{
  "id": "32",
  "title": "The skin repair environment",
  "summary": "Altered skin matrix and trophic support may prevent durable replacement of lost sensory terminals.",
  "source_title": "Altered skin matrix and trophic support restrict terminal replacement",
  "page": "32_skin_repair_environment.html",
  "source_markdown": "sources/32_skin_repair_environment.md",
  "structured_data": "data/32_skin_repair_environment.json",
  "snapshot_date": "2026-09-14",
  "source_review_date": "2026-09-14",
  "scope_label": "Early type 2 diabetes",
  "scope_exception": false,
  "source_headline_markdown": "**Best judgment: about 16%; uncertainty range (sensitivity): 0.8–62%.** A skin contribution to impaired axon growth is plausible and experimentally supported. The main uncertainty is whether correcting it supplies enough durable replacement during ordinary early diabetes to prevent material human terminal loss. These are literature-informed subjective probabilities, not a measured treatment effect or calibrated success rate. Reviewed 14 September 2026. [Probability method](probability_method.md) · [Collection index](README.md).",
  "probability_percent": {
    "skeptical": 0.8421875,
    "base": 16.016,
    "favorable": 62.137125
  },
  "probability_meaning": "Subjective belief in the complete causal proposition. Scenario endpoints are sensitivity products, not confidence limits.",
  "causal_proposition_markdown": "In adults with type 2 diabetes diagnosed within five years and no clinical DPN at entry, an altered distal skin scaffold and trophic environment impedes replacement of peptidergic C-fiber terminals. Correcting the specified environment from entry would add replacements equivalent to at least **20% of five-year reference net terminal loss**, and avert at least **20% of that net loss**, in a baseline-defined group accounting for at least **10% of reference-care incident DPN**.\n\nThis is a prevention claim about failed replacement. Recovery after capsaicin denervation, nerve injury or established DPN informs the mechanism but is a different outcome. Reduced destruction without additional replacement does not satisfy this particular proposition.\n\n**The bounded environment.** The compartment is distal-leg dermis and epidermal basement membrane, including the extracellular interface of keratinocytes and dermal fibroblasts with the target arbor. The scaffold panel comprises **collagens I, III and IV, laminins and fibronectin**, including their local amount, spatial organization and covalent modifications. The trophic panel comprises extracellular **NGF, BDNF, NT-3, NT-4, GDNF, neurturin and artemin**, distinguishing mature proteins, precursors and matrix-bound fractions. SEC31A-mediated secretion and keratinocyte SIRT1 activity are possible upstream regulators, not required individual defects.\n\nThis explicitly bounded assessment does not encompass every hypothetical repellent, skin immune process or epithelial signal. Whole-nerve regeneration cannot identify the contribution of this skin compartment.\n\n**R1's exposure definition.** In a qualifying group, at least 10% of mapped terminal-interface sites must have a panel protein abundance, precursor/mature ratio or specified matrix-adduct occupancy outside its matched nondiabetic 2.5th–97.5th percentile interval and at least 50% from its positive reference median, for at least 30 cumulative days in the first year. Matrix adducts measured here are fructosyl-lysine, CML, CEL, MG-H1, G-H1, glucosepane and pentosidine. Measurements distinguish matrix-bound protein from unrelated intracellular expression. Site proportions are calculated within each person and then averaged equally across people.\n\nThe deviation must occur before the connected arbor's first ≥5% net physical deficit lasting 30 days. An isolated change in collagen hue, SEC31A RNA or GAP43 staining does not establish this exposure criterion. The numerical boundary specifies a substantial, persistent deviation without defining it by harm; it is not an established biological threshold.\n\n**Q — reference scaffold and trophic substitution.** For five years, an ideal local controller maintains the listed extracellular proteins' joint spatial, chemical and concentration profiles at an intact nondiabetic reference profile matched for age, sex, skin site and mechanical loading. Reference profiles are assigned independently of outcomes and preserve observed covariance; Q does not assemble an arbitrary set of independently favorable component values. It replaces deficient as well as excessive components, including precursor forms and normal matrix modifications. Repeated reference measurements allow normal aging and activity-dependent changes.\n\nThe ideal capability is molecularly selective substitution without cutting or denervating skin. Q does not replace neurons, glia, vessels or immune cells, alter neuronal growth machinery directly, or normalize circulating glucose, insulin or lipids. The comparator receives contemporary diabetes care without Q. Other initiating exposures remain comparable. Downstream changes in inflammation, vascular behavior, skin mechanics, sensory activity and neuronal signaling are allowed. Removing compensatory support, disrupting useful matrix adaptations, or adding new terminal destruction counts against Q. It is not a controller that removes only whichever components prove harmful.\n\n**Groups and endpoints.** Candidate groups are the whole eligible population, males, females, baseline BMI ≥30 kg/m², baseline HbA1c ≥7.5%, and a group with the specified panel deviation at ≥10% of interface sites during a 30-day baseline observation. No additional intersections or future responder definitions are allowed.\n\nTrack fixed territories with stable peptidergic identity and physical anatomy. Let D be cumulative terminal disappearances and R be newly established terminal entries over five years, each divided by baseline terminal amount. A replacement's subsequent disappearance is counted in D. Staining intensity alone is not disappearance or entry. With positive mean reference loss L₀ = mean(D₀ − R₀), define:\n\n- **B_R = (mean R_Q − mean R₀)/L₀:** added replacement relative to reference net loss.\n- **B_T = [L₀ − mean(D_Q − R_Q)]/L₀:** net preservation, including subsequent losses.\n- **C = P(group | incident DPN under reference care):** case coverage.\n\nHumans receive equal weight. Incident DPN means new bilateral length-dependent nerve signs with objective small-fiber loss or abnormal conduction, excluding another cause. Every claim retains only groups satisfying all predecessors; a complete NO must exclude every remaining candidate group. Exposure in one group and benefit in another cannot complete the chain.",
  "claims": [
    {
      "id": "R1",
      "source_id": "R1",
      "claim": "A group has the specified persistent panel deviation before substantial terminal deficit.",
      "claim_markdown": "A group has the specified persistent panel deviation before substantial terminal deficit.",
      "conditional_percent": {
        "skeptical": 50.0,
        "base": 80.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 80.0,
      "source_cumulative_display": null,
      "reason_markdown": "Human skin remodeling and diabetic trophic changes favor an exposure. Exact early compartment, dose and ordering remain sparsely measured.",
      "question": "Do skin matrix proteins or extracellular growth factors show persistent abnormal amounts, processing, or glycation before terminal loss?"
    },
    {
      "id": "R2",
      "source_id": "R2",
      "claim": "In a group satisfying R1, Q adds ordinary-life replacements with B_R ≥0.05.",
      "claim_markdown": "In a group satisfying R1, Q adds ordinary-life replacements with B_R ≥0.05.",
      "conditional_percent": {
        "skeptical": 25.0,
        "base": 55.0,
        "favorable": 85.0
      },
      "cumulative_base_percent": 44.0,
      "source_cumulative_display": null,
      "reason_markdown": "Coculture and skin-directed intervention support an extrinsic effect. Human spontaneous renewal, panel sufficiency and clean compartment attribution remain unresolved.",
      "question": "Would restoring the skin's matrix and growth-factor environment add replacements equivalent to at least 5% of five-year net terminal loss under reference care?"
    },
    {
      "id": "R3",
      "source_id": "R3",
      "claim": "In a group satisfying R1–R2, added replacement reaches B_R ≥0.20.",
      "claim_markdown": "In a group satisfying R1–R2, added replacement reaches B_R ≥0.20.",
      "conditional_percent": {
        "skeptical": 35.0,
        "base": 65.0,
        "favorable": 90.0
      },
      "cumulative_base_percent": 28.6,
      "source_cumulative_display": null,
      "reason_markdown": "Given a real contribution, large substrate and mouse structural effects make materiality plausible. Low ordinary renewal demand or other growth bottlenecks can keep it small.",
      "question": "Would restoring the skin's matrix and growth-factor environment add replacements equivalent to at least 20% of five-year net terminal loss under reference care?"
    },
    {
      "id": "R4",
      "source_id": "R4",
      "claim": "In a group satisfying R1–R3, net preservation also reaches B_T ≥0.20.",
      "claim_markdown": "In a group satisfying R1–R3, net preservation also reaches B_T ≥0.20.",
      "conditional_percent": {
        "skeptical": 55.0,
        "base": 80.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 22.88,
      "source_cumulative_display": null,
      "reason_markdown": "Established extra entries usually add some retained innervation. Continued destruction, fragile new branches and lost compensatory support can erase the gain.",
      "question": "Would restoring the skin's matrix and growth-factor environment also avert at least 20% of five-year net terminal loss after subsequent disappearance is counted?"
    },
    {
      "id": "R5",
      "source_id": "R5",
      "claim": "A group satisfying R1–R4 has C ≥0.10.",
      "claim_markdown": "A group satisfying R1–R4 has C ≥0.10.",
      "conditional_percent": {
        "skeptical": 35.0,
        "base": 70.0,
        "favorable": 90.0
      },
      "cumulative_base_percent": 16.016,
      "source_cumulative_display": null,
      "reason_markdown": "Skin exposures are common, but the sufficiently correctable combination may be uncommon among incident cases. Existing cohorts do not measure this denominator.",
      "question": "Does the benefiting subgroup account for at least 10% of incident DPN cases under reference care?"
    }
  ],
  "sections": [
    {
      "heading": "The causal claim and its scope",
      "markdown": "In adults with type 2 diabetes diagnosed within five years and no clinical DPN at entry, an altered distal skin scaffold and trophic environment impedes replacement of peptidergic C-fiber terminals. Correcting the specified environment from entry would add replacements equivalent to at least **20% of five-year reference net terminal loss**, and avert at least **20% of that net loss**, in a baseline-defined group accounting for at least **10% of reference-care incident DPN**.\n\nThis is a prevention claim about failed replacement. Recovery after capsaicin denervation, nerve injury or established DPN informs the mechanism but is a different outcome. Reduced destruction without additional replacement does not satisfy this particular proposition.\n\n**The bounded environment.** The compartment is distal-leg dermis and epidermal basement membrane, including the extracellular interface of keratinocytes and dermal fibroblasts with the target arbor. The scaffold panel comprises **collagens I, III and IV, laminins and fibronectin**, including their local amount, spatial organization and covalent modifications. The trophic panel comprises extracellular **NGF, BDNF, NT-3, NT-4, GDNF, neurturin and artemin**, distinguishing mature proteins, precursors and matrix-bound fractions. SEC31A-mediated secretion and keratinocyte SIRT1 activity are possible upstream regulators, not required individual defects.\n\nThis explicitly bounded assessment does not encompass every hypothetical repellent, skin immune process or epithelial signal. Whole-nerve regeneration cannot identify the contribution of this skin compartment.\n\n**R1's exposure definition.** In a qualifying group, at least 10% of mapped terminal-interface sites must have a panel protein abundance, precursor/mature ratio or specified matrix-adduct occupancy outside its matched nondiabetic 2.5th–97.5th percentile interval and at least 50% from its positive reference median, for at least 30 cumulative days in the first year. Matrix adducts measured here are fructosyl-lysine, CML, CEL, MG-H1, G-H1, glucosepane and pentosidine. Measurements distinguish matrix-bound protein from unrelated intracellular expression. Site proportions are calculated within each person and then averaged equally across people.\n\nThe deviation must occur before the connected arbor's first ≥5% net physical deficit lasting 30 days. An isolated change in collagen hue, SEC31A RNA or GAP43 staining does not establish this exposure criterion. The numerical boundary specifies a substantial, persistent deviation without defining it by harm; it is not an established biological threshold.\n\n**Q — reference scaffold and trophic substitution.** For five years, an ideal local controller maintains the listed extracellular proteins' joint spatial, chemical and concentration profiles at an intact nondiabetic reference profile matched for age, sex, skin site and mechanical loading. Reference profiles are assigned independently of outcomes and preserve observed covariance; Q does not assemble an arbitrary set of independently favorable component values. It replaces deficient as well as excessive components, including precursor forms and normal matrix modifications. Repeated reference measurements allow normal aging and activity-dependent changes.\n\nThe ideal capability is molecularly selective substitution without cutting or denervating skin. Q does not replace neurons, glia, vessels or immune cells, alter neuronal growth machinery directly, or normalize circulating glucose, insulin or lipids. The comparator receives contemporary diabetes care without Q. Other initiating exposures remain comparable. Downstream changes in inflammation, vascular behavior, skin mechanics, sensory activity and neuronal signaling are allowed. Removing compensatory support, disrupting useful matrix adaptations, or adding new terminal destruction counts against Q. It is not a controller that removes only whichever components prove harmful.\n\n**Groups and endpoints.** Candidate groups are the whole eligible population, males, females, baseline BMI ≥30 kg/m², baseline HbA1c ≥7.5%, and a group with the specified panel deviation at ≥10% of interface sites during a 30-day baseline observation. No additional intersections or future responder definitions are allowed.\n\nTrack fixed territories with stable peptidergic identity and physical anatomy. Let D be cumulative terminal disappearances and R be newly established terminal entries over five years, each divided by baseline terminal amount. A replacement's subsequent disappearance is counted in D. Staining intensity alone is not disappearance or entry. With positive mean reference loss L₀ = mean(D₀ − R₀), define:\n\n- **B_R = (mean R_Q − mean R₀)/L₀:** added replacement relative to reference net loss.\n- **B_T = [L₀ − mean(D_Q − R_Q)]/L₀:** net preservation, including subsequent losses.\n- **C = P(group | incident DPN under reference care):** case coverage.\n\nHumans receive equal weight. Incident DPN means new bilateral length-dependent nerve signs with objective small-fiber loss or abnormal conduction, excluding another cause. Every claim retains only groups satisfying all predecessors; a complete NO must exclude every remaining candidate group. Exposure in one group and benefit in another cannot complete the chain."
    },
    {
      "heading": "Reasoning from the biology",
      "markdown": "A growth cone needs a traversable substrate, adhesion and signaling that permits extension and retention. Fibroblast collagen secretion can alter the substrate; keratinocytes can change ligand supply. Excess matrix deposition, chemical modification and reduced trophic activity are different disturbances. Collagen staining does not directly measure stiffness, and increased ligand expression does not specify receptor delivery or signaling.\n\nThe relevant alternatives are:\n\n`skin environment → less successful terminal replacement → net denervation`\n\nand\n\n`intrinsic neuronal injury or systemic exposure → poor regrowth and altered skin`.\n\nBoth can occur together. A neuronal deficit can make a moderately abnormal substrate consequential; it need not be absent for an extrinsic intervention to help. Q's effect asks what the environment contributes with the same initiating neuronal state, allowing all downstream interaction.\n\nA challenge that abruptly removes most terminals creates much greater renewal demand than ordinary life. Slow return after that challenge does not establish the fraction of spontaneous diabetic loss caused by failed replacement. Nor does a marker of attempted growth count as a replacement rate.\n\nFor illustration, if five-year reference net loss is 30% of baseline terminals, six percentage points of additional retained replacement meet the 20% materiality definition. If only five percentage points can be added, the maximum contribution is 16.7%. If six percentage points are added but half disappear again, net preservation is only 10%. These are accounting examples, not fitted human turnover estimates. [Calculation](../../../findings/damage_calibration_skin_quantitative.md)."
    },
    {
      "heading": "Evidence that moves the judgment",
      "markdown": "**Human skin and crossed-species coculture support an extrinsic mechanism.** Kan reported reduced diabetic skin innervation, altered collagen organization and increased fibroblast-associated SEC31A. Human dermal fibroblasts exposed to TGF-β1 changed collagen secretion and supported less rat DRG neurite growth; SEC31A silencing reversed impaired outgrowth. The accessible Figure 5 describes **six independent coculture experiments**, not six human donors, and TGF-β1 conditioning at 1, 10 and 50 ng/mL for 48 hours before four days of coculture.\n\nThis is more specific than a whole-nerve expression association. However, the recovered abstract and figure captions do not establish the human donor denominator, matched early-diabetes exposure, all culture doses or animal-linked raw outcomes. Collagen color/organization is not a direct mechanical measurement. The result supports causal substrate regulation, without supplying an early human exposure-response curve or spontaneous replacement fraction. [Kan et al., 2022](https://doi.org/10.1002/ana.26347).\n\n**Human regeneration is impaired, but the skin-specific attribution is unresolved.** Khoshnoodi reported **36 type 2, 11 type 1 and 10 control participants**, with normal proximal-thigh IENFD and sural conduction at entry. After 48-hour 0.1% capsaicin exposure, day-180 recovery was **58% of baseline in type 2 versus 91% in controls**, SD 21 and 25 percentage points. An unadjusted aggregate contrast is −33 points, approximate interval **−52 to −14**.\n\nType 2 participants were **13.7 years older** on average; their mean diagnosed duration was 4.9 years, not an enrollment restriction to ≤5 years. The adjusted initial-regrowth contrast, control minus type 2, was **0.044 fibers/mm/day (−0.008 to 0.097)**. Baseline distal-leg preservation was not established. Attrition and an internally inconsistent control row further limit precise rate reconstruction. The result supports challenge-repair vulnerability before recognized neuropathy, without separating skin from neuron or measuring everyday replacement. [Khoshnoodi et al., 2019](https://pmc.ncbi.nlm.nih.gov/articles/PMC6801164/); [aggregate checks](../../../findings/damage_calibration_skin_quantitative.md).\n\n**A skin-directed mouse intervention increases epidermal innervation.** O'Brien induced keratinocyte SIRT1 overexpression after high-fat feeding, using doxycycline **200 mg/kg of food** for two months. Figure 8G shows approximately **17.4 versus 31.6 fibers/mm**, with six visible observations per arm. Conditional on those being independent mice, digitization gives **+14.2 fibers/mm**, approximate interval **6.4–22.0**. A one-native-pixel systematic reading difference changes the contrast by about 0.87 fibers/mm.\n\nThe comparison is induced overexpression plus doxycycline versus the same construct without doxycycline; a doxycycline-only control is not described in the recovered main experiment. Glucose tolerance showed no detectable improvement, which is not proof of identical metabolic exposure. The clearest SIRT1/BDNF knockout phenotypes concerned Meissner corpuscles and Aβ touch fibers; free endings were relatively unaffected by those deletions. The authors left the overexpression-associated free-ending increase mechanistically unresolved. Thus this is positive structural evidence for skin influence, with treatment attribution, fiber subtype and repair-versus-prevention gaps. Its 82% higher endpoint density is **not** 82% prevention of diabetic loss. [O'Brien et al., 2024](https://pmc.ncbi.nlm.nih.gov/articles/PMC11449144/).\n\n**A recent human fibroblast study limits a broad cell-defect story.** Bentzen studied **30 donors**: 5 controls, 7 with diabetes without DPN, 7 with painless DPN and 11 classified with painful DPN. Diabetes duration averaged roughly 11–13 years. Passage-3/4 fibroblasts maintained at 5.5 mM glucose did not show detected group differences in mitochondrial respiration, membrane potential, ROS or measured IL-6/CXCL8 output. Respiration assays themselves used 10 mM glucose.\n\nThese small, selected groups do not establish equivalence. Culture can remove an exposure-dependent state; the study did not test neuron outgrowth on the donors' intact matrix. It weakens a claim of universal, persistent basal fibroblast mitochondrial dysfunction, while leaving induced collagen/trophic defects open. It is established-disease evidence, not a longitudinal test of R1. [Bentzen et al., 2026](https://pmc.ncbi.nlm.nih.gov/articles/PMC13005869/).\n\n**Growth-marker ratios do not resolve successful replacement.** Bönhof compared 32 recent type 2 participants, 34 painful and 32 painless DPN participants, and 50 controls. The dermal GAP43/PGP9.5 length ratio was **1.18 in painful DPN, 1.07 in painless DPN and 1.02 in controls**. Increased marker expression relative to remaining fibers is compatible with attempted repair during net loss. It cannot establish preserved replacement capacity; a shrinking denominator alone can raise a ratio. [Bönhof et al., 2017](https://pubmed.ncbi.nlm.nih.gov/28914336/).\n\nCarroll provides a useful independent endpoint check in **22 carpal-tunnel patients**, selected as 11 with increased PGP9.5 IENFD after decompression and 11 without, plus 11 healthy controls. GAP43 labeling did not distinguish those groups. Its correlation with six-month IENFD change was **−0.08 before surgery and −0.02 afterward**. Diabetes was excluded, and sampling could miss transient earlier signals. This does not invalidate GAP43's biology; it shows that a single marker measurement need not measure completed human reinnervation. [Carroll et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC9668135/).\n\nA previous public-data analysis of **77 established-DPN sural-nerve donors** likewise did not identify a useful skin-repair expression predictor: the combined score had AUROC **0.497** among 49 extreme regeneration/degeneration groups. SEC31A alone reached 0.603 with a wide, nonsignificant adjusted contrast. Wrong tissue, myelinated-nerve outcome and absent skin exposure mean this is a failed proxy, not a causal skin null. [Existing analysis](../../../model_effort/integrative_model_v1/review/repair_biology_findings.md)."
    },
    {
      "heading": "Probabilities of the necessary claims",
      "markdown": "Each probability conditions on all previous claims being true in at least one common permitted group, given the assessed evidence. No independence of biological steps is assumed.\n\n| Claim | Additional proposition | Best probability | Uncertainty range (sensitivity) | Main reason for the range |\n|---|---|---:|---:|---|\n| R1 | A group has the specified persistent panel deviation before substantial terminal deficit. | 80% | 50–95% | Human skin remodeling and diabetic trophic changes favor an exposure. Exact early compartment, dose and ordering remain sparsely measured. |\n| R2 | In a group satisfying R1, Q adds ordinary-life replacements with B_R ≥0.05. | 55% | 25–85% | Coculture and skin-directed intervention support an extrinsic effect. Human spontaneous renewal, panel sufficiency and clean compartment attribution remain unresolved. |\n| R3 | In a group satisfying R1–R2, added replacement reaches B_R ≥0.20. | 65% | 35–90% | Given a real contribution, large substrate and mouse structural effects make materiality plausible. Low ordinary renewal demand or other growth bottlenecks can keep it small. |\n| R4 | In a group satisfying R1–R3, net preservation also reaches B_T ≥0.20. | 80% | 55–95% | Established extra entries usually add some retained innervation. Continued destruction, fragile new branches and lost compensatory support can erase the gain. |\n| R5 | A group satisfying R1–R4 has C ≥0.10. | 70% | 35–90% | Skin exposures are common, but the sufficiently correctable combination may be uncommon among incident cases. Existing cohorts do not measure this denominator. |\n\nR2 is not the unconditional probability that human regeneration is impaired. It assumes the defined early skin deviation exists and prices the added causal replacement effect. R3 assumes that effect already exists, and R4 asks whether it survives competing destruction. Neither COL1 expression, SEC31A elevation nor the SIRT1–BDNF branch is an additional compulsory discount.\n\nThe preferred product is `0.80 × 0.55 × 0.65 × 0.80 × 0.70 = 0.16016`. A separate whole-claim assessment also places this around one chance in six: there is credible extrinsic biology and a relevant structural intervention, but no direct human measurement of ordinary replacement rescued before clinical DPN. The evidence does not distinguish 15% from 20% sharply."
    },
    {
      "heading": "Ideal experiments that would resolve each claim",
      "markdown": "These specify conceptual tests with unlimited measurement and intervention capability.\n\n**R1 — occurrence and timing.** Follow the full eligible population and matched nondiabetic references, measuring the named extracellular panel and continuously mapping connected physical arbors through year one. Resolve every allowed group using equal human weighting. **YES** if at least one group meets the exact deviation, site fraction, duration and pre-deficit condition. **NO** if none does. Measuring intracellular RNA, enrolling only established DPN or missing the relevant exposure window leaves R1 unresolved.\n\n**R2 — causal replacement.** In every group still satisfying R1, randomize intact, initially mapped skin to Q or reference care from entry; maintain the specified joint reference profile for five years and verify compartment specificity. Track actual entries, later disappearances and stable fiber identity. Bilateral/site experiments may estimate local effects, but the group endpoint must represent its defined distal skin territory and use humans as independent units. **YES** if any retained group has positive L₀ and B_R ≥0.05. **NO** if all fail. A marker rise, incomplete Q delivery or an imposed denervation challenge does not resolve this test.\n\n**R3 — replacement magnitude.** In the same intervention comparison, enumerate all groups satisfying R1–R2 and calculate B_R without substituting endpoint density for entries. **YES** if any reaches 0.20; **NO** if all remain below it. A wide interval crossing the boundary is unresolved, not a negative result.\n\n**R4 — durable net benefit.** Continue the same physical tracking through year five, including losses of new branches and all downstream costs. **YES** if any group satisfying R1–R3 also has B_T ≥0.20. **NO** if none does. Extra replacement combined with equal or greater destruction fails this claim even if a growth assay is positive.\n\n**R5 — coverage.** In a representative reference-care cohort, count incident DPN in all baseline groups still satisfying R1–R4 and in the full eligible population. **YES** if at least one group contains ≥10% of reference-care incident cases; **NO** if all contain less. Cases prevented by Q are not used as the denominator, and groups cannot be formed from eventual response."
    },
    {
      "heading": "Uncertainty and evidence updates",
      "markdown": "The **0.8–62% sensitivity range** comes from coherent complete interpretations:\n\n- **Challenge-specific or poorly correctable:** [50,25,35,55,35]% → **0.84%**. Apparent skin effects depend on severe denervation or artificial conditioning; endogenous early deviations and ordinary replacement benefit are smaller, and added branches often disappear.\n- **Preferred:** [80,55,65,80,70]% → **16.02%**. Extrinsic defects contribute, but their early human magnitude and coverage remain open.\n- **Common and limiting:** [95,85,90,95,90]% → **62.14%**. Natural early matrix/trophic deviations substantially constrain frequent replacement, Q removes the relevant limitation, and most added innervation persists.\n\nThese are not independent row distributions or weighted empirical models. Other plausible combinations matter: a strong but uncommon substrate defect [85,80,90,85,30]% gives **15.61%**; common abnormalities with small durable contributions [90,65,35,70,90]% give **12.90%**. Broad exposure is not enough to establish materiality.\n\nThe main uncertainty is the ordinary-life causal replacement effect. Sampling intervals for the mouse figure and human capsaicin contrast address neither its human transfer nor population coverage. Choosing 5%, 20% and 10% boundaries also affects the proposition; the accounting examples show how modest absolute replacement can cross or miss them.\n\n**Hypothetical upward evidence:** direct human early-diabetes tracking with selective panel correction, measured extra entries, retention and representative coverage could support [90,85,80,90,80]%, giving **44.06%**. **Hypothetical downward evidence:** complete target correction that improves severe challenge regrowth but leaves ordinary entries and net loss nearly unchanged could support [80,20,35,55,60]%, giving **1.85%**.\n\nChanging R2 alone to 85% or 20% gives **24.75% or 5.82%**. Resolving it with certainty leaves **29.12%** if the remaining conditionals stay unchanged. Those are arithmetic sensitivities, not expected information gains."
    },
    {
      "heading": "Overlap and boundaries",
      "markdown": "[Matrix glycation](03_extracellular_matrix_glycation.md) targets a particular chemical modification and also includes distal endoneurial matrix. Q here alters a broader protein/trophic panel but is restricted to skin and has different possible costs; the two complete propositions are not logically identical. [NGF support](14_ngf_trophic_support.md) overlaps one ligand branch, while [intrinsic growth restriction](44_intrinsic_growth_restriction.md) concerns neuronal machinery. Skin-cell senescence can be upstream of extracellular changes, but is not necessary for this claim. Their probabilities must not be added as separate prevention fractions."
    }
  ],
  "source_claim_table_markdown": "| Claim | Additional proposition | Best probability | Uncertainty range (sensitivity) | Main reason for the range |\n|---|---|---:|---:|---|\n| R1 | A group has the specified persistent panel deviation before substantial terminal deficit. | 80% | 50–95% | Human skin remodeling and diabetic trophic changes favor an exposure. Exact early compartment, dose and ordering remain sparsely measured. |\n| R2 | In a group satisfying R1, Q adds ordinary-life replacements with B_R ≥0.05. | 55% | 25–85% | Coculture and skin-directed intervention support an extrinsic effect. Human spontaneous renewal, panel sufficiency and clean compartment attribution remain unresolved. |\n| R3 | In a group satisfying R1–R2, added replacement reaches B_R ≥0.20. | 65% | 35–90% | Given a real contribution, large substrate and mouse structural effects make materiality plausible. Low ordinary renewal demand or other growth bottlenecks can keep it small. |\n| R4 | In a group satisfying R1–R3, net preservation also reaches B_T ≥0.20. | 80% | 55–95% | Established extra entries usually add some retained innervation. Continued destruction, fragile new branches and lost compensatory support can erase the gain. |\n| R5 | A group satisfying R1–R4 has C ≥0.10. | 70% | 35–90% | Skin exposures are common, but the sufficiently correctable combination may be uncommon among incident cases. Existing cohorts do not measure this denominator. |"
}