Plain-Language Research Summary: A Blood Test for Nerve-Fibre Loss in CMT

In 2021 I argued that all types of CMT are “axonal”. Even in the “demyelinating” types, what drives disability over time is the loss of nerve fibres (axons), not the slowing of nerve signals. In CMT1A, for example, disability tracks axon loss rather than nerve conduction velocity.¹ That raises a practical question. If axon loss is what matters, can we measure it more directly than with a nerve conduction study or a clinical exam? One candidate is a blood test for a protein called neurofilament light chain, or NfL.

This is a plain-language summary of a 2025 study of NfL and three other blood markers in CMT, with two earlier studies for context.

What NfL Is

Neurofilaments are part of the internal scaffolding of nerve cells, and they are especially concentrated in axons. When axons are damaged, NfL is released and can be measured in blood or spinal fluid. Very sensitive tests, such as the single molecule array (Simoa) used in this study, can detect the tiny amounts in plasma. NfL isn’t specific to CMT, or to any one disease. It rises with axon damage from many causes, and it rises with age.²

The Question

Researchers in Riga, Latvia, working with the clinical neurochemistry laboratory in Gothenburg, Sweden, asked two things:

1. Are four blood markers higher in people with CMT than in healthy people?
2. And do they track how severe someone’s CMT is?

Besides NfL, the researchers measured GFAP, a protein found in some Schwann cells (the cells that make myelin in the peripheral nerves), and GDF15 and FGF21, two markers linked to muscle and mitochondrial stress.

What Was Done

This was a cross-sectional study. Each person gave one blood sample and had one assessment, at one point in time. Severity was measured with the CMT Neuropathy Score, second version (CMTNSv2), a 0–36 scale that combines symptoms, a neurological exam and two nerve conduction measurements. Scores of 1–10 count as mild, 11–20 as moderate and above 20 as severe.³ The laboratory staff who ran the blood tests didn’t know which samples came from people with CMT.

Who and How Many

• 44 people with genetically confirmed CMT and 44 healthy controls of similar age and sex. The average age was about 39.
• Based on nerve conduction, 31 had demyelinating CMT and 13 had axonal CMT.
• 24 had the CMT1A duplication of PMP22 and six had GJB1 variants (CMTX1). The rest were spread across other genes, including MPZ, MFN2, AARS1 and HSPB1.
• The median CMTNSv2 score was 12, in the moderate range. People with brain or spinal cord abnormalities were excluded.

What It Found

• All four markers were higher in the CMT group. Median NfL was 12.3 pg/mL in CMT and 5.25 pg/mL in controls, more than twice as high.⁴
• NfL rose with severity. The correlation with the CMTNSv2 was moderate: about 0.46 to 0.47 on a scale where 1 would be perfect. (The abstract and the results section give slightly different figures.) GFAP correlated more weakly, at 0.31.
• NfL also rose with age (0.38). That matters, because NfL rises with age in everyone.
• NfL and GFAP weren’t significantly different between the demyelinating and axonal groups. The authors suggest that after years of disease, myelin and axon damage are bound up together. That fits the point of my earlier post.
• GDF15 tracked age and disease duration. FGF21 was higher in CMT but didn’t correlate with anything the authors examined.

A correlation around 0.46 means people with higher severity scores tended to have higher NfL, with a lot of overlap. You couldn’t read a person’s severity from their NfL level.

What It Can’t Tell You

• Whether NfL changes as CMT progresses. One snapshot can’t show that. The authors themselves note that the CMTNS shows only small yearly changes, which is why they want a more sensitive marker.
• Whether NfL responds to treatment. Nobody in the study was treated.
• Much about specific types. Apart from CMT1A, the subgroups were tiny. The 13 people with axonal CMT carried variants in at least six different genes.
• Anything about you as an individual. The groups overlap, NfL rises with age, and it rises with axon damage from any cause. This is a research tool, not a test that tells one person how they’re doing.
• How solid the smaller findings are. The study made many comparisons in a small sample, and the paper doesn’t report correcting for that.

Two Other Studies Worth Knowing

London, 2018. Sandelius and colleagues measured plasma NfL in 75 people with CMT and 67 age-matched controls, also at a single point in time. Median NfL was 26.0 pg/mL in CMT and 14.6 pg/mL in controls. It correlated with the Rasch-modified CMT Examination Score (0.43) and Neuropathy Score (0.37). It was raised in the CMT1A, SPTLC1 and GJB1 subgroups, and in both demyelinating and axonal forms.⁵ The absolute numbers come from different laboratories and populations, so don’t compare 26.0 in London with 12.3 in Riga.

Riga, 2023. The same Latvian group followed people over time. They enrolled 101 people with CMT and 64 controls and retested 73 patients and 28 controls three years later. NfL was higher in CMT, and on average it rose over three years in both patients and controls. In 22 of the 73 patients, though, it went down. The change in NfL didn’t correlate with severity (r = 0.228, p = 0.052). The authors concluded that NfL “is not a biomarker for CMT progression.”⁶

So the 2025 paper’s conclusion, that NfL and GFAP “can be reliable disease indicators in future research,” should be read alongside its own group’s three-year result. NfL tells groups of people with CMT apart from groups without it, and loosely tracks severity at one point in time. It doesn’t yet track change in a person. A 2026 set of trial-design recommendations from clinicians, patient organizations and industry suggests NfL could still be useful in trials: a baseline level as a sign of prognosis, or a fall in NfL as a sign that a drug is having a biological effect.⁷ Neither would show that a treatment helps people feel or function better.

Who Funded It

The Latvian Science Council funded the 2025 study (project lzp-2021/1-0327, “Discovering biomarkers of disease progression and variability in Charcot–Marie-Tooth neuropathy”). Henrik Zetterberg’s work is supported by a long list of public and charitable funders, mostly for dementia research. Zetterberg declared advisory and consulting roles with many pharmaceutical and diagnostics companies, and he is a co-founder of Brain Biomarker Solutions in Gothenburg AB. The paper says these are outside the submitted work. The other authors declared no conflicts.⁸

The 2018 London study was funded by the Wellcome Trust, the UK Medical Research Council and the US NIH-funded Inherited Neuropathy Consortium, among others. Zetterberg and Kaj Blennow, who co-authored all three papers, both disclosed that they co-founded Brain Biomarker Solutions.⁹

Bottom Line

NfL is a real signal: it’s reliably higher in people with CMT, and it loosely tracks how severe their CMT is. It isn’t yet a way to tell whether your CMT is getting worse, or whether a treatment works. For that, trials still lean on clinical scales, functional tests, and muscle MRI, each with its own problems.¹⁰

Why It Matters to Me

Nerve conduction studies are still the most common (and often poorly understood) diagnostic tech for CMT and related neuromuscular diseases. The technology and diagnostic model behind them date back to the 1960s-70s. I’ve had one done annually for quite some time, never on a machine with an edition of Windows newer than Windows ME! Nerve conduction velocity and amplitude don’t map to clinical, physical symptoms, they don’t track mobility loss, and they don’t have much to do with the patient’s lived experience. Hopefully that limited NCV data will increasingly be supplemented with the results of new test methods, like muscle biopsies and blood tests. Diagnostic tech for CMT is due for big updates and reforms.

Notes

1. Krajewski, K. M., R. A. Lewis, D. R. Fuerst, C. Turansky, S. R. Hinderer, J. Garbern, J. Kamholz, and M. E. Shy. 2000. “Neurological Dysfunction and Axonal Degeneration in Charcot-Marie-Tooth Disease Type 1A.” Brain 123 (7): 1516–27. https://doi.org/10.1093/brain/123.7.1516. ↩︎
2. Pretkalnina, Dace, Elizabete Kenina, Linda Gailite, Dmitrijs Rots, Kaj Blennow, Henrik Zetterberg, and Viktorija Kenina. 2025. “Evaluating Plasma Biomarkers NfL, GFAP, GDF15, and FGF21 as Indicators of Disease Severity in Charcot–Marie–Tooth Patients.” Frontiers in Neurology 15: 1490024. https://doi.org/10.3389/fneur.2024.1490024. ↩︎
3. Murphy, Sinéad M., David N. Herrmann, Michael P. McDermott, et al. 2011. “Reliability of the CMT Neuropathy Score (Second Version) in Charcot-Marie-Tooth Disease.” Journal of the Peripheral Nervous System 16 (3): 191–98. https://doi.org/10.1111/j.1529-8027.2011.00350.x. The 0–36 range and severity bands are as described in Pretkalnina et al. 2025, Methods. ↩︎
4. Pretkalnina et al. 2025, Table 2. https://doi.org/10.3389/fneur.2024.1490024. ↩︎
5. Sandelius, Åsa, Henrik Zetterberg, Kaj Blennow, Rocco Adiutori, Andrea Malaspina, Matilde Laura, Mary M. Reilly, and Alexander M. Rossor. 2018. “Plasma Neurofilament Light Chain Concentration in the Inherited Peripheral Neuropathies.” Neurology 90 (6): e518–e524. https://doi.org/10.1212/WNL.0000000000004932. ↩︎
6. Setlere, Signe, Arta Grosmane, Natalja Kurjane, Linda Gailite, Dmitrijs Rots, Kaj Blennow, Henrik Zetterberg, and Viktorija Kenina. 2023. “Plasma Neurofilament Light Chain Level Is Not a Biomarker of Charcot–Marie–Tooth Disease Progression: Results of 3-Year Follow-Up Study.” European Journal of Neurology 30 (8): 2453–60. https://doi.org/10.1111/ene.15858. ↩︎
7. Abrams, Charles K., Sue Bruhn, Joshua Burns, et al. 2026. “Clinical Development of Therapies for Charcot–Marie–Tooth Disease: Recommendations for Trial Design, Endpoints, and Regulatory Pathways.” Journal of the Peripheral Nervous System 31 (3): e70146. https://doi.org/10.1111/jns.70146. ↩︎
8. Pretkalnina et al. 2025, Funding and Conflict of interest statements. https://doi.org/10.3389/fneur.2024.1490024. ↩︎
9. Sandelius et al. 2018, Study funding and Disclosure sections. https://doi.org/10.1212/WNL.0000000000004932. ↩︎
10. McCray, Brett A., and Vera Fridman. 2024. “Clinical Outcome Assessments and Biomarkers in Charcot-Marie-Tooth Disease.” Neurology 103 (12): e210120. https://doi.org/10.1212/WNL.0000000000210120. ↩︎

Sources

• Abrams, Charles K., Sue Bruhn, Joshua Burns, Rebekah Fleming, Katherine Forsey, Vera Fridman, David N. Herrmann, et al. “Clinical Development of Therapies for Charcot–Marie–Tooth Disease: Recommendations for Trial Design, Endpoints, and Regulatory Pathways.” Journal of the Peripheral Nervous System 31, no. 3 (2026): e70146. https://doi.org/10.1111/jns.70146.
• Krajewski, K. M., R. A. Lewis, D. R. Fuerst, C. Turansky, S. R. Hinderer, J. Garbern, J. Kamholz, and M. E. Shy. “Neurological Dysfunction and Axonal Degeneration in Charcot-Marie-Tooth Disease Type 1A.” Brain 123, no. 7 (2000): 1516–27. https://doi.org/10.1093/brain/123.7.1516.
• McCray, Brett A., and Vera Fridman. “Clinical Outcome Assessments and Biomarkers in Charcot-Marie-Tooth Disease.” Neurology 103, no. 12 (2024): e210120. https://doi.org/10.1212/WNL.0000000000210120.
• Murphy, Sinéad M., David N. Herrmann, Michael P. McDermott, Steven S. Scherer, Michael E. Shy, Mary M. Reilly, and Davide Pareyson. “Reliability of the CMT Neuropathy Score (Second Version) in Charcot-Marie-Tooth Disease.” Journal of the Peripheral Nervous System 16, no. 3 (2011): 191–98. https://doi.org/10.1111/j.1529-8027.2011.00350.x.
• Pretkalnina, Dace, Elizabete Kenina, Linda Gailite, Dmitrijs Rots, Kaj Blennow, Henrik Zetterberg, and Viktorija Kenina. “Evaluating Plasma Biomarkers NfL, GFAP, GDF15, and FGF21 as Indicators of Disease Severity in Charcot–Marie–Tooth Patients.” Frontiers in Neurology 15 (January 2025): 1490024. https://doi.org/10.3389/fneur.2024.1490024.
• Sandelius, Åsa, Henrik Zetterberg, Kaj Blennow, Rocco Adiutori, Andrea Malaspina, Matilde Laura, Mary M. Reilly, and Alexander M. Rossor. “Plasma Neurofilament Light Chain Concentration in the Inherited Peripheral Neuropathies.” Neurology 90, no. 6 (2018): e518–e524. https://doi.org/10.1212/WNL.0000000000004932.
• Setlere, Signe, Arta Grosmane, Natalja Kurjane, Linda Gailite, Dmitrijs Rots, Kaj Blennow, Henrik Zetterberg, and Viktorija Kenina. “Plasma Neurofilament Light Chain Level Is Not a Biomarker of Charcot-Marie-Tooth Disease Progression: Results of 3-Year Follow-Up Study.” European Journal of Neurology 30, no. 8 (2023): 2453–60. https://doi.org/10.1111/ene.15858.