Study partner profile effects on CDR-SB change in anti-amyloid therapy evaluation

Area of Science:

• Neurology and Clinical Trials
• Biomarker Research in Neurodegenerative Diseases

Background:

• The Clinical Dementia Rating Sum of Boxes (CDR-SB) is a key outcome measure in anti-amyloid therapy (AAT) clinical trials.
• CDR-SB integrates data from both participants and their study partners.
• The influence of study partner characteristics on CDR-SB scores and their 18-month change interpretation in Alzheimer's disease (AD) trials is not fully understood.

Purpose of the Study:

• To quantify the impact of study partner characteristics and simulated follow-up imbalance on 18-month CDR-SB change.
• To assess the reliability of CDR-SB as an outcome measure in AD trials, particularly for AAT evaluation.

Main Methods:

• Utilized the NACC Uniform Data Set for an AD-primary early symptomatic cohort (n=15,061).
• Employed linear mixed-effects calibration models and Monte Carlo simulations to estimate CDR-SB change components.
• Analyzed observed profile changes, simulated follow-up imbalance, and tipping-point scenarios in AD-primary, amyloid-positive, and trial-like cohorts.

Main Results:

• Observed study partner profile changes resulted in negligible cohort-level CDR-SB differences (mean 0.0014 points).
• Simulated follow-up imbalance (10-50% reassignment) generated small differences (0.014 to 0.071 points).
• Achieving a 0.45-point difference (Clarity AD benchmark) required >100% net imbalance, feasible in 48% of iterations; biomarker-confirmed cohorts showed less stability due to imprecision.

Conclusions:

• Study partner profile variations and simulated imbalance had a minimal impact on CDR-SB scores in a large AD-primary cohort.
• The observed CDR-SB differences were small relative to established benchmarks like the Clarity AD trial.
• Systematic collection of study partner data and sensitivity analyses are recommended for observational and external-comparator studies evaluating AAT.