Skip to main content
. 2022 May 13;145(10):3546–3557. doi: 10.1093/brain/awac175

Table 4.

Accuracy of plasma p-tau181 in discriminating donors with and without autopsy-confirmed Alzheimer’s disease stratified by blood draw greater than and less than 5 years before death

AUC 95% CI P-value
Blood draw less than 5 years before death (n = 45)
Autopsy-confirmed Alzheimer’s disease
 Model 1 0.83 0.71–0.95 <0.01
 Model 2 0.85 0.74–0.97 <0.01
 Model 3 0.91 0.83–1.00 <0.01
Blood draw greater than 5 years before death, n = 58
Autopsy-confirmed Alzheimer’s disease
 Model 1 0.65 0.51–0.80 0.049
 Model 2 0.71 0.58–0.85 <0.01
 Model 3 0.77 0.65–0.90 <0.01

Model 1 examined plasma p-tau181 alone. Model 2 was based on predicted probabilities from binary logistic regression that included plasma p-tau181, age at death, years between last blood draw and death, sex and APOE ε4 status. Model 3 was also based on predicted probabilities from binary logistic regression that included Model 2 covariates in addition to global CDR score at time of blood draw.