Neuropsychiatric Symptom Burden across Neurodegenerative Disorders and its Association with Function

Objective Neuropsychiatric symptoms (NPS) are prevalent in neurodegenerative disorders, however, their frequency and impact on function across different disorders is not well understood. We compared the frequency and severity of NPS across Alzheimer's disease (AD) (either with mild cognitive impairment or dementia), Cerebrovascular disease (CVD), Parkinson's disease (PD), frontotemporal dementia (FTD), and amyotrophic lateral sclerosis (ALS), and explored the association between NPS burden and function. Methods We obtained data from Ontario Neurodegenerative Disease Research Initiative (ONDRI) that included following cohorts: AD (N = 111), CVD (N = 148), PD (N = 136), FTD (N = 50) and ALS (N = 36). We compared the frequency and severity of individual NPS (assessed by the neuropsychiatric inventory questionnaire) across cohorts using generalized estimating equations and analysis of variance. Second, we assessed the relationship of NPS burden with instrumental (iADLs) and basic (ADLs) activities of living across cohorts using multivariate linear regression while adjusting for relevant demographic and clinical covariates. Results Frequency of NPS varied across cohorts (χ2(4) = 34.4, p < .001), with post-hoc tests showing that FTD had the greatest frequency as compared to all other cohorts. The FTD cohort also had the greatest severity of NPS (H(4) = 34.5, p < .001). Further, there were differences among cohorts in terms of the association between NPS burden and ADLs (F(4,461) = 3.1, p = 0.02). Post-hoc comparisons suggested that this finding was driven by the FTD group, however, the differences did not remain significant following Bonferroni correction. There were no differences among cohorts in terms of the association between NPS burden and IADLs. Conclusions NPS frequency and severity are markedly greater in FTD as compared to other neurodegenerative diseases. Further, NPS burden appears to be associated differently with function across neurodegenerative disorders, highlighting the need for individualized clinical interventions.


Abstract
Objective: Neuropsychiatric symptoms (NPS) are prevalent in neurodegenerative disorders, however, their frequency and impact on function across different disorders is not well understood. We compared the frequency and severity of NPS across Alzheimer's disease (AD) (either with mild cognitive impairment or dementia), Cerebrovascular disease (CVD), Parkinson's disease (PD), frontotemporal dementia (FTD), and amyotrophic lateral sclerosis (ALS), and explored the association between NPS burden and function.
Methods: We obtained data from Ontario Neurodegenerative Disease Research Initiative (ONDRI) that included following cohorts: AD (N = 111), CVD (N = 148), PD (N = 136), FTD (N = 50) and ALS (N = 36). We compared the frequency and severity of individual NPS (assessed by the neuropsychiatric inventory questionnaire) across cohorts using generalized estimating equations and analysis of variance. Second, we assessed the relationship of NPS burden with instrumental (iADLs) and basic (ADLs) activities of living across cohorts using multivariate linear regression while adjusting for relevant demographic and clinical covariates.
Results: Frequency of NPS varied across cohorts (χ 2 (4) = 34.4, p < .001), with post-hoc tests showing that FTD had the greatest frequency as compared to all other cohorts. The FTD cohort also had the greatest severity of NPS (H (4) = 34.5, p < .001). Further, there were differences among cohorts in terms of the association between NPS burden and ADLs (F (4,461) = 3.1, p = 0.02). Post-hoc comparisons suggested that this finding was driven by the FTD group, however, the differences did not remain significant following Bonferroni correction. There were no differences among cohorts in terms of the association between NPS burden and IADLs.
Conclusions: NPS frequency and severity are markedly greater in FTD as compared to other neurodegenerative diseases. Further, NPS burden appears to be associated differently with function across neurodegenerative disorders, highlighting the need for individualized clinical interventions.
Keywords aging, aging and memory, Alzheimer's disease, cohort study, geriatric psychiatry Introduction Neuropsychiatric symptoms (NPS) constitute a major challenge for patients with Alzheimer's disease (AD), cerebrovascular disease (CVD), Parkinson's disease (PD), frontotemporal dementia (FTD), and amyotrophic lateral sclerosis (ALS). 1,5 NPS commonly include apathy, depression, anxiety, irritability, agitation, aggression, delusions, hallucinations, as well as sleep and appetite disturbances. 1,5 NPS occur almost invariably at some point during the course of neurodegenerative illness. Among patients with dementia, the 5-year period prevalence of at least one symptom has been estimated to be up to 97%, with the most common being depression, anxiety, and apathy. 6 Moreover, the presence of these symptoms has been associated with accelerated disease progression, reduced quality of life, and greater disability. 1,7 Interestingly, the prevalence of particular neuropsychiatric symptoms differs between specific neurodegenerative disease cohorts. 2,8,9 Varied symptom presentations across cohorts also pose a significant challenge in the evaluation and management of NPS, as treatment needs to be individualized. 10 Importantly, NPS has been associated with cognitive impairment, with a higher symptom burden predictive of poorer cognitive function. 11,12 Specifically, depression and sleep disturbances in AD, while delusions and abnormal eating behaviours in FTD, were negatively associated with cognition. 12 Moreover, NPS in patients with AD and CVD have been associated with impairments in activities of daily living (ADLs). 8,13,14 However, most studies lack participant stratification into specific neurodegenerative disease cohorts. 13 Thus, data assessing NPS and their association with function across diverse, well-defined neurodegenerative disease cohorts is needed to better understand this relationship. 15 The primary objective of this study was to compare the burden of NPS among AD, CVD, PD, FTD, and ALS cohorts. We hypothesized that these groups will have distinct NPS frequency and severity. We further aimed to explore the association between NPS and function in each cohort while adjusting for demographic and other clinical factors.

Methods
Data used in this study are from Phase 1 of the Ontario Neurodegenerative Disease Research Initiative (ONDRI)a longitudinal observational cohort study addressing the phenotypic sequelae of degenerative cognitive impairment. The sample included participants with AD (mild cognitive impairment (MCI) or dementia due to AD), CVD, PD, FTD, or ALS. Detailed inclusion/exclusion criteria of ONDRI participants were reported previously. 16,17 For MCI, participants with MCI only due to AD were included and they met the National Institute on Aging-Alzheimer's Association (NIA-AA) core clinical criteria for amnestic single or multiple domain mild cognitive impairment. 18 Other causes of cognitive impairment were ruled out by standardized work up, including brain imaging. 16 All participants underwent comprehensive evaluations to assess cognition, function, and neuropsychiatric symptoms. Of the enrolled participants, only those providing neuropsychiatric symptom score, and one of either basic or instrumental activities of daily living score described below were included in analyses.

Neuropsychiatric Symptoms and Cognition
The neuropsychiatric inventory questionnaire (NPI-Q) was used to assess neuropsychiatric symptoms. 19 Study partners of participants reported on the presence and severity of 12 neuropsychiatric symptoms including delusions, hallucinations, agitation/ aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability, motor disturbances, nighttime behaviours, and appetite/eating. Total scores for symptom severity (maximum of 36) and partner distress (maximum of 60) were computed by adding the individual symptom scores. 20 Montreal cognitive assessment (MoCA) was used to assess cognition. 21

Function
Lawton-Brody instrumental activities of daily living (iADL) scale was used to measure participants' ability to function independently on instrumental activities of daily living. Study partners rated the participants' ability across eight iADL tasks, ranging from totally independent to totally dependent, including telephone use, shopping, food preparation, housekeeping, laundering, use of transportation, managing medications, and financial management, for a maximum possible total score of 23. 22 In the event that a given participant did not engage in an activity pre-morbidly, the activity's score was deducted from that participant's total possible score (maximum of 23). This produced a modified total score for each participant, by which each participant's score was divided to compute an iADL percent score.
The physical self-maintenance scale was used to measure participants' ability to function independently on basic activities of daily living (ADLs). Study partners rated the participants' ability to function independently across six basic ADLs: feeding, dressing, grooming, ambulation, bathing, and toileting, for a maximum possible total score of 24. 23 ADL percent scores were computed for each participant by dividing their total score by 24.

Motor Function
The ALS functional rating scale-revised (ALSFRS-R) was used in the ALS cohort, 24 and the Unified Parkinson's Disease Rating Scale (MDS-UPDRS, part III) was used in the PD cohort to assess motor symptoms. 25

Statistical Analyses
Statistical analyses were performed using SPSS (IBM Corp, SPSS v25). Chi-square or Fisher's exact test (expected cell count less than five) were used to compare categorical measures among cohorts. Standardized residuals (z scores) were evaluated for significant outcomes. For all continuous outcomes, means and standard deviations were computed, normality was assessed using the Shapiro-Wilk test and by visual inspection of data histograms and quantile-quantile plots. Generalized estimating equations (GEE) were used to compare NPS frequency among cohorts, by evaluating binary outcomes (presence of individual NPI-Q symptoms) for individuals within each cohort while accounting for correlation of multiple symptoms measure from the same individual. The GEE covariance structure was specified as "unstructured" as the structure of the within-subject covariance was unknown. One-way analysis of variance (ANOVA) was used to compare NPS severity among cohorts. As applicable, Kruskal-Wallis test was used for non-normally distributed continuous data. Post hoc analyses were completed using Bonferroni testing. To investigate the relationships among NPI-Q total score, MoCA and ADL/iADL scores, correlations were computed using Pearson's or Spearman's correlation coefficient as appropriate, first across combined patient cohorts and further within each cohort. To explore whether the dependence of ADL and iADL percent scores on NPI-Q differed between cohorts, we carried out multivariable linear regression with ADL or iADL score as the dependent variable, NPI-Q and cohort membership as the independent variables with an interaction term, and with age, education, and MoCA as additional independent variables. Subsequently, to explore the association between NPS and function within each cohort, independent of demographic variables and cognition, linear regression was performed with total ADL or iADL percent scores as the dependent variables and NPI-Q total score, age, education, and MoCA scores as the independent variables. AD cohort participants were divided into two groups (MCI or dementia) and analyzed according to diagnosis at the time of study initiation. ALSFRS-R and MDS-UPDRS were included as covariates to control for motor disease burden in ALS and PD cohorts. Bootstrapping using 1,000 samples was performed with all linear regressions to improve interpretability independent of distribution assumptions. Only participants with complete data across measures of interest were included in regressions. Statistical significance was set at α = 0.05.
Of the total sample, 111 (88%) of AD participants, 148 (92%) of CVD participants, 136 (97%) of PD participants, 50 (94%) of FTD participants, and 37 (93%) of ALS participants provided sufficient data for inclusion in this study. Of the 38 excluded participants, 36 were missing NPI-Q scores, and two were missing functional outcomes measures (both basic and instrumental activities of daily living). Of the missing data, three participants (7.8%) had missing data due to verbal refusal to provide data. The remainder of the missing data was for unspecified reasons (i.e., not related to an administrative error, technical challenges, or otherwise). The demographics for included participants are presented in Table 1.
Participants with FTD suffered from the greatest frequency of aggression, anxiety, apathy, disinhibition, irritability, euphoria, appetite changes, and motor disturbances. FTD and PD participants showed the greatest frequency of abnormal night-time behaviours. Notably, there were no significant differences among cohorts with respect to the frequency of depression, hallucinations, or delusions. Further details, with included participant counts, are provided in Supplementary  Table 1. Participants in the FTD cohort also experienced the greatest overall NPS severity (mean (SD) NPI-Q severity score, AD = 3.7 (4.0), ALS = 3.1 (3.7), FTD = 8.2 (6.2), PD = 3.5 (3.9), CVD = 3.1 (3.9); H (4) = 34.5, p < .001). Further, partners of participants with FTD reported the greatest degree of overall partner distress (mean (SD) NPI-Q distress score, AD      there was no significant correlation between NPI-Q and ADL percent score. Summary correlation data and number of included participants from each cohort are presented in Table 3.

Discussion
This study conducted a comprehensive evaluation of NPS and their association with function in large, well-defined cohorts of patients across a wide spectrum of neurodegenerative disorders. As hypothesized, we found that the frequency of NPS differed significantly between cohorts, with FTD demonstrating the greatest frequency of NPS. The FTD cohort also had the greatest severity of NPS. Further, there were differences between cohorts in terms of the association between NPS burden and ADLs, and this finding seemed to be largely driven by the FTD cohort. Cohort-stratified analyses revealed a significant association between NPS and activities of living (both basic and instrumental) in AD, FTD, PD, and CVD cohorts. When examined in the presence of covariates (age, education, cognition, and motor function), NPS alone was associated with iADLs in FTD and CVD, whereas NPS, age, and cognition were associated with iADLs in AD. NPS and age were together associated with iADLs in PD. Furthermore, NPS alone were associated with ADLs in FTD and CVD, and together with motor disease burden in PD. Motor disease burden alone was associated with ADLs in ALS. NPS and age were together associated with ADLs in AD.
In line with our work, previous studies have revealed differences in the frequency and severity of NPS among neurodegenerative disease cohorts. In one observational study, participants with FTD had more frequent appetite disturbances, disinhibition, and abnormal motor behaviours as    compared to AD and CVD. 26 This finding is consistent with our data, and in addition, we observed higher frequencies of nearly all other NPS in our FTD cohort as compared to other cohorts. Notably, the previous study had a smaller sample size and higher frequency of overall NPS relative to our sample. 26 In another large study, FTD was associated with greater overall NPS burden when compared to AD, CVD, and DLB cohorts. 27 While the authors did not present the frequency of individual NPS within each cohort at baseline, 27 these findings parallel the results observed in our data. Of note, the behavioural variant subtype was disproportionally represented in our FTD cohort, likely accounting for some of these differences, as previously suggested. 3,27 Relationship Between NPS and Function Interestingly, our data suggested that the dependence of ADL scores on NPS differed significantly between cohorts. While pairwise comparisons did not remain significant after correction, these data support the notion that with greater disease burden, as ADLs become affected, the relationship between NPS and function becomes increasingly clinically relevant. The most notable group differences were observed when comparing FTD cohort to ALS and CVD cohorts.
While cohort-level analyses demonstrated associations between NPS with iADL and ADL across all cohorts, this was not observed in the ALS cohort, possibly due to partial confounding by motor disturbance, consistent with previous data. 28 Further, our finding in the AD cohort that NPS, cognition, and age together were associated with iADL score, while NPS alone predicted ADL score, suggests that NPS are more closely linked with functional ability at more advanced stages of illness, when basic ADLs become compromised. Additionally, the association of NPS alone with function for both iADL and ADLs in FTD suggests that NPS are the main drivers of functional impairment in FTD throughout the duration of illness. Among PD participants, motor symptom severity demonstrated a significant relationship with ADL scores, but not with iADL scores, suggesting that motor symptoms play a greater role in functional outcomes as disease burden progresses and basic ADLs become compromised. Taken together, these results highlight the need for developing effective clinical interventions for patients experiencing NPS due to neurodegenerative disorders. These should include improved use of existing non-pharmacological and pharmacological approaches, and development of novel therapeutics to improve care for NPS. 1,29,33 Limitations The present study has several limitations. First, this present study was cross-sectional, and we are thus unable to draw conclusions about longitudinal predictions based on our models. Second, the FTD group included participants with bvFTD, PPA, CBS, and PSP, and stratified analyses by these subtypes was not completed due to limited sample sizes. This should be investigated in future studies with larger cohorts. Third, while our diagnostic assessment for patients included standardized criteria and neuroimaging, assessment of biomarkers was not incorporated. Fourth, patients recruited into the study were those with mild-to-moderate illness, potentially reducing the observed range of functional impairment and NPS severity. Therefore, the generalizability of our findings is limited to patients in early or mid-stages of disease. Fifth, unequal cohort sizes included in our data may have resulted in reduced statistical power.

Conclusion
The present study demonstrates significant differences in the frequency and severity of neuropsychiatric symptoms across neurodegenerative disease cohorts, with the greatest frequency and severity of NPS observed in FTD. Further, NPS burden appears to be associated differently with function across neurodegenerative disorders. These findings elucidate the importance of NPS and their impact on function across the spectrum of neurodegenerative disorders. Considering the irreversibility of cognitive decline, the effective and individualized management of NPS may improve function and other clinical outcomes such as patient and caregiver quality of life. 34 Magstim, and in-kind research accounts from Scientific Brain Training Pro. He participated in 2021 in an advisory board for Biogen Canada Inc. E.F. has received personal compensation for serving on a PSP scientific advisory board for Biogen, for serving as a section editor for NeuroImage Clinical, and for serving as a course director for the AAN annual meeting. M.F. is listed on a patent related to methods and kits for differential diagnosis of Alzheimer disease vs frontotemporal dementia using blood bio-