Neurofilament light chain in spinal fluid and plasma in multiple system atrophy: a prospective, longitudinal biomarker study

Abstract

Purpose: There is a critical need for reliable diagnostic biomarkers as well as surrogate markers of disease progression in multiple system atrophy (MSA). Neurofilament light chain (NfL) has been reported to potentially meet those needs. We therefore sought to explore the value of NfL in plasma (NfL-p) in contrast to cerebrospinal fluid (NfL-c) as a diagnostic marker of MSA, and to assess NfL-p and NfL-c as markers of clinical disease progression.

Methods: Well-characterized patients with early MSA (n = 32), Parkinson's disease (PD; n = 21), and matched controls (CON; n = 15) were enrolled in a prospective, longitudinal study of synucleinopathies with serial annual evaluations. NfL was measured using a high-sensitivity immunoassay, and findings were assessed by disease category and relationship with clinical measures of disease progression.

Results: Measurements of NfL-c were highly reproducible across immunoassay platforms (Pearson, r = 0.99), while correlation between NfL-c and -p was only moderate (r = 0.66). NfL was significantly higher in MSA compared with CON and PD; the separation was essentially perfect for NfL-c, but there was overlap, particularly with PD, for NfL-p. While clinical measures of disease severity progressively increased over time, NfL-c and -p remained at stable elevated levels within subjects across serial measurements. Neither change in NfL nor baseline NfL were significantly associated with changes in clinical markers of disease severity.

Conclusions: These findings confirm NfL-c as a faithful diagnostic marker of MSA, while NfL-p showed less robust diagnostic value. The significant NfL elevation in MSA was found to be remarkably stable over time and was not predictive of clinical disease progression.

Keywords: Biomarker; Multiple system atrophy; Neurofilament light chain; Parkinson’s disease.

Figure 1.. Neurofilament light chain correlation analyses.

Correlation of NfL values in all subjects at baseline derived previously using conventional ELISA versus those derived using the high-sensitivity platform in this study (A). Correlation of baseline NfL in CSF versus NfL in plasma derived using the same high-sensitivity platform (B). There was remarkable agreement of values derived in CSF using different platforms, while the correlation between NfL in CSF and NfL in plasma was highly significant but of only moderate strength.

Figure 2.. Comparison of neurofilament light chain between disease categories.

NfL in MSA, PD, and CON subjects in CSF (A) and in plasma (B). There was perfect separation of MSA from CON subjects and near perfect separation from PD using NfL in CSF. There was also good separation of groups using NfL in plasma, but greater overlap with MSA, particularly with PD.

Figure 3.. ROC curves.

ROC curves for NFL-c differentiating MSA from CON subjects (A) and differentiating MSA from PD (B), as well as ROC curves for NfL-p differentiating MSA from CON (C) and from PD (D). NfL-c provided near perfect separation of MSA from the other groups, while the accuracy of separation was good but less robust for NfL-p.

Figure 4.. Progression of UMSARS scores in MSA patients over time.

There was progressive and significant increase of all components of UMSARS (I, II, Total, and IV) over time.

Figure 5.. Longitudinal observations of neurofilament light chain.

Spaghetti plots illustrating longitudinal change of NfL in CSF (top panels) and plasma (bottom panels) over serial measurements across disease categories. There was no significant change of NfL-c or NfL-p over time for either disease category, including MSA.