Cerebrospinal Fluid Biomarkers in Cerebral Amyloid Angiopathy: New Data and Quantitative Meta-Analysis

Abstract

Background: To evaluate the diagnostic accuracy of cerebrospinal fluid (CSF) biomarkers in patients with probable cerebral amyloid angiopathy (CAA) according to the modified Boston criteria in a retrospective multicentric cohort.

Methods: Beta-amyloid 1-40 (Aβ40), beta-amyloid 1-42 (Aβ42), total tau (t-tau), and phosphorylated tau 181 (p-tau¹⁸¹) were measured in 31 patients with probable CAA, 28 patients with Alzheimer's disease (AD), and 30 controls. Receiver-operating characteristics (ROC) analyses were performed for the measured parameters as well as the Aβ42/40 ratio to estimate diagnostic parameters. A meta-analysis of all amenable published studies was conducted.

Results: In our data Aβ42/40 (AUC 0.88) discriminated best between CAA and controls while Aβ40 did not perform well (AUC 0.63). Differentiating between CAA and AD, p-tau¹⁸¹ (AUC 0.75) discriminated best in this study while Aβ40 (AUC 0.58) and Aβ42 (AUC 0.54) provided no discrimination. In the meta-analysis, Aβ42/40 (AUC 0.90) showed the best discrimination between CAA and controls followed by t-tau (AUC 0.79), Aβ40 (AUC 0.76), and p-tau¹⁸¹ (AUC 0.71). P-tau¹⁸¹ (AUC 0.76), Aβ40 (AUC 0.73), and t-tau (AUC 0.71) differentiated comparably between AD and CAA while Aβ42 (AUC 0.54) did not. In agreement with studies examining AD biomarkers, Aβ42/40 discriminated excellently between AD and controls (AUC 0.92-0.96) in this study as well as the meta-analysis.

Conclusion: The analyzed parameters differentiate between controls and CAA with clinically useful accuracy (AUC > ∼0.85) but not between CAA and AD. Since there is a neuropathological, clinical and diagnostic continuum between CAA and AD, other diagnostic markers, e.g., novel CSF biomarkers or other parameters might be more successful.

Keywords: Alzheimer’s dementia (AD); Boston criteria; cerebral amyloid angiopathy (CAA); cerebrospinal fluid (CSF); high-precision electro-chemiluminescence immunoassay (ECLIA).

FIGURE 1

Flow chart of study selection for the meta-analysis.

FIGURE 2

Box- and scatterplots of the CSF biomarker values in this study per diagnostic group. (A) Aβ40, (B) Aβ42, (C) t-tau, and (D) p-tau¹⁸¹. Red dots indicate the pathologically confirmed CAA samples. All concentrations in pg/ml. The box encompasses 50% of the samples and the whiskers extend 1.5 quartiles to each side of the box. Significant differences are indicated including p-values.

FIGURE 3

Receiver-operating characteristics curves for the CSF parameters in our study and the meta-analyses. (A) ROC plots of Aβ40, Aβ42, t-tau, p-tau¹⁸¹, and Aβ42/40 in this study for: controls vs. AD, controls vs. CAA, AD vs. CAA. (B) The same plots as in (A) for the meta-analysis including this study. (C) And excluding this study. CSF parameters corresponding to the lines are color-coded at the bottom of the plots. Solid lines: parameter with the highest AUC and parameters with an AUC not significantly different from the highest AUC. Dashed line: parameters with a significantly smaller AUC compared to the one with the highest AUC. Text insert: AUC, 95% confidence interval of the AUC, sensitivity, specificity of the parameter with the highest AUC and, if applicable name of additional parameter with identical AUC.