The Bolus Theory: From Intravascular Leak to Focal Endothelial Destruction

Serious injuries after COVID-19 vaccination — myopericarditis, thrombosis with thrombocytopenia, stroke, vasculitis, capillary leak, sudden death — have been investigated one diagnosis at a time, each with its own proposed mechanism. This paper argues that they share a single cause, and that the cause is the route of delivery rather than the substance delivered. The argument proceeds by elimination. Six observations are set out that any explanation must satisfy: that some recipients are seriously harmed while most are not, with no recipient characteristic identifying which; that the same diagnoses follow products sharing no antigen, platform or adjuvant; that the muscle holding the entire dose at its highest concentration is minimally transfected and minimally damaged; that the lesions are focal, with destroyed endothelium beside intact endothelium; that circulating antigen is present in quantities too small, and too uniformly distributed, to draw the boundary of a focal lesion; and that free circulating antigen does not separate the harmed from the unharmed. Taken together these exclude the payload, lot-to-lot variation, circulating protein, immune complexes, autoimmunity, the lymphatic route, and focality generated by the immune response rather than by the delivery. What remains is that during ordinary intramuscular injection a variable fraction of the dose enters the bloodstream directly at the injection site, not as one bolus but as a train of concentrated micro-boluses, which are carried downstream, transfect bounded populations of endothelial cells, and cause their destruction by the cytotoxic response they provoke. On this account what differs between recipients is principally the rate at which material enters the vascular system: injection speed sets the concentration of each micro-bolus, and concentration rather than total quantity determines whether a focal lesion forms. Two tables compiled for this paper record, for fourteen injected products against thirteen categories of adverse event, the published individual-patient case reports that exist for each pairing, with every citation verified against the PubMed record. Four of the twelve event categories are documented after every one of the thirteen intramuscular and subcutaneous products; six are documented after twelve or thirteen of them. The products involved include live attenuated virus, inactivated split and whole virus, bacterial toxoid, recombinant protein, virus-like particle, bacterial polysaccharide and conjugate, lipid-nanoparticle mRNA, adenoviral vector, and an obsolete nerve-tissue preparation — sharing no antigen, no platform, no adjuvant and no excipient. Seven predictions that could defeat the account are stated, together with the experiments that would test them and have not been performed. Conjectures are labelled as such throughout, and a limitations section sets out what the evidence does not support. The paper does not claim that the payload is harmless, nor that any particular fraction of a given dose leaks. The practical implication requires no new product and no new equipment: if the account is correct, injection speed and the delay before needle withdrawal are the variables that determine whether a focal lesion forms at all.