Bipolar disorder and dementia: where is the link?


Professor Emmanuel Haffen, MD, PhD, Department of Clinical Psychiatry, University Hospital of Besançon, Besançon 25000, France. Email:


Cognitive disorders appearing in the course of bipolar disease have been identified, and recent studies have defined the neuropsychological characteristics of this pathology, which includes attention, executive function, memory and language disorders. However, questions remain concerning the appearance of dementia symptoms over the course of bipolar disorder in certain patients: is it a chance association or is there a connection between bipolar disorders and dementia? If the latter hypothesis is considered, what is the nature of the dementia, which might be considered as a dementia specific to bipolar disorder? Current clinical, neuropsychological and cerebral imaging data are inconclusive, but similarities with frontotemporal dementia might be highlighted. Functional imaging studies might provide answers as well as more specific tests in neuropsychology. The cause of cognitive damage in bipolar disease also raises questions concerning a neurodevelopmental or neurodegenerative process, because several factors seem to influence cognition and these two processes might occur simultaneously. Long-term studies are necessary to determine whether cognitive deterioration in bipolar disease is stable or progressive. There might also be different neurobiological subgroups of patients with bipolar disease.


The onset of cognitive illness in the long-term outcome of psychiatric pathologies has already been studied.1 Cognitive decline in patients with a unipolar or bipolar affective disorder is not only a consequence of social or psychological factors related to chronic illnesses, but also seems to increase the risk of developing dementia compared with patients with other chronic illnesses, regardless of alcohol or drug use or abuse.2 Research over the past 20 years has focused primarily on cognitive function in schizophrenia, so cognitive impairment and its impact in this area are relatively well-known. The study of cognitive functioning in bipolar disorder is more recent. There is uncertainty about the outcome of cognitive impairment in the course of the disease with the emergence of dementia-like syndromes that raise questions about the link between bipolar disorder and dementia. Is it a chance association or is there an etiopathogenic link, a connection between bipolar disorder and dementia? If the latter, what is the nature of this specific dementia in bipolar disorder, which seems to resemble frontotemporal dementia (FTD)? Without claiming to answer these questions definitively, we present two clinical cases involving an outcome of severe or moderate cognitive impairment in bipolar disorder, echoing the diagnostic features of bipolar disorder and FTD described recently in the literature.


The first case concerned a 62-year-old man referred to our department and treated for a severe depressive episode, gradually deteriorating over 8 months. We observed a lack of compliance with therapy dating 15 days before his hospitalization. Previous daily treatment included mirtazapine 15 mg,escitalopram 20 mg and valpromide 600 mg, as well as lorazepam and propericiazine. He was single, childless, had lived alone since the death of his parents, and had worked freelance until 2 years previously after completing further study. Medical history of psychiatric disorder showed an affective disorder going back approximately 30 years, with melancholic depressive episodes interspersed with hypomanic episodes. The family reported a series of manic or depressive episodes. A diagnosis was made of bipolar disorder type II. His medical history was as follows: unobstructive hypertensive cardiomyopathy, a preventive cholecystectomy and a strong family history of cardiopathy. There was no history of substance or alcohol abuse. On admission, the patient's symptoms included abulia, apragmatism, anorexia, almost total insomnia and morbid preoccupations. The family had noticed impaired comprehension, and problems with memory and disorientation for approximately 2 years, with a gradual reduction in autonomy. The patient had little awareness of his disorder. We observed spatial-temporal disorientation, difficulty concentrating and ideational perseverations. Appropriate psychoactive drugs were prescribed during hospitalization, which led to a rapid improvement in mood and sleep patterns, and reduced anxiety. However, cognitive and memory problems persisted, associated with significant behavioural problems including withdrawal, apathy and abulia. Daily treatment included 75 mg clomipramine, 900 mg valpromide and 35 mg clonazepam. A computed tomography scan showed cortical-subcortical cerebral atrophy, which was considered abnormal given the age of the patient. A magnetic resonance imaging (MRI) found advanced generalised cortical atrophy, which was abnormal for the patient's age, and a discrete chronic vascular leukoencephalopathy and no signs of recent ischaemia (Fig. 1).

Figure 1.

Patient 1. Axial T2 FLAIR: advanced generalized cortical atrophy and discrete chronic vascular leukoencephalopathy without signs of recent ischaemia.

To eliminate the possibility of a degenerative disease, a lumbar puncture was required to carry out a cytobacteriological examination of the spinal fluid, electrophoresis tau and β-amyloid. The results were inconclusive. To eliminate vascular disease, a cervical Doppler ultrasound showed atheromatous plaques justifying permanent anticoagulation treatment, but no haemodynamically significant lesions. An initial cognitive evaluation found an overall cognitive alteration (Table 1). A multidisciplinary consultation at the Centre for Memory, Resources and Research concluded a mesial frontal lobe syndrome with slow progressive change, but with a recent rapid deterioration. An increase in cognitive disorders, identified by deteriorating results after a second neuropsychological assessment, was observed a year later (Table 1). The patient's mood was considered ‘stable’ and there was no clinical evidence suggesting depression. No anxiety, sleep or appetite problems were observed, but the patient showed a certain indifference towards himself and his surroundings. Sphincter conduct disorders were found. The patient did not complain of mood or cognitive disorders. Daily treatment combined 150 mg venlafaxine, 900 mg valpromide, 3 mg clonazepam and 10 mg propericiazine at bedtime. The patient presented with rapidly progressive general cognitive impairment, with impairment of executive functions and frontal syndrome. The diagnosis of dementia seems obvious, but its cause and characteristics raise questions. The disease progressed quickly, with a loss of 5 points in the Mini-Mental State Examination over the course of a year. This patient had not previously complained of cognitive impairment and his care only began when the disturbances became visible and disturbing for his family. The behavioural element was significant, with emotional blunting and a loss of initiative and interest. A diagnosis of Alzheimer's disease was eliminated due to his age, progression of cognitive disorders and the results of imaging and spinal tap. There were no signs suggestive of Lewy body disease. The clinical picture strongly resembled FTD according to diagnostic criteria by Neary et al.,3 but the rapid progression and the neuropsychological assessment (normal denomination) make this diagnosis less clear.

Table 1.  Neuropsychological test results for patient 1 followed for bipolar disorder for 30 years
TestsFirst evaluation12 months later
  1. DMS, delayed matching to sample; DRS, Dementia Rating Scale.

Mini-Mental State Examination18/3013/30
Memory Impairment Screen score4/83/8
Clock Drawing Task0/50/5
Isaacs Set Test66
Crossing Off Test7253
Buschke Cued Recall Test score  
 Immediate recall8/168/16
 Free recall3/480/48
 Total recall31/4826/48
Trail-Making Test score  
 Part AFailedFailed
 Part BFailedFailed
Copy6/6Not carried out
Picture naming  
 30 items29/3028/30
 80 items80/80Not carried out
Test for matching categories4/10Not carried out
Mattis DRS86/144Not carried out
DMS 48 set 1Not carried out60
DMS 48 set 2Not carried out60
Frontal Assessment Battery scoreNot carried out6
Luria's graphic series  
 CopyNot carried outnormal
 ModelNot carried outperseveration
BrixtonNot carried out51/55
Hayling test  
 Part ANot carried out15/15
 Part BNot carried outFailed


The second case involved a woman aged 77 years, who was followed for 30 years for bipolar disorder type I. This patient was single, childless and retired, having held a position of responsibility in the public sector. The onset of the disease began around the age of 35 years. The disease was marked by numerous hospital admissions for depressive, mixed and manic episodes, sometimes with psychotic (delusional and hallucinatory) features. Multiple treatments had been undertaken in the past – mood stabilizers, antidepressants, antipsychotics and electroconvulsive therapy sessions. The patient's condition was satisfactory between 1995 and 2008, with no hospitalization during this period. The patient's clinical history was marked by a hospital admission in 2008 after a manic episode with delusions of persecution, then 2 months later, mood changes involving depressive mood with sadness, abulia, anhedonia, clinophilia, psychomotor slowdown and anorexia. Her condition improved after a change in treatment. Despite an increase in antidepressant treatment, depressive symptoms soon recurred involving abulia, anhedonia, carelessness, clinophilia and anorexia with delusions of persecution. The woman and her entourage then reported a progressive loss of autonomy and increasing difficulties managing her paperwork. Altogether, these symptoms led to her being readmitted to hospital. After her mood was stabilized, the patient complained of having no memory of recent events and difficulty speaking, which lasted for approximately a year. The patient also complained of visual hallucinations: she said she could see shadows and people; these episodes seemed to be ephemeral. An initial assessment consisting of cerebral MRI, a neuropsychological assessment and a speech assessment were carried out (Table 2). The MRI found cortical with subcortical atrophy spreading predominantly over the bilateral frontotemporo-island with an internal bilateral temporal atrophy (Fig. 2). A scan carried out 15 years previously had already found cortical and sub-tentorial atrophy. The speech therapy assessment highlighted a difficulty in finding words without any trouble understanding that could evoke primary progressive aphasia. Given these cognitive impairments associated with lesions detected by the MRI scan, onset of dementia could not be ruled out and could thus explain the occurrence of mood relapses for 1 year. However, brain atrophy is common in long-term bipolar disorder, and because cognitive disorders in the patient could be attributed to residual symptoms of mood episodes, no definitive diagnosis could be made.

Table 2.  Neuropsychological test results for patient 2 followed for bipolar disorder for 30 years
TestsFirst evaluation6 months later12 months later
  1. DMS, delayed matching to sample.

Memory Impairment Screen Score7/87/8Not carried out
Clock Drawing Task0/50/5Not carried out
Isaacs Set Test18924
Crossing off Test116130185
Mini-Mental State Examination26/3020/3027/30
Buschke Cued Recall Test score   
 Immediate Recall14/1616/1614/16
 Free recall14/4819/4826/48
 Total recall35/4837/4846/48
Trail-Making Test score   
 Part A825865
 Part B313360212
Picture naming score   
 30 items25/3022/3029/30
 80 items72/8069/8071/80
Copy score6/66/66/6
Test for matching categories10/10Not carried out10/10
DMS 48 Set 1 (%)9287100
Phonological FluencyNot carried out512
Rey-Osterrieth figure copyNot carried outNot carried out36/36
Visual Object and Space Perception Battery   
 Shape detection screening testNot carried out20/2019/20
 Incomplete lettersNot carried out17/2019/20
 SilhouettesNot carried out8/2019/20
Figure 2.

Patient 2. Axial T2 FLAIR: subcortical atrophy predominantly over the bilateral fronto-temporo-island with internal bilateral temporal atrophy and discrete chronic vascular leukoencephalopathy.

Less than 2 months after discharge, another hospital admission was required before depressive symptoms and persistent problems in autonomy recurred, despite her living in a more secure environment. During hospitalization, a rapid improvement in mood and a gradual recovery of autonomy was observed. The patient expressed the need for more company and was sent to a nursing home. She was again admitted to hospital 3 months later as a result of continuous falls and deterioration in her general condition with temporo-spatial disorientation. After eliminating an acute organic cause, a second dementia assessment was undertaken in view of the progression of the disorder. A perfusion scintigraphy brain scan and a DaTSCAN (Fig. 3) evaluating dopaminergic neurotransmission using Neurolite was carried out. The perfusion scan showed bilateral mesial superior frontal hypofixation and frontotemporal bilateral junction extended towards the upper left frontal region. There was also a low uptake in internal temporal bilateral hypofixation, occipital higher bilateral and left parietal hypofixations. There was a discrete uptake in the left striatum and an asymmetry in fixation of the thalamus on the right side. There were many scattered bilateral hypoperfusions in the brain. Scintigraphy of dopaminergic neurotransmission showed homogeneous bilateral fixation and normal intensity in the striatum. No scintigraphic argument was found in favour of a dopaminergic neurotransmission disorder. Dementia with Lewy bodies was then suggested in view of a progressive cognitive decline (Table 2) and the presence of parkinsonian syndrome, which was characterized by bilateral extrapyramidal signs. The absence of impaired dopaminergic neurotransmission might be misinterpreted as a result of the presence of atrophy. After 3 months in hospital and changes in treatment (withdrawal of divalproex sodium allowing levodopa to be stopped and lamotrigine to be started), an improvement in mood was observed and akathisia disappeared. A controlled evaluation did not find any argument in favour of a dysexecutive syndrome, or of visual-spatial, visual-perceptual or aphasic deficits (Table 2). The language evaluation carried out at the same time showed an improvement compared with the previous assessment. However, phasic disorders were highlighted, as well as difficulties in visual recognition and a partial alteration of the semantic system. Discussing these results in multidisciplinary consultations eliminated the hypothesis of dementia with Lewy bodies on the basis of the evolution of cognition, the disappearance of extrapyramidal syndrome and the DaTSCAN data. The hypothesis of an underlying degenerative process could not be eliminated, given the persistence of phasic and praxis disorders, despite a significant improvement in assessments. The hypometabolism detected by scintigraphy was related to the spread of atrophy on MRI images. The hypothesis of cognitive disorders related to mood changes in a person aged 77 years followed for 40 years for bipolar disorder was considered in view of this atypical case study.

Figure 3.

Patient 2. Perfusion scintigraphy brain scan:1–3 bilateral mesial superior frontal and left orbitofrontal hypofixation. DaTSCAN:4 homogenous bilateral fixation.


These two clinical situations show the difficulty in analysing the relationship between bipolar disorder and cognitive impairment. We suspect there is a specific dementia profile following bipolar disorder, but a chance association with FTD is not excluded. Both clinical situations described in the present case report show an outcome of bipolar disease in patients with a high sociocultural level leading to a potentially considerable disability. The exact nature of these dementia disorders remains difficult to establish despite data from imaging, neuropsychological assessments, orthophonics and the opinion of the multidisciplinary memory centre. The current question concerns whether or not dementia following bipolar disease exists. We assume there would be no return ad integrum of cognitive functions during intercritical phases. The literature leads us to believe that the persistence of a cognitive deterioration during the euthymic phase and progressive cognitive damage can end in dementia. This ‘specific’ dementia would be characterized by cognitive achievement centred on attention, executive functions, verbal memory and language with a behavioural element of the frontal type.4 However, similarities might be observed between the outcome of cognitive deterioration in bipolar disorder and FTD, as follows.

From a clinical point of view

FTD manifests itself in changes of personality and behaviour, and resembles a psychiatric pathology, causing diagnostic errors.5

Clinical confirmation of the beginning of FTD appears difficult in patients followed for bipolar disorder. Apathy is frequently found in patients with dementia following a bipolar disorder,6 and is a dominant behavioural frontal symptom of FTD regardless of the degree of damage caused by dementia.7

From a neuropsychological point of view

Cognitive damage including difficulty sustaining attention, executive function, verbal memory and language disorders has been observed in bipolar disorder. These disorders might lead to a clinical presentation of frontal-subcortical dementia. A frontal behavioural deficit might be noted together with social disability.8 These symptoms are associated with frontal disorders and the social cognition present in FTD.

From an imaging point of view

Diverse damage has been described in bipolar disorder. The volumetric changes shown by MRI show an atrophy of the frontal and temporal regions.9 This damage causes frontal and temporal circuit dysfunction. Therefore, certain cognitive deficits might be associated with underlying cerebral abnormalities.10 Functional imaging showed predominant hypoperfusion in the frontotemporal region. These data can be strongly associated with the frontal and social cognition disturbances found in FTD.

However, dementia syndrome arising as a result of bipolarity as described by Lebert does not seem to correspond to the criteria of the main types of dementia, so we suspect specific dementia in bipolar disorder.6 The argument against FTD involves the behavioural aspect. The frontal signs in dementia following bipolar disease would be present but less severe than in FTD, and we would find self-neglect and emotional indifference less frequently. The differences between dementia following a bipolar disorder and FTD or frontotemporal lobar degeneration (FTLD) could relate to outcome and prognosis.

The progression towards dementia seems fast in the first patient and progressive in the second patient. Can we therefore assume that FTD occurred by chance in the first case and that there is a link between bipolar disorder and dementia in the second case? The concept of clinical heterogeneousness is present in these two pathologies, but it is difficult to establish precise diagnostic criteria, particularly in FDT. The changing diagnostic criteria of FTD are constantly being discussed, with increasing knowledge of molecular genetics and biochemistry.11 Future studies, in particular with functional neuroimaging data, should allow for progress in the field of frontotemporal atrophies.

In the study by Velakoulis et al.,12 the possibility that schizophrenia or bipolar disorder is in fact the first stage of FTD is reinforced by the possible overlapping of clinical data, neuropsychology and neuroimaging. These two pathologies could be linked by a pathological process in the same cerebral region. So a subgroup within FTLD could be defined, allowing a connection to be made with the dementia outcome in some bipolar patients. The hypothesis of a different psychotic bipolar subgroup of patients in the neurobiological aspect of bipolar disease has been proposed.13

The source of the cognitive damage that appears in bipolar disease remains uncertain. Doubts persist on whether cognitive deficits are pre-existing in the disease, which would support the neurodevelopmental hypothesis, or if they are the consequence of the impact of the disease, supporting the hypothesis of a neurodegenerative process.

Several factors linked to the disease seem to influence cognition: clinical variables (number of episodes, in particular mania, number of hospitalizations and duration of the disease), hormonal biological factors, and pharmacological and diagnostic variables.14

Concerning the cumulative neurological ‘toxicity’ of the thymic episodes

A study in Denmark summarizing the registers of hospitalizations for a thymic disorder between 1970 and 1999 showed a risk of dementia that increases with the number of thymic episodes having led to hospitalization. This risk seems greater in unipolar depressive disorder than in bipolar disorder. Therefore, every episode increases the risk of a diagnosis of dementia in depressive disorder by 13% and by 6% in bipolar disorder.15 The possibility of a direct link between manic depressive disorder and dementia is raised, which supports the hypothesis of neurotoxicity in the combined thymic episodes.9,16 The most dangerous type of thymic episode is also controversial: manic17 and depressive.10 However, compared with the bipolar I subtype, patients with bipolar II disorder show a similar pattern of functional impairment, suggesting that disability in bipolar II patients is just as significant as in bipolar I patients. This might be explained by the fact that bipolar II patients experience more residual depressive symptoms overall, which points to the importance of treating depressive residual symptoms.18

Concerning the effects of treatments

The role of psychotropic treatments prescribed during the disease must be taken into account: lithium, anticonvulsants, antidepressants, antipsychotics and tranquillizers. These various psychotropic drugs can modify cognition, but it is important to distinguish the symptoms being explored, such as a decline in attentiveness or sedative effects, from side-effects and iatrogenic effects. Lithium, which has been used for a long time in bipolar disease, has been seen as both neurotoxic and neuroprotective.

Iatrogenic effects as a potential aggravating factor cannot be the only cause, but must nevertheless be taken into account. Pursuing treatment with lithium seems to be associated with a decrease in the risk of developing dementia.19 Current neuroanatomical and neurobiological data show favourable neuroprotective properties of lithium, in particular on hippocampus cells.20 A recent meta-analysis showed that lithium treatment appears to have limited and minor negative effects on cognition.21

Studies in general support the neurodegenerative model which, while not corresponding perfectly, seems more suitable; cognitive dysfunction being acknowledged as being correlated to psychosocial functioning, although it is still possible that both hypotheses are compatible.22

Long-term studies are necessary to be able to conclude and specify the outcome: stable or progressive cognitive deterioration.

The study of Mur et al.23 shows deficits after a 2-year follow up of bipolar euthymic patients. Cognitive disorders seem stable as patients age from the point of view of the altered functions and their severity, and the lack of change in cerebral morphology suggests a progressive neurotoxic process.24 Few studies have looked at premorbid functioning to find out whether certain cognitive deficits existed before disorders begin.

A preliminary study carried out in Spain suggests that a subgroup of bipolar and schizophrenic patients presents with a gene transfer involved in neuronal migration and that these changes can predict the seriousness of frontal deficits in both diseases. Therefore, a percentage of bipolar patients might present changes associated with neurodevelopment.25 It is important to differentiate between subgroups of patients in terms of their neuropsychological profile, in particular concerning psychotic symptoms, which cause cognitive damage.26,27 Schizo-affective patients present with a greater cognitive impairment than bipolar patients without psychotic symptoms and healthy controls.28

Bipolar disorders are therefore associated with persistent deficits in terms of attention, verbal memory, and learning and executive functions. These deficits are probably the result of a combination of genetic factors and environmental risk factors, therefore relating to a neurodevelopmental and degenerative process. The role of treatments on cognition is still under debate, but the first aim in treating bipolar disease patients is to assure thymic stabilization.29

These clinical observations describe progression towards a dementia process for patients suffering from a bipolar disorder for at least 30 years. The cognitive profile is distinct, with damage to memory, language, executive and behavioural function, and associated with cerebral abnormalities in imaging. The literature is rich in cognitive impairments in bipolar disorder, but the question of a dementia syndrome outcome remains to be addressed.

A ‘specific dementia’ in bipolar disorder should be described in future studies. However, symptomatology, neuropsychological assessments and imaging currently underline similarities with FTLD, particularly FTD. Functional imaging data could support this hypothesis. This specific dementia does not appear to worsen considerably and might be confused with a form of FTLD.

Establishing a differential diagnosis of FTD in the outcome of a bipolar disorder remains difficult and a chance association between these two pathologies cannot be excluded altogether.

Specific neuropsychological tests should be developed for the diagnosis of this disorder, as the current battery of tests fails to discriminate effectively.


The authors thank Frances Sheppard (Centre for Clinical Investigation, University Hospital of Besançon) for her assistance with the English writing.