Repeated plasma p-tau217 measurements to monitor clinical progression heterogeneity

Abstract

Introduction: Heterogeneity of clinical progression in Alzheimer's disease (AD) complicates the assessment of disease progression and treatment effects in trials. This study evaluates the potential of plasma phosphorylated tau-217 (p-tau217) to capture this heterogeneity.

Methods: We used k-means clustering to analyze cognitive trajectories in amyloid beta -positive (Aβ+) cognitively normal (CN) and mild cognitive impairment (MCI) participants from two independent cohorts. Cohort 1 included 186 participants (71 CN, 115 MCI; 507 observations) and Cohort 2 included 207 participants (64 CN, 144 MCI; 781 observations), both with up to 10 years of follow-up.

Results: Three progression clusters emerged in both cohorts: stable cognition, slow decline, and rapid decline-each including cases initially classified as CN or MCI. Baseline plasma p-tau217 was linked to progression risk in both cohorts, whereas longitudinal increases in Cohort 1 were steepest in rapid decliners.

Discussion: Plasma p-tau217 may aid in capturing clinical heterogeneity and support stratification and monitoring of disease progression in clinical trials.

Highlights: k-Means found stable, slow, and rapid cognitive decline clusters in amyloid beta-positive (Aβ+) cases. Higher baseline plasma phosphorylated tau-217 (p-tau217) levels predicted faster cognitive decline. Longitudinal increases in plasma p-tau217 were steepest in rapid decliners. Plasma p-tau217 tracks clinical progression heterogeneity in Aβ+ cases. Cognitive stage and amyloid alone may miss severity and risk in early-stage Alzheimer's disease.

Keywords: Alzheimer's disease; amyloid; cerebrospinal fluid; clinical heterogeneity; clinical progression; clinical trials; plasma; p‐tau217.

FIGURE 1

Illustration of the cognitive clusters derived from k‐means in both Cohort 1 and Cohort 2. (A) Cohort 1 and (B) Cohort 2 illustrate trajectories over time for each cluster. (C) Cohort 1 and (D) Cohort 2 illustrate trajectories over time for each cluster split by baseline cognitive stage (cognitively normal [CN]/subjective cognitive decline [SCD] or mild cognitive impairment [MCI]). Bands associated with each regression line illustrated 95% confidence intervals. (E) Cohort 1 and (F) Cohort 2 illustrate numbers and percentages of CN/SCD and MCI at baseline within each cluster.

FIGURE 2

Between cluster comparison of plasma phosphorylated tau‐217 (p‐tau217) in Cohort 1 and Cohort 2. (A) Cohort 1 and (B) Cohort 2 show between‐cluster comparisons of plasma p‐tau217, with red dots indicating means. Cohort 1 utilized the University of Gothenburg in‐house p‐tau217 assay, whereas Cohort 2 utilized a commercially available kit by Janssen. Both used Quanterix Simoa HD‐X setups. All statistical tests were two‐sided and false‐discovery rate was used to adjust for multiple comparisons in each cohort. Due to differing assays and the range of pg/mL between assays, (D) shows the similarities of mean fold changes for cognitive decline clusters as compared to the stable cognition clusters in both cohorts. (C) Shows steeper plasma p‐tau217 increases over time in slow‐ and rapid‐decline clusters compared to the stable cognition cluster, with 95% confidence intervals. Unstandardized beta coefficients (b) illustrate individual slopes for each cluster. Created in BioRender. Kirsebom, B. (2025) https://BioRender.com/o54z804.