Objective:

Develop and apply a standardized and reproducible definition of rapidly progressive dementia (RPD) in clinical and research settings. 

Background:

A lack of clear diagnostic criteria for RPD has resulted in wide-ranging definitions and wider-ranging estimates of RPD frequency (1-30%). A standardized and reproducible definition of RPD that clearly differentiates patients with RPD from individuals with “typically progressive dementia” is needed to promote collaborative multi-institutional research.

Design/Methods:

The Clinical Dementia Rating® (CDR) is a validated and reproducible measure of cognitive function derived via semi-structured interview with a patient and knowledgeable collateral source. We labelled patients as rapidly progressive if they progressed from CDR 0 (cognitively normal) to ≥1 (dementia) within 1 year or CDR 0 to ≥2 (complete dependence on others for activities of daily living) within 2 years. Proposed criteria were applied to patients referred for the evaluation of suspected RPD to Mayo Clinic Florida (Jacksonville, FL) and Washington University School of Medicine (St. Louis, MO), and to participants with dementia included within the National Alzheimer’s Coordination Center (NACC) dataset.

Results:

71/87 (81.6%) of Mayo Clinic patients and 84/169 (49.7%) of Washington University patients met the proposed definition of RPD (χ²=24.5, p < 0.0001). Neurodegenerative (including prion disease) and autoimmune diseases were the most common causes of RPD at both centers, accounting for 59/155 (38.1%) and 52/155 (33.5%) of total cases, respectively. 901/32211 (2.8%) patients in the NACC dataset met RPD criteria. Almost all cases were attributed to neurodegenerative diseases (894/901, 99.2%).The criteria selected a distinct population when considering time-to-CDR in the NACC dataset (Mann-Whitney U statistic: 131,000, p < 0.0001). 

Conclusions:

Our CDR-based RPD definition can identify patients with exceptional rates of decline across multiple clinics and research sites. Future studies will compare the performance of a CDR-based definitions versus other time- or test-based criteria for RPD.