Mass azithromycin distribution and antibiotic resistance in the gut and nasopharynx: a cluster-randomized trial.
Doan T, Yan D, Arzika AM, Abdou A, Maliki R, Aichatou B, Bello IM, Beidi D, Galo N, Harouna N, Karamba AM, Mahamadou S, Abarchi M, Ibrahim A, Zhong L, Chen C, Liu Y, Yu D, Abraham T, Cheng AS, Peterson B, Oldenburg CE, Porco TC, Arnold BF, Hinterwirth A, Lebas E, O'Brien KS, Lietman TM
- DOI
- 10.1038/s41591-026-04217-9
- Record issued
- 2026-08-10
- Engine
- 7.29.0
- Exported
- 2026-09-22
Prepared by Alpha1. This document is confidential: it is intended for the recipient it was shared with and must not be redistributed. The live record at alpha1science.com/verify/fea1448f-796e-44ac-8577-aa0b1a30fdb5 is authoritative.
How this rating was calculated
- IntegrityIntegrity concern ×2−1★
- ReportingBiological variables partially met−0.25★
- CitationsUnresolved reference−0.25★
- 01Methods and results do not match
The Methods state that if a pairwise comparison was not significant, subsequent comparison P values were not reported, yet the Results report P = 0.01 for the child-vs-infant comparison after the infant-vs-placebo comparison was non-significant (P = 0.66), contradicting the stated gatekeeping procedure.
“However, communities whose children were treated with azithromycin had macrolide resistance determinants 1.13-fold higher (95% CI: 1.02–1.23-fold, P = 0.01) than communities where only infants were treated with azithromycin”
Statistical analysisFind in source
This Kaimen Rigor review uses Kaimen Rigor reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.
The paper is a well-designed cluster-randomized trial with comprehensive reporting in most dimensions. Two minor gaps are present: imprecise p-value reporting (thresholds instead of exact values) and incomplete reporting of biological variables (weight, health status, demographics). An internal inconsistency in the hierarchical testing procedure is noted.
Three independent runs of the same model were synthesized; where they diverged, the evidence was weighed and the more conservative rating was adopted when the evidence supported it. The statistics verification covered only 5 tests (all consistent) and cannot speak to the correctness of the model-derived fold changes. The citation check flagged one Dryad DOI as not found in the registry, which should be verified.
Numerical inconsistencies
1 finding · worst lowValues that contradict each other or are impossible for the stated sample: recomputed p-values and test statistics, GRIM/GRIMMER checks on summary numbers, percentages against their own counts, totals against their parts, and estimates against their own confidence intervals.
- Internal contradictions in the reported numbersAssessed
Recomputed 5 tests: 5 consistent, 0 inconsistent; 5 via agent-written checks.
- CONSISTENTreported p < .010 · recomputed p = .002Reviewer 1Gut macrolide AMR fold change, child–azithromycin vs placebo: p from estimate + 95% CI on log scale.
“Macrolide resistance determinants were 1.16-fold as high (95% confidence interval (CI): 1.06–1.28-fold, P < 0.01)”
Taken as given: 1.16 is the reported fold change and 1.06–1.28 its two-sided 95% CI for the gut child-vs-placebo comparison; the CI is on the log (fold-change) scale and symmetrical; the p-value is two-tailed, derived from the estimate and its CI via the t distribution as stated in MethodsMethod: Two-tailed p recomputed from the log fold-change estimate and its 95% CI.How we recomputed it: pCI(1.16, 1.06, 1.28, 1) - CONSISTENTreported p = .010 · recomputed p = .010Reviewer 1Gut macrolide AMR fold change, child–azithromycin vs infant–azithromycin.
“followed by child–azithromycin compared with infant–azithromycin (1.13, 95% CI: 1.02–1.23; P = 0.01)”
Taken as given: 1.13 is the reported fold change and 1.02–1.23 its two-sided 95% CI for the gut child-vs-infant comparison; the CI is on the log scale; the p-value is two-tailed from the t distributionMethod: Two-tailed p recomputed from the log fold-change estimate and its 95% CI.How we recomputed it: pCI(1.13, 1.02, 1.23, 1) - CONSISTENTreported p = .660 · recomputed p = .446Reviewer 1Gut macrolide AMR fold change, infant–azithromycin vs placebo.
“compared to communities whose children were treated with placebo (1.04-fold, 95% CI: 0.94–1.15, P = 0.66)”
Taken as given: 1.04 is the reported fold change and 0.94–1.15 its two-sided 95% CI; the CI is on the log scale; the p-value is two-tailed from the t distributionMethod: Two-tailed p recomputed from the log fold-change estimate and its 95% CI.How we recomputed it: pCI(1.04, 0.94, 1.15, 1) - CONSISTENTreported p = .100 · recomputed p = .076Reviewer 1Nasopharyngeal macrolide AMR fold change, child–azithromycin vs placebo.
“the mean fold difference in macrolide AMR was 2.14-fold (95% CI: 0.93–4.99-fold, P = 0.10; Fig. )”
Taken as given: 2.14 is the reported fold change and 0.93–4.99 its two-sided 95% CI; the CI is on the log scale; the p-value is two-tailed from the t distributionMethod: Two-tailed p recomputed from the log fold-change estimate and its 95% CI.How we recomputed it: pCI(2.14, 0.93, 4.99, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewer 1Gut ermF fold change, child–azithromycin vs placebo.
“There was a 7.11-fold as high (95% CI: 3.89–12.91-fold, P < 0.001) in ermF in the child–azithromycin arm compared to the placebo arm”
Taken as given: 7.11 is the reported fold change and 3.89–12.91 its two-sided 95% CI; the CI is on the log scale; the p-value is two-tailed from the t distributionMethod: Two-tailed p recomputed from the log fold-change estimate and its 95% CI.How we recomputed it: pCI(7.11, 3.89, 12.91, 1)
- lowinternal contradictionThe stated hierarchical gatekeeping rule (if a pairwise comparison is non-significant, subsequent comparison P values are not reported) appears inconsistent with the reported results, since the infant-vs-placebo comparison (P = 0.66, priority 2) is non-significant yet the downstream child-vs-infant comparison (priority 3) is reported as P = 0.01.
If a pairwise comparison was not significant with an alpha of 0.05, then subsequent comparison P values were not reported. ... 1.13-fold higher (95% CI: 1.02–1.23-fold, P = 0.01) than communities where only infants were treated
Methodsreviewer’s wording
Overstated conclusions
1 finding · worst lowConclusions that reach past what the paper's own results support — including a significance claim that no longer holds when the statistic is recomputed, and efficacy resting on an unvalidated surrogate endpoint.
- Conclusions only partially backed by the presented evidenceAssessed
12 major claims checked against the paper's own evidence: all adequately supported.
- partialReviewer 2The selection of ermF with azithromycin MDA has the potential to facilitate the persistence of resistant organisms within hosts to cause disease.The increase in ermF is demonstrated, but the causal link to disease persistence is explicitly speculative and not tested.Evidence: ermF increased 7.11-fold (95% CI 3.89–12.91) child vs placebo; the paper states this is speculative.
“The selection of ermF with azithromycin MDA has the potential to facilitate the persistence of resistant organisms within hosts to cause disease.”
Discussion ¶4Find in source - supportedReviewers 1, 2, 3The trial met its primary AMR endpoint for the gut but not for the nasopharynx.The gut macrolide burden was significantly higher in the child arm vs placebo (P < 0.01), while the nasopharyngeal comparisons were all non-significant, matching the claim.Evidence: Primary endpoint results: gut fold change 1.16 (95% CI 1.06–1.28, P < 0.01) child vs placebo; nasopharynx 2.14 (95% CI 0.93–4.99, P = 0.10).
“The trial met its primary AMR endpoint for the gut but not for the nasopharynx.”
AbstractFind in source - supportedReviewer 1Selection of macrolide resistance determinants in the gut was driven by increases in ermF.The paper reports a 7.11-fold increase in gut ermF in the child arm vs placebo (P < 0.001) and states other macrolide ARGs were not different, directly supporting the claim.Evidence: ergF results: 7.11-fold (95% CI 3.89–12.91, P < 0.001) child vs placebo; other macrolide ARGs not different across arms.
ermF ... drove the changes seen at the antibiotic class level in the gut
Resultsreviewer’s wording - supportedReviewer 1Resistance to other classes of antibiotics was not observed in the gut or nasopharynx.No significant differences were found for any of the 20 nonmacrolide antibiotic classes evaluated in either niche, backing the claim.Evidence: Secondary endpoints: no notable differences for any nonmacrolide antibiotic classes across all comparisons in gut or nasopharynx.
“There were no notable differences for any nonmacrolide antibiotic classes across all comparisons”
ResultsFind in source - supportedReviewers 1, 2, 3Close monitoring of AMR should be an essential component of MDA for childhood mortality.The demonstrated gut macrolide resistance selection supports the recommendation for AMR surveillance as an essential programmatic component; the claim is framed as a policy recommendation consistent with the evidence.Evidence: Primary gut AMR endpoint and ermF selection results.
“Close monitoring of AMR should be an essential component of MDA for childhood mortality.”
AbstractFind in source - supportedReviewer 1This study confirmed the selection of macrolide resistance and associated microbiome effects when azithromycin MDA is administered in regions concurrently receiving SMC.Gut macrolide resistance selection and differential abundance of taxa (e.g., Bacteroides/Prevotella) were observed, consistent with the summary statement.Evidence: Primary gut AMR endpoint and differential abundance results (e.g., 5.70-fold Porphyromonas gingivalis in nasopharynx).
“This study confirmed the selection of macrolide resistance and associated microbiome effects when azithromycin MDA is administered in regions concurrently receiving SMC.”
DiscussionFind in source - supportedReviewer 2Four semiannual azithromycin treatments to children aged 1–59 months led to a selection of macrolide resistance determinants in the gut.The significant fold increase in macrolide resistance determinants supports the claim.Evidence: Primary endpoint: 1.16-fold (95% CI 1.06–1.28, P<0.01) child vs placebo; ermF 7.11-fold (95% CI 3.89–12.91, P<0.001).
“four semiannual azithromycin treatments to children aged 1–59 months led to a selection of macrolide resistance determinants in the gut”
Discussion ¶1Find in source - supportedReviewer 2A difference in resistance genes to other classes of antibiotics was not observed in either the gut or the nasopharynx.The reported secondary analyses show no significant differences for nonmacrolide classes.Evidence: Secondary endpoints: no notable differences for any nonmacrolide antibiotic classes across all comparisons (Fig. 4, Extended Data Tables 1 and 2).
“a difference in resistance genes to other classes of antibiotics was not observed in either the gut or the nasopharynx”
Discussion ¶1Find in source - supportedReviewer 3The gut macrolide AMR burden was highest in child–azithromycin compared with placebo.The fold change of 1.16 (95% CI 1.06-1.28, P<0.01) supports this claim.Evidence: Results, Primary endpoints, Figure 3a.
Macrolide resistance determinants were 1.16-fold as high (95% confidence interval (CI): 1.06–1.28-fold, P < 0.01) in communities whose children aged 1–59 months were treated with azithromycin compared to placebo.
Resultsreviewer’s wording - supportedReviewer 3There were no statistically significant differences in macrolide AMR selection in the nasopharynx between arms.All pairwise comparisons for nasopharyngeal macrolide resistance had CIs spanning 1 and p-values >0.05.Evidence: Results, Primary endpoints, Figure 3c.
In the nasopharynx, macrolide resistance was not statistically significant between treatment groups.
Resultsreviewer’s wording - supportedReviewer 3Azithromycin MDA did not result in co-selection of resistance to other antibiotic classes in the gut or nasopharynx.Secondary endpoints for non-macrolide antibiotic classes showed no notable differences, as shown in Figure 4 and Extended Data Tables.Evidence: Results, Secondary endpoints, Figure 4.
There were no notable differences for any nonmacrolide antibiotic classes across all comparisons.
Resultsreviewer’s wording - supportedReviewer 3The selection of macrolide resistance in the gut was predominantly driven by an increase in ermF.The paper reports a significant fold change for ermF (7.11-fold, 95% CI 3.89-12.91, P<0.001) in the child-azithromycin vs placebo comparison, and this is the only macrolide ARG that showed significant differences.Evidence: Results, Additional prespecified and exploratory outcomes, Figure 3b.
There was a 7.11-fold as high (95% CI: 3.89–12.91-fold, P < 0.001) in ermF in the child–azithromycin arm compared to the placebo arm.
Resultsreviewer’s wording
Data authenticity concerns
1 finding · worst mediumAn adversarial read for patterns associated with data that may not be genuine: results that look too clean, implausibly large effects, duplicated data or images, and methods that do not match the results reported.
- Methods and results do not matchAssessed
2 integrity concerns flagged (0 high).
- mediummethod result mismatchThe Methods state that if a pairwise comparison was not significant, subsequent comparison P values were not reported, yet the Results report P = 0.01 for the child-vs-infant comparison after the infant-vs-placebo comparison was non-significant (P = 0.66), contradicting the stated gatekeeping procedure.
“However, communities whose children were treated with azithromycin had macrolide resistance determinants 1.13-fold higher (95% CI: 1.02–1.23-fold, P = 0.01) than communities where only infants were treated with azithromycin”
Statistical analysisFind in source
Reporting gaps
1 finding · worst mediumRequired detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.
- Biological variables underreported (sex, age, strain)Assessed
Prior work (MORDOR, WHO guidelines, AVENIR mortality) is cited, and the rationale for studying AMR in the context of MDA is clearly explained. The limitations of prior research (e.g., smaller samples, different settings, concurrent SMC) are discussed in the Introduction and Discussion, showing how this study addresses them.
“That study, however, lacked a placebo arm for all 5 years, and AMR spillover among communities may have been a confounder and cannot be ruled out.”
“Repeated semiannual azithromycin mass drug administration (MDA) to children has been shown to reduce all-cause childhood mortality. However, antibiotic resistance is a major public health concern as the program is being implemented in sub-Saharan Africa.”
Randomization was response-adaptive at the community level using R. Blinding was double-blind with allocation concealment. A priori power analysis was provided for sample size. Inclusion and exclusion criteria are detailed for communities and individuals. Outlier handling is not explicitly discussed, but the primary analysis uses non-parametric Wilcoxon tests, which are robust to outliers, and samples were pooled at the village level, making outlier detection less applicable.
“The randomization sequence was generated by a biostatistician who was aware of the group assignments, using the R software (R Foundation for Statistical Computing).”
“all other study personnel, participants, laboratory personnel and all field workers responsible for administering and sample collection were masked to treatment allocation.”
“50 communities per group would provide 80% power to detect an approximately 2-fold difference in any pairwise comparison, assuming an alpha of 0.05 and a s.d. of the log base 2 read count of 1.88.”
“Aside from the UCSF biostatistician and data analyst responsible for the randomization, all other study personnel, participants, laboratory personnel and all field workers responsible for administering and sample collection were masked to treatment allocation.”
“we estimated that 50 communities per group would provide 80% power to detect an approximately 2-fold difference in any pairwise comparison, assuming an alpha of 0.05 and a s.d. of the log base 2 read count of 1.88.”
“The randomization unit was the grappe, which is the smallest administrative unit in Niger and approximately equivalent to a community.”
“All other study personnel, participants, laboratory personnel and all field workers responsible for administering and sample collection were masked to treatment allocation.”
“For load of genetic resistance determinants in nasopharyngeal swabs, we estimated that 50 communities per group would provide 80% power to detect an approximately 2-fold difference in any pairwise comparison, assuming an alpha of 0.05 and a s.d. of the log base 2 read count of 1.88.”
Sex is reported in Table 1 for each arm. Age is reported as median (IQR) and age categories. However, weight is not reported (only weight-based dosing is mentioned), and health status is not described beyond the inclusion criteria (weight ≥3000g, no known allergies). Demographics include age, sex, and geographic region but not race/ethnicity or comorbidities. The study is in Niger, so race/ethnicity may be less relevant, but it is not reported.
“As the trial was conducted at the community level and all analyses targeted community effects, both sexes were included in all analyses. Data were not reported disaggregated by sex.”
“Inclusion criteria for individual participants were: age 1–59 months and weighing at least 3,000 g”
The paper states ethical approval from the University of California, San Francisco Committee for Human Research and the Comité National Éthique pour la Recherche en Santé in Niger. Informed consent was obtained from guardians (written) and community leaders (verbal). The trial was conducted according to Good Clinical Practice guidelines and the Declaration of Helsinki.
“We obtained ethical approval for the study from the University of California, San Francisco Committee for Human Research and the Comité National Éthique pour la Recherche en Santé in Niger.”
“written informed consent was obtained from the guardians for sample collection, which was documented electronically.”
“This trial was conducted according to Good Clinical Practice guidelines and adhered to the principles of the Declaration of Helsinki.”
“We obtained ethical approval for the study from the University of California, San Francisco Committee for Human Research and the Comité National Éthique pour la Recherche en Santé in Niger.”
“written informed consent was obtained from the guardians for sample collection, which was documented electronically.”
“This trial was conducted according to Good Clinical Practice guidelines and adhered to the principles of the Declaration of Helsinki.”
“We obtained ethical approval for the study from the University of California, San Francisco Committee for Human Research and the Comité National Éthique pour la Recherche en Santé in Niger.”
“Additional verbal consent was obtained from community leaders, and written informed consent was obtained from the guardians for sample collection, which was documented electronically.”
“This trial was conducted according to Good Clinical Practice guidelines and adhered to the principles of the Declaration of Helsinki.”
Azithromycin is identified with manufacturer (Pfizer, donated), formulation, and dose (single 20 mg/kg oral suspension), satisfying the investigational-product standard. Sequencing reagents (NEBNext, KAPA HyperExplore) are named. Software is identified with versions (R 4.3.1, Kraken2 2.1.2, Bracken 2.5, DESeq2, Topconfects, vegan, BWA, MEGARes 3.0). Antibodies, cell lines, mycoplasma, and organisms criteria are not applicable to this human drug trial.
“Interventions involved a single 20 mg kg − 1 dose of oral azithromycin or placebo suspension offered to children aged 1–59 months at study months 0, 6, 12 and 18.”
“All statistical analyses were conducted in R (R Foundation for Statistical Computing, version 4.3.1).”
“Interventions involved a single 20 mg kg − 1 dose of oral azithromycin or placebo suspension offered to children aged 1–59 months at study months 0, 6, 12 and 18.”
“Sequencing libraries were performed using the NEBNext Ultra II DNA Library Prep Kit (New England Biolabs) per manufacturer’s recommendations.”
“All statistical analyses were conducted in R (R Foundation for Statistical Computing, version 4.3.1).”
“Sequencing libraries were performed using the NEBNext Ultra II DNA Library Prep Kit (New England Biolabs) per manufacturer’s recommendations.”
“All statistical analyses were conducted in R (R Foundation for Statistical Computing, version 4.3.1).”
Tests (Wilcoxon rank-sum, t-test, PERMANOVA, DESeq2) are named. Assumptions are not formally verified, but the primary analysis uses non-parametric tests. Effect sizes are reported with 95% CIs. Software is identified. Data presentation includes figures with error bars (95% CI) and per-group n. Some p-values are reported as thresholds (P < 0.01) rather than exact values, but the paper uses a hierarchical testing procedure that only reports significance. No arithmetic inconsistencies were detected in the demographic table.
“We reported the fold change for each comparison, defined as the difference in means of transformed and normalized reads, along with 95% CI derived from the two-sample t -test and two-sided Wilcoxon P values.”
“Macrolide resistance determinants were 1.16-fold as high (95% confidence interval (CI): 1.06–1.28-fold, P < 0.01)”
“communities whose children were treated with azithromycin had macrolide resistance determinants 1.13-fold higher (95% CI: 1.02–1.23-fold, P = 0.01)”
“pairwise Wilcoxon rank-sum tests were performed between each treatment arm.”
“There were no detectable differences among the communities whose infants were treated only with azithromycin compared to communities whose children were treated with placebo (1.04-fold, 95% CI: 0.94–1.15, P = 0.66).”
“Pairwise Wilcoxon rank-sum tests were performed between each treatment arm.”
The data availability statement provides a BioProject accession (PRJNA1337442) for sequencing reads and a Dryad DOI (10.5061/dryad.p8cz8wb48) for limited deidentified data. Managed access is available for additional requests. Custom analysis code is not mentioned, but the paper relies on well-documented standard software (R, Kraken2, etc.).
“Requests can be made to the PIs, Tom.Lietman@ucsf.edu and/or Kieran.Obrien@ucsf.edu, and will be addressed within 120 days.”
“Nonhost sequencing reads for all pooled samples are available at BioProject ID PRJNA1337442”
“Limited deidentified individual information is available via Dryad at 10.5061/dryad.p8cz8wb48”
“BioProject ID PRJNA1337442”
Methods include complete descriptions of study design, interventions, sample collection, and analysis. Trial registration (NCT04224987) is provided. A reporting summary is linked. All prespecified outcomes (primary and secondary) are reported. Limitations are extensively discussed in the Discussion section. Conclusions are measured and do not overstate the findings. Funding sources and a conflict of interest statement are included.
“ClinicalTrials.gov registration: NCT04224987 (http://clinicaltrials.gov/ct2/show/NCT04224987)”
“Further information on research design is available in the linked to this article.”
“ClinicalTrials.gov registration: NCT04224987”
“This study has multiple limitations inherent to simple, large, randomized controlled trials.”
“This work was funded by the Bill and Melinda Gates Foundation (INV-002454: T.M.L.), the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (R01AI175250: T.M.L. and K.S.O’B.) and an unrestricted grant from Research to Prevent Blindness.”
“ClinicalTrials.gov registration: NCT04224987”
Registered (1 ID: ClinicalTrials.gov). No reporting guideline cited.
Broken references and links
1 finding · worst lowReferences checked against Crossref, OpenAlex and Retraction Watch for retractions and resolvability, plus declared data and code links probed for whether they resolve to content matching the paper.
- References not resolvable to a published paperRecomputed
Checked 66 references by DOI: 1 verified — 1 DOI unresolved, 64 no DOI (shown, not verified).
- UNRESOLVED10.5061/dryad.p8cz8wb48Supplementary data for “Effect of mass azithromycin distribution on antibiotic resistance in the gut and nasopharynx: a cluster-randomized trial”Cited DOI does not resolve to any Crossref record.
- NO DOIAzithromycin to reduce childhood mortality in sub-Saharan AfricaNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOICause-specific mortality of children younger than 5 years in communities receiving biannual mass azithromycin treatment in Niger: verbal autopsy results from a cluster-randomised controlled trialNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOILonger-term assessment of azithromycin for reducing childhood mortality in AfricaNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMass azithromycin distribution to prevent child mortality in Burkina Faso: the CHAT randomized clinical trialNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIAzithromycin to reduce mortality—an adaptive cluster-randomized trialNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIWHO Guideline on Mass Drug Administration of Azithromycin to Children Under Five Years of Age to Promote Child SurvivalNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMass drug administration of azithromycin: an analysisNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIEthical challenges in mass drug administration for reducing childhood mortality: a qualitative studyNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIWHO Policy Recommendation: Seasonal Malaria Chemoprevention (SMC) for Plasmodium falciparum Malaria Control in Highly Seasonal Transmission Areas of the Sahel Sub-region in AfricaNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIModulation of fecal microbiota and reductions in fecal antibiotic resistance genes (ARGs) driven by Weissella-fermented feed in growing pigsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIDetection of tetQ and ermF antibiotic resistance genes in Prevotella and Porphyromonas isolates from clinical specimens and resident microbiota of humansNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIPrevalence of antimicrobial resistance genes in Bacteroides spp. and Prevotella spp. Dutch clinical isolatesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIInsights from Bacteroides species in children with type 1 diabetesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMass azithromycin distribution and community microbiome: a cluster-randomized trialNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMacrolide resistance in MORDOR I—a cluster-randomized trial in NigerNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIGut microbiome alteration in MORDOR I: a community-randomized trial of mass azithromycin distributionNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMacrolide and nonmacrolide resistance with mass azithromycin distributionNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIGut resistome of preschool children after prolonged mass azithromycin distribution: a cluster-randomized trialNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIProlonged mass azithromycin distributions and macrolide resistance determinants among preschool children in Niger: a sub-study of a cluster-randomized trial (MORDOR)No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIThe role of ‘spillover’ in antibiotic resistanceNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIHost range of the ermF rRNA methylase gene in bacteria of human and animal originNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIComplete nucleotide sequence and transcription of ermF, a macrolide-lincosamide-streptogramin B resistance determinant from Bacteroides fragilisNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIAzithromycin and erythromycin susceptibility and macrolide resistance genes in Prevotella from patients with periodontal diseaseNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIDiversity of antimicrobial resistance genes in Bacteroides and Parabacteroides strains isolated in GermanyNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIThe prevalence of antibiotic resistance genes in Bacteroides fragilis group strains isolated in different European countriesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIBacteroides: the good, the bad, and the nitty-grittyNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMechanisms of reduced susceptibility and genotypic prediction of antibiotic resistance in Prevotella isolated from cystic fibrosis (CF) and non-CF patientsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIDistribution, trends, and antimicrobial susceptibility of Bacteroides, Clostridium, Fusobacterium, and Prevotella species causing bacteremia in Japan during 2011–2020: a retrospective observational study based on national surveillance dataNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOICommensal Bacteroides species induce colitis in host-genotype-specific fashion in a mouse model of inflammatory bowel diseaseNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIDefining dysbiosis for a cluster of chronic diseasesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMobile elements carrying ermF and tetQ genes in gram-positive and gram-negative bacteriaNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIEscherichia coli as reservoir for macrolide resistance genesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMobile genetic elements in the genus Bacteroides, and their mechanism(s) of disseminationNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOICultivation of anaerobic bacteria: foundations and principlesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOITemporal trends in phenotypic macrolide and nonmacrolide resistance for Streptococcus pneumoniae nasopharyngeal samples up to 36 months after mass azithromycin administration in a cluster-randomized trial in NigerNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIThe antimicrobial susceptibility of Porphyromonas gingivalis: genetic repertoire, global phenotype, and review of the literatureNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIAntibiotic susceptibility and resistance genes in oral clinical isolates of Prevotella intermedia, Prevotella nigrescens, and Prevotella melaninogenicaNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIEmergence of antibiotic-resistant Porphyromonas gingivalis in United States periodontitis patientsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIPresence and antibiotic resistance of Porphyromonas gingivalis, Prevotella intermedia, and Prevotella nigrescens in childrenNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMacrolide resistance mechanisms in Gram-positive cocciNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIAntimicrobial resistance following azithromycin mass drug administration: potential surveillance strategies to assess public health impactNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOILimitations of current techniques in clinical antimicrobial resistance diagnosis: examples and future prospectsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOINeonatal antibiotic exposure impairs child growth during the first six years of life by perturbing intestinal microbial colonizationNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIAltering the intestinal microbiota during a critical developmental window has lasting metabolic consequencesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIAntibiotic use in children is associated with increased risk of asthmaNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIIntestinal microbiome is related to lifetime antibiotic use in Finnish pre-school childrenNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIHow colonization by microbiota in early life shapes the immune systemNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMetabolic and metagenomic outcomes from early-life pulsed antibiotic treatmentNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIThe invisible extinctionNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIHigh-throughput sequencing of pooled samples to determine community-level microbiome diversityNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIUnveiling microbial diversity: harnessing long-read sequencing technologyNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIDeep longitudinal lower respiratory tract microbiome profiling reveals genome-resolved functional and evolutionary dynamics in critical illnessNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOILongitudinal analyses of infants’ microbiome and metabolome reveal microbes and metabolites with seemingly coordinated dynamicsNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOICARD 2017: expansion and model-centric curation of the comprehensive antibiotic resistance databaseNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIIdentification of acquired antimicrobial resistance genesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIFLASH: a next-generation CRISPR diagnostic for multiplexed detection of antimicrobial resistance sequencesNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMEGARes: an antimicrobial resistance database for high throughput sequencingNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIMEGARes and AMR++, v3.0: an updated comprehensive database of antimicrobial resistance determinants and an improved software pipeline for classification using high-throughput sequencingNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIImpact of database choice and confidence score on the performance of taxonomic classification using Kraken 2No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIImproved metagenomic analysis with Kraken 2No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIBracken: estimating species abundance in metagenomics dataNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIvegan: Community ecology packageNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOIModerated estimation of fold change and dispersion for RNA-seq data with DESeq2No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
- NO DOITopconfects: a package for confident effect sizes in differential expression analysis provides a more biologically useful ranked gene listNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
2 data/code links checked; 2 live.
- dataBioProjectLIVEHTTP 200http://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA1337442Resolves to BioProject (data repository).
- datahttp://clinicaltrials.gov/ct2/show/NCT04224987LIVEHTTP 200Resolves, but the content could not be matched to the paper.
Copyediting
12 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 12 minor suggestions below.
12 copyedit issues flagged: mostly typo, grammar, clarity.
- MINORtypoResults (microbiome paragraph)“differential abundance analysis of the gut micriobiome identified notable taxa”→ Change 'micriobiome' to 'microbiome'.Spelling error.
- MINORtypoResults (ARGs paragraph)“e rmF , an ARG that encodes a methyltransferase”→ Remove the stray leading 'e' so the sentence begins 'ermF, an ARG that encodes...'.Stray character before the gene name.
- MINORgrammarResults, Primary endpoints“Macrolide resistance determinants were 1.16-fold as high (95% confidence interval (CI): 1.06–1.28-fold, P < 0.01)”→ Consider 'were 1.16-fold higher' for smoother phrasing.Awkward comparative construction.
- MINORgrammarResults, ARGs paragraph“There was a 7.11-fold as high (95% CI: 3.89–12.91-fold, P < 0.001) in ermF”→ Consider 'There was a 7.11-fold increase in ermF'.Incomplete comparative construction.
- MINORclarityAbstract“1: child–azithromycin) azithromycin to 1–59-month olds”→ Rephrase for clarity, e.g., 'azithromycin to children aged 1–59 months'.Awkward construction in the arm description.
- MINORtypoResults, paragraph 6“micriobiome”→ microbiomeTypo in 'differential abundance analysis of the gut micriobiome'
- MINORtypoResults, Additional prespecified and exploratory outcomes“e rmF”→ ermFStray 'e' before 'rmF'
- MINORconsistencyAbstract and Figure 3“1.04×”→ 1.04Inconsistent use of multiplication sign in fold-change notation; elsewhere '1.16' without ×.
- MINORgrammarMethods, Study setting“received between 2 to 4 monthly distributions”→ received from 2 to 4 monthly distributionsNon-standard preposition usage.
- MINORclarityMethods, Statistical methods“If a pairwise comparison was not significant with an alpha of 0.05, then subsequent comparison P values were not reported.”→ Consider clarifying that this rule applies only to the fixed-sequence primary analysis, since the Results report later comparisons.This statement conflicts with the reported P value for the child-vs-infant comparison after a non-significant earlier comparison.
- MINORconsistencyAbstract, primary endpoints“The co-primary outcomes included changes in gut and nasopharynx macrolide AMR.”→ Consider changing 'nasopharynx macrolide AMR' to 'nasopharyngeal macrolide AMR' for consistency with elsewhere in the paper.Minor stylistic issue.
- MINORclarityResults, Main study outcomes“The prespecified primary outcomes for this AMR study included separate analyses of the load of macrolide genetic resistance determinants in pooled community-level rectal and nasopharyngeal samples collected after four semiannual distributions.”→ Consider rephrasing to: 'The prespecified primary outcomes for this AMR study were the load of macrolide genetic resistance determinants in pooled community-level rectal and nasopharyngeal samples collected after four semiannual distributions.'Slightly redundant wording.
Post-publication audit: The published paper is robustly reported, but two issues warrant attention by readers: (1) an inconsistency between the stated hierarchical testing procedure and the reported Results (child-vs-infant P=0.01 reported after a non-significant comparison), and (2) imprecise p-value reporting for the primary endpoint. No erratum is required, but an independent re-analysis should verify exact p-values and assess the impact of the gatekeeping deviation.
- 1.HIGHrigorClarify the hierarchical testing procedure in the Methods and Results: either explain why the child-vs-infant comparison (P=0.01) is reported despite the non-significant infant-vs-placebo comparison (P=0.66) violating the stated gatekeeping rule, or explicitly label it as an exploratory analysis.This is an internal contradiction that undermines the transparency of the primary analysis and could affect the interpretation of significance.
- 2.HIGHdata codeVerify that the Dryad dataset (10.5061/dryad.p8cz8wb48) is accessible and correct, as the citation verification component flagged it as not found in the registry.A broken or incorrect data DOI would prevent readers from accessing the underlying deidentified data, violating the data availability commitment.
- 3.HIGHstatisticsReport exact p-values (e.g., P=0.004) instead of thresholds (P<0.01, P<0.001) for the primary gut macrolide AMR endpoint and the ermF comparison in the Results and Figure 3.Exact p-values allow readers to assess the precision of the evidence and are required for independent verification.
- 4.HIGHreportingAdd baseline weight and health status (e.g., presence of comorbidities) to Table 1, as currently only age and sex are tabulated and the weight eligibility threshold is the sole health indicator.Complete demographic and health status data are essential for assessing the comparability of study arms and generalizability.
- 5.HIGHdata codePublish the custom analysis code (R/DESeq2/Topconfects pipeline) in a version-controlled public repository (e.g., GitHub with a DOI) and cite it in the Data availability section.Reproducibility of the results requires access to the exact analysis scripts, not just the raw data and software names.
- 6.HIGHreportingExplicitly state whether the primary analysis was intention-to-treat or per-protocol, and describe how missing data (e.g., samples that failed QC) were handled in the Methods, Statistical methods section.The analysis population definition is a standard reporting requirement for RCTs and affects the interpretation of results.
- 7.MEDIUMreportingName the specific reporting guideline followed (e.g., the CONSORT extension for cluster-randomized trials) in the Methods or Reporting Summary, rather than referencing only a generic linked summary.Explicit adherence to a recognized guideline facilitates verification of completeness by reviewers and readers.
- 8.MEDIUMcopyeditCorrect the typo 'micriobiome' to 'microbiome' in the Results section (differential abundance analysis of the gut microbiome).Minimal typo correction for clarity.
- 9.MEDIUMcopyeditRemove the stray leading 'e' before 'ermF' in the Results section (stray character in 'e rmF').Typographical error that could cause confusion.
- 10.MEDIUMcopyeditRephrase awkward grammar: change 'were 1.16-fold as high' to 'were 1.16-fold higher' and 'There was a 7.11-fold as high' to 'There was a 7.11-fold increase' in the Results.Improves readability and clarity of reported effect sizes.
- 11.MEDIUMcopyeditIn the Abstract, change 'nasopharynx macrolide AMR' to 'nasopharyngeal macrolide AMR' for consistency with the Methods and Results.Consistency in terminology improves clarity.
- 12.LOWreportingAdd a brief statement about outlier handling in the Methods, even if the primary non-parametric tests are robust to outliers and samples were pooled.Addresses a minor reporting gap to satisfy the design criterion.
The star rating is the report’s one-glance summary. Every paper starts at 5★ and loses stars for the concrete problems the review finds — so a rating is never a vague average, it’s a running total you can read line by line under “How this rating was calculated.”
- Reporting — 8 dimensionseach dimension that fully fails−½★
- each dimension partially met−¼★
- Statistics · Integrity · Claimseach serious problem−1★
- each medium problem−½★
- Citationseach retracted or unverifiable reference−¼★
- Copyeditonly when the manuscript needs a full edit−½★
The rating never drops below 1★, and a demonstrable critical failure (an impossible statistic, a proven ethics violation) caps it at 1★ on its own — so the stars can never look healthy when the verdict is CRITICAL.
The rating draws on a panel of agents. Three independent Kaimen Rigor reviewers grade the eight dimensions below across several independent passes (the shown verdict is their majority vote — steadier than any single run), isolate the paper’s major claims and check its own evidence backs them, and flag integrity concerns. Alongside them, a citation agent resolves every reference against Crossref, OpenAlex, and Retraction Watch; a statistics agent recomputes reported tests; and rule-based checks verify that declared data/code links actually resolve. Full text is required — an abstract-only submission is not analyzed.
Graded against NIH, MDAR, ARRIVE 2.0, CONSORT, EQUATOR, and RRID guidelines. A dimension that doesn’t apply to the study type is skipped, never penalized.
This Kaimen Rigor review is model-assisted and is not a substitute for formal expert review. It complements human evaluation by surfacing potential methodological concerns — verify each finding against the source.