Abstract
BACKGROUND: In dementia research, affective neuropsychiatric symptoms (NPS) - depression, anxiety, and apathy - remain understudied. Improving strategies to accurately identify clinically relevant NPS is essential for more robust research. OBJECTIVES: We sought to determine how often objective metrics and clinical gestalt metrics agree on NPS presence or absence. We further sought to determine optimal cut-offs for affective NPS presence/absence using the Neuropsychiatric Inventory Questionnaire (NPI-Q) severity ratings. METHODS: We assessed agreement for NPS presence/absence among 5 different depression metrics, 4 anxiety metrics, and 2 apathy metrics via Jaccard indices using the National Alzheimer's Coordinating Centers (NACC) dataset. Analysis included exploring different NPIQ severity rating thresholds of >0, >1, >2, and 0 and >1. RESULTS: NPIQ cut-off >1 for presence and =0 for absence of an NPS led to the best agreement with other metrics. However, there was poor agreement for NPS presence across depression metrics (6%) and across anxiety metrics (7%). Choice of metric could greatly skew the frequency of an NPS being present. All 3 affective NPS were more common in Lewy Body Disorder compared to Alzheimer's Disease or Vascular Cognitive Impairment, regardless of metric. CONCLUSIONS: Though NPIQ severity rating cut-off choice should depend on study design, using a severity score of >1 for presence and =0 for absence may best fit clinical gestalt for affective NPS. Lewy Body Disorders present with more affective NPS than other common dementia etiologies. Future consensus on criteria for depression and anxiety syndromes in dementia may improve their identification.