Conflicting cerebrospinal fluid biomarkers and progression to dementia due to Alzheimer’s disease

Panagiotis Alexopoulos, Lukas Werle, Jennifer Roesler, Nathalie Thierjung, Lena Sophie Gleixner, Igor Yakushev, Nikolaos Laskaris, Stefan Wagenpfeil, Philippos Gourzis, Alexander Kurz, Robert Perneczky, for the Alzheimer’s Disease Neuroimaging Initiative (adni)Alzheimer 2016 · DOI

Abstract

The risk of progression to dementia due to AD differs between patients with different CSF biomarker constellations.

Introduction

Mild Cognitive Impairment (MCI) often serves as a precursor to Alzheimer’s Disease (AD), yet the trajectory for individual patients remains highly variable. Clinicians typically rely on a "pathological CSF signature"—characterized by decreased Amyloid-beta 1–42 ($A\beta_{42}$) and elevated total Tau (t-Tau) and phosphorylated Tau (p-Tau)—to predict progression. While diagnostic guidelines are clear for fully positive or negative profiles, nearly 45% of patients present with "conflicting" biomarker constellations, falling into a clinical "grey zone." This study investigates how these mixed profiles influence the risk of progressing to AD dementia.

Methods and Subgroup Classification

Researchers analyzed a multicentre cohort of 469 patients from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). To account for "borderline" cases, a 20% standard deviation range around standard cut-offs was defined. Patients were categorized into five subgroups based on their biomarker "constellation":

  1. $MCI_{Non+}$: No positive biomarkers.
  2. $MCI_{All+}$: All three markers positive.
  3. $MCI_{A\beta+}$: Positive $A\beta_{42}$ but negative/borderline Tau.
  4. $MCI_{A\beta+T+}$: Positive $A\beta_{42}$ plus one positive Tau marker.
  5. $MCI_{T+}$: Negative/borderline $A\beta_{42}$ but at least one positive Tau marker (often termed SNAP: Suspected non-AD pathophysiology).

The complex 3D biomarker data was visualized in 2D using Non-negative Matrix Factorization (NNMF), allowing researchers to see how these subgroups cluster and overlap.

NNMF visualization of MCI subgroups based on CSF biomarker constellations.

Key Results: The Hierarchy of Risk

The study followed patients for up to 8 years, with 159 individuals progressing to AD dementia. Cox regression models revealed a clear spectrum of risk. The $MCI_{All+}$ group faced the highest risk (Hazard Ratio [HR] = 4.399) compared to the $MCI_{Non+}$ reference.

Notably, Amyloid-beta positivity emerged as the primary driver of progression. Patients with positive $A\beta_{42}$ but negative Tau ($MCI_{A\beta+}$) had a risk (HR = 2.697) statistically similar to those who also had one positive Tau marker ($MCI_{A\beta+T+}$, HR = 2.887).

Cumulative survival curves for progression to Alzheimer’s dementia across biomarker subgroups.

The survival analysis illustrated in Figure 6 shows that while $MCI_{All+}$ patients progress most rapidly, any Amyloid-positive constellation significantly increases risk compared to the $MCI_{Non+}$ and $MCI_{T+}$ (SNAP) groups.

Discussion and Clinical Implications

The high prevalence of conflicting biomarker results (45%) suggests that the "typical" AD signature is far from universal in the early stages of cognitive decline. The findings emphasize that Amyloid-beta is the "anchor" for risk; even without signs of neuronal injury (Tau), $A\beta_{42}$ positivity confers significant risk. Conversely, the $MCI_{T+}$ group showed a lower risk of progressing to Alzheimer's-specific dementia, suggesting their cognitive decline may involve non-AD pathologies like vascular disease.

While limited by the lack of histopathological confirmation and the specific nature of the ADNI cohort, these results advocate for updated diagnostic criteria. Future guidelines should explicitly incorporate these conflicting constellations to provide more accurate prognostic "likelihood grades" for patients in the clinical grey zone.

Summary updated

Abstract

The risk of progression to dementia due to AD differs between patients with different CSF biomarker constellations.
BibTeX
@inproceedings{alexopoulos2016conflicting,
  author = {Panagiotis Alexopoulos and Lukas Werle and Jennifer Roesler and Nathalie Thierjung and Lena Sophie Gleixner and Igor Yakushev and Nikolaos Laskaris and Stefan Wagenpfeil and Philippos Gourzis and Alexander Kurz and Robert Perneczky and for the Alzheimer’s Disease Neuroimaging Initiative (ADNI)},
  title = {{Conflicting cerebrospinal fluid biomarkers and progression to dementia due to Alzheimer’s disease}},
  booktitle = {Alzheimer},
  year = {2016},
  doi = {10.1186/s13195-016-0220-z},
}
Figure