Standardizing "Rapid": Applying the Clinical Dementia Rating to Define Rapidly Progressive Dementia

Rapidly progressive dementia (RPD) accounts for 3%-4% of dementia cases yet presents a disproportionate clinical challenge given the wide differential diagnosis and the need to quickly identify potentially treatment-responsive causes. The lack of a standardized definition of RPD challenges generalization of published findings and precludes multisite studies. We sought to determine whether a standardized definition, based on the Clinical Dementia Rating (CDR), could reliably distinguish individuals with a broad spectrum of causes of RPD early in the symptomatic course.

RPD was defined as dementia (global CDR ≥1) developing within 1 year or incapacitation (global CDR ≥2) within 2 years of symptom onset. Patients with suspected RPD were referred from inpatient and outpatient settings at Mayo Clinic in Florida (Jacksonville, FL) and Washington University in St. Louis (Saint Louis, MO) and prospectively evaluated (RaPID cohort). Retrospective data from participants assessed across 46 research centers were also analyzed (National Alzheimer's Coordinating Center [NACC] data set). Characteristics and rates of progression (main outcome) were compared between typically progressive and RPD. Univariable and multivariable analyses were performed using Mann-Whitney U/Kruskal-Wallis (continuous) and Fisher exact/χ² tests (binary), respectively. Our definition of RPD was compared against definitions identified through a scoping review conducted using a validated 6-stage approach.

In the RaPID cohort, 185 of 248 (74.6%) patients met our CDR-based RPD definition (median [range] age-at-symptom onset: 68.8 years [22.0-90.7]; 45.4% female). Autoimmune (55/185, 29.7%), neurodegenerative (51/185, 27.6%), and prion diseases (35/185, 18.9%) predominated. Among 20,418 NACC participants with cognitive impairment, 836 (4.1%) met our RPD definition (median [range] age-at-symptom onset: 74 years [24-103]; 57.1% female). Neurodegenerative diseases predominated, especially Alzheimer disease (485/836, 58.0%). Individuals with RPD declined substantially faster (ΔCDR sum-of-boxes/year: RaPID, 14.4 ± 5.3 vs 4.3 ± 4.3, p < 0.001; NACC, 6.7 ± 3.1 vs 1.7 ± 1.4, p < 0.001). Our scoping review identified 22 studies that applied 10 distinct definitions of RPD. Uniquely, our CDR-based definition clearly distinguished individuals with a broad spectrum of causes of RPD, assessed across multiple sites/settings, supporting early diagnosis with minimal resource requirements.

The CDR-based definition of RPD effectively identified individuals with a broad spectrum of causes of RPD early in the symptomatic course. The need for training and time to reliably administer the CDR may limit application of the proposed definition of RPD. Validation in additional cohorts is required to support generalizability.