Bronchial and Systemic Relationships of Haemophilus in Chronic Obstructive Pulmonary Disease

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript investigates associations between COPD exacerbation phenotype (frequency and hospitalization) and the bronchial microbiome assessed from sputum using 16S rRNA gene sequencing, with an emphasis on Haemophilus taxa and systemic inflammatory proteins, particularly IL-8. The topic is clinically relevant, and the attempt to connect bronchial community structure with systemic biomarkers is potentially valuable. However, the current framing and analytical detail leave substantial uncertainty about whether the reported differences reflect true biological shifts or are confounded by disease severity, treatment exposures, and measurement limitations intrinsic to sputum 16S profiling. In its current form, I consider the work promising but not yet sufficiently supported for acceptance without msignificantrevision.

Strengths
The cohort is reasonably sized for an exploratory microbiome study in clinically stable COPD (n=72), and the manuscript focuses on a biologically plausible axis (Haemophilus–IL-8) while also describing broader community shifts (e.g., increased Pseudomonas in frequent exacerbators). Separating exacerbation frequency from hospitalization is conceptually appropriate because these capture distinct aspects of clinical course, and the reported correlation between them is appropriately noted. The observation that H. parainfluenzae—rather than H. influenzae—dominates among Haemophilus taxa in this dataset could be important if taxonomic assignment is robust.

Major scientific and methodological concerns
Study design and confounding by severity and treatment need tighter control. The results indicate that all patients with ≥2 exacerbations required at least one hospitalization, and the frequency of exacerbations and hospitalizations was inversely correlated with FEV1%. This raises a central interpretive concern: differences attributed to “frequent exacerbator” status may broadly reflect more advanced disease and its associated therapies, rather than an independent effect of exacerbation propensity. At minimum, the manuscript should present stratified baseline characteristics by group (infrequent vs frequent; hospitalized vs non-hospitalized), including FEV1%, GOLD stage if available, ICS/LABA/LAMA use, chronic macrolide use, recent systemic steroids, and antibiotic exposure windows. More importantly, the primary microbiome comparisons should be re-analyzed with multivariable models that adjust for lung function (FEV1%), smoking status/intensity, age, and key medication exposures. Without this, it isn't easy to interpret the observed reduction in Haemophilus as a phenotype-specific finding rather than confounding.

The definitions of “clinically stable” and “proximity to recent infections” must be explicit. The manuscript statthat es participants were “clinically stable,” but does nspecifyide the stability window (e.g., exacerbationsion and no antibiotics/systemic steroids for 4–8 weeks). Because both microbiome composition and systemic inflammatory proteins are susceptible to recent antimicrobial and steroid exposure, the stability definition is not a minor detail; it is foundational. Please specify the exact criteria, and report the time since the last exacerbation and the time since the last antibiotic/systemic steroid course. If these data are unavailable, this limitation should be elevated and interpretation tempered accordingly.

Sputum as a proxy for the “bronchial microbiome” requires more vigorous justification and more robust quality control reporting. Many taxa highlighted as “common” in COPD sputum (e.g., Prevotella, Porphyromonas, Leptotrichia, Veillonella) are also abundant oral commensals; sputum frequently contains variable oral contamination. The manuscript should clarify the sputum collection method (spontaneous vs induced), whether a mouth rinse was used, and whether any microscopic quality criteria (e.g., squamous epithelial cell thresholds) were applied. Ideally, negative controls and contaminant filtering should be described. Without these controls, the reported positive correlations between H. parainfluenzae and oral anaerobes could reflect sampling depth along the oral–lower airway gradient rather than ecological coupling in the lower airway.

Species-level inference for Haemophilus from 16S rRNA sequencing is a critical vulnerability that must be addressed directly. The central narrative contrasts H. parainfluenzae and H. influenzae, yet 16S regions commonly used in amplicon studies often lack the resolution to reliably differentiate closely related Haemophilus species. The manuscript must specify which 16S region was sequenced, which classifier/database was used, and provide evidence that the pipeline can distinguish H. parainfluenzae from H. influenzae in that region (e.g., by in silico validation against reference sequences or by reporting exact ASV sequences and their nearest-neighbor matches). If robust discrimination is not feasible, the claims should be reframed to the genus level or supported with orthogonal validation (e.g., species-specific qPCR, culture with MALDI-TOF, or metagenomic sequencing). As it stands, the biological interpretation hinges on a taxonomic resolution that may not be justified.

Statistical approach should account for compositional data properties and multiple testing. The manuscript reports differences in relative abundance and numerous correlations among taxa and with IL-8. Relative abundance data are compositional and zero-inflated; standard correlation and univariate testing can produce spurious associations. Please clarify the differential abundance method used (if it is simple rank-based tests on RAs, that is not sufficient). Consider applying compositionally aware approaches (e.g., ANCOM/ANCOM-BC, ALDEx2, or MaAsLin2 with appropriate transformations) and report effect sizes with confidence intervals. For the correlation network, a method designed for compositional data (e.g., SparCC-type or proportionality metrics) would be more defensible than raw Spearman correlations on relative abundances. Additionally, the manuscript appears to test many taxa and many proteins; explicit control of false discovery rate is necessary, and p-values should be reported alongside FDR-adjusted q-values.

The biomarker analysis needs a clearer assay description and a principled multiplicity strategy. Fifty-eight proteins were assessed, and 47 exceeded LOD. The manuscript focuses on IL-8, but it is not clear whether IL-8 was pre-specified as the primary biomarker (based on mechanism) or selected post hoc after screening. If IL-8 is a primary endpoint, state this and justify biologically. If not, then the association with IL-8 should be presented as exploratory and corrected for the number of biomarkers tested. Please also report assay platform, sample handling, transformation/normalization, and whether covariates (age, smoking, FEV1%, comorbidities) were included in protein–microbiome association models.

The interpretation should be tightened to avoid causal implications and to reconcile its directionality with prior literature. The manuscript notes that some studies reported increased Haemophilus in frequent exacerbators and higher mortality, while others found declines in common taxa with severity. Given that your finding is “Haemophilus under-represented in frequent exacerbators” with increased Pseudomonas, a plausible alternative explanation is that more severe COPD and/or repeated antibiotic exposure selects for Pseudomonas and depresses Haemophilus and other commensals. This interpretation needs to be explicitly considered, especially if hospitalization and antibiotic exposure are correlated. The discussion should separate three concepts: exacerbation propensity, disease severity, and treatment intensity, and avoid implying that frequent exacerbations “cause” the observed microbiome changes based on a cross-sectional design.

Minor comments and clarity issues
The manuscript would benefit from presenting Table 1 stratified by exacerbator group and hospitalization group, rather than only overall summary statistics. This is necessary for readers to judge comparability and confounding. The criteria for defining “common taxa” should be explicitly stated (prevalence threshold, abundance threshold, and rationale), and the complete list of taxa analyzed should be transparent (e.g., supplementary tables with prevalence and abundance distributions). The beta-diversity analysis should specify the distance metric, the ordination method, and the exact statistical test (PERMANOVA, permutation count, and whether dispersion was tested). It would also be helpful to report whether sequencing depth differed between groups and how rarefaction/normalization was handled.

From a reporting standpoint, it would strengthen the manuscript to include a concise data and code availability statement, including the deposition of raw reads (SRA/ENA) and metadata sufficient to reproduce the group definitions and covariate adjustment.

Author Response

Done.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Comments:

1. The Introduction is somewhat lengthy and could be more sharply focused. For example, the first sentence (Lines 93–100) lists multiple genera and encapsulates several ideas in one long sentence. Consider breaking complex sentences into shorter ones for clarity. Additionally, the background could be streamlined to emphasize why this study is needed.A more concise introduction focusing on the knowledge gap (e.g., inconsistent findings on Haemophilus in COPD exacerbations) would improve focus.

2. The last paragraph of the Introduction appropriately states the study’s aim and objectives. However, it does not explicitly mention the study’s hypothesis or summarize the main expected findings. For instance, if the authors hypothesized that H. parainfluenzae has an anti-inflammatory association in COPD (as suggested later in the Discussion), this should be stated clearly. Also, consider briefly foreshadowing the key result (e.g., “we found that frequent exacerbations are associated with a distinct microbiome profile and lower H. parainfluenzae relative abundance, correlating with higher IL-8 levels”).

3. The definitions of “exacerbation” and “hospital admission” are clearly given (Lines 501–509), which is excellent. It might help to explicitly state the timeframe (which you did: “previous year”) in the Results as well when first describing frequent vs. infrequent exacerbators, just to remind readers.

4. The results state that H. influenzae and H. parainfluenzae were identified at species level in many subjects (e.g., 43 and 52, respectively). However, standard 16S regions often have limited discriminatory power for closely related Haemophilus species. The methods should explicitly state: sequenced region, classifier/database, confidence thresholds, and whether ASVs were used. Otherwise, the paper should temper “species identification” language (or validate with species-specific qPCR/culture/metagenomics).

5. A negative correlation is reported between H. parainfluenzae relative abundance and blood IL-8 (and similar for several co-occurring genera).This cannot support “anti-inflammatory role” claims without adjusting for major confounders (FEV1%, smoking burden, inhaled corticosteroids, prior antibiotics, comorbidities, exacerbation history). In COPD, IL-8 is strongly linked to disease severity and systemic inflammation; the observed relationship may reflect a “health-associated” oral-commensal community rather than a direct protective effect.

6. Some referenced studies are over a decade old, which raises concern that more recent evidence might have been overlooked. For instance, references 13 (Larsson, 2008) and 14 (Bhowmik, 2000) are quite dated for discussions of inflammatory markers in COPD. Similarly, Ref. 24 (Chung, 2001) on cytokines is over 20 years old, and Ref. 26 (ATS spirometry standards, 1995) could be updated to a more recent standard or guideline.

7. Discussion – lack of strengths and limitations section. 

8. The manuscript currently has no clear future perspectives or forward-looking statements. It would strengthen the Discussion to include 1–2 sentences on what comes next.

9. The data are presented in multiple tables and figures clearly. However, the manuscript might benefit from a summary figure or schematic that integrates the key findings.

Author Response

See adjoined archive with the comments.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

In this manuscript, the authors investigate the effect of the existence of H. parainfluenzae along the course of COPD. I went through the manuscript and found the following points:

1.  The authors need to write the whole name of COPD in the title. I found the abbreviation difficult to follow (regarding readers away from the field).
2. The list of authors is incomplete. It ends with the word “and”.
3. The introduction is very small. The authors need to expand it in more detail.
4. Some phrases are difficult to understand. For example, the “in the first second” in the results. What does this stand for?
5. What are the criteria for evaluation of the exacerbations?
6. In Table 1, the heading of the table is not clear. What do N and 72 mean in the heading of this table? And the font is not identical!
7. What are the criteria for non-admitted patients?
8. Why are some taxa of Hemophilus not identified? And in Table 4, the relative frequency value is unclear. What does it mean?
9. The authors need to indicate the rationale for referring to IL-8 in their findings. I find that just jumping without indicating the reason is not helpful.
10. The discussion should be modified to include information about the current therapeutic situation in this kind of disease.
11. Regarding the ethical approval for this study, the authors need to provide the approval number.
12. How was the bacterial identity determined? And how were the proteins measured? The authors should indicate this in detail.

Author Response

See adjoined file for responses.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have responded adequately to all of my comments. 

Author Response

We thank the reviewer and we are glad to see that our responses has been approved by him/her.

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors for their answer.

The introduction is still very short and do not give a nice presentation for the topic. In addition, the problem with identifying the species level is not fully addressed.  

Author Response

1.- The introduction is still very short and do not give a nice presentation for the topic.

Following the suggestion of the reviewer, the introduction has been extended and significant changes in two paragraphs has been included in the revised version of the manuscript, in order to improve the focus of the research and its justification. The changes in the manuscript can be found in lines 83-101.

2.- The problem with the identification at species level is not fully addressed.

We agree with the reviewer that the identification of Haemophilus taxa at species level raises some difficulties that need a comment and a more detailed explanation. The limitations paragraph in the discussion section has been modified and extended to address this issue. Full atention to this problem is given in the revised version of the manuscript (linees 291-311) and two additional references has been included in the manuscript (numbers 28 and 29).

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors for addressing my comments. However, I still do not understand why “Hae-Mophilus” is written like this in the title. The Introduction is still very small and non-informative.

Author Response

1.- I still do not understand why Hae-Mophilus is written like this in the title.

Avoiding any misstyping, the title of the revised version of the manuscript is "Bronchial and systemic relationships of Haemophilus in chronic obstructive pulmonary disease".

2.- The introduction is still very small and non-informative.

Considering the comment of the reviewer, the introduction has been extended significantily in the revised version of the manuscript, with a new paragraph on the differences between Haemophilus influenzae and parainfluenzae in the clinicall setting of COPD, their different impacts on host inflammatory responses, and the specific characteristics of Hamophilus influenzae, that favor the formation of biofilms, allowing the persistence of the microorganism in the bronchial tree, as bronchial colonization. The new paragraph is located in lines 90-121 of the introduction section, and its has been expanded with 6 additonal references (16, 17, 20, 21, 22, 23). 

Round 4

Reviewer 3 Report

Comments and Suggestions for Authors

The authors addressed my comments. I would like to thank them.