By Makoto Ishii, MD, PhD
Assistant Professor of Neuroscience and Neurology, Feil Family Brain and Mind Research Institute, Department
of Neurology, Weill Cornell Medical College
SYNOPSIS: In individuals with subjective cognitive decline, multiple biomarkers of neurodegeneration were found to add predictive values beyond amyloid and tau biomarkers; however, the various neurodegeneration biomarkers were not equivalent and should not be used interchangeably.
SOURCE: Ebenau JL, Pelkmans W, Verberk IMW, et al. Association of CSF, plasma, and imaging markers of neurodegeneration with clinical progression in people with subjective cognitive decline. Neurology 2022;98:e1315-e1326.
Abnormal amyloid-beta aggregation, neurofibrillary tau tangles, and neurodegeneration are the classic hallmarks of Alzheimer’s disease (AD), and, until recently, these pathological changes could be identified only in postmortem brains. With the development of fluid and imaging biomarkers, these pathological changes now can be identified in vivo in individuals, even prior to cognitive symptoms. This has led to the establishment of the ATN classification (A: amyloid-beta; T: tau; N: neurodegeneration), where amyloid-beta pathology can be measured by cerebrospinal fluid (CSF) amyloid-beta or amyloid positron emission tomography (PET) and tau pathology by CSF phosphorylated tau or tau PET.
Although amyloid-beta and tau biomarkers are rapidly becoming established, there are a variety of proposed biomarkers of neurodegeneration, including atrophy on magnetic resonance imaging (MRI), hypometabolism on fluorodeoxyglucose (FDG) PET, CSF total tau, and blood-based biomarkers, such as neurofilament light (NfL) and glial fibrillary acidic protein (GFAP). Since the neurodegeneration biomarkers used in the ATN classification vary considerably and may reflect different aspects of neurodegeneration, it is not clear if the different neurodegeneration biomarkers are equivalent and, therefore, interchangeable. Importantly, whether neurodegeneration markers have predictive value for clinical progression beyond amyloid-beta and tau biomarkers has not been established.
Ebenau and colleagues sought to address these gaps in our knowledge by evaluating neurodegeneration biomarkers in a longitudinal study of 401 individuals with subjective cognitive decline from the Alzheimer Center Amsterdam. Structural MRI 3D T1-weighted images, non-fasting blood samples, and lumbar puncture for CSF analysis were obtained. Mini-Mental State Examination (MMSE) was assessed annually and served as the longitudinal measure of global cognition. The following neurodegeneration biomarkers were examined: CSF total tau, medial temporal atrophy visual rating on MRI, hippocampal volume on MRI, serum NfL, and serum GFAP.
The authors found that the various neurodegeneration biomarkers were modestly to moderately correlated (range, -0.28 to 0.58). Serum NfL and GFAP were strongly correlated (range, 0.58; P < 0.01), while CSF phosphorylated tau and total tau were very strongly correlated (range 0.89; P < 0.001). Adjusting for age and sex resulted in a drastic reduction in the coefficients, but the correlation between CSF phosphorylated tau and total tau remained very strong. Using uncorrected models, all neurodegeneration biomarkers predicted clinical progression to mild cognitive impairment (MCI) or dementia. Adding covariates of age, sex, CSF amyloid-beta, and CSF phosphorylated tau to the model attenuated all hazard ratios (HR).
However, hippocampal volume, serum NfL, and serum GFAP added significant predictive value beyond the amyloid-beta and phosphorylated-tau biomarkers (HR, 1.52 [95% confidence interval (CI), 1.11 to 2.09]; HR, 1.51 [95% CI, 1.05 to 2.17]; HR, 1.50 [95% CI, 1.04 to 2.15], respectively). CSF total tau was excluded from the final analysis because of its collinearity with CSF phosphorylated tau. Finally, although CSF total tau, hippocampal volume, and serum GFAP all predicted MMSE slope or cognitive decline over time, only hippocampal volume added predictive value beyond amyloid-beta and phosphorylated tau biomarkers.
Although the recent advances in biomarkers to classify AD and predict clinical progression in individuals with no or minimal clinical symptoms have been truly remarkable, there is room for improvement. In this present study, the authors found that using neurodegeneration biomarkers of hippocampal volume, serum NfL, or serum GFAP improved prediction of clinical progression beyond simply using amyloid-beta and phosphorylated tau biomarkers. An important finding was that not all neurodegeneration biomarkers added predictive value. Medial temporal visual rating scores were found to be too crude a measure to accurately predict decline in cognitively normal individuals. Moreover, CSF total tau commonly is used as a neurodegeneration biomarker, but CSF total tau may not be an appropriate independent biomarker of neurodegeneration since it strongly correlated with CSF phosphorylated tau.
A major strength of this study is the use of multiple neurodegeneration biomarkers from different modalities in a relatively large study population. However, there are notable limitations. As noted by the study authors, the list of neurodegeneration markers examined was not exhaustive and notably did not include FDG-PET or other MRI atrophy measures. Furthermore, the study population consisted of individuals with subjective cognitive decline presenting at a memory clinic. This may limit the ability to generalize the findings to other populations. Also, many of the biomarkers used do not have optimal cutoff values, which may result in different results, depending on which cutoff values are used. Until a validated cutoff for the biomarkers is established, this may limit the general use of these biomarkers on an individual level.
Finally, a low percentage of individuals (16%) had clinical progression to MCI or dementia during this study. This was attributed by the authors to the relatively short follow-up duration of 3.8 years and the overall younger age (mean 60.9 years). Future studies with longer follow-up duration and older study population will be needed to verify and validate the findings from this study.
Despite any limitations, this study eloquently highlights the importance of selecting appropriate neurodegeneration biomarkers when using the ATN framework and that the different neurodegeneration biomarkers should not be treated as equivalent. Unfortunately, this study could not find a single “most suitable” biomarker for neurodegeneration, which would be useful when bringing these biomarkers to more general clinical practice. Finally, identifying additional biomarkers beyond the ATN framework, including vascular, inflammatory, and metabolic biomarkers, should remain a research priority, since these other classes of biomarkers could add additional diagnostic and predictive value beyond ATN.