{
  "id": "35",
  "title": "Rapid glycemic correction",
  "summary": "Rapid correction of prolonged severe hyperglycemia may cause acute small-fiber injury that slower correction could partly prevent.",
  "source_title": "Rapid correction of chronic hyperglycemia precipitates structural small-fiber injury",
  "page": "35_rapid_glycemic_correction.html",
  "source_markdown": "sources/35_rapid_glycemic_correction.md",
  "structured_data": "data/35_rapid_glycemic_correction.json",
  "snapshot_date": "2026-09-14",
  "source_review_date": "2026-09-14",
  "scope_label": "Post-correction neuropathy · Six months",
  "scope_exception": true,
  "source_headline_markdown": "**Best judgment: about 16%; uncertainty range (sensitivity): 0.4–62%.** The acute clinical syndrome is more convincing than the claim that slowing correction preserves newly endangered peptidergic terminals. Human biopsies were obtained after symptoms began; prospective cohorts and the rat structural results constrain a broadly applicable injury interpretation. These are subjective probabilities of the complete causal proposition, not observed treatment success rates. Reviewed 14 September 2026. [Probability method](probability_method.md) · [Collection index](README.md).",
  "probability_percent": {
    "skeptical": 0.39375,
    "base": 16.016,
    "favorable": 62.137125
  },
  "probability_meaning": "Subjective belief in the complete causal proposition. Scenario endpoints are sensitivity products, not confidence limits.",
  "causal_proposition_markdown": "Among adults with chronically poor glycemic control, rapid glucose correction can cause an acute neuropathic syndrome with new physical small-fiber loss. Replacing a specified rapid correction with a specified slower correction would prevent at least **20% of six-month net peptidergic C-fiber terminal loss** in a baseline-defined group accounting for at least **10% of new post-correction small-fiber neuropathy cases**.\n\n**Population exception.** Eligible adults have type 1 or type 2 diabetes, HbA1c ≥10% on measurements at least six months apart, and no clinical small-fiber neuropathy at entry. Subclinical fiber loss is allowed and measured. This is an acute, exposure-defined question; its denominator and six-month horizon differ from ordinary five-year incident type 2 DPN. A result in newly diagnosed diabetes does not fully test prolonged exposure.\n\n**The two interventions.** R is an ideal glucose controller that reduces the person's baseline 24-hour mean plasma glucose linearly to 8 mmol/L over six weeks. Q reaches the same target linearly over 18 weeks. Both then maintain that mean through week 26. The within-day deviations around the scheduled mean follow the same bounded template, with no glucose below 3.9 mmol/L. These are definitions of the causal comparison, not proposed clinical treatment schedules or validated safe rates.\n\nThe controller changes glucose itself without a direct action on nerves, insulin receptors or inflammatory pathways. Baseline nutrition, medication strategy, diabetes history and other exposures are matched at assignment. Glucose-dependent adaptation, endogenous hormonal responses, microvascular changes and repair remain free to differ. The slower arm necessarily has greater cumulative glucose exposure; its resulting damage counts against Q. Holding that exposure equal would define a different intervention. This exact claim does not require true hypoglycemia.\n\n**Clinical and structural quantities.** The acute syndrome is new neuropathic pain increasing by more than three points on a 0–10 scale and/or new objectively verified autonomic dysfunction, developing over two weeks, requiring clinical attention and beginning within eight weeks of starting the controller. Ascertainment does not use the rate of HbA1c reduction as a diagnostic criterion.\n\nTrack physical terminal entries in fixed distal-leg territories from before correction. Stable peptidergic identity is established independently of current calcitonin gene-related peptide (CGRP) expression. Loss is cumulative disappearance minus newly established entries, including any subsequent loss of replacements. It is normalized to baseline terminal amount and averaged equally across people. L_R and L_Q denote six-month peptidergic losses. With L_R > 0, preservation is **B = (L_R − L_Q)/L_R**.\n\nA new post-correction small-fiber neuropathy case requires new bilateral distal small-fiber signs plus new physical skin-fiber loss below an age-adjusted reference boundary by week 26, without another cause; it does not require the glucose-rate criterion. **C** is the fraction of those cases under R belonging to the qualifying baseline group. Pain alone and an autonomic questionnaire threshold alone do not establish that case.\n\nCandidate groups are the whole eligible population, type 1 diabetes, type 2 diabetes, baseline HbA1c ≥13%, diabetes duration ≥5 years, and a documented baseline history of intentional insulin withholding for weight control. No additional intersections or future responder groups are allowed. Each successive claim requires at least one candidate satisfying every preceding requirement. Groups cannot be exchanged between links. A complete NO excludes every candidate still eligible under earlier links.",
  "claims": [
    {
      "id": "R1",
      "source_id": "R1",
      "claim": "R increases eight-week acute-syndrome incidence by at least one absolute percentage point compared with Q.",
      "claim_markdown": "R increases eight-week acute-syndrome incidence by at least one absolute percentage point compared with Q.",
      "conditional_percent": {
        "skeptical": 45.0,
        "base": 80.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 80.0,
      "source_cumulative_display": "80%",
      "reason_markdown": null,
      "question": "Does lowering mean glucose to 8 mmol/L over six weeks, versus 18 weeks, raise the eight-week risk of acute neuropathic pain or autonomic dysfunction by at least one percentage point?"
    },
    {
      "id": "R2",
      "source_id": "R2",
      "claim": "That group has positive mean new pan-small-fiber net loss under R by week 12, and R causes at least one additional baseline percentage point of that loss compared with Q.",
      "claim_markdown": "That group has positive mean new pan-small-fiber net loss under R by week 12, and R causes at least one additional baseline percentage point of that loss compared with Q.",
      "conditional_percent": {
        "skeptical": 20.0,
        "base": 55.0,
        "favorable": 85.0
      },
      "cumulative_base_percent": 44.0,
      "source_cumulative_display": "44%",
      "reason_markdown": null,
      "question": "Does six-week glucose correction cause new small-fiber loss by week 12, with at least one extra percentage point of baseline terminals lost compared with 18-week correction?"
    },
    {
      "id": "R3",
      "source_id": "R3",
      "claim": "The same group has positive six-month peptidergic loss under R and Q gives B ≥5%.",
      "claim_markdown": "The same group has positive six-month peptidergic loss under R and Q gives B ≥5%.",
      "conditional_percent": {
        "skeptical": 50.0,
        "base": 80.0,
        "favorable": 95.0
      },
      "cumulative_base_percent": 35.2,
      "source_cumulative_display": "35.2%",
      "reason_markdown": null,
      "question": "Would 18-week rather than six-week glucose correction avert at least 5% of six-month net peptidergic terminal loss?"
    },
    {
      "id": "R4",
      "source_id": "R4",
      "claim": "Its six-month peptidergic preservation reaches B ≥20%.",
      "claim_markdown": "Its six-month peptidergic preservation reaches B ≥20%.",
      "conditional_percent": {
        "skeptical": 35.0,
        "base": 70.0,
        "favorable": 90.0
      },
      "cumulative_base_percent": 24.64,
      "source_cumulative_display": "24.64%",
      "reason_markdown": null,
      "question": "Would 18-week rather than six-week glucose correction avert at least 20% of six-month net peptidergic terminal loss?"
    },
    {
      "id": "R5",
      "source_id": "R5",
      "claim": "It accounts for C ≥10% of the specified new post-correction small-fiber neuropathy cases under R.",
      "claim_markdown": "It accounts for C ≥10% of the specified new post-correction small-fiber neuropathy cases under R.",
      "conditional_percent": {
        "skeptical": 25.0,
        "base": 65.0,
        "favorable": 90.0
      },
      "cumulative_base_percent": 16.016,
      "source_cumulative_display": "16.016%",
      "reason_markdown": null,
      "question": "Does the benefiting subgroup account for at least 10% of new post-correction small-fiber neuropathy cases under six-week glucose correction?"
    }
  ],
  "sections": [
    {
      "heading": "The causal claim and its scope",
      "markdown": "Among adults with chronically poor glycemic control, rapid glucose correction can cause an acute neuropathic syndrome with new physical small-fiber loss. Replacing a specified rapid correction with a specified slower correction would prevent at least **20% of six-month net peptidergic C-fiber terminal loss** in a baseline-defined group accounting for at least **10% of new post-correction small-fiber neuropathy cases**.\n\n**Population exception.** Eligible adults have type 1 or type 2 diabetes, HbA1c ≥10% on measurements at least six months apart, and no clinical small-fiber neuropathy at entry. Subclinical fiber loss is allowed and measured. This is an acute, exposure-defined question; its denominator and six-month horizon differ from ordinary five-year incident type 2 DPN. A result in newly diagnosed diabetes does not fully test prolonged exposure.\n\n**The two interventions.** R is an ideal glucose controller that reduces the person's baseline 24-hour mean plasma glucose linearly to 8 mmol/L over six weeks. Q reaches the same target linearly over 18 weeks. Both then maintain that mean through week 26. The within-day deviations around the scheduled mean follow the same bounded template, with no glucose below 3.9 mmol/L. These are definitions of the causal comparison, not proposed clinical treatment schedules or validated safe rates.\n\nThe controller changes glucose itself without a direct action on nerves, insulin receptors or inflammatory pathways. Baseline nutrition, medication strategy, diabetes history and other exposures are matched at assignment. Glucose-dependent adaptation, endogenous hormonal responses, microvascular changes and repair remain free to differ. The slower arm necessarily has greater cumulative glucose exposure; its resulting damage counts against Q. Holding that exposure equal would define a different intervention. This exact claim does not require true hypoglycemia.\n\n**Clinical and structural quantities.** The acute syndrome is new neuropathic pain increasing by more than three points on a 0–10 scale and/or new objectively verified autonomic dysfunction, developing over two weeks, requiring clinical attention and beginning within eight weeks of starting the controller. Ascertainment does not use the rate of HbA1c reduction as a diagnostic criterion.\n\nTrack physical terminal entries in fixed distal-leg territories from before correction. Stable peptidergic identity is established independently of current calcitonin gene-related peptide (CGRP) expression. Loss is cumulative disappearance minus newly established entries, including any subsequent loss of replacements. It is normalized to baseline terminal amount and averaged equally across people. L_R and L_Q denote six-month peptidergic losses. With L_R > 0, preservation is **B = (L_R − L_Q)/L_R**.\n\nA new post-correction small-fiber neuropathy case requires new bilateral distal small-fiber signs plus new physical skin-fiber loss below an age-adjusted reference boundary by week 26, without another cause; it does not require the glucose-rate criterion. **C** is the fraction of those cases under R belonging to the qualifying baseline group. Pain alone and an autonomic questionnaire threshold alone do not establish that case.\n\nCandidate groups are the whole eligible population, type 1 diabetes, type 2 diabetes, baseline HbA1c ≥13%, diabetes duration ≥5 years, and a documented baseline history of intentional insulin withholding for weight control. No additional intersections or future responder groups are allowed. Each successive claim requires at least one candidate satisfying every preceding requirement. Groups cannot be exchanged between links. A complete NO excludes every candidate still eligible under earlier links."
    },
    {
      "heading": "Biological logic and alternatives",
      "markdown": "A nerve adapted to prolonged hyperglycemia could respond poorly to a rapid change in substrate delivery, perfusion, inflammatory signaling or trophic support. That could produce transient dysfunction, physical degeneration, or both. The exposure-level proposition does not silently require one particular mediator.\n\nThree distinctions determine the structural inference. First, small-fiber pain can begin in already damaged nerves through altered firing or regeneration. Second, a low fiber count obtained after pain begins does not date the original loss. Third, reduced CGRP staining can mean less peptide expression in surviving fibers. Neither symptoms nor peptide labeling alone establishes disappearance.\n\nSlower correction could prevent an early injury yet yield no six-month net advantage if rapid correction permits faster subsequent repair, or if prolonged hyperglycemia injures the slower arm. Conversely, a few highly susceptible people could experience substantial injury without meeting the coverage requirement. The clinical syndrome, new structural injury, target subtype, durable magnitude and coverage therefore receive separate judgments."
    },
    {
      "heading": "Evidence that moves the judgment",
      "markdown": "**A striking human referral association supports an acute trigger, with major selection and definition limits.** Gibbons and Freeman evaluated 954 tertiary neuropathy referrals, including 910 people with diabetes. Among 168 with an HbA1c fall ≥2 percentage points over three months, **104 met TIND criteria**. Among 742 below that threshold, **32 reported abrupt pain or autonomic symptoms**. The latter were not formal TIND cases because rapid HbA1c decline was part of the case definition.\n\nThe descriptive acute-syndrome contrast is **61.9% versus 4.3%, risk ratio 14.35 (approximate 95% interval 10.02–20.56)**. It is not a general-population risk estimate or a randomized effect. The reported approximately 20% risk with a 2–3-point fall and >80% with a >4-point fall apply to this referred sample. Defining TIND partly by the exposure also prevents a straightforward TIND-versus-no-TIND comparison across that boundary.\n\nThe 104 cases included **76 type 1 and 28 type 2 patients**. Their mean HbA1c changed **14.6→7.5% and 13.3→8.5%**, respectively. Mean ages were about 25 and 51 years. **62/76 type 1 cases** had a history of intentional insulin withholding/eating-disorder treatment; six type 2 cases reported substantial weight loss with severe dietary restriction. These features support susceptibility but confound a universal glucose-rate interpretation.\n\nPain began within weeks, tracked the size of the HbA1c fall, and occurred with insulin, oral agents and dietary change. This makes a single drug toxicity less satisfactory as a complete explanation. However, starting HbA1c, prior exposure duration, nutritional history, referral probability and achieved change remain entangled. Normal conduction in **56/74 tested cases** is compatible with small-fiber dysfunction but cannot count skin terminals. [Gibbons and Freeman, 2015](https://pmc.ncbi.nlm.nih.gov/articles/PMC4285188/).\n\nAs a selection check, the observed exposure–syndrome odds ratio is 36.05. Even equal source-population syndrome risks could produce that referral table if the relative case-versus-noncase referral odds differed between exposure groups by this factor. This calculation does not assert that the actual bias was that large. It shows why a large, precise referral association cannot supply the population causal probability by itself. [Calculations](../../../findings/damage_calibration_tind_quantitative.md).\n\n**Human biopsies establish abnormal structure after onset, not the amount newly lost during correction.** A 16-person series included nine type 1 and seven type 2 patients, with mean HbA1c changes **15.5→6.4% and 13.0→7.5%**. All developed severe pain within weeks. The results report distal-leg biopsies in **eight people**, all abnormal or borderline, with fiber swellings. The methods mention ten biopsied subjects, an unresolved reporting discrepancy.\n\nBiopsies were obtained within five months after pain began, without a pre-correction skin baseline. Thus existing diabetic damage and newly induced loss cannot be separated. **Three repeat biopsies one year later showed increased density**. That supports repair after the syndrome; it does not reconstruct the missing initial trajectory. Pain took a mean **15 months, range 12–28**, to decrease by at least half, so this is not evidence of complete recovery within the six-month claim window.\n\nThe series also contains a type 2 patient who lost 21 pounds during severe calorie restriction and seven type 1 women with remote insulin withholding. It does not establish a pure glucose-rate effect independent of all nutritional susceptibility. No peptidergic physical-fate assay was performed. [Gibbons and Freeman, 2010](https://pmc.ncbi.nlm.nih.gov/articles/PMC3057039/).\n\nLonger follow-up of **26 type 1 TIND patients** found improvement among 19 maintaining stable glycemia and worsening among seven with unstable control over eight years. It is observational follow-up after TIND, with no randomized correction rate or pre-event structural baseline. These same-center series should not be counted as three independent replications of incident injury. [Gibbons, 2017](https://doi.org/10.1016/j.jdiacomp.2017.01.010).\n\n**Prospective clinical evidence does not reproduce a general high-risk curve.** In a 2022 pilot, 60 patients were screened and 21 agreed to repeated examinations. Thirteen had an HbA1c decrease ≥2 points, averaging **11.2→6.4%**; eight had little change. Only **one of 13** experienced a painful clinical episode, an observed fraction **7.7% (Wilson interval 1.4–33.3%)**.\n\nThe rapid-improvement group mostly had recently diagnosed diabetes: the reported duration was 0.23 years after excluding one longstanding case, versus about 20 years in the comparison group. Thirteen of the 21 already had clinical sensory neuropathy. This is neither a clean matched rate comparison nor our prolonged-hyperglycemia, clinically intact population. It had no skin biopsies. Its limited recruitment and mostly null functional tests do not establish structural equivalence. [Hoffmann et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8892777/).\n\nThe 2026 Korean prospective report is a stronger challenge to transporting the referral curve. **Boramae: 71 people, HbA1c 11.3→6.9%; TRIPLE-AXEL: 79 newly diagnosed, drug-naïve type 2 patients, 9.5→6.4%.** Neuropathy indices did not worsen overall. Abnormal COMPASS-31 autonomic symptom scores fell **26→21 and 5→1**, respectively.\n\nSix Boramae participants developed an abnormal score, chiefly orthostatic symptoms without increased neuropathic pain. This is **6/71 = 8.5% of the whole cohort**, or **6/45 = 13.3% (6.3–26.2%) among those initially below the threshold**. Eleven therefore recovered below the threshold; the paired exact P value is 0.332. The TRIPLE-AXEL summary permits zero or one incident questionnaire abnormality, not a known count of strict TIND cases.\n\nAn autonomic questionnaire is not the clinical TIND definition, a peptidergic structural endpoint or a comprehensive measure of transient symptoms between visits. Mean HbA1c reduction also does not reveal how many individuals crossed a particular rate boundary. Nevertheless, these prospectively observed cohorts materially weaken a broadly frequent, severe neuropathy response to rapid improvement. The accessible primary abstract lacks the individual trajectories needed for a harmonized structural analysis. [Choe et al., 2026](https://doi.org/10.1016/j.diabres.2026.113388).\n\n**The rat experiment does not establish physical peptidergic loss.** Baum randomized 38 male BB/OKL rats to insulin schedules, **12/10/16** in three groups: adequate treatment from diabetes onset, persistently insufficient treatment, or insufficient treatment for one month followed by intensification. The entire intensification group did not show worse terminal nerve-conduction results than the adequately treated group.\n\nA subsequent split by achieved HbA1c decline found slower conduction and more macrophages in larger responders. That split was not randomized. In it, **CGRP-positive density was 19 versus 28**, a 32% reduction, whereas **PGP9.5 total density was 27 versus 25**, an 8% increase; the total-density comparison was not significant. This combination permits changed peptide expression or subtype composition and does not demonstrate disappearance of peptidergic terminals.\n\nThe paper reports seven rats per histology group but also conflicting specimen/subgroup totals, so a unique independent-animal structural interval cannot be reconstructed from the summaries. Seventy-five sections are not 75 animals. The absent significant total-density difference is not proof of no structural effect. Moreover, one month of inadequate control does not reproduce years of human hyperglycemia, and the study did not randomize two correction rates after identical preceding exposure. [Baum et al., 2021](https://pmc.ncbi.nlm.nih.gov/articles/PMC7913916/).\n\n**The acute claim cannot be extended to chronic harm from better control.** An individual-participant analysis of four large type 2 trials, 27,049 participants, reported a nerve-event hazard ratio of **1.02 (0.93–1.13) during year one** and **0.98 (0.87–1.09) over five years**. These largely annual, broader neuropathy outcomes are poorly suited to acute small-fiber events. They constrain a claim of substantial persistent nerve harm from intensive control, while leaving a selected acute syndrome possible. [Zoungas et al., 2017](https://doi.org/10.1016/S2213-8587(17)30104-3)."
    },
    {
      "heading": "Probabilities of the necessary claims",
      "markdown": "Every probability after R1 is conditional on **all preceding claims being true**. The low/best/high entries belong to the complete interpretations below, not independent distributions. No empirical data determine these exact percentages.\n\n| # | Necessary claim in at least one permitted group | Conditional YES: low / best / high | Best cumulative |\n|---|---|---:|---:|\n| R1 | R increases eight-week acute-syndrome incidence by at least one absolute percentage point compared with Q. | 45 / **80** / 95% | 80% |\n| R2 | That group has positive mean new pan-small-fiber net loss under R by week 12, and R causes at least one additional baseline percentage point of that loss compared with Q. | 20 / **55** / 85% | 44% |\n| R3 | The same group has positive six-month peptidergic loss under R and Q gives B ≥5%. | 50 / **80** / 95% | 35.2% |\n| R4 | Its six-month peptidergic preservation reaches B ≥20%. | 35 / **70** / 90% | 24.64% |\n| R5 | It accounts for C ≥10% of the specified new post-correction small-fiber neuropathy cases under R. | 25 / **65** / 90% | **16.016%** |\n\nR1 remains relatively high because of the close temporal pattern, exposure gradient and occurrence across treatment types. The prospective studies and definition-dependent selection prevent near-certainty for the exact rate comparison. R2 is the principal gap: abnormal post-event structure and possible regeneration favor real injury, but no paired randomized human structural trajectory or convincing rate-specific animal disappearance result settles it.\n\nConditional on actual rate-induced pan-small-fiber injury, peptidergic involvement is plausible, giving R3 a comparatively high price. It still requires six-month retention, not only early injury or current peptide expression. R4 discounts the possibility that structural loss largely reflects chronic diabetes or is repaired by six months. R5 concerns the acute case denominator, which is narrower than all DPN but has not been measured in a suitable inception cohort.\n\nThe best product is **0.80 × 0.55 × 0.80 × 0.70 × 0.65 = 0.16016**, about 16%. R1–R2 jointly receive 44%; R3–R5 receive 36.4% conditional on that. Splitting subtype, magnitude and coverage does not itself impose a penalty. The whole-claim judgment reflects the missing new-loss comparison despite substantial clinical plausibility."
    },
    {
      "heading": "Uncertainty and numerical updates",
      "markdown": "The **lower interpretation**, [45,20,50,35,25]%, gives **0.39375%**. It treats much of the clinical association as susceptibility/selection, allows acute sensitization without new structural loss, and expects little durable rate-specific benefit.\n\nThe **upper interpretation**, [95,85,95,90,90]%, gives **62.1371%**. It assumes a real adaptation injury concentrated after prolonged severe exposure, interprets the prospective cohorts as mostly outside that susceptibility state, and expects meaningful peptidergic retention in a sufficiently common acute-case subgroup.\n\nThese are jointly coherent sensitivity scenarios, not combinations of unrelated row extremes. The displayed range is about **61.7 percentage points wide**; it is not a confidence interval and has no justified model weights or variance. The preferred estimate gives the new prospective and animal constraints substantive weight without converting missing pre-correction biopsies into a negative result. Differences of a few percentage points are not supportable.\n\nOther interpretations matter. A **large effect in a very rare susceptible group**, [90,80,90,90,20]%, gives **11.664%**. A **frequent syndrome with a weak structural component**, [90,25,65,35,75]%, gives **3.8391%**. Clinical frequency and terminal injury cannot be represented by one uncertainty knob.\n\nA hypothetical matched, pre-correction randomized trajectory showing new pan-small-fiber loss, stable-identity peptidergic preservation and an adequate case denominator could move the rows together to [90,85,90,85,75]%, giving **43.8919%**. Precise preserved physical anatomy despite the clinical syndrome, across the eligible groups, could instead yield [75,20,40,35,50]%, giving **1.05%**.\n\nChanging only R2 from 55% to 85% gives **24.752%**; reducing it to 20% gives **5.824%**. Resolving R2 YES with certainty while freezing the rest gives at most **29.12%**, not the expected benefit of doing more research. Existing clinical abstracts, post-event biopsies and aggregate rat staining cannot resolve that link."
    },
    {
      "heading": "Ideal experiments that would resolve each claim",
      "markdown": "These are ideal comparisons defining truth, with unlimited measurement and selective glucose-control capability. They are not a clinical protocol. Use the entry population, finite groups, R/Q trajectories and endpoints above. Assign independent people to the controllers, measure actual glucose continuously, follow all participants regardless of symptoms, and image physical nerve identity before correction and through week 26. Neither repeated fibers nor visits inflate human n. Estimate complete-population quantities without sampling or ascertainment error.\n\nA later claim is evaluated only among groups satisfying the complete earlier prefix. If earlier truth is unknown, its conditional is unresolved. A failed controller, incomplete anatomy, missing follow-up or insufficient precision is **unresolved**, not NO.\n\n**R1 — rate-specific acute syndrome.** With matched baseline exposures and the glucose schedules achieved, determine eight-week syndrome incidence under R and Q using glucose-blind symptom and objective autonomic adjudication. **YES:** at least one group has an absolute R-minus-Q incidence difference ≥0.01. **NO:** every candidate difference is below 0.01. The referral series approaches timing and clinical characterization but not randomized exposure; prospective cohorts approach unbiased follow-up but lack this harmonized rate comparison.\n\n**R2 — newly caused physical small-fiber loss.** Within R1-qualified groups, map all small-fiber terminal entries before intervention and repeatedly through week 12, distinguishing death/disappearance from transient loss of marker expression. Calculate mean net loss from each arm's own baseline. **YES:** a group has positive R loss and R-minus-Q loss ≥0.01 of baseline terminal amount. **NO:** no R1-qualified group meets both conditions. Post-onset human biopsies demonstrate neither baseline subtraction nor the Q contrast. The rat CGRP/PGP9.5 discrepancy illustrates why stable anatomical tracking is required.\n\n**R3 — peptidergic involvement with retained benefit.** In groups meeting R1–R2, identify each peptidergic unit before correction and determine L_R, L_Q and B at week 26. Allow natural repair and prolonged-glucose costs to operate. **YES:** some group has L_R >0 and B ≥0.05. **NO:** every prefix-qualified group has nonpositive reference loss or B <0.05. Pan-neuronal biopsies and CGRP expression do not resolve physical subtype preservation or six-month net benefit.\n\n**R4 — material magnitude.** Use the same complete six-month fate counts and intervention, with no additional protective treatment. **YES:** some R1–R3 group has B ≥0.20. **NO:** all have B <0.20. For example, if R loses 10% of baseline terminals, Q must lose at most 8% to pass. Early pain relief, a transient count difference or a 20% peptide-label change cannot establish this comparison.\n\n**R5 — coverage of new acute-exposure cases.** In a complete eligible-population R arm, ascertain the defined new structural small-fiber neuropathy cases by week 26 and their baseline group membership. Combine those counts with the already established R1–R4 groups. **YES:** at least one has C ≥0.10. **NO:** no such group reaches 0.10, or there are no reference cases. This requires a population denominator, not the proportion of specialist referrals, COMPASS-31 conversions or the proportion of treated people who report pain."
    },
    {
      "heading": "Overlap and boundary",
      "markdown": "[True hypoglycemic energy failure](41_hypoglycemic_energy_failure.md) requires low glucose that this comparison excludes. [Glycemic variability](42_glycemic_variability.md) concerns oscillation; here the within-day deviation template is held constant while the mean falls. Hypoxia, inflammation and regeneration may mediate the exposure effect without being individually established by it.\n\nThe complete judgment concerns a specified acute structural consequence of correcting chronic severe hyperglycemia. It does not provide a safe rate, support continued poor control, or estimate what fraction of ordinary five-year DPN is caused by treatment."
    }
  ],
  "source_claim_table_markdown": "| # | Necessary claim in at least one permitted group | Conditional YES: low / best / high | Best cumulative |\n|---|---|---:|---:|\n| R1 | R increases eight-week acute-syndrome incidence by at least one absolute percentage point compared with Q. | 45 / **80** / 95% | 80% |\n| R2 | That group has positive mean new pan-small-fiber net loss under R by week 12, and R causes at least one additional baseline percentage point of that loss compared with Q. | 20 / **55** / 85% | 44% |\n| R3 | The same group has positive six-month peptidergic loss under R and Q gives B ≥5%. | 50 / **80** / 95% | 35.2% |\n| R4 | Its six-month peptidergic preservation reaches B ≥20%. | 35 / **70** / 90% | 24.64% |\n| R5 | It accounts for C ≥10% of the specified new post-correction small-fiber neuropathy cases under R. | 25 / **65** / 90% | **16.016%** |"
}