It has been known for many years that frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS) share some common genetic basis. Indeed, many families demonstrate autosomal-dominant inheritance patterns of FTD, ALS, and/or FTD-ALS complex. The causative mutation was identified simultaneously by two groups in 2011 as being a hexanucleotide repeat on chromosome 9, the
C9orf72 mutation.
1,2 This major discovery has generated substantial interest in the scientific community, with over 700 publications on this topic since the initial reports.
The prevalence of FTD is 15–22 per 100,000, with around 20% of cases secondary to autosomal dominant mutations.
3 The
C9orf72 mutation was found to probably be the most common genetic form of FTD, being present in 7%−12% of subjects.
4 Of significant interest, pathogenic repeat expansion has been identified in 2%−5% of apparently sporadic FTD cases,
5 i.e., in patients without family history suggestive of dominant inheritance.
The most common clinical presentations of
C9orf72 mutations are behavioral variant FTD (bvFTD), ALS, or a combination of both (FTD-ALS).
4 The bvFTD syndrome presents with variable combinations of behavioral symptoms, including apathy, disinhibition, loss of empathy, stereotyped/repetitive behavior, and hyperorality.
6 Cognitive impairment in bvFTD due to
C9orf72 mutations includes the typical deficits in social cognition and executive function but can also involve memory disturbances and parietal lobe deficits.
4,7 ALS is often associated with early behavioral and cognitive changes, even when not meeting criteria for FTD-ALS.
8 In ALS due to
C9orf72 mutations, the phenotype is clinically indistinguishable from idiopathic ALS at the motor level, but with probably higher rates of behavioral and cognitive disturbances.
4,9 The language variants of FTD spectrum diseases, referred to as primary progressive aphasias (PPA), can also be associated with behavioral and psychiatric disturbances, although to a lesser degree than bvFTD.
10 PPAs are a more uncommon presentation of
C9orf72, with most reported cases being nonfluent or semantic variants.
7,11Of particular interest to the neuropsychiatric community, phenotypic characterization of clinical cohorts of patients with
C9orf72 have hinted toward an unusually high frequency of late-onset psychotic disorders as the initial presentation, sometimes many years prior to the onset of more typical FTD or ALS symptoms.
7,12 Subsequent reports have described a wide variety of presentations, such as mania and bipolar disorder, major depressive episodes with catatonic features, and obsessive-compulsive disorder. These studies raise concrete and important questions for clinical practice across multiple settings, particularly in adult psychiatry: 1) How common are these presentations? 2) Could some cases of late-onset psychosis be explained by
C9orf72 mutation? 3) Who should get genetic testing? Although there are still no disease-modifying treatments for FTD or ALS secondary to
C9orf72 mutations, identifying the mutation in patients has major implications for prognosis and family counseling.
As an attempt to resolve these important clinical issues, the Committee on Research of the American Neuropsychiatric Association (ANPA CoR) has decided to undertake a comprehensive review of predominant psychiatric presentations of C9orf72 mutations to shed light on this topic and suggest clinical recommendations to clinicians.
Basic Genetics
The
GGGGCC (also called
G4C2) repeat expansion is located in chromosome 9 open reading frame 72 (
C9orf72) gene on the short arm of the chromosome. Although the exact function remains uncertain, studies have identified possible protein isoforms that are thought to contribute to the regulation of endosomal trafficking.
13 The
C9orf72 mutation has three transcription variants, depending on whether the hexanucleotide repeat expansion is located in the promoter region (variant 1) or in intron 1 (variants 2 and 3).
1 In healthy controls, the normal hexanucleotide repeat size is usually up to 10 and on rare occasions up to 30 repeats.
1, 14–16 Repeat sizes between 20 and several hundred have been identified in both healthy controls and patients.
4,15,17–22 Repeat lengths of more than several hundred are very likely to be pathogenic for ALS or FTD.
1,4,23 The exact pathogenic number of repetition is difficult to determine for many reasons. The hexanucleotide repeat expansion in somatic tissues and blood is usually unstable (that is, within an individual, various organ tissues may exhibit different repeat lengths). Therefore, intermediate repeat lengths in blood or somatic tissue might be present in an individual with pathogenic repeat length expansion in the brain.
14 In addition, there is no direct connection between the number of repeats and the clinical phenotype severity, and data are not consistent on whether there is an association with the age of onset of the disease.
15 Furthermore, the mutation has been identified in up to 0.15% of the healthy population in the United Kingdom.
15The familial cases demonstrate autosomal dominant pattern of inheritance. Age of onset, even within a family, varies widely. Some kindred appear to exhibit an anticipation-like phenomenon.
24 However, studying anticipation in affected families has been challenging, given that the repeat expansion is unstable and variable in size across the cells. Cases of incomplete penetrance have also been reported.
25FTD cases secondary to
C9orf72 mutations are associated with TAR DNA binding proteins (TDP-43) pathology types A and B, in addition to specific p62/SQSTM1-positive neuronal cytoplasmic inclusions containing dipeptide repeat (DPR) proteins from the translation of the abnormal gene.
4,26 The precise pathogenic mechanism of
C9orf72 hexanucleotide repeat expansion remains uncertain. Potential explanations include loss of function, gain of function, or a combination of both.
1.
Loss-of-function mechanisms: The
C9orf72 protein has been shown to regulate endosomal trafficking and autophagia in neuronal cell lines, primary cortical neurons, and human spinal cord motor neurons.
13 C9orf72 is also homologous to DENN proteins (differentially expressed in normal and neoplasia proteins), which regulate membrane trafficking function by activating RAB GTPases.
20,27 The observation of lower levels of normal
C9orf72 transcript variants in the brain, spinal cord, and blood in patients with
C9orf72 expansions compared with healthy controls or patients without the
C9orf72 expansions underlies the hypothesized loss-of-function mechanism.
1,23,28–302.
Gain-of-function mechanisms: The toxic products of
C9orf72 hexanucleotide repeat expansion’s transcription (toxic RNA repeat structures) and translation (toxic DPR proteins) are suggested as the main players in the gain-of-function mechanisms. The hexanucleotide repeat expansion’s toxic RNAs have been found in the regions affected by the disease
31–34 and are believed to bind and sequester RNA-binding proteins.
28,35–42 However, the role of RNA-binding protein sequestration in disease pathogenesis is yet to be illuminated.
The unconventional repeat-associated non-ATG translation of the
C9orf72 hexanucleotide repeat expansion results in the formation of DPR proteins.
43 DPRs have been shown to be toxic in a range of model organisms
33,39,44 and cell lines.
34 DPRs were also localized in the affected brain regions of FTD-ALS patients.
45,46 The dysregulation of nucleocytoplasmic transport and the consequent neurodegeneration are thought to be the main pathogenic mechanisms of PDR proteins’ toxicity.
13,27In summary, past studies suggest that C9orf72 expansion results in both loss- and gain-of-function mechanisms, but further studies are warranted to elucidate the relative contribution of these mechanisms and the molecular pathways that lead to neurodegeneration.
Methods
The goal of this project was to review the literature related to psychiatric presentations of C9orf72 and the frequency of mutations in primary psychiatric disorders, with the ultimate objective being to provide concrete guidance to clinicians on how to approach this clinical problem. The main focus of the review was to identify key clinical features to improve diagnostic recognition in clinical practice, including when genetic testing should be obtained or not. Management and treatment aspects were reviewed in the identified articles, but this was a secondary objective given the very limited literature on the topic.
A PubMed search was conducted including all English language articles up to July 2016. A keyword strategy was used including the following search terms:
“C9orf72” AND
•
“Psychosis” OR “Schizophrenia”
•
“Bipolar Disorder” OR “Mania”
•
“Anxiety” OR “Obsessive-Compulsive Disorder (OCD)”
A complementary Google Scholar search with the terms “C9orf72” AND “Psychiatry” was performed to identify any missing articles (one additional article was identified). All abstracts obtained by the keyword search were reviewed by the primary investigator (S.D.) to exclude articles that were not relevant to one of the two objectives and to delete duplications. A total of 43 articles were identified with the following breakdown: schizophrenia/psychosis (N=25), bipolar disorder/mania (N=4), depression (N=2), anxiety/OCD (N=2), catatonia (N=1), and psychiatry (N=9 articles not identified by single-symptom search terms). Articles were divided into two categories according to the objectives: 1) articles related to the psychiatric presentations of C9orf72 mutations and 2) articles related to the prevalence of C9orf72 mutations in cohorts of patients with primary psychiatric disorders. All articles were systematically reviewed by the coauthors to produce this report. Results were initially presented at the 27th Annual Meeting of the American Neuropsychiatric Association in San Diego (March 2016). The level of evidence varied depending on subtopics. There were good-quality case-control and epidemiological studies for psychosis (section 4.1) and for the prevalence of C9orf72 mutations in primary psychiatric disorders (section 5). The evidence was of lower quality for other psychiatric symptoms, consisting either of small case series or single case reports.
Clinical Presentations
Psychosis
The frequency of psychotic symptoms (delusions and/or hallucinations) in sporadic FTD is relatively low, with a recent estimate of a 10% prevalence.
47 Reports of increased frequency of psychotic symptoms in the early stages of FTD secondary to
C9orf72 mutations emerged from multiple sources in 2012,
7,12 one year after the discovery of the mutation. A total of 10 studies documenting the prevalence of psychosis in cohorts of patients with FTD-related
C9orf72 mutations have been identified (
Table 1). A few case reports were also identified.
48–54 These studies have consistently reported a marked increased frequency of psychosis at the onset of FTD or preceding more classical FTD symptoms in patients with
C9orf72 mutations.
7,12,55–61 These studies showed a prevalence of psychotic symptoms in the range of 21% to 56%. Although published studies strongly support this increased rate of psychosis, there are unexplained geographic discrepancies in prevalence, including much lower rates of psychiatric disturbances recently reported by a German consortium at the last International Conference on Frontotemporal Dementia.
62Other than the overall increased prevalence of psychosis, a few additional key features emerged from these studies. First psychotic symptoms can include both delusions and hallucinations in all sensory modalities, together or separated. Second, there is a high frequency of somatic delusions (e.g., foreign object in body, pregnancy) in addition to unexplainable somatic symptoms or preoccupations (e.g., pain) that do not reach delusional intensity.
7,48,51,59,61 Delusion subtypes that have been reported also include persecution, jealousy, grandiose, and mystical/religious.
51 We are not aware of any study that has directly compared the nature of delusions between primary psychotic disorders or bipolar disorder versus FTD due to
C9orf72, but other than a possibly higher prevalence of somatic delusions found in some
7 but not all
51 studies, the content of delusions appears to be relatively similar between these disorders. Third, psychotic symptoms often precede the appearance of more typical FTD features by 1–5 years.
51,56 Finally, although this is more anecdotal, reports consistently describe poor response and adverse reactions to antipsychotic medications.
59,63These studies on psychotic symptoms also highlighted important features of the atypical aspects of FTD secondary to
C9orf72 mutations. More than a third of cases did not have a family history of FTD-ALS; therefore, in clinical practice these cases would be considered as sporadic FTD.
7,57 In addition, multiple subjects did not show significant atrophy on MRI in the initial assessments.
57,56 When atrophy is present, the pattern is not restricted to frontotemporal areas, with frequent involvement of parietal regions, cerebellum, and thalamus.
64 In the absence of those typical neuroimaging findings, one study showed that 62% did not meet research criteria for probable bvFTD.
6,55 Solje et al.
60 similarly showed that among 36 patients with some bvFTD symptoms due to a
C9orf72 mutation, 19% did not meet the full research criteria for possible bvFTD and 36% did not meet criteria for probable bvFTD. In addition, 17.6% of subjects had a normal FDG-PET or SPECT,
60 which are thought to be more sensitive for bvFTD diagnosis than MRI. The lower sensitivity of diagnostic criteria in this population could also be in part related to decreased frequency of apathy (the most common symptom in sporadic bvFTD)
65 and higher emotional warmth at the onset.
56,61 In addition, patients with
C9orf72 have a higher rate of family psychiatric history,
55 a factor that has been shown to bias clinicians toward missing FTD diagnoses.
66 Factoring all these aspects, clinicians need to have a high index of suspicion in patients with late-onset psychotic disorders in order to identify potential
C9orf72 cases, because they cannot rely solely on the absence of the usual bvFTD clinical features described in the diagnostic criteria to exclude this possibility.
Mania and Bipolar Disorder
In 2013, Floris et al. published a case report of a patient who presented with bipolar I disorder at age 42 with both episodes of mania and hypomania.
67 He had a family history of bipolar disorder in one uncle, and his symptoms responded well to lithium. At age 64 he presented with more classical bvFTD symptoms (repetitive behavior, disinhibition, attentional and executive function deficits), delusions, and mild Parkinsonism. Although many of these symptoms could be explained by a manic episode, MRI demonstrated frontal atrophy, and a genetic test showed more than 70 repeats of the
C9orf72 region, confirming a diagnosis of definite bvFTD. While this could have been coincidental, other reports have described relatively similar situations. One case in the Galimberti et al.
57 study presented with agitation/mania. In addition, cases of
C9orf72 have been identified in cohorts of typical bipolar disorder patients (see the section on the prevalence of
C9orf72 mutations in primary psychiatric disorders).
24,68 This includes a 35-year-old patient with typical lithium-responsive bipolar I disorder whose father suffered from bipolar I disorder at age 62, followed by bvFTD at 66.
24 A
C9orf72 mutation was also found in a 37-year-old male with major depressive, mixed and manic episodes who also responded to lithium. Although the numbers remain small, these reports suggest that late-onset bipolar I disorder (particularly manic episodes) can rarely be the initial clinical manifestation of
C9orf72 mutations and that these episodes tend to be lithium responsive.
Major Depressive Disorder (MDD)
MDD is the most common psychiatric diagnosis preceding the recognition of bvFTD
66; however, there are few reports related to
C9orf72 mutations. One study showed a 30% rate of mood/affective symptoms in mutation cases, but with little information as to what was included in this classification.
60 One case report described a man who suffered a traumatic brain injury at age 40 with permanent cognitive and judgment deficits.
69 At age 66 he developed depressive symptoms and made a suicide attempt in the context of grief. The course was complicated by retarded catatonia that responded to antidepressants, aripiprazole augmentation, lorazepam, and ECT. This was followed by a relapse 6 months later with added impulsivity, restlessness, repetitive behaviors, and REM sleep behavior disorder with progressive deterioration. A brain MRI showed frontal atrophy, and a
C9orf72 expansion was confirmed. His mother had suffered from an unidentified progressive neuropsychiatric syndrome over a 20-year period. One patient presented at age 46 with an agitated depression that evolved into bvFTD with significant psychotic symptoms and intolerance to antipsychotics.
63 At autopsy she had extensive DPR aggregates but minimal TDP-43. Of note, her son, who suffered from intellectual disability, died at age 26 of unrelated causes but was found to be a carrier and have DPR aggregates without TDP-43 at autopsy.
Using a postmortem approach, Bieniek et al.
70 tested for the mutation in 31 brains without macroscopic atrophy in which the clinical diagnosis had been dementia (N=3), depression (N=6), or both, which they referred to as “depressive pseudodementia” (N=22). Of those 22 “pseudodementia” cases, two were found to be positive for the
C9orf72 mutation and had C9RANT neuronal inclusions, with the most severe deposits in the cerebellum. In one patient, major depression started at age 52, and over time the patient developed unexplained toe numbness (at age 57), drug-induced parkinsonism, and catatonia (at age 65) and died at age 66. His brain MRI was normal at age 61 but showed mild atrophy at 65. His father suffered from MDD and Parkinsonism. The other case had MDD at age 66, followed by attentional and short-term recall deficits that were clinically thought to be secondary to Alzheimer’s disease.
Anxiety and OCD
Very few studies were found on the topic of anxiety. One case series of 19 patients with
C9orf72 mutations reported a 33% rate of anxiety as part of the initial presentation, but without specifying the nature of this anxiety.
11 We also identified two case reports of OCD as the initial presentation of bvFTD due to
C9orf72 mutations,
71,72 which is also a known presentation of sporadic bvFTD.
73Catatonia
Catatonia has been previously reported in patients suffering from bvFTD.
74 Two cases were found in patients with
C9orf72 mutations, both described above.
69,70Suicide
Patients suffering from bvFTD are usually unaware of the severity of their deficits; therefore, suicide attempts are rare. We identified one case report of a patient with bvFTD due to
C9orf72 with minimal cognitive deficits who was first brought to clinical attention after a serious suicide attempt.
75 A patient described in the MDD section also attempted suicide as part of a depressive syndrome.
69 Suicidal ideations were also reported in multiple members of a Swedish family.
59 However, there are insufficient data to conclude whether the frequency of suicide attempt is higher in patients with the mutation than in sporadic cases. Of note, a recent study did not find any
C9orf72 mutation in a postmortem cohort of 109 victims of suicide (without FTD) in Finland.
76Prevalence of C9ORF72 Mutations in Primary Psychiatric Disorders
In the context of the relatively high prevalence of psychiatric symptoms at the onset of C9orf72 mutations and the frequently atypical nature of those patients, various groups have investigated whether the mutation could be the cause of some cases of patients with primary psychiatric disorders. We identified studies that screened large clinical populations for the mutation in both schizophrenia/schizoaffective disorder and bipolar disorder but not in other major psychiatric disorders.
We identified six studies in schizophrenia and schizoaffective disorders spanning North America, Europe, and Asia (
Table 2).
77–82 Four of those studies did not identify a single case in a total of 1410 subjects. One study from Germany and Italy identified two cases out of 297 patients.
82 One patient had paranoid and grandiose ideations with a thought disorder starting at age 33. There was a family history of a sister with dementia and Parkinsonism and unspecified dementia in the mother. The other patient had onset of paranoid ideations and mood symptoms at age 44 (which is atypically late for schizophrenia), with a family history of schizophrenia in his mother. A recent American study found three positive cases out of 697, including one patient with treatment-refractory childhood-onset schizophrenia.
86 Pulling together all six studies, the prevalence of
C9orf72 mutation is estimated at below 0.1% in patients with typical schizophrenia or schizoaffective disorders.
A total of four American and European studies have investigated the frequency of the mutation in large cohorts of patients with bipolar disorder (
Table 2).
24,68,79,83 Two cases were identified out of 862 patients. One of those patients had a family history of a father with late- onset bipolar disorder progressing to FTD (described above).
68 This amounts to a prevalence of approximately 0.1%, which is similar to the rate of the mutation found in healthy controls in one study.
15Mechanistic Hypotheses
The exact function of the
C9orf72 gene remains to be elucidated; therefore, the mechanism by which the mutation increases the prevalence of psychiatric prodromes in FTD is unknown. However, there are a few interesting hypotheses. First, the atrophy pattern is more atypical in patients with
C9orf72 mutation compared with sporadic cases, with more severe involvement of the cerebellum and the thalamus, two structures that could contribute to the different clinical presentations.
4,64 In addition, preliminary results from the GENFI study have suggested that volume loss in the insula, parietal lobe, thalamus, and cerebellum could start more than 10 years prior to bvFTD clinical symptoms.
84 The cerebellum has a key role in modulating thoughts, affect, and behavior
85; therefore, involvement of this structure could contribute to psychiatric phenotypes. Of interest, one study found a higher expression of the
C9orf72 gene in the cerebellum of patients suffering from schizophrenia.
86 One study reported a correlation between the degree of precuneus atrophy and the presence of delusions among patients with
C9orf72 mutation.
51 A small study has also shown potential differences in network connectivity in patients with
C9orf72 compared with sporadic FTD. Both groups showed an increase in default mode network connectivity, but of lesser magnitude in
C9orf72 compared with sporadic cases,
87 which could theoretically have a link to the more preserved emotional warmth observed in the early stages.
61 There is also one study that reported a link between a 10-pair deletion prior to the
C9orf72 expansion and the expression of psychotic symptoms, but this finding has not yet been replicated.
88While this is not a mechanistic explanation, Downey et al.
89 have run a series of experiments of body schema perception tasks in patients with the mutation compared with sporadic FTD and controls.
95 The idea to conduct these tasks came from the high frequency of somatic complaints and even somatic delusions in these patients. Results demonstrated altered body schema processing in patients with the mutation in various tasks such as two-point discrimination, body part illusions, and self- versus non-self-differentiation.
Clinical Implications
This review identified a few key findings as they pertain to the psychiatric presentations of C9orf72 mutations. First, among the psychiatric prodromes, psychotic symptoms seem to be the most common, including various combinations of delusions or overinvested ideas and multimodal hallucinations. There is a high prevalence of unexplained somatic preoccupations, sometimes of delusional intensity. Multiple subtypes of delusions have also been reported, including persecutory, jealousy, religious/mystical, and grandiose. These symptoms are usually not responsive to antipsychotics in this context. Albeit less frequent, it appears that late-onset bipolar disorder with manic episodes can also be part of the prodromal phase of C9orf72 genetic FTD, and reported cases point to good therapeutic response to lithium. Other forms of mood disturbances—including recurrent depressive episodes with or without catatonia, dysphoria with suicidal thoughts or attempts, and depression-related cognitive disturbances (“pseudodementia”)—are also possible.
There are various factors that further increase the challenge of correctly identifying cases of C9orf72 mutations in patients with late-onset psychiatric presentations:
1.
Psychiatric symptoms can precede typical bvFTD features by up to 4–5 years.
2.
Progression of symptoms can be slow over many years.
55,903.
Neuroimaging can be normal in the initial phase of the disease.
55,60,914.
Many subjects do not have a positive family history (either no cases or only cases of apparent primary psychiatric disorders).
Regarding point 1, Block et al.
72 proposed a model in which in some cases the psychiatric disturbance could be a prodrome of FTD during the phase of functional synaptic changes and neurotransmitter instability (i.e., psychiatric features precede FTD), whereas in others it is the result of synapse involution and cell death, thus explaining why psychiatric features start at the beginning of FTD in parallel to more typical dementia features.
It should also be mentioned that even when cognitive and neurological signs are present, they are often not restricted to the prototypical description of bvFTD or PPA as per current diagnostic criteria.
4 Indeed, patients can have early deficits in learning and recall
7 or perceptual-motor parietal dysfunction, which is incompatible with current
DSM-5 frontotemporal neurocognitive disorder criteria D requesting the relative sparing of learning and memory and perceptual-motor function. There are also cases with Parkinsonism (usually a late feature), Huntington disease-like phenotypes, and cerebellar dysfunction.
4,15,55,92 These clinical features parallel the atypical aspects of atrophy patterns that involve more parietal, cerebellar, and subcortical areas than do sporadic bvFTD. Integrating these findings together, it is unsurprising that the sensitivity of clinical bvFTD criteria is low in this population.
55,62Of note, increased rates of psychotic symptoms in FTD are not restricted to
C9orf72 mutations. There is a reported association with the rare frontotemporal lobar degeneration (FTLD) FUS pathology.
47 GRN mutations are also to an extent associated with increased psychiatric symptoms, but a recent study by Snowden et al.
61 confirmed that the rate of psychotic symptoms, particularly somatic delusions, is higher in
C9orf72 than in
GRN mutations.
The other clear finding from this review is that
C9orf72 mutations are very rare in patients with typical
DSM-5 schizophrenia or schizoaffective disorder (<0.1%) and bipolar disorder (approximately 0.1%). It is therefore not advisable to test patients at random. The prevalence of
C9orf72 mutation in cases of late-onset psychotic disorder or mania remains unknown, but on the basis of the epidemiology of late-onset psychotic disorders (incidence approximately 12.6 out of 100,000),
93 C9orf72 mutations would explain only a minority of cases. However, restricting testing to patients meeting diagnostic criteria for bvFTD will clearly miss cases over periods of many years until the dementia becomes more evident. Although there are currently no curative treatments, establishing a diagnosis of neurodegenerative disease is important for families to plan personal and financial affairs prior to severe cognitive decline. It is also important to avoid potentially deleterious interventions such as high-dose antipsychotics in psychosis due to
C9orf72 mutations. It will be even more crucial to identify these patients when potential therapeutic interventions come along, while not causing prohibitive costs by testing patients at large.
The first step toward improved recognition of cases is for clinicians to be aware of this mutation and the variety of clinical presentations, including isolated psychiatric syndromes.
73 In all patients with late-onset mania, psychosis, and mood disorder with catatonia or cognitive deficits, a detailed family history should be obtained, including screening for early-onset dementia, FTD, ALS, Parkinsonism, and unexplained neuropsychiatric syndromes. Patients should have at minimum a screening cognitive assessment (e.g., Montreal Cognitive Assessment) and elemental neurological examination. Clinical symptoms of dementia, including symptoms of bvFTD and PPA, should be elicited from patients and relatives. Neuroimaging should include at minimum a cerebral CT scan, but ideally a brain MRI should be obtained to assess for early signs or cortical and subcortical atrophy. If there is a suspicion of bvFTD or other early-onset dementia, a consultation in behavioral neurology or neuropsychiatry should be obtained.
In patients with late-onset psychiatric presentations but without further clinical evidence of FTD, the question of who should be tested for
C9orf72 mutation remains open. In current practice, clinicians often wait until the appearance of clinical evidence of dementia, but the literature clearly shows that FTD features can be delayed by 4–5 years. There are currently no agreed-upon guidelines on who should be tested for the mutation in patients with late-onset psychiatric disturbances. Among patients with clinical diagnoses of frontotemporal lobar spectrum disorders, Wood et al.
94 have identified patients with high risk of genetic mutations as having either ≥1 first-degree relative (FDR) with FTLD/ALS or one second-degree relative with FTLD/ALS and one FDR with other types of dementia or Parkinson’s disease or ≥2 FDRs with early-onset (≤65 years old) dementia or Parkinson’s. However, restricting testing to only high-risk settings will miss cases, as many patients with
C9orf72 mutations have family history only of apparent primary psychiatric disorders.
Practices will vary according to local availability of genetic consultation and tests, but we propose the following division in terms of C9orf72 testing for patients presenting with late-onset (after 40 years of age) psychosis, bipolar disorder, or MDD with catatonia or cognitive deficits:
Testing should be obtained if:
1.
There is a history of first-degree relative with confirmed C9orf72 mutation or with FTD spectrum disease or ALS;
2.
They meet criteria for the high-risk category of Wood et al.
94;
Testing should be considered and discussed with the patient if:
3.
There is a family history of late-onset bipolar or psychotic disorder or other unspecified progressive neuropsychiatric disturbance;
4.
There is progressive deterioration with cognitive decline or emerging features of bvFTD, Parkinsonism, or both.
Based on the current literature, C9orf72 testing should not be obtained in patients with DSM-5 diagnosis of schizophrenia, schizoaffective disorder, or bipolar disorder with onset prior to the age of 40 unless there is a proven genetic mutation in a first-degree relative.
We recommend a genetic consultation and counseling prior to proceeding with genetic testing. Testing by polymerase chain reaction (PCR) and Southern blot is the gold standard, as the standard repeat-primed PCR cannot distinguish repeats larger than 30–50, and blinded tests in different laboratories have showed inaccuracies.
95Conclusions
Clinicians need to be aware that late-onset psychosis and bipolar disorder can be the initial prodromal phase of FTD due to C9orf72 mutations. The presentations are heterogeneous and can be difficult to identify, given their homology to primary psychiatric disorders, frequently normal imaging, and delayed appearance of more typical FTD features. Physicians need to have a high index of suspicion and elicit detailed family history of neurodegenerative diseases in those patients. This review also highlights the numerous gaps in our current knowledge (e.g., specific pathogenic mechanism, link between the mutation and psychiatric symptoms, geographic discrepancies in prevalence of psychosis, management), and much research remains to be done. Despite these limitations, we encourage psychiatrists, neurologists, geriatricians, and primary care physicians to familiarize themselves with this topic, as they have a key role in the identification of cases in inpatient and outpatient settings.