Skip to main content
Specific Cognitive/Behavioral Domains Predict Neuropsychiatric Symptoms in Severe Dementia
Rocky Mountain Psychological Association
  • William Rozum, Utah State University
  • Bryce Cooley, Utah State University
  • Alexandria Richens, Utah State University
  • Joshua Matyi, Utah State University
  • Elizabeth Vernon, Utah State University
  • JoAnn Tschanz, Utah State University
Document Type
Conference Poster
Salt Lake City, Utah
Publication Date
Background: Neuropsychiatric symptoms (NPS) have high prevalence in Alzheimer’s disease and related disorders (ADRD), with nearly 100% of individuals experiencing some type of symptom over the course of dementia (Tschanz et al, 2011). The occurrence of NPS is highly variable and fluctuates in severity (Tschanz et al., 2016). Their occurrence differs by type of dementia and increases over time (Kazui et al., 2016). Although risk factors for NPS in ADRD have been studied (e.g., Steinberg et al., 2014; Treiber et al, 2008), greater understanding of the nature of NPS and their triggers is needed to inform care management strategies (Gauthier et al., 2010). While much research has examined NPS in mild-to-moderate dementia, fewer studies have examined NPS in severe dementia. We investigated the cognitive correlates of NPS in patients with severe dementia in a community-based sample, examining whether impairments in specific cognitive or behavioral domains were more predictive of specific NPS. We hypothesized that poorer cognitive abilities would be associated with more severe NPS (e.g., agitation) and higher cognitive scores with affective symptoms in severe dementia. Methods: Eighty-nine (27%) out of 328 total participants of a longitudinal study of dementia progression (the Cache County Dementia Progression Study) met criteria for severe dementia: Mini-Mental State Exam (MMSE) score of ≤10 or Clinical Dementia Rating of 3 (severe). Forty-eight (54%) of these individuals completed the Severe Cognitive Impairment Profile (SCIP), which assessed the following domains: Comportment, Attention, Language, Memory, Motor, Conceptualization, Arithmetic, and Visuospatial abilities. NPS were assessed by caregiver report using the Neuropsychiatric Inventory (NPI). The NPI assesses delusions, hallucinations, depression, anxiety, irritability, apathy, agitation/aggression, judgement, aberrant motor behaviors, euphoria, sleep and appetite. Demographic information, overall health, place of residence (private home, assisted living facility and nursing home), and dementia duration were also assessed. NPI severity scores (intensity x frequency) were summed across domains to yield a total NPI score (Total NPI-12) and domain clusters of psychotic symptoms (hallucinations and delusions), affective symptoms (depression, anxiety, and irritability), apathy, and agitation/aggression were examined. Bivariate correlations between SCIP domain scores and Total NPI-12 and the domain clusters were examined. SCIP domain scores that were significantly correlated with NPI scores in bivariate analyses were entered into multiple regression models. Covariates tested included the age at which severe dementia criteria was met, the duration of dementia from age of onset, gender, place of residence, overall health and years of education. Results: Mean (SD) age and education were 86.23 (6.12) and 13.13 (3.13), respectively. Total NPI-12 scores showed significant correlations with the SCIP sub scores of comportment ( r = -0.36, p = 0.017) and memory (r = - 0.31, p = 0.047). Apathy significantly correlated with comportment (r = -0.38, p = 0.010) while agitation/aggression correlated with conceptualization (r = -0.41, p = 0.007), language (r = -0.36, p = 0.017), memory (r = -0.48, p = 0.001), and visuospatial ability (r = -0.31, p = 0.045). In multiple regression models (with inclusion of significant covariates), total NPI-12 scores were significantly associated with comportment (β = -1.32, SE = 0.56, p = 0.02); apathy was significantly associated with comportment (β = -0.01, SE = 0.02, p = 0.003); and agitation/aggression was significantly associated with memory (β = -0.43, SE = 0.12, p = 0.001). NPI affective and psychotic scores were not associated with any SCIP domains. Conclusion: In this sample of individuals with severe dementia, we found several cognitive or behavioral domains were associated with NPS. Poorer abilities in Comportment, which consisted of responses to social questions (e.g., greetings) were associated with more severe apathy, and poorer abilities in conceptualization, language, memory and visuospatial skills were associated with more severe agitation/aggression. With the latter, multiple regression models found only memory scores to independently predict agitation/aggression, reflecting moderate correlation between cognitive domains. Our results suggest that poor cognitive abilities may increase vulnerability to NPS, possibly as a result of impaired comprehension of activities and events in the environment. Cognitive testing may be useful to identify those at greatest risk for NPS. Furthermore, environmental manipulations that aim to decrease the complexity and therefore degree of stimulation for persons with dementia to a level more appropriate to their level of cognitive function may help reduce the occurrence of NPS in severe dementia.
Citation Information
William Rozum, Bryce Cooley, Alexandria Richens, Joshua Matyi, et al.. "Specific Cognitive/Behavioral Domains Predict Neuropsychiatric Symptoms in Severe Dementia" Rocky Mountain Psychological Association (2017)
Available at: