Complement-mediated injury

In adults with type 2 diabetes diagnosed within five years and no clinical DPN at entry, terminal complement assembles on vascular or glial cells connected to peripheral peptidergic C-fibers and forms conducting C5b-9 membrane-attack complexes before structural injury. Selectively preventing C9 incorporation into those complexes from entry would avert at least 20% of five-year net terminal-axon loss, in an independently defined group accounting for at least 10% of incident DPN under reference care.

This concerns C9-dependent membrane injury, including sublytic effects. Classical, lectin and alternative initiation are alternative routes into it. Deficient CD59 protection, including glycation inactivation, is one possible enabling condition; neither CD59 deficiency nor one initiating complement pathway is mandatory. C3a/C5a receptor signaling without membrane attack is outside this price.

Population and endpoint. Reference care is contemporary diabetes care without the selective intervention, with comparable initiating glycemic, lipid, nutritional and mechanical exposures. Identify adult peptidergic, unmyelinated sensory neurons at entry by combined molecular and anatomical classification. Follow their physical distal skin arbors; reduced peptide expression or staining alone is not disappearance.

For each person, net terminal loss is physical destruction minus replacement over five years, divided by baseline terminal amount. A new terminal subsequently destroyed contributes to both totals. Average people equally. Define:

B = (mean reference net loss − mean loss with Q) / mean reference net loss.

The reference mean must be positive. B measures relative preservation of structure, not pain relief, conduction velocity, myelin thickness or staining intensity.

Finite groups and case coverage. Candidates are the whole eligible population, males, females, baseline BMI ≥30 kg/m², baseline HbA1c ≥7.5%, and a baseline local-terminal-complement-positive group. The latter has the R1 molecular event in at least 10% of mapped target units during a 30-day entry observation. No additional intersections or post-treatment responder groups qualify. Incident DPN means new bilateral length-dependent peripheral nerve signs with objective small-fiber loss or abnormal conduction, after excluding another cause. Coverage is C = P(group membership | incident DPN under reference care).

Connected cells and events. A target unit includes the identified neuron, its Remak Schwann cells and satellite glia, and endothelial or mural cells of vessels demonstrably supplying these structures. Q acts on those vascular and glial cells; neuronal membranes, blood cells, CNS cells and unrelated vessels are outside its direct scope. Damage can reach the target terminal through vascular supply, glial support or local downstream inflammation.

R1 is direct assembly of C5b, C6, C7 and C8 into a surface-associated C5b-8 complex on a named vascular or glial cell during the first follow-up year. At least 10% of baseline target units, averaged with equal person weight within a candidate group, must contain an event before structural injury in the connected arbor. R2 is addition of C9 to the same complex and formation of a C9-dependent conducting membrane pore within 24 hours, with at least 10% of units completing that sequence before injury.

For this timing condition, structural injury begins at the first ≥5% net physical terminal deficit from entry maintained for 30 days in the tracked arbor. Individual terminal turnover without that sustained deficit does not itself establish injury. The criterion applies in the reference course; an intervention that prevents injury is not penalized for eliminating its timing landmark. These numerical boundaries define the proposition rather than estimate biological constants.

Selective intervention Q. From entry for five years, prevent C9 incorporation and pore formation by terminal complement specifically on the named vascular and glial membranes. This molecular intervention leaves upstream complement activation, C3 fragment opsonization, C3a/C5a production, C5b-8 formation, CD59's other functions and other pore systems directly intact. Residual C5b-8 effects remain possible. Q does not independently improve systemic glucose, supply trophic factors or restore perfusion.

Membrane repair, vascular tone, barrier function, inflammatory recruitment, infection consequences, glial metabolism and terminal replacement remain free to respond. Their benefits and harms all count in B. The intervention therefore targets a defined event, rather than whichever complement effects happen to be harmful.

Source assessment

Best judgment: about 8%; skeptical–favorable sensitivity: 0.14%–56%. These are subjective probabilities of the complete human proposition, not treatment effects or confidence limits. Diabetic nerve complement deposition is well supported; conducting membrane injury before human terminal loss is much less directly established. Reviewed 14 September 2026. Probability method · Collection index.

Reasoning from the biology

Complement abundance, cleavage, complex assembly, deposition and membrane conductance are different observations. Plasma leakage can increase C3, C9 and immunoglobulins in nerve without proving active attack on a living cell. A terminal-complex neoepitope demonstrates more than total C9 protein, but a deposited or nonconducting complex is still not a demonstrated membrane pore.

The priced sequence is:

terminal-complement assembly → C9-dependent membrane pores → vascular/glial consequences → terminal destruction minus replacement.

Pores can kill a cell, perturb ions or provoke signaling while it survives. Endothelial dysfunction could reduce supply or alter the blood–nerve barrier; glial injury could interrupt support. These are alternatives, not mandatory serial requirements. Membrane removal and replacement, residual complement regulation, collateral blood supply and glial compensation can prevent a pore event from causing appreciable terminal loss.

Complement also participates in clearance and repair. Broad inhibition can alter those functions and suppress C3a/C5a signaling, making its result different from selective C9 prevention. A benefit from anti-C5 or anti-MASP-2 treatment cannot automatically be assigned to membrane attack.

Human CD59 provides a concrete susceptibility mechanism. Its K41/H44 region can undergo glycation that impairs complement restriction. Early purified-protein experiments used 0.5 M sugar, far above diabetic glucose concentrations, so they establish chemistry rather than physiological kinetics. A later diabetic mouse experiment expressing human wild-type CD59 or a glycation-resistant H44Q mutant supplied an in-vivo functional comparison in erythrocytes. Human CD59's glycation motif differs from ordinary mouse CD59, limiting a simple mouse-to-human exposure translation. Neither experiment demonstrated diabetic nerve-terminal preservation. Acosta et al., 2000, Sahoo et al., 2017.

Evidence that moves the judgment

Human nerve deposition is a substantial, replicated association. Rosoklija and colleagues studied 15 diabetic and 18 other chronic neuropathy biopsies, with similar reported age and neuropathy severity. C3d was present in 15/15 versus 3/18; C5b-9 in approximately 14/15 versus 3/18. The many stained vessels within a nerve do not increase the independent sample size. These were established neuropathies, without a selective intervention or pre-loss observation. Rosoklija et al., 2000.

Yell and colleagues provide the larger accessible donor table: 56/63 people with diabetes versus 7/29 confirmed nondiabetic people had positive endoneurial microvascular C5b-9 staining. An additional 25 had unknown diabetes status and should not silently become confirmed controls. The public-table odds ratio is 25.14, but the more interpretable single-feature staining-grade AUROC is 0.847. A donor bootstrap gives approximately 0.756–0.926. This classifies diabetes among selected neuropathy biopsies; it does not predict incident DPN among people with diabetes. Yell et al., 2018.

The simple alternatives matter. Microvascular sclerosis alone gives AUROC 0.763, and age 0.629. The staining-minus-sclerosis difference is 0.084, with a paired bootstrap interval −0.015 to 0.183. Excluding six pediatric controls leaves 63 diabetic and 23 nondiabetic adults: staining AUROC 0.833, sclerosis 0.762, age 0.533. Thus childhood control diagnoses do not explain the whole association, but the stain has not demonstrated a reliable discrimination gain over sclerosis. Prior donor-level adjustment retained a diabetes association after age and sclerosis; adjustment still cannot establish causal direction.

The disease selection is consequential: diabetic biopsies included amyotrophy, autoimmune conditions and possible CIDP as well as sensory-motor polyneuropathy. In the latter category, 34/37 were stain-positive, so autoimmune diagnoses alone do not explain the result. Strong deposition was especially associated with chronic vascular sclerosis. The same paper describes strong perineurial and larger-artery staining in diabetic and nondiabetic tissues as apparently nonpathogenic. This makes “C5b-9 stain equals destructive pore” an untenable shortcut. Yell et al., 2018.

Glycated CD59 adds mechanism, but not an independent nerve cohort. Qin and colleagues found glycated CD59 associated with complement deposition in human diabetic nerve and kidney. Their nerve specimens were a subset of the Rosoklija series: 12 diabetic and 14 nondiabetic biopsies. They are not a second independent replication. Reduced complement restriction was tested functionally in erythrocytes from 12 diabetic and six nondiabetic people, rather than the nerve cells whose staining was measured. The tissue association and blood-cell functional result are mutually informative, but they do not identify early nerve-cell CD59 failure or its contribution to terminal loss. Qin et al., 2004.

An older immunofluorescence study found immunoglobulin/complement-associated staining in 6/16 diabetic versus 16/53 other neuropathies and considered trapping among the explanations. Its assay was not the same terminal-complex neoepitope endpoint, so it is neither a direct refutation of the later C5b-9 findings nor evidence that all deposition is active injury. 1985 biopsy study.

The recent fascicle proteomics remain suggestive rather than causal. Schwarz and colleagues sampled 15 independent people: four nondiabetic controls, four diabetic people without the identified sciatic fascicular lesions and seven with lesions. All underwent amputation for ischemia and/or infection. Two of the four diabetic people without those lesions already had neuropathy. These groups do not represent early diabetes without DPN followed to incident disease. The article also gives inconsistent sex counts between its table and results prose, so its apparent demographic balance cannot be taken on trust. Schwarz et al., 2025.

Lesioned fascicles contained more C3/C9 and less CD59 protein, together with extracellular and acute-phase proteins and structural abnormalities. The supplementary intensities use 5/6/55 proteomic samples, not 5/6/55 independent donors. Donor-level intensities and their mapping were not publicly available in the accessed release. Replacing those sample counts with donor counts while retaining fascicle-level standard deviations would not repair the analysis. Total C9 abundance does not prove MAC assembly, and abundance-based pathway enrichment cannot identify classical versus alternative activation. Plasma entry, inflammation and changing cellular composition remain competing explanations. The study informs established injury and barrier disruption, without establishing the first-year pore sequence.

Accessible intervention evidence does not yet isolate diabetic structural MAC injury. A disclosed anti-MASP-2 mouse experiment used 12 animals per group, weekly 1 mg/kg antibody from seven to 24 weeks of age, and hot-plate tests at weeks 17, 18 and 20. Treatment reduced the diabetic delay in thermal response relative to saline and isotype controls. The final prose mistakenly calls this increased reaction time; the results describe decreased latency. This is a patent-disclosed behavioral comparison, without a reported terminal-density endpoint in that example. Upstream lectin-pathway inhibition also affects processes beyond C9 pores. It supports a possible complement-dependent functional effect, not the priced structural prevention fraction. US20110311549A1, Example 51.

Rare inherited CD59 deficiency and CD59-deficient mouse nerve abnormalities demonstrate that loss of complement protection can injure peripheral nerve. They do not establish a common diabetic subgroup or specifically peptidergic terminal loss. Likewise, C3a/C5a receptor localization and acute compound-action-potential effects concern receptor signaling and excitability, not membrane attack or structural preservation. CD59 nerve study, 2023, complement-receptor study, 2025.

New blood-marker discrimination is not prospective nerve causation. A 2025 study compared 82 healthy controls, 82 diabetic people without microvascular complications and 82 with complications. Glycated-CD59 AUROC for the two diabetic groups was 0.849 (0.789–0.908). Only 12 belonged to the neuropathy-only subgroup; the composite also included kidney, retinal and mixed complications. Renal function, inflammatory measures and disease duration differed. Recruitment was prospective, but the reported outcome was prevalent complications. The high AUROC does not establish local functional CD59 loss or prevention of incident DPN. Plasma GCD59 study, 2025.

An existing analysis of 77 established-DPN donors in GSE148059 found no discriminating complement transcript score for the 52-week myelinated-fiber course after its seven-score multiplicity check. The exact C3+C9−CD59 score had permutation p≈0.746, family-adjusted p=1.0; compartment adjustment did not rescue it. This constrains that expression-based route. RNA abundance cannot measure complement cleavage or pores, and the public RNA specimen's timing is unresolved; the analysis is not a validated baseline predictor or a negative early-C-fiber intervention.

Probabilities of the logical claims

Conditionals assume preceding claims for the same candidate group and connected sequence. If several groups qualify, a later row requires at least one of those groups to continue satisfying the chain.

See the claims and probabilities table above.

R1

Repeated human terminal-complement deposition supports occurrence; early timing and exact living cell surfaces remain unobserved.

R2

CD59 functional evidence and complement biology support insertion, while deposited complexes can be nonconducting and membrane repair can prevail.

R3

Vascular/glial injury is biologically capable of affecting terminals; selective diabetic structural rescue is missing.

R4

Extensive late vascular deposition permits a material role, but does not establish its preventable fraction or durable benefit.

R5

The exposure may be common in diabetes; the prevalence of early, functionally damaging membrane attack remains unknown.

Assembly is not defined by damage. Pore conductance is not defined by terminal loss. R3 adds a structural effect; R4 adds its magnitude; R5 adds population coverage. None requires a particular upstream activator or CD59 mechanism to be true.

Ideal experiments that would resolve each claim

Use a representative early-T2 population, the entry-defined groups, complete mapping of connected target units and nonperturbing physical tracking over five years. The idealizations are direct molecular-complex identification, conductance measurement, lineage tracing and a sustained membrane-specific Q. Normal terminal turnover is distinguished from the stated sustained net-deficit landmark.

R1 — assembly, with no predecessor assumed. Observe C5b-8 molecular assembly on the named living vascular or glial cell surfaces and its time relative to connected-arbor injury. YES: one candidate group meets the first-year, pre-injury and 10%-of-units criteria. NO: complete observation shows all groups fail at least one criterion. A C9 protein increase, a late deposit or an unresolved cell surface leaves R1 unresolved.

R2 — pore formation conditional on R1. Follow the same complexes through C9 addition and measure C9-dependent conductance in the same membranes, using selective C9 blockade to establish identity. YES: one compatible group has at least 10% of units complete the 24-hour, pre-injury sequence. NO: all compatible groups fail it, including complexes confined to nonconducting deposits or extracellular material. Staining without functional membrane measurement leaves R2 unresolved. Sublytic pores qualify; cell death is not required.

R3 — structural causation conditional on R1–R2. Randomly allocate people in each compatible group to reference care or Q from entry. Verify the specified membrane target and direct selectivity throughout follow-up. Count physical terminal destruction and replacement while allowing all downstream consequences. YES: one compatible group has a positive reference-loss denominator and B≥0.05. NO: all faithfully implemented comparisons have B<0.05, including net harm. Pain improvement, anti-C5 treatment that also suppresses C5a, missing structural follow-up or incomplete target engagement leaves this exact comparison unresolved.

R4 — magnitude conditional on R1–R3. Use the same groups, Q and five-year endpoint. YES: at least one R3 group reaches B≥0.20. NO: every R3 group has 0.05≤B<0.20. A short-lived structural difference does not qualify if the five-year boundary fails. An interval straddling 0.20 is unresolved.

R5 — coverage conditional on R1–R4. Ascertain baseline membership and reference-care incident DPN across the representative population. YES: one R4 group has C≥0.10. NO: every R4 group has C<0.10. Missing incident ascertainment is unresolved. Do not create a new union or define membership by observed benefit.

Exact population values determine truth; finite data can remain unresolved near a boundary.

Uncertainty and sensitivity

Best conditionals [60,65,60,60,60]% give 8.424%, reported as about 8%. Skeptical [30,30,25,25,25]% and favorable [90,90,85,90,90]% give 0.140625% and 55.7685%. These coordinated scenarios have no assigned weights, justified variance or empirical human calibration.

A mainly deposited, late-injury interpretation, [40,40,35,40,40]%, gives 0.896%. Common early pores with transferable structural consequences, [80,80,80,80,85]%, give 34.816%. The large span reflects unresolved human timing and causal specificity rather than uncertainty about whether diabetic biopsies stain more often.

Direct early human pore evidence could move R2 from 65% to 85%, giving 11.016% with other judgments unchanged. Evidence that most qualifying surface complexes remain nonconducting could move it to 30%, giving 3.888%. Merging R1 and R2 must yield 39%, preserving the headline; an extra row alone does not justify another penalty.

Lowering the preservation, coverage or unit-prevalence boundary weakens the proposition. Moving the injury landmark later also relaxes the timing requirement, while requiring activation before every individual turnover event would strengthen it substantially. Exact repricing of those alternative definitions is unsupported.

The quantitative finding (background note outside this collection) gives donor-level discrimination, its simple comparators and the limits of the older transcript/proteomic analyses. No observed score supplies an empirical probability that Q will prevent human terminal loss.

Overlap and boundary

This route overlaps microvascular supply, PKC-related vascular injury, glial support, TNF/TLR4 inflammation and macrophage clearance. CD59 glycation also connects it to glycation biology, without making extracellular matrix glycation necessary.

One causal sequence can satisfy several descriptions. Their probabilities and prevented fractions cannot be added. Complement-dependent established-nerve repair and autoimmune demyelination remain distinct from the early-T2 prevention proposition.