Oral single-active pantothenic acid and HbA1c in adults with type 2 diabetes

Complete submitted research report

# Oral single-active pantothenic acid and HbA1c in adults with type 2 diabetes

**TASK-1034 · R01-074 · Evidence cutoff/self-review date: 2026-09-17** Efficacy grade **?**, score **null (unscored, not zero)** · Safety **Caution** · Document quality **A (separate from efficacy)** Result `completed_with_uncertainty` · Content readiness `ready_with_uncertainty` · New site identity unassigned. This is the complete English report and exactly matches `en.body_markdown` in the verdict JSON.

## 1. Thirty-second answer

Within the disclosed search, **no directly eligible result established how much oral single-active pantothenic acid changes HbA1c versus a comparator in adults with type 2 diabetes.** A B5 randomized trial with a registration and posted results exists, but its reported glucose measure is fasting blood sugar, not HbA1c. Dietary/plasma pantothenic-acid associations and HbA1c changes with B5-containing mixtures do not isolate the effect of pantothenic acid. [S01–S05]

This is not evidence of zero effect, equivalence to placebo, or absence of all human research. These materials do not establish an individual dose or expected HbA1c reduction. Research doses are not instructions to stop diabetes medication or replace standard care. The `? / null` judgment is limited to the unverified direct result for this question; it is not a probability of treatment failure or a safety certification.

## 2. Separating the research question from study facts

The question is **single-active pantothenic acid → oral administration → adults with type 2 diabetes → a between-group HbA1c difference at a defined time**. Actual placebo, usual care or active comparators, common medications and lifestyle management must be identified study by study. Absolute HbA1c levels in percent, changes in percentage points, adjusted between-group effects and within-group changes are distinct.

Calcium or sodium pantothenate also requires verification of salt mass, active pantothenic-acid amount and co-active ingredients. Pantothenic acid and pantethine are not merged as one intervention. Topical or injected panthenol/dexpanthenol, multivitamins, foods, dietary associations and animal/cellular/mechanistic evidence cannot establish the isolated effect in this question. Deficiency correction is distinguished from supplementation without confirmed deficiency.

The proposed population, formulation, duration and comparator are questions, not verified trial facts. **Actual dose, active amount, duration and comparator for an eligible HbA1c comparison remain unverified.** Separately, the adjacent IRCT *protocol* specifies B5 250 mg/day, eight weeks, placebo and metformin/insulin. That does not verify actual exposure, the complete single-active composition, or adherence. [S01]

Fasting/postprandial glucose, OGTT, insulin, HOMA, clamps, continuous glucose monitoring, complications and clinical events are separate endpoints. Their changes or statistical significance are not converted into an HbA1c effect.

## 3. Input identity, completion history and reuse

The input ZIP contains 1,954,171 bytes, 57 files and 10,657,680 uncompressed bytes, with SHA-256 `5b2efdce90d6898281acd25893f9fec75eded3e076c3b710052e88b58147322b`. All 56 member hashes declared by its manifest were checked; the manifest's own hash is in the external inventory. The separately attached prompt is byte-identical to the prompt inside the ZIP. The common/R01/NUT instructions, original rubric/calculator, codebook/categories, full 3,149-record index, eight candidate originals and exact prior five-task records were read and applied. All 57 original files are preserved byte-for-byte under ASCII paths in the completed archive, with original Korean paths mapped in provenance. [P01]

The supplied current record shows completion through range 73: 68 published, five exact duplicates and 27 pending. TASK-1029→3145, 1030→3146, 1031→3147, 1032→3148 and 1033→3149 were reconciled with the six-route deployment receipts. Reading those receipts was not a fresh deployment or independent live-site check. Null IDs in prior original manifests describe historical handoff states; those originals were preserved and current completion records took precedence. [P01]

| Reused ID | Reused scope and reason this is not an exact duplicate | |---|---| | 028 | Broad legacy skin/hair material and route/claim boundaries, not direct HbA1c efficacy | | 3099 | Acne and single/combined-product mapping; different endpoint | | 3100 | LDL and pantethine distinction; different endpoint/molecular identity | | 3101 | Fatigue/FSS and nutrition/mixture boundaries; different endpoint | | 3102 | Oral exposure/diarrhea safety extraction, not regraded here | | 3147 | Chronic-ulcer epithelialization and prior safety sources; different endpoint | | 3148 | Stratum-corneum hydration and oral/topical distinction; different endpoint | | 3149 | TEWL and prior search/safety/limitations; different endpoint |

Ten related entries were located in the full index. Additional entries 366 (pantethine lipids) and 2466 (panthenol-containing eyelid wipe) were assessed at index-boundary level only; their unsupplied full originals were not claimed as read. No exact completed claim was identified in the provided scope, so the operation is `new`. Candidate presence or absence was not treated as a novelty guarantee. Completed B6 and skin judgments were not researched, graded or translated again. [P01]

## 4. Searches actually performed and access limits

Searches were performed on 2026-09-17 using `pantothenic acid`, `pantothenate`, `calcium pantothenate`, `sodium pantothenate`, `vitamin B5` and Korean pantothenic-acid terms with `type 2 diabetes`, `HbA1c`, `glycated/glycosylated hemoglobin`, trial/randomized/placebo, and termination/retraction/correction terms. Public web search led to official PubMed indexing, registrations and publisher sources; NIH ODS and NGSP were checked. Exact queries and search modes are retained in `search_log.json`.

Some searches returned imprecise or irrelevant results. Native PubMed search URLs and the Europe PMC API failed. Authenticated exhaustive searches of Embase, Cochrane and Web of Science, the entire WHO registry network, author contact and individual-data requests were not performed. Unknown database-wide result totals were not fabricated as PRISMA counts. Five core clinical/observational cohorts were classified in the evidence table and zero directly eligible effect datasets were verified; these are **counts in this review's evidence table**, not counts of all research worldwide.

IRCT registration/results PDF and a revision HTML page, Alarifi's full text/tables and Lee's full text/figures were inspected. Tak's full publisher HTML was read. Shoura's parsed PDF full text was accessible, but table-image verification failed. For Donati's pantethine paper, official bibliographic metadata were located but the abstract/full report was unavailable. No matched correction/retraction notice was verified in the retrieved materials and searches; this is not exhaustive historical clearance. No publisher PDF is claimed to have been archived locally as exact original bytes. [S01–S06]

## 5. Expert evidence table

These are human studies, but all are **ineligible for estimating the isolated oral single-active pantothenic-acid HbA1c effect** specified here. Their reported facts are retained with different exclusion reasons. A registration, PDF and HTML for the same study were not counted as independent cohorts.

| Study | Actual design, denominators and intervention/comparator | Endpoint, result and interpretation | |---|---|---| | S01 IRCT20230702058641N2 | T2D, registered parallel randomized double-blind trial. Target48; results table19+19. Protocol B5 chewable250 mg/day for8weeks versus shape/taste/color-matched placebo, both with metformin/insulin | Registered primary outcomes are depression/anxiety/stress. FBS changes: control−0.89±7.33, intervention−5.84±21.47; between-change P=.352, units omitted. No HbA1c registration/result verified. Complete single-active composition also unverified | | S02 Shoura2025 | Cross-sectional plasma biomarkers. Methods state four groups of35: healthy, obese non-diabetic, T2D, T2D+CVD. No administered supplement | Plasma PA–HbA1c correlation r=−0.318, P=.0003. Not an administration effect or treatment duration. Table-image/denominator/unit concerns unresolved | | S03 Alarifi2025 | 196 adults with T2D; HbA1c<8% n66, >=8% n130. Dietary recalls, no assigned intervention | Dietary B5 5.59±2.99 versus4.71±2.76 mg/day, P=.05. Age/sex-adjusted elevated-HbA1c OR2.16, 95%CI1.04–4.47, P=.03. Direction/model caution; not a supplement HbA1c mean difference | | S04 Tak2023 | Single-arm mixed nutritional drink replacing breakfast for8weeks.33 recruited,26 analyzed; no concurrent comparator | HbA1c7.23±0.82%→7.03±0.69%, P=.041. Arithmetic within-person change−0.20 percentage points; not an isolated B5 or between-group effect | | S05 Lee2016 | Mixed fermented product versus placebo in crossover trial.36 randomized,31 completed; two12week periods and12week washout. B5 label1.8 mg/day plus multiple other ingredients | HbA1c changes+0.09±0.4 versus+0.01±0.6 percentage points, P=.48. Same31 participants cross over; not a single-B5 trial. Non-significance is not B5 ineffectiveness or equivalence |

The ± figures are source-reported SDs, not SEs or CIs. S03's CI describes an observational odds ratio, not an HbA1c mean difference. The simple−4.95 change difference for S01 and+0.08 percentage-point difference for S05 are retained only in the calculation audit, not presented as verified adjusted HbA1c effects. [S01–S05]

### 5.1 What the closest B5 registered trial does and does not establish

IRCT was registered while recruiting on2024-02-01 and posted results on2024-09-21. Recruitment-complete and results-posted status were verified. Actual recruitment/completion date fields are empty; no separate journal DOI/full report was verified. A parsed/rendered generated-PDF date discrepancy was not interpreted as a substantive update. Subtracting the38 table participants from a target48 does **not establish ten withdrawals**. The investigator list under a mislabeled participant-flow heading is not a patient flow diagram. [S01]

The protocol includes T2D patients aged30–65. Actual diagnostic criteria, disease duration, baseline HbA1c, renal function, diet and complete co-morbidity distribution are insufficiently reported. Baseline treatment counts are control metformin11/insulin8 versus intervention metformin15/insulin4, but drug doses, treatment changes and lifestyle adherence remain unverified. Trial exclusion rules are not transferred into medication-stopping advice for readers. [S01]

B5 nutritional-status values conflict between the baseline table,97.5±30.1/91.4±56.0, and the results-table baseline,5.97±1.30/4.91±0.56, with no verified units, method or deficiency threshold. The control group's nutritional-status change sign also disagrees with endpoint-minus-baseline arithmetic. These were not repaired into a deficiency-correction trial or reliable total-intake measure. Shehab is the named manufacturer, but salt, active amount, complete composition and supply agreement remain unverified. [S01]

Blocks of four and participant/investigator blinding are described, but actual concealment implementation, ITT, missing-data handling and adherence are not verified. Summary phase2 versus structured phase3 is another source conflict. FBS does not appear in the registered primary/secondary hierarchy; its prespecification was not established. A blank adverse-events section does not mean zero events or proven hypoglycemia safety. The registry describes100% public academic support from Shahre-kord University of Medical Sciences, not verified absence of every conflict. [S01]

### 5.2 Observational HbA1c associations are not deficiency-treatment effects

Shoura's diabetes definition includes fasting glucose>=126 mg/dL, HbA1c>=48 mmol/mol or diabetes medication. eGFR<15 was excluded, which does not establish control of every renal-function state. Plasma PA ELISA was not confused with the HbA1c assay. Parsed-table denominator and NGSP/IFCC concerns were quarantined as unresolved without image verification; no confirmed author error or corrected clinical baseline is published here. STDF46040 funding, STDF/EKB open-access support and the authors' no-competing-interest declaration were recorded. [S02]

Alarifi estimated diet using two nonconsecutive24hour recalls. Mean HbA1c was9.18±2.1%; the study's8% strata are not universal individual treatment targets. The dietary table shows lower intake with higher HbA1c, whereas the regression odds-ratio direction differs. Information was insufficient to resolve linear/logistic wording, variable coding, multiplicity and residual confounding by medications, energy intake and other factors. Neither favorable direction was selected as causal evidence. The no-external-funding statement and acknowledgement of Shaqra University research support are both retained. [S03]

NIH ODS distinguishes plasma pantothenic acid from better intake/status indicators. Consequently, a low plasma value/tertile or dietary estimate does not independently establish clinical deficiency or show that supplementation will improve HbA1c. [S08]

### 5.3 Misinterpretations prevented in mixtures and crossover designs

Tak's intervention replaced breakfast with two daily200 mL/200 kcal packs. The400 mL/400 kcal is **product exposure**, not a pantothenic-acid dose. Nutrients, energy and meal substitution change together, with26 per-protocol participants. Three screening exclusions, three voluntary withdrawals and one protocol violation were separated. Daesang company funding and the authors' no-conflict declaration were recorded separately. A narrative of no serious adverse events with data not shown is not absence of all adverse events or proof of single-B5 safety. [S04]

Lee screened82, randomized36 in19/17 sequences, and analyzed31 in16/15 sequences. The same31 participants experience both conditions; they are not62 independent people. The sequence calendar spans36weeks, but active-product exposure per completer is12weeks, distinct from24weeks across both dosing periods and12weeks of washout. Five withdrawal reasons were elevated triglycerides1, insulin initiation2, cerebral ischemia1 and taste1; these are not assigned as adverse effects of isolated B5. [S05]

The product contains fermented vegetables/fruits/nuts, chromium/zinc and several vitamins. B5 is labeled1.8 mg/day; the salt/active-acid basis is unspecified. PDF text lost the micro symbol; original page images established chromium100 μg and B12 0.75 μg rather than milligrams. No unverified elemental/salt equivalence was imposed for metal ingredients. [S05]

A Hills–Armitage method addressing period effects is described, but paired covariance, a specific adjusted-effect CI and residual carryover results were unavailable. The two condition SDs and sample counts were not treated as independent parallel arms to fabricate a CI. Empty-bottle counts/weights monitored adherence, without a verified quantitative adherence percentage. The manufacturer produced/supplied both active and placebo and funded the trial and first author. Gastrointestinal/chest symptoms can overlap across conditions and were not summed into unique-person rates or isolated-B5 risks. [S05]

## 6. HbA1c measurement and numerical audit

| Source | Verified assay/specimen/time information | Remaining limits | |---|---|---| | IRCT | No HbA1c item in posted results | This does not establish that HbA1c was never measured. Assay, baseline/end values, units and calibration are unreported | | Shoura | Eight-hour fasting heparin plasma,6000rpm for15min and−80°C describe biomarker handling | Not transferred to HbA1c handling/assay. PA ELISA is separate; table images and unit correspondence unverified | | Alarifi | HbA1c immunoturbidimetry, Roche Cobas6000;10hour fasting EDTA,8–10 inversions, analysis within1hour | Exact reagent, NGSP/IFCC traceable calibration and variant-specific adjustments unreported | | Tak | HbA1c HPLC HLC-723G8, fasting EDTA, baseline/eight weeks | Instrument name alone does not establish mode, standardization or absence of variant interference | | Lee | Baseline/four/twelve weeks in each period,12hour fast; HbA1c in percent | HbA1c instrument/method unreported. HemoCue is glucose and ELISA is insulin, not HbA1c |

The official equation is `NGSP % = 0.09148 × IFCC mmol/mol + 2.152`. Its inverse is `(NGSP % − 2.152)/0.09148`; the intercept is not added to changes, SDs or SEs. One percentage-point difference corresponds to approximately10.931 mmol/mol under this equation. This is a unit audit, not generation of unobserved trial results. A relative percent change is a ratio to baseline, not an absolute percentage-point change. [S10]

Shortened red-cell survival, anemia/iron deficiency, transfusion, hemoglobin variants, renal disease and erythropoietin can affect HbA1c measurement or interpretation; some effects depend on the assay. Whether these were measured or corrected must be checked in each study. This general caution **does not establish pantothenic-acid-specific HbA1c assay interference**, which was not verified here. [S09]

An eligible target between-group HbA1c effect,95%CI and MCID remain unverified. Clinical-event absolute risk, ARR or NNT were not generated from HbA1c surrogate data. Non-significant P values were not used as equivalence evidence, and no clinical-importance threshold was invented.

## 7. Safety and total exposure

**The safety label is Caution, separate from its explanation.** The Rao healthy-adult oral-exposure extraction was reused from3102 and3147–3149. Forty unique participants comprise32 single-dose and eight repeated-dose participants. At the paper's labeled5,000 mg fasting single dose, diarrhea occurred in1/8; the same eight later had0/8 under fed conditions. This was not a randomized placebo comparison, so no protective food effect or causal rate was calculated. Zero of eight during2,000 mg/day for14days does not establish long-term safety. Twenty-two days of follow-up were not rewritten as22days of dosing; salt/active amount and total dietary exposure were not guessed. Funding and company-related employment/consultancy/shareholding disclosures were retained. [S07; P01]

Current NIH ODS information notes that a tolerable upper intake level has not been established and that large doses can cause diarrhea/gastrointestinal discomfort. Its statement about no known clinically relevant medication interactions is not verification of every diabetes-drug combination, renal-function state or prolonged high-dose exposure. A sensitization-related oral-exposure eczema case was reused only at prior bibliographic-map level, without assigning a rate, mechanism or risk for this population. [S08; S11]

Unreported IRCT adverse events, mixture-associated symptoms/serious-event statements and short-term healthy-adult data were not pooled into a single-B5 risk estimate. Pregnancy/lactation, children, hepatic/renal disease, long-term total intake, co-medication and hypoglycemia risk remain uncertain in this scope. This is not personal dosing or treatment-change advice.

## 8. Classification, efficacy grade, score and document quality

The existing kind is **S** and category **blood-sugar**. R01/NUT and GLY are work/module/internal question identifiers, not new site categories or official clinical codes. New site ID, slug, URL, first-publication timestamp and semantic codes are all null and were not issued. [P01]

The original stage-0 rubric and unchanged `등급도출.py` were actually executed. Inputs use `claim_type=B` and surrogate endpoint `S`. The legacy `no_human_study=true` gate is **limited to this oral single-active/T2D/between-group HbA1c question within the verified search**. Related human studies are separately marked as existing; universal absence is false. The flag was not automatically inferred from access failure: actual compound, population, comparator and endpoint exclusions were assessed. Replication, independence, effect, bias and precision axes remain unscored null. Proposed and final grade are both `?`, with no validation errors. No numeric anchor was invented for `?`, so **score=null**, not zero. [P01; audits/grading_calculation.json]

This efficacy judgment describes the currently verified direct-evidence level; it is not official GRADE, treatment probability or a clinical success rate. **Document quality A** is the author's separate assessment that the requested traceability, boundaries, missingness explanations, bilingual content and format are complete. It does not mean strong efficacy evidence, independent external review, journal certification or an error-free guarantee.

## 9. Unresolved matters and revision triggers

Result status is `completed_with_uncertainty`, content readiness is `ready_with_uncertainty`, and content verification is `completed_with_declared_scope`. Weak direct evidence or inaccessible full text does not leave this manuscript pending. Conversely, unknown numbers are not filled in or expanded into a universal no-research claim.

Revision triggers include a T2D comparison with verified single-active composition/salt/active amount and HbA1c results; IRCT HbA1c data, registry history, complete composition or valid patient flow; inaccessible full text/table images; correction/retraction; or new information changing numerical or classification boundaries. Any subsequently assigned site ID/URL should be retained when updating, with before/after values, evidence and dates logged. Initial public `corrections=[]` is distinct from this submission's prepublication audit.

The same author checked sources/registration, numbers, units, comparators, denominators, safety, original calculator, classification, bilingual text and fields. Disk-wrapper rereading and ZIP restoration/repacking are **separate technical validation**, not proof of clinical truth. No clinical re-search, regrading or translation-review TODO is handed to Codex. No server access, deployment, ledger update or automatic start of task75 was performed.

## Sources and input traceability

The following addresses identify sources and locations. This is not a reproduction of journal full texts; factual extraction and author interpretation are distinguished. Detailed access limits, funding, product supply and study IDs are in `sources.json`;43 study-level extraction fields are in `study_extraction.json`.

- **S01** IRCT20230702058641N2. Registration/results; last substantive update2024-09-21. https://en.irct.ir/trial/74085/pdf - **S02** Shoura et al. Scientific Reports2025;15:32549. DOI10.1038/s41598-025-19271-5. https://www.nature.com/articles/s41598-025-19271-5.pdf - **S03** Alarifi et al. Italian Journal of Food Science2025;37(4):269–278. DOI10.15586/ijfs.v37i4.3145. https://www.itjfs.com/index.php/ijfs/en/article/view/3145/1755 - **S04** Tak et al. Nutrition Research and Practice2023;17(2):241 onward. DOI10.4162/nrp.2023.17.2.241. https://www.e-nrp.org/DOIx.php?id=10.4162/nrp.2023.17.2.241 - **S05** Lee et al. Food & Nutrition Research2016;60:30298. DOI10.3402/fnr.v60.30298. https://foodandnutritionresearch.net/index.php/fnr/article/download/1006/3895/ - **S06** Donati et al. Clin Ter1989;128(6):411–422, PMID2524328. Pantethine bibliographic lead only. https://pubmed.ncbi.nlm.nih.gov/2524328/ - **S07** Rao et al.2021. Exact-input healthy-adult oral calcium-pantothenate PK/safety extraction reused. https://www.gavinpublishers.com/article/view/the-pharmacokinetics-of-orally-administered-calcium-pantothenate-in-healthy-adults - **S08** NIH ODS. Pantothenic Acid, updated2026-05-01. https://ods.od.nih.gov/factsheets/PantothenicAcid-HealthProfessional/ - **S09** NGSP. Factors that Interfere with HbA1c Test Results, updated2026-06-23. https://ngsp.org/factors.asp - **S10** NGSP. IFCC Standardization. https://ngsp.org/ifccngsp.asp - **S11** Hemmer et al.1997, PMID9255501. Reused sensitization/oral-exposure bibliographic record only. https://pubmed.ncbi.nlm.nih.gov/9255501/ - **P01** Exact current input ZIP, prompt, original calculator/rubric, codebook, full index and completion records. Byte-exact copies and original path/SHA mappings are in `provenance/input_inventory.json` and `provenance/input_path_map.json`; prior originals were not modified.