Disease progression in dementia with Lewy bodies: A longitudinal study on clinical symptoms, quality of life and functional impairment

Abstract Background and Objectives Dementia with Lewy Bodies (DLB) is a heterogeneous disease, with variable signs and symptoms across multiple domains. We aimed to identify associations with rate of change in cognition, everyday functioning (IADL) and quality of life (QoL). Methods We included 121 DLB patients (69 ± 6 yrs, 14%F, MMSE: 25 ± 3) in our prospective cohort (follow‐up 2 ± 1 yrs). We described progression of symptoms and cognitive decline over time. Mixed models were used to investigate whether changes in symptoms were associated to changes in IADL (FAQ), QoL (QoL‐AD) and caregiver burden (ZBI). Last, we investigated whether baseline symptoms and biomarkers predicted decline in cognition (MMSE), IADL (FAQ) and QoL (QoL‐AD). Results Parkinsonism and RBD were most frequently present early in the disease course, while hallucinations were more likely to develop in a later stage. MMSE (annual change β ± SE = −2.06 ± 0.23), QoL‐AD (−1.03 ± 0.20), and FAQ (3.04 ± 0.30) declined over time. Increasing severity of clinical symptoms was associated to increases in FAQ, QoL‐AD and caregiver burden. Baseline clinical symptoms were not predictive for decline in these outcomes. By contrast, AD co‐pathology (CSF pTau/Aβ42 ratio) was associated to steeper decline in MMSE (−1.23 ± 0.54). Medial temporal atrophy (−0.81 ± 0.26) and global cortical atrophy (−0.73 ± 0.36) predisposed for decline in QoL‐AD. Conclusions Our findings imply that underlying disease processes, rather than clinical symptomatology aid in predicting decline. These findings are relevant for treatment strategies and the development of DLB specific outcome measures.

have been related to impairment in daily living (IADL). 2,3 DLB has an unfavorable prognosis when compared to Alzheimer's disease (AD). Patients with DLB have more rapid cognitive decline, higher mortality risk and earlier nursing home admission. [4][5][6][7] Former studies estimating disease trajectories in DLB were often limited to estimating global cognitive decline, while it is debatable whether this is the most relevant measure of disease burden from the patients' perspective. When providing treatment, the ultimate goal is to optimize patients' QoL and limit impairment in daily living. As such, it is essential to gain insight into how progression of symptoms affects these outcomes. Increased understanding of the contribution of symptoms to disease burden could give targets for symptomatic treatment.
In the present study, we described the progression in symptoms and cognitive domains over time. We aimed to assess how changes in symptoms are associated with changes in functional outcomes and QoL. Last, we aimed to identify factors that influence progression in functional outcomes (cognition, IADL and QoL).

and MCI due to Lewy
Bodies (MCI-LB) (n = 30) 8 from the DEvELOP cohort, 2 embedded within the Amsterdam Dementia cohort. 9 In short, patients were referred to Alzheimer Center Amsterdam and underwent diagnostic screening for dementia that included neurological, physical and neuropsychological evaluation, brain imaging, laboratory work and lumbar puncture. 9 Diagnoses were made in a multidisciplinary meeting. In the case of DLB or MCI-LB diagnosis, patients were invited to participate in DEvELOP. Exclusion criteria were severe physical or life-threatening conditions and nursing home admittance.
All patients gave written informed consent for use of their clinical data. The local medical ethics committee approved the study.

| Clinical follow-up
The first patient was included in 2016 and follow-up is still ongoing.
Patients were invited for annual follow-up, during which the clinical workup was repeated. All patients were invited for at least one 1 year follow-up and 92% of the patients that were still available for followup had 2 follow-up visits. For MCI-LB patients, their diagnosis was re-evaluated during these follow-up visits. Progression to dementia was defined as having two or more impaired cognitive domains at neuropsychological assessment, accompanied by interference in daily living.

| Clinical measures
Within DEvELOP, several questionnaires were administered to evaluate core and suggestive symptoms, IADL and QoL. The operationalization of symptoms and outcomes used in this study are summarized in Table 1 24,27 Time dependent tests were inverted so that higher scores reflect better performance. We calculated z-scores using the baseline data as a reference. Next, we created composite scores by calculating the average Z-score of each test in the domain.

| Neuropsychological assessment
Composite scores were only calculated if at least two tests in the corresponding domain were available.

| Baseline biological measures
MRI scanning was performed according to the standardized dementia protocol and visual assessment of atrophy was performed by experienced neuroradiologists. 9 Medial temporal lobe atrophy (MTA) was rated using coronal T1-weighted images on a 5-point scale (0-4). 28 Global cortical atrophy (GCA) was rated on axial images using a 4-point scale (0-3). 29 Baseline MRI was available for n = 109 (90%) patients.
Dopamine transporter imaging ( 123 FP-CIT(DAT)-SPECT) was performed on discretion of the clinician to confirm diagnosis and was available for n = 95 (79%) patients. The SPECT imaging protocol has been described in detail in a previous report. 31 Visual assessments and age-matched BRs were taken into account in determining whether the scan was normal or abnormal. 31 EEGs were recorded as standard screening of the memory clinic using a digital EEG system and software (Brain RT®; OSG b.v.). The EEG registrations were visually assessed by certified neurophysiologists and were scored according to a standardized visual rating scheme. 32 From this scheme, the 5 point scale to assess the severity of EEG abnormalities was used, in which a severity score of higher than 1 indicated an abnormal EEG. EEG was available for n = 81 (67%) patients.

| Statistical analyses
Analyses were performed in

| Longitudinal symptoms, functional outcomes and cognition
At baseline, mean UPDRS-III was 21 � 12 and 75% of patients ful- For suggestive symptoms, we found that at baseline, nearly half of the patients (49%) had urinary problems and 37% had signs of constipation. Suggestive neuropsychiatric symptoms were common, 60% of the caregivers reported apathy (NPI frequency * severity: 2.9 � 3.1), 45% reported anxiety (NPI frequency * severity:

| Changes in symptoms relate to changes in IADL and QoL
We applied LMM on repeated delta scores per individual to assess whether changes in symptoms were related to changes in functional outcomes ( Table 2)    per year, similar as found in previous studies in DLB 35,36 and in AD. 37 When looking at individual cognitive domains, we found prominent decline on tests that address attention and processing speed. Most previous studies evaluating cognitive decline in DLB were limited by the use of screening tools that are particularly sensitive for memory impairment, while such tools might fail to capture decline in cognitive domains that are most affected in DLB. 38 The trajectories of the MMSE might be an underrepresentation of the actual decline in cognitive functioning. Our findings underline the importance of carefully assessing all cognitive domains, especially attention.

| Factors that influence disease progression
Impaired attentional processing is also a crucial feature involved in the core symptom fluctuations. 39 Interestingly, although attention showed significant decline over time, the presence of fluctuations that the findings are dependent on the measurements that were used, but it is not clear whether all measures are equally responsiveness to change. There is no consensus on how to measure DLB specific symptomatology and progression. 38 A DLB-specific, standardized set of outcomes would highly improve evaluation of disease progression and would enhance the potential of comparison with other studies.
To conclude, we presented the trajectories of cognitive domains and symptoms in our prospective DLB cohort. We found clinically relevant associations with disease burden. Biological markers could help in predicting which patients are at risk for worse prognosis. We propose that DLB specific outcome measures include a combination of cognitive measures, specifically tests for attention and processing speed, and functional measures, like IADL and QoL, as those capture clinically relevant aspects of the disease spectrum and are sensitive for decline.