{
  "id": "03",
  "title": "Extracellular matrix glycation",
  "summary": "Glycation of the surrounding matrix may prevent sensory terminals from replacing ordinary losses.",
  "source_title": "Glycated extracellular matrix prevents terminal replacement",
  "page": "03_extracellular_matrix_glycation.html",
  "source_markdown": "sources/03_extracellular_matrix_glycation.md",
  "structured_data": "data/03_extracellular_matrix_glycation.json",
  "snapshot_date": "2026-09-14",
  "source_review_date": "2026-09-14",
  "scope_label": "Early type 2 diabetes",
  "scope_exception": false,
  "source_headline_markdown": "**About 15% probability; uncertainty range (sensitivity) 0.8–55%.** The central product is 15.5%. Modified growth substrates can restrict sensory axon growth, and people with diabetes show impaired reinnervation before clinical neuropathy. The largest uncertainty is whether naturally glycated human matrix causes enough failed replacement during ordinary life to account for material terminal loss. Literature reviewed 14 September 2026.",
  "probability_percent": {
    "skeptical": 0.819,
    "base": 15.47,
    "favorable": 55.233
  },
  "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, without confirmed distal polyneuropathy and with initially preserved distal-leg innervation, nonenzymatic glycation of extracellular matrix impedes replacement of peptidergic C-fiber terminals. Selectively normalizing the specified matrix modifications before prospective loss would add replacements equivalent to at least 20% of reference five-year net loss **and** avert at least 20% of that net loss, in a baseline subgroup containing at least 10% of incident DPN cases under reference care.\n\nThe matrix is the extracellular protein scaffold in distal-leg dermis, its epidermal basement membrane, and connected distal endoneurial sheaths that growing axons traverse. The named proteins are collagen I, III and IV, laminins and fibronectin. The specified modifications are fructosyl-lysine, carboxymethyl-lysine, carboxyethyl-lysine, methylglyoxal- and glyoxal-derived hydroimidazolones (MG-H1 and G-H1), glucosepane and pentosidine. They include early sugar additions, non-crosslinking advanced glycation end products (AGEs), and crosslinks; these are not interchangeable measures of stiffness.\n\nThe ideal intervention restores these protein-bound adduct occupancies to matched nondiabetic values and maintains them there for five years. It preserves the proteins' initial amounts, amino-acid sequences, normal enzymatic crosslinks and other initiating exposures. Changes in adhesion, physical organization, stiffness, receptor signaling and matrix turnover caused by that chemical correction may occur. It does not replace the entire diabetic skin, remove all free carbonyl compounds, or suppress every RAGE ligand. Removing named chemical modifications is the idealized capability; a present drug's description as an “AGE breaker” does not establish this specificity.\n\nAllowed baseline groups are the whole eligible population, HbA1c ≥7.5%, or a group with at least one specified local protein-bound adduct above its matched nondiabetic 95th percentile. The exposure requirement is a group mean local occupancy above that reference for at least 90 cumulative days in year one, beginning before prospective terminal loss. Groups are fixed by pretreatment characteristics, never future response. The same group, connected matrix territory and intervention must satisfy the complete chain. These explicit subgroup and persistence rules bound the question; they are not measured toxicity thresholds.\n\nTrack a fixed skin territory using stable peptidergic identity. Let D be terminal disappearances and R be new terminal entries over five years, each divided by baseline terminal number. Track each entry and its subsequent fate, counting a later disappearance in D. Net loss is L = D − R. With a positive mean reference loss L₀, define:\n\n- **Replacement contribution B_R = (mean R₁ − mean R₀)/mean L₀.** This measures extra entries relative to the amount of reference loss.\n- **Net preservation B_T = (mean L₀ − mean L₁)/mean L₀.** This includes any compensating loss of new or existing terminals.\n\nBoth must reach 0.20. Increased short-lived sprouts alone fail the net-preservation requirement; pure reduction of destruction without additional replacement fails the replacement claim. Coverage C is the fraction of reference-care incident DPN cases in a qualifying group, requiring C ≥0.10. Incident DPN means new bilateral length-dependent sensory signs plus objective small-fiber loss or abnormal nerve conduction, without another neuropathy. Recovery after an imposed lesion and repair of established DPN inform the mechanism but are different outcomes.",
  "claims": [
    {
      "id": "R1",
      "source_id": "R1",
      "claim": "At least one allowed group has the specified persistent pre-loss matrix glycation.",
      "claim_markdown": "At least one allowed group has the specified persistent pre-loss matrix glycation.",
      "conditional_percent": {
        "skeptical": 65.0,
        "base": 85.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 85.0,
      "source_cumulative_display": null,
      "reason_markdown": "Rat endoneurial chemistry and human skin measurements support natural accumulation. Early human adduct occupancy at the actual terminal growth interface is less well established.",
      "question": "Is glycation of the extracellular matrix around growing sensory terminals persistently elevated before terminal loss?"
    },
    {
      "id": "R2",
      "source_id": "R2",
      "claim": "In a group satisfying R1, matrix-specific normalization produces additional ordinary-life replacements with B_R ≥0.05.",
      "claim_markdown": "In a group satisfying R1, matrix-specific normalization produces additional ordinary-life replacements with B_R ≥0.05.",
      "conditional_percent": {
        "skeptical": 20.0,
        "base": 50.0,
        "favorable": 80.0
      },
      "cumulative_base_percent": 42.5,
      "source_cumulative_display": null,
      "reason_markdown": "Adult sensory-neuron substrate experiments supply a causal growth effect. Human dose matching, matrix specificity and replacement during unprovoked early disease remain unresolved.",
      "question": "Would correcting matrix glycation 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, additional replacement reaches B_R ≥0.20.",
      "claim_markdown": "In a group satisfying R1–R2, additional replacement reaches B_R ≥0.20.",
      "conditional_percent": {
        "skeptical": 35.0,
        "base": 65.0,
        "favorable": 85.0
      },
      "cumulative_base_percent": 27.625,
      "source_cumulative_display": null,
      "reason_markdown": "Some imposed substrate effects are large. Ordinary turnover and competing growth limitations could make even a real human contribution small.",
      "question": "Would correcting matrix glycation 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": 45.0,
        "base": 80.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 22.1,
      "source_cumulative_display": null,
      "reason_markdown": "Material extra replacement will often preserve the population, but excess disappearance or short-lived growth could cancel it. This row assumes the added replacements already exist.",
      "question": "Would correcting matrix glycation also avert at least 20% of five-year net terminal loss, after accounting for subsequent terminal disappearance?"
    },
    {
      "id": "R5",
      "source_id": "R5",
      "claim": "A group satisfying R1–R4 contains C ≥0.10 of reference incident cases.",
      "claim_markdown": "A group satisfying R1–R4 contains C ≥0.10 of reference incident cases.",
      "conditional_percent": {
        "skeptical": 40.0,
        "base": 70.0,
        "favorable": 90.0
      },
      "cumulative_base_percent": 15.47,
      "source_cumulative_display": null,
      "reason_markdown": "Glycation is common, but the prevalence of a subgroup whose loss is materially replacement-limited is unknown.",
      "question": "Does the benefiting subgroup account for at least 10% of incident DPN cases under reference care?"
    }
  ],
  "sections": [
    {
      "heading": "The causal proposition",
      "markdown": "In adults with type 2 diabetes diagnosed within five years, without confirmed distal polyneuropathy and with initially preserved distal-leg innervation, nonenzymatic glycation of extracellular matrix impedes replacement of peptidergic C-fiber terminals. Selectively normalizing the specified matrix modifications before prospective loss would add replacements equivalent to at least 20% of reference five-year net loss **and** avert at least 20% of that net loss, in a baseline subgroup containing at least 10% of incident DPN cases under reference care.\n\nThe matrix is the extracellular protein scaffold in distal-leg dermis, its epidermal basement membrane, and connected distal endoneurial sheaths that growing axons traverse. The named proteins are collagen I, III and IV, laminins and fibronectin. The specified modifications are fructosyl-lysine, carboxymethyl-lysine, carboxyethyl-lysine, methylglyoxal- and glyoxal-derived hydroimidazolones (MG-H1 and G-H1), glucosepane and pentosidine. They include early sugar additions, non-crosslinking advanced glycation end products (AGEs), and crosslinks; these are not interchangeable measures of stiffness.\n\nThe ideal intervention restores these protein-bound adduct occupancies to matched nondiabetic values and maintains them there for five years. It preserves the proteins' initial amounts, amino-acid sequences, normal enzymatic crosslinks and other initiating exposures. Changes in adhesion, physical organization, stiffness, receptor signaling and matrix turnover caused by that chemical correction may occur. It does not replace the entire diabetic skin, remove all free carbonyl compounds, or suppress every RAGE ligand. Removing named chemical modifications is the idealized capability; a present drug's description as an “AGE breaker” does not establish this specificity.\n\nAllowed baseline groups are the whole eligible population, HbA1c ≥7.5%, or a group with at least one specified local protein-bound adduct above its matched nondiabetic 95th percentile. The exposure requirement is a group mean local occupancy above that reference for at least 90 cumulative days in year one, beginning before prospective terminal loss. Groups are fixed by pretreatment characteristics, never future response. The same group, connected matrix territory and intervention must satisfy the complete chain. These explicit subgroup and persistence rules bound the question; they are not measured toxicity thresholds.\n\nTrack a fixed skin territory using stable peptidergic identity. Let D be terminal disappearances and R be new terminal entries over five years, each divided by baseline terminal number. Track each entry and its subsequent fate, counting a later disappearance in D. Net loss is L = D − R. With a positive mean reference loss L₀, define:\n\n- **Replacement contribution B_R = (mean R₁ − mean R₀)/mean L₀.** This measures extra entries relative to the amount of reference loss.\n- **Net preservation B_T = (mean L₀ − mean L₁)/mean L₀.** This includes any compensating loss of new or existing terminals.\n\nBoth must reach 0.20. Increased short-lived sprouts alone fail the net-preservation requirement; pure reduction of destruction without additional replacement fails the replacement claim. Coverage C is the fraction of reference-care incident DPN cases in a qualifying group, requiring C ≥0.10. Incident DPN means new bilateral length-dependent sensory signs plus objective small-fiber loss or abnormal nerve conduction, without another neuropathy. Recovery after an imposed lesion and repair of established DPN inform the mechanism but are different outcomes."
    },
    {
      "heading": "Biological logic and alternatives",
      "markdown": "Glucose and reactive carbonyls can modify extracellular proteins faster than those proteins are replaced. A modified integrin-binding site may weaken growth-cone anchoring and signaling; altered crosslinking can also change access and mechanical resistance. Neither a higher total AGE concentration nor a stiffer tissue proves that the particular sites contacted by growing axons are impaired. Different matrix proteins, sites and trophic conditions may produce different effects. RAGE signaling is one possible consequence, not a required mediator.\n\nReduced extension becomes failed replacement only if terminals need replacing while that matrix is present. A large regeneration defect after capsaicin or nerve crush does not measure ordinary replacement demand. Intrinsic neuronal growth restriction, inadequate trophic signals, glial support or continued axonal destruction can produce the same slow increase in fiber density. Density curves alone cannot identify entry and disappearance separately.\n\nThe scale matters. If reference five-year net loss is 30% of baseline terminals, the materiality definition requires at least 6% of baseline number in additional replacements and at least 6 percentage points less net loss. If only 5% of baseline number could be replaced through this route, even perfect correction could preserve at most 16.7% of reference loss. These are illustrative arithmetic scenarios, not measured human turnover rates."
    },
    {
      "heading": "Conditional probabilities",
      "markdown": "Each row assumes all earlier requirements hold in at least one common allowed group. The cumulative event retains only groups satisfying every preceding requirement. Adhesion and growth signaling are explanatory alternatives; they are not separately discounted when the intervention's replacement effect is measured directly.\n\n| Claim | Additional proposition | Best probability | Uncertainty range (sensitivity) | Reason for estimate and breadth |\n|---|---|---:|---:|---|\n| R1 | At least one allowed group has the specified persistent pre-loss matrix glycation. | 85% | 65–95% | Rat endoneurial chemistry and human skin measurements support natural accumulation. Early human adduct occupancy at the actual terminal growth interface is less well established. |\n| R2 | In a group satisfying R1, matrix-specific normalization produces additional ordinary-life replacements with B_R ≥0.05. | 50% | 20–80% | Adult sensory-neuron substrate experiments supply a causal growth effect. Human dose matching, matrix specificity and replacement during unprovoked early disease remain unresolved. |\n| R3 | In a group satisfying R1–R2, additional replacement reaches B_R ≥0.20. | 65% | 35–85% | Some imposed substrate effects are large. Ordinary turnover and competing growth limitations could make even a real human contribution small. |\n| R4 | In a group satisfying R1–R3, net preservation also reaches B_T ≥0.20. | 80% | 45–95% | Material extra replacement will often preserve the population, but excess disappearance or short-lived growth could cancel it. This row assumes the added replacements already exist. |\n| R5 | A group satisfying R1–R4 contains C ≥0.10 of reference incident cases. | 70% | 40–90% | Glycation is common, but the prevalence of a subgroup whose loss is materially replacement-limited is unknown. |\n\nBase [85,50,65,80,70]% gives **15.47%**. Coordinated skeptical [65,20,35,45,40]% gives **0.819%**; favorable [95,80,85,95,90]% gives **55.233%**. Merging the replacement-causation and replacement-magnitude rows yields 32.5%, preserving the joint probability. A separate whole-claim judgment is plausible but more likely false than true: the substrate mechanism has direct experimental support, while human natural-dose attribution and ordinary replacement demand remain unmeasured. The evidence does not distinguish 15% from 20% reliably."
    },
    {
      "heading": "Evidence supporting and challenging the claim",
      "markdown": "**Adult sensory neurons respond to the substrate, but not uniformly.** Duran-Jimenez measured increased glycation products in diabetic rat endoneurial matrix over 3–24 weeks, with 6–8 diabetic animals and four controls per time point. Growth assays used four independent adult rat DRG culture experiments. Laminin or fibronectin coatings were exposed to 500 μM methylglyoxal for 24 hours, washed three times, then received neurons for 18 hours. With nerve growth factor (NGF), digitized longest-neurite means on control versus modified laminin were approximately **481 versus 308 μm**, while fibronectin gave **326 versus 324 μm**. Branching/density retained approximately 40% and 78%, respectively. Aminoguanidine during modification prevented substantial inhibition. The separate protein-chemistry preparations were dialyzed; that does not establish removal of free methylglyoxal from the washed growth coatings. Modification occupancy in those coatings was not matched to an early human nerve sample. The experiment supports substrate-dependent restriction and trophic compensation, not one universal matrix-quality multiplier. [Duran-Jimenez 2009](https://doi.org/10.2337/db09-0320); [quantitative support](../../../findings/damage_calibration_matrix_quantitative.md).\n\n**Adhesion loss need not imply the same growth effect.** Bennmann modified laminin or collagen IV with 1 mM methylglyoxal for four hours. PC12 cell adhesion fell about 80–90% in three independent experiments, yet images showed no obvious neurite-growth difference across modified/unmodified borders. That visual comparison is not a precise equivalence test, and the surviving adherent cells may be selected. Directly modifying the cells instead suppressed subsequent neurite outgrowth by over 90%, in two independent experiments. This separates extracellular substrate effects from direct cellular glycation and cautions against inferring one from the other. [Bennmann 2014](https://doi.org/10.1371/journal.pone.0112115).\n\n**A human-derived cell model supplies a positive at artificial exposure.** Differentiated SH-SY5Y neuroblastoma cells had reduced neurite length after 72 hours with 100 μg/mL glycated collagen, in three independent assays with technical triplicates; viability was preserved. The collagen was prepared with 200 mM ribose, 160 mM glucose and 200 mM threose for seven days, then digested and solubilized. This is a modified-collagen preparation acting on one tumor-derived cell line, not intact adult human peptidergic axons on a naturally glycated matrix. The reported soluble preparation does not isolate physical substrate resistance. [Primary study, 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8773477/). A 2025 protocol from this model measures capsaicin-induced calcium sensitization; it does not add a human structural prevention result. [Primary protocol](https://www.jove.com/t/68036/advanced-glycation-end-products-sensitize-human-sensory-like-neuron).\n\n**Human prospective associations are real but do not identify the mediator.** In the DCCT/EDIC skin-biopsy study, 165 participants with type 1 diabetes were free of the study's clinical and autonomic neuropathy outcomes at baseline; 74 subsequently developed the clinical-or-autonomic composite. In a selected multivariable model, collagen MG-H1 had adjusted OR **2.3 per SD, 95% CI 1.4–3.7**, whereas furosine had OR **1.6, 0.8–3.2** after glycemic and other risk-factor adjustment. The correlated-marker selection and composite endpoint prevent treating this as a specific matrix-to-peptidergic mechanism estimate. [Genuth 2015](https://pmc.ncbi.nlm.nih.gov/articles/PMC4274803/).\n\nIn ZODIAC, 662 people with type 2 diabetes without baseline monofilament neuropathy contributed 66 incident cases over mean 3.1 years. Skin autofluorescence predicted neuropathy with adjusted OR **1.50, 95% CI 1.05–2.14**. Adjustment for current HbA1c cannot fully remove cumulative glycemic exposure. Autofluorescence mixes signals and does not locate the causal protein or adduct. This study provides temporal relevance to incident disease, but its monofilament outcome is not peptidergic terminal replacement. [Gerrits 2008](https://pure.rug.nl/ws/files/2699820/Gerrits_2008_Diab_Care.pdf).\n\n**Recent human imaging remains an association between proxies.** A 2025 study compared 29 people with type 2 diabetes and neuropathy, 33 without neuropathy, and ten controls. Among diabetic participants, skin autofluorescence correlated with sciatic fractional anisotropy at r=−0.49; after age and renal-function adjustment, r=−0.30, P=.039. In the subgroup without neuropathy, adjusted r=−0.36, P=.075. Diabetes duration averaged 8.6 years without and 12.5 years with neuropathy. Skin autofluorescence alone discriminated existing neuropathy with AUROC **0.697, 95% CI 0.534–0.860**. This extends the human association to nerve imaging but measures neither matrix adduct identity nor terminal replacement, and does not establish the early five-year exposure claim. [Mooshage 2025](https://doi.org/10.1007/s00062-024-01493-1).\n\n**Human reinnervation is impaired, without identifying extracellular causation.** Polydefkis analyzed 31 controls and 19 diabetic participants, including eleven without and eight with neuropathy, after 48-hour capsaicin denervation. Mean reinnervation rates were **0.177, 0.100 and 0.040 fibers/mm/day**, respectively. The diabetes association survived baseline fiber-density adjustment. The diabetic group mixed ten type 1 and nine type 2 participants, had median diabetes duration 15 years, and was older than controls. It establishes a preclinical-neuropathy regeneration deficit, not necessarily a defect early after diabetes onset or a glycated-matrix cause. [Polydefkis 2004](https://doi.org/10.1093/brain/awh175).\n\nThe later six-month series included 36 type 2 and eleven type 1 participants without neuropathy and ten controls. The adjusted initial-rate contrast, control minus type 2, was **0.044 fibers/mm/day, 95% CI −0.008 to 0.097**. Published control baseline, final density and change summaries are internally inconsistent, so they do not support an exact reconstructed paired effect. Repeated density measurements still cannot distinguish continued disappearance from restricted entry. [Khoshnoodi 2019](https://pmc.ncbi.nlm.nih.gov/articles/PMC6801164/).\n\n**A drug label does not supply a selective matrix test.** In eight-week STZ-diabetic mice, alagebrium 1 mg/kg/day during the final two weeks improved vascular responses but left mean conduction velocity approximately **43 versus 43 m/s** in treated versus untreated diabetic groups, twelve mice/group. Tail-flick latency was approximately **6.8 versus 7.2 seconds**, seven/group. Neither endpoint measured terminal structure, and the study did not demonstrate normalization of the specified local matrix adducts. The result limits claims of readily translated functional benefit but does not precisely exclude the matrix-replacement mechanism. [Demiot 2006](https://doi.org/10.2337/db05-1433). A glycated reconstructed-skin study reported nerve-network disruption and limited alagebrium rescue, but the accessible abstract does not recover independent n or isolate matrix effects from prolonged glyoxal exposure to cells. It receives little additional weight. [Cadau 2015](https://pubmed.ncbi.nlm.nih.gov/25771012/).\n\nNo selective human matrix-normalization experiment with terminal entry and disappearance was located through the review date. Current skin chemistry, imaging and regeneration studies cannot be combined into a measured attributable fraction merely because their directions agree."
    },
    {
      "heading": "Ideal experiments that would resolve each claim",
      "markdown": "These are truth definitions, not proposed work. Assume noninjurious repeated local chemistry, selective restoration of the named protein-bound adducts, and stable five-year tracking of human peptidergic terminal entries and disappearances. Sample independent people in the defined early type 2 population. Use fixed baseline groups and matched nondiabetic reference distributions. Randomize matrix correction versus intact matrix, with comparable glucose, lipids, insulin, renal function, ordinary injury exposure and diabetes care; permit downstream matrix, immune, glial and neuronal responses. Verify adduct occupancy, protein abundance and preserved non-glycation chemistry. Reintroduce the original adduct pattern in a restoration arm and use a second independent correction to exclude intervention artifacts. No imposed denervation is part of the resolving prevention comparison.\n\n**R1 — persistent pre-loss matrix exposure.** With no predecessor assumed, measure each specified adduct in mol per mol parent protein at the terminal growth interfaces through year one, alongside stable terminal tracking. YES requires at least one fixed group's mean occupancy above the matched 95th percentile for ≥90 cumulative days, beginning before prospective loss. NO requires every allowed group to fail that exposure, duration or timing boundary. A fluorescent signal without chemical identity, unlocalized bulk tissue, incomplete pre-loss sampling or uncertainty spanning the boundary remains unresolved. Rat nerve chemistry and human skin measurements supply natural glycation evidence, but not the complete early human localized trajectory. Injury is not assumed by this row.\n\n**R2 — a small ordinary replacement effect.** Assuming R1, randomize the specified five-year matrix correction in each surviving group and quantify R₁−R₀ with the reference net-loss denominator. YES is B_R ≥0.05 in at least one group; NO is B_R <0.05 in all. At L₀=0.30, an additional 0.02 of baseline terminals entering gives B_R=6.7% and passes. Verify chemical specificity and restoration; permit trophic signaling and growth responses to change. Incomplete correction, uncertain entry identity, a nonpositive reference-loss denominator or an interval spanning 5% remains unresolved. Adult culture experiments establish growth capability, but not replacement during ordinary human early disease. This row does not establish material magnitude or net preservation.\n\n**R3 — material additional replacement.** Assuming R1–R2, estimate the same B_R against 0.20. At L₀=0.30, at least 0.06 additional entries per baseline terminal count passes; an additional 0.02 fails this row while satisfying R2. YES is B_R ≥0.20 in a surviving group, NO is below 0.20 in all, and finite bounds crossing 20% remain unresolved. Ordinary terminal demand is measured directly in this experiment, not added as a separate assumed repair requirement. Existing forced-injury and substrate assays do not measure that demand or this fraction.\n\n**R4 — material net preservation.** Assuming R1–R3, separately measure disappearances D as well as entries R and compare B_T with 0.20. At L₀=0.30, intervention loss ≤0.24 passes. An intervention adding 0.06 entries but also adding 0.03 disappearances gives B_R=20% yet B_T=10%: R3 YES, R4 NO. YES requires B_T ≥0.20 in a surviving group; NO requires less in all. Failure to track the fate of entrants, peptide-expression changes without structures, or uncertainty spanning 20% remains unresolved. Net density recovery studies supply a broader structural outcome but cannot establish the separate replacement and disappearance quantities required here.\n\n**R5 — reference-case coverage.** Assuming R1–R4, follow a representative reference-care population for five years with pretreatment group classification and incident-DPN ascertainment. Estimate C, the share of reference cases in each qualifying group. YES is at least one C ≥0.10; NO is C <0.10 for all. For example, 1,500 of 10,000 reference cases passes and 500 fails with adequate precision. Do not add overlapping groups unless their union satisfies the prior effect criteria. Selected chemistry samples, missing incidence or bounds spanning 10% remain unresolved. Prospective AGE associations do not supply the prevalence of a group meeting the matrix-specific replacement criteria."
    },
    {
      "heading": "Why uncertainty remains broad",
      "markdown": "The skeptical scenario treats glycation mainly as a record of historical exposure, with intrinsic growth restriction and continuing destruction explaining human regeneration deficits. The favorable scenario assigns a substantial part of those deficits to naturally modified growth interfaces, with enough ordinary turnover for correction to preserve terminals. The span is a coordinated sensitivity analysis of subjective judgments, not a confidence interval on either the experimental effect or the probability. No defensible probability distribution or variance has been inferred from the number of studies.\n\nHypothetically, a selective human correction at measured natural occupancy that increases ordinary replacement could raise R2 from 50% to 80%, taking the complete probability to **24.8%** with other inputs fixed. Precisely excluding a 5% replacement contribution despite verified correction could lower R2 to 10%, giving **3.1%**. A new autofluorescence association alone would not justify either update. Changing the required 20% preservation, 10% case coverage, adduct list or persistence boundary changes the scientific question; it is distinct from uncertainty about the stated proposition.\n\nThe exposure overlaps [carbonyl stress](05_methylglyoxal_carbonyl_stress.md), and some effects may pass through [RAGE](04_rage_signaling.md). The replacement outcome also overlaps [skin repair environment](32_skin_repair_environment.md) and [intrinsic growth restriction](44_intrinsic_growth_restriction.md). Those mechanisms are neither independent alternatives nor compulsory serial factors in this probability."
    }
  ],
  "source_claim_table_markdown": "| Claim | Additional proposition | Best probability | Uncertainty range (sensitivity) | Reason for estimate and breadth |\n|---|---|---:|---:|---|\n| R1 | At least one allowed group has the specified persistent pre-loss matrix glycation. | 85% | 65–95% | Rat endoneurial chemistry and human skin measurements support natural accumulation. Early human adduct occupancy at the actual terminal growth interface is less well established. |\n| R2 | In a group satisfying R1, matrix-specific normalization produces additional ordinary-life replacements with B_R ≥0.05. | 50% | 20–80% | Adult sensory-neuron substrate experiments supply a causal growth effect. Human dose matching, matrix specificity and replacement during unprovoked early disease remain unresolved. |\n| R3 | In a group satisfying R1–R2, additional replacement reaches B_R ≥0.20. | 65% | 35–85% | Some imposed substrate effects are large. Ordinary turnover and competing growth limitations could make even a real human contribution small. |\n| R4 | In a group satisfying R1–R3, net preservation also reaches B_T ≥0.20. | 80% | 45–95% | Material extra replacement will often preserve the population, but excess disappearance or short-lived growth could cancel it. This row assumes the added replacements already exist. |\n| R5 | A group satisfying R1–R4 contains C ≥0.10 of reference incident cases. | 70% | 40–90% | Glycation is common, but the prevalence of a subgroup whose loss is materially replacement-limited is unknown. |"
}