Kidney failure and uremic exposure

In adults with diabetes and advanced chronic kidney disease, reduced renal handling of a finite solute panel produces abnormal blood and local nerve exposures before further structural loss. Correcting those exposures would preserve at least 20% of five-year net peptidergic C-fiber terminal loss in a baseline group accounting for at least 10% of new or progressive distal neuropathy in this renal-disease population.

This is a secondary-exposure/progression scope exception. Include T1D and T2D, existing neuropathy, and either eGFR <30 mL/min/1.73 m² for ≥3 months or maintenance dialysis. It does not claim that mild albuminuria causes early incident DPN, or that diabetes must have caused the kidney disease.

The panel has ten members: potassium, urea, β2-microglobulin, indoxyl sulfate, p-cresyl sulfate, p-cresyl glucuronide, hippurate, phenylacetylglutamine, indoleacetate and trimethylamine N-oxide. The last seven cover candidate retained metabolites also produced through gut/host metabolism. β2-microglobulin represents a specific middle-sized protein, not every molecule removed by filtration.

Other renal consequences—anemia, altered erythropoietin or vitamin-D production, phosphate, acidosis, nutrient deficiency, volume overload and treatment-related ischemia—are alternatives outside this panel. They can mediate downstream effects of its correction, but cannot independently make the panel hypothesis true.

Q — circulating panel correction. For five years, a selective blood controller supplies/removes only these ten species to reproduce their joint concentration trajectories in a diabetes reference population with preserved renal function: eGFR ≥90 and UACR <30 mg/g for ≥3 months, matched for sex, age within five years, diabetes type, diabetes duration within five years, HbA1c within one percentage point and BMI within 3 kg/m². For each participant, draw one complete reference trajectory from that matched distribution; preserve its temporal and cross-solute relationships rather than independently choosing favorable values.

For protein-bound metabolites, control both free and total concentration with selective exchange that leaves albumin amount and unrelated ligands unchanged. Potassium correction includes the full daily trajectory, not only predialysis sampling. Hold initially assigned dialysis, other treatment and nutrition comparable. Renal function, circulation, inflammation, nutrient handling and subsequent physiology may respond; all downstream costs and compensation count. Q is not kidney transplantation or a particular dialysis membrane.

Source and timing. R1 requires at least one panel member to be abnormal for ≥30 cumulative days in year one in ≥10% of a qualifying group. For potassium, abnormal means ≥4.8 mmol/L; for other panel members, the free concentration must exceed both twice the matched-reference median and its 95th percentile. These are operational exposure boundaries, not validated injury thresholds.

At least half of that excess must be attributable to deficient renal handling. Define the source comparison by restoring only that solute's renal extraction/clearance function to its matched-reference concentration-dependent function, keeping its nonrenal production and initial distribution equal. For potassium, this includes renal excretion control; for filtered proteins it includes renal uptake/catabolism. Other kidney functions remain initially unchanged. This prevents high microbial production alone from satisfying a renal-retention claim.

R2 requires the same abnormal solute to create ≥50% excess local free exposure relative to its matched reference, or ≥0.3 mmol/L excess for potassium, with Q removing at least half of that excess. Locate it at the sensory neuron, axon, terminal, connected supporting cell or its local microvessel interface. The local condition must involve units serving ≥10% of mapped terminals in ≥10% of people, for ≥30 cumulative days in year one, before a further ≥5% physical net deficit lasting 30 days. A high plasma value without that connected local exposure does not suffice.

Outcome and groups. Track stable-identity peptidergic terminals in fixed bilateral distal-leg and dorsal-foot territories. Loss means physical disappearance minus newly established entries, also counting later loss of replacements. Peptide downregulation is not disappearance. Normalize to baseline terminal amount, average territories within people and people equally, and compare five-year losses L₀ and L_Q. For L₀ >0, B = (L₀ − L_Q)/L₀.

A case is new bilateral length-dependent neuropathy with objective small-fiber loss or abnormal conduction, or established neuropathy with a further ≥10% baseline-normalized physical pan-small-fiber deficit. Include renal-related second-cause neuropathy; symptoms alone do not define structural progression. C is the fraction of reference-care cases belonging to the qualifying group.

The six candidate groups are all eligible people, T1D, T2D, maintenance-dialysis patients, nondialysis patients with eGFR 15–29, and people with baseline potassium ≥5.0 mmol/L. Membership precedes Q. There are no additional intersections or future-response groups. Each later claim requires at least one candidate satisfying all preceding requirements; NO excludes every remaining candidate.

Source assessment

Best judgment: about 24%; uncertainty range (sensitivity): 0.68–69%. Advanced kidney dysfunction can expose nerves to abnormal circulating chemicals. Human skin denervation and a potassium intervention support this possibility, but neither identifies the five-year peptidergic benefit of correcting a specified renal exposure mixture. These are literature-informed subjective probabilities of that complete claim. Reviewed 14 September 2026. Probability method · Collection index.

Biological reasoning

Renal retention changes exposure without requiring hyperglycemia. Potassium can depolarize nerve membranes; prolonged depolarization may increase energetic load or disturb transport. Retained organic solutes may act directly, through local vascular injury, or through supporting cells. These are alternative routes. A single named receptor, oxidant or cell type is not mandatory.

Electrical dysfunction can improve within hours without regrowth. Conversely, damaged terminals may fail to recover even after a circulating exposure normalizes. Neither acute excitability nor persistence of an old deficit determines whether earlier correction prevents further loss.

Diabetes and CKD also share age, vascular disease, cumulative metabolic exposure, medication and nutritional differences. A nondiabetic CKD cohort removes diabetes as a necessary cause, but does not isolate the retained panel. A kidney transplant changes far more than clearance and can add immunosuppressant neurotoxicity.

Primary evidence

There is substantial human small-fiber pathology without diabetes. Chao studied 40 consecutive nondiabetic CKD patients, mean age 60.7 years: two had stage 3, six stage 4 and 32 stage 5 disease. Distal-leg intraepidermal density was 2.8±2.0 versus 8.6±2.8 fibers/mm in age/sex-matched controls, approximately 67% lower. 27/40 had density below the reference threshold; an ordinary Wilson interval for this selected case-series proportion is approximately 52–80%.

The independent unit is the patient, not sections or fibers. The accessible primary material does not establish the control n sufficiently to reconstruct a unique effect interval, so none is invented. Density correlated with renal-disease duration after adjustment for age, sex and BMI. There was no significant density difference between symptomatic and asymptomatic patients.

This is a direct physical small-fiber anchor for advanced CKD, stronger than an isolated pain association. It remains a cross-section against healthy controls, without exposure-specific correction, a diabetic renal comparison or peptidergic-specific fate tracking. The 67% density gap is not a 67% preventable renal-toxin fraction. Chao et al., 2011.

Selective potassium timing affects human nerve physiology. In four dialysis patients, investigators held serum potassium near 5.0 mmol/L for the first three hours of a modified dialysis session while other solutes were removed, then used low-potassium dialysis for four more hours. Median motor-nerve excitability remained abnormal during the potassium clamp. In routine sessions in the same people, potassium and excitability improved after three hours.

This is a useful within-person intervention linking potassium to membrane depolarization. It does not demonstrate physical injury, C-fiber loss or the effect of a five-year controller. Four participants do not become a large n through repeated time points. Arnold et al., 2014.

A randomized potassium strategy slowed a clinical neuropathy score. A 24-month study randomized 47 stage 3–4 CKD patients, 23 to restriction and 24 to control. Five withdrew before treatment; 21 per arm entered the analysis. Fourteen analyzed participants had CKD attributed to diabetes. Dialysis, transplantation and eGFR <15 were excluded.

The intervention targeted serum potassium ≤4.5 mmol/L using dietary restriction, adding 15–30 g/day sodium polystyrene sulfonate when needed. Control participants received equally frequent dietary contact, with potassium intervention above 6.0 mmol/L. Mean potassium during follow-up was approximately 4.6 versus 4.8 mmol/L.

Total neuropathy score increased 0.4±2.2 versus 2.8±3.3 points. From the published change SDs and n=21/arm, the advantage is 2.4 points, approximate 95% interval 0.64–4.16, standardized difference about 0.86. Table 4 gives the control final score as 8.9, whereas the abstract says 8.6; the table and reported change agree.

Gait speed improved more, but the sural-amplitude comparison was nonsignificant, P=0.19, without a precise structural bound. No skin-terminal measurement was made. Diet plus a binding resin is not a unique potassium-only intervention, and the trial could not estimate a diabetes-specific effect reliably. It nevertheless provides randomized evidence beyond an association with CKD severity. Arnold et al., 2017; calculations (background note outside this collection).

The larger dialysis trial was negative, with limited sustained exposure separation. FINESSE randomized 124 people, 63 to hemodiafiltration and 61 to high-flux hemodialysis, with mean follow-up 41 months. Seventy-three percent had neuropathy at baseline. Modified total neuropathy score worsened 1.7 versus 1.2 points out of 28; the treatment-minus-control difference was +0.5, 95% interval −0.7 to +1.7. No diabetes interaction was detected.

This interval argues against a two-point average clinical advantage of the tested strategy. It permits smaller effects and does not measure peptidergic structural preservation. Twenty-four participants had baseline data only; there were 31 deaths, 16 versus 15. Models used available yearly assessments without imputation, so “124 randomized” does not mean 124 complete four-year nerve trajectories.

Per-session β2-microglobulin removal improved 66.3% versus 57.0%, yet predialysis concentration differed only −1.4 mg/L, interval −4.0 to +1.2, with means 28.9 versus 30.3. Predialysis potassium, phosphate and small-solute clearance did not differ materially. This is a genuine clinical null for added hemodiafiltration, but not a demonstrated normalization of the full panel's continuous exposure. Clearance during treatment and concentration over the interval are different quantities. Kang et al., 2021.

A free-solute association supports investigation, not a resolved toxin. Lai studied 54 predialysis CKD patients, 27 with and 27 without diabetes, with mean eGFR near 31. Diabetic participants had worse conduction-amplitude scores despite similar mean free indoxyl sulfate: 0.24 versus 0.21 µg/mL. Total concentrations were 4.0 versus 3.9 µg/mL, so only about 5–6% of the group mean total was free.

Free indoxyl sulfate correlated with worse amplitude score, reported r=0.39. The selected regression retained free indoxyl sulfate and an oxidative-stress marker. But its coefficient/SE, 5.04/2.65, implies an ordinary two-sided P near 0.063, versus printed 0.035; a mediation direct-effect table also has an incompatible coefficient/SE/P combination. Cross-sectional mediation does not establish endothelial causation even if the statistics are repaired.

The study measures circulating free dose and a large-fiber functional association. It lacks local concentrations, random exposure manipulation and a physical terminal endpoint. Protein binding also prevents treating a total plasma dose as the freely available concentration used in a culture experiment. Lai et al., 2022.

Transplantation does not provide a clean renal-exposure rescue. In a T1D series, 18 successful islet-after-kidney recipients were followed against nine kidney-only recipients for up to six years. Conduction improved within the islet group, but cross-sectional group comparisons did not establish a significant difference. Skin analyses emphasized AGE/RAGE staining; they did not show a quantified randomized peptidergic terminal-regeneration contrast. Both groups already had a kidney transplant. The added intervention changes glycemia, insulin and other endocrine signals, rather than isolating renal clearance. Del Carro et al., 2007.

A 2026 study examined 160 kidney-transplant recipients, a median 6.1 years after transplantation. It identified 84 with large-fiber polyneuropathy and seven with small-fiber neuropathy; the latter was classified using symptoms, examination and quantitative sensory testing, not skin density. Higher urea remained associated with large-fiber neuropathy, OR 1.12 per reported unit, interval 1.04–1.23, in a selected multivariable model.

This cross-section cannot determine how many nerves were protected by transplantation or why abnormalities persisted. The separate 2025 diagnostic-score publication used the same SENS cohort and is not an independent replication. A null association with time since transplantation does not prove either transplant protection or irreversible injury. Nolte et al., 2026.

The existing within-diabetes metabolite null has a different scope. In the public 97-donor ADDITION-Denmark contrast, the seven organic metabolites in this panel did not distinguish established DPN from diabetes without DPN: family permutation P=0.907; indoxyl-sulfate geometric-mean ratio 0.902, interval 0.724–1.123. Participant renal covariates were unavailable.

This limits claims that a large unadjusted circulating difference is already demonstrated in that cohort. It does not test advanced CKD, free endoneurial dose, potassium or β2-microglobulin, and is not an equivalence test for Q's effect. Likewise, a SUDOSCAN-derived “renal” score partly computed from skin conductance is not independent renal evidence. Public-data checks (background note outside this collection); renal-score correction (background note outside this collection).

Conditional judgments

Each row after R1 conditions on all preceding requirements. The low/best/high columns are inputs to complete coherent sensitivity scenarios, not independent marginal distributions.

See the claims and probabilities table above.

R1 is supported by the known concentration/clearance physiology in advanced disease. R2 is plausible, particularly for potassium and vascular interfaces, but local free exposures and sustained Q engagement are incompletely measured. R3 is lower because the actual structural anchor is observational, while interventions mainly measure electrical or clinical function.

Given some retained structural effect, neither a large preserved fraction nor a small one is established. That supports 65% for R4 rather than treating every biochemical correction as material. R5 benefits from the deliberately exposed advanced-CKD scope, but referral prevalence and prevalent old neuropathy are not the required case denominator.

The headline is 0.95 × 0.85 × 0.65 × 0.65 × 0.70 = 0.238818125. Source plus local exposure jointly receive 80.75%; the remaining structural/magnitude/coverage requirements jointly receive 29.575% conditional on that. More rows do not themselves justify a smaller probability.

Sensitivity and possible updates

The lower interpretation, [80,45,25,30,25]%, gives 0.675%. It attributes much CKD-associated denervation to preexisting disease, nutrition, medication or vascular injury outside the panel; potassium changes electrical function while retained solutes have little durable structural contribution.

The upper interpretation, [99,95,90,90,90]%, gives 68.56245%. It treats human denervation and potassium intervention as views of a substantial causal exposure burden, with negative dialysis results explained by poor sustained panel separation and limited endpoints.

The range is about 67.9 percentage points wide. It is sensitivity to jointly plausible interpretations, not a confidence interval or a calibrated probability distribution. Trial sampling uncertainty, selective follow-up, exposure engagement, endpoint transfer and subgroup coverage are different uncertainties. The preferred scenario gives the randomized positive and negative findings weight while refusing to equate either with terminal-fate evidence.

A reversible electrical effect with little structural preservation, [98,95,30,35,80]%, gives 7.8204%. A large structural effect confined to rare severe retention, [98,90,85,85,20]%, gives 12.7449%. These alternatives are not captured by merely shifting every row together.

Hypothetical sustained panel correction with direct human peptidergic preservation and sufficient case coverage could produce [99,95,90,85,85]%, or 61.1560%. A precise structural null despite complete local engagement could give [95,90,20,25,40]%, or 1.71%.

Changing R3 alone to 90% gives 33.0671%; reducing it to 25% gives 9.1853%. Certainty about that one link permits at most 36.74125% with other judgments fixed. This is an arithmetic ceiling, not the expected value of more research.

Ideal experiments that would resolve each claim

Use the defined population, six groups, ten-solute intervention, matched reference law and complete physical-fate measurements. Independent people are the inference units. Observe replacements, later loss, systemic responses and vital status; selective survivor follow-up cannot substitute for complete outcomes. These ideal comparisons define the proposition rather than recommend a clinical protocol.

Each later test is restricted to groups satisfying the entire prior prefix. Exact population values decide YES or NO. Inadequate solute control, unidentified free exposure, incomplete follow-up or intervals crossing a boundary are unresolved.

R1 — renal source of abnormal exposure. Measure complete production, distribution, renal extraction/excretion and circulating free/total trajectories. Compare the actual renal handling function with the specified normal-renal function for each candidate solute while initial nonrenal production is identical. YES: at least one group meets the ≥10% people, ≥30-day abnormal-concentration and ≥50% renal-attribution conditions. NO: none does. Low eGFR alone establishes neither a particular panel excess nor its source fraction; human concentration studies approach only part of this comparison.

R2 — connected local delivery. Map the same solute from circulation to the specified neuron/support-cell/microvascular interface, then compare Q with reference care before the connected further deficit. YES: a group meets the local ≥50% excess (≥0.3 mmol/L for potassium), ≥50% correction, prevalence, unit coverage, duration and temporal-order conditions. NO: every R1-qualified group fails at least one. The four-person potassium clamp approaches selective dose/physiology; it does not establish the mapped C-terminal exposure history.

R3 — retained peptidergic effect. Compare five years of Q with reference care in R1–R2-qualified groups. Track stable-identity physical disappearance, replacement and later disappearance. YES: at least one group has L₀ >0 and B ≥0.05. NO: none meets both. Chao provides actual human skin pathology; the potassium and FINESSE trials provide assignment with incomplete exposure or endpoint matching. Neither is this comparison.

R4 — material magnitude. Use the same complete trajectories and allow downstream compensation and costs. YES: an R1–R3 group reaches B ≥0.20. NO: every such group remains below 0.20. For a reference loss of 30% of baseline terminals, Q must preserve at least six baseline percentage points. A 67% cross-sectional density gap, a two-point score change, or improved clearance cannot be substituted for this fraction.

R5 — renal-population coverage. In the complete reference-care advanced-CKD population, identify all new/progressive cases and their preintervention group membership. YES: an R1–R4 group contains ≥10% of those cases. NO: none does, or there are no reference cases. Include secondary renal neuropathy but distinguish it from isolated symptoms. The percentage with old neuropathy on dialysis or after transplantation does not answer this prospective coverage question.

Relation to other claims

The organic panel overlaps gut-derived exposure, but its source intervention and population differ. A gut-output correction in early T2D does not imply benefit from renal-panel correction in advanced CKD, or vice versa. Potassium may act through ion loading; local vascular or oxidative injury are alternatives, not additional compulsory rows.

Functional B12 deficiency is outside this panel; renal elevation of MMA cannot stand in for a cofactor deficit. The remaining scientific uncertainty is specific: whether sustained correction of the ten defined exposures retains human peptidergic terminals at material scale, beyond the clinical and electrical effects already observed.