Neuropsychiatric Symptoms and Functional Decline in Alzheimer's Disease and Lewy Body Dementia

Miguel Germán Borda, MD; Dag Aarsland, MD, PhD; Diego Alejandro Tovar-Rios, MSC; Lasse M. Giil, MD, PhD; Clive Ballard, MD, PhD; Maria Camila Gonzalez, MD; Kolbjørn Brønnick, PhD; Guido Alves, MD, PhD; Ketil Oppedal, PhD; Hogne Soennesyn, MD, PhD; Audun Osland Vik-Mo, MD, PhD


J Am Geriatr Soc. 2020;68(10):2257-2263. 

In This Article


Setting and Participants

This is a Norwegian longitudinal cohort study with yearly assessment of patients referred to dementia clinics in Hordaland and Rogaland counties, titled the Dementia Study of Western Norway (DemVest).[18] To select patients with mild dementia only, a Mini-Mental State Examination (MMSE) score of at least 20 or a Clinical Dementia Rating score of 1 was required for inclusion.[19] Exclusion criteria at study entry were moderate or severe dementia, delirium, previous bipolar or psychotic disorder, terminal illness, or a recently diagnosed major somatic disease, which could significantly impact cognition, function, or study participation.

For the purposes of this study, only people diagnosed with AD (n = 111) and LBD (n = 85) with information regarding 5-year functional status and NPSs were included. Twenty-six patients with other dementia diagnoses were excluded from this analysis, yielding a total of 196 participants. During the 5 years of study, 106 patients died and 24 were dropouts (lost during the follow-up or partially missing due to lacking Rapid Disability Rating Scale (RDRS-2) and/or Neuropsychiatric Inventory (NPI)) (Supplementary Appendix S1).


Dementia. The diagnosis of dementia was established according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, and was further classified as AD or LBD according to the National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer's Disease and Related Disorders Association. Dementia with Lewy bodies was determined according to the revised consensus criteria, and Parkinson's disease dementia (PDD) was determined according to the recommendations from the movement disorder society task force.[20–22] Caregivers completed the Informant Questionnaire on Cognitive Decline in the Elderly, a questionnaire shown to be a reliable and valid instrument to detect dementia,[23] and the clinician completed the Clinician Dementia Rating scale.[24] Two of the investigators independently applied the diagnostic criteria. In cases of disagreement, and in patients fulfilling more than one set of operationalized diagnostic criteria, the final ascertainment was made based on consensus. A pathological diagnosis was ascertained in 31 AD and 20 LBD patients, proving diagnostic accuracy above 80%.[25] Further information regarding the DemVest study can be found elsewhere.[18,26,27]

Functional Status. ADLs were evaluated using the first 13 activities from the RDRS-2: (1) eating, (2) making simple food (e.g., sandwiches), (3) cooking dinner and adhering to a diet, (4) mobilization, - inside/outside (with or without aids), (5) daily personal care (including brushing teeth, combing hair, and maintaining personal hygiene), (6) bathing/showering, (7) dressing (including finding clothes), (8) toilet usage (including occasional clothing and cleaning), (9) usage of mobile telephone, (10) buying food and other necessary items, (11) handling money and paying bills, (12) having a financial overview plan ahead and writing tax returns, and (13) taking medications as prescribed.[28] Each item was scored from 1 to 4 (alone = 1, with some help = 2, with much help = 3, and not performed = 4). The final ADL score was calculated by adding all the items and then dividing by 13, where higher scores indicate worse function.[28]

Neuropsychiatric Symptoms. The validated Norwegian version of the clinician-rated 12-item NPI was used to interview family or caregivers, and the nursing home (NH) version NPI–Nursing Home was used after participants were referred to NHs (65% of participants were living in a NH after 5 years). The items included in the NPI are delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability, motor disturbance, nighttime behaviors, and appetite/eating.[29,30] All assessments were completed by the medical research staff after a structured interview of the informant who had the most day-to-day contact with the patient. The 12 items were registered as present or not during the past 4 weeks, and if present, scored according to their frequency (score = 1–4) and severity (score = 1–3), with higher scores representing higher frequency/severity. Here, we report the frequency × severity score for the 12 individual items with a maximum possible score of 144. We evaluated the sum of frequency x severity scores at the time of diagnosis and its longitudinal course over 5 years.

Confounding Variables. We included demographic factors, such as sex and age, in analyses. Comorbidities were assessed using the Cumulative Illness Rating Scale (CIRS), a systematic scale of comorbidity based on patient and informant reports and medical records.[31] Cognition was evaluated using the MMSE. The MMSE 5 years' longitudinal course was used for adjustments.[19,27]

Statistical Analysis

The descriptive analyses were performed by estimating percentages for categorical variables, and means and standard deviations for quantitative variables and groups were compared using the Pearson chi-squared test and t-student test, respectively. To estimate the association between NPSs on the longitudinal functional trajectory over 5 years, we applied a random coefficient model to take into account the variability between patients with mild dementia and stratifying by diagnosis (i.e., AD and LBD).[32] We divided the NPI-total score into quartiles (baseline = first quartile = 0–5 (reference), second quartile = 6–16, third quartile = 17–30, and fourth quartile = 31–89) (longitudinal = first quartile = 0–7 (reference), second quartile = 7–16, third quartile = 16–30, and fourth quartile = 30–108). We used both NPI-total at baseline only and NPI-total assessed longitudinally for association with functional trajectories. Based on an a priori review of the literature, we adjusted the models for the following covariates: baseline age, sex, comorbidities at baseline (CIRS score), MMSE assessed longitudinally, and time. The trend over time was modeled using a quadratic trajectory. A random intercept and a random slope with an unstructured variance–covariance matrix were used to model the variance according to the lowest Bayesian Information Criterion (BIC). We centered time at 3 years, and all other continuous variables at their mean at baseline. We considered significance at P < .05 to evaluate the variables in the model. STATA 15® and R software were used to perform all statistical analyses.