State transitions in amyotrophic lateral sclerosis: mechanistic convergence and biomarker-guided stratification

Biological heterogeneity is a major obstacle to therapeutic development in amyotrophic lateral sclerosis (ALS). Cross-sectional pathway labels do not show whether a mechanism is active or treatment-relevant in an individual patient.

We conducted a structured critical narrative review informed by the Scale for the Assessment of Narrative Review Articles. MEDLINE, Embase, PubMed, and CENTRAL were searched for English-language literature published from 1 January 2015 to 7 August 2026; seminal earlier studies were identified by citation tracking. Human genetic, pathological, longitudinal biomarker, multi-omic, imaging, neurophysiological, and interventional evidence was prioritized. Evidence was appraised by human proximity, temporality, replication, mechanistic specificity, and treatment relevance.

Evidence supports interacting RNA/proteostasis, nucleocytoplasmic-transport, glial-inflammatory, and energetic-metabolic processes, but current evidence does not yet support the use of any candidate state for individual treatment assignment. The most mature temporal example is loss of TDP-43 splicing repression before neurofilament elevation in genetically at-risk individuals. Nucleocytoplasmic-transport evidence is strongest in C9orf72 disease; inflammatory and metabolic signals are confounded by disease burden and systemic illness; and the gut-brain axis remains a context-dependent amplifier. We propose prespecified transition rules, evidence grades, and biomarker-by-treatment interaction tests.

A state-transition model becomes clinically useful when longitudinal measurements add value beyond clinical predictors and injury markers and prospectively identify differential treatment benefit.