Addition of the Aβ42/40 ratio to the cerebrospinal fluid biomarker profile increases the predictive value for underlying Alzheimer's disease dementia in mild cognitive impairment

Abstract

Background: Cerebrospinal fluid (CSF) biomarkers have been used to increase the evidence of underlying Alzheimer's disease (AD) pathology in mild cognitive impairment (MCI). However, CSF biomarker-based classification often results in conflicting profiles with controversial prognostic value. Normalization of the CSF Aβ42 concentration to the level of total amyloid beta (Aβ), using the Aβ42/40 ratio, has been shown to improve the distinction between AD and non-AD dementia. Therefore, we evaluated whether the Aβ42/40 ratio would improve MCI categorization and more accurately predict progression to AD.

Methods: Our baseline population consisted of 197 MCI patients, of which 144 had a follow-up ≥ 2 years, and comprised the longitudinal study group. To establish our own CSF Aβ42/40 ratio reference value, a group of 168 AD-dementia patients and 66 neurological controls was also included. CSF biomarker-based classification was operationalized according to the framework of the National Institute of Aging-Alzheimer Association criteria for MCI.

Results: When using the core CSF biomarkers (Aβ42, total Tau and phosphorylated Tau), 30% of the patients fell into the high-AD-likelihood (HL) group (both amyloid and neurodegeneration markers positive), 30% into the low-AD-likelihood group (all biomarkers negative), 28% into the suspected non-Alzheimer pathophysiology (SNAP) group (only neurodegeneration markers positive) and 12% into the isolated amyloid pathology group (only amyloid-positive). Replacing Aβ42 by the Aβ42/40 ratio resulted in a significant increase in the percentage of patients with amyloidosis (42-59%) and in the proportion of interpretable biological profiles (61-75%), due to a reduction by half in the number of SNAP cases and an increase in the proportion of the HL subgroup. Survival analysis showed that risk of progression to AD was highest in the HL group, and increased when the Aβ42/40 ratio, instead of Aβ42, combined with total Tau and phosphorylated Tau was used for biomarker-based categorization.

Conclusions: Our results confirm the usefulness of the CSF Aβ42/40 ratio in the interpretation of CSF biomarker profiles in MCI patients, by increasing the proportion of conclusive profiles and enhancing their predictive value for underlying AD.

Keywords: Alzheimer’s disease; Aβ42/40 ratio; CSF biomarkers; Mild cognitive impairment.

Fig. 1

Comparison of MCI biomarker-based subgroups with clinical follow-up. Amongst mild cognitive impairment patients who progressed to Alzheimer’s disease during clinical follow-up (MCI-AD), the percentage of different biomarker-based subgroups was compared. Biomarker-based subgroups determined using either CSF Aβ42, t-Tau and p-Tau (gray bars) or Aβ42/40 ratio, t-Tau and p-Tau (black bars). LL low-AD likelihood, HL high-AD likelihood, IAP isolated amyloid pathology, SNAP suspected non-Alzheimer pathology; Aβ42 42-aminoacid isoform of amyloid beta, Aβ40 40-aminoacid isoform of amyloid beta

Fig. 2

Kaplan–Meier survival curves for probability of conversion to AD according to different MCI biomarker-based subgroups. Number of individuals at risk at each time interval shown below the graphs. MCI subgroups determined taking into account CSF t-Tau and p-Tau levels and either Aβ42 (a) or the Aβ42/40 ratio (b). Log-rank (Mantel–Cox) p < 0.001 for both. AD Alzheimer’s disease, LL low-AD likelihood, HL high-AD likelihood, IAP isolated amyloid pathology, SNAP suspected non-Alzheimer pathology