Work

When Immune-Mediated Disease Can Resolve: Stratification by Antigen Source and an Umbrella Protocol with Subtraction as a Diagnostic Endpoint

Year: 2026
Type: preprint
Abstract: Background. It has been argued in Nature Immunology that therapeutic development in immune-mediated disease is organised around clinical signs and symptoms, that the field should move to mechanisms, and that a mechanistic marker is needed to separate the pathways diseases share from those that are their own. The markers proposed so far are cellular and cytokine-based. Where the removal of a triggering factor has been discussed as a therapeutic principle, it has been confined to infectious agents and tumours. Proposal. A different marker is proposed: where the antigen comes from, and whether its source can be removed. On this criterion, 68 immune-mediated diseases for which an antigen and a mechanism of exposure are documented fall into five functional groups — exogenous removable antigen (23 diseases, model: insulin autoimmune syndrome), infectious persistence (11, model: Chagas cardiomyopathy), endogenous antigen modified by a metabolic process (10, model: rheumatoid arthritis), structural exposure (22, model: sympathetic ophthalmia) and defective clearance (2, model: systemic lupus erythematosus). A sixth group gathers diseases in which assessment has not identified a source, genetic defects of immune regulation, and any disease of the first five groups once its damage has become irreversible. The groups are ordered by a therapeutic lever that decreases from the first to the sixth. A group is an outcome of diagnosis, not an attribute of a disease: the same diagnosis can place two patients in two groups. Protocol. An umbrella design is proposed for adults with any immune-mediated disease, in which the stratification variable is the source of the antigen and the intervention changes with the group. Two arms are specified for a first study. In group I the intervention is subtraction of the identified source, and its result is the diagnostic endpoint that confirms or corrects the assignment; a single-arm two-stage design is used. In group III the intervention is exogenous D-beta-hydroxybutyrate as ketone monoester, 10–25 g twice daily with meals for twelve weeks, added to usual care and compared with a matched placebo drink; blood beta-hydroxybutyrate is measured as the exposure variable. Patients whose subtraction fails are reassessed and reassigned, so that the initial assignment ceases to be a wager. Falsification. Five results that would show the framework to be wrong are stated in advance, including the one that tests the classification itself: if removal of a suspected source works equally well in groups where the model predicts it should not, the grouping has no value. Sixth in a series on the AHDI hypothesis (Antigenic Homeostasis Driven Immunopathology).
Language: English
Subfield: Immunology
Open Access status: green
OpenAlex ID: W7237355435
Cites:
Cited by:
Related to: