Clinical Use of Dementia Biomarkers to Evaluate Cognitive Deficits in People With Primary Psychiatric Disorders

Chronic primary psychiatric disorders (PPD), including schizophrenia, schizoaffective disorder, bipolar disorder, and treatment-refractory depression, are associated with accelerated cognitive decline in later life, frequently leading to premature loss of autonomy.^1,2 Individuals living with PPD are not considered to have formal neurodegenerative disorders, but they may have well-documented cognitive disturbances associated with their psychiatric conditions. During acute psychiatric episodes, cognitive symptoms are often more severe in people with PPD than in people with mild stages of dementia.^3,4 In their chronic stable phases, people with schizophrenia have mild to moderate multidomain cognitive weaknesses (1 to 2 SD below the mean), whereas people with bipolar disorder have milder cognitive deficits (0.5 to 1 SD below the mean).⁵ These symptoms are multifactorial, reflecting the effects of the disease (eg, cognitive symptoms can be present at the onset of psychosis), lifestyle risk factors (eg, smoking, inactivity, social isolation), as well as the side effects of certain medications.

A subset of individuals have more pronounced cognitive decline in midlife or late life, and they must be assessed clinically to determine whether symptoms are the result of the primary psychiatric process or an additional neurodegenerative process, such as Alzheimer disease (AD), frontotemporal dementia (FTD), or Lewy body dementia (LBD). These individuals are encountered frequently in memory clinics⁶ and often represent a major clinical challenge because the effects of severe and persistent mental illnesses (SPMI) confound the usual assessment tools, including neuropsychologic testing. Objective biomarkers play a crucial role in the clinical assessment, but large knowledge gaps remain in the context of psychiatric disorders. These clinical biomarkers can be divided into those that evaluate neurodegeneration and those directly targeting molecular pathology.

Neurodegeneration Biomarkers

Brain MRI

MRI plays an important role in the assessment of cognitive deficits in people with SPMI but with major limitations. MRI serves 3 major functions: excluding potentially treatable causes of dementia (eg, tumor, infection, normal-pressure hydrocephalus), quantifying the burden of vascular damage, and identifying signs of specific neurodegenerative pathology (eg, hippocampal atrophy, frontotemporal atrophy). Analysis of MRI scans in people with chronic psychiatric illnesses is complicated by the higher proportion of vascular risk factors in this group (eg, smoking; suboptimal medical management of chronic diseases, such as hypertension),⁷ leading to an increased burden of white matter hyperintensities in addition to the mild brain volume loss and ventricular enlargement associated with those conditions.⁸ Despite these limitations, careful review of regional atrophy, ideally using structured visual rating scales, can assist in the identification of clearly abnormal volume loss confirming a degenerative process. MRI scans can also reveal signs of rare white matter or metabolic diseases that present as adult-onset progressive neuropsychiatric conditions.⁹

Brain FDG-PET

[18F]Fluorodeoxyglucose positron emission tomography (FDG-PET) can be used to evaluate regional patterns of hypometabolism as measured by reduced glucose uptake. FDG-PET is generally used as a second-line test after structural imaging, given its higher sensitivity to early neurodegenerative processes, including AD, FTD, and LBD, and its capacity to differentiate those etiologies.¹⁰ The general assumption is that people with PPD should have normal FDG-PET results, allowing the differentiation of dementia subtypes. However, studies have cast doubts on this assumption, finding a high frequency (up to 45%) of mildly abnormal findings among people with PPD.¹¹ There is a knowledge gap in the context of aging in people with SPMI, as there is no reported large-scale database of FDG-PET profiles in this population. Overall, FDG-PET is a useful test in this clinical context, particularly when results are within normal limits, to exclude a dementia process, given its high sensitivity in the early stages. However, mildly abnormal or ambiguous findings should not be interpreted as proof of a neurodegenerative process in the context of SPMI or severe acute psychiatric symptoms.¹²

NfL and GFAP

Serum or cerebrospinal fluid (CSF) neurofilament light chain (NfL) is a robust marker of axonal damage that can be used clinically to identify neurodegeneration across several pathologies.¹³ NfL is emerging as having a potential role in the differential diagnosis of PPD and dementia.¹⁴ Clinical use is becoming feasible with the availability of normative data across the healthy lifespan and its reliable measurement as a blood-based marker (as opposed to requiring CSF analysis).¹⁵ Studies have shown that NfL can differentiate PPD from behavioral variant FTD (bvFTD) with accuracies greater than 80%.¹² In the SPMI context, it is important to recognize that NfL levels increase with age (with a more pronounced effect after age 60 to 70)¹⁵ and that people with PPD have slightly higher mean NfL levels compared with controls.¹⁶ Therefore, the accuracy of the test is likely to decrease in older individuals with PPD.

Glial fibrillary acidic protein (GFAP) is found in glial cells and plays a role in their activation and neuroinflammation after brain injury or degeneration, being released into biofluids.¹⁷ GFAP is another candidate biomarker for differential diagnosis between PPD and neurodegenerative diseases,¹⁸ but some studies have shown a significant increase of GFAP in PPD, such as major depressive disorder, without neurodegeneration. The role of GFAP in this clinical context remains uncertain, and it is not recommended for clinical use.

Molecular Pathology Biomarkers and Genetics

Amyloid Biomarkers

Soluble and insoluble amyloid plaques can be detected reliably and measured with CSF biomarkers (eg, lower AΒ42 or AΒ42:AΒ40 ratio) and the various amyloid PET tracers, respectively.¹⁹ Based on available data, it does not appear that people with SPMI have a higher prevalence of amyloid pathology compared with the general population. Therefore, these tests can be used to identify AD in the context of psychiatric symptoms.²⁰ The interpretation of amyloid biomarkers requires clinical judgment; although their absence excludes AD pathology, the presence of abnormal amyloid does not necessarily imply that it is the cause of the cognitive deficits observed in an individual. Indeed, an older individual with SPMI could have mild cognitive deficits attributable to the multifactorial effect of schizophrenia and asymptomatic amyloid plaques without having started a neurodegenerative cascade leading to clinical AD. Amyloid positivity must be interpreted factoring in age, given the higher proportion of asymptomatic amyloid plaques in older people, and testing is generally not recommended in people older than 75. The causative role of the positive amyloid biomarker in an individual can be inferred based on the nature of cognitive symptoms (eg, explainable by PPD or not) and the presence or absence of other neurodegeneration or tau biomarkers.

Tau Biomarkers

Tau proteins (tau and phospho-tau) can be measured in CSF or with tau PET tracers, including flortaucipir F-18 and a variety of other research tracers.²¹ The accumulation of tau is closely related to neurodegeneration and cognitive symptoms in AD, and, therefore, constitutes an excellent biomarker for the differential diagnosis. There are several relevant caveats in the context of psychiatric disorders. As for NfL and GFAP, one should not assume that people with SPMI will have a similar baseline level of tau as healthy individuals. Studies are emerging showing that tau accumulation could be accelerated in chronic psychiatric illnesses and may be related to premature cognitive decline.²² Whereas tau PET tracers are effective at capturing the degeneration and neurofibrillary tangles of AD, they do not bind accurately to the various tau subtypes of frontotemporal lobar degeneration. Therefore, a negative result with tau biomarkers would not exclude an FTD diagnosis, for example. New tracers are under development, such as the [18F]Florzolotau tau tracer, which could have better accuracy for frontotemporal lobar degeneration,²³ but they are not yet ready for clinical use.

Genetic Testing

Genetic testing in the context of neurocognitive disorders is used to identify causative sequence variations in people with confirmed diagnoses and relevant family history rather than to determine whether an individual has a neurodegenerative process. The one relevant exception is a hexanucleotide repeat expansion within C9orf72, which is an autosomal dominant genetic cause of FTD and amyotrophic lateral sclerosis (ALS).²⁴C9orf72 is a complex sequence variation resulting in both loss of function and gain of function, leading to protracted heterogenous psychiatric prodromes before reaching the clinical FTD or ALS stage associated with abnormal brain imaging findings.²⁵ Therefore, genetic testing for C9orf72 can be considered in people with atypical or late-onset psychiatric symptoms when they have a first-degree family history of FTD, ALS, or early-onset dementia.¹² Testing for other FTD sequence variations should be considered, because psychiatric prodromes have been described at a lower frequency in other mutations, including progranulin (GRN).

Integrative Clinical Approach

The assessment of neurocognitive disorders in the context of chronic psychiatric illness follows the usual clinicopathologic approach, but also takes into account several confounding factors (Figure). The assessment requires detailed cognitive testing and a psychiatric assessment to determine symptom severity and stability. Potential symptoms of bvFTD must be differentiated from those of PPD. If objective cognitive deficits are identified, they must be interpreted in light of current psychiatric disturbances to determine whether they are explainable by the psychiatric pathology. For example, mild executive dysfunction and poor memory retrieval could be easily explained by chronic schizophrenia, whereas naming and semantic deficits in an individual with major depression would require investigations to rule out a comorbid neurodegenerative condition. This decision requires integrative clinical judgment and cannot be replaced by simple cutoff levels on cognitive tests. Specialized neuropsychologic tests evaluating social cognition, such as facial emotion or famous people recognition, can be helpful in the context of disambiguating PPD from bvFTD or right temporal variant FTD.²⁶

The temporal progression of symptoms is another key feature to assess. The probability of a neurodegenerative disorder is low if symptoms are static or fluctuate over long periods, but more likely if cognitive deficits are clearly progressing over time.

In the absence of a clear cognitive disorder, it is preferable to focus on improving psychiatric treatments (eg, minimizing anticholinergic side effects) and psychosocial support rather than prescribing excessive tests that could yield ambiguous false-positive results. If a formal cognitive disorder is suspected, the initial step should include the usual basic laboratory profile and structural brain imaging, preferably MRI, but CT scanning can be used depending on access and the level of clinical suspicion. If a diagnosis is not established after this first round of investigations, a brain FDG-PET is an appropriate next step. Negative FDG-PET results generally end investigations, unless the clinical profile and family history indicate a possible genetic mutation associated with FTD.

In case of an ambiguous family history, the C9orf72 expansion test can be sufficient, because the C9orf72 expansion is the most frequent cause of atypical psychiatric prodromes. However, if a clear first-degree family history of FTD exists, it is advisable to obtain the entire FTD genetic panel, including GRN and MAPT.

In case of an ambiguous or positive FDG-PET scan, next steps can include investigation into molecular amyloid and tau biomarkers, if justified by the clinical context. Amyloid and tau biomarkers have a higher yield in younger individuals, if potential genetic implications exist, to determine whether anti-amyloid treatments are indicated or for potential access to clinical trials. Individual preferences and values must be considered in the decision to pursue investigation into more advanced biomarkers.

Throughout all investigations, it is crucial to remember that subtle or mildly abnormal findings, including borderline hypometabolism on FDG-PET, mild diffuse cortical volume loss, and modestly elevated NfL levels, can be related to the primary psychiatric illness and should not be overinterpreted as proof that the individual will experience further functional decline.

Knowledge Gaps and Future Directions

Chronic PPD are complex illnesses that have multifactorial effects on the brain. The clinical interpretability of dementia biomarkers in individuals with PPD is hindered by the lack of data on the effects these diseases have on the brain in older people. The clinical assumption that neurodegeneration biomarkers should be normal in older individuals has been shown repeatedly to be erroneous. There is a clear need for large research cohorts to study brain aging using biomarkers in people with chronic psychiatric disorders. In addition, several upcoming biomarkers (eg, blood-based AD markers [eg, pTau181, pTau217, pTau231], skin alpha-synuclein in LBD) need to be integrated into clinical care.²⁷ A wider range of people with psychiatric disorders must be included in research cohorts to improve clinical care and avoid discriminating against this population.