# Riboflavin (vitamin B2) × 161.3-km race finishing time in experienced adult ultramarathon runners

Verdict 3124 · Evidence Check

Stored-material review date: 2026-10-04  
Original evidence date: 2026-09-16  
Review-criteria version: v3124-B2-ULTRA-FINISH-1.0  
Public page/material preparation date: 2026-11-07

Clinical papers were not read directly; existing review summaries were checked. No new clinical-literature search was performed. Unverified does not mean that research is absent or that there is no effect.

## Assessment scope

- Population: Experienced adult ultramarathon runners participating in the 2016 WSER; finishing-time data restricted to study completers
- Ingredient, formulation and route: Riboflavin (vitamin B2), ordinary form as reported in the original · Reported as an investigator-prepared capsule with riboflavin as the single active nutrient; batch and excipient documentation unverified · Oral
- Dose: 100 mg before the start and 100 mg at 90 km; planned total of 200 mg in one race
- Comparator: Color-matched placebo containing 95 mg maltodextrin and 5 mg of 10% beta-carotene powder per capsule; race diet and usual care not standardized
- Measured outcome: 161.3 km finishing time (hours) at race completion; Table 1 study completers
- Duration: One 161.3 km ultramarathon race; not daily dosing over several weeks
- Questions outside this scope: Pain, 400 m recovery, CK and laboratory time to exhaustion are preserved but do not replace this table's time outcome.

## Checklist items

### N01 · Which population, intervention, distance and time comparison does this table address?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The original claim concerns 161.3 km race finishing time in experienced adult ultramarathon runners. The table presents a one-race comparison of oral single-ingredient B2 versus a color-matched placebo. subclaims=[]; pain, 400 m recovery, CK, vitamin status and laboratory time trials are separate outcomes.
- Why unverified: The identity and inclusion/exclusion boundaries of the independent claim have been linked to the current ID and exact original wording. This condition being met does not establish a reduction in finishing time.
- Reason for missing information: There is no gap in identity linkage, but causal effect and precision remain unverified in separate rows.
- Required materials and conditions: Linkage of the current claim, population, formulation/route, comparator and endpoint to the excluded endpoints on the same page
- Materials to obtain first: Linkage of the current claim, population, formulation/route, comparator and endpoint to the excluded endpoints on the same page
- Further research: In subsequent research, prespecify the finishing-time question for the same distance separately from the recovery question.
- Conditions for verification: Link the original /ko/effect precisely to the single-race finishing-time question and explain whether parent and subordinate claims overlap.
- Why this condition is needed: Prevents favorable recovery or nutritional-status results from being transferred to efficacy for finishing time.
- Application exceptions: Pain, 400 m recovery, CK and laboratory time to exhaustion are preserved but do not replace this table's time outcome.
- Remaining uncertainty: There is no gap in identity linkage, but causal effect and precision remain unverified in separate rows.
- Referenced methods: Question-clarification principles in M03 §3.1, §3.2.1–3.2.4; detailed separation rules are self-designed; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed
- Related sources: S01

### N02 · Are the study's single-ingredient riboflavin and exact product specifications verified?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The study reports oral capsules containing riboflavin as the single active nutrient. Documentation of the exact salt, product batch and complete excipients has not been verified. Results for B-vitamin combinations or other formulations are not treated as the same intervention.
- Why unverified: The molecule, single active nutrient and route are preserved, but the full product specifications have not been verified.
- Reason for missing information: The salt, batch, complete excipients and documentation of measured active amount/quality are unverified.
- Required materials and conditions: Documentation of the trial capsule's molecule/salt, batch, active amount, complete excipients and quality
- Materials to obtain first: Documentation of the trial capsule's molecule/salt, batch, active amount, complete excipients and quality
- Further research: Disclose batch specifications and analytical data for the trial product and directly compare interventions with the same specifications.
- Conditions for verification: Record the reported molecule, single-ingredient status and route separately from unverified specifications.
- Why this condition is needed: Needed to prevent results for B-vitamin combinations or other formulations from being linked as the same intervention.
- Application exceptions: Do not extend the report of riboflavin as a single active nutrient to verified documentation of the exact salt, manufacturing batch or complete excipients.
- Remaining uncertainty: The salt, batch, complete excipients and documentation of measured active amount/quality are unverified.
- Referenced methods: M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed
- Related sources: S01

### N03 · To what extent do the actual runners match the population in the original claim?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The actual data concern experienced adult ultramarathon participants in the 2016 Western States Endurance Run. Finishing time is restricted to 32 study completers with complete data. Exact age eligibility limits and renal/hepatic-disease and pregnancy subgroups are unverified.
- Why unverified: The race setting, experienced adults and actual time-analysis population have been linked, but detailed selection criteria and special-population coverage have not been verified.
- Reason for missing information: Age and inclusion/exclusion details, baseline racing ability, and renal/hepatic-disease and pregnancy subgroups are unverified.
- Required materials and conditions: 2016 WSER inclusion/exclusion criteria, age range, experience, baseline racing ability and special-population data
- Materials to obtain first: 2016 WSER inclusion/exclusion criteria, age range, experience, baseline racing ability and special-population data
- Further research: Subsequent trials should specify racing experience and selection criteria and predefine population-specific applicability limits.
- Conditions for verification: Identify the race-participation setting, selected time-analysis population and unverified detailed population criteria.
- Why this condition is needed: Needed to prevent experienced race participants from being generalized to all athletes or adults in general.
- Application exceptions: If exact age/disease criteria are unverified, do not invent criteria beyond the original adult-population scope.
- Remaining uncertainty: Age and inclusion/exclusion details, baseline racing ability, and renal/hepatic-disease and pregnancy subgroups are unverified.
- Referenced methods: M03 §3.2.1; detailed population criteria are self-designed
- Related sources: S01

### N04 · Is it verified whether this is correction of deficiency or additional supplementation on top of adequate intake?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Although presented as a supplementation question, the population has not been verified as deficient or adequate through measurements of baseline B2 markers, deficiency thresholds and total dietary intake. Unverified status is not used as confirmation of nondeficiency.
- Why unverified: The necessary baseline measurement/intake records could not be verified within the currently retained scope.
- Reason for missing information: Baseline markers, deficiency criteria, total dietary B2, concurrent supplements and group-specific data are unverified.
- Required materials and conditions: Baseline B2 markers/deficiency thresholds, total dietary B2, concurrent supplements and baseline data by group
- Materials to obtain first: Baseline B2 markers/deficiency thresholds, total dietary B2, concurrent supplements and baseline data by group
- Further research: Measure baseline B2, diet and concurrent supplementation, and plan prespecified effect-modification analyses by deficiency status.
- Conditions for verification: Distinguish actually measured baseline status from unverified nutritional/dietary dimensions.
- Why this condition is needed: Needed to prevent unreported baseline status from being converted into confirmed absence of deficiency or direct evidence for deficiency treatment.
- Application exceptions: Record the supplementation intent, but do not establish deficiency or adequacy in a population whose testing has not been verified.
- Remaining uncertainty: Baseline markers, deficiency criteria, total dietary B2, concurrent supplements and group-specific data are unverified.
- Referenced methods: M03 §3.2.1; detailed population criteria are self-designed; M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.1–3.2.4; details of dose, duration and effect modification are self-designed
- Related sources: S01

### N05 · Are the two administrations in one race distinguished from long-term daily dosing?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The plan is to administer oral B2 100 mg 0.5–1 hour before the start and 100 mg at 90 km. The planned total in one 161.3 km race is 200 mg. This is not a trial of long-term daily use of 200 mg or 400 mg.
- Why unverified: The planned route, amount and timing in the actual manuscript have been linked to correction QA02. This is distinct from the N06 question of whether everyone actually absorbed the planned amount.
- Reason for missing information: There is no gap in the planned specifications. Actual administration/absorption and the optimal dose remain separately unverified.
- Required materials and conditions: Administration specifications, oral route, timing before the start/at 90 km, and planned total amount
- Materials to obtain first: Administration specifications, oral route, timing before the start/at 90 km, and planned total amount
- Further research: Fix the timing within the same race, and design studies directly comparing administration timing/doses as separate questions.
- Conditions for verification: Link the plan for 100 mg 0.5–1 hour before the start and 100 mg at 90 km to exposure in one race.
- Why this condition is needed: Needed to prevent the exploratory search term of 400 mg or a daily dose from replacing the actual intervention.
- Application exceptions: The planned total and each person's actual intake are separate; do not label this a long-term 200 mg/day trial.
- Remaining uncertainty: There is no gap in the planned specifications. Actual administration/absorption and the optimal dose remain separately unverified.
- Referenced methods: M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.1–3.2.4; details of dose, duration and effect modification are self-designed
- Related sources: S01

### N06 · Did everyone actually receive the planned second dose?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Omission of the second dose and vomiting were reported, and analgesic use also occurred. The current records cannot verify that every runner fully received or absorbed the two planned capsules.
- Why unverified: The occurrence of deviations is preserved, but individual actual administered amounts and the scope of their analytical handling could not be verified.
- Reason for missing information: The number of participants with omitted doses and their timing, individual actual amounts, absorbed amounts, and prespecified analyses of adherence/assignment effects are unverified.
- Required materials and conditions: Individual administration/omission, vomiting, time available for absorption, adherence and analytical-handling records
- Materials to obtain first: Individual administration/omission, vomiting, time available for absorption, adherence and analytical-handling records
- Further research: Record administration/deviations for each individual and analyze prespecified effects of assignment separately from effects of adherence.
- Conditions for verification: Link verified omissions/vomiting to gaps in actual administered amounts and limits on assessing their analytical impact.
- Why this condition is needed: Treating the planned amount as actual exposure in everyone changes the comparison for the selected intervention.
- Application exceptions: Do not estimate the amount absorbed after vomiting or the number of participants with omitted doses if the manuscript does not provide them.
- Remaining uncertainty: The number of participants with omitted doses and their timing, individual actual amounts, absorbed amounts, and prespecified analyses of adherence/assignment effects are unverified.
- Referenced methods: M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.3, M08 §8.2.2–8.2.3; self-developed application; M08 §8.2.2, M15 §15.3.1; self-developed application
- Related sources: S01

### N07 · Are powder mass and active amount distinguished for the color-matched placebo?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The placebo contains 95 mg maltodextrin and 5 mg of 10% beta-carotene powder per capsule. The retained nominal calculation is 0.5 mg/capsule of active beta-carotene, not an active amount of 5 mg. Race diet and usual care are not standardized.
- Why unverified: The reported ingredients and historical nominal calculation were verified, but measured batch content and physiological equivalence/inertness comparisons were not.
- Reason for missing information: The placebo's actual batch content, physiological effects/nonspecific differences, and balance of care are unverified.
- Required materials and conditions: Placebo batch/ingredients, maltodextrin and beta-carotene powder specifications, and assessment of physiological effects
- Materials to obtain first: Placebo batch/ingredients, maltodextrin and beta-carotene powder specifications, and assessment of physiological effects
- Further research: Disclose the control capsule's ingredients/color and measured content, and assess nonspecific differences from the intervention.
- Conditions for verification: Attribute the reported 95 mg/5 mg composition and historical nominal 0.5 mg separately, without assuming physiological equivalence.
- Why this condition is needed: The actual comparison can be explained only after identifying the placebo and its nominal powder mass.
- Application exceptions: Do not read 5 mg of 10% powder as 5 mg of active beta-carotene, or convert a nominal calculation into measured content.
- Remaining uncertainty: The placebo's actual batch content, physiological effects/nonspecific differences, and balance of care are unverified.
- Referenced methods: M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.3, M08 §8.2.2–8.2.3; self-developed application
- Related sources: S01

### N08 · Are the diet, usual-care and analgesic differences actually internally consistent?

Verification status: Conflicting information  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Race diet and usual care were not standardized. For medication use during the race in the placebo group, the text records 6 participants, Table 1 records 36% of 14 participants, and the sum across medication types is 5 participants. QA04 leaves these unreconciled and preserves reconciled count=null.
- Why unverified: The original text, table and total referring to the same medication-user count could not be verified as consistent. This is an internal conflict in the source data, not a difference between analytical estimates.
- Reason for missing information: A consistent medication-user count, whether the same people overlap, balance of common care/diet, and analytical handling are unverified.
- Required materials and conditions: Diet/race care in both groups, medication types/participant counts, original table/text and corrections, and analytical handling
- Materials to obtain first: Diet/race care in both groups, medication types/participant counts, original table/text and corrections, and analytical handling
- Further research: Cross-check individual concomitant-use data against the original table, and preplan recording/handling of common care and rescue treatment.
- Conditions for verification: Preserve the verified lack of standardized care and the conflicting placebo medication-user counts, and leave the unresolved value as null.
- Why this condition is needed: Prevents concomitant use and internal data conflicts from being attributed to B2's contribution.
- Application exceptions: If the text, table and sum across medication types differ, do not select one or calculate a correction.
- Remaining uncertainty: A consistent medication-user count, whether the same people overlap, balance of common care/diet, and analytical handling are unverified.
- Referenced methods: M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed; M03 §3.2.3, M08 §8.2.2–8.2.3; self-developed application; M08 §8.2.2, M15 §15.3.1; self-developed application
- Related sources: S01

### N09 · Has the randomized title been distinguished from actual alternating allocation?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Actual allocation alternated by order of arrival at registration; it was not a random sequence. This differs from the randomized wording in the S04 conference title; QA03 records randomized=false according to the detailed Methods of the original journal article.
- Why unverified: The actual design identity was attributed to the retained detailed methods/corrections. Meeting the allocation-identity condition does not mean confounding was controlled or that this was a confirmatory RCT.
- Reason for missing information: The identity of alternating allocation has been linked, but limitations in concealment and baseline balance/adjustment remain.
- Required materials and conditions: Allocation by registration-arrival order, sequence generation, baseline balance and related report identities
- Materials to obtain first: Allocation by registration-arrival order, sequence generation, baseline balance and related report identities
- Further research: Subsequent comparisons should disclose the random sequence and its actual implementation, and assess balance in baseline racing ability.
- Conditions for verification: Label alternating allocation and its predictability according to the current detailed allocation record.
- Why this condition is needed: The word randomized in the title cannot replace the nonrandom procedure in the detailed Methods.
- Application exceptions: Do not remove nonrandomized data or treat verification of design identity as satisfaction of causality requirements.
- Remaining uncertainty: The identity of alternating allocation has been linked, but limitations in concealment and baseline balance/adjustment remain.
- Referenced methods: M08 §8.1–8.2.3; self-developed application; M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed
- Related sources: S01, S04

### N10 · Was it impossible to predict or change the allocation sequence?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Alternating allocation by registration-arrival order is predictable. Whether allocation concealment was effectively implemented has not been verified in the current records.
- Why unverified: The allocation procedure is known, but the information needed to assess concealment implementation/effectiveness is unverified.
- Reason for missing information: Registration personnel, sequence storage/exposure, the possibility of allocation changes and concealment-audit information are unverified.
- Required materials and conditions: Allocation personnel/concealment implementation, registration process, any breaches and baseline characteristics
- Materials to obtain first: Allocation personnel/concealment implementation, registration process, any breaches and baseline characteristics
- Further research: Separate responsibility for sequence generation, storage and implementation from registration procedures, including an audit of the allocation process.
- Conditions for verification: Record the allocation actually verified separately from the unverified effectiveness of concealment.
- Why this condition is needed: Needed to prevent the predictability of alternating allocation from being automatically offset by assumed successful concealment.
- Application exceptions: Unverified concealment information does not establish the magnitude of concealment failure or the occurrence of manipulation.
- Remaining uncertainty: Registration personnel, sequence storage/exposure, the possibility of allocation changes and concealment-audit information are unverified.
- Referenced methods: M08 §8.1–8.2.3; self-developed application
- Related sources: S01

### N11 · Are participant/onsite-researcher blinding and noticing urine color distinguished?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Blinding of participants and onsite researchers was reported. Urine color was noticed by 3 participants in the B2 group and 1 in the placebo group, so fully maintained blinding cannot be verified. The actual time-recording assessors and checks of maintained blinding are also unverified.
- Why unverified: The blinding report and cases of awareness were verified, but sufficient information on actual maintenance and the assessment stage is lacking.
- Reason for missing information: The timing of awareness, allocation guesses, and the relationship of assessors to time recording/blinding are unverified.
- Required materials and conditions: Blinding implementation, timing of noticing urine color, allocation guesses, assessors and checks of maintained blinding
- Materials to obtain first: Blinding implementation, timing of noticing urine color, allocation guesses, assessors and checks of maintained blinding
- Further research: Collect allocation guesses and the timing of awareness, and specify blinding of those recording race times.
- Conditions for verification: Preserve reported blinding together with unverified maintenance and cases of awareness.
- Why this condition is needed: Needed to prevent the mere presence of a color-matched placebo from being used to claim perfect blinding.
- Application exceptions: Record whether blinding was maintained even for an objective time outcome, but do not estimate the magnitude of time-related bias solely from urine-color awareness.
- Remaining uncertainty: The timing of awareness, allocation guesses, and the relationship of assessors to time recording/blinding are unverified.
- Referenced methods: M08 Table 8.2.a, M03 §3.2.4; self-developed application; M03 §3.2.3, M08 §8.2.2–8.2.3; self-developed application
- Related sources: S01

### N12 · What denominators do 44 participants, 37 participants and 32 participants each represent?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Allocation/starting counts were 22 participants in the B2 group and 22 participants in the placebo group; finishing counts were 20 participants and 17 participants; and complete-data counts in Table 1 were 18 participants and 14 participants. The actual denominator for the finishing-time means is 32 participants, not 44 participants or 37 participants.
- Why unverified: The group-specific stages and denominators for the Table 1 means are clearly linked. They have not been renamed randomized n.
- Reason for missing information: There is no gap in the identity of the flow counts. Reasons for losses and analysis of the overall effect remain addressed in N13.
- Required materials and conditions: Group-specific allocation/starting, finishing and data-completeness flow, and the identity of the Table 1 analysis
- Materials to obtain first: Group-specific allocation/starting, finishing and data-completeness flow, and the identity of the Table 1 analysis
- Further research: Disclose the full participant flow and the actual analysis denominator for each outcome using consistent identifiers.
- Conditions for verification: Link the stages of 22/22 starters, 20/17 finishers and 18/14 with complete data.
- Why this condition is needed: Needed to prevent starter or finisher counts from being substituted as the denominators for the table's means.
- Application exceptions: Do not rename alternately allocated participant counts as randomized n or label the results as means for 37 participants.
- Remaining uncertainty: There is no gap in the identity of the flow counts. Reasons for losses and analysis of the overall effect remain addressed in N13.
- Referenced methods: M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptations are self-designed; M08 §8.2.2, M15 §15.3.1; self-developed application
- Related sources: S01

### N13 · How were nonfinishing and incomplete data after finishing handled?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The retained records distinguish 7 nonfinishers among 44 starters from 5 finishers with incomplete data. Historical arithmetic gives group-specific nonfinisher counts of 2 participants/5 participants and incomplete-data counts after finishing of 2 participants/3 participants. These subtraction records are not new calculations, and the comparison of 32 participants does not recover the effect in all 44 assigned participants.
- Why unverified: Loss stages and retained group-specific counts have been linked, but the reasons, handling and estimation of the effect of interest in all starters could not be verified.
- Reason for missing information: Individual reasons for loss, analysis of all participants, handling of selection/missingness and appropriate sensitivity analyses are unverified.
- Required materials and conditions: Individual reasons for nonfinishing/missing data, the analysis plan for the effect in all assigned participants, and raw data
- Materials to obtain first: Individual reasons for nonfinishing/missing data, the analysis plan for the effect in all assigned participants, and raw data
- Further research: Define the prespecified effect of interest including nonfinishers and missing-data sensitivity analyses, and report the full data flow.
- Conditions for verification: Separate nonfinishing from incomplete data among finishers, and retain the unverified status of estimation for all starters.
- Why this condition is needed: Needed to prevent different reasons for loss from being combined or selected completers from being interpreted as the effect in all assigned participants.
- Application exceptions: Do not arbitrarily assign times to nonfinishers without finishing times; retain historical subtraction values as separate calculation records.
- Remaining uncertainty: Individual reasons for loss, analysis of all participants, handling of selection/missingness and appropriate sensitivity analyses are unverified.
- Referenced methods: M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptations are self-designed; M08 §8.2.2, M15 §15.3.1; self-developed application
- Related sources: S01

### N14 · For what distance, at what time point and with what device was time measured?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The outcome is finishing time (in hours) at completion of one 161.3 km ultramarathon race. It is the value for study completers in Table 1, not a standardized laboratory time trial or a 400 m recovery run. Timing-device and recording/adjudication details are unverified.
- Why unverified: The endpoint, distance, unit and completion time point were verified, but actual recording/verification procedures are unverified within the retained scope.
- Reason for missing information: The timing device, recording personnel/adjudication, measurement error and adjustment for baseline course ability are unverified.
- Required materials and conditions: Official race-completion time, timing device/adjudication, course conditions, record verification and Table 1 location
- Materials to obtain first: Official race-completion time, timing device/adjudication, course conditions, record verification and Table 1 location
- Further research: Disclose official timing records and course/adjudication procedures, and audit measurement errors/differences.
- Conditions for verification: Distinguish the time unit at actual 161.3 km race completion from unverified recording procedures.
- Why this condition is needed: Needed to prevent field-race finishing time from being treated as equivalent to a laboratory test or recovery run.
- Application exceptions: If device/adjudication details are unverified, do not declare measurement complete merely because the outcome is objective.
- Remaining uncertainty: The timing device, recording personnel/adjudication, measurement error and adjustment for baseline course ability are unverified.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M08 Table 8.2.a, M03 §3.2.4; self-developed application
- Related sources: S01

### N15 · Was finishing time a prespecified primary efficacy endpoint?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The study focuses on pain and recovery, and finishing time appears in a characteristics table. A registration/planned hierarchy establishing finishing time as a prespecified primary efficacy endpoint could not be verified.
- Why unverified: The fact that an outcome was reported does not by itself establish a prespecified primary (1st) endpoint or confirmatory analysis.
- Reason for missing information: Registration/protocol/SAP and the finishing-time hierarchy and analysis/multiple-comparison plan are unverified.
- Required materials and conditions: Registration/protocol/SAP, endpoint/time-point hierarchy, the table's purpose and reasons for changes
- Materials to obtain first: Registration/protocol/SAP, endpoint/time-point hierarchy, the table's purpose and reasons for changes
- Further research: Fix the primary/supporting role of finishing time, testing hierarchy and handling of multiple comparisons before the race.
- Conditions for verification: Preserve the recovery focus of the study and the location of time in the characteristics table, and indicate that the prespecified hierarchy is unverified.
- Why this condition is needed: Needed to prevent time in a characteristics table from being relabeled as a prespecified primary (1st) efficacy outcome.
- Application exceptions: Do not convert failure to acquire the plan into confirmed selective reporting or lack of registration.
- Remaining uncertainty: Registration/protocol/SAP and the finishing-time hierarchy and analysis/multiple-comparison plan are unverified.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M08 Table 8.2.a, M04 §4.6.3; self-developed application
- Related sources: S01

### N16 · Are the Table 1 means and dispersion measures attributed to the same population/units?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: In Table 1, the 18 B2 participants have 27.26 ± 2.36 hours and the 14 placebo participants have 27.08 ± 2.88 hours, expressed as mean±SD. This SD is not a between-group CI or SE.
- Why unverified: The original means, dispersion notation and group-specific denominators were accurately linked to the retained results/inputs. Meeting the numerical-attribution condition is separate from establishing causal efficacy.
- Reason for missing information: There is no gap in the identity of the original table summaries. Original between-group CI/P/adjusted values remain unverified in N18.
- Required materials and conditions: Group-specific complete-data n, mean time, SD notation, and original table/units
- Materials to obtain first: Group-specific complete-data n, mean time, SD notation, and original table/units
- Further research: Disclose the dispersion definition for each outcome together with precise numerical values/denominators.
- Conditions for verification: Preserve 27.26±2.36 hours for B2 n=18 and 27.08±2.88 hours for placebo n=14 as mean±SD.
- Why this condition is needed: Needed to prevent SD from being read as SE or CI, or linked to a different denominator.
- Application exceptions: Do not use mean±SD as individual differences/between-group CI, or extract precision beyond the table summaries.
- Remaining uncertainty: There is no gap in the identity of the original table summaries. Original between-group CI/P/adjusted values remain unverified in N18.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptations are self-designed
- Related sources: S01

### N17 · Has nonsignificance for time been kept from being converted into an effect of 0 or equivalence?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The retained original table/manuscript states that a significant between-group difference in finishing time was not identified. Alternating allocation and selected-finisher analysis prevent this from establishing an effect of 0, equivalence or exclusion of important time benefit.
- Why unverified: The nonsignificance statement and its interpretation limits were linked to the same time endpoint and preserved. This condition being met does not establish no effect.
- Reason for missing information: Original exact P/CI, important-difference criteria and causal analyses are unverified.
- Required materials and conditions: The original table's no-difference statement, coherent between-group estimates/intervals and a validated important-difference criterion
- Materials to obtain first: The original table's no-difference statement, coherent between-group estimates/intervals and a validated important-difference criterion
- Further research: Prespecify coherent between-group intervals and important-difference criteria to provide interpretation beyond significance.
- Conditions for verification: Link the retained nonsignificant result and current interpretation limits to the same endpoint.
- Why this condition is needed: Needed to prevent nonsignificant results from being favorably omitted or expressed as definitive absence of effect.
- Application exceptions: Do not obscure nonsignificance with favorable results for other recovery outcomes, or equate nonsignificance with exclusion of important benefit.
- Remaining uncertainty: Original exact P/CI, important-difference criteria and causal analyses are unverified.
- Referenced methods: M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed; M15 §15.3.1–15.3.3; self-developed application
- Related sources: S01

### N18 · Are author-reported between-group CI, exact P and adjusted estimates available?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Author-reported finishing-time CI, exact P and adjusted causal effect are null in the retained stored summary. Historical t-reconstruction values are recorded separately and do not fill these original values.
- Why unverified: The required originally reported estimates/tests and adjusted values could not be verified and cannot be replaced with existing descriptive calculations.
- Reason for missing information: Author CI, exact P, adjusted model and the original estimate of the overall effect of interest are unverified.
- Required materials and conditions: CI for the original time outcome, exact test P, adjusted model/estimand and results table
- Materials to obtain first: CI for the original time outcome, exact test P, adjusted model/estimand and results table
- Further research: Disclose the original between-group CI/P and model/denominator for the prespecified effect of interest on finishing time.
- Conditions for verification: Separate verification status for original CI, exact P and adjusted effect from historical recalculation.
- Why this condition is needed: Needed to prevent recalculated summary statistics from being used as original author reports or adjusted causal estimates.
- Application exceptions: If an original value is null, do not fill the gap even when historical calculations exist.
- Remaining uncertainty: Author CI, exact P, adjusted model and the original estimate of the overall effect of interest are unverified.
- Referenced methods: M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptations are self-designed; M08 §8.2.2, M15 §15.3.1; self-developed application; M15 §15.3.1–15.3.3; self-developed application
- Related sources: S01

### N19 · Are the sign and descriptive meaning of the retained mean difference preserved?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The retained mean difference is B2−placebo +0.18 hours (+10.8 minutes); a positive value is descriptively in the slower direction. The original retained values are 0.18000000000000327 hours and 10.800000000000196 minutes. This is arithmetic on selected-sample means, not a new estimate of causal harm or significant delay.
- Why unverified: The inputs, direction and original values of the retained calculation were accurately distinguished from the rounded reader-facing expression. No new calculation was performed in this work.
- Reason for missing information: The causal difference for all assigned participants/with adjustment and its original interval are unverified.
- Required materials and conditions: Historical arithmetic inputs, B2−placebo direction, original mean-difference values/limits and calculation provenance
- Materials to obtain first: Historical arithmetic inputs, B2−placebo direction, original mean-difference values/limits and calculation provenance
- Further research: Subsequent research should report the prespecified direction and between-group estimates/intervals for the overall effect of interest.
- Conditions for verification: Specify the direction of the historical +0.18 hours (+10.8 minutes) and its limitation as a descriptive value for a selected sample.
- Why this condition is needed: Needed to prevent a positive time difference from being converted into new causal harm or significant delay.
- Application exceptions: Perform no new calculations in this work, and retain both rounded displays and the original stored floating-point values.
- Remaining uncertainty: The causal difference for all assigned participants/with adjustment and its original interval are unverified.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M08 §8.2.2, M15 §15.3.1; self-developed application; M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed
- Related sources: S01

### N20 · Can the assumptions of the equal-variance t-reconstructed interval be verified?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The historical pooled-variance unpaired t model is a descriptive model for the selected sample. The retained 95% CI is [−1.7108254419152404, 2.0708254419152468] hours, P = 0.8471597784689339, SE = 0.925843873613083 hours, and df = 30. Equal variance is an explicit calculation assumption, not a verified author-used t subtype. It does not remedy missingness/confounding.
- Why unverified: The historical calculation's model, values and assumptions were linked separately from the unverified original CI. This meets a calculation-preservation condition, not a causal-precision condition.
- Reason for missing information: The original author interval, t subtype and a coherent estimate accounting for all assigned participants/confounding are unverified.
- Required materials and conditions: Original retained pooled-calculation values, inputs, assumptions, selected sample and distinction from original reporting
- Materials to obtain first: Original retained pooled-calculation values, inputs, assumptions, selected sample and distinction from original reporting
- Further research: Disclose participant-level data and a prespecified model to examine direct estimation that accounts for selection/nonrandomization limitations.
- Conditions for verification: Attribute the existing CI/P/SE/df solely to an assumption-dependent descriptive model.
- Why this condition is needed: Needed to preserve existing arithmetic without promoting it to an original CI or causal precision.
- Application exceptions: Do not establish that the authors used the equal-variance t subtype. Reconstruction does not remedy missingness/confounding.
- Remaining uncertainty: The original author interval, t subtype and a coherent estimate accounting for all assigned participants/confounding are unverified.
- Referenced methods: M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptations are self-designed; M08 §8.2.2, M15 §15.3.1; self-developed application; M15 §15.3.1–15.3.3; self-developed application
- Related sources: S01

### N21 · Were the unequal-variance sensitivity results also preserved as a separate model?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The historical Welch unpaired t sensitivity analysis uses an unequal-variance model. The retained 95% CI is [−1.7761840334977923, 2.136184033497799] hours, P = 0.8512055048234594, SE = 0.9496732938644559 hours, and df = 24.925799135526574. These results use assumptions different from pooled analysis; they are neither a source-data conflict nor author CI.
- Why unverified: Both models were preserved and attributed to their respective assumptions; the conclusion was not changed through new recalculation/model selection.
- Reason for missing information: Gaps concerning selection/confounding, the original author interval and the overall effect of interest remain.
- Required materials and conditions: Retained Welch values, assumptions, inputs, differences from pooled results and verification status of the original interval
- Materials to obtain first: Retained Welch values, assumptions, inputs, differences from pooled results and verification status of the original interval
- Further research: Prespecify variance assumptions and sensitivity analyses for missingness/confounding, and report model-specific results transparently.
- Conditions for verification: Link the historical Welch CI/P/SE/df to that model and explain the remaining causal limitations.
- Why this condition is needed: Needed to prevent model selection that presents only favorable intervals or relabels them as author CI.
- Application exceptions: Do not classify differences between sensitivity-model results as a conflict within the same source data. Perform no new recalculation.
- Remaining uncertainty: Gaps concerning selection/confounding, the original author interval and the overall effect of interest remain.
- Referenced methods: M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed; M08 §8.2.2, M15 §15.3.1; self-developed application; M15 §15.3.1–15.3.3; self-developed application
- Related sources: S01

### N22 · Is there a basis for deciding how much faster a time constitutes important benefit?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: A validated important difference (MCID) for the same race/runners is null. The retained record has equivalence_demonstrated=false, and reconstructed intervals allow both faster and slower times. NNT is null because this continuous-time/selected descriptive comparison has no eligible binary-risk/time estimate.
- Why unverified: The relationship between an important difference and a coherent original interval could not be verified, so exclusion of important benefit or equivalence cannot be claimed.
- Reason for missing information: Justified important-difference and equivalence thresholds and a prespecified precision design are unverified.
- Required materials and conditions: Justified important-difference and equivalence thresholds for the same race/population, and appropriate original estimation intervals
- Materials to obtain first: Justified important-difference and equivalence thresholds for the same race/population, and appropriate original estimation intervals
- Further research: Justify an important difference in the runner/race context and prespecify a precision/equivalence design incorporating it.
- Conditions for verification: Distinguish the unverified important difference, unestablished equivalence and inapplicability of NNT.
- Why this condition is needed: Needed to prevent CX in the existing score or nonsignificance from being used to exclude important benefit.
- Application exceptions: Do not create universal thresholds for minutes, participant numbers or duration, or force NNT onto a continuous-time question.
- Remaining uncertainty: Justified important-difference and equivalence thresholds and a prespecified precision design are unverified.
- Referenced methods: M15 §15.3.1–15.3.3; self-developed application; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed
- Related sources: S01

### N23 · Is a comparative effect accounting for baseline racing ability and selection verified?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Adjustment for baseline racing ability on an equivalent course and adjusted causal estimation have not been verified. The mean difference from predictable alternating allocation and selected-completer analysis cannot be regarded as estimating the intervention effect in all starters.
- Why unverified: Baseline balance/adjustment and full-analysis data needed for the causal effect of interest could not be verified.
- Reason for missing information: Baseline racing ability, balance in nutrition/care, covariates/models and the effect in all assigned participants are unverified.
- Required materials and conditions: Baseline racing ability, experience on the same course, nutritional/care covariates and prespecified adjustment/analysis of all participants
- Materials to obtain first: Baseline racing ability, experience on the same course, nutritional/care covariates and prespecified adjustment/analysis of all participants
- Further research: Measure racing ability before allocation and prespecify the effect of interest and handling of covariates/losses.
- Conditions for verification: Link the unverified adjusted effect and baseline balance to limitations on the current effect of interest.
- Why this condition is needed: Needed to prevent the selected-completer mean difference from being read as the effect of B2 in all assigned participants.
- Application exceptions: Lack of adjustment does not establish the magnitude of bias; do not adjust values using a new model.
- Remaining uncertainty: Baseline racing ability, balance in nutrition/care, covariates/models and the effect in all assigned participants are unverified.
- Referenced methods: M08 §8.1–8.2.3; self-developed application; M08 §8.2.2, M15 §15.3.1; self-developed application
- Related sources: S01

### N24 · Have the official registration, protocol and SAP been compared with the original report?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: S01 registration_id is null; registration was not reported in the original at the time, and the external record was recorded as unverified. Ethics approval or the existence of an article is not reading the full registration, protocol or SAP. Access to the relevant external materials was not attempted in this work; they were not acquired or read.
- Why unverified: Full texts needed to actually compare the prospective plan with the original results were not acquired/read in this work. This differs from absence of registration or confirmed selective reporting.
- Reason for missing information: Original registration identity, full text, plan version/date and endpoint/analysis change history are unverified.
- Required materials and conditions: Full original registration/protocol/SAP, versions/dates and endpoint/analysis comparisons
- Materials to obtain first: Full original registration/protocol/SAP, versions/dates and endpoint/analysis comparisons
- Further research: Disclose original plans and change histories, and compare outcome-specific plan adherence/reasons for changes.
- Conditions for verification: Indicate the scope of plans acquired/read and the scope not compared according to the actions actually taken in this work.
- Why this condition is needed: Needed to prevent ethics approval or study existence from being treated as reading the registration/plans.
- Application exceptions: S01 registration_id=null does not prove absence of an external registration that is currently unverified.
- Remaining uncertainty: Original registration identity, full text, plan version/date and endpoint/analysis change history are unverified.
- Referenced methods: Material collection/selection principles in M04 §4.5–4.6; distinctions between access and reading and the hyperlink contract are self-designed; M04 §4.5, §4.6.3; limits on the current material conversion follow the user-defined scope and self-designed rules; M08 Table 8.2.a, M04 §4.6.3; self-developed application
- Related sources: S01

### N25 · Are the conference title and original journal article being counted as independent replications?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The HTML/PDF/index versions of the S01 original article are the same report. S04 is a probable conference companion on the same investigators/intervention topic, but full-text/exact-identity comparison is incomplete. It is not counted as a separate cohort or independent replication; randomized in the conference title differs from the alternating allocation in S01.
- Why unverified: The probable family link and exclusion of duplicate counting were verified, but the exact identity of the full conference text is unverified.
- Reason for missing information: The full S04 abstract and exact family matching of participants/time points are unverified.
- Required materials and conditions: S01/S04 original reports, full conference text, cohort/time-point/participant linkage and design wording
- Materials to obtain first: S01/S04 original reports, full conference text, cohort/time-point/participant linkage and design wording
- Further research: Link participants, registrations and time points across related reports, and disclose a map of unique families and outcome locations.
- Conditions for verification: Preserve probable-companion status, lack of acquired full text and conflicting design wording in the title; do not count it as a separate replication.
- Why this condition is needed: Needed to prevent a companion report from the same investigators/intervention from being counted as two race trials.
- Application exceptions: Do not overstate a probable same-family link as a confirmed finding about participant independence/overlap.
- Remaining uncertainty: The full S04 abstract and exact family matching of participants/time points are unverified.
- Referenced methods: Distinction between studies and reports in M04 §4.6.1–4.6.2; cohort/outcome linkage is a self-developed application; M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed; M04 §4.6.1–4.6.2, M15 §15.3.1; self-developed application
- Related sources: S01, S04

### N26 · Are positive signals for pain/400 m recovery being substituted as time evidence?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Favorable signals for some pain and 400 m recovery results from the same study are preserved. These outcomes differ from finishing time for the 161.3 km race and do not replace the nonsignificant finishing-time result. CK and vitamin status are also separate endpoints.
- Why unverified: Favorable adjacent outcomes and the boundary of the selected time outcome were preserved together according to the manuscript. No new numerical values or intervals for positive signals were created.
- Reason for missing information: This table does not independently adjudicate recovery efficacy. Precise numerical values/hierarchy for those signals were not provided in this current summary.
- Required materials and conditions: Endpoint-specific results, time points, tests/hierarchy and finishing-time separation records from the same study
- Materials to obtain first: Endpoint-specific results, time points, tests/hierarchy and finishing-time separation records from the same study
- Further research: Prespecify race performance and post-race recovery separately, and report endpoint-specific estimates/intervals and multiple comparisons.
- Conditions for verification: Distinguish the endpoints of some recovery/pain signals from the actual time outcome.
- Why this condition is needed: Needed to preserve favorable results without obscuring the no-difference result for the selected finishing-time outcome.
- Application exceptions: Do not add numbers/CI to recovery signals for which the currently retained manuscript supplies no numerical values.
- Remaining uncertainty: This table does not independently adjudicate recovery efficacy. Precise numerical values/hierarchy for those signals were not provided in this current summary.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed; M03 §3.2.4; self-defined extension boundary
- Related sources: S01

### N27 · Does time to exhaustion with a combination product in nonathletes answer the single-ingredient B2 race-time question?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: S02 Ex PLUS is a combination of multiple vitamins, minerals, taurine and inositol. It involved 32 healthy nonathletes, a 28-day intervention and a 14-day washout, testing time to exhaustion at 85% VO2max. Do not combine it as field-race finishing-time evidence for single-ingredient B2.
- Why unverified: The actual differences in ingredients, population, duration and endpoint in the current materials have been linked, allowing the exclusion role to be explained. This status does not confirm Ex PLUS or B2 efficacy.
- Reason for missing information: A direct comparison in the same race and independent replication of single-ingredient B2 have not been verified. Independent verification of the reported registration NCT05586295 is recorded as false.
- Required materials and conditions: S02 ingredient table, population, exposure/washout, actual performance test and registration/original results
- Materials to obtain first: S02 ingredient table, population, exposure/washout, actual performance test and registration/original results
- Further research: Design a direct comparison in the same ultramarathon population differing only in single-ingredient B2.
- Conditions for verification: Record the boundary of a combination product in 32 healthy nonathletes, 28 days, a 14-day washout and time to exhaustion at 85% VO2max.
- Why this condition is needed: Needed to prevent differences in composition, population and endpoint in indirect evidence from being combined as replications of single-ingredient B2.
- Application exceptions: Do not isolate the B2 component's effect from a combination product in nonathletes or convert it into race finishing time.
- Remaining uncertainty: A direct comparison in the same race and independent replication of single-ingredient B2 have not been verified. Independent verification of the reported registration NCT05586295 is recorded as false.
- Referenced methods: M03 §3.2.1; detailed population criteria are self-designed; M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M04 §4.6.1–4.6.2, M15 §15.3.1; self-developed application
- Related sources: S02

### N28 · Have the swimming article's eligibility and numerical values been verified?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: S05 is a candidate identified through the title and bibliography of a 1984 study of B2 supplementation, nutritional status and performance in elite swimmers. Incomplete access to the original/DOI resolution at the time prevented verification of age, exact endpoint, dose, design and effect/CI. The external original remains unread in the current work.
- Why unverified: Bibliographic identity is preserved, but the eligibility and precise numerical values of the performance data are unverified.
- Reason for missing information: Original-article methods/results, participant age, denominators and effect/intervals are unverified.
- Required materials and conditions: Full original 1984 article, age/ingredient/dose, design, performance test, denominators and effect/intervals
- Materials to obtain first: Full original 1984 article, age/ingredient/dose, design, performance test, denominators and effect/intervals
- Further research: In a subsequent review that acquires the original, verify the actual sport, test and population, and separate evidence according to directness.
- Conditions for verification: Display title/bibliographic identity separately from unverified eligibility/numerical values.
- Why this condition is needed: Needed to prevent bibliographic discovery from being promoted to a performance effect or independent replication.
- Application exceptions: Do not use historical inaccessibility of the original as absence of the article/study; distinguish it from access not attempted in the current work.
- Remaining uncertainty: Original-article methods/results, participant age, denominators and effect/intervals are unverified.
- Referenced methods: Material collection/selection principles in M04 §4.5–4.6; distinctions between access and reading and the hyperlink contract are self-designed; M03 §3.2.1; detailed population criteria are self-designed; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M04 §4.6.1–4.6.2, M15 §15.3.1; self-developed application
- Related sources: S05

### N29 · Have the Winters1992 population and the swimming study been separated?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The title of S06 Winters1992 concerns B2 requirements/exercise adaptation in older women. QA01 separated it from the 1984 elite-swimming candidate. Access at the time was limited to official-index search metadata; no direct performance numbers were imported.
- Why unverified: The retained population/question correction and the boundary preventing numerical transfer were linked. Verification of the original article's results was not claimed.
- Reason for missing information: Exact exposure, endpoints and denominators/numerical values in the S06 original/abstract are unverified.
- Required materials and conditions: S06 bibliography, original article/abstract, exercise-adaptation/nutrient-requirement endpoints and QA01 correction
- Materials to obtain first: S06 bibliography, original article/abstract, exercise-adaptation/nutrient-requirement endpoints and QA01 correction
- Further research: Subsequent original-source review should first verify each population, endpoint and ingredient, then extract by independent question.
- Conditions for verification: Separate the exercise-adaptation/requirement question in older women from the elite-swimming candidate.
- Why this condition is needed: Prevents a mistaken population link in exploratory queries from being carried over to numerical values from another article.
- Application exceptions: Do not fill in actual performance numbers or effects solely from title verification.
- Remaining uncertainty: Exact exposure, endpoints and denominators/numerical values in the S06 original/abstract are unverified.
- Referenced methods: M03 §3.2.1; detailed population criteria are self-designed; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed
- Related sources: S06, S05

### N30 · Is post-exercise B2 status in adult men an effect of supplementation→time?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: S07 addresses the direction of exercise affecting B2 status in adult men. Historical access was limited to the publisher landing page, and this status response is not a direct result showing that B2 supplementation reduces race time.
- Why unverified: The boundary explaining the direction and endpoint differences in the current title/facts has been linked.
- Reason for missing information: Actual exposure, tests, time points and results/denominators in the original have not been verified in the current work.
- Required materials and conditions: S07 original article, exposure/outcome direction, nutritional-status tests/time points and actual comparison
- Materials to obtain first: S07 original article, exposure/outcome direction, nutritional-status tests/time points and actual comparison
- Further research: Design the performance-improvement question separately as a comparison differing only in B2 exposure.
- Conditions for verification: Indicate the exercise→B2 status direction shown by the current title/facts and its difference from the performance question.
- Why this condition is needed: Prevents nutritional-status responses to exercise from being mixed with supplement effects on performance in the reverse direction.
- Application exceptions: Do not convert facts about status endpoints into supplementation effects or clinical deficiency treatment.
- Remaining uncertainty: Actual exposure, tests, time points and results/denominators in the original have not been verified in the current work.
- Referenced methods: M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M03 §3.2.4; self-defined extension boundary
- Related sources: S07

### N31 · Is the exercise challenge in women by fitness level a comparison of oral B2 versus placebo?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The S09 female data compare exercise challenge, fitness groups and B2 status; they are not a randomized oral B2 versus placebo supplementation trial. Use of B2-containing supplements was excluded, and the original Methods registration number was recorded as NL7891. Historical access to portions of abstract/Methods copies is separate from verification of the full official registration.
- Why unverified: The retained original Methods' exposure direction, exclusions and registration number were linked to their differences from the direct performance question. This does not mean the entire original/registration was read in this work.
- Reason for missing information: The full official registration and complete raw data/results were not verified in this work. This is not the same single-ingredient B2 time comparison.
- Required materials and conditions: S09 original abstract/Methods, exposure/exclusion criteria, results, and original location and registration of NL7891
- Materials to obtain first: S09 original abstract/Methods, exposure/exclusion criteria, results, and original location and registration of NL7891
- Further research: A performance-comparison question for single-ingredient B2 requires a separate intervention/control trial in the same population.
- Conditions for verification: Link the exclusion of B2-containing supplement use, exercise challenge, status endpoint and original NL7891 separately.
- Why this condition is needed: Needed to prevent exercise-challenge observations from being called a single-ingredient B2 supplementation trial.
- Application exceptions: Verifying a registration number in the original Methods is separate from reading the full official registration.
- Remaining uncertainty: The full official registration and complete raw data/results were not verified in this work. This is not the same single-ingredient B2 time comparison.
- Referenced methods: M03 §3.2.1; detailed population criteria are self-designed; M03 §3.2.2–3.2.3; reverse-direction markers and test adaptations are self-designed; M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are self-designed; M04 §4.5, §4.6.3; limits on the current material conversion follow the user-defined scope and self-designed rules
- Related sources: S09

### N32 · Has discovery of the female study's registration been distinguished from the adjacent TCTR entry?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: NL-OMON24874 and NL-OMON48355 in S10 are discovery-level candidate aliases, and the detailed official registration was inaccessible at the time. QA05 identified TCTR20190907005 as an adjacent different entry, not the female study's registration. Unverified registration details were not used to fill in the S09 methods.
- Why unverified: The retained correction of mistaken identity has been linked, but exact comparison of official details/aliases is unverified.
- Reason for missing information: Comparison of official identities, full registration/version and methods between the OMON candidates and NL7891 is unverified.
- Required materials and conditions: Official full text/identity of NL7891 and OMON candidates, comparison with the original Methods, and discovery-page structure
- Materials to obtain first: Official full text/identity of NL7891 and OMON candidates, comparison with the original Methods, and discovery-page structure
- Further research: Compare the identities of the official registration and original article using source materials, and mark uncertain aliases.
- Conditions for verification: Preserve separation of the adjacent TCTR entry, nonacquisition of official details and the S09 original Methods basis.
- Why this condition is needed: Needed to prevent registration numbers in adjacent search results from being used to fill in the original study's identity/methods.
- Application exceptions: OMON aliases remain at the discovery level; do not promote unread official details to verified methods.
- Remaining uncertainty: Comparison of official identities, full registration/version and methods between the OMON candidates and NL7891 is unverified.
- Referenced methods: Distinction between studies and reports in M04 §4.6.1–4.6.2; cohort/outcome linkage is a self-developed application; material collection/selection principles in M04 §4.5–4.6; distinctions between access and reading and the hyperlink contract are self-designed; M04 §4.5, §4.6.3; limits on the current material conversion follow the user-defined scope and self-designed rules
- Related sources: S10, S09

### N33 · Are repeated citations in reviews and limited searching being read as independent replication or absence of studies?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The retained direct-comparison family is UF2016 alone, and the recorded number of verified independent replications is 0. The S08 review's repeated citation of Hoffman is not a new trial. Historical searching comprised 63 general-web queries and 3 failed native database searches, without assurance of exhaustive searching. No new search was conducted in this work.
- Why unverified: Families/repeated citations and the actual historical search scope have been linked, but independent replication of the same question and exhaustive searching for unpublished work/corrections have not been verified.
- Reason for missing information: Direct independent replication and the exhaustive status of unpublished work, registrations and corrections are unverified. This does not prove study absence or current up-to-dateness.
- Required materials and conditions: Unique-family/report map, actual search records, registration/unpublished/correction scope and full review article
- Materials to obtain first: Unique-family/report map, actual search records, registration/unpublished/correction scope and full review article
- Further research: Subsequent reviews should transparently record the scope of searches, failures, registrations and corrections, and link independent comparisons addressing the same question.
- Conditions for verification: Display the single direct family, unverified independent replication, repeated citations in reviews and search limitations together.
- Why this condition is needed: Needed to prevent one family and an incomplete search from being converted into replication success/failure or 0 studies worldwide.
- Application exceptions: The 63 web queries and 3 failures are historical records, not new successful searches or current exhaustive verification.
- Remaining uncertainty: Direct independent replication and the exhaustive status of unpublished work, registrations and corrections are unverified. This does not prove study absence or current up-to-dateness.
- Referenced methods: Distinction between studies and reports in M04 §4.6.1–4.6.2; cohort/outcome linkage is a self-developed application; M04 §4.6.3, M15 §15.3.1–15.3.3; status distinctions are self-designed; M04 §4.5, §4.6.3; limits on the current material conversion follow the user-defined scope and self-designed rules; M04 §4.6.1–4.6.2, M15 §15.3.1; self-developed application
- Related sources: S01, S04, S08

### N34 · What is the scope of foundation/public support and no-conflict declarations?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: S01 records support from the Western States Endurance Run Foundation, Rossi Family Foundation and VA resources/facilities, and the authors' declaration of no competing interests. Support roles, all donor provenance and product provision are unverified. Historical I2 is not certification of independent replication, external review or absence of influence.
- Why unverified: The level at which funding/declarations were verified is preserved, but information covering the full scope of independence is lacking.
- Reason for missing information: Support roles, donors, product provision and the existence of independent study families/external review are unverified.
- Required materials and conditions: Funding/product provision, donor provenance, support roles, author declarations/employment and family map
- Materials to obtain first: Funding/product provision, donor provenance, support roles, author declarations/employment and family map
- Further research: Disclose support roles, product provision, all interests and linkage to independent study families.
- Conditions for verification: Explain actual funding/declarations and the boundaries of unverified support information.
- Why this condition is needed: Needed to prevent reported declarations from being used as proof of external review, absence of bias or independent replication.
- Application exceptions: Do not fill gaps concerning donors, product provision or roles in foundation/public support with an assumption of no corporate influence.
- Remaining uncertainty: Support roles, donors, product provision and the existence of independent study families/external review are unverified.
- Referenced methods: Chamgap's own independence/attribution criteria. The official methods sections in this pilot do not validate every detail of this item; distinction between studies and reports in M04 §4.6.1–4.6.2; cohort/outcome linkage is a self-developed application
- Related sources: S01

### N35 · What can be concluded about vomiting after the 90 km dose?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The original record reports that two participants in the B2 group vomited within 1 hour after the 90 km dose. The temporal relationship of an event after dosing is preserved, but B2 causation and comparative risk/incidence are not established.
- Why unverified: The participant count and temporal signal were verified, but comparative ascertainment/denominators, recovery/re-exposure and alternative explanations were not sufficiently verified.
- Reason for missing information: Systematic harm denominators, equivalent placebo ascertainment, discontinuation/recovery/re-exposure and alternative explanations/concomitant use during the race are unverified.
- Required materials and conditions: Number of participants with vomiting, onset/recovery, administration/discontinuation/re-exposure, alternative causes, and placebo ascertainment/denominators
- Materials to obtain first: Number of participants with vomiting, onset/recovery, administration/discontinuation/re-exposure, alternative causes, and placebo ascertainment/denominators
- Further research: Systematically collect exposure/event timing, discontinuation/recovery, alternative explanations and concomitant use, and plan comparable harm analyses.
- Conditions for verification: Preserve the record of 2 participants in the B2 group within 1 hour after administration at 90 km as a temporal signal.
- Why this condition is needed: Needed to separate the occurrence of an event after dosing from comparative risk/causation.
- Application exceptions: Do not infer the amount absorbed after vomiting or other causes, and do not convert 2 participants into an incidence rate.
- Remaining uncertainty: Systematic harm denominators, equivalent placebo ascertainment, discontinuation/recovery/re-exposure and alternative explanations/concomitant use during the race are unverified.
- Referenced methods: M19 §19.1.2–19.2.4; self-developed application; M19 §19.1.2–19.2.4; detailed distinctions are self-designed; design-specific harm questions in M19 §19.2; details of temporal relationships are self-designed; design considerations in M19 §19.2.2–19.2.4; detailed confounding records are self-designed; Chamgap's own interpretation boundary. Informed by M19 principles, not execution of an official causality tool
- Related sources: S01

### N36 · Do the 3 participants versus 1 participant noticing urine color represent harm incidence?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The retained counts of participants who noticed urine color are 3 in the B2 group and 1 in the placebo group. This record signals a limitation in maintained blinding; it does not establish participant-level incidence of systematically collected adverse events or clinical harm.
- Why unverified: The awareness record was verified, but actual ascertainment denominators/definitions, timing and information for clinical-harm interpretation are unverified.
- Reason for missing information: Awareness-ascertainment denominators, timing, relationships to allocation guesses/unblinding and clinical-event classification are unverified.
- Required materials and conditions: Definition/timing of awareness, participant counts, ascertainment scope, actual denominators in both groups and blinding checks
- Materials to obtain first: Definition/timing of awareness, participant counts, ascertainment scope, actual denominators in both groups and blinding checks
- Further research: Collect awareness/allocation guesses identically in both groups and disclose ascertainment denominators and onset timing.
- Conditions for verification: Link the report of 3 participants versus 1 participant to limits on interpreting blinding maintenance/safety.
- Why this condition is needed: Needed to prevent awareness counts from being converted into systematic adverse-event incidence or a confirmed proportion with allocation unblinded.
- Application exceptions: Do not automatically substitute the table's time-analysis denominator or the starter denominator for the awareness-ascertainment denominator.
- Remaining uncertainty: Awareness-ascertainment denominators, timing, relationships to allocation guesses/unblinding and clinical-event classification are unverified.
- Referenced methods: M08 Table 8.2.a, M03 §3.2.4; self-developed application; M19 §19.1.2–19.2.4; self-developed application; M19 §19.1.2–19.2.4; detailed distinctions are self-designed
- Related sources: S01

### N37 · Were adverse events, serious events and discontinuations due to harm systematically counted?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: Group-specific denominators and ascertainment scope for systematic adverse events, serious adverse events and discontinuations due to harm have not been verified. The existence of vomiting/color-awareness reports is distinct from complete ascertainment or absence of events.
- Why unverified: The required systematic definitions, ascertainment and exposure/loss/participant-level denominators could not be verified. Unverified information was not converted into 0 events.
- Reason for missing information: AE/SAE/discontinuation definitions, actual group-specific denominators/durations, ascertainment/detection and reasons for dropout are unverified.
- Required materials and conditions: AE/SAE definitions in both groups, ascertainment procedures, participants/events, exposure time, discontinuations and loss denominators
- Materials to obtain first: AE/SAE definitions in both groups, ascertainment procedures, participants/events, exposure time, discontinuations and loss denominators
- Further research: Track exposure and losses among all assigned participants, and record events, participants and discontinuations using the same procedures in both groups.
- Conditions for verification: Retain currently verified signals and specific gaps in systematic ascertainment/denominators.
- Why this condition is needed: Needed to prevent a few event records and efficacy data from being converted into complete safety assurance.
- Application exceptions: Do not use unreported/unverified information as 0 events or proof of no serious events.
- Remaining uncertainty: AE/SAE/discontinuation definitions, actual group-specific denominators/durations, ascertainment/detection and reasons for dropout are unverified.
- Referenced methods: M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptations are self-designed; M19 §19.1.2–19.2.4; self-developed application; M19 §19.1.2–19.2.4; detailed distinctions are self-designed; M19 §19.1.2–19.2.4, M04 §4.6.3; self-developed application; Chamgap's own interpretation boundary. Informed by M19 principles, not execution of an official causality tool
- Related sources: S01

### N38 · Have the meanings of no established UL and no known interactions been appropriately limited?

Verification status: Partly verified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The S03 NIH ODS context records that no UL was established because of limited harm data. This does not prove that high doses cannot cause adverse effects. No known clinically relevant drug interactions is likewise not extended to a guarantee of safety for all concomitant use.
- Why unverified: The retained general official context was verified, but it is not direct verification of all harms/concomitant use in the same race/at the same dose. The full official text was not read in this work.
- Reason for missing information: The latest official update and direct harm information for the same exposure/population and concomitant use are unverified.
- Required materials and conditions: S03 official safety/interaction original text, update date, evidence scope and data on the same exposure/population
- Materials to obtain first: S03 official safety/interaction original text, update date, evidence scope and data on the same exposure/population
- Further research: Review harm data addressing the same dose/duration/concomitant use and official updates separately.
- Conditions for verification: Indicate that the lack of an established UL reflects limited harm data and specify the scope of verification of known interactions.
- Why this condition is needed: Needed to prevent the lack of an established upper limit from being read as unlimited high-dose safety.
- Application exceptions: Do not count a general official document as a direct comparative harm result for this race/all concomitant use.
- Remaining uncertainty: The latest official update and direct harm information for the same exposure/population and concomitant use are unverified.
- Referenced methods: M04 §4.5, §4.6.3; limits on the current material conversion follow the user-defined scope and self-designed rules; M19 §19.1.2–19.2.4; self-developed application; Chamgap's own interpretation boundary. Informed by M19 principles, not execution of an official causality tool
- Related sources: S03

### N39 · Does the evidence also answer dose/duration, special-population and test-interference questions beyond one race?

Verification status: Unverified  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: One-race data do not verify pregnancy, lactation, children, renal/hepatic disease, long-term high-dose use or test interference. They do not answer questions about the optimality of other doses/repeated races or persistence of effects after discontinuation. Study doses are not personal dosing recommendations or instructions to replace symptom assessment/treatment during a race.
- Why unverified: Direct evidence for this extended scope could not be verified in the current retained-material review. Extension conditions were not promoted merely because a narrow time result exists.
- Reason for missing information: Direct evidence on long-term use, vulnerable groups, dose comparisons, concomitant use, test interference and persistent effects is unverified.
- Required materials and conditions: Direct evidence on long-term/repeated dosing, vulnerable populations/renal or hepatic conditions, pregnancy/lactation/children, concomitant use and test interference
- Materials to obtain first: Direct evidence on long-term/repeated dosing, vulnerable populations/renal or hepatic conditions, pregnancy/lactation/children, concomitant use and test interference
- Further research: For each needed extension, separately design a study with the appropriate population, duration and focus on harms/interactions or test interference.
- Conditions for verification: Distinguish the direct one-race scope from unverified long-term/special-population, interaction and interference questions.
- Why this condition is needed: Needed to prevent brief race data from being generalized to overall safety, optimal dosing or efficacy in all runners.
- Application exceptions: Do not require every extension study for the narrow finishing-time table, but explicitly mark unverified information in extension claims.
- Remaining uncertainty: Direct evidence on long-term use, vulnerable groups, dose comparisons, concomitant use, test interference and persistent effects is unverified.
- Referenced methods: M03 §3.2.1–3.2.4; details of dose, duration and effect modification are self-designed; M19 §19.1.2–19.2.4; self-developed application
- Related sources: S01, S03

### N40 · Has this summary review been distinguished from historical full-text and verdict records?

Verification status: Met  
Reading scope: Summary reading: the specified locations in the retained 3124 stored summary were actually read. Full articles/PDFs, appendices, registrations/protocols, raw data and corrections were not read in this work.

- Current findings: The materials actually read in this work are the Korean/structured evidence, bibliography, corrections and historical records in the currently retained 3124 stored summary. Original HTML/PDF access covering 8 pages and other partial access on 2026-09-16 are historical reports; access to external full texts was not attempted in this work, and they were not acquired or read. The original D·30/Caution and historical calculations/self-review are preserved; no new grade, verdict, search or calculation was produced.
- Why unverified: Original bytes/hashes, dates, current/historical reading and the scope of original-verdict preservation were recorded according to the actions actually taken. This table is not independent clinical/official certification or site publication.
- Reason for missing information: Actual reading and complete holdings of current external literature, registrations, appendices and historical separate files were not verified.
- Required materials and conditions: Current original/snapshot SHA, scope actually read, historical access/calculations, existing grade/score and reading receipts
- Materials to obtain first: Current original/snapshot SHA, scope actually read, historical access/calculations, existing grade/score and reading receipts
- Further research: In a subsequent case where new evidence is permitted, preserve original values/previous originals together with dates, access and locations/versions before reviewing linkage conditions.
- Conditions for verification: Separately record the original stored summary's bytes/date, current summary reading, nonreading of the original article and preservation of D/30/Caution.
- Why this condition is needed: Needed to prevent full-text access/self-review on 2026-09-16 from being represented as original-source reading or independent certification today.
- Application exceptions: Distinguish stored-original-file/hash null values from unexamined historical files, and do not claim a new download/reassessment was performed.
- Remaining uncertainty: Actual reading and complete holdings of current external literature, registrations, appendices and historical separate files were not verified.
- Referenced methods: Material collection/selection principles in M04 §4.5–4.6; distinctions between access and reading and the hyperlink contract are self-designed; M04 §4.5, §4.6.3; limits on the current material conversion follow the user-defined scope and self-designed rules
- Related sources: S01, S02, S03, S04, S05, S06, S07, S08, S09, S10

## Evidence and sources

### S01

A Placebo-Controlled Trial of Riboflavin for Enhancement of Ultramarathon Recovery

[Source](https://link.springer.com/article/10.1186/s40798-017-0081-4)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: 27.26 ± 2.36 hours in 18 B2 participants versus 27.08 ± 2.88 hours in 14 placebo participants (mean±SD); nonsignificant table result. Alternating allocation and selected finishers; original CI/exact P/adjusted effect unverified.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full article text on the web | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full article PDF | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full registration web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full protocol/SAP | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Raw data | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Correction materials | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S02

A functional evaluation of anti-fatigue and exercise performance improvement following vitamin B complex supplementation in healthy humans, a randomized double-blind trial

[Source](https://www.medsci.org/v20p1272.htm)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: Ex PLUS is a combination of vitamins/minerals/taurine/inositol; it involved 32 healthy nonathletes, a 28-day intervention, a 14-day washout and time to exhaustion at 85% VO2max. This is not a field-race time result for single-ingredient B2.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full article text on the web | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full registration web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S03

Riboflavin — Health Professional Fact Sheet

[Source](https://ods.od.nih.gov/factsheets/Riboflavin-HealthProfessional/)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: The lack of an established UL due to limited harm data does not mean unlimited safety. No known clinical drug interactions is not a guarantee of safety for every combination.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full official web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S04

A Randomized Controlled Trial of Riboflavin for Enhancement of Ultramarathon Recovery

[Source](https://digitalcommons.wku.edu/ijesab/vol9/iss5/45/)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: A probable companion with the same topic/investigators/intervention; randomized in its title differs from the alternating allocation in the original article. Exact full-abstract/identity comparison is incomplete; not counted as independent replication.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full conference abstract | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S05

The effects of a riboflavin supplementation on the nutritional status and performance of elite swimmers

[Source](https://www.sciencedirect.com/science/article/pii/S0271531784800056)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: A title concerning supplementation, nutritional status and performance in elite swimmers was identified, but age, endpoint, dose, design and effect/CI are unverified. Not counted as direct race-time evidence or independent replication.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full article text | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full article abstract | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S06

Riboflavin requirements and exercise adaptation in older women

[Source](https://pubmed.ncbi.nlm.nih.gov/1503064/)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: Concerns B2 requirements/exercise adaptation in older women, not an elite-swimming study. No decisive performance numbers were imported.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full article text | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full article abstract | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S07

The effect of exercise on the riboflavin status of adult men

[Source](https://www.cambridge.org/core/journals/british-journal-of-nutrition/article/effect-of-exercise-on-the-riboflavin-status-of-adult-men/FF6E34C4E3FA4F2F425DE633E9A6E5CF)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: The B2 status response to exercise is not direct evidence that B2 supplementation reduces race time. Original details/numerical values unverified.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full article text | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S08

Vitamin Supplementation in Sports: A Decade of Evidence-Based Insights

[Source](https://www.mdpi.com/2072-6643/18/2/213)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: A repeated citation of Hoffman is not another trial. Opening the original failed at the time, so it was recorded only at the search/reference-map level.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full review article | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S09

The Effect of a Single Bout of Exercise on Vitamin B2 Status Is Not Different between High- and Low-Fit Females

[Source](https://www.mdpi.com/2072-6643/13/11/4097)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: An exercise-challenge comparison that excluded use of B2-containing supplements. The registration number in the original Methods is NL7891. This is not a randomized oral B2 versus placebo race-time result.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full article text | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full article abstract | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full registration web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

### S10

Mitochondrial Coupling Efficiency, Respiration and Vitamin B2 status in Untrained and Endurance-trained Young Females

[Source](https://onderzoekmetmensen.nl/en/node/24874/pdf)

Materials actually read: Summary reading: the relevant bibliography/facts, Korean text and historical access records in the currently stored manuscript were read. External source materials were not read in this work.  
Findings recorded in the summary: OMON aliases remain at the discovery level; official details were inaccessible at the time. The adjacent TCTR entry was not assigned to the female study, and unverified registration material was not used to fill in methods.  
Record level: Report-specific results, direction, exclusion and discovery roles from the retained manuscript/extraction dated 2026-09-16. Not direct verification of original clinical materials in this work.  
Independent replication: The single direct UF2016 family, probable companion/repeated citations and different questions are distinguished. No new verification of external independent replication or unique participant counts was performed in the current work.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Full official registration web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |
| Full web document | Currently unverified (no external lookup in this work) | Access not attempted | Not read | Distinguish actual reading of the retained stored summary from the history of full/partial access and failures on 2026-09-16. The existence of a link/bibliographic record does not constitute current reading of the full original text. |

## Review criteria and methods

These are Chamgap's own application conditions, informed by the relevant sections of official methodological guidance. They do not constitute official assessment-tool certification or a new efficacy score.

- M03 [Cochrane Handbook 6.5 (2024), chapter updated 2023-08](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-03) · §3.1, §3.2.1–3.2.3, MECIR C5/C7/C13 · Clearly record the question, eligibility criteria and reasons for changes, and distinguish variations in the intervention. · Self-developed application reusing the directly reviewed methodological principles/versions and actual HTML copies/SHAs from 1585. No new reading of full methodological texts or official execution/certification in this work
- M04 [Cochrane Handbook 6.5.1 (2025), chapter updated 2025-03](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-04) · §4.5, §4.6.1–4.6.3, MECIR C42 · Distinguish and link studies and reports, and do not omit a study solely because its outcomes are not reported. · Self-developed application reusing the directly reviewed methodological principles/versions and actual HTML copies/SHAs from 1585. No new reading of full methodological texts or official execution/certification in this work
- M08 [Cochrane Handbook 6.5 (2024), chapter updated 2019-10](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-08) · §8.1–8.2.3, Table 8.2.a · Hold the outcome and effect of interest fixed, then assess randomization, deviations, missing data, measurement and selective reporting. · Self-developed application reusing the directly reviewed methodological principles/versions and actual HTML copies/SHAs from 1585. No new reading of full methodological texts or official execution/certification in this work
- M15 [Cochrane Handbook 6.5 (2024), chapter updated 2023-08](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-15) · §15.2.4, §15.3.1–15.3.3 · Distinguish effect estimates, intervals, importance and applicability. Sample size alone does not determine precision. · Self-developed application reusing the directly reviewed methodological principles/versions and actual HTML copies/SHAs from 1585. No new reading of full methodological texts or official execution/certification in this work
- M19 [Cochrane Handbook 6.5 (2024), chapter updated 2019-10](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-19) · §19.1.2, §19.2.1–19.2.4 · Questions about harms may require designs, populations and follow-up different from efficacy questions. Case reports do not establish incidence. · Self-developed application reusing the directly reviewed methodological principles/versions and actual HTML copies/SHAs from 1585. No new reading of full methodological texts or official execution/certification in this work

## What the statuses mean

- Met: The selected preservation/linkage condition can be explained. Distinct from confirmation of efficacy
- Partly verified: Some required information is linked, but specific gaps remain
- Unverified: Required information/conditions could not be verified within the currently permitted retained-material review scope
- Conflicting information: Unresolved conflict among original text/table/total records concerning the same medication-user count. No arbitrary correction
- Not applicable: Nonapplicability of an extension question absent from the original claim. Distinct from nonreading or absence of effect

## Change history

2026-11-07 · Public page/materials 1.0. Public-facing displays and data were prepared from the existing 40 items. No new clinical search, change of verification status or scoring was performed.

## Material-specific source addresses

- S01-M01 · Full article text on the web: Material URL unverified; external material unread.
- S01-M02 · Full article PDF: Material URL unverified; external material unread.
- S01-M03 · Full registration web document: Material URL unverified; external material unread.
- S01-M04 · Full protocol/SAP: Material URL unverified; external material unread.
- S01-M05 · Raw data: Material URL unverified; external material unread.
- S01-M06 · Correction materials: Material URL unverified; external material unread.
- S02-M01 · Full article text on the web: Material URL unverified; external material unread.
- S02-M02 · Full registration web document: Material URL unverified; external material unread.
- S03-M01 · Full official web document: Material URL unverified; external material unread.
- S04-M01 · Full conference abstract: Material URL unverified; external material unread.
- S04-M02 · Full web document: Material URL unverified; external material unread.
- S05-M01 · Full article text: Material URL unverified; external material unread.
- S05-M02 · Full article abstract: Material URL unverified; external material unread.
- S06-M01 · Full article text: Material URL unverified; external material unread.
- S06-M02 · Full article abstract: Material URL unverified; external material unread.
- S07-M01 · Full article text: Material URL unverified; external material unread.
- S07-M02 · Full web document: Material URL unverified; external material unread.
- S08-M01 · Full review article: Material URL unverified; external material unread.
- S09-M01 · Full article text: Material URL unverified; external material unread.
- S09-M02 · Full article abstract: Material URL unverified; external material unread.
- S09-M03 · Full registration web document: Material URL unverified; external material unread.
- S10-M01 · Full official registration web document: Material URL unverified; external material unread.
- S10-M02 · Full web document: Material URL unverified; external material unread.