capacity of patients with BPD, and exerts a negative effect on the global prognosis [2].

In the last 2 years, the only experimental study conducted in elderly BPD is the first to compare cognition in older patients with schizophrenia and bipolar I disorder, employing stringent definitions of symptomatic remission and euthymia [3▪▪]. They compared the cognitive abilities of 67 community-living patients with schizophrenia, controlling for their state of symptomatic remission (20 were in symptomatic remission, 47 were not in symptomatic remission) with those of 74 euthymic patients with bipolar I disorder. Cognitive dysfunction is of particular concern in later life, as the interaction between aging and previously present cognitive deficits may lead to amplification of impairments. Older patients with schizophrenia are not inordinately prone to the development of Alzheimer's disease, and long-term stability of cognitive functioning over time is well established. In patients with BPD, premorbid development is relatively normal, whereas disease exacerbations are associated with increasing cognitive dysfunction (see the review of Lewandowski et al. [4] and Vieta et al. [5]). The net outcome of the diverging trajectories of schizophrenia and BPD may be that in later life the two disorders approach one another in severity of cognitive impairments. In this study, schizophrenia and bipolar I patients were impaired relative to the healthy controls (with mostly large effect sizes) in verbal memory, executive function and verbal fluency. Differences in cognitive performance between remitted schizophrenia patients and bipolar patients were notably limited, with neither of the two groups demonstrating a systematically better performance. Although this last finding should be considered provisional because of the small sample size of the remitted schizophrenia group, it may be tentatively concluded that in later life schizophrenia and BPD share major phenotypic similarity in their trait-related cognitive impairments. In this older sample, repetition of the analyses including only bipolar patients who had a history of psychosis had no substantial impact on the between-group comparisons. Adjustment for premorbid intelligence quotient (IQ) did not essentially change the results. Finally, the widespread and differential use of psychotropic medication among the patient groups prevented evaluating their influence on cognitive functioning. The conclusion of this study is that, in older age, community-living patients with schizophrenia and with BPD suffer from severe and pervasive cognitive deficits. The finding that cognitive impairments in the symptomatically remitted patients with schizophrenia were largely similar to those of the euthymic bipolar I patients is in line with the evidence for key similarities in the profile of trait-related cognitive dysfunction in the two disorders, but warrants replication in larger samples. In addition, future research using longitudinal study designs may clarify the eventual impact of differential mortality between schizophrenia and BPD on late life cognitive outcome, as well as elucidate the factors contributing to cognitive deterioration from a lifespan perspective.

Box 1

Box 1

We describe below studies made in nonelderly bipolar patients that evaluate factors potentially involved in the development of dementia in BPD. Bearden et al. [6] evaluated the relationship between neurocognition and occupational function in 79 previously employed adults with bipolar I disorder. The objective of the study was to investigate the reasons why many patients with BPD do not regain their premorbid level of occupational functioning even after mood episodes have been remitted. Better neurocognitive function in multiple domains, particularly in the domains of episodic memory, attention/working memory, and executive function, and improvement in these domains over time were strongly predictive of subsequent occupational recovery. Further research in this area is critical to pinpoint the causes of long-term functional disability in patients with BPD and better target treatments capable of enhancing functional outcome (see also the review of Vieta et al. [5]).

Solé et al. [7] evaluated whether 43 strictly defined adult euthymic bipolar type II patients, a less studied group, would present neurocognitive disturbances. They also evaluated the impact of these disturbances on functional outcome as measured by the Social and Occupational Functioning Assessment Scale. These patients were compared with 42 demographically and educationally matched healthy individuals. In accordance with previous studies, this report provides further evidence that the bipolar type II patients also present cognitive impairments in attention, learning and verbal memory, and executive functions. In these bipolar type II patients, psychosocial functioning was mainly affected by subthreshold-depressive symptoms and executive functioning.

These two latter studies regarding functionality in BPD individuals exemplify some of the challenges in establishing functional impairment directly linked to cognitive decline, which is the cornerstone to diagnose dementia. BPD patients often present with socio-occupational and psychosocial problems associated with symptomatology and comorbidities such as alcohol and drug dependence. Once BPD is usually characterized by oscillations of mood that influence functionality it is difficult to establish a continuous decrease in functionality as seen in dementia, especially in Alzheimer's disease.

Osher et al. [8] compared neuropsychological functioning of 51 euthymic adult bipolar patients with 162 elders with mild cognitive impairment (MCI) and 495 healthy adult controls (older than the bipolar patients). Bipolar and MCI groups showed statistically equivalent functioning in memory, executive function, verbal function and information processing speed. In the domains of visual–spatial processing, attention and motor skills, the MCI group outperformed the bipolar group. In every domain, the healthy control group outperformed both the bipolar and the MCI groups. Neurocognitive function did not appear to be strongly influenced by medication class, although a larger sample would have allowed more precise comparisons. The level of cognitive impairment found in euthymic bipolar patients has some similarities to the level of impairment found in patients diagnosed with MCI, even though the bipolar individuals were, on average, over 30 years younger. However, we must emphasize that comparison groups were paired neither on age nor on diagnosis.

Most studies suggest that generalized, rather than specific, cognitive impairment characterizes euthymic BPD. Age, illness duration, education and clinical course may moderate these broad cognitive impairments (see review of Mann-Wrobel et al. [9]). Differently from this evidence, in the work of Xu et al. [10] unipolar depression-remitted patients had impairment in processing speed, visual memory, verbal fluency and executive function, whereas bipolar type I-remitted and type II-remitted patients showed cognitive dysfunction only in the domains of processing speed and visual memory.

Episodic memory impairment is a robust correlate of familial risk for schizophrenia and BPD; still much is unknown about the processes that underlie this deficit and how they may be implicated in BPD and schizophrenia. In a small sample of patients, familial predisposition to schizophrenia, compared with that of BPD, was associated with significantly greater impairment in cognitive control processes during episodic memory retrieval, with some evidence of specificity for schizophrenia in connection with mechanisms relating to increased forgetting [11].

The Digit Symbol Test may be a valid endophenotype, highly specific for differentiating BPD patients and first-degree relatives from controls [12]. This study also demonstrated that impaired processing speed is familial and does not appear to be a consequence of symptoms, premorbid IQ or level of education, but intrinsically associated with the disorder. This impaired psychomotor performance speed is in accordance with literature and may well reflect underlying vulnerability to BPD.

Back to Top | Article Outline


There are few studies exploring neurobiological basis for cognitive decline in BPD. Barbosa et al. [13▪] studied the concept that brain-derived neurotrophic factor (BDNF) and proinflammatory molecules are important contributors to the pathophysiology of BPD, and an imbalance in peripheral levels of these molecules may be implicated in the cognitive decline observed in BPD patients. The relationship of executive performance through the Frontal Assessment Battery of 25 BPD type I euthymic adult patients and plasma levels of BDNF and tumor necrosis factor (TNF)-α was investigated in relation to 25 age-matched and sex-matched healthy controls. BDNF was altered in BPD but did not correlate with executive functioning. Executive function was correlated with age and the Mini-mental State Examination, but not with BDNF plasma levels, and neither executive function nor BDNF level was influenced by psychiatric and clinical comorbidities or by medications in use. Interestingly, TNF-α plasma levels correlated with inhibitory control (as part of the executive function battery) in BPD patients.

Reduced plasma or cerebrospinal fluid levels of β-amyloid (Aβ) are one of the most important downstream markers of Alzheimer's disease pathology. Piccinni et al. [14▪] investigated changes of plasma levels of Aβ40 and Aβ42 and/or of Aβ40/Aβ42 ratio in a group of depressed BPD patients, and the possible relationship between the clinical course of the disorder and Aβ levels. Sixteen BPD patients suffering from a major depressive episode were compared with a control group. The Aβ42 mean plasma levels in the patients’ BPD were significantly lower. Correlations were found with time since first episode of BPD and number of affective episodes. Therefore, changes in plasma levels of different Aβ peptides might represent a useful tool to identify the risk for cognitive decline in bipolar patients.

Back to Top | Article Outline


There have been very few controlled trials with enhancement in cognition as a primary outcome in patients with BPD. Although the reasons for persistent cognitive impairment in BPD are not well understood, one possible strategy for improvement is to augment dopaminergic activity. Burdick et al. [15▪] addressed this issue in this first controlled trial of pramipexole, a dopamine agonist, with cognition as the primary outcome measure in patients with BPD. Previously, this group showed data regarding safety and efficacy issues for pramipexole for treatment-resistant bipolar depression, but with significant improvement of attention and processing speed as secondary findings. The authors conducted an 8-week, double-blind, placebo-controlled trial of pramipexole versus placebo in 50 stable adult outpatients with BPD. No significant benefit for pramipexole over placebo was found. However, after multivariate analysis the group of euthymic BPD patients (18 within placebo and 16 with pramipexole) presented differences in two neurocognitive measures regarding processing speed and working memory (P = 0.003). The modest cognitive beneficial effect of pramipexole in the subgroup of euthymic patients with BPD deserves replication in future trials, given some important methodological limitations.

Back to Top | Article Outline


There is undoubtedly a growing interest in research on the cognitive impairment and dementia of BPD, and the number of publications in this field has increased considerably over the past years; however, methodological differences across different studies and the course of the disease itself may preclude the generalization of findings. In this review over the past 2 years of research, we found that cognitive impairment in BPD approaches the impairment seen in other disorders such as schizophrenia and MCI, but develops in a different course. The level of proinflammatory factor (TNF) was correlated with cognitive function status. The level of β-amyloid or amyloid peptides was correlated with the presence of depressive symptoms and the clinical course of this disease. In light of the findings from the pramipexole study for the treatment of cognitive impairment in BPD, we first need to better understand the clinical presentation and evolution of the cognitive decline seen in BPD to achieve successful treatments for this issue.

Back to Top | Article Outline


The authors receive financial support from FAPESP (grant n° 02/13633–7 and 2009/52825–8), Associação Beneficiente Alzira Denise Hertzog da Silva (ABADHS), CNPq (grant n° 554535/2005–0), Alzheimer's Association (NIRG-08–90688).

Back to Top | Article Outline

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline


Papers of particular interest, published within the annual period of review, have been highlighted as:

  • ▪ of special interest
  • ▪▪ of outstanding interest

Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 132).

Back to Top | Article Outline


1. Robinson LJ, Ferrier IN. Evolution of cognitive impairment in bipolar disorder: a systematic review of cross-sectional evidence. Bipolar Disord 2006; 8:103–116.
2. Depp CA, Moore DJ, Sitzer D, et al. Neurocognitive impairment in middle-aged and older adults with bipolar disorder: comparison to schizophrenia and normal comparison subjects. J Affect Disord 2007; 101:201–209.
3▪▪. Meesters PD, Schouws S, Stek M, et al. Cognitive impairment in late life schizophrenia and bipolar I disorder. Int J Geriatr Psychiatry 2012. doi: 10.1002/gps.3793. [Epub ahead of print]

It is the only study in elder bipolar patients and gives an interesting comparison of cognitive decline in schizophrenia and BPD.

4. Lewandowski KE, Cohen BM, Ongur D. Evolution of neuropsychological dysfunction during the course of schizophrenia and bipolar disorder. Psychol Med 2011; 41:225–241.
5. Vieta E, Popovic D, Rosa AR, et al. The clinical implications of cognitive impairment and allostatic load in bipolar disorder. Eur Psychiatry 2012. [Epub ahead of print]
6. Bearden CE, Shih VH, Green MF, et al. The impact of neurocognitive impairment on occupational recovery of clinically stable patients with bipolar disorder: a prospective study. Bipolar Disord 2011; 13:323–333.
7. Solé B, Bonnin CM, Torrent C, et al. Neurocognitive impairment and psychosocial functioning in bipolar II disorder. Acta Psychiatr Scand 2012; 125:309–317.
8. Osher Y, Dobron A, Belmaker RH, et al. Computerized testing of neurocognitive function in euthymic bipolar patients compared to those with mild cognitive impairment and cognitively healthy controls. Psychother Psychosom 2011; 80:298–303.
9. Mann-Wrobel MC, Carreno JT, Dickinson D. Meta-analysis of neuropsychological functioning in euthymic bipolar disorder: an update and investigation of moderator variables. Bipolar Disord 2011; 13:334–342.
10. Xu G, Lin K, Rao D, et al. Neuropsychological performance in bipolar I, bipolar II and unipolar depression patients: a longitudinal, naturalistic study. J Affect Disord 2012; 136:328–339.
11. Christodoulou T, Messinis L, Papathanasopoulos P, Frangou S. The impact of familial risk for schizophrenia or bipolar disorder on cognitive control during episodic memory retrieval. Psychiatry Res 2012; 197:212–216.
12. Daban C, Mathieu F, Raust A, et al. Is processing speed a valid cognitive endophenotype for bipolar disorder? J Affect Disord 2012; 139:98–101.
13▪. Barbosa IG, Rocha NP, Huguet RB, et al. Executive dysfunction in euthymic bipolar disorder patients and its association with plasma biomarkers. J Affect Disord 2012; 137:151–155.

This study evaluated the correlation between cognitive function and plasma BDNF and TNF.

14▪. Piccinni A, Origlia N, Veltri A, et al. Plasma β-amyloid peptides levels: a pilot study in bipolar depressed patients. J Affect Disord 2012; 138:160–164.

This study evaluated plasma levels of Aβ in depressed patients with BPD.

15▪. Burdick KE, Braga RJ, Nnadi CU, et al. Placebo-controlled adjunctive trial of pramipexole in patients with bipolar disorder: targeting cognitive dysfunction. J Clin Psychiatry 2012; 73:103–112.

bipolar disorder; cognitive impairment; dementia; neurobiology; treatment

© 2013 Lippincott Williams & Wilkins, Inc.