The clinical phenotype of psychiatric-onset prodromal dementia with Lewy bodies: a scoping review

When compared to Alzheimer’s dementia (AD), dementia with Lewy bodies (DLB) is associated with poorer outcomes, including increased mortality, caregiver burden and the need for earlier residential care [1]. These observations probably reflect the comorbid parkinsonian and neuropsychiatric features typical of DLB. The recently revised diagnostic criteria for DLB highlight an aggressive cognitive decline in combination with four core clinical features, namely parkinsonism, cognitive fluctuations, visual hallucinations and Rapid Eye Movement Sleep Behaviour Disorder (RBD) [2]. The presence of at least two of these core features or one core feature and an indicative biomarker (polysomnographic confirmation of REM sleep without atonia, abnormal iodine MIBG myocardial scintigraphy or reduced dopamine transporter uptake in the basal ganglia), allow for a diagnosis of probable DLB to be made clinically. In addition, the diagnostic criteria also recognise a range of supportive clinical features (e.g. constipation, hyposmia, and daytime somnolence) that can assist in making the diagnosis [2].

Given its mixed clinical phenotype, the accurate diagnosis of DLB can be delayed by several years [3]. This delay is likely to impact on successful disease-modifying strategies, which rely on a diagnosis at the earliest point in time [4]. Therefore, as seen with other neurodegenerative conditions, there has been an increasing emphasis on the identification of prodromal DLB. DLB is a synucleinopathy with a differential cell loss occurring across multiple neurotransmitters and neural pathways which can occur decades before clinical diagnosis [5]. Thus, prodromal DLB is likely to be characterised by a range of differing clinical presentations encompassing both cognitive and non-cognitive symptoms. To characterise these potential phenotypes, the International DLB Consortium proposed research criteria for prodromal DLB in 2020 [6]. They identified three subtypes of prodromal DLB characterised by (i) a mild cognitive impairment-onset; (ii) a delirium-onset; and (iii) a psychiatric-onset.

Whilst the evidence to support the research criteria proposed for prodromal mild cognitive impairment (MCI) with Lewy bodies (MCI-LB) was relatively robust, describing a non-amnestic phenotype often with associated RBD, the Consortium recognised that there was insufficient evidence to propose formal criteria for the delirium- and psychiatric-onset presentations.

Clearly, significant challenges exist when studying the symptomatic overlap between primary late-onset psychiatric disorders and the prodromal symptoms of DLB. For example, psychomotor slowing, antipsychotic-induced bradykinesia, and cognitive impairments are frequently seen in psychiatric conditions and may be very difficult to differentiate clinically from the early parkinsonism and neuropsychiatric impairment occurring in the context of an evolving DLB. Indeed, it has been reported that over 40% of patients who develop DLB present with a concurrent psychiatric symptom [7].

Despite the need to characterise the existence of a psychiatric-onset prodromal presentation of DLB, to the best of our knowledge, no previous reviews evaluating the current literature have been published. Thus, in this review, we aimed to identify those publications that have previously described patients who might satisfy the proposed psychiatric-onset DLB and determine the presence and frequency of clinical features and biomarkers that might allow us to better stratify the risk of their transition to dementia. It is hoped that establishing such a framework could help in proposing diagnostic guidelines and in the design of future prospective studies focussing on these patients.

As shown in Table 1, of the 14 articles included in the final review, two were cohort studies and another 12 were case reports, which accounted for a total of 69 patients initially presenting with a psychiatric illness, of which 52 were subsequently diagnosed with DLB.

The largest study identified in this review was conducted by Takahashi et al. [9], which recruited 35 patients with major depressive disorder (MDD) diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV-TR) [9] (mean age 71 ± 9.7, males 8, females 27). The primary purpose of this study was to evaluate the utility of the ventilatory response to hypercapnia (VRH) as a potential novel biomarker of autonomic dysfunction in patients that might have an early neurodegenerative condition. The study included patients (over 50 years) who required admission into a psychiatric unit with a Mini-Mental State Examination score of over 24/30 and the presence of bradykinesia but the absence of other extrapyramidal signs (i.e. scores of ≥ 1 for bradykinesia but 0 for the remaining 13 items on the Unified Parkinson’s Disease Rating Scale part III [10]). The initial presentation of MDD occurred in these patients prior to their recruitment into this study. The interval between initial presentation with MDD and recruitment into the study was not specified. Patients underwent a detailed neuropsychological assessment battery, as well as autonomic function testing that included ¹²³I-meta-iodobenzylguanidine (MIBG) scintigraphy, heart rate variability, orthostatic hypotension assessments and VRH. These patients were followed up annually and assessed for mood and cognitive state. They required at least three annual assessments to be included in the study. The examination period for this cohort ranged between 3 and 8 years during which 18/35 patients were diagnosed with DLB (mean age: 61 years; range 50–76 years; male: 4, female: 14).

The second cohort study evaluated here was conducted by Utsumi et al. [11]. This study described both the clinical and biomarker characteristics of 21 patients (mean age 62 ± 6.3 years, range 54–76 years, 7 males, 14 females) who presented with a psychotic illness requiring at least one hospital admission who went on to develop DLB. Patients needed to demonstrate no overt impairment in their activities of daily living or functional status for at least 2 years after their first psychotic episode. To be recruited into this cohort, patients needed to have at least one positive radiological biomarker that was indicative of DLB (i.e. (123I-N-ω-fluoropropyl-2β -carbomethoxy-3β-(4-iodophenyl) nortropane single photon emission computed tomography (FP-SPECT) and/or ¹²³I-meta-iodobenzylguanidine (MIBG) scintigraphy) or a brain perfusion SPECT, as a supportive biomarker of DLB. This study presented data on psychiatric symptoms, core features of DLB and biomarker results. These 21 patients developed DLB after a mean prodromal period of 9.1 (± 4.5, range 2–19) years at a mean age of 71 years (± 6.1, range 60–82).

The remaining 12 articles were case studies that described a total of 13 patients (5 females, 8 males). Their mean age at presentation of their first psychiatric illness was 69 years (± 12; range 53–88). This group had a mean prodromal period of 6.2 years (± 5.5, range 1–17) and were diagnosed with DLB at a mean age of 75 years (± 9.0, range 59–89).

Altogether, these 14 articles described 52 patients who developed DLB (mean age at initial psychiatric presentation 63 years, range 53–88, males 19: females 33). Takahashi et al. [9] did not report the interval between first psychiatric presentation and the diagnosis of DLB. However, for the remaining 34 patients, the mean time for conversion to DLB was 8 (± 5) years from the onset of first psychiatric symptoms to a diagnosis of DLB, which occurred at a mean age of 73 (± 7.5) years. Both patients who presented with the core clinical feature (CS1 and CS2, see below), transitioned to DLB within 1 year. The available data showed no significant difference in age at presentation or time to conversion from initial onset of psychiatric illness to the development of DLB between the males and females.

The study by Utsumi et al. [11] was the only manuscript that reported a comparator group of patients who presented with MDD but did not transition to DLB over the course of the study. In this study, 18 patients who converted to DLB, compared to the 17 who did not, were younger at the onset of their psychiatric illness, although this difference was not statistically significant (61 ± 8.9 years versus 69 ± 9.2 years). A female predominance was seen in both groups (13/17 for DLB non-converters versus 14/18 for converters).

This review reports the existing literature in relation to the possible existence of a psychiatric-onset presentation of prodromal DLB. To date, a total of 14 articles have reported 52 cases of non-demented patients presenting with an initial psychiatric diagnosis (> 50 years) who developed DLB over a variable period of 1–19 years (mean 8 years). Overall, most of these patients had recurring bouts of psychiatric symptoms that typically presented with depression, but commonly developed psychotic features before manifesting severe cognitive decline and the core features that led to the eventual diagnosis of DLB.

Strikingly, whilst there is a well-known male predominance in the prevalence of DLB, these studies highlight that patients presenting with prodromal psychiatric features were disproportionately female (63%). This female predominance is typically seen in late-onset depression and psychosis [24, 25] but there are few longitudinal studies that have tracked the progression of potential neurodegeneration in these cases. Indeed, it has recently been proposed that female DLB patients are more likely to present with psychotic symptoms compared to males who typically present more frequently with RBD [26].The majority of patients included in this review were Japanese (Cohort 1 and 2 totalling 39 patients, 75%) which accounted for the female predominance. Previous studies of Japanese DLB patients have noted a female predominance [26, 27], and therefore it is unclear whether ethnicity accounts for the unexpected gender difference. The data presented here also point towards psychiatric-onset DLB being a female-predominant phenotype compared to that of the MCI-LB prodrome, which has been shown to be more common in males [28].

This study highlights the difficulty in identifying patients with prodromal DLB who initially present with a psychiatric illness based purely on clinical clues. Core features tended to occur later in the course of the disease, and the majority of patients in the case studies initially were managed in the community. It is well known that mimics of the core symptoms reported in DLB can occur in patients with psychiatric disorders. For example, the presence of psychomotor retardation or paratonia can be mistaken for parkinsonism. Similarly, commonly used antidepressants such as selective serotonin reupdate inhibitors can precipitate RBD [29].

The utility of biomarkers in identifying these patients requires further evaluation. Whilst FP-CIT and MIBG were the most common biomarkers utilised in this study with sensitivities close to those demonstrated in previous studies [30], we are not aware of any work that has investigated the utility of these biomarkers in isolated depression or psychosis, where there has been a lower index of suspicion. Furthermore, studies to date have focussed on differentiating AD from DLB, rather than identifying patients who may progress to DLB from an isolated psychiatric disease.

This review has a number of limitations which impacts the confidence we can have in our conclusions. Firstly, the two cohort studies accounted for 75% (39/52) of patients and selection bias resulting from their respective inclusion criteria, may have skewed the presence or absence of clinical features (i.e. presence of bradykinesia in Cohort 1 and psychotic symptoms in Cohort 2). Furthermore there was inconsistent reporting relating to the timing and number of core features and biomarkers between studies. Interpretation of psychiatric features was also challenging, given the differences in terminology and reporting that has occurred over time with the evolution of psychiatric diagnostic criteria. Similarly, the lack of a standardised assessment battery and the reliance on clinical criteria may also have impacted on the ability to evaluate cases accurately. Together, these limitations emphasise the need for more high-quality and ideally multicentre prospective studies with larger case numbers undergoing standardised assessments. Unfortunately, given the lack of a comparison arm in most of the published studies, it is not possible to comment on the predictive ability of individual affective symptoms (depression, anxiety, apathy) in relation to future DLB risk. Future studies looking at patient groups with well-characterised affective symptoms measured by validated scales, but not necessarily carrying a formal diagnosis of anxiety or depression may also offer further insights into the psychiatric prodrome of DLB.

There have been no studies to date investigating the neuropathology of Prodromal DLB. Studies investigating the neuropathology of late-life depression have shown a mixture of Alzheimer's (Amyloid plaques), Lewy bodies and vascular pathology [31‐33], although none of these studies describe patients who had developed DLB. Work characterising the neuropathology of patients with psychiatric-onset DLB to determine the potential influence of mixed comorbidities (e.g. synucleinopathy, amyloid, tau) could provide a greater understanding of the underpinning neurobiological processes.

Despite the limitations, the results above provide some practical starting points that may be useful for clinicians to increase their awareness of this condition and assist recruitment and design of future studies. Firstly, the onset of new psychiatric symptoms in patients (especially females), over the age of 50 years should trigger the routine investigation for other core and supportive clinical features that might otherwise not be considered, such as parkinsonism, RBD, hyposmia, orthostatic hypotension, and constipation [34, 35]. Secondly, relevance should be given to such patients who experience multiple or recurrent bouts of psychiatric symptoms and/or delirium-like episodes. Thirdly, the transition from presenting with affective to psychotic features appears to be of particular importance in the risk of developing DLB and should raise the suspicion in treating clinicians for underlying neurodegeneration. In addition, the presence of any of the above features should trigger caution in physicians with respect to the prescribing of typical antipsychotics that may increase morbidity. Finally, we are entering a new era where we may be able to consider more sensitive biomarkers of disease from techniques such as α-synuclein seed amplification [36, 37], which may become more routine than established tests such as PSG, MIBG or FP-CIT(DaTscan).

In summary, much work is still needed to improve our understanding of this entity given the relative paucity small number of cases reported in the literature to date. Advancing this field will take a concerted international effort, and interdisciplinary collaboration between psychiatry, primary care and neurology will also be essential. It is hoped that a greater understanding of prodromal DLB will lead to earlier and more accurate diagnosis with the potential for more meaningful treatment strategies, as well as avoidance of morbidity associated with inappropriate neuroleptic prescription in this cohort.