P-Tau217 Continues to Show Benefit as Predictor of Cognitive Decline in Preclinical Alzheimer Disease
Plasma P-tau217 was associated with cognitive decline across several cognitive tests in 2 different cohorts and with conversion to AD dementia in the large-scale, longitudinal BioFINDER-1 cohort.
Oskar Hansson, MD, PhD
A recently published prospective population-based prognostic study evaluating combinations of different plasma biomarkers showed that phosphorylated tau 217 (P-tau217) was the best marker to predict cognitive decline in patients with preclinical Alzheimer disease (AD). Investigators concluded that plasma P-tau217 may be used as a complement to cerebrospinal fluid (CSF) or PET for participation selection in clinical trials of novel disease-modifying therapies.
Senior investigator Oskar Hansson, MD, PhD, professor of neurology, Lund University and Skåne University Hospital, and colleagues evaluated data from 2 prospective longitudinal cohort studies: the Swedish BioFINDER-1 and the Wisconsin Registry for Alzheimer Prevention (WRAP). A total of 171 amyloid-ß-positive, cognitively unimpaired participants were included in the main analyses, with biomarkers such as P-tau181, P-tau217, P-tau231, glial fibrillary filament protein, and neurofilament light measured in plasma. Additional CSF biomarkers in the BioFINDER-1 cohort and PiB PET uptake in the WRAP cohort were also evaluated.
The Mini-Mental State Examination (MMSE) and modified Preclinical Alzheimer Cognitive Composite (mPACC) over a median of 6 years made up the primary outcome of the study, with conversion to AD dementia as a secondary end point. Models were adjusted for age, sex, years of education, apolipoprotein ε4 allele status, and baseline cognition. In the BioFINDER-1 cohort, after adjustment, plasma P-tau217 was found to be the strongest biomarker to predict cognitive decline in both the mPACC (R2 = 0.41 vs 0.23 for the covariates-only model; P <.001) and the MMSE (R2 = 0.34 vs 0.04 for the covariates-only model; P <.001), yielding significantly improved model fits. Additionally, P-Tau217 was the strongest CSF biomarker to predict cognitive decline for both tests (mPACC: R2 = 0.37; P <.001; MMSE: R2 = 0.24; P <.001).
The inclusion of plasma P-tau217 (ß = –0.098 [SE, 0.018]; P <.001), APOE4 status (ß = 0.110 [SE, 0.050]; P = .03), sex (ß = 0.090 [SE, 0.051]; P = .08) and baseline mPACC scores (ß = 0.078 [E, 0.019]; P <.001) was found to be the optimal biomarker combination to predict mPACC slopes. For the MMSE, the best combination model included plasma P-tau217 (ß = –0.400 [SE, 0.065]; P <.001), and CAS Aß42/40 (ß = 0.090 [SE, 0.058]; P = .13).
In total, 30 of the 118 amyloid-ß-positive individuals from the BioFINDER-1 cohort who were evaluated for dementia converted to AD dementia at follow-up. Baseline plasma P-tau217 was associated with significant conversion to AD dementia compared with no conversion or conversion to non-AD dementias (HR, 2.03%; 95% CI, 1.57-2.63; P <.001). Plasma P-tau217 was also the only biomarker included in an optimal survival model.
READ MORE: Solanezumab Fails Phase 3 A4 Study of Preclinical Alzheimer Disease
To validate findings from the BioFINDER-1 analysis, predictive models of cognitive decline in the WRAP cohort showed that both plasma P-tau217 (R2 = 0.13; P = .01) and amyloid PET (R2 = 0.10; P = .02) were associated with the mPACC and improved the model fit when compared with the covariates-only model. Using plasma P-tau217, baseline mPACC score, and APOE4 status from BioFINDER-1, findings from WRAP showed that only P-tau217 was a significant predictor of mPACC slopes (ß = –0.045 [SE, 0.015]; P = .005). Compared with the covariates-only model (R2 = 0.24), plasma P-tau217 (R2 = 0.29; P = .046) was also associated with MMSE slopes, but amyloid PET (R2 = 0.28; P = .07) was not.
In simulated trials of amyloid-ß-positive cognitively unimpaired individuals, the inclusion of P-tau217 produced substantial reductions in sample sizes. Using the BioFINDER-1 cohort, the relative sample sizes when using mPACC slopes as the outcome were 79% when including the 3 highest quartiles of baseline plasma P-tau217, 55% when including the 2 highest, and 42% when including the highest quartile.
"This finding suggests that plasma P-tau217 could increase the power of early-stage AD trials, which is a logical extension of how biomarkers have been integrated in previous AD trials," Hansson et al wrote. "Clinical trials are needed to assess whether specific treatment principles are actually effective. A possible caveat is that individuals with steeper declines in cognition could hypothetically have conditions that have progressed beyond certain disease events for which certain treatments are less effective."
