Abstract
The proposed reconceptualization of frontotemporal dementia by Looi et al. (2012) should be of interest to every reader, not just those who practise neuropsychiatry or old-age psychiatry. There are implications for clinical practice and for the evolution of diagnostic criteria.
The traditional encapsulation of a neurodegenerative disorder, presenting with frontal lobe damage and distinguishable from Alzheimer’s disease (AD), known as Pick’s disease (Lishman, 1987), has evolved greatly over the past 20 years. The overarching terms frontotemporal dementia (FTD) and frontotemporal lobar degeneration are now used to describe three clinically heterogeneous neurodegenerative disorders associated with focal atrophy of the anterior temporal lobe or the orbitomedial frontal lobe. There are two that present with decline in language function, progressive non-fluent aphasia and semantic dementia, and another termed behavioural variant frontotemporal dementia (bvFTD), which is the focus of the paper by Looi et al. (2012).
The general psychiatrist may well encounter FTD in clinical practice. A particularly striking feature is that FTD is the second most common cause of younger onset dementia, with an estimated prevalence of 15 cases per 100 000 between the ages of 45 and 64 (Ratnavalli et al., 2002). FTD presents with a range of non-specific symptoms that are suggestive of impairment of social cognition and emotional regulation, rather than overt neurocognitive impairment. These symptoms are associated with disturbance of the respective frontostriatal circuits – apathy, emotional blunting and social withdrawal (anterior cingulate circuit); impulsivity, disinhibition and changes in eating habits (orbitofrontal circuit); and functional decline from impaired problem solving and memory (dorsolateral prefrontal circuit).
Its presentation with standard psychiatric symptoms may lead to FTD being diagnosed as one of the major mental illnesses in younger adults. A Melbourne group has documented a subset of FTD cases that had attracted a previous diagnosis of schizophrenia or bipolar disorder; almost all had presented to psychiatric services by the age of 60 and more than half by the age of 40 (Velakoulis et al., 2009). The authors hypothesized that FTD presenting with psychosis may have been a non-specific effect of a neurodegenerative insult in a younger brain. Another possibility was that shared localization of neuropathology led to FTD presenting as an organic phenocopy of schizophrenia.
This highlights the conundrum of basing psychiatric diagnosis predominantly on clinical grounds. Unique among disorders treated by psychiatrists, the dementias are clinicopathological entities, defined by a clinical phenotype and specific neuropathological changes such as amyloid deposition in AD or tau protein in FTD (Dubois et al., 2010). As brain pathology cannot be directly investigated in vivo, formal diagnosis of the dementias remains reliant on symptomatology and neurocognitive examination.
To move beyond clinically based, probabilistic diagnoses, psychiatry needs to find valid biomarkers, the basis of the paper by Looi et al. (2012). Apart from their application as a diagnostic tool, biomarkers may also be used to stage disease, inform prognosis, and predict and monitor response to a therapeutic intervention (Biomarkers Definitions Working Group, 2001). Indeed, a decade ago, an agenda for integrating basic science and neuroscience research into classificatory systems was set up as a challenge for the authors of the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5; Charney et al., 2002). At this stage, none of the proposed DSM-5 diagnostic revisions has recommended the use of specific biomarkers.
The DSM-5 Neurocognitive Disorders Work Group has proposed the criterion of a clear decline in neurocognitive performance on formal testing, or equivalent clinical evaluation, for diagnosing the dementias (American Psychiatric Association, 2011). In this regard, the Mini Mental State Examination (MMSE) is not sensitive in screening for FTD, but the Addenbrooke’s Cognitive Examination Revised (ACER) is and can be used in the clinical setting. The difficulty with proposing specific patterns of neurocognitive change as a diagnostic aid in the dementias is that, in FTD, patients with behavioural change continue to perform well in early disease, and deficits in episodic memory may be more common than previously thought (Piguet et al., 2011).
The proposed DSM-5 criteria for the dementias, which regards evidence of neuroimaging changes as desirable rather than essential in increasing diagnostic certainty, stands in contrast to other criteria recently proposed by international groups that have included in vivo markers of pathological changes (Dubois et al., 2010; Rascovsky et al., 2007). One set of criteria has noted that AD is ‘a clinicobiological entity that requires biomarkers for formal diagnosis’ (Dubois et al., 2010), while another has advocated biomarkers to increase the certainty of diagnosis of FTD from ‘possible’ to ‘probable’ (Rascovsky et al., 2007). The main biomarkers proposed by both criteria are similar and may be used to help distinguish AD and FTD – structural MRI (medial temporal atrophy vs frontal and/or anterior temporal atrophy, respectively), single photon emission computed tomography (SPECT) (temporoparietal hypoperfusion vs frontal hypoperfusion) and cerebrospinal fluid (CSF) biomarkers (the ratio of tau protein to amyloid beta is lower in FTD).
So is this confidence in using biomarkers in dementia diagnosis justified? The history of psychiatry is littered with promising biomarkers that have eventually disappointed, notably the dexamethasone suppression test for depression. Potential problems include validity, cost, availability of clinical experts trained in the interpretation of these biomarkers and relevance to the early or even preclinical stages of disease. Critics have argued that a biological measure that has not yet been empirically validated should be termed a ‘biomeasure’, not a biomarker (Kraemer et al., 2002).
Looi et al. (2012) have proposed frontostriatal dysfunction as a specific endophenotype – an intervening variable that potentially marks the path between genotype and the behaviour of interest (Gottesman and Gould, 2003) – for FTD. Yet, striatal atrophy is not specific to FTD. Functional abnormalities of the striatum and greater corticostriatal circuitry also exist in certain mood disorder subtypes, with evidence of striatal hypoactivity in unipolar depression (Marchand and Yurgelun-Todd, 2010), and reduced caudate nucleus volume in older people with depression (Hickie et al., 2007). It has been hypothesized that schizophrenia is associated with elevated dopamine function in those areas of the striatum most involved in dorsolateral prefrontal cortex information processing (Kegeles et al., 2010).
In conclusion, while the specific conceptualization of FTD has advanced significantly over recent times, the general use in psychiatry of biomarkers to improve clinical validity remains very much in its infancy. In particular, it is unclear whether the frontostriatal changes advocated as a marker of FTD represent a primary or early pathological process of the frontal cortex, with secondary or later pathology in the striatum, or vice versa.
See Review by Looi et al., 2012, 46(5): 422-434