# Long-term factorial supplementation with garlic extract plus steam-distilled garlic oil × Reduction of gastric-cancer incidence Verdict 3060 · Evidence Check Stored-material review date: 2026-10-06 Original evidence cut-off: 2026-09-08 Review-criteria version: 1.0 Public page/material preparation date: 2026-10-25 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: Residents at high risk of gastric cancer in Linqu, Shandong, China. The whole trial included 3,365 people; analysis denominators for the garlic groups remain unverified. - Ingredient, formulation and route: Oral combined supplementation with clove-derived garlic extract and steam-distilled garlic oil. Exact product, manufacturing and component specifications remain unverified. - Dose: Exact daily mg, administration frequency, actual intake and adherence remain unverified. - Comparator: Placebo for the garlic factor. H. pylori treatment and vitamins are separate randomized factors; detailed adjustment and interactions remain unverified. - Measured outcome: Cumulative gastric-cancer incidence during follow-up. Gastric-cancer mortality is a separate endpoint; the incidence OR and mortality HR are not interchangeable. - Duration: Supplementation during 1995~2003 for 7.3 years; follow-up through 2017 for 22.3 years. The full follow-up is not treated as the duration of intake. - Questions outside this scope: Gastric-cancer mortality is not substituted for incidence, and their event counts are not added.; Results of the H. pylori-treatment or vitamin factors are not attributed to garlic.; Raw garlic, powder, oil alone, aged garlic extract, black garlic, fermented garlic and isolated SAC are not treated as this combination.; Whole-trial counts and the OR are not used to calculate group denominators, absolute risks or NNT.; Exact dose, optimal dose, subgroups, persistence after stopping and safety at ordinary food intakes are separate questions. ## Checklist items ### N01 · Which garlic factor and which gastric-cancer outcome are compared? Verification status: Met Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original claim is reduced gastric-cancer incidence; the main endpoint is cumulative gastric-cancer incidence during follow-up. This concerns the retained summary comparing the combined garlic factor with its placebo. - Why unverified: The claim boundary is checked; this does not establish preventive efficacy or full-text verification. - Reason for missing information: The claim boundary is checked; this does not establish preventive efficacy or full-text verification. - Required materials and conditions: Combined garlic supplementation and its placebo, high-risk population, cumulative incidence, and supplementation and follow-up intervals - Materials to obtain first: The original protocol and report specifying the trial question, factor-specific comparisons and incidence definition - Further research: A product-specification-matched comparative study stating the incidence question separately - Conditions for verification: Specify cumulative gastric-cancer incidence for the combined garlic factor versus its placebo, with other factors and endpoints excluded. - Why this condition is needed: Needed to avoid combining incidence, mortality and other randomized factors from the same trial into a single efficacy claim. - Application exceptions: Mortality and H. pylori-treatment questions are retained as related background and do not replace this incidence comparison. - Remaining uncertainty: The claim boundary is checked; this does not establish preventive efficacy or full-text verification. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N02 · What was the trial combination of garlic extract and oil? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: This is oral combined supplementation with Allium sativum clove-derived extract and steam-distilled garlic oil. Detailed specifications and equivalence to other formulations are absent from the retained summary. - Why unverified: The combination is identified, but product, manufacturing and component specifications remain unverified. - Reason for missing information: The combination is identified, but product, manufacturing and component specifications remain unverified. - Required materials and conditions: Ingredients, specifications, manufacture, content and formulation of the extract and steam-distilled oil combination - Materials to obtain first: Manufacturing, component quality and supply material for the trial product - Further research: A direct formulation comparison with a fixed product-matched combination - Conditions for verification: Identify the actual trial specifications and combination of both components and record differences from other formulations. - Why this condition is needed: Needed to avoid transferring this result to garlic oil alone, raw garlic or AGE. - Application exceptions: Do not equate different extracted, aged or fermented formulations merely because they share a plant source. - Remaining uncertainty: The combination is identified, but product, manufacturing and component specifications remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N03 · Are the exact daily dose and actual intake of the combination linked? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original Korean/English wording that exact daily mg require checking in the original source is retained. Actual intake also remains unverified. - Why unverified: Exact mg, frequency and adherence information are absent from this retained scope. - Reason for missing information: Exact mg, frequency and adherence information are absent from this retained scope. - Required materials and conditions: Daily mg of each component, administration frequency, actual exposure and adherence - Materials to obtain first: The trial-product label, dosing schedule, and group adherence and discontinuation material - Further research: A product-matched dose study linking actual exposure - Conditions for verification: Distinguish nominal amounts of each component from actual intake records; do not infer missing quantities. - Why this condition is needed: Needed to avoid filling the phrase trial specifications with an exact dose or actual intake. - Application exceptions: Do not import mg, mL or SAC content from other garlic studies into this trial. - Remaining uncertainty: Exact mg, frequency and adherence information are absent from this retained scope. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N04 · What eligibility criteria defined the high-risk Linqu residents? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: This describes the whole-trial 3,365 residents at high risk of gastric cancer in Linqu, Shandong, China. Composition by age, lesions and infection remains unverified. - Why unverified: The region and high-risk population are retained, but specific eligibility and composition remain unverified. - Reason for missing information: The region and high-risk population are retained, but specific eligibility and composition remain unverified. - Required materials and conditions: Gastric-cancer risk, age, baseline lesions, infection status, inclusion/exclusion criteria and recruitment setting - Materials to obtain first: Baseline tables, infection and lesion composition, and group balance - Further research: An external replication matched to the product in another risk group or setting - Conditions for verification: Connect the actual composition of the selected high-risk population with the scope of the table. - Why this condition is needed: Needed to avoid transferring a local high-risk-population result to preventive food effects for everyone. - Application exceptions: Preserve mixed infection status, baseline lesions, composition and differences. - Remaining uncertainty: The region and high-risk population are retained, but specific eligibility and composition remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N05 · How are the randomized factorial structure and actual allocation connected? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original record describes long-term follow-up of a blinded randomized factorial trial with a 2×2×2/2×2 structure. Sequence generation, concealment and detailed baseline balance remain unverified. - Why unverified: The structure is stated, but the actual factor-allocation procedure has not been compared with the full text. - Reason for missing information: The structure is stated, but the actual factor-allocation procedure has not been compared with the full text. - Required materials and conditions: The 2×2×2/2×2 allocation structure, factor application by infection status, concealment and baseline balance - Materials to obtain first: The protocol, allocation flow, methods and baseline tables - Further research: A product-matched randomized factorial trial disclosing allocation details - Conditions for verification: Link the factor-specific allocation structure and garlic comparison of interest to their original definitions. - Why this condition is needed: Needed to avoid treating the randomized-factorial label as confirmation of all allocation procedures and detailed comparisons. - Application exceptions: Do not flatten the factorial design into an ordinary parallel 2-group comparison or label unknown allocation details as not applicable. - Remaining uncertainty: The structure is stated, but the actual factor-allocation procedure has not been compared with the full text. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N06 · What was the garlic-factor placebo and how were other factors handled? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The comparator is placebo for the garlic factor, with separate H. pylori-treatment and vitamin assignments. Detailed adjustment and interactions remain unverified. - Why unverified: The comparison direction is identified, but placebo details and the full factor analysis are unavailable. - Reason for missing information: The comparison direction is identified, but placebo details and the full factor analysis are unavailable. - Required materials and conditions: Placebo composition, blinding, and comparison/analysis handling of H. pylori and vitamin assignments - Materials to obtain first: Placebo specifications, factor layout, and analysis and interaction plans - Further research: A study prespecifying the main-factor analysis and interaction reporting - Conditions for verification: Distinguish the actual garlic-placebo comparison from the scope of adjustment, stratification and interaction handling for other assigned factors. - Why this condition is needed: Needed to avoid mixing other factors’ effects into the garlic comparison. - Application exceptions: Do not delete garlic results because other factors exist; identify the comparison of interest and its handling. - Remaining uncertainty: The comparison direction is identified, but placebo details and the full factor analysis are unavailable. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N07 · Are H. pylori-treatment results distinguished from garlic results? Verification status: Met Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The retained text identifies H. pylori treatment as a separate randomized factor and does not merge it with garlic’s incidence effect. - Why unverified: This checks the attribution boundary. It is not new verification of infection-treatment or interaction effects. - Reason for missing information: This checks the attribution boundary. It is not new verification of infection-treatment or interaction effects. - Required materials and conditions: Allocation and results for the infection-treatment factor and its relationship to the garlic comparison - Materials to obtain first: Infection status, treatment-factor results, factor-specific estimates and interaction tables - Further research: Prespecified effect-modification assessment by infection status and treatment - Conditions for verification: Record incidence comparisons for the H. pylori-treatment factor and the garlic factor separately. - Why this condition is needed: Needed to avoid attributing infection-treatment effects to garlic’s effect on gastric-cancer incidence. - Application exceptions: Do not add separate randomized-factor results; preserve the existence of related material. - Remaining uncertainty: This checks the attribution boundary. It is not new verification of infection-treatment or interaction effects. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N08 · Are vitamin-factor results also kept separate from garlic? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original Korean/English description is retained: vitamins were a separate randomized factor. Numerical interactions with the garlic comparison remain unverified. - Why unverified: Separate attribution is identified, but concurrent-exposure and interaction material is unavailable. - Reason for missing information: Separate attribution is identified, but concurrent-exposure and interaction material is unavailable. - Required materials and conditions: Vitamin assignment, dose, concurrent exposure, results and interactions - Materials to obtain first: Vitamin specifications, factor layout, adherence, and interaction plans/results - Further research: A study separating prespecified concurrent-factor interactions from standalone comparisons - Conditions for verification: Attribute vitamin-factor and garlic-factor results separately and record unverified concurrent exposure. - Why this condition is needed: Needed to avoid counting vitamin results from a multifactor trial as independent replication of garlic efficacy. - Application exceptions: Distinguish vitamin and garlic assignments without describing unverified interactions as absent. - Remaining uncertainty: Separate attribution is identified, but concurrent-exposure and interaction material is unavailable. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N09 · Are 7.3 years of supplementation distinguished from 22.3 years of follow-up? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Supplementation lasted 7.3 years during 1995~2003, with 22.3 years of follow-up through 2017. The whole follow-up is not converted into the supplementation period. - Why unverified: The two periods are identified, but post-discontinuation exposure and period-specific estimates remain unverified. - Reason for missing information: The two periods are identified, but post-discontinuation exposure and period-specific estimates remain unverified. - Required materials and conditions: End of supplementation during 1995~2003, follow-up through 2017, post-discontinuation exposure and event timing - Materials to obtain first: Supplementation, discontinuation and follow-up schedules, with period-specific exposure/analysis tables - Further research: Long-term follow-up prespecifying actual post-discontinuation exposure and effects over time - Conditions for verification: Link supplementation and follow-up separately; use only directly reported additional exposure and period-specific outcomes. - Why this condition is needed: Needed to avoid converting long follow-up into equally long intake or confirmation of a sustained effect. - Application exceptions: Retain post-discontinuation follow-up without inferring actual intake after stopping or period-specific effects. - Remaining uncertainty: The two periods are identified, but post-discontinuation exposure and period-specific estimates remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N10 · How were cumulative gastric-cancer incidence, diagnosis and detection collected? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The cumulative gastric-cancer-incidence endpoint is identified, but diagnostic/detection procedures and assessor-blinding details are absent from the retained summary. - Why unverified: A clinical-event name is available, but the full definition and detection conditions remain unverified. - Reason for missing information: A clinical-event name is available, but the full definition and detection conditions remain unverified. - Required materials and conditions: First-incidence definition, diagnostic criteria, detection/follow-up methods, assessor blinding and cumulative time point - Materials to obtain first: Diagnostic/event-adjudication criteria, endoscopy/pathology, detection schedules and blinding material - Further research: A long-term study disclosing standard event adjudication and group-specific detection conditions - Conditions for verification: Link diagnosis, detection, cumulative interval and group-specific measurement conditions to the selected incidence outcome. - Why this condition is needed: Needed to avoid treating an event name alone as confirmation of diagnostic accuracy or equivalent detection. - Application exceptions: Incidence and mortality are both clinical events but do not substitute for the same event definition or analysis. - Remaining uncertainty: A clinical-event name is available, but the full definition and detection conditions remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N11 · Which set do the 3,365 participants and 151 incidence events describe? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original record gives a whole-trial 3,365 participants and 151 incidence events. Allocation/analysis counts and event counts for each garlic group remain unverified. - Why unverified: Whole-trial totals are identified, but group totals for this comparison are unavailable. - Reason for missing information: Whole-trial totals are identified, but group totals for this comparison are unavailable. - Required materials and conditions: Whole-trial allocation, follow-up and incidence totals, and analysis denominators for the garlic groups - Materials to obtain first: Factor-specific allocation, completion, analysis and incidence flow tables - Further research: Reporting that discloses every group’s analysis denominator and event count - Conditions for verification: Keep supplied whole-trial counts at their original level; retain missing group denominators and events separately as unverified. - Why this condition is needed: Needed to avoid converting whole-trial counts into each group’s analysis count or an incidence denominator. - Application exceptions: Distinguish events from people and do not double-count the same participants across factors or reports. - Remaining uncertainty: Whole-trial totals are identified, but group totals for this comparison are unavailable. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N12 · What measure do the garlic-incidence OR and 95% CI represent? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original incidence OR0.81 and 95%CI0.57~1.13 are retained. The detailed incidence model and P value remain unverified in the current material. - Why unverified: The measure and interval are available, but the exact model/adjustment and original result table have not been compared. - Reason for missing information: The measure and interval are available, but the exact model/adjustment and original result table have not been compared. - Required materials and conditions: Incidence OR0.81, 95%CI0.57~1.13, model and comparison direction - Materials to obtain first: Original incidence result tables, model/adjustment definitions and group totals - Further research: A replication prespecifying the incidence estimand and its uncertainty - Conditions for verification: Link the incidence OR and interval to their endpoint, period and comparison, retaining unverified model details. - Why this condition is needed: Needed to avoid converting the OR into RR, HR, absolute risk or a mortality value. - Application exceptions: Preserve the original measure and interval without calculating new CI, P or effect values. - Remaining uncertainty: The measure and interval are available, but the exact model/adjustment and original result table have not been compared. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N13 · Can group incidence and an absolute risk difference be checked directly? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Incidence, denominators, absolute risk differences and NNT for each garlic group are absent from the retained summary. - Why unverified: Absolute risk is not newly calculated from the OR and whole-trial counts alone. - Reason for missing information: Absolute risk is not newly calculated from the OR and whole-trial counts alone. - Required materials and conditions: Group events, denominators and directly reported absolute differences for the same follow-up/analysis set - Materials to obtain first: Group events, follow-up denominators, and direct absolute differences and intervals - Further research: A long-term study reporting group absolute risks with uncertainty - Conditions for verification: Display an absolute preventive magnitude only when matching group denominators and directly reported absolute values are linked. - Why this condition is needed: Needed to avoid inventing prevention counts and absolute differences using the whole-trial 151 events or the OR. - Application exceptions: Do not back-calculate or extrapolate unverified absolute values, NNT or baseline risk. - Remaining uncertainty: Absolute risk is not newly calculated from the OR and whole-trial counts alone. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N14 · Does the incidence OR describe assignment or adherence effects? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The retained record provides a randomized factorial design and an OR, but ITT/PP, group analysis sets and adjustment variables remain unverified. - Why unverified: The allocation description is not connected to the analysis definition for the outcome of interest in this material. - Reason for missing information: The allocation description is not connected to the analysis definition for the outcome of interest in this material. - Required materials and conditions: ITT/PP, actual analysis sets, assignment/adherence comparison, adjustment variables and factor handling - Materials to obtain first: SAP, analysis sets/exclusions, adjustment methods and the actual incidence model - Further research: A study prespecifying assignment and adherence effects separately - Conditions for verification: Connect the effect of interest to the original analysis set and adjustment definitions and state the unverified scope. - Why this condition is needed: Needed to avoid automatically filling ITT analysis from randomization. - Application exceptions: Do not choose ITT, PP or a model when the original material does not specify the analysis. - Remaining uncertainty: The allocation description is not connected to the analysis definition for the outcome of interest in this material. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N15 · How were long-term losses and deaths handled in the incidence analysis? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Long follow-up and a separate 94 gastric-cancer deaths are mentioned, but incidence-analysis material on loss, censoring and competing-event handling is unavailable. - Why unverified: The retained scope does not establish how losses and deaths were handled. - Reason for missing information: The retained scope does not establish how losses and deaths were handled. - Required materials and conditions: Group losses/reasons, event detection, competing deaths/censoring and sensitivity analyses - Materials to obtain first: Group follow-up flow, death/censoring handling and sensitivity analyses - Further research: Follow-up prespecifying and disclosing long-term missingness and competing events - Conditions for verification: Connect actual handling of loss, censoring and competing deaths to its influence on analysis denominators and original results. - Why this condition is needed: Needed to avoid treating long follow-up and a total event count as confirmation of no missingness influence. - Application exceptions: Do not add deaths as incidence events or infer the competing-event method. - Remaining uncertainty: The retained scope does not establish how losses and deaths were handled. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N16 · Is citation identifier NCT00339768 linked to actual verification of prior registration? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The citation lists NCT00339768, but the structured registration field says unverified. The original registration was not read in this conversion. - Why unverified: An identifier clue exists, but prospective dates, trial identity and prior plans remain unverified. - Reason for missing information: An identifier clue exists, but prospective dates, trial identity and prior plans remain unverified. - Required materials and conditions: Registration identity, registration/change dates, recruitment start, registry versions, protocol/SAP - Materials to obtain first: Original NCT00339768 registration history, protocol and incidence/mortality SAP - Further research: A study disclosing change history and the prior incidence analysis - Conditions for verification: Check original registry versions, dates and trial identity, distinguishing plans actually compared from parts not compared. - Why this condition is needed: Needed to avoid promoting an identifier string into confirmation of prospective registration or a successful plan comparison. - Application exceptions: Preserve both the unverified structured registration field and citation identifier; lack of acquisition does not establish nonregistration. - Remaining uncertainty: An identifier clue exists, but prospective dates, trial identity and prior plans remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N17 · Were prespecified incidence/mortality endpoints and analysis hierarchy compared? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: This original verdict selects incidence as its main question and specifies mortality separately. The trial’s prespecified endpoint hierarchy/models were not compared in this conversion. - Why unverified: The manuscript boundary is identified, but actual agreement with prior plans remains unverified. - Reason for missing information: The manuscript boundary is identified, but actual agreement with prior plans remains unverified. - Required materials and conditions: Planned incidence/mortality endpoints and models, time points, changes and multiplicity rationale - Materials to obtain first: Endpoints, timing, multiplicity and change history in registration/protocol/SAP - Further research: A long-term confirmatory study prespecifying incidence and mortality separately - Conditions for verification: Compare original plans with actual incidence and mortality analyses separately, retaining unavailable parts. - Why this condition is needed: Needed to avoid substituting a favorable other endpoint for the selected incidence analysis. - Application exceptions: Record the manuscript’s main question separately from verification of the trial’s prespecified primary/secondary endpoints. - Remaining uncertainty: The manuscript boundary is identified, but actual agreement with prior plans remains unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N18 · Is the gastric-cancer-mortality HR substituted for the incidence OR? Verification status: Met Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Incidence OR0.81/95%CI0.57~1.13 and mortality HR0.66/0.43~1.00 are different endpoints in the same retained report. The 151 incidence events and 94 deaths are not added. - Why unverified: This checks endpoint separation. It does not newly confirm reduced mortality or prevention of incidence. - Reason for missing information: This checks endpoint separation. It does not newly confirm reduced mortality or prevention of incidence. - Required materials and conditions: Separate endpoints, models and interpretation for incidence 151 events/OR and mortality 94 deaths/HR - Materials to obtain first: Original incidence/mortality result tables, models and prior plans - Further research: A study defining and reporting each endpoint as a separate question in advance - Conditions for verification: Attribute incidence OR0.81(0.57~1.13) and mortality HR0.66(0.43~1.00) to their respective endpoints. - Why this condition is needed: Needed to avoid transferring a borderline signal in another endpoint into confirmed prevention of incidence. - Application exceptions: Retain mortality values as originally reported without merging their events, intervals or measure into incidence. - Remaining uncertainty: This checks endpoint separation. It does not newly confirm reduced mortality or prevention of incidence. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N19 · Was nonsignificant incidence converted into an effect of 0 or equivalence? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original record contains OR0.81 with 95%CI0.57~1.13 and nonsignificant/null wording. No new effect of 0, equivalence or important-difference criterion is created. - Why unverified: The original report is retained, but prior importance/equivalence hypotheses remain unverified. - Reason for missing information: The original report is retained, but prior importance/equivalence hypotheses remain unverified. - Required materials and conditions: Original OR/CI, reported wording and equivalence/important-difference criteria - Materials to obtain first: Prespecified important differences, equivalence hypotheses/interval criteria and original result tables - Further research: An incidence study prespecifying the effect magnitude of interest and precision - Conditions for verification: Preserve original observations and intervals without adding unreported equivalence, effect of 0 or importance criteria. - Why this condition is needed: Needed to retain original null/nonsignificant wording without newly concluding an exact effect of 0 or equivalence. - Application exceptions: Retain original D34/E0 and null wording as a historical verdict; this is not a new clinical grade. - Remaining uncertainty: The original report is retained, but prior importance/equivalence hypotheses remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N20 · Are a trial’s factors, languages and follow-up reports counted as independent replications? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The retained 1 evidence summary and 1 citation for Li WQ et al.BMJ2019/PMID31511230·PMC6737461·DOI10.1136/bmj.l5016 are linked as clues to the same report. No global unique-trial count is established. - Why unverified: Current family clues exist, but a complete search for other reports and independent replication was not performed in this conversion. - Reason for missing information: Current family clues exist, but a complete search for other reports and independent replication was not performed in this conversion. - Required materials and conditions: Shared-participant relationships among reports, factors and time points of the Shandong trial - Materials to obtain first: A study-family crosswalk connecting registration, reports and subsequent follow-up - Further research: A separate product-matched replication with independent funding and investigators - Conditions for verification: Identify currently linked reports/families; leave unestablished global unique-study and participant counts null. - Why this condition is needed: Needed to avoid adding the same participants’ results or Korean/English/combined summaries as independent studies or patients. - Application exceptions: Linking a family in the retained 1 report is separate from establishing a unique-trial count across the literature. - Remaining uncertainty: Current family clues exist, but a complete search for other reports and independent replication was not performed in this conversion. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N21 · Should inter-study heterogeneity be calculated within the one retained family? Verification status: Not applied Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Only one report is directly retained. Its factors, incidence/mortality and Korean/English descriptions are not independent units for inter-study heterogeneity. - Why unverified: Not applied only to inter-study comparison within this retained single-family question; this does not establish 0 studies or absence of replication. - Reason for missing information: Not applied only to inter-study comparison within this retained single-family question; this does not establish 0 studies or absence of replication. - Required materials and conditions: Actually distinct study families and condition-specific results - Materials to obtain first: Other distinct product-matched studies/reports and comparisons of their conditions - Further research: Review of independent-family result distributions for the same incidence question - Conditions for verification: Do not create inter-study heterogeneity values within the current single-report clue; distinguish unverified separate families. - Why this condition is needed: Needed to avoid turning factors, languages or endpoints within a retained report into multiple studies and inventing heterogeneity. - Application exceptions: Inter-study statistical comparison is not applied to this within-family question; independent replication across the literature remains unverified. - Remaining uncertainty: Not applied only to inter-study comparison within this retained single-family question; this does not establish 0 studies or absence of replication. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N22 · Does the record directly answer dose, infection-status and post-discontinuation questions? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Exact daily dose, direct dose comparisons, prespecified subgroups and period-specific effects after stopping remain unverified in the retained summary. - Why unverified: The main incidence comparison does not fill extension questions. - Reason for missing information: The main incidence comparison does not fill extension questions. - Required materials and conditions: Direct dose comparisons, prespecified subgroups/interactions and period-specific estimates - Materials to obtain first: Direct dose comparisons, infection/lesion plans and interactions, and period-specific estimates - Further research: A study separating prespecified effect modification, dose and post-discontinuation follow-up - Conditions for verification: Connect an extension only when its direct comparison, prior plan, denominator and interval are available. - Why this condition is needed: Needed to avoid extending the overall incidence comparison to optimal dose, specific groups or persistence after stopping. - Application exceptions: Do not require every extension study for the narrow incidence question; retain each extension separately as unverified. - Remaining uncertainty: The main incidence comparison does not fill extension questions. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N23 · Does an incidence question need a bridge from surrogates to events? Verification status: Not applied Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The selected outcome is cumulative gastric-cancer incidence, a clinical event. Mortality is separate. - Why unverified: This limited nonapplication concerns a surrogate-to-incidence bridge for incidence itself; it does not establish preventive efficacy. - Reason for missing information: This limited nonapplication concerns a surrogate-to-incidence bridge for incidence itself; it does not establish preventive efficacy. - Required materials and conditions: Distinction between selected cumulative gastric-cancer incidence and surrogate/mortality questions - Materials to obtain first: Direct material and original plans for separate mortality/other-event questions - Further research: A study independently verifying any clinical endpoint to which the claim is extended - Conditions for verification: Keep incidence as the directly selected endpoint without extending it to mortality or other events. - Why this condition is needed: Needed to avoid imposing a surrogate bridge on a question already concerning clinical incidence. - Application exceptions: A surrogate-to-incidence bridge is unnecessary for incidence itself, but incidence results do not guarantee reduced mortality. - Remaining uncertainty: This limited nonapplication concerns a surrogate-to-incidence bridge for incidence itself; it does not establish preventive efficacy. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N24 · Is public funding the same as checking product/author conflicts? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The retained record states NCI and Chinese public funding and original I2. Product provision, author employment/patents and disclosure details remain unverified. - Why unverified: Funding names alone do not newly establish overall independence or replication. - Reason for missing information: Funding names alone do not newly establish overall independence or replication. - Required materials and conditions: NCI and Chinese public funding, product provision, employment/patents and declared conflicts - Materials to obtain first: Original funding, product provision, author employment/patents and COI disclosures - Further research: A product-specification-matched replication with separate funding and investigators - Conditions for verification: Explain identified funding statements and unverified product/author relationships at their respective original locations. - Why this condition is needed: Needed to avoid converting public funding into proof of no conflicts or independent replication. - Application exceptions: Retain original I2 and independent-funding wording without newly certifying all disclosures or product provision. - Remaining uncertainty: Funding names alone do not newly establish overall independence or replication. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N25 · What are the actual reading levels for summaries, citations, full text and registration? Verification status: Met Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: This conversion reads the retained summary and an identical copy. Bibliographic identifiers/URLs exist, but external abstracts, full texts, registration and supplements were not accessed. - Why unverified: Reading-state separation is met; this does not mean external-material conditions are met. - Reason for missing information: Reading-state separation is met; this does not mean external-material conditions are met. - Required materials and conditions: Existence, public availability, access, acquisition, actual reading, dates and original locations/links for each material - Materials to obtain first: Actual acquisition/reading receipts and original locations for the paper, supplements, registration and SAP - Further research: Compare subsequently acquired material with each condition and link update history - Conditions for verification: Link the summary-reading label to current retained-summary citation locations; keep external full text and registration unread according to actual actions. - Why this condition is needed: Needed to avoid converting the presence of PMID, PMCID, DOI, NCT or URLs into actual reading of the material. - Application exceptions: Do not describe unattempted access as failure or nonpublic status; separate the retained summary from external material. - Remaining uncertainty: Reading-state separation is met; this does not mean external-material conditions are met. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N26 · Were specific Korean/English values and remaining generic wording preserved? Verification status: Met Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Specific Korean incidence/mortality values, the generic English evidence table and the historical 1 correction record are all retained unchanged. Generic English wording is not interpreted as successful full-text verification. - Why unverified: Original preservation is met; this does not claim all original English fields were corrected in this conversion. - Reason for missing information: Original preservation is met; this does not claim all original English fields were corrected in this conversion. - Required materials and conditions: Original Korean/English/combined-summary/evidence/citation values and the English-placeholder correction event - Materials to obtain first: Original text before/after the historical correction and report-specific English evidence fields - Further research: When new material is authorized, distinguish evidence updates from language corrections - Conditions for verification: Preserve every original literal, type and correction event and link specific observations to their actual original locations. - Why this condition is needed: Needed to avoid filling generic confirmation wording with missing specific values or verification actions. - Application exceptions: Preserve both the historical correction of 1 field and remaining generic English wording; do not newly revise the original. - Remaining uncertainty: Original preservation is met; this does not claim all original English fields were corrected in this conversion. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N27 · Are original review date, reading date and search scope distinguished? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The original verdict-review date is 2026-09-08. The retained summary has no separate evidence cut-off or material-specific access dates; no new search was performed in this conversion. - Why unverified: The original date is present, but historical search cut-off and material-specific access records remain unverified. - Reason for missing information: The original date is present, but historical search cut-off and material-specific access records remain unverified. - Required materials and conditions: Original review date 2026-09-08, actual reading date, inputs used, and search/correction/retraction-check scope - Materials to obtain first: Existing search records, report-specific access records and correction/retraction history - Further research: Compare the existing question with new evidence within an authorized follow-up scope - Conditions for verification: Link every absence/unverified statement to the retained scope and date; do not convert no search into 0 studies. - Why this condition is needed: Needed to avoid backdating the original review date as a current access date or search cut-off. - Application exceptions: No new clinical search was conducted, so do not claim absence of new papers or a complete current literature check. - Remaining uncertainty: The original date is present, but historical search cut-off and material-specific access records remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N28 · Are original D34/E0 and current item states distinguished? Verification status: Met Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Original D34, Caution, all axes and the Codex independent-cross-check wording are preserved as history. No full-text verification, new clinical score or probability is created in this conversion. - Why unverified: Separation of historical and current states is met; this is not independent recertification of the original verdict. - Reason for missing information: Separation of historical and current states is met; this is not independent recertification of the original verdict. - Required materials and conditions: Original grade, score, axes, historical Codex cross-check and the scope of current rows - Materials to obtain first: Versions, original text and decisions for the original verdict, axes and corrections - Further research: Link only affected conditions to a separate version when an evidence-update event arises - Conditions for verification: Preserve historical values separately; explain row states from their observations and reasons for uncertainty. - Why this condition is needed: Needed to avoid inheriting historical scores/axes as automatic fulfillment, probability or clinical certification. - Application exceptions: Preserve original D34/Caution/B-H-R1-I2-E0-B0-C0 and wording without new grading or calculation. - Remaining uncertainty: Separation of historical and current states is met; this is not independent recertification of the original verdict. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N29 · How were actual harms collected for this long-term combined exposure? Verification status: Partly verified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: The retained narrative cautions about bleeding, gastrointestinal discomfort and interactions from concentrated extract plus oil over years. Specific event counts, collection systems and discontinuation totals are unavailable. - Why unverified: Caution wording is retained, but direct harm-event definitions and totals remain unverified. - Reason for missing information: Caution wording is retained, but direct harm-event definitions and totals remain unverified. - Required materials and conditions: Definitions, systematic collection, severity and discontinuation for bleeding, gastrointestinal discomfort and interactions - Materials to obtain first: Original-trial harm-collection tables, severity, discontinuation and event definitions - Further research: A study systematically collecting harms for the same combined exposure - Conditions for verification: Distinguish actual event definitions, collection and discontinuation from the level of the original caution narrative. - Why this condition is needed: Needed to avoid turning caution wording into an event frequency or confirmed causal events. - Application exceptions: Do not erase incidence results because harms are unreported or equate the combination with ordinary-food-intake safety. - Remaining uncertainty: Caution wording is retained, but direct harm-event definitions and totals remain unverified. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N30 · Are harm timing, stopping, recovery and re-exposure known? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Only 7.3 years of supplementation and 22.3 years of follow-up are provided; individual harm-event timelines, recovery and re-exposure information are unavailable. - Why unverified: Trial periods cannot be converted into a harm-onset/recovery timeline. - Reason for missing information: Trial periods cannot be converted into a harm-onset/recovery timeline. - Required materials and conditions: Actual exposure start/stop, harm onset/recovery, re-exposure and observation intervals - Materials to obtain first: Exposure, onset, stopping and recovery dates for individual safety events - Further research: Long-term safety-event follow-up without requiring re-exposure - Conditions for verification: Distinguish periods from event timelines; do not fill unverified stopping, improvement or re-exposure. - Why this condition is needed: Needed to avoid treating trial periods alone as confirmation of temporal harm causality. - Application exceptions: Do not require personal re-exposure for this incidence-efficacy question or infer absent original event records. - Remaining uncertainty: Trial periods cannot be converted into a harm-onset/recovery timeline. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N31 · Were harm frequency versus placebo and its denominator checked directly? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Garlic-factor placebo is the comparator for incidence. Group harm events, safety-analysis denominators and observation time remain unverified. - Why unverified: Neither an incidence comparator nor the whole-trial N is used to invent harm frequency. - Reason for missing information: Neither an incidence comparator nor the whole-trial N is used to invent harm frequency. - Required materials and conditions: Group harm-person/event counts, exposure/observation denominators, placebo comparisons and intervals - Materials to obtain first: Group harm/discontinuation people and event counts, exposure time and direct comparison values - Further research: A product-matched comparative study disclosing safety denominators and events - Conditions for verification: Display comparisons/frequencies only from direct reports linking harm events, group denominators and observation time. - Why this condition is needed: Needed to avoid using the whole-trial 3,365 participants as every harm denominator. - Application exceptions: Caution wording is not a comparative-risk value; do not extrapolate background incidence. - Remaining uncertainty: Neither an incidence comparator nor the whole-trial N is used to invent harm frequency. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N32 · Could concurrent exposure, selection or omissions change harm attribution? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Other randomized factors are identified, but concomitant drugs, baseline harm risks, differential event omissions/detection and adjustment information are unavailable. - Why unverified: Do not establish harm causality or absence of harm without checking alternative explanations and reporting limits. - Reason for missing information: Do not establish harm causality or absence of harm without checking alternative explanations and reporting limits. - Required materials and conditions: Other randomized factors, concomitant drugs, baseline conditions, event reporting/detection and follow-up losses - Materials to obtain first: Concomitant exposure, baseline conditions, group reporting/omissions and factor-specific harm handling - Further research: A safety study jointly addressing exposure, alternative explanations and event omissions - Conditions for verification: Explain the verified scope of alternative/concurrent exposures, event collection and losses, with attribution limits. - Why this condition is needed: Needed to avoid treating missing concurrent-exposure and omission information as no alternative cause or no harm. - Application exceptions: Do not newly establish H. pylori or vitamin factors as harm causes; record only identified concurrent exposure. - Remaining uncertainty: Do not establish harm causality or absence of harm without checking alternative explanations and reporting limits. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ### N33 · What is known about independent harm records and long-term/special-population safety? Verification status: Unverified Reading scope: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. - Current findings: Direct safety families, group events/denominators and rare/special-population results remain unverified. Only existing caution wording is retained. - Why unverified: Long-term incidence follow-up and safety of other formulations or food intakes do not substitute for safety certification of this combination. - Reason for missing information: Long-term incidence follow-up and safety of other formulations or food intakes do not substitute for safety certification of this combination. - Required materials and conditions: Distinct safety events/study families, rare/long-term harms, vulnerable populations and concurrent-drug exposure - Materials to obtain first: Independent safety reports for this combination, long-term/rare events, special populations and concurrent-exposure material - Further research: Long-term/special-population safety follow-up linking product specifications and exposure - Conditions for verification: Keep the family, period and population scope of direct safety material separate from unverified replication/vulnerable-group information. - Why this condition is needed: Needed to avoid extending a long-follow-up incidence study to safety for every rare event and population. - Application exceptions: Requoting a summary is not an independent event; do not require every safety study for this narrow incidence question. - Remaining uncertainty: Long-term incidence follow-up and safety of other formulations or food intakes do not substitute for safety certification of this combination. - Referenced methods: Existing method principles distinguish population, formulation, dose, timing, endpoint, comparison, denominators, overlap, funding and harms in the retained report. Detailed conditions are Chamgap’s own application. - Related sources: S1 ## Evidence and sources ### S1 Li WQ et al. BMJ. 2019;366:l5016. PMID: 31511230. PMCID: PMC6737461. DOI: 10.1136/bmj.l5016. NCT00339768. [PubMed](https://pubmed.ncbi.nlm.nih.gov/31511230/) Materials actually read: The retained Korean/English summaries and case-specific condition/application records were read. External PubMed, DOI, full-text and registration material remain unread, with access not attempted. Findings recorded in the summary: The retained incidence result is OR0.81, 95%CI0.57~1.13. The separate mortality result is HR0.66 (0.43~1.00). The whole-trial 3,365 participants, 151 gastric-cancer incidence events and 94 gastric-cancer deaths are preserved at their respective levels. Record level: The incidence OR and mortality HR are separate reported values. The nonsignificant incidence description is not converted into an effect of 0 or equivalence. Group denominators and events, exact models, ITT/PP and the prespecified endpoint hierarchy remain unverified. Independent replication: The retained evidence and citation link to the BMJ2019 report by Li WQ and colleagues. Factors, incidence/mortality, time points and Korean/English summaries are not double-counted as independent studies or participants. Global unique-trial counts, participant totals and independent replication remain unverified. | Material | Public availability | Access status | Actually read | Notes | |---|---|---|---|---| | Retained verdict summary | Retained verdict summary read | Retained copy acquired; summary read | Stored existing review summary read | Relevant content in the retained Korean/English summary, evidence and citation was read. This does not establish acquisition or reading of the original clinical paper. BMJ2019 by Li WQ and colleagues is a report linked in the retained summary. External abstracts, full texts, registration, protocols and analysis plans were not read in this conversion, and independent clinical validation was not performed. The material check date is the original date retained in the summary and condition records. The original verdict-review date does not establish a search cut-off or historical access dates for individual materials; external access was not attempted in this conversion. | | PubMed bibliographic/abstract gateway | Bibliographic link retained; external material unread | External access not attempted; not acquired | Not read | This is the PubMed link from the retained citation. The current response, bibliographic identity, abstract content and public access were not checked. BMJ2019 by Li WQ and colleagues is a report linked in the retained summary. External abstracts, full texts, registration, protocols and analysis plans were not read in this conversion, and independent clinical validation was not performed. The material check date is the original date retained in the summary and condition records. The original verdict-review date does not establish a search cut-off or historical access dates for individual materials; external access was not attempted in this conversion. | | PMC full-text gateway | Full-text gateway; external material unread | External access not attempted; not acquired | Not read | This gateway was constructed from retained citation identifier PMC6737461. The current response, full-text and supplement identity, and reading remain unverified. BMJ2019 by Li WQ and colleagues is a report linked in the retained summary. External abstracts, full texts, registration, protocols and analysis plans were not read in this conversion, and independent clinical validation was not performed. The material check date is the original date retained in the summary and condition records. The original verdict-review date does not establish a search cut-off or historical access dates for individual materials; external access was not attempted in this conversion. | | Trial registration identifier gateway | Registration identifier gateway; original registration unread | External access not attempted; not acquired | Not read | This gateway was constructed from retained citation identifier NCT00339768. The original registration field remains unverified, as do registration dates, versions, trial identity and prior plans. BMJ2019 by Li WQ and colleagues is a report linked in the retained summary. External abstracts, full texts, registration, protocols and analysis plans were not read in this conversion, and independent clinical validation was not performed. The material check date is the original date retained in the summary and condition records. The original verdict-review date does not establish a search cut-off or historical access dates for individual materials; external access was not attempted in this conversion. | | Protocol and statistical analysis plan | Existence, public availability and location unverified; unread | External access not attempted; not acquired | Not read | Existence, public availability, location, prior endpoints, analysis models and change history for the protocol and SAP remain unverified. Access was not attempted; this does not establish absence of registration or nonpublic status. BMJ2019 by Li WQ and colleagues is a report linked in the retained summary. External abstracts, full texts, registration, protocols and analysis plans were not read in this conversion, and independent clinical validation was not performed. The material check date is the original date retained in the summary and condition records. The original verdict-review date does not establish a search cut-off or historical access dates for individual materials; external access was not attempted in this conversion. | ## 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 conditions and reasons for changes, and distinguish variations in interventions. · Retained method-application records were reused. External originals were not reread and method content was not reassessed. - 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 studies from reports and link them; do not omit a study solely because results are unreported. · Retained method-application records were reused. External originals were not reread and method content was not reassessed. - 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 · Define the outcome and effect of interest, then examine randomization, deviations, missing data, measurement and selective reporting. · Retained method-application records were reused. External originals were not reread and method content was not reassessed. - 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. · Retained method-application records were reused. External originals were not reread and method content was not reassessed. - 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 · Harm questions may require different designs, populations and follow-up from efficacy questions. Case reports do not establish incidence rates. · Retained method-application records were reused. External originals were not reread and method content was not reassessed. ## What the statuses mean - Partly verified: Some required facts are known, but the stated conditions involving analyses, populations, formulation or denominators remain unverified. - Unverified: Direct information is unverified in the retained record. This does not establish an effect of 0, no studies or absence of harm. - Review incomplete: Reading of the material or review of a condition is incomplete. This does not establish absence of studies or efficacy. - Not reported: The needed information was not found in material actually read. This is distinguished from unread material. - Inaccessible: Access was actually attempted but the material was not acquired. This is distinguished from access not attempted. - Insufficient evidence matching the conditions: Evidence directly matching the defined question and conditions is insufficient. This does not establish absence of all research. - Not applicable: A specific condition does not apply to the question. This does not substitute for unread or unverified material. - Conflicting data: This covers discrepancies in material about the same question, denominator and timing. Differences between separate questions are not automatically treated as conflicts. - Met: The retained identity, observation or scope required by this item was checked within stated limits. This does not establish preventive efficacy or full-text certification. - Not applied: An exception whose application premise is not met within the retained scope and selected question. It is limited to inter-study heterogeneity within the single retained report in N21, and a surrogate-to-incidence bridge for the clinical incidence endpoint in N23. It does not substitute for unread material, absence of evidence or confirmed efficacy, and requires review if its premise changes. ## Change history 2026-10-25 · Public page/materials 1.0. The public view and materials were prepared from the existing 33 items. No new clinical search, verification-state change or scoring was performed.