Bipolar Disorder Among Patients Diagnosed With Frontotemporal Dementia

Behavioral variant frontotemporal dementia (bvFTD) is a progressive neurodegeneration of the frontal and anterior temporal lobes that usually presents in mid-life, with alterations in personality and emotional regulation (1, 2). bvFTD, which is the second most common dementia among patients <65 years old (3), includes progressive decline in executive and other cognitive abilities, as well as behavioral changes (4). The earliest personality and emotional features tend to be apathy or abulia, disinhibition or impulsivity, and loss of sympathy or empathy (2); however, a wide range of behavioral symptoms can occur, including manic-like behavior and symptoms (5, 6). Many, if not most, patients with bvFTD present to psychiatrists who have to distinguish early bvFTD from primary psychiatric disorders (7).

Some patients with bvFTD can appear similar to patients with bipolar disorder on initial presentation (8–10). Patients can manifest with euphoria or emotional outbursts, talkative or pressured speech, racing thoughts, task-oriented behavior, increased energy, and psychomotor activity, as well as irritability, distractibility, and disinhibition (6, 8, 11). Conversely, the manifestations of bipolar disorder include not only manic hypomanic or depressive episodes (12, 13) but also neurocognitive impairments in complex attention, executive functions, and other changes similar to bvFTD (4, 14–16).

This overlap in symptoms of bipolar disorder and bvFTD suggests possible shared origins and mechanisms between these two disorders and raises the question of whether bipolar disorder precedes or leads to bvFTD (17). Indeed, a number of investigators have reported a link between the two disorders (18, 19), yet the exact association is unclear. We hypothesized that beyond shared presenting symptoms, manic or hypomanic episodes can be a prodrome of bvFTD in some patients, and late-onset bipolar disorder can present as a bvFTD mimic or FTD phenocopy in others (20, 21). Accordingly, we reviewed our clinical base of well-diagnosed patients referred for bvFTD who also had a psychiatric diagnosis of bipolar disorder and reviewed the literature on the association between these two disorders.

Methods

Participants

The participants presented for evaluation at a university clinic specializing in bvFTD. All patients had neurobehavioral changes suggestive of bvFTD and underwent a comprehensive evaluation, laboratory assessment, and MRI. In addition, most had fluorodeoxyglucose positron emission tomography (FDG-PET) of the brain. A total of 137 patients met initial clinical criteria for bvFTD (possible, probable, or definite) for the present review (2). To be included in this study, patients diagnosed with bvFTD had to have a knowledgeable informant and an initial history of progressive deterioration of behavior or cognition from their prior normal baseline level. Changes in behavior or cognition included three of the following symptoms: apathy, disinhibition, loss of empathy, repetitive behaviors, dietary changes, and a predominant dysexecutive cognitive profile (2). Patients met criteria for possible bvFTD unless neuroimaging showed frontal or anterior temporal changes supportive of a diagnosis of probable bvFTD or there was definite bvFTD on neuropathology. Among the 137 patients with bvFTD, we further identified all those who had a prior diagnosis of bipolar disorder made by a psychiatrist. This study was approved by the institutional review board of the University of California Los Angeles.

Procedures

For the subgroup of patients with a prior psychiatric diagnosis of bipolar disorder, we abstracted historical evidence of the previous psychiatric symptoms that led to their bipolar disorder diagnosis, focusing on mania, hypomania, or a history of elevated, expansive, or irritable mood (22). We evaluated criteria for manic or hypomanic episodes that included three of the following symptoms: inflated self-esteem or grandiosity, decreased need for sleep, more talkative than usual or feeling pressured to continue talking (“press of speech”), flight of ideas or subjective experience that thoughts are racing, distractibility, increase in goal-directed activity or psychomotor agitation, and excessive involvement in activities that have a high potential for painful consequences. Additionally, we assayed criteria for depression or depressive episodes and family history suspicious for mood disorders.

Results

Among the 137 patients, 14 (10.2%) had a preceding psychiatric history reported to be indicative of a bipolar spectrum disorder. On review, three of these patients did not meet criteria for bipolar disorder but had a history of other psychiatric symptoms that led to their bipolar disorder diagnosis, including suicidal behavior, paranoia, impulsivity, antisocial behavior, and obsessive-compulsive behavior. The remaining 11 patients comprised two subcategories: the prodromal and phenotype subgroups.

Prodromal Group

Eight patients (males, N=6; females, N=2) had an identifiable past history of mania (bipolar I disorder, N=6) or hypomania (bipolar II disorder, N=2) occurring 6–12 years before an established diagnosis of progressive bvFTD. The mean age at onset of mania or hypomania (48.25 years [SD=13.0]) preceded the mean age at onset of bvFTD (57.25 years [SD=11.9]) by an average of 9 years. For one patient, we could not verify any past psychiatric and family histories because of the patient’s immigrant status and loss of family members to war. The other seven patients had a positive family psychiatric history, usually for depression, and one patient had a chromosome 9 open reading frame 72 (C9orf72) mutation. Of the eight patients in the prodromal group, four had prominent bitemporal hypometabolism with or without frontal involvement on FDG-PET, and one other had less prominent temporal lobe hypometabolism. Most of the changes were bilateral at the time of imaging, and no clear laterality difference was discerned. Details of the demographic and clinical characteristics of the prodromal group are presented in Table 1. Patients with autopsies (N=2) exhibited FTD changes with tau positive intranuclear inclusions, involving the frontal and temporal neocortex in one patient and inferior temporal lobes in the other patient.

Phenotype Group

Three additional patients (males, N=2; female, N=1), comprising the phenotype group, did not meet criteria for progressive bvFTD because they did not have frontotemporal changes on FDG-PET and did not manifest clinical worsening of FTD behaviors or neurocognitive measures when followed for 3–7 years. The mean age at onset of manic episodes (58.87 years [SD=3.21]) preceded the mean age at onset of bvFTD (60.0 years [SD=1.0]) by about 1 year.

Additional Observations

Other characteristics of patients with mania or hypomania were obtained. All patients with manic episodes underwent psychiatric hospitalization, and most manic episodes were with psychotic features. All except two patients were taking various antipsychotic medications, and only one patient was taking lithium. Depressive episodes were present in six patients with subsequent progressive bvFTD but not in the three nonprogressive patients, all three of whom had severe, recurrent manic episodes with hospitalization. Only one patient approached being a rapid cycler (patient 8). Another patient was reported by relatives to be cyclothymic during his college years (patient 4). This same patient had his initial and persistent manic episode precipitated by antidepressant medication, but there was no other evidence of substance-abuse-induced mania or drug-induced mania. Otherwise, all manic and hypomanic episodes were of relatively late onset, with the possible exception of those for patient 4, as well as for one patient whose psychiatric history could not be verified (patient 6).

Discussion

We found a heterogeneous relationship of bipolar disorder with bvFTD. The symptoms of mania or hypomania and bvFTD have been shown to overlap, and there are reports of manic behavior as an initial manifestation of bvFTD (7, 9–11, 23–25). Given this overlap of symptoms, we examined patients in a bvFTD cohort and found that 10% carried a prior diagnosis of bipolar disorder, far greater than the proposed worldwide prevalence for bipolar spectrum disorders of 2.4% (26). On further analysis of their psychiatric histories, only 11 (8%) patients met criteria for bipolar disorder; among these, three had nonprogressive bvFTD phenocopies (20, 21). The results suggest that late-onset bipolar disorder can be a prodrome of bvFTD and also lead to cognitive or behavioral changes that mimic bvFTD.

Investigators have established manic or manic-like symptoms as potential presenting features of bvFTD (6, 9, 11, 27). There are reports of manic behaviors specifically with bilateral (right >left) temporal lobe degeneration in bvFTD, even with relative sparing of the frontal lobes (6, 28–30). Other reports have described secondary mania from strokes and neurological lesions involving the right inferolateral frontal and basotemporal regions (29, 31, 32). Hence, manic symptoms and bvFTD may share a common neuroanatomical dysfunction, pointing to a potential link between bipolar disorder and this dementia (33, 34).

Bipolar disorder may be a prodrome to bvFTD. Among the major psychiatric syndromes, bipolar disorder is most frequently associated with dementia (35); among major dementias, bvFTD is most commonly preceded by psychiatric symptoms (7). Although a recent meta-analysis of longitudinal (1–9 years) studies did not find significant differences in rates of cognitive decline between patients with bipolar disorder and control subjects (36), many other studies from around the world have reported an increased risk of dementia among cohorts of patients with bipolar disorder, particularly those with bipolar I disorder (37, 38), compared with control subjects, with hazard ratios ranging from 2.36 to 7.52 (37, 39–44). The dementia syndrome in these patients may be more similar to bvFTD than other dementias, such as Alzheimer’s disease (45). In one large epidemiological study of more than 1,000 patients with FTD, the investigators reported a background of mood disorders in 19% of patients, many with bipolar disorder (46) (Table 2); compared with other major types of dementia, patients with bvFTD are more than twice as likely to have a diagnosis of bipolar disorder (7). Although patients with lifelong bipolar disorder may go on to develop FTD (18), it is late-onset bipolar disorder that carries the most significant risk for developing bvFTD (39, 43). Most of our patients, as well as other study subjects from the literature, exhibited a prodromal relationship of several years between episodes of mania and the development of clinical and FDG-PET evidence of bvFTD (8). Our results, along with those from other studies, suggest that relatively late-onset bipolar disorder can precede typical progressive bvFTD.

Other reports of late-onset bipolar disorder have described the emergence of a nonprogressive bvFTD mimic (10, 47, 48). This is consistent with the bvFTD phenocopy syndrome, which lacks the neuropathology of FTD in most autopsied cases (20, 21). This is also consistent with findings for three of the patients in our study, as well as patients from other studies, with bipolar I disorder who developed bvFTD-like symptoms without progression or neuroimaging changes over subsequent years of follow-up (47). For example, Vorspan et al. (25) described a 54-year-old patient with bipolar disorder with relapsing-remitting agitated manic episodes and symptoms of bvFTD who was followed up for 7 years without progression of symptoms. Severe or numerous manic and depressive episodes may result in frontally predominant cognitive and behavioral injury and an eventual nonprogressive dementia mimic of bvFTD (10, 30, 35, 41, 42, 49, 50). Episodes of mania and depression that are close in time to the onset of the dementia, which was the case for the three nonprogressive patients in our study, may leave a phenotype of bvFTD in their wake, even when euthymic. Patients with bipolar disorder may have relatively stable cognitive changes in executive functions (working memory and cognitive control or cognitive flexibility and response inhibition) and verbal memory (4, 51–53), and most routine executive function tests may fail to distinguish the cognitive changes of bipolar disorder from bvFTD (54, 55). The etiology of cognitive changes from manic or depressive episodes could be due to accelerated aging in cognitive control processes (40, 56–58), prefrontal structural or functional abnormalities (59–62), injury from neuroinflammatory activity or oxidative stress (63, 64), frontal lobe and posterior white matter abnormalities (65–67), or alterations in connectivity, networks, or interoceptive circuits (64, 68).

A third but apparently rare association of bipolar disorder with bvFTD is through a common genetic predisposition. Among patients with familial bvFTD, psychiatric symptoms are the initial manifestations for up to one-half of individuals with the C9orf72 mutation, a nucleotide repeat disorder, and one-quarter of those with a progranulin (GRN) mutation, with decreased progranulin levels (1, 69). Some investigators have found a history of bipolar disorder and subsequent bvFTD with C9orf72 (70, 71), similar to one of our patients, and with GRN mutations (19, 72). Additionally, some investigators have considered an expanding number of C9orf72 nucleotide repeats in familial bipolar disorder and psychosis leading to bvFTD (48, 70, 72, 73), but others have failed to identify a link to bipolar disorder among relatives of patients with the C9orf72 protein or a clear relationship with the number of repeats (74, 75). To our knowledge, there is no reported association with bipolar disorder in microtubular-associated protein tau mutations (76), the third common genetic form of bvFTD. For C9orf72 and GRN mutations, the literature raises the hypothesis that the close link between bipolar disorder and bvFTD may have a genetic predisposition but in relation to other factors, either environmental or additional genetic factors, such as variants of the glycogen synthase kinase 3β gene (77).

There are important limitations to this study. First, it is a retrospective review, albeit of carefully evaluated patients referred to a program specializing in bvFTD. Nevertheless, our study is clinical, observational, and descriptive. Hence, there are a number of potential confounding factors to consider, such as selection bias on referral to a specialty clinic. Second, as with any retrospective study in which information is historical, our study is subject to questions of reliability and completeness, which may be particularly a factor for psychiatric and behavioral symptoms. Wherever possible, we compared the secondary reports with hospitalization records available for many of the manic episodes. Another factor is the disproportionate number of males evident in our total sample and present across the different subgroups. Finally, the imaging results of necessity included imaging reports from other institutions and facilities where direct analysis of brain scans was not available.

Conclusions

The relationship of bipolar disorder with bvFTD is heterogeneous. Manic or hypomanic episodes may indicate a prodrome heralding bvFTD, possibly with early right >left temporal involvement. Late-onset bipolar disorder may also present as a bvFTD mimic or bvFTD phenocopy that is often mild, nonprogressive, and without further cognitive decline or definitive neuroimaging or neuropathological changes. Much less frequently, there appears to be an association of bipolar disorder with a shared genetic predisposition from mutations in the C9orf72, GRN, or other genes. These are preliminary conclusions; however, they can lead to more research on the concepts of a late-onset bipolar disorder prodrome of bvFTD and late-onset bipolar disorder leaving a nonprogressive bvFTD phenotypic profile.