Determinants of Changes in the Clinical Dementia Rating Scale: Reversion and Conversion.

Background: At the Knight Alzheimer Disease Research Center (ADRC), clinicians are blinded to cognitive data and prior assessments to ensure that ratings are free of circularity and examiner bias. It is common for cognitive status to fluctuate between normal and abnormal in longitudinal studies of aging and dementia, even when using the gold standard Clinical Dementia Rating (CDRâ) scale. CDR fluctuation is common when clinical symptoms are mild. It was hypothesized reverters, defined as individuals who change from CDR 0.5 to CDR 0, would have worse cognition, more abnormal biomarkers, and more likely to progress to CDR>0 compared to participants who remain unimpaired ("CN"). Method: Cognitively normal participants with at least three annual visits were included (n = 924, Table 1). Participants were classified as CN, reverters, and converters to stable dementia ("converters"). The Knight PACC was used to compare cognition between groups. Participants (n = 187) with cerebrospinal fluid or positron imaging tomography biomarkers were classified as amyloid "positive" or "negative". The visit of first CDR>0 was defined as the "index visit" and the visit prior was defined as the "pre‐index visit." Linear models evaluated relationships between biomarker status, APOE ε4 status, and cognition at the pre‐index and index visit. Logistic regression evaluated predictors of converting to a dementia status ("conversion") and reverting from CDR 0.5 back to CDR 0 ("reversion"). Result: Of 924 participants, 127 (14%) were reverters and 142 (15%) were converters. Converters had the lowest PACC scores of the three groups at the pre‐index and index visit, with reverters having slightly better PACC scores than converters but significantly worse scores than CN (Figure 1). Predictors of conversion and reversion included older age, male gender, APOE ε4 carriage, and worse PACC scores. Amyloid positivity did not predict conversion or reversion. Reverters exhibited more longitudinal cognitive decline compared to CN but the magnitude of decline was less than converters (Figure 2). Conclusion: CDR reversion is a risk factor for future conversion to CDR>0 and is associated with older age, APOE ε4 carriage, and poor cognitive performance. CDR reversion appears to occupy a transitional phase in disease progression between cognitive stability and consistent progression to stable dementia. [ABSTRACT FROM AUTHOR]