Review Reports - Microbiome Diversity in Pancreatic Surgery: Associations with Preoperative Stenting and Postoperative Outcomes

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Manuscript entitled “Microbiome Diversity in Pancreatic Surgery: Associations with Preoperative Stenting and Postoperative Outcomes”, by Laura Oelschlägel, Johannes Klose, Markus Glaß, Stefan Moritz, Bogusz Trojanowicz, Jörg Kleeff, and Artur Rebelo. This manuscript addresses a clinically meaningful and timely question at the intersection of pancreatic surgery and translational microbiome research. The multi-site sampling strategy is particularly interesting, and the distinction between DNA-based and RNA-based communities has the potential to add real value to the field.

Comments:

1. The patient and sample flow needs to be presented much more transparently. The manuscript states that 72 patients were enrolled and 412 samples collected, but only 58 patients and 224 samples were ultimately analyzed. This is a substantial reduction, and it raises the possibility of selection bias at both the patient and sample level. I would strongly recommend adding a clear flow diagram showing how many patients and samples were excluded at each stage and for what reason, such as failed extraction, low biomass, insufficient read depth, contamination concerns, or missing clinical metadata.

2. The definitions of the clinical endpoints require more precision. The paper links microbial patterns to postoperative morbidity, including severe complications and infectious complications, but the main text does not define these outcomes clearly enough. It should be stated explicitly over what postoperative time window complications were recorded, which events were counted as infectious complications, and whether standardized criteria such as Clavien–Dindo, CDC definitions, or ISGPS criteria were used where applicable. A dedicated table summarizing the number and type of postoperative complications would make the clinical context much easier to follow.

3. The regression analyses currently appear underadjusted for a surgical cohort of this complexity. Although the models include sex, age, BMI, diabetes, and smoking, important perioperative confounders seem to be missing. In pancreatic surgery, postoperative outcomes are strongly influenced by factors such as procedure type, diagnosis, operative complexity, blood loss, pancreatic texture and duct size, neoadjuvant therapy, perioperative antibiotic exposure, preoperative cholangitis, and biliary instrumentation. Since preoperative stenting is central to the manuscript’s hypothesis, it is particularly important that its role be handled more rigorously in multivariable models. At the very least, I would encourage the authors to perform sensitivity analyses that include additional clinical covariates where available or to restrict selected analyses to more homogeneous subgroups, such as pancreaticoduodenectomy patients only or stented versus non-stented patients.

4. The reported association between higher Enterococcus faecalis abundance and lower odds of severe complications is intriguing, but in its current form it is difficult to interpret and may be overemphasized. Biologically, this finding is somewhat counterintuitive, and the effect estimate itself is not easy to understand unless the authors specify the exact unit of abundance increase used in the model. It is also unclear whether this association would remain significant after correction for multiple testing, given the number of taxa examined. I would therefore recommend a more cautious presentation, including clarification of the scaling, graphical display of abundance distributions by outcome group, and sensitivity analyses using more robust compositional-data approaches or models that handle sparsity and zero inflation more appropriately. Adjustment for stenting and antibiotic exposure would be particularly important here because Enterococcus is closely linked to both.

5. Since preoperative biliary stenting is one of the central explanatory variables in the manuscript, it needs to be characterized in much more detail. At present, “stenting” is treated as a relatively simple binary feature, but its microbiological consequences may differ considerably depending on whether the stent was plastic or metal, how long it had been in place, how many ERCP procedures were performed, whether exchanges were necessary, and whether cholangitis occurred before surgery.

6. The manuscript would also benefit from language polishing. Several sections contain long and complex sentences, and in places the writing feels denser than necessary. Because the paper sits at the interface of surgery, microbiome science, and statistical analysis, improved clarity would make it more accessible to a broader readership.

7. It would be helpful if the authors stated more directly whether the study is intended as exploratory and hypothesis-generating. Given the single-center setting, the modest sample size, and the complexity of the data structure, a clear statement about the exploratory nature of the work would help frame the findings appropriately.

Author Response

Dear Editor and Reviewers,

We would like to sincerely thank you for your careful evaluation of our manuscript and for your constructive and insightful comments. We greatly appreciate the time and effort invested in reviewing our work. The suggestions provided were extremely helpful and allowed us to improve the clarity, methodological transparency, and interpretation of our results substantially.

In the revised version of the manuscript, we have carefully addressed all comments raised by the reviewers. We clarified methodological aspects, refined the description of clinical endpoints, adjusted the interpretation of microbiome–outcome associations, and strengthened the discussion of limitations and the exploratory nature of the study. In addition, several sections of the manuscript were revised for improved readability and structure.

Below, we provide a detailed point-by-point response to each comment. All changes made in the manuscript are indicated accordingly.

Comments and Suggestions for Authors

Comments:

The patient and sample flow needs to be presented much more transparently. The manuscript states that 72 patients were enrolled and 412 samples collected, but only 58 patients and 224 samples were ultimately analyzed. This is a substantial reduction, and it raises the possibility of selection bias at both the patient and sample level. I would strongly recommend adding a clear flow diagram showing how many patients and samples were excluded at each stage and for what reason, such as failed extraction, low biomass, insufficient read depth, contamination concerns, or missing clinical metadata.

Answer: Thank you for your comment. To better illustrate the workflow, we have created a flowchart intended to clarify the reduction of incoming samples for the reader.

We have added Figure 1 to the Methods section.

The definitions of the clinical endpoints require more precision. The paper links microbial patterns to postoperative morbidity, including severe complications and infectious complications, but the main text does not define these outcomes clearly enough. It should be stated explicitly over what postoperative time window complications were recorded, which events were counted as infectious complications, and whether standardized criteria such as Clavien–Dindo, CDC definitions, or ISGPS criteria were used where applicable. A dedicated table summarizing the number and type of postoperative complications would make the clinical context much easier to follow.

Answer: Thank you for your comment. We specified the inclusion criteria for the binary logistic regression in greater detail in order to better justify the exclusion of eight patients. Furthermore, the two outcomes were defined more precisely, and the time period of data collection was also indicated. To improve clarity, we created a table showing the distribution of all infectious complications as well as a table describing the distribution of the Clavien–Dindo classifications within this study population.

In section “Binary Logistic regression” we therefore added the following: Models were adjusted for potential confounders (sex, age, BMI, diabetes, smoking status) and included 50 patients, as some patient information was partially missing in the retrospective data collection. Postoperative complications were classified according to the Clavien–Dindo system and categorized as severe (Clavien–Dindo ≥III) or non-severe. Infectious complications were analyzed collectively (yes/no), with no dis-tinction between local and systemic infections. Table 1 summarizes the infectious complications observed in this cohort, classified according to CDC criteria, whereas Table 2 shows the distribution of the Clavien–Dindo classifications. Only the period up to hospital discharge was considered. Given the exploratory nature of the analysis, no formal correction for multiple testing was applied.

The regression analyses currently appear underadjusted for a surgical cohort of this complexity. Although the models include sex, age, BMI, diabetes, and smoking, important perioperative confounders seem to be missing. In pancreatic surgery, postoperative outcomes are strongly influenced by factors such as procedure type, diagnosis, operative complexity, blood loss, pancreatic texture and duct size, neoadjuvant therapy, perioperative antibiotic exposure, preoperative cholangitis, and biliary instrumentation. Since preoperative stenting is central to the manuscript’s hypothesis, it is particularly important that its role be handled more rigorously in multivariable models. At the very least, I would encourage the authors to perform sensitivity analyses that include additional clinical covariates where available or to restrict selected analyses to more homogeneous subgroups, such as pancreaticoduodenectomy patients only or stented versus non-stented patients.

Answer: Thank you for this important comment. We agree that postoperative outcomes in pancreatic surgery are influenced by a variety of perioperative factors, including procedure type, operative complexity, pancreatic texture, duct size, neoadjuvant therapy, and biliary instrumentation. These variables are indeed clinically relevant and may act as potential confounders. However, the total study cohort comprised only 50 patients. Given this limited sample size, the number of covariates that can be included in multivariable logistic regression models is restricted in order to avoid model overfitting and unstable estimates. Based on statistical guidance and commonly accepted recommendations for regression modeling in small datasets, the number of variables in the model was therefore limited to a maximum of five covariates. For this reason, confounders were selected a priori based on general clinical relevance for postoperative outcomes rather than on statistical testing within the cohort. We included age, sex, BMI, diabetes, and smoking status as broadly established risk factors for postoperative morbidity. While we acknowledge that additional pancreas surgery–specific variables may also influence outcomes, incorporating a larger number of covariates in this relatively small dataset would likely reduce the robustness and interpretability of the model. Similarly, subgroup analyses (e.g., restricting analyses to pancreaticoduodenectomy or stratifying by stenting status) would have resulted in substantially smaller subgroups and further reduced statistical power, thereby limiting the reliability of the findings.

We acknowledged the problem in the limitation part of the discussion:

Owing to the relatively small cohort size, adjustment in the multivariable models had to be restricted to a limited number of general clinical risk factors, which may have limited the ability to fully account for procedure-specific perioperative confounders.

The reported association between higher Enterococcus faecalis abundance and lower odds of severe complications is intriguing, but in its current form it is difficult to interpret and may be overemphasized. Biologically, this finding is somewhat counterintuitive, and the effect estimate itself is not easy to understand unless the authors specify the exact unit of abundance increase used in the model. It is also unclear whether this association would remain significant after correction for multiple testing, given the number of taxa examined. I would therefore recommend a more cautious presentation, including clarification of the scaling, graphical display of abundance distributions by outcome group, and sensitivity analyses using more robust compositional-data approaches or models that handle sparsity and zero inflation more appropriately. Adjustment for stenting and antibiotic exposure would be particularly important here because Enterococcus is closely linked to both.

Answer: Thank you for your comment. The frequencies of the bacterial species are given as percentages (scale 0–100). Accordingly, when interpreting the odds ratio (OR), a 1-unit increase corresponds to an increase of 1 percentage point. Since most values are below 1%, this scaling is biologically unrealistic; therefore, the OR must be interpreted with caution. We rescaled the values such that a 1-unit increase corresponded to an increase of 0.01 percentage points. As a result, the OR became approximately 1, which illustrates that the findings should be interpreted with great caution and are primarily exploratory in nature.

We added in the discussion section:

(OR 0.89 per 1 percentage point increase, 95%-CI 0.80–0.99, p=0.035)

However, most observed relative abundances were far below 1 %, meaning that a 1 %-point increase represents an unrealistically large change compared with typical patient differences. Realistic changes correspond to minimal effects on risk. Rescaling the predictor indicates that effects corresponding to typical within-sample variation are very small (OR ≈ 0.998 per 0.01% increase), supporting a negligible practical impact. Therefore, this finding should be interpreted cautiously.

Since preoperative biliary stenting is one of the central explanatory variables in the manuscript, it needs to be characterized in much more detail. At present, “stenting” is treated as a relatively simple binary feature, but its microbiological consequences may differ considerably depending on whether the stent was plastic or metal, how long it had been in place, how many ERCP procedures were performed, whether exchanges were necessary, and whether cholangitis occurred before surgery.

Answer: Thank you for this important comment. We agree that the microbiological consequences of preoperative biliary stenting may vary depending on factors such as stent type, duration of placement, number of ERCP procedures, stent exchanges, and the occurrence of preoperative cholangitis. However, given the relatively small cohort size (n = 50), further stratification of the stenting variable into multiple subgroups would have resulted in very small group sizes and substantially reduced statistical power. For this reason, preoperative biliary stenting was intentionally modeled as a binary variable (stented vs. non-stented) to allow a more robust statistical analysis within this dataset. Importantly, in our cohort the vast majority of implanted stents were plastic stents, resulting in very limited variability with regard to stent type. In addition, a considerable proportion of patients received their biliary stents at referring hospitals prior to transfer for surgery. Consequently, detailed information on stent characteristics (e.g., exact type, duration of placement, number of ERCP procedures, or stent exchanges) was not consistently available for all patients.

We have clarified this aspect in the manuscript and acknowledge that a more detailed characterization of biliary stenting may provide additional insights in the limitation part of the discussion:

In addition, preoperative biliary stenting was analyzed as a binary variable, although its microbiological consequences may vary depending on stent-specific characteristics and procedural circumstances.

The manuscript would also benefit from language polishing. Several sections contain long and complex sentences, and in places the writing feels denser than necessary. Because the paper sits at the interface of surgery, microbiome science, and statistical analysis, improved clarity would make it more accessible to a broader readership.

Answer: Thank you for this advise. We therefore carefully revised the language throughout the manuscript to improve readability and reduce sentence complexity where possible.

It would be helpful if the authors stated more directly whether the study is intended as exploratory and hypothesis-generating. Given the single-center setting, the modest sample size, and the complexity of the data structure, a clear statement about the exploratory nature of the work would help frame the findings appropriately.

Answer: Thank you for this helpful suggestion. We agree that the study should be interpreted as exploratory. We have therefore clarified in the Discussion that, given the single-center design and the limited cohort size, the findings should be considered hypothesis-generating and require validation in larger cohorts.

We added: Given the single-center design and the limited cohort size, the present analysis should be considered exploratory and hypothesis-generating, warranting validation in larger, prospectively collected multicenter cohorts to overcome the limitations of small sample size and heterogeneity

We again thank the Editor and Reviewers for their thoughtful and constructive feedback, which has helped us to substantially improve the quality and clarity of the manuscript. We believe that the revisions have strengthened both the methodological transparency and the interpretation of the results, particularly with respect to the exploratory nature of the microbiome analyses and the clinical relevance of the findings.

We hope that the revised version of the manuscript is now suitable for publication and we would be pleased to provide any additional clarification if required.

With kind regards,

on behalf of all authors

Artur Rebelo and Laura Oelschlägel

Reviewer 2 Report

Comments and Suggestions for Authors

I have reviewed the manuscript “Microbiome Diversity in Pancreatic Surgery: Associations with Preoperative Stenting and Postoperative Outcomes”. This paper investigates the composition of the microbiome across multiple anatomical compartments (bile, pancreatic fluid, duodenum, tumour tissue, and healthy pancreatic tissue) in patients undergoing pancreatic surgery. Using 16S rRNA gene sequencing (both DNA- and RNA-based approaches), the authors analyze 224 samples from 58 patients to evaluate microbial diversity and assess associations with preoperative biliary stenting and postoperative outcomes.

The study reports that microbial composition differs across anatomical sites, with significant beta-diversity differences between duodenal, bile, and pancreatic fluid samples. Preoperative biliary stenting was associated with statistically significant shifts in microbial composition, although with small effect sizes. Additionally, Enterococcus faecalis was identified as being associated with a reduced risk of severe postoperative complications (Clavien–Dindo ≥ III).

The topic is timely and clinically relevant, particularly in the context of increasing interest in the role of microbiota in surgical outcomes and cancer biology. The multi-compartment sampling and combined DNA/RNA analysis represent important strengths.

Major Concerns

Interpretation of microbiome–outcome associations

The association between Enterococcus faecalis and reduced postoperative complications is an interesting finding but remains insufficiently supported
The reported effect size is modest (OR = 0.89, p = 0.035) and derived from a relatively small cohort (n = 50 in regression analysis),
Only relative abundance was used, which may bias interpretation,
No correction for multiple testing is described despite multiple taxa being evaluated.

This raises the possibility of a type I error.

The authors should explicitly state that this finding is exploratory and hypothesis-generating rather than confirmatory.

Dominant role of interindividual variability

A key result of the study is that interindividual variability explains the majority of microbiome variance (R² up to ~0.75–0.97 in several analyses).

Despite this, the manuscript places considerable emphasis on group-level differences (e.g., stent vs. no stent), which show:

low explanatory power (R² often ~0.03–0.05)
overlapping clustering in nMDS plots (Figures 3–5)

The conclusions should be reframed to emphasize that patient-specific factors dominate microbiome composition, while group-level effects are modest.

Statistical significance vs. biological relevance

Several findings reach statistical significance (e.g., PERMANOVA p-values), but effect sizes are consistently small.

For example:

stent vs. no stent comparisons show significant p-values but low R² (~4–5%)
anatomical differences are statistically significant but explain only a minor fraction of variance

The manuscript would benefit from a clearer distinction between statistical significance and biological relevance, as current interpretations may overstate the importance of these findings.

Cohort heterogeneity and subgroup analyses

The study includes multiple disease entities (pancreatic cancer, cholangiocarcinoma, chronic pancreatitis), introducing biological heterogeneity.

Additionally:

subgroup sizes are small (e.g., pancreatitis n = 7, CCA n = 11)
some analyses rely on limited sample numbers

The authors should better justify pooling strategies and more clearly acknowledge potential confounding effects between diagnosis and microbiome composition.

Methodological transparency

Important methodological details are primarily presented in the Supplementary Materials.

Key elements that should be summarized in the main text include:

sequencing and bioinformatic pipeline (e.g., filtering, ASV assignment)
contamination control procedures
criteria for taxonomic classification

Providing these details in the main Methods section would improve reproducibility and transparency.

Minor Comments

Interpretation of alpha diversity

Although no statistically significant differences were observed (p > 0.05), the manuscript refers to “clear trends.” These should be interpreted cautiously and not overstated.

nMDS visualization and interpretation

Figures 3–6 show relatively high stress values (~0.2–0.27), indicating limited reliability of ordination.

The authors should explicitly acknowledge this limitation and avoid overinterpreting spatial clustering patterns.

Language and structure

The manuscript is generally well written, but some sections (particularly Results) are overly descriptive, especially regarding taxonomic composition. Consider condensing repetitive elements to improve readability.

Terminology - the term “Escherichia/Shigella” should be briefly clarified in the main text. Additionally, the distinction between DNA-based (total microbiome) and RNA-based (active microbiome) analyses should be introduced earlier for clarity.

5 Some figures are redundant or provide limited additional insight given the high overlap between groups. The authors may consider streamlining figure presentation.

Overall Assessment and Recommendation

This manuscript addresses a relevant and emerging topic and includes several methodological strengths, particularly the multi-site sampling and combined DNA/RNA analysis.

However, the current version tends to overinterpret statistically significant findings with limited biological effect sizes and does not sufficiently emphasize the dominant role of interindividual variability.

I recommend major revision, with particular focus on:

tempering causal and clinical interpretations
clearly distinguishing statistical from biological significance
improving methodological transparency
reframing conclusions to reflect the exploratory nature of key findings.

With these revisions, the manuscript could represent a valuable contribution to the field.

Comments on the Quality of English Language

English needs to be checked.

Author Response

Dear Editor and Reviewers,

Below, we provide a detailed point-by-point response to each comment. All changes made in the manuscript are indicated accordingly.

Comments and Suggestions for Authors

Major Concerns

Interpretation of microbiome–outcome associations

The association between Enterococcus faecalis and reduced postoperative complications is an interesting finding but remains insufficiently supported
The reported effect size is modest (OR = 0.89, p = 0.035) and derived from a relatively small cohort (n = 50 in regression analysis),
Only relative abundance was used, which may bias interpretation,
No correction for multiple testing is described despite multiple taxa being evaluated.

This raises the possibility of a type I error.

The authors should explicitly state that this finding is exploratory and hypothesis-generating rather than confirmatory.

Answer:

We added a statement in section “Binary logistic regression”: Given the exploratory nature of the analysis, no formal correction for multiple testing was applied

And also in the discussion: Considering the relatively small cohort, the evaluation of multiple taxa without correction for multiple testing, and the use of relative abundance data, the possibility of a type I error cannot be excluded. Therefore, this finding should be regarded as exploratory and hypothesis-generating, requiring validation in larger cohorts.

Dominant role of interindividual variability

A key result of the study is that interindividual variability explains the majority of microbiome variance (R² up to ~0.75–0.97 in several analyses).

Despite this, the manuscript places considerable emphasis on group-level differences (e.g., stent vs. no stent), which show:

low explanatory power (R² often ~0.03–0.05)
overlapping clustering in nMDS plots (Figures 3–5)

The conclusions should be reframed to emphasize that patient-specific factors dominate microbiome composition, while group-level effects are modest.

Answer: Thank you for pointing that out. In the respective sections of the PERMANOVA analyses, the high degree of inter-individual variability has already been highlighted.

In contrast, interindividual differences explained the majority of variance (R²=0.751; p=0.001).

These findings indicate a statistically significant but modest effect of preoperative stenting on the biliary microbiome, while interindividual differences remain the primary driver of microbial variability

But we added in the discussion a statement: … In contrast, interindividual variability emerged as the predominant factor shaping microbial composition.

Statistical significance vs. biological relevance

Several findings reach statistical significance (e.g., PERMANOVA p-values), but effect sizes are consistently small.

For example:

stent vs. no stent comparisons show significant p-values but low R² (~4–5%)
anatomical differences are statistically significant but explain only a minor fraction of variance

The manuscript would benefit from a clearer distinction between statistical significance and biological relevance, as current interpretations may overstate the importance of these findings.

Answer: Thank you for the comment.

We added in the discussion section:

Although diversity analyses revealed statistically significant differences, their biological impact appears limited due to the small effect sizes. In contrast, interindividual variability emerged as the predominant factor shaping microbial composition.

Cohort heterogeneity and subgroup analyses

The study includes multiple disease entities (pancreatic cancer, cholangiocarcinoma, chronic pancreatitis), introducing biological heterogeneity.

Additionally:

subgroup sizes are small (e.g., pancreatitis n = 7, CCA n = 11)
some analyses rely on limited sample numbers

The authors should better justify pooling strategies and more clearly acknowledge potential confounding effects between diagnosis and microbiome composition.

Answer: We added the following section to the Discussion addressing the above-mentioned critique

To increase statistical power for multivariable regression analyses, patients with different disease entities were pooled. This approach introduces biological heterogeneity and may confound associations between diagnosis and microbiome composition, which should be considered when interpreting group-level differences

Methodological transparency

Important methodological details are primarily presented in the Supplementary Materials.

Key elements that should be summarized in the main text include:

sequencing and bioinformatic pipeline (e.g., filtering, ASV assignment)
contamination control procedures
criteria for taxonomic classification

Providing these details in the main Methods section would improve reproducibility and transparency.

Answer: Thank you for pointing that out. We therefore elaborated on the aforementioned points in more detail in the Methods section to improve clarity and understanding

Following rigorous quality control, 224 samples were included for taxonomic classification and downstream analyses. A detailed workflow including exclusion criteria is shown in Figure 1. Taxonomic classification was performed based on ASVs using the RDP database

Minor Comments

Interpretation of alpha diversity

Although no statistically significant differences were observed (p > 0.05), the manuscript refers to “clear trends.” These should be interpreted cautiously and not overstated.

Answer: We adjusted the wording accordingly in the results section of alpha diversity analysis

However, descriptive differences between groups were observed (Figures 2).

nMDS visualization and interpretation

Figures 3–6 show relatively high stress values (~0.2–0.27), indicating limited reliability of ordination.

The authors should explicitly acknowledge this limitation and avoid overinterpreting spatial clustering patterns.

Answer: Thank you for the suggestion. We adjusted the wording of the NMD interpretations so that the text conveys a more descriptive tone.

For example we used “pattern/groupings” instead of “cluster”

Language and structure

Answer: Thank you for pointing that out. We attempted to shorten some sections.

Terminology - the term “Escherichia/Shigella” should be briefly clarified in the main text. Additionally, the distinction between DNA-based (total microbiome) and RNA-based (active microbiome) analyses should be introduced earlier for clarity.

Answer: We added an explanation in the Methods section regarding the Escherichia/Shigella terminology. Furthermore, we revised the text to state that the distinction between DNA and RNA samples is already introduced in the Methods section. To also address the comment about making the text more accessible to a broader readership, we additionally explain the different implications of DNA and RNA analyses in relation to microbial activity.

Method section:

Samples were analyzed using 16S rRNA gene sequencing, with both DNA and RNA extracted from each sample whenever possible to assess the total and potentially active microbial fractions.

For some bacterial genera, species-level assignment was not possible due to the limited taxonomic resolution of the 16S rRNA gene in the V1–V2 region. This particularly applies to the family Enterobacteriaceae. For example, Escherichia coli and Shigella cannot be distinguished based on their 16S rRNA sequences and are hereafter referred to as “Escherichia/Shigella”.

Result-section: alpha diversity analysis

Additionally, discrepancies were observed between the results of separate analyses of DNA-based (total microbiome) and RNA-based (active microbiome) samples.

5 Some figures are redundant or provide limited additional insight given the high overlap between groups. The authors may consider streamlining figure presentation.

Overall Assessment and Recommendation

This manuscript addresses a relevant and emerging topic and includes several methodological strengths, particularly the multi-site sampling and combined DNA/RNA analysis.

I recommend major revision, with particular focus on:

tempering causal and clinical interpretations
clearly distinguishing statistical from biological significance
improving methodological transparency
reframing conclusions to reflect the exploratory nature of key findings.

With these revisions, the manuscript could represent a valuable contribution to the field.

We hope that the revised version of the manuscript is now suitable for publication and we would be pleased to provide any additional clarification if required.

With kind regards,

on behalf of all authors

Artur Rebelo and Laura Oelschlägel

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have addressed all of my concerns satisfactorily.