{
  "id": "07",
  "title": "Insufficient axonal energy",
  "summary": "Too little usable ATP in distal sensory axons may limit maintenance and terminal survival.",
  "source_title": "Insufficient usable ATP in distal sensory axons",
  "page": "07_bioenergetic_insufficiency.html",
  "source_markdown": "sources/07_bioenergetic_insufficiency.md",
  "structured_data": "data/07_bioenergetic_insufficiency.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 35%; coordinated sensitivity scenarios: 4–77%.** These are subjective probabilities of the complete human causal proposition, not the fraction of DPN caused by energy failure or a treatment-effect interval. Reviewed 14 September 2026. [Probability method](probability_method.md) · [Collection index](README.md).",
  "probability_percent": {
    "skeptical": 4.05,
    "base": 33.75,
    "favorable": 77.16375
  },
  "probability_meaning": "Subjective belief in the complete causal proposition. Scenario endpoints are sensitivity products, not confidence limits.",
  "causal_proposition_markdown": "In adults diagnosed with type 2 diabetes within five years and without clinical DPN, inadequate local ATP availability in peptidergic C-fiber axons contributes to subsequent loss of their skin terminals. Selectively correcting the local energetic state from before structural injury would reduce five-year net terminal loss by at least 20%, in a baseline-describable subgroup accounting for at least 10% of reference-care incident DPN cases.\n\nThis is a **convergent intracellular maintenance mechanism**. It does not assert a particular fuel shortage, mitochondrial enzyme defect or loss of AMPK signaling. The target compartments are the peripheral axon and terminal cytosol, including the sites serving pumps, motors, translation and membrane maintenance. A soma-only deficit, low bulk-nerve OCR or abnormal mitochondrial morphology does not establish a deficit in these compartments. Glial, vascular and somatic changes may create the axonal deficit; correcting those cells is not the intervention priced here.\n\nUsable ATP has two relevant dimensions: free ATP available to ATP-dependent reactions, and the chemical free energy released by ATP hydrolysis. Let A be free ATP activity, and let Φ be the positive magnitude of ATP-hydrolysis free energy calculated from measured ATP, ADP and phosphate activities at the local pH, free Mg and temperature. Total extracted ATP, ATP/AMP ratio and adenylate energy charge are different quantities. A smaller total ATP pool need not reduce Φ, and a maintained ATP concentration need not preserve Φ when ADP or phosphate rises.\n\nFor each measurement time and ordinary activity state, obtain matched nondiabetic fifth-percentile reference values A05 and Φ05. Define deficit burdens as the time-averaged positive shortfalls max(0, 1−A/A05) and max(0, 1−Φ/Φ05), measured over year one. R1 requires at least one group-mean burden above the corresponding nondiabetic 95th percentile, documented in intact axonal segments before the subsequent structural losses under consideration. These two alternatives distinguish kinetic substrate availability from phosphorylation potential; they are not independent mandatory causes. Burden integrates recurrent brief shortages as well as chronic depression. Loss of tissue itself cannot generate a valid deficit measurement: use surviving-compartment activities, not ATP per original biopsy mass. Include every baseline participant; do not select neurons or people by their eventual response.\n\nAllowed baseline groups are the whole eligible population, HbA1c at least 7.5%, or axonal A or Φ below its matched fifth percentile during baseline ordinary-activity sampling. Groups cannot be defined as future decliners or responders. Baseline group membership and the year-one descriptive deficit are different measurements. The same group and connected axons must satisfy the complete chain.\n\nFor each person, five-year net loss L is terminal disappearances minus newly established terminals, divided by baseline terminal count. Track stable peptidergic identity and actual structures rather than a change in peptide staining. A new terminal that subsequently disappears enters both counts. For positive reference-care mean loss L0, define B=(L0−L1)/L0. A smaller contribution is B≥0.05; materiality is B≥0.20. Coverage C is the share of **reference-care incident cases** belonging to the qualifying baseline group, and must be at least 0.10. Incident DPN means new bilateral length-dependent sensory signs with objective small-fiber loss or abnormal nerve conduction after excluding other neuropathy causes. These thresholds are chosen definitions, not estimated constants.",
  "claims": [
    {
      "id": "R1",
      "source_id": "R1",
      "claim": "Intact distal axons show the defined excess burden of low A or low Φ before subsequent terminal loss.",
      "claim_markdown": "Intact distal axons show the defined excess burden of low A or low Φ before subsequent terminal loss.",
      "conditional_percent": {
        "skeptical": 45.0,
        "base": 75.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 75.0,
      "source_cumulative_display": "75%",
      "reason_markdown": null,
      "question": "Do intact distal axons have unusually frequent or severe shortages of free ATP or energy available from ATP hydrolysis before terminal loss?"
    },
    {
      "id": "R2",
      "source_id": "R2",
      "claim": "Correcting those local energetic deficits with E reduces five-year net terminal loss by at least 5%.",
      "claim_markdown": "Correcting those local energetic deficits with E reduces five-year net terminal loss by at least 5%.",
      "conditional_percent": {
        "skeptical": 40.0,
        "base": 75.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 56.25,
      "source_cumulative_display": "56.25%",
      "reason_markdown": null,
      "question": "Would directly restoring local ATP availability and ATP-hydrolysis energy avert at least 5% of five-year net terminal loss?"
    },
    {
      "id": "R3",
      "source_id": "R3",
      "claim": "The same E benefit reaches at least 20%.",
      "claim_markdown": "The same E benefit reaches at least 20%.",
      "conditional_percent": {
        "skeptical": 45.0,
        "base": 75.0,
        "favorable": 90.0
      },
      "cumulative_base_percent": 42.1875,
      "source_cumulative_display": "42.19%",
      "reason_markdown": null,
      "question": "Would directly restoring local ATP availability and ATP-hydrolysis energy avert at least 20% of five-year net terminal loss?"
    },
    {
      "id": "R4",
      "source_id": "R4",
      "claim": "A group satisfying R1–R3 accounts for at least 10% of reference-care incident DPN cases.",
      "claim_markdown": "A group satisfying R1–R3 accounts for at least 10% of reference-care incident DPN cases.",
      "conditional_percent": {
        "skeptical": 50.0,
        "base": 80.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 33.75,
      "source_cumulative_display": "33.75%",
      "reason_markdown": null,
      "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 diagnosed with type 2 diabetes within five years and without clinical DPN, inadequate local ATP availability in peptidergic C-fiber axons contributes to subsequent loss of their skin terminals. Selectively correcting the local energetic state from before structural injury would reduce five-year net terminal loss by at least 20%, in a baseline-describable subgroup accounting for at least 10% of reference-care incident DPN cases.\n\nThis is a **convergent intracellular maintenance mechanism**. It does not assert a particular fuel shortage, mitochondrial enzyme defect or loss of AMPK signaling. The target compartments are the peripheral axon and terminal cytosol, including the sites serving pumps, motors, translation and membrane maintenance. A soma-only deficit, low bulk-nerve OCR or abnormal mitochondrial morphology does not establish a deficit in these compartments. Glial, vascular and somatic changes may create the axonal deficit; correcting those cells is not the intervention priced here.\n\nUsable ATP has two relevant dimensions: free ATP available to ATP-dependent reactions, and the chemical free energy released by ATP hydrolysis. Let A be free ATP activity, and let Φ be the positive magnitude of ATP-hydrolysis free energy calculated from measured ATP, ADP and phosphate activities at the local pH, free Mg and temperature. Total extracted ATP, ATP/AMP ratio and adenylate energy charge are different quantities. A smaller total ATP pool need not reduce Φ, and a maintained ATP concentration need not preserve Φ when ADP or phosphate rises.\n\nFor each measurement time and ordinary activity state, obtain matched nondiabetic fifth-percentile reference values A05 and Φ05. Define deficit burdens as the time-averaged positive shortfalls max(0, 1−A/A05) and max(0, 1−Φ/Φ05), measured over year one. R1 requires at least one group-mean burden above the corresponding nondiabetic 95th percentile, documented in intact axonal segments before the subsequent structural losses under consideration. These two alternatives distinguish kinetic substrate availability from phosphorylation potential; they are not independent mandatory causes. Burden integrates recurrent brief shortages as well as chronic depression. Loss of tissue itself cannot generate a valid deficit measurement: use surviving-compartment activities, not ATP per original biopsy mass. Include every baseline participant; do not select neurons or people by their eventual response.\n\nAllowed baseline groups are the whole eligible population, HbA1c at least 7.5%, or axonal A or Φ below its matched fifth percentile during baseline ordinary-activity sampling. Groups cannot be defined as future decliners or responders. Baseline group membership and the year-one descriptive deficit are different measurements. The same group and connected axons must satisfy the complete chain.\n\nFor each person, five-year net loss L is terminal disappearances minus newly established terminals, divided by baseline terminal count. Track stable peptidergic identity and actual structures rather than a change in peptide staining. A new terminal that subsequently disappears enters both counts. For positive reference-care mean loss L0, define B=(L0−L1)/L0. A smaller contribution is B≥0.05; materiality is B≥0.20. Coverage C is the share of **reference-care incident cases** belonging to the qualifying baseline group, and must be at least 0.10. Incident DPN means new bilateral length-dependent sensory signs with objective small-fiber loss or abnormal nerve conduction after excluding other neuropathy causes. These thresholds are chosen definitions, not estimated constants."
    },
    {
      "heading": "What correcting energy availability means",
      "markdown": "The ideal intervention E supplies external chemical energy to regenerate ATP locally when A or Φ would fall below its matched normal range. It maintains both quantities within that range during ordinary activity, without changing systemic glycemia, lipid exposure, oxygen delivery or trophic input. It can replenish an inadequate adenylate pool if needed, with the necessary matter and energy accounted for. It does not directly repair mitochondria, silence SARM1, replace proteins, restore cargo transport or suppress inflammation. Those processes may change as consequences of corrected intracellular ATP availability.\n\nThis requires capabilities unavailable today, but not energy generation without fuel. An orthogonal ATP-regeneration system driven by an explicitly supplied external energy source, local feedback measurement, and selective delivery to axoplasm defines the intended comparison. Maintain extracellular ATP/adenosine exposure at its reference trajectory to exclude an extracellular purinergic treatment. Intracellular nucleotide sensing may respond. Thus the proposition concerns the consequences of correcting intracellular ATP availability, including its signaling consequences; it does not identify which ATP-consuming process mediates preservation.\n\nATP supply from oxidative phosphorylation, glycolysis, phosphotransfer and supporting-cell substrates is interchangeable only to the extent that it reaches the relevant compartment at the needed time. At a steady state, production equals consumption; a persistent claim that actual consumption indefinitely exceeds production violates mass balance. Injury could instead arise because ATP-dependent work slows, demand is curtailed, or the free-energy state settles below that needed for normal performance. A falling respiratory capacity does not prove this. For example, capacity 100 with demand 50 still meets demand after a 30% capacity reduction, leaving reserve 20. Those are illustrative units, not measured axonal rates.\n\nFailure of one source can be compensated by another. Conversely, transport distance, local demand or adenylate loss can matter without lower mitochondrial maximum capacity. Hence “reduced capacity,” “increased demand” and “loss of reserve” are possible explanations of R1, not additional necessary rows. Pumps, motors, proteostasis, membrane renewal and nucleotide sensing are alternative downstream mediators. The causal structural comparison determines whether the observed deficit actually matters."
    },
    {
      "heading": "Probability decomposition",
      "markdown": "Each row is conditional on all previous rows in the same baseline group and connected axons. Retain only groups satisfying the cumulative requirements. No independence assumption is made.\n\n| Claim | Additional proposition | Skeptical / best / favorable | Cumulative best |\n|---|---|---:|---:|\n| R1 | Intact distal axons show the defined excess burden of low A or low Φ before subsequent terminal loss. | 45 / **75** / 95% | 75% |\n| R2 | Correcting those local energetic deficits with E reduces five-year net terminal loss by at least 5%. | 40 / **75** / 95% | 56.25% |\n| R3 | The same E benefit reaches at least 20%. | 45 / **75** / 90% | 42.19% |\n| R4 | A group satisfying R1–R3 accounts for at least 10% of reference-care incident DPN cases. | 50 / **80** / 95% | 33.75% |\n\nR1 receives substantial but incomplete support from animal nerve energetics and the physical demands of axonal maintenance. Human pre-loss local ATP/free-energy measurements remain absent. R2 is high conditional on a real deficit, but does not assume that every below-normal value is limiting or that restoring ATP reverses all other injuries. R3 addresses magnitude after accepting a smaller causal effect. R4 is relatively high because the route can receive several common diabetic inputs, while its actual incident-case coverage remains unmeasured. These are elicited judgments; no paper estimates these four conditional probabilities."
    },
    {
      "heading": "Evidence and its scope",
      "markdown": "**Mitochondrial capacity changes are real in several diabetic models, but they are not ATP flux.** Roy Chowdhury and colleagues measured DRG enzyme activities in STZ mice. At four weeks, diabetic/control ratios for complex-I-linked activity, complex IV and citrate synthase were 1.03, 0.94 and 0.95; at fourteen weeks they were 0.71, 0.79 and 0.85. Groups contained 5–8 animals depending on assay. Adult diabetic-rat cultures had reduced spare capacity, while coupling efficiency was slightly higher; changing glucose acutely to 25 mM did not reproduce the full phenotype. Resveratrol 5 mg/kg/day during weeks 9–18 improved several metabolic measures and skin nerve profiles (n=10–11), without correcting hyperglycemia. AMPK constructs and inhibitors modified outgrowth in 3–4 culture replicates. These supply established-disease metabolic and structural interventions, but resveratrol is not an ATP clamp and the assays are not paired local ATP–terminal trajectories. [Roy Chowdhury et al., 2012](https://pmc.ncbi.nlm.nih.gov/articles/PMC3359752/).\n\n**Other models retain abundant bulk energy.** Alloxan-diabetic rat sciatic nerve had increased ATP, phosphocreatine and energy charge at the reported 3-, 6- and 26-week sampling periods, with elevated ketones and no changed lactate/pyruvate ratio. This primary abstract supplies a directionally important counterexample to universal whole-nerve ATP depletion, but not compartment-specific ATP activities; the accessible record did not provide the table-specific independent n or uncertainty. A bulk increase cannot exclude a distal deficit, and that possibility cannot be treated as an observed result. [Thurston et al., 1995](https://pubmed.ncbi.nlm.nih.gov/7859940/).\n\nThe 24-week db/db tracer study found reduced glucose-derived TCA labeling in sciatic nerve (n=8/group for glucose), while labeled pyruvate bypassed part of the difference. This supports altered substrate handling, without determining whether glycolysis, alternative fuels or lower demand maintain ATP. Bolus labeling and static metabolite amounts do not measure the proposed A/Φ state. [Sas et al., 2016](https://insight.jci.org/articles/view/86976).\n\n**ATP and AMPK must be measured separately.** In a 2025 db/db study, four mice/group supplied the nucleotide panel and three/group the AMPK blot. AdipoRon, 30 mg/kg from age 16–20 weeks, was associated with improved sciatic-nerve metabolic/signaling measures. Digitized source means give diabetic/control ATP about 0.73, AMP about 3.11, and pAMPK/AMPK about 0.62; treatment raised ATP and pAMPK while lowering AMP. This directly challenges an inference that low AMPK necessarily indicates high ATP or low AMP. Whole nerve, absent ADP, and incomplete mapping of the assay's concentration axis prevent a local free-energy calculation. [Strengthening monocarboxylate transporters by adiponectin receptor agonist, 2025](https://doi.org/10.1186/s12964-025-02326-5).\n\nIn the IMS32 Schwann-cell experiment, one-hour pyruvate withdrawal at 15 mM glucose retained about 82% of ATP and 80% of total adenylates while calculated energy charge rose by 0.013. Its metabolite table lacks replicate dispersion and an adequately explained cell denominator; AMPK was not measured. This establishes an important distinction among observables, not a demonstrated neuronal energy-sensing mechanism. [Yako et al., 2021](https://doi.org/10.1038/s41598-021-98082-w).\n\n**Large ATP loss does not determine axonal fate by itself.** In primary mouse sensory neurons, 50 μM CCCP reduced ATP to less than 15% of control within 4–6 hours in both wild-type and Sarm1-null neurons. Yet Sarm1-null axons survived; adding 5 mM 2-deoxyglucose defeated that protection. The combined-treatment structural panel used three experimental replicates, not three human donors. This separates a mitochondrial insult, residual glycolytic support and execution of degeneration. It does not imply ATP is dispensable, or that a selective ATP restoration would fail. [Summers et al., 2014](https://pmc.ncbi.nlm.nih.gov/articles/PMC4087211/).\n\nSimilarly, NMNAT3 protected mouse DRG axons against 2.5 μM rotenone despite substantial ATP decline. Control ATP fell about 55% at six hours and 83% at 96 hours; NMNAT3 cultures started with and retained slightly higher ATP. A 20 mM 2-deoxyglucose experiment lowered ATP by 37% at 24 hours without degeneration in the later 48-hour images. ATP was measured in six wells with technical duplication across two experiments, not six animals. These observations reject a requirement to preserve the original ATP level; they cannot exclude an unmeasured nonlinear local threshold or combine mismatched assay times into a universal ATP–survival curve. [Press and Milbrandt, 2008](https://doi.org/10.1523/JNEUROSCI.0525-08.2008).\n\n**Human cellular evidence supplies temporal possibility, with a different insult.** Human iPSC-derived dopamine neurons exposed to 100 μM vacor showed adenylate-pool contraction before detectable fragmentation at four hours. Digitized mean ATP+ADP+AMP was 57% of vehicle in untreated cells and 103% with SARM1 ASO. The displayed measurements represent four donor lineages with unequal repeats across differentiations, not nine independent people. The published and reconstructed measurements are normalized to pellet protein and do not recover free distal ATP activities. They show that nucleotide loss can precede visible injury and be prevented by upstream SARM1 intervention; they do not isolate ATP restoration or model incident DPN. [Loreto et al., 2025](https://pmc.ncbi.nlm.nih.gov/articles/PMC7617922/).\n\n**Human skin places altered mitochondria in the affected compartment.** Hamid and colleagues studied seven controls, seven diabetic people without DPN and seven with DPN, with similar mean ages around 53 years. Distal-leg IENFD was 5.8±1.2, 3.8±2.0 and 1.0±0.7 fibers/mm respectively, SD. Diabetic groups averaged about 70–77 months of diabetes; this was cross-sectional. The mitochondrial-size analysis concerned fluorescently segmented organelles in remaining fibers, not respiratory capacity or ATP. Neither individual organelles nor multiple images increase n beyond seven/group. The result supports compartment relevance and a pathology association, without establishing energetic causation or timing. [Hamid et al., 2014](https://pmc.ncbi.nlm.nih.gov/articles/PMC4241807/).\n\n**Mitochondrial transfer is supportive evidence with a donor limitation.** Xu and colleagues showed mouse nerve injury after impaired glia-to-neuron mitochondrial transfer and protection in several injury models. Human diabetic SGC MYO10 staining was lower at the donor level (four controls/five diabetic donors). The functional transfer panel has only two versus three donor experiments; the ten cultures/group in a human OCR panel are not ten independent donors. The mouse co-culture source traces also show that lower total or basal OCR can coexist with higher oligomycin-sensitive respiration, so one scalar cannot convert OCR into ATP supply. Transfer from damaged glia can harm neurons. This supports a role for organelle quality and support, without proving that ATP shortage mediates the structural effects. [Xu et al., 2026](https://doi.org/10.1038/s41586-025-09896-x).\n\n**Human structural rescue is possible, but its energetic mediation is unknown.** The topical pirenzepine trial randomized 58 people with established T2 DPN; 57 entered modified intention-to-treat analysis. At 24 weeks the pooled treated-ankle IENFD change was +1.8 versus −0.6 fibers/mm, with reported baseline-adjusted between-group p=.012. Other efficacy outcomes did not differ in that analysis. This supports local structural plasticity, not prevention or an ATP-specific effect. [Sivadasan et al., 2025](https://pmc.ncbi.nlm.nih.gov/articles/PMC12721300/). Its mechanistic mouse/culture programme links M1 antagonism to AMPK, mitochondrial spare capacity and outgrowth, but the AMPK factorials do not measure respiration and outgrowth jointly. [Calcutt et al., 2017](https://www.jci.org/articles/view/88321). The separate completed trial NCT04786340 has no recovered numerical outcome report; it cannot be counted as replication. [Trial record](https://clinicaltrials.gov/study/NCT04786340).\n\nNo accessible human study directly restores the defined local A/Φ state before DPN and measures five-year subtype-stable terminal loss. Broad antioxidant, metabolic or trophic interventions are neither equivalent positives nor equivalent negatives for E. Rare inherited respiratory-chain disorders establish that mitochondrial defects can cause neuropathy, but do not establish common diabetic energy limitation; mitochondrial effects on calcium, redox and biosynthesis must still be separated from ATP availability."
    },
    {
      "heading": "Ideal experiments that would resolve each claim",
      "markdown": "These are ideal truth conditions, not claims of present feasibility. Use a population-representative early-T2 cohort and an age-, sex- and activity-matched nondiabetic reference group. Baseline subgroup assignment precedes intervention and outcome observation. People are the independent units. Noninjurious local measurement must preserve the terminals being tracked. Measure A, ADP, phosphate, pH, Mg and Φ through ordinary activity and rest; include local spatial gradients and transient shortages. Hold other initiating exposures comparable, including reference diabetes care, glucose, oxygenation, lipids, trophic signals and extracellular purines. Intracellular responses downstream of E may change. All comparisons concern the same connected axons and eligible baseline group.\n\n**R1 — a pre-loss local energetic deficit.** There are no assumed predecessors. Measure the two defined year-one deficit burdens in intact peripheral axon/terminal compartments, together with subsequent structural trajectories. Directly sample ATP-dependent work and total ATP-production/consumption flux to interpret compensation, but do not substitute these for A/Φ. YES requires at least one permitted group's mean burden above its matched nondiabetic 95th-percentile boundary, with the abnormal state preceding the structural losses attributed to it. NO requires both burdens at or below their boundaries for every permitted group. Account for all baseline participants, including early losses, rather than retaining only intact survivors at year one. Boundary-spanning simultaneous intervals, missing high-demand periods, indirect probes without calibration or measurements only after fragmentation are unresolved. R1 does not yet establish causal damage. Existing studies supply animal bulk pools, capacities and human pathology, not this human temporal measurement.\n\n**R2 — a smaller structural contribution.** Assuming R1 for a permitted baseline group, randomize E versus sham from cohort entry for five years. Confirm continuous local energetic correction at rest and during demand, with both A and Φ in their matched normal range. Quantify the external energy and adenylate delivery and exclude direct effects on extracellular purines or other initiating exposures. Use a second method that achieves the same intracellular energetic trajectory; restoration of the original low-energy trajectory can distinguish correction from an unrelated device effect. Measure baseline terminal number, disappearances and successful entries with stable peptidergic identity. YES requires B≥0.05; NO requires B<0.05, with a confidence interval wholly on the appropriate side. A result such as L0=.30 and L1=.28 gives B=.067: it can establish this smaller contribution while failing R3. Inadequate engagement, a normal soma but still-deficient terminal, imprecise effect estimation or marker-only rescue leaves R2 unresolved. Current metabolic treatments do not implement E, even when structural rescue is observed.\n\n**R3 — material preservation.** Assuming R1–R2, estimate B in the same five-year randomized comparison, preserving the reference loss denominator and propagating its uncertainty. YES is B≥0.20; NO is B<0.20 at adequate precision. With L0=.30, L1=.24 lies exactly on the truth boundary; .23 meets and .28 fails it. This separates magnitude from the already-established smaller effect. Nonpositive true reference loss fails the specified relative-loss proposition; a noisy near-zero estimate remains unresolved. Improved conduction or pain cannot resolve this structural claim. Existing intervention studies address different doses, stages and mechanisms, so none estimates this conditional magnitude directly.\n\n**R4 — common enough to matter.** Assuming a baseline group satisfies R1–R3, measure its share C of all incident DPN cases in the population-representative reference-care arm. Use the same five-year follow-up and outcome definition, accounting for follow-up loss and competing death when defining the reference incident-case population. YES requires C≥0.10; NO requires C<0.10 for every otherwise qualifying group, with sufficiently precise intervals. A group containing 15% of incident cases qualifies and one containing 3% does not, regardless of how large its individual benefit is. A selected rare-disease series, an enriched trial alone or the fraction of remaining cases among treated people cannot resolve C. Existing human cohorts do not supply the needed energetic classification plus incident-case coverage."
    },
    {
      "heading": "Calibration and remaining uncertainty",
      "markdown": "The best conditional inputs [75,75,75,80]% multiply to 33.75%, rounded to about 35%. The skeptical scenario [45,40,45,50]% assumes that bulk mitochondrial findings mostly reflect compensation or established damage, that correcting local ATP leaves major injury routes active, and that material benefit is restricted. Its product is 4.05%. The favorable scenario [95,95,90,95]% assumes a recurrent pre-loss deficit common across diabetic inputs, effective energetic rescue and broad case coverage; its product is 77.16%. These are coordinated interpretations of shared evidence, not draws from independent row distributions.\n\nThe human pre-loss deficit and the effect of selective energetic correction are the major unknowns. If only R1 fell from 75% to 40%, the product would be 18%; if it rose to 95%, 42.75%. Actual compartment-resolved causal evidence would also change R2 and possibly the later rows. Combining R2 and R3 gives 56.25% conditional on R1 and leaves the headline unchanged. Splitting supply, reserve and demand into more prose does not justify further discounts.\n\nThe definition permits transient shortages through integrated burden rather than requiring a permanently depleted ATP pool. A broader definition including soma-only energy defects would be a different proposition; stronger clinical evidence for repair in established DPN does not automatically raise early-prevention probability. Replacing the 20% terminal threshold by 10% broadens the event and replacing it by 30% narrows it. Existing data cannot price that entire response curve or a credible distribution of human effect sizes."
    },
    {
      "heading": "Relationship to other hypotheses",
      "markdown": "This route can mediate [fatty-acid injury](08_fatty_acid_delivery.md), [ischemia](15_microvascular_ischemia.md), [glial metabolic support](17_schwann_metabolic_support.md), [mitochondrial dynamics](25_mitochondrial_dynamics.md) and [PARP/NAD loss](28_parp_nad_depletion.md). [SARM1](27_sarm1_axon_destruction.md) can both create energetic collapse and determine whether stressed axons fragment. These are overlapping relationships, not independent votes for ATP limitation. [Glucose-driven respiratory overload](06_glucose_mitochondrial_overload.md) can be false while local ATP becomes limiting; neither hypothesis is inferred from the other."
    }
  ],
  "source_claim_table_markdown": "| Claim | Additional proposition | Skeptical / best / favorable | Cumulative best |\n|---|---|---:|---:|\n| R1 | Intact distal axons show the defined excess burden of low A or low Φ before subsequent terminal loss. | 45 / **75** / 95% | 75% |\n| R2 | Correcting those local energetic deficits with E reduces five-year net terminal loss by at least 5%. | 40 / **75** / 95% | 56.25% |\n| R3 | The same E benefit reaches at least 20%. | 45 / **75** / 90% | 42.19% |\n| R4 | A group satisfying R1–R3 accounts for at least 10% of reference-care incident DPN cases. | 50 / **80** / 95% | 33.75% |"
}