The neuropsychology and neurobiology of late-onset schizophrenia and very-late-onset schizophrenia-like psychosis: A critical review

Highlights

• LOS and VLOSP are characterized by diffuse and mildly progressive cognitive dysfunction.

• The nature and progression of neuropsychological deficits differ from changes in neurodegenerative processes.

• Neurobiological underpinnings consist of generalized brain atrophy and increased ventricle-to-brain ratio, white matter pathology and functional changes suggesting a fronto-subcortico-temporal pattern of impairment.

• The contribution of neurobiological mechanisms in LOS and VLOSP may be modulated by premorbid (psychosocial and genetic) vulnerability.

• More research is needed to draw firm conclusions.

Abstract

Objective

The current review discusses neuropsychological profiles and the longitudinal course of cognitive dysfunction in Late Onset Schizophrenia (LOS) and Very-late-onset schizophrenia-like psychosis (VLOSLP), and attempts to clarify its neurobiological underpinnings.

Method

A systematic literature search resulted in 29 publications describing original research on the neuropsychology of LOS/VLOSLP and 46 studies focussing on neurobiology.

Results

Although mildly progressive cognitive impairment is usually present, only a subgroup of LOS/VLOSLP develops dementia during a 10-year follow-up succeeding the onset of psychosis. This coincides with the absence of neuropathological evidence for neurodegeneration in many cases. Cognitive deterioration is characterized by deficits in (working) memory, language, psychomotor speed and executive functioning. Underlying neurobiological changes encompass white matter pathology, increased ventricle-to-brain ratio (VBR) with coinciding atrophy and hypo-metabolism of frontal, temporal and subcortical areas.

Conclusions

Multiple changes in neurobiology and cognition contributing to LOS/VLOSLP may reflect stress-related accelerated brain aging rather than neurodegenerative pathology. Their involvement in the onset of illness, however, might be inversely proportional to pre-existing (psychosocial and/or genetic) vulnerability to psychosis.

Introduction

Although schizophrenia is generally considered as a disease with onset in adolescence, several studies report on individuals who first experienced psychotic symptoms in late life in the absence of a mood disorder or neurological illness (Binbay et al., 2012, Howard et al., 2000, Sharma et al., 2014). An international expert consensus referred to psychosis with an onset after 40 years of age as late-onset schizophrenia (LOS). Very-late-onset schizophrenia-like psychosis (VLOSLP) was delineated as a condition with onset of psychotic symptoms after 60 years (Howard et al., 2000). Although these age ranges seem arbitrary, they result from expert discussion and are hypothesized to reflect differing subtypes of schizophrenia-like illnesses with separate mechanisms contributing to onset and maintenance of symptoms (Howard et al., 2000, Palmer et al., 2001).

The one-year prevalence of schizophrenia in individuals aged 45–65 years is 0.6% (Keith et al., 1991), whereas the community prevalence for VLOSLP ranges from 0.1% to 0.5% (Howard et al., 2000). However, in two samples of 8010 Dutch and 1777 British patients aged 60 years or older who were admitted to hospital, there was a linear trend in the association between increasing age and first onset of non-organic, non-affective psychosis. Specifically, the annual incidence of VLOSLP increased by 11% with each 5 year increase in age (van Os et al., 1995). As the older age groups are the fastest growing segment in the world population, healthcare may thus increasingly be confronted with a first episode of psychosis in elderly patients. Hence, research into this highly incapacitating disorder is needed. Even more so as prior studies suggest that research on schizophrenia with onset before the age of 45 years (Average age Onset Schizophrenia, AOS), is not always generalizable towards LOS or VLOSLP. For instance, in contrast with AOS, there is a female preponderance in VLOSLP and LOS (Howard et al., 2000). Furthermore, there is a lower morbid risk in relatives of individuals with VLOSLP in comparison with AOS, suggesting a lower genetic risk factor for LOS/VLOSLP (Howard et al., 1997). Finally, clinical features differ, with the most striking differences between AOS and VLOSLP (Kerssens et al., 2006). VLOSLP is characterized by positive psychotic symptoms such as delusions and (multimodal) hallucinations. Specifically, partition delusions – concerning the unwanted permeability of borders – and paranoid delusions are highly prominent (Hanssen et al., 2015). On the other hand, formal thought disorder is usually absent. Furthermore, there are fewer negative symptoms such as affective flattening in VLOSLP and LOS compared to AOS (Howard et al., 2000). These findings regarding epidemiology and phenomenology point to (partly) diverging etiological mechanisms in AOS and VLOSLP. They may also be associated with different cognitive profiles, as for instance negative symptoms and disorganization are related to executive dysfunction more often than reality distortion in AOS (Frith, 1996). Moreover, female predominance may mask group differences with respect to social cognition between AOS and LOS/VLOSLP as there are known gender differences in social cognition in healthy controls, suggesting better performance in females (Green et al., 2015).

Some researchers have disputed the diagnostic validity of LOS or VLOSLP (Andreasen, 1999, Brodaty et al., 2003). They suggest that particularly VLOSLP encompasses a prodromal phase of a neurodegenerative disease such as Alzheimer’s disease (AD) with predominantly memory (and language) dysfunction (Brendan and Petersen, 2007, Fichman et al., 2011), frontotemporal degeneration (FTD) with important executive dysfunction and impaired social cognition in the behavioral variant (Hutchinson and Mathias, 2007), or dementia with Lewy bodies (LBD) which is initially primarily characterized by perceptual and executive deficits (Collerton et al., 2003) according to systematic review and/or meta-analysis. Neurodegenerative illnesses are certainly more prevalent in late life and are often accompanied by behavioral and psychological symptoms, such as depression or psychosis (Savva et al., 2009). Moreover, these symptoms may occur before cognitive decline becomes apparent. Coincidently, mild cognitive impairment (MCI) is frequently observed in LOS and VLOSLP, thus creating an overlap in clinical presentation of LOS or VLOSLP and neurodegenerative disorders.

Still, research has shown that VLOSLP, though it may be associated with an increased risk of dementia compared to normally aging older adults (Kørner et al., 2009), does not invariably predict cognitive or functional decline (Lagodka and Robert, 2009, Rabins and Lavrisha, 2003). In many cases there is a non-progressive dysfunction, reminiscent of the static encephalopathy of AOS. Nevertheless, in late life a static cognitive dysfunction accompanied by psychotic symptoms might not only represent a (latent) neurodevelopmental disorder surfacing as a result of environmental triggers, it is possibly combined with stress-related accelerated neurobiological aging which may increase pre-existent vulnerability to psychopathology. Finally, non-progressive cognitive impairment in later life is often also associated with cerebrovascular disease.

Differential diagnosis between an incident dementia or a static (neurodevelopmental and/or cerebrovascular) encephalopathy in late onset psychosis is valuable with respect to early treatment possibilities. Neuropsychological assessment is one possible non-invasive and sensitive diagnostic tool that may provide insight into the characteristic cognitive and related neurobiological profiles in LOS and VLOSLP as opposed to neurodegeneration (Frith, 1996, Zakzanis et al., 2001), and it may also further our understanding of the (socio)cognitive and neurobiological mechanisms that cause an onset of psychosis in late life and sustain it (Frith, 1996, Phillips et al., 1997). However, common clinical practice mostly utilizes neuropsychological assessment in conjunction with functional or structural brain imaging as proper diagnostic research encompasses information from different and compatible sources (Ruff, 2003). Whereas neuropsychological assessment is usually conducted for the purpose of differential diagnosis, prediction of functional potential, and measuring treatment response or disease course, clinical correlation with imaging findings can be complementary as imaging studies can specify the location of many structural and functional brain changes leading to cognitive changes (Harvey, 2012). Notably, correlations may sometimes be low or absent as brain changes evident in imaging can be associated with nearly normal cognitive functioning due to compensatory neurological mechanisms, while individuals with no lesions detectable on imaging can have substantial cognitive and functional limitations, leading to the suggestion that neuropsychological research may sometimes detect subtle disease-related changes before these become apparent on imaging (Harvey, 2012). In summary, neuropsychological research allows for quantitative descriptions of the patient’s cognitive status, and combined with imaging results it may increase our understanding of brain–behaviour relationships (Ruff, 2003). However, there have been little studies directly correlating neuropsychological results with imaging research, electro-encephalogram (EEG) or event related potential (ERP) in LOS or VLOSLP.

In the current review, we will therefore first try to identify a neuropsychological profile which is characteristic of VLOSLP or LOS. Additionally, we will look at research concerning its typical evolution and possible underlying neuroanatomical or functional changes. We will review study results and implications for clinical practice. To conclude, we will suggest possibilities for future research.