Care partner‐informed meaningful change thresholds for the Clinical Dementia Rating‐Sum of Boxes for trials of early Alzheimer's disease

INTRODUCTION Consensus definitions of meaningful within‐patient change (MWPC) on the Clinical Dementia Rating‐Sum of Boxes (CDR‐SB) are needed. Existing estimates use clinician‐rated anchors in clinically diagnosed Alzheimer's disease (AD) populations. Incorporating the care partner perspective offers important insights, and evaluating biomarker‐confirmed cohorts aligns estimates with ongoing trials. METHODS Anchor‐based analyses were conducted to evaluate MWPC on the CDR‐SB in early AD (Tauriel; NCT03289143) using Caregiver Global Impression of Change in memory or daily activities. RESULTS Across time points and anchors, mean CDR‐SB changes associated with the “somewhat worse” category ranged from 1.50 to 2.12 in early AD, 1.07 to 2.06 in mild cognitive impairment‐AD, and 1.79 to 2.25 in mild AD. DISCUSSION The proposed ranges are appropriate to define meaningful progression on the CDR‐SB in similar cohorts and support the interpretation of treatment benefit through MWPC analyses. Thresholds should be calibrated to the context of use; lower/higher thresholds may be applicable in studies of earlier/later disease over shorter/longer durations. Highlights Within‐patient CDR‐SB change thresholds are provided using caregiver‐rated anchors. 1.5 to 2.5 points may be an appropriate range in early AD trials of similar durations. Cumulative distribution function plots illustrate the benefit of a given treatment. When selecting thresholds, the target population and study design should be considered.