Before Symptoms Appear, Blood Already Bears the Traces of ALS: 19 Proteins Estimate the Onset Window

The challenge of treating amyotrophic lateral sclerosis (ALS) lies not only in its rapid progression, but also in the fact that nerve damage often precedes noticeable symptoms. A study published in *Nature Medicine* shows that, in people who carry ALS-associated pathogenic variants but have not yet developed the disease, changes in plasma proteins may reveal years in advance that the disease is gradually approaching, providing early-intervention research with a timeline that still requires calibration.

The research team analyzed 516 serial plasma samples from 137 participants, including 33 who progressed from an asymptomatic state to clinical disease, 35 patients with ALS, 10 carriers of pathogenic variants who had not yet converted, and 59 controls. The study used a high-throughput platform to measure more than 5,000 proteins; the peer-reviewed final paper reported that 92 of them showed significant changes before clinical conversion, involving biological pathways related to skeletal muscle, the extracellular matrix, neurofilaments, and inflammatory signaling.

The team then developed a core panel of 19 proteins from the candidate markers to determine whether symptoms would appear within six months, one year, two years, three years, or five years after each blood draw. The areas under the curve obtained through five-fold cross-validation ranged from 0.80 to 0.89, indicating that the model had some discriminatory ability; when estimating the time to clinical conversion, the mean absolute error was approximately 1.6 years. The tool is not intended to determine exactly when an individual will develop the disease, but to narrow a previously highly uncertain risk window to a range more suitable for study design.

Current monitoring of early ALS risk relies largely on neurofilament light chain (NfL/NEFL), which typically rises after axonal damage accelerates. The new panel includes NEFL as well as EDA2R, CALCA, and several proteins associated with muscle or neural function; over longer prediction periods, its overall performance was better than that of NEFL alone. However, when predicting conversion within six months, the two performed similarly, reflecting the sharp rise in NEFL shortly before disease onset, which remains a strong and direct short-term signal.

The UK Biobank provided partial external validation: in cross-sectional data, the researchers reproduced the pre-onset upward trends in proteins including NEFL, EDA2R, and CA3, while an available 15-protein subset also outperformed NEFL alone in estimating time to conversion. However, the biobank lacked exact symptom-onset dates, so the researchers could only estimate them retrospectively from relevant hospitalization records, producing a “pseudo-longitudinal” trajectory; its mean absolute error was approximately 2.75 years, and it cannot be regarded as a complete independent clinical validation.

The most direct application of these findings may be to help ALS prevention trials select carriers of pathogenic variants who are more likely to develop the disease in the near term, reducing the challenge of participants having to wait years for a clinical event. The panel may also serve as a research marker for tracking disease progression. However, the discovery cohort included only 33 confirmed converters and is primarily applicable to populations with clearly defined genetic risk; whether the model can maintain its performance across genotypes, populations, testing platforms, and sporadic ALS still requires confirmation in larger prospective studies.

The research team has made available the analysis code, summary statistics, and data needed to reproduce the figures, covering protein changes, prediction models, and the UK Biobank validation workflow; patient-level and processed data were not released because of privacy and data-governance considerations. Beyond independent replication, the next step will be to translate the exploratory protein platform into a standardized and reproducible test with clearly defined clinical thresholds before it can move from a research tool toward trial enrollment or individual risk assessment.