Abstract
While psychiatric misdiagnosis is well-known in behavioral variant frontotemporal dementia (bvFTD), a systematic evaluation of standardized criteria for psychiatric disorders in bvFTD is still missing. Our aim was to define frequency and character of DSM-IV psychiatric disorders among patients with probable and definite bvFTD compared to possible bvFTD, other neurodegenerative diseases, and psychiatric diagnoses, using MINI-International Neuropsychiatric Interview. We additionally compared psychiatric prodromes between these groups. Subjects were participants of the late-onset frontal lobe (LOF) study, a longitudinal multicenter study. In each patient, after baseline diagnostic procedure, a neurologist and geriatric psychiatrist made a joint clinical diagnosis. Independently, a structured diagnostic interview according to DSM-IV and ICD-10 criteria (MINI-Plus) was performed by a trained professional blinded to clinical diagnosis. Out of 91 patients, 23 with probable and definite bvFTD, 3 with possible bvFTD, 25 with a non bvFTD neurodegenerative disease, and 40 with a clinical psychiatric diagnosis were included. Overall frequency of formal current and past psychiatric disorders in probable and definite bvFTD (21.7% current, 8.7% past) did not differ from other neurodegenerative diseases (12.0% current, 16.0% past) or possible bvFTD (66.7% current, 66.7% past), but was less than in patients with a clinical psychiatric diagnosis (57.5% current, 62.5% past; p < 0.01). In probable and definite bvFTD unipolar mood disorders were most common. Formally diagnosed psychiatric disorders are not overrepresented in probable bvFTD, suggesting that psychiatric misdiagnosis in bvFTD can be reduced by strictly applying diagnostic criteria. In suspected bvFTD close collaboration between neurologists and psychiatrists will advance diagnostics and subsequent treatment.
Keywords
INTRODUCTION
The behavioral variant of frontotemporal dementia (bvFTD) is the most prevalent form of FTD and is associated with progressive degeneration of the frontal lobes, anterior temporal lobes, or both [1, 2]. Alterations in social cognition represent the earliest and core symptoms of bvFTD resulting in emotional disengagement and socially inappropriate responses or activities. As is apparent in revised consortium criteria, neuropsychiatric symptoms including apathy, stereotyped and impulsive behavior overshadow cognitive disabilities [1, 3–5]. Consequently, both other neurodegenerative diseases and psychiatric disorders are crucial in the challenging differential diagnosis.
The distinction between bvFTD and Alzheimer’s disease (AD) has become easier by the use of biomarkers that are able to identify underlying AD pathology, such as the amyloid-β (Aβ) and tau [6, 7]. Differentiating bvFTD from psychiatric disorders, however, is still difficult, particularly since biomarkers for bvFTD are less robust [8]. However, the current clinical criteria require that “if behavioral disturbance is better accounted for by a psychiatric diagnosis, a diagnosis of bvFTD has to be excluded” [5]. Previous studies indicate that as a result of symptomatic overlap between bvFTD and psychiatric disorders, bvFTD patients are clinically often mistaken for psychiatric patients and vice versa [9–11]. One study found that in comparison to other neurodegenerative diseases, bvFTD patients receive a prior psychiatric diagnosis significantly more often (50.7%) than patients with AD (23.1%), semantic dementia (24.4%), or progressive nonfluent aphasia (11.8%) [12]. Although psychiatric misdiagnosis is a well-known phenomenon in bvFTD, a systematic evaluation of standardized criteria for a psychiatric disorder in bvFTD patients is still missing, potentially leading to a clinical under- or overestimation of psychiatric disorders in bvFTD patients. Additionally, though specifically psychotic disorders and bipolar disorders emerged as relatively common psychiatric misdiagnoses in bvFTD, this has never been systematically evaluated through applying the DSMIV and ICD-10 criteria [11–13].
Recent studies showed increased rates of psychiatric symptoms prior to clinical AD, suggesting that especially depressive symptoms appear before a threshold of neurodegeneration is passed [14, 15]. A similar mechanism might be present in bvFTD, with the presence of prominent psychiatric and behavioral symptoms before clinical onset of bvFTD described as mild behavioral impairment [16, 17], but this has not been studied so far.
In this study we aimed at defining the frequency of current psychiatric disorders according to DSMIV and ICD-10 criteria in patients with probable and definite bvFTD in comparison to patients with possible bvFTD, other neurodegenerative diseases, and psychiatric diagnoses, all exhibiting a late-onset frontal lobe syndrome. At baseline we set out to determine the specific psychiatric disorders according to DSMIV and ICD-10 criteria in these groups. Additionally we aimed at studying the frequency and character of past psychiatric disorders according to DSMIV and ICD-10 criteria in these groups. All patients were followed during a period of 2 years and the clinical diagnosis after 2 years was used as the gold standard.
MATERIALS AND METHODS
Patients
Subjects were participants of the ongoing late-onset frontal lobe (LOF) study, a longitudinal multicenter prospective follow-up study aiming to identify (prodromal) bvFTD among a cohort of patients with frontal neuropsychiatric features. An objective of the LOF study is to enforce early diagnosis of bvFTD and to discern the bvFTD prodrome from a broad spectrum of clinically relevant differential diagnosis [18]. All patients were recruited through the memory clinic of the Alzheimer center VUmc Amsterdam and the Old Age Psychiatry Department of GGZinGeest Amsterdam, the Netherlands (inpatient and outpatient) between April 2011 and June 2013. Patients were directed to these specialized health care institutions by primary care physicians or a medical specialist. Patients were only included if they were aged between 45 and 75 and if they had a Frontal Behavior Inventory score of 11 of higher and/or a Stereotype Rating Inventory of 10 or higher. Exclusion criteria were as follows: 1) an already established diagnosis of dementia or a psychiatric diagnosis that could explain behavioral problems, 2) Mini Mental State Exam score less than 18, 3) traumatic brain injury, mental retardation or drug or alcohol abuse in medical history, 4) lack of reliable informant, 5) insufficient communicative skills of either patient or the closest informant, 6) acute onset of behavioral problems, 7) clinically apparent aphasia or semantic dementia, 8) MRI contra-indications [18].
A total of 137 patients were included at baseline. After two years, 8 patients turned out to have a neurologic or general disease like multiple sclerosis or obstructive sleep apnea but without a neurodegenerative or clinical psychiatric diagnosis and were therefore excluded from this study, as were patients with subjective complaints or relational problems without a formal diagnosis (n = 11) and patients with vascular mild cognitive impairment (n = 3). Nine participants did not complete the MINI-Plus due to patients’ tiredness and logistic problems. 15 patients were lost to follow up and therefore excluded. The remaining 91 patients were included. Out of these patients, 23 patients had a probable or definite bvFTD diagnosis (19 probable bvFTD, 4 definite bvFTD out of which 1 patient with histopathological confirmation postmortem, 2 patients with C9ORF72 repeat expansion, and 1 patient with a GRN mutation), 3 patients had possible bvFTD and 25 patients had another neurodegenerative disease (7 patients with AD, 4 with vascular dementia, 4 with dementia with Lewy bodies, 5 with progressive supranuclear palsy, 2 with semantic dementia, 1 with corticobasal degeneration, 1 with Huntington’s disease, and 1 with a cerebellar dementia). A total number of 40 patients received a clinical psychiatric diagnosis (2 patients with schizophrenia, 12 with a major depressive disorder, 4 with minor depressive disorder, 7 with a bipolar disorder, 3 with an autism spectrum disorder, 1 with obsessive compulsive disorder, 1 with a general anxiety disorder, 10 with personality problems).
Diagnostic procedure
Before inclusion, informed consent was obtained from all participants or, in case of incompetence of giving a fully informed consent, obtained from the caregiver or legal representative. All patients underwent a standardized assessment, including medical history and family history, informant-based history, physical, neurological and psychiatric examinations, neuropsychological assessment, laboratory tests, and magnetic resonance imaging (MRI) of the brain acquired on a 3T Signa HDxt scanner (GE Medical Systems, Milwaukee, WI) following a standard MRI protocol for dementia. In case of normal or insufficiently explanatory MRI results (not explaining frontosubcortical dysfunction), an [18F]FDG-PET scan was performed EXACT HR
scanner (Siemens/CTI, Knoxville, TN). Neurological and psychiatric evaluation was done by both a neurologist as well as an experienced geriatric psychiatrist (YP, NP, PS, AD, CK, MS). In a multidisciplinary consensus meeting the neurologist and psychiatrist determined the specific diagnosis, and whether a diagnosis of probable or definite bvFTD, possible bvFTD, other neurodegenerative disease, or a psychiatric disorder was applicable. Diagnoses were based on the National Institute on Aging-Alzheimer’s Association guidelines for AD, the National Institute of Neurological Disorders and Stroke and Association Internationale pour la Recherché et l ‘Enseignement en Neurosciences (NINDS-AIREN) criteria for vascular dementia, the International Consensus Diagnostic Criteria for dementia with Lewy bodies, the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition for psychiatric disorders, and the International bvFTD Criteria Consortium criteria for bvFTD [5, 19–21]. All included subjects were genetically screened for C9orf repeat expansion, given the great symptomatic overlap with psychiatric disorders and long disease courses that have been described in this mutation type [22]. A diagnosis of possible bvFTD was only applied when a clinical psychiatric diagnosis was excluded. All patients were followed during a period of 2 years and the clinical diagnosis after 2 years was used as the gold standard.
Measurements
At baseline we set out to determine the specific psychiatric disorders according to DSMIV and ICD-10 criteria in the diagnostic groups, using the MINI-International Neuropsychiatric Interview (MINI-Plus). Additionally we defined the frequency and character of past psychiatric disorders according to DSMIV and ICD-10 criteria in these groups, using the MINI-Plus.
The MINI-Plus is a structured diagnostic interview, developed to assess the psychiatric diagnoses of patients according to DSMIV and ICD-10 criteria. It is widely used in patients with neuropsychiatric symptoms and it has been found to have a good sensitivity, specificity and inter-rater and test-retest reliability [23]. It contains 16 DSMIV psychiatric disorders including inter alia unipolar and bipolar mood disorders, psychotic disorders like schizophrenia and anxiety and obsessive compulsive disorders. It does not include autism spectrum disorders and personality disorders besides the antisocial personality disorder. The MINI-International Neuropsychiatric Interview (MINI-Plus) incorporates both criteria for current psychiatric disorders according to DSMIV and ICD-10 (present psychiatric disorders) as well as criteria for past psychiatric disorders according to DSMIV and ICD-10 criteria (psychiatric disorders in previous history).
All included patients underwent a MINI-Plus interview, performed by trained clinicians who were blind for the clinical diagnosis (FG, WK). These clinicians did not have information about previous medical history neither other advance medical information. They worked separately from the clinicians involved in the diagnostic procedures. The MINI-Plus result was not included in the clinical evaluation.
Statistical analyses
Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS for Windows; IBM, Armonk, NY), version 21. Group differences on sociodemographic variables were investigated using independent t-tests and χ2 tests. For the main analysis, the clinical diagnosis was the independent variable and the formal psychiatric disorder was the dependent variable. Differences between groups for categorical variables were evaluated using χ2 tests and Fisher’s exact tests. A p-value of <0.05 was considered statistically significant, except otherwise indicated.
Ethical considerations
The study was approved by the Medical Ethics Committee of the VU University Medical Center, Amsterdam. All participants signed a written consent form.
RESULTS
Patient characteristics
Demographic data of the diagnostic groups are displayed in Table 1. Patients were predominantly Caucasian (>90%) and male (>60%). Regarding disease duration, the groups did not differ (p > 0.05). In the group of patients with a psychiatric diagnosis, the patients were significantly younger (p < 0.05) and had significantly more males than in the other groups.
Fulfillment of DSMIV and ICD-10 criteria for current psychiatric disorder
The overall frequency of probable and definite bvFTD patients who fulfilled criteria for a current psychiatric disorder according to DSMIV and ICD-10 criteria was 21.7%. This frequency did not differ from patients with other neurodegenerative diseases (12.0%) or possible bvFTD (66.7%), but was significantly lower than patients with a clinical psychiatric diagnosis (57.5%, p < 0.01; Table 2a). Out of 3 patients with genetic bvFTD (2 with C9ORF72 repeat expansion and 1 with GRN mutation), 1 patient fulfilled criteria for a current psychiatric disorder according to DSMIV and ICD-10 criteria (33.3%).
Probable and definite bvFTD patients fulfilled the formal criteria most commonly for unipolar mood disorders (13.0%). Patients with a psychiatric diagnosis fulfilled criteria for a bipolar disorder significantly more often than patients with probable and definite bvFTD (F 0.04). Some patients fulfilled criteria for multiple psychiatric disorders according to DSMIV and IC-10 criteria at the same time.
Fulfillment of DSMIV and ICD-10 criteria for past psychiatric disorder
The overall frequency of any past psychiatric disorder according to DSMIV and ICD 10 criteria in probable and definite bvFTD was 8.7%. This did not differ significantly from patients with other neurodegenerative diseases (16.0%) or possible bvFTD (66.7%), but was significantly less than patients with a clinical psychiatric diagnosis (62.5%, p < 0.01; Table 2b).
Unipolar mood disorders were the most prevalent past psychiatric disorder in probable and definite bvFTD (8.7%), but significantly less prevalent than in patients with a psychiatric diagnosis (42.5%, p 0.005). Patients diagnosed with probable and definite bvFTD did not significantly differ from patients with other neurodegenerative diseases or possible bvFTD regarding past unipolar mood disorders. Patients with a psychiatric diagnosis fulfilled criteria for a past bipolar disorder significantly more often than patients with probable and definite bvFTD (F 0.04). As shown in Table 2b, some patients fulfilled DSMIV and ICD-10 criteria for multiple past psychiatric disorders.
DISCUSSION
Contrary to our expectations we found that despite large clinical symptomatic overlap between bvFTD and psychiatric disorders, formal psychiatric disorders are not overrepresented in probable bvFTD. Of 91 patients with a late-onset frontal syndrome, the overall frequency of psychiatric disorders according to DSMIV and ICD 10 criteria in 23 probable bvFTD patients (21.7%) did not differ from patients with other neurodegenerative diseases (n = 25) or possible bvFTD (n = 3), but was less than in patients with a clinical psychiatric diagnosis (n = 40). Although in our group of patients with a psychiatric diagnosis the male gender was more prevalent than in the other groups, fulfillment of DSMIV and ICD-10 criteria was not higher in male psychiatric patients than in female psychiatric patients (57.6% in male versus 57.1% in female patients). Previous literature showed that bvFTD is still underdiagnosed and often initially mistaken for psychiatric illnesses [24, 25]. It was found that consensus criteria for bvFTD and other neurodegenerative diseases are crucial to differentiate bvFTD from other dementias, but psychiatric misdiagnoses must still be ruled out [25, 26]. In contrast to a previous study, using retrospective chart reviews, reporting that 50.7% of bvFTD patients receive a prior psychiatric diagnosis [12], our results show that correctly applying DSMIV and ICD-10 criteria in bvFTD gives a lower rate of formal psychiatric disorders in bvFTD (21.7%).
The most common psychiatric disorders according to DSMIV and ICD-10 criteria that probable bvFTD patients fulfilled were unipolar mood disorders (13.0%). A recent meta-analysis including 29 studies showed that depressive mood and its manifestations are recognized in approximately one third (33%) of patients with bvFTD [26]. However, the majority of these studies about the prevalence of (comorbid) depression in bvFTD are based upon reports of depressed mood only [28]. Studies applying formal DSMIV and ICD-10 criteria are limited and suggest lower rates [29, 30]. Lopez et al. prospectively evaluated DSM-IIIR criteria in 20 patients with frontotemporal dementia (six autopsy-proven) and found that 25% (n = 5) met DSM-III-R criteria of a major depression [29]. Gregory et al. found 3 bvFTD patients out of 15 who reported sadness, but only one met criteria for DSM-IV major depression [30]. The discrepancy between prevalence of depressive mood in bvFTD and the prevalence of depressive and dysthymic disorder according to DSMIV and ICD-10 criteria may highlight the risk of overdiagnosing depressive disorder in bvFTD when not using formal criteria.
It is remarkable that in our cohort, patients with a psychiatric diagnosis fulfilled criteria for a bipolar disorder relatively often (17.5%) and significantly more often than patients with probable or definite bvFTD. This is probably related to the inclusion criteria of our specific cohort with only patients with the frontal lobe syndrome, also in the subgroup of patients with a psychiatric diagnosis. Previous studies discovered a clinical overlap between the frontal lobe syndrome and bipolar disorder, regarding social cognition, executive disturbances and behavioral profiles [31–33]. A previous study also found that patients with bvFTD who are initially misdiagnosed with a psychiatric diagnosis are more likely to receive a diagnosis of bipolar disorder than patients with other neurodegenerative diseases [11, 12]. The current study shows that by correctly applying DSMIV and ICD-10 criteria, the prevalence of bipolar disorder in bvFTD is not higher than in patients with other neurodegenerative diseases.
Regarding past psychiatric disorders, results show that patients with probable bvFTD fulfill DSMIV and ICD-10 criteria for a past psychiatric disorder less often than patients with a clinical psychiatric diagnosis (8.7% versus 62.5%), but remarkably not more than patients with other neurodegenerative diseases or possible bvFTD. Previous studies suggested that many patients develop neuropsychiatric symptoms before impending dementia [14, 15]. One study found that in the presence of cognitive decline, mild behavioral impairment and neuropsychiatric symptoms have similar risks to convert to FTD as AD [16]. This is the first study about preceding psychiatric disorders according DSMIV and ICD-10 criteria in a frontal cohort, revealing that formal preceding psychiatric disorders are as present in bvFTD as other neurodegenerative diseases, but not redundant.
In addition, we found that depression and dysthymic disorder were the most frequent past DSMIV and ICD-10 disorders among probable bvFTD patients. This is an intriguing result, especially in the context of recent findings about the prodromal phase of bvFTD and AD. Mendez et al. already suggested that depression may be a prodrome of frontotemporal dementia and a possible familial risk factor [28]. Similar to studies of late-onset AD and autosomal dominant AD, our results suggest increased rates of depression prior to onset of bvFTD or when a threshold of neurodegeneration is passed [14, 15]. These findings have implications for the early detection and treatment of patients with probable bvFTD and might influence pathogenic concepts on the concurrence of bvFTD and psychiatric disorders.
There are some limitations in this study. First of all, although the MINI-Plus for psychiatric diagnosis was performed from sections A through Z, autism spectrum disorders and personality disorders, except antisocial personality disorder, are not categorized in this interview. For this reason patients clinically diagnosed with a psychiatric diagnosis did not meet criteria for a DSMIV disorder in more than 57.5% of cases. It is conceivable that more bvFTD patients or patients with another neurodegenerative disease would meet criteria for a psychiatric disorder if the MINI-Plus interview included more personality disorders and autism spectrum disorders. Second, recent studies showed that psychiatric symptoms seem to occur more often in genetic FTD, especially in patients with C9ORF72 repeat expansion, which ensured us to test all our patients for this mutation [34]. However, in our study we had only 3 patients with genetic bvFTD (2 with C9ORF72 repeat expansion and 1 with GRN mutation) which made it impossible to examine significant differences between genetic bvFTD and sporadic bvFTD regarding the prevalence of psychiatric disorders. Another constraint is that we cannot exclude that lack of insight and memory problems have limited the sensitivity, specificity and reliability of the MINI plus interview in bvFTD patients. The prevalence of neuropsychiatric symptoms may thus have been underestimated. Lack of insight, however, is common among psychiatric patients for which the interview is validated [36–38]. Besides, also patients with other neurodegenerative diseases often deal with memory problems. The period of time from the clinical diagnosis to the past psychiatric disorder was not always exactly known. Future studies focusing on the prodromal phase of bvFTD are warranted to better understand whether these past psychiatric disorders represent a prodromal stage or comorbidity.
Even though there are limitations, this is the first study which systematically and prospectively demonstrates that formal psychiatric disorders are not overrepresented in probable bvFTD. Despite large clinical symptomatic overlap between bvFTD and psychiatric disorders, misdiagnosis in bvFTD could be limited by correctly applying diagnostic criteria for psychiatric disorders. In suspected bvFTD close collaboration between neurologists and psychiatrists will advance diagnostics and treatment.
