# Clinical NEC prevention practices drive different microbiome profiles and functional responses in the preterm intestine.

- **DOI:** [10.1038/s41467-023-36825-1](https://doi.org/10.1038/s41467-023-36825-1)
- **Rating:** Limited (2.7 out of 5)
- **Verdict:** Multi-kingdom profiling maps preterm gut differences across three NICU regimens, but center-confounded treatment bundles prevent attribution to individual preventive practices.
- **Framework version:** v5.13
- **Pipeline version:** v0.4
- **Evaluation date:** 2026-11-13
- **Evaluation page:** https://www.nabu.science/paper/10.1038%2Fs41467-023-36825-1

## Key claims

1. We demonstrate that preventive regimens have the highest impact on development and maturation of the gastrointestinal microbiome, enabling the establishment of a resilient microbial ecosystem that reduces pathogenic threats in at-risk preterm infants.
   - Support: Tentative
   - Type: Primary finding
   - Basis: center-confounded observational comparisons cannot establish regimen-driven resilience or reduced pathogenic threats
   - Location: Abstract
2. Engraftment of [Bifidobacterium longum subsp. infantis] NCDO 2203 is associated with a substantial reduction of microbiome-associated antibiotic resistance as compared to regimens using probiotic Lactobacillus rhamnosus LCR 35 or no supplementation.
   - Support: Tentative
   - Type: Secondary
   - Basis: bundled gentamicin exposure and unnormalized genomic hits limit interpretation of the observed resistance contrast
   - Location: Abstract
3. Crucially, the beneficial effects of Bifidobacterium longum subsp. infantis NCDO 2203 supplementation depends on simultaneous feeding with HMOs.
   - Support: Tentative
   - Type: Secondary
   - Basis: no comparison isolates simultaneous probiotic and HMO feeding to establish the asserted dependence
   - Location: Abstract
4. Although both probiotic species, B. infantis NCDO 2203 and L. acidophilus NCDO 1748, were administered in equal amounts in K, L. acidophilus NCDO 1748 was substantially lower in abundance than the predominant B. infantis NCDO 2203
   - Support: Supported
   - Type: Secondary
   - Basis: the within-Klagenfurt relative-abundance comparison directly documents unequal abundance despite equal administered doses
   - Location: Results, paragraph 8

## Strengths

- **Integrated profiling extends the cohort resource** Bacterial, archaeal, fungal and viral profiles are combined with functional metagenomics and NMR metabolomics. These measurements extend the cohort's earlier bacteriome analysis with genomic and metabolic descriptions.
  - Dimension: contribution
  - Location: Methods, Sample processing; Results, Figs. 1–5
- **Contamination controls and genomic strain confirmation** The workflow includes 16 negative controls, decontam filtering and multiple MAG quality checks. Sequence and genome comparisons support identification of the administered probiotic strains in stool.
  - Dimension: methodological rigour
  - Location: Methods, DNA extraction, sequencing, and metabolomics
- **Discussion explicitly identifies major interpretive limits** The Discussion states that changing treatment components cannot be disentangled and that iRep cannot establish colonization. It also acknowledges the small cohort, short observation period and low biomass of early samples.
  - Dimension: positioning
  - Location: Discussion, iRep interpretation and limitations

## Limitations

- **Treatment bundles remain inseparable from centers** Each NICU uses a different combination of probiotics, gentamicin, antifungals and feeding. The analysis cannot distinguish probiotic effects from these co-varying exposures, including the absence of gentamicin at the center with fewer resistance-gene hits.
  - Dimension: methodological rigour
  - Location: Table 1; Results, opening PERMANOVA analysis and resistance analysis; Discussion, limitations
- **Causal and synergy claims exceed comparisons** The Abstract asserts HMO-dependent probiotic benefit and regimen-driven ecosystem resilience. No group directly tests the probiotic-plus-predominantly-human-milk combination, and the observational center contrasts do not establish those causal effects.
  - Dimension: methodological rigour
  - Location: Abstract; Table 1; Discussion, HMO interpretation
- **Virulence statement reverses its reported totals** The Results describe fewer pathogenic factors in Klagenfurt while reporting 193 hits there, versus 64 in Graz and 173 in Leoben. Klagenfurt therefore has the highest aggregate count, although normalized per-infant burden cannot be determined from these totals.
  - Dimension: reporting
  - Location: Results, virulence-factor analysis; Fig. 5d

## Reliability checks

Caveats (4 of 4 checks).

- **Results check:** Noted
  - We demonstrate that preventive regimens have the highest impact on development and maturation of the gastrointestinal microbiome, enabling the establishment of a resilient microbial ecosystem that reduces pathogenic threats in at-risk preterm infants: "K infants had fewer pathogenic factors in their microbiomes than G and L infants" (Results, virulence factor analysis paragraph (Fig. 5d)) vs "G: 64 hits, K: 193 hits, L: 173 hits" (Results, virulence-factor analysis). The printed aggregate K count of 193 exceeds both G's 64 and L's 173, contradicting the stated lower-count direction without establishing normalized per-infant comparisons. → Not supported by the paper's own data
  - Minor reporting notes (2):
    - Descriptive: "at time point 7, a significantly decreased amount for K samples was detected as compared to G and L" (Results, human milk and HMO conversion) vs "2′FL, K:G q = 0.091, K:L q = 0.033; 3′FL, K:G q = 0.062, K:L q = 0.020" (Results, human milk and HMO conversion). The K:G q-values of 0.091 and 0.062 exceed 0.05, while the corresponding K:L values of 0.033 and 0.020 are below 0.05. → Conclusion holds
    - Descriptive: "Similarly, genes involved in metabolism of di- and oligosaccharides were reduced in K samples, but not significantly (Fig. 4a)." (Results, paragraph 16) vs "Di- and Oligosaccharides: K:G q = 0.057, K:L q < 0.001, G:L q = 0.949;" (Figure 4a caption). The blanket nonsignificance statement conflicts with the caption's significant K:L comparison, although K:G is not significant at 0.05. → Conclusion holds
- **Reporting and transparency:** Noted. Ethics approval, parental consent, trial registration, and sequencing-data and code resources are declared. The supplied text contains a malformed Gibberella q-value and an incorrect HMO figure reference.
  - Ethics approval: declared · Data availability: declared · Code availability: declared
- **Reference integrity:** Noted
  - 95 references in manuscript
  - 84 of 87 checkable references found in an index
  - 8 are books, websites or datasets — counted, but not index-checkable
  - Could not be verified automatically — queued for manual review:
    - Karthikeyan, G. & Bhat, B. V. The PiPS (Probiotics in Preterm Infants Study) Trial—controlling the confounding factor of crosscontamination unveils significant benefits. Indian Pediatr. 54, 162 (2017). (Crossref bibliographic top candidate ratio 0.47 — no plausible match)
    - Ramsey, M., Hartke, A. & Huycke, M. The physiology and metabolism of Enterococci. From Commensals to Leading Causes of Drug Resistant Infection. In: Enterococci: From Commensals to Leading Causes of Drug Resistant Infection. Boston: Massachusetts Eye and Ear Infirmary (2014). (Ambiguous: title ratio 0.68 via crossref_bibliographic sits between the divergence (0.30) and match (0.70) thresholds; too close to call either way, so it is recorded as unresolved rather than judged.)
    - Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 37, 852–857 (2019). (Ambiguous: title ratio 0.52 via crossref_bibliographic sits between the divergence (0.30) and match (0.70) thresholds; too close to call either way, so it is recorded as unresolved rather than judged.)
  - Citation details diverge for 1 reference:
    - #81 Wickham, H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. https://doi.org/10.18637/jss.v035.b01 (2016). (DOI resolves to: “<b>ggplot2:</b>Elegant Graphics for Data Analysis” — Valero-Mora, 2010 (Journal of Statistical Software). Differs from the DOI’s record on: year, authors) https://doi.org/10.18637/jss.v035.b01
  - References not verified automatically are resolved manually within 24 hours.

## Post-publication record

- **Status:** Clean
- **Summary:** No retraction or expression of concern found in Crossref (includes Retraction Watch).
- **Last checked:** 2026-11-13
