Abstract
INTRODUCTION: Assessment of short-term intra- and inter-individual variability for Alzheimer's disease (AD) plasma biomarkers is essential for clinically relevant interpretation of biomarker levels. We hypothesized that the variability of plasma tau phosphorylated at threonine 217 (p-tau217) could be reduced by combining it with a tau marker, plasma brain-derived tau (BD-tau), as the p-tau217/BD-tau ratio. METHODS: Three consecutive blood samples were collected from memory clinic patients within 36 days. Patients were dichotomized by cerebrospinal fluid (CSF) amyloidosis (Aβ+ = 29, Aβ- = 18). We compared intra- and inter-individual variability (coefficient of variation [CV]) in the plasma p-tau217/BD-tau ratio with p-tau217 alone and tested if kidney function, glycated hemoglobin, and body mass index (BMI) affected the variability. Finally, we compared the p-tau217/BD-tau ratio with CSF p-tau217. RESULTS: We found that for Aβ+ individuals, the intra-individual variability of the plasma p-tau217/BD-tau ratio (CV 7.1% [95% confidence interval {CI} 5.6;8.4]) was lower than for p-tau217 alone (CV 9.4% [95% CI 7.4;11.5]). At the group level, the variability in the p-tau217/BD-tau ratio was reduced in both Aβ+ (CV 15.1% [95% CI 11.7;18.7]) and Aβ- (CV 18.4% [95% CI 13.0;23.8]) individuals compared to p-tau217 alone (Aβ+ CV 19.1 [15.0;23.4], Aβ- 27.1 [18.4;36.0]). Adjusting for estimated glomerular filtration rate, hemoglobin A1C, and BMI further reduced the inter-individual variability of p-tau217/BD-tau in the Aβ+ group. CSF p-tau217 showed higher correlation with plasma p-tau217/BD-tau (rho = 0.53, p = 0.0005) than with p-tau217 alone (rho = 0.37, p = 0.02). DISCUSSION: Our findings suggest that using the ratio of plasma p-tau217 to plasma BD-tau and accounting for the influence of peripheral confounders improves biomarker stability, which is important for the interpretation of longitudinal biomarker changes and to prevent misclassification. HIGHLIGHTS: The plasma p-tau217/BD-tau ratio lowered short-term intra- and, especially, inter-individual variability compared to the variability in plasma p-tau217 alone.Plasma BD-tau did not correlate with eGFR, HbA1c, or BMI, while plasma p-tau217 was significantly negatively associated with BMI.Adjusting for eGFR, HbA1c, and BMI further reduced the inter-individual variability of p-tau217/BD-tau.Additionally, CSF p-tau217 correlated better with plasma p-tau217/BD-tau than with p-tau217 alone.