{
  "id": "13",
  "title": "C-peptide deficiency",
  "summary": "Loss of C-peptide in type 1 diabetes may remove a source of support for peripheral nerve terminals.",
  "source_title": "C-peptide deficiency removes nerve support",
  "page": "13_c_peptide_deficiency.html",
  "source_markdown": "sources/13_c_peptide_deficiency.md",
  "structured_data": "data/13_c_peptide_deficiency.json",
  "snapshot_date": "2026-09-14",
  "source_review_date": "2026-09-14",
  "scope_label": "Type 1 diabetes",
  "scope_exception": true,
  "source_headline_markdown": "**Best judgment: about 20%. Uncertainty range (sensitivity): 1–64%.** The conditional product is 18.6%. This is a **type 1 diabetes exception**, not a probability for the default type 2 population. Animal unmyelinated-fiber intervention evidence supports a structural role, but no human trial establishes prevention of terminal loss. These are subjective probabilities of a causal claim, not treatment effects or confidence intervals. Reviewed 14 September 2026. [Probability method](probability_method.md) · [Collection index](README.md).",
  "probability_percent": {
    "skeptical": 1.4,
    "base": 18.5625,
    "favorable": 63.954
  },
  "probability_meaning": "Subjective belief in the complete causal proposition. Scenario endpoints are sensitivity products, not confidence limits.",
  "causal_proposition_markdown": "In adults with type 1 diabetes, little endogenous C-peptide and no clinical DPN, physiological replacement of missing C-peptide before injury would reduce five-year net peptidergic skin-terminal loss by at least 20% in a baseline-describable group covering at least 10% of reference-care incident DPN cases in this population. This concerns prevention; restoring function or structure after established neuropathy is supporting evidence with a different time course.\n\nC-peptide is the 31-amino-acid connecting segment released when human proinsulin is processed into insulin. The specified replacement is intact, native-sequence human C-peptide, with normal free termini. Pegylated peptide, proinsulin, an isolated terminal fragment and insulin are different interventions. Native degradation products may mediate downstream effects of the intact replacement; identifying a particular receptor is not required for the whole-peptide causal claim.\n\nEligibility requires type 1 diabetes diagnosed at least one year earlier, no clinical DPN and mixed-meal-stimulated plasma C-peptide below 0.20 nmol/L on two baseline assessments. This defines low endogenous secretion; it is not itself an uncertain causal requirement. Exclude another identified cause of neuropathy, and analyze kidney clearance when interpreting peptide exposure. Type 1 disease duration need not be less than five years: the relevant stage is before clinical neuropathy. The case-coverage denominator excludes type 2 diabetes.\n\nThe relevant connected nerve unit includes the peptidergic neuron's soma, peripheral axon and skin terminals, its associated satellite/Remak Schwann cells, and the microvascular bed supplying those structures. A circulating peptide can act on endothelium or blood cells and still support the nerve indirectly. The proposition does not silently claim a neuron-exclusive receptor mechanism. Direct nerve signaling, sodium–potassium pump activity, local perfusion and altered trophic-factor availability are possible alternatives or interacting mediators.\n\nR1 specifies missing **local exposure**, beyond the blood-based eligibility definition: mean intact C-peptide availability at the extracellular surface of at least one of those connected compartments is below the age-, sex-, BMI- and kidney-function-matched nondiabetic fifth percentile for ≥90 cumulative days in year one, before the losses attributed to that deficit. Measure intact peptide in nmol/L, separating it from cross-reacting proinsulin, fragments and drug conjugates. For a blood-cell or luminal endothelial route, the corresponding compartment is local perfusing plasma. Low plasma C-peptide makes this likely but does not directly measure every tissue pool or its persistence.\n\nPermitted baseline groups are the whole eligible population, HbA1c ≥7.5%, type 1 diabetes duration ≤10 years, or fasting C-peptide <0.03 nmol/L. A group's identity is fixed before outcomes, not defined by later peptide recovery or treatment response. Follow all baseline participants and early terminal losses. The same group and connected nerve route must satisfy every requirement.\n\nThe ideal intervention Q replaces only the missing intact peptide, adjusting systemic delivery to reproduce the matched nondiabetic median arterial concentration profile across fasting, meals and overnight periods for five years. Normal distribution and clearance are allowed to operate; verify exposure at the implicated nerve unit. It does not supply a large bolus and call its daily average physiological. The reference arm receives matched carrier and usual diabetes care without peptide replacement.\n\nGlucose trajectories and circulating insulin exposure are matched between arms using ideal independent feedback control. Thus the estimand excludes an effect mediated solely by lowering systemic glucose, while allowing C-peptide to alter local insulin responsiveness. Baseline renal/vascular disease, diet and other treatments remain balanced; C-peptide-dependent changes in local perfusion, inflammation, pump activity, trophic signals and renewal may occur. Freezing those mediators would disable the biological alternatives being tested. Physiological peptide replacement is well defined even while its receptor remains uncertain.\n\nFor each participant, L is terminal disappearances minus newly established terminals over five years, divided by baseline terminal count. A new terminal that later disappears is counted in both terms. Stable peptidergic identity distinguishes true axonal changes from altered substance-P or CGRP staining. If reference-care mean loss L0 is positive, relative preservation is B=(L0−L1)/L0. Smaller structural support is B≥0.05 and material support is B≥0.20. Coverage C is the qualifying group's share of **reference-care incident DPN cases** in the eligible type 1 population, with C≥0.10. Incident DPN means new bilateral length-dependent sensory signs with objective small-fiber loss or abnormal nerve conduction, excluding other causes. The exposure duration, group cutoffs and effect boundaries define this assessment; they are not established biological thresholds.",
  "claims": [
    {
      "id": "R1",
      "source_id": "R1",
      "claim": "The defined persistent local intact-peptide deficiency exists before attributed loss.",
      "claim_markdown": "The defined persistent local intact-peptide deficiency exists before attributed loss.",
      "conditional_percent": {
        "skeptical": 70.0,
        "base": 90.0,
        "favorable": 99.0
      },
      "cumulative_base_percent": 90.0,
      "source_cumulative_display": "90%",
      "reason_markdown": "Severe secretory deficiency strongly predicts missing systemic input; exact local exposure and pre-loss persistence are less directly measured.",
      "question": "Is intact C-peptide persistently deficient at the connected nerve unit before terminal loss in type 1 diabetes?"
    },
    {
      "id": "R2",
      "source_id": "R2",
      "claim": "With R1 true, Q yields B≥0.05 in the same connected population.",
      "claim_markdown": "With R1 true, Q yields B≥0.05 in the same connected population.",
      "conditional_percent": {
        "skeptical": 25.0,
        "base": 55.0,
        "favorable": 85.0
      },
      "cumulative_base_percent": 49.5,
      "source_cumulative_display": "49.5%",
      "reason_markdown": "Rat C-fiber preservation and sequence-sensitive physiological effects support causation; human structural benefit and compensation remain unresolved.",
      "question": "Would replacing missing intact C-peptide at physiological levels avert at least 5% of five-year net terminal loss?"
    },
    {
      "id": "R3",
      "source_id": "R3",
      "claim": "With R1–R2 true, preservation reaches B≥0.20.",
      "claim_markdown": "With R1–R2 true, preservation reaches B≥0.20.",
      "conditional_percent": {
        "skeptical": 20.0,
        "base": 50.0,
        "favorable": 80.0
      },
      "cumulative_base_percent": 24.75,
      "source_cumulative_display": "24.75%",
      "reason_markdown": "Animal structural effects can be substantial, but the largest human replacement trial failed its conduction endpoint and supplies no structural magnitude.",
      "question": "Would replacing missing intact C-peptide at physiological levels avert at least 20% of five-year net terminal loss?"
    },
    {
      "id": "R4",
      "source_id": "R4",
      "claim": "With R1–R3 true, a qualifying baseline group has C≥0.10.",
      "claim_markdown": "With R1–R3 true, a qualifying baseline group has C≥0.10.",
      "conditional_percent": {
        "skeptical": 40.0,
        "base": 75.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 18.5625,
      "source_cumulative_display": "18.5625%",
      "reason_markdown": "The reference population already selects deficient secretion, making adequate coverage plausible; the causal subgroup could still be small.",
      "question": "Does the benefiting subgroup account for at least 10% of incident DPN cases in the eligible type 1 diabetes population?"
    }
  ],
  "sections": [
    {
      "heading": "The causal proposition and population",
      "markdown": "In adults with type 1 diabetes, little endogenous C-peptide and no clinical DPN, physiological replacement of missing C-peptide before injury would reduce five-year net peptidergic skin-terminal loss by at least 20% in a baseline-describable group covering at least 10% of reference-care incident DPN cases in this population. This concerns prevention; restoring function or structure after established neuropathy is supporting evidence with a different time course.\n\nC-peptide is the 31-amino-acid connecting segment released when human proinsulin is processed into insulin. The specified replacement is intact, native-sequence human C-peptide, with normal free termini. Pegylated peptide, proinsulin, an isolated terminal fragment and insulin are different interventions. Native degradation products may mediate downstream effects of the intact replacement; identifying a particular receptor is not required for the whole-peptide causal claim.\n\nEligibility requires type 1 diabetes diagnosed at least one year earlier, no clinical DPN and mixed-meal-stimulated plasma C-peptide below 0.20 nmol/L on two baseline assessments. This defines low endogenous secretion; it is not itself an uncertain causal requirement. Exclude another identified cause of neuropathy, and analyze kidney clearance when interpreting peptide exposure. Type 1 disease duration need not be less than five years: the relevant stage is before clinical neuropathy. The case-coverage denominator excludes type 2 diabetes.\n\nThe relevant connected nerve unit includes the peptidergic neuron's soma, peripheral axon and skin terminals, its associated satellite/Remak Schwann cells, and the microvascular bed supplying those structures. A circulating peptide can act on endothelium or blood cells and still support the nerve indirectly. The proposition does not silently claim a neuron-exclusive receptor mechanism. Direct nerve signaling, sodium–potassium pump activity, local perfusion and altered trophic-factor availability are possible alternatives or interacting mediators.\n\nR1 specifies missing **local exposure**, beyond the blood-based eligibility definition: mean intact C-peptide availability at the extracellular surface of at least one of those connected compartments is below the age-, sex-, BMI- and kidney-function-matched nondiabetic fifth percentile for ≥90 cumulative days in year one, before the losses attributed to that deficit. Measure intact peptide in nmol/L, separating it from cross-reacting proinsulin, fragments and drug conjugates. For a blood-cell or luminal endothelial route, the corresponding compartment is local perfusing plasma. Low plasma C-peptide makes this likely but does not directly measure every tissue pool or its persistence.\n\nPermitted baseline groups are the whole eligible population, HbA1c ≥7.5%, type 1 diabetes duration ≤10 years, or fasting C-peptide <0.03 nmol/L. A group's identity is fixed before outcomes, not defined by later peptide recovery or treatment response. Follow all baseline participants and early terminal losses. The same group and connected nerve route must satisfy every requirement.\n\nThe ideal intervention Q replaces only the missing intact peptide, adjusting systemic delivery to reproduce the matched nondiabetic median arterial concentration profile across fasting, meals and overnight periods for five years. Normal distribution and clearance are allowed to operate; verify exposure at the implicated nerve unit. It does not supply a large bolus and call its daily average physiological. The reference arm receives matched carrier and usual diabetes care without peptide replacement.\n\nGlucose trajectories and circulating insulin exposure are matched between arms using ideal independent feedback control. Thus the estimand excludes an effect mediated solely by lowering systemic glucose, while allowing C-peptide to alter local insulin responsiveness. Baseline renal/vascular disease, diet and other treatments remain balanced; C-peptide-dependent changes in local perfusion, inflammation, pump activity, trophic signals and renewal may occur. Freezing those mediators would disable the biological alternatives being tested. Physiological peptide replacement is well defined even while its receptor remains uncertain.\n\nFor each participant, L is terminal disappearances minus newly established terminals over five years, divided by baseline terminal count. A new terminal that later disappears is counted in both terms. Stable peptidergic identity distinguishes true axonal changes from altered substance-P or CGRP staining. If reference-care mean loss L0 is positive, relative preservation is B=(L0−L1)/L0. Smaller structural support is B≥0.05 and material support is B≥0.20. Coverage C is the qualifying group's share of **reference-care incident DPN cases** in the eligible type 1 population, with C≥0.10. Incident DPN means new bilateral length-dependent sensory signs with objective small-fiber loss or abnormal nerve conduction, excluding other causes. The exposure duration, group cutoffs and effect boundaries define this assessment; they are not established biological thresholds."
    },
    {
      "heading": "Conditional judgments and causal logic",
      "markdown": "Peptide deficiency alone does not prove that normal peptide supplies indispensable support. Compensation may make the deficiency harmless, or the effect may be limited to nerve physiology. Q's structural effect tests nonredundant support directly. A known GPR146 receptor, pump correction, nitric-oxide response and NGF increase are not four mandatory probability discounts. Conversely, a vibration-threshold change cannot substitute for structural preservation.\n\nThe complete proposition is R1 AND R2 AND R3 AND R4 for at least one common qualifying group and connected route. Each row after the first is conditional on all preceding rows. R2 and R3 divide small from material effects of the same intervention rather than repeat the same boundary.\n\n| Claim | Additional proposition | Best conditional YES | Uncertainty range (sensitivity) | Cumulative best | Reason |\n|---|---|---:|---:|---:|---|\n| R1 | The defined persistent local intact-peptide deficiency exists before attributed loss. | 90% | 70–99% | 90% | Severe secretory deficiency strongly predicts missing systemic input; exact local exposure and pre-loss persistence are less directly measured. |\n| R2 | With R1 true, Q yields B≥0.05 in the same connected population. | 55% | 25–85% | 49.5% | Rat C-fiber preservation and sequence-sensitive physiological effects support causation; human structural benefit and compensation remain unresolved. |\n| R3 | With R1–R2 true, preservation reaches B≥0.20. | 50% | 20–80% | 24.75% | Animal structural effects can be substantial, but the largest human replacement trial failed its conduction endpoint and supplies no structural magnitude. |\n| R4 | With R1–R3 true, a qualifying baseline group has C≥0.10. | 75% | 40–95% | 18.5625% | The reference population already selects deficient secretion, making adequate coverage plausible; the causal subgroup could still be small. |\n\nThe best product is 0.90×0.55×0.50×0.75=0.185625, rounded to about 20%. Nearby distinctions such as 15% versus 20% are not supported by empirical calibration."
    },
    {
      "heading": "Evidence and what it measures",
      "markdown": "**The largest human trial provides a real efficacy constraint.** Wahren and colleagues randomized 250 people with type 1 diabetes and mild-to-moderate neuropathy to weekly pegylated C-peptide at 0.8 mg (71 participants), 2.4 mg (73) or placebo (106) for 52 weeks. Reported plasma levels reached 1.8–2.2 and 5.6–6.8 nmol/L. Sural conduction improved by 1.0 m/s in the combined active groups and 1.2 m/s under placebo: the treatment–placebo mean contrast is **−0.2 m/s**, not a positive 1.0-m/s treatment effect. The clinical neuropathy score was unchanged. [Wahren et al., 2016](https://pubmed.ncbi.nlm.nih.gov/26884473/).\n\nVibration perception threshold improved by 4.5 µm under active treatment versus 0.1 µm under placebo, a **4.4-µm between-group improvement**. This supports a possible sensory-functional action. It is not small-fiber structure. The registry confirms established clinical neuropathy and bilateral abnormal sural conduction at entry, with conduction as the primary endpoint; it contains no posted result tables or skin-structural endpoint. The full statistical report was inaccessible, so no confidence interval is reconstructed by assuming that the abstract's unexplained ± values are standard errors. The slightly negative conduction point estimate does not establish harm. [NCT01681290](https://clinicaltrials.gov/study/NCT01681290).\n\nPegylation and delivery profile limit equivalence to native Q; measured circulating exposure does not establish normal tissue distribution or target action. These are reasons the trial is incomplete for this proposition, not evidence that it would have succeeded with another formulation. Its failed primary outcome should still lower confidence in a large broadly transferable effect.\n\n**Smaller native-peptide trials were encouraging but did not measure structure.** Ekberg's 2003 trial randomized 49 asymptomatic type 1 participants; 46 completed three months of 600 nmol/day native peptide in four subcutaneous doses or placebo. Mean diabetes duration was ten years. Baseline sural conduction was 50.9 versus 54.2 m/s in matched healthy comparators. The active group improved by 2.7 m/s, with a reported between-group p<0.05; vibration improved, heat/cold perception did not. The often-cited roughly 80% correction is 2.7 divided by the 3.3-m/s baseline deficit. It is neither a placebo-adjusted preservation fraction nor a count of saved terminals. [Ekberg et al., 2003](https://pubmed.ncbi.nlm.nih.gov/12540632/).\n\nA six-month, three-arm study of native peptide at 1.5 or 4.5 mg/day reported 139 protocol completers, mean diabetes duration 30.6 years, with clinical impairment in 86%. Positive results included more conduction responders and a 1.0-m/s improvement versus placebo in the less severely affected subgroup of 70 participants. Some other reported improvements were within active groups. The primary abstract does not establish the subgroup's prespecification or a full-cohort structural effect. The registry's enrollment field of 180 is not interchangeable with 139 completers. [Ekberg et al., 2007](https://pubmed.ncbi.nlm.nih.gov/17192336/) · [NCT00278980](https://clinicaltrials.gov/study/NCT00278980).\n\n**There is genuine unmyelinated-fiber animal evidence.** In spontaneous type 1 diabetic BB/Wor rats, native C-peptide replacement from diabetes onset prevented sural C-fiber atrophy, degeneration and fiber loss without correcting hyperglycemia. NGF/neurotrophin pathways and substance-P/CGRP content also improved. This is closer to the structural prevention claim than a pain or conduction study. However, sural unmyelinated axons are not longitudinally identified human peptidergic skin terminals, and changed peptide content alone cannot identify subtype survival. Exact animal counts and numerical treatment effects were not available from the accessible primary abstract. [Kamiya et al., 2004](https://pubmed.ncbi.nlm.nih.gov/15497155/).\n\nA subsequent treatment study from four to seven months of diabetes compared continuous replacement by osmopump with once-daily injection. Continuous delivery improved unmyelinated axon number and area, while daily injection had only marginal effects on structural and molecular measures. This is direct evidence that exposure duration can matter, although it does not validate the five-year human profile or demonstrate that pegylated peptide failed for that reason. Its complete numerical tables were inaccessible. [Kamiya et al., 2006](https://pubmed.ncbi.nlm.nih.gov/17130507/).\n\nOther BB/Wor work reported prevention of nodal abnormalities and improved myelinated-fiber regeneration with replacement from onset or months five to eight. These studies share a model and research program, so their biological relatedness should not be mistaken for multiple independent human replications. [Sima et al., 2001](https://pubmed.ncbi.nlm.nih.gov/11508275/).\n\n**Growth of control nerves can exaggerate an apparent loss phenotype.** A separate BB/Wor natural-history study used five animals per group for morphometry. Although diabetic animals developed increasing relative deficits and degenerative profiles, their absolute myelinated and unmyelinated fiber numbers did not change across the observed period; control nerves continued to mature. C-peptide in blood was 43 versus 710 pmol/L at two months, verifying severe deficiency in this model. This was not a replacement experiment. [Kamiya et al., 2009](https://doi.org/10.1155/2009/793281).\n\nThe arithmetic matters: control counts rising from 100 to 160 while diabetic counts stay at 100 create a 37.5% relative deficit without net disappearance in the diabetic group. This illustration does not erase histological degeneration or exclude impaired replacement; it explains why a cross-sectional fiber-number ratio is not automatically the specified net-loss prevention fraction in adults.\n\n**A vascular route is plausible; a specific receptor remains disputed.** In STZ rats, two weeks of C-peptide after six weeks of diabetes corrected about 62% and 78% of motor and sensory conduction deficits. Scrambled peptide did not work. Nitric-oxide-synthase inhibition blocked the conduction benefit and attenuated improved nerve perfusion, supporting a sequence-sensitive, nitric-oxide-dependent physiological route. It did not measure terminal preservation or prove a neuron-specific receptor. [Cotter et al., 2003](https://pubmed.ncbi.nlm.nih.gov/12829651/).\n\nGPR146 knockdown blocked C-peptide-induced cFos in KATOIII gastric tumor cells, motivating a proposed signaling complex. A subsequent study found no C-peptide response in GPR146-expressing CHO cells in mass-redistribution or β-arrestin assays up to 33 µM, and no internalization in its imaging assay. Different cellular contexts leave room for a required partner, but the receptor identity cannot be treated as settled. Neither result resolves the whole-peptide nerve effect; naming that receptor is unnecessary to the present Q. [Yosten et al., 2013](https://doi.org/10.1530/joe-13-0203) · [Lindfors et al., 2020](https://doi.org/10.1016/j.bmcl.2020.127208).\n\n**Recent observational evidence does not supply the missing causal comparison.** A 2026 retrospective study of 288 adults with type 2 diabetes reported less worsening of foot electrochemical skin conductance in the highest versus lowest C-peptide-change tertile: adjusted hazard ratio 0.31, 95% CI 0.14–0.69. The exposure was C-peptide at the last hospitalization minus baseline, measured over the same follow-up as the outcome. It is not a baseline exposure; excluding first-year events does not by itself make it one. The selected 126.8-pmol/L cutoff was optimized in the same data. Conductance reflects sudomotor function, not peptidergic terminal structure, and changing secretion also changes insulin physiology. This finding does not establish type 1 replacement efficacy or a causal 10% subgroup. [Primary cohort report, 2026](https://pmc.ncbi.nlm.nih.gov/articles/PMC13123299/).\n\nThe DCCT establishes that type 1 clinical neuropathy can be prevented by intensive diabetes treatment: 1,441 participants and a reported 64% reduction in confirmed neuropathy over five years. It neither isolated C-peptide nor established complete C-peptide absence in all participants; it cannot decide the independent peptide mechanism. [DCCT neuropathy report](https://pubmed.ncbi.nlm.nih.gov/7887548/)."
    },
    {
      "heading": "Ideal experiments that would resolve each claim",
      "markdown": "These are definitions of ideal resolution, not proposed studies. Recruit representative independent humans meeting the stated type 1 eligibility, determine permitted baseline-group membership, and track stable peptidergic terminals for five years. Randomize native Q versus carrier with matched glucose and circulating insulin profiles before attributed loss. Verify arterial and local intact-peptide profiles across the day, mass/sequence identity, degradation products and absence of insulin or other active contaminants. Use independent peptide manufacture and interruption/restoration of replacement to distinguish peptide effects from delivery artifacts. Permit the downstream nerve and vascular effects specified above.\n\nRetain the same group and connected route across requirements. YES requires at least one retained combination satisfying the row; NO requires every retained combination to fail. Use simultaneous precision bounds for the finite permitted groups and compartment comparisons. Equality belongs to the ≥ side for duration, benefit and coverage. Incomplete engagement, missing early losses, nonpositive reference mean loss or unstable subtype labeling leaves structural preservation unresolved.\n\n**R1 — local exposure, no predecessors.** Measure intact extracellular C-peptide in the connected neuronal, glial and microvascular compartments from baseline through year one, alongside matched nondiabetic reference distributions. YES is a qualifying mean below the local fifth percentile for ≥90 cumulative days before the attributed injury. NO requires adequate spatial/temporal measurements showing all candidates fail the level or duration criterion. A low blood meal test alone does not settle a retained tissue pool; a 20-day deficit fails the persistent-exposure claim. Peptide assays unable to distinguish intact peptide and proinsulin, or imprecise duration, remain unresolved. Human secretion measurements and the rat blood data approximate exposure, not all of this local comparison.\n\n**R2 — structural support, assuming R1.** Compare Q and carrier from baseline in the same R1-qualified group and connected route. Count disappearances and successful new terminals to estimate B. YES is B≥0.05; NO is B<0.05 with sufficiently narrow bounds across retained combinations. L0=0.30 and L1=0.28 gives B=0.067 and satisfies this row. Accurate upper bounds below 0.05 establish NO despite improved vibration threshold or blood flow. A new peptide bolus, incomplete tissue delivery, glycemic imbalance or peptide-expression readout alone is unresolved. The structural rat studies support this comparison more directly than the human functional trials, but do not replace it.\n\n**R3 — material magnitude, assuming R1–R2.** Use that same intervention, group, structural endpoint and five-year interval. YES is B≥0.20; NO under the predecessors is 0.05≤B<0.20. L0=0.30 and L1=0.24 reaches materiality, whereas L1=0.28 meets only R2. A precision interval spanning 0.20 remains unresolved; statistical significance against zero is insufficient. None of the clinical trials estimated this effect, and the rat replacement ratios cannot be translated into it without accounting for maturation, loss and renewal.\n\n**R4 — coverage, assuming R1–R3.** In a representative reference-care cohort of the specified deficient type 1 population, ascertain baseline group membership and all incident DPN cases. Estimate C=P(qualifying group | reference-care incident DPN). YES is C≥0.10 for a retained group; NO requires all qualifying groups to have C<0.10 with adequate precision. An otherwise convincing effect restricted to 2% of cases fails. Coverage intervals spanning 10%, using a type 2 denominator, or selecting people by future C-peptide recovery remain unresolved. A trial's recruited subgroup size is not this population quantity."
    },
    {
      "heading": "Sensitivity and evidence that would move the assessment",
      "markdown": "The coordinated skeptical interpretation uses [70,25,20,40]% and gives **1.4%**: the local deficit exists less consistently than plasma suggests, physiological actions are largely redundant or functional, and adult structural benefit is too small or rare. The favorable interpretation uses [99,85,80,95]% and gives **63.95%**: missing peptide reaches most connected units, continuous native replacement transfers the rat structural effect, and the qualifying population is broad. The approximately 63-percentage-point span describes sensitivity to these interpretations, not a probability distribution or sampling confidence interval. Its breadth reflects human transfer, formulation, structural magnitude and subgroup coverage, not the number of published papers.\n\nTwo additional scenarios distinguish the alternatives. A chiefly functional effect, [90,25,40,75]%, gives 6.75%. An early native-peptide effect that established-disease conjugate trials poorly capture, [95,75,65,85]%, gives 39.37%. The latter remains possible, but it is not established by explaining away trial failure.\n\nHypothetically, selective physiological replacement showing human structural preservation could raise R2 from 55% to 80%, increasing the whole estimate to **27%** with the remaining rows fixed. Repeated successful replacement with B<0.05 could lower R2 to 20%, giving **6.75%**. Evidence about magnitude or coverage should move those rows as well. Merging R1–R2 into their 49.5% joint probability preserves the headline; adding an obligatory GPR146 row would instead change the proposition.\n\nExisting accessible evidence cannot quantify five-year human peptidergic preservation. Full numerical animal replacement tables and the detailed 2016 clinical analysis were inaccessible, so their missing precision is not counted as a null. Lowering the benefit or coverage boundary with Q fixed enlarges the claim's truth set; changing peptide form, delivery profile or glucose matching defines a different claim. This mechanism may overlap [insulin/IGF support](12_insulin_igf_trophic_deficit.md), [NGF support](14_ngf_trophic_support.md) or [microvascular perfusion](15_microvascular_ischemia.md), without allowing the probabilities to be added. [Numerical support](../../../findings/damage_calibration_cpeptide_quantitative.md)."
    }
  ],
  "source_claim_table_markdown": "| Claim | Additional proposition | Best conditional YES | Uncertainty range (sensitivity) | Cumulative best | Reason |\n|---|---|---:|---:|---:|---|\n| R1 | The defined persistent local intact-peptide deficiency exists before attributed loss. | 90% | 70–99% | 90% | Severe secretory deficiency strongly predicts missing systemic input; exact local exposure and pre-loss persistence are less directly measured. |\n| R2 | With R1 true, Q yields B≥0.05 in the same connected population. | 55% | 25–85% | 49.5% | Rat C-fiber preservation and sequence-sensitive physiological effects support causation; human structural benefit and compensation remain unresolved. |\n| R3 | With R1–R2 true, preservation reaches B≥0.20. | 50% | 20–80% | 24.75% | Animal structural effects can be substantial, but the largest human replacement trial failed its conduction endpoint and supplies no structural magnitude. |\n| R4 | With R1–R3 true, a qualifying baseline group has C≥0.10. | 75% | 40–95% | 18.5625% | The reference population already selects deficient secretion, making adequate coverage plausible; the causal subgroup could still be small. |"
}