Inflammatory Biomarker Profiles in Very Preterm Infants within the Context of Preeclampsia, Chorioamnionitis, and Clinically Diagnosed Postnatal Infection

Round 1

Reviewer 1 Report

The article "Inflammatory Biomarker Profiles in Very Preterm Infants within the Context of Preeclampsia, Chorioamnionitis and Clinically Diagnosed Postnatal Infection" by Edwald  et al describes cytokine profiles in venous cord blood in periviable and very preterm infants. Although there is some data on the cytokines and chemokines profiled in this cohort, there is only scant data in the periviable and extreme preterm population. The article is succinct and well written. In particular the correlation of these biomarkers with histopathologic analysis and the presence or absence of funisitis has potential clinical utility.  I have no major comments or issues but several minor comments listed below. 

 

Minor comments

- would recommend that the cytokine/chemokine data is also stratified by mode of delivery and the presence or absence of labor- this would add another figure and clinically useful information

-When referring to placental pathology consistent with chorioamnionitis, would recommend that this is specified in the text to separate out histopathologic chorioamnionitis and clinical chorioamnionitis to clearly differentiate the 2 entities

- Does the institution use cord sampling for blood cultures in the diagnosis and treatment of early onset sepsis? If so, the results would be relevant and interesting to the study. 

- it is interesting that the authors stratified spontaneous preterm labor into the suspected infection group and cervical insufficiency into the infection not suspected group when the clinical presentation of these overlaps significantly. I would  recommend that the justification for this is discussed with the addition of a line in the text. 

- I recommend describing the inclusion of the institution specific protocols for antibiotic treatment, multiple institutions use a 48 hour antibiotic rule out and then will discontinue antibiotics if there is no evidence of infection due to increased risks associated with postnatal antibiotic exposure and NEC. As this is a single institution study, it is relevant to describe if 7 days of antibiotics are used universally. 

- the postnatal infections should be differentiated with clinical terms of early onset sepsis (<7 days) and late onset sepsis (>7 days). In particular late onset sepsis is associated with a variety of other risk factors including necrotizing enterocolitis, spontaneous intestinal perforation,  TPN, ventilation and is not strongly associated with obstetric factors. It is important to revise the paragraph in line 123 to reflect this. The current reading suggests that the 2 infants with late onset sepsis acquired this through horizontal transfer due to prenatal infection- this is highly unlikely with these pathogens

Author Response

We appreciate your comments and improved the manuscript by completing the following specific updates.

1. The article is succinct and well written. In particular the correlation of these biomarkers with histopathologic analysis and the presence or absence of funisitis has potential clinical utility.

Thank-you!

2. Recommend stratifying by mode of delivery and labor; this would add clinically useful information

We incorporated this data into Table 2. Notably, Cesarean delivery and absence of labor were associated with significantly higher MCP-1 levels, consistent with the increased MCP-1 levels in pregnancies complicated by preeclampsia. We also note increased CRP levels associated with labor. These associations are now described in the results and discussion sections beginning on lines 174 and 281, respectively.

3. In the text, separate histopathologic chorioamnionitis and clinical chorioamnionitis to clearly differentiate the entities

We appreciate this recommendation and have updated the text.

4. The authors stratified spontaneous preterm labor into the suspected infection group and cervical insufficiency into the infection not suspected group when the clinical presentation of these overlaps significantly. I would recommend that the justification for this is discussed with the addition of a line in the text.

We updated the text on line 105, “While there can be overlap in the presentation of cervical insufficiency and spontaneous preterm labor, the absence of uterine contractions or labor clinically distinguished the former from the latter.”

5. Does the institution use cord sampling for blood cultures in the diagnosis and treatment of early onset sepsis? Describe the institution specific protocol for antibiotic treatment (are 7 days of antibiotics used universally or does the institution use a 48 hour antibiotic rule out and then discontinue antibiotics if there is no evidence of infection).

Line 108 now clarifies, “Our institution does not collect universal blood cultures at delivery. In the case of suspected infection based on the prenatal history and postnatal status, blood cultures may be obtained. When blood cultures are obtained, antibiotics are universally initiated, but there is an automatic stop order placed and more than 48 hours of antibiotics are only administered for persistent concern for infection or culture-proven infection.”

6. The postnatal infections should be differentiated with clinical terms of early onset sepsis (<7 days) and late onset sepsis (>7 days). It is important to revise the paragraph in line 140 to reflect this. The current reading suggests that the 2 infants with late onset sepsis acquired this through horizontal transfer due to prenatal infection.

We agree the second paragraph of the results was confusing, and it has been revised on lines 140 to 151 to clearly distinguish early-onset and late-onset sepsis.

Reviewer 2 Report

Materials and Methods:
Lines 102 ff: "Continuous variables were analyzed by the Kruskal-Wallis test ..... One-way ANOVA was utilized to compare biomarkers  ...."
My comment: The choice of an appropriate statistical doesn't depend from the fact whether the variable of interest is a biomarker or not. Rather, it depends on the distribution. For variable approximately normally distributed ANOVA is appropriate.

The sentence "Factors that we found significant .... were contrasted by two-way ANOVA with the Holm-Sidak test ...." is not clear to me. What is the second factor in the 2-way ANOVA? Why do you use Holm-Sidak test only in the 2-way ANOVA? Furthermore: Why do you a significance level of 0.01 in addition to Holm-Sidak correction?

Table 1: You should add that for "Prenatal Infection suspected" multiple answers are possible.

Results, line 142: You present means togther with standard errors. Therefore, you should give the number of measurements.

Table 2 and text: p values should be presented uniformly. It is not senseful just to write: "p < 0.05" or "p < 0.01" and to give no p values for non significant results. 

2 way ANOVAs: Has the interaction term been investigated? I think this would be informative.

Figure 2: MCP-1 and IL-6 values are presented. What about CRP?
Figure 3: What about IL-6? For these investigations with 2 factors (i.e. suspected prenatal infection AND PET), 2 way ANOVAs would be useful. In my opinion, the term "post-hoc subgroup analysis" is misleading (as mentioned in Line 169). For instance, figure 3a: Why do you compare the subgroups "PET / no PET" only for infants with "no suspected prenatal infection" and why do you compare the Chorio-subgroups (yes / no) only in the population of infants with suspected prenatal infection. This is not clear

Discussion, lines 251 and 252: "It is notable that ..... the directionality .. paralleled the significant increase in MCP-1 ...". It would be helful for the reader to refer on the corresponding figures 3a and 3b. Furthermore, exact p values sould be given.

 

 

 

 

N/A

Author Response

We appreciate your comments and improved the manuscript by completing the following specific updates.

 1. Line 121: "Continuous variables were analyzed by the Kruskal-Wallis test ..... One-way ANOVA was utilized to compare biomarkers ...." For variables approximately normally distributed, ANOVA is appropriate.

We agree. The continuous variables are analyzed by ANOVA, and that is now reflected in the text.

2. Line 125: What is the second factor in the 2-way ANOVA?

Postnatal age was the second factor, and that is now clarified in the methods and results section.

3. Line 129: Why do you a significance level of 0.01 in addition to Holm-Sidak correction?

We clarified that the high level of significance was used for unplanned subgroup analyses to minimize the risk of type 1 error.

4. Table 1: You should add that for "Prenatal Infection suspected" multiple answers are possible.

We have added, “In some cases, prenatal infection was suspected based on the presence of more than one diagnosis.”

5. Line 163: You present means with standard errors. Therefore, you should give the number of measurements.

We have provided the number of measurements (i.e., N = 12).

6. Table 2: p values should be presented uniformly.

 We have included all the specific P values.

7. Figure 2: MCP-1 and IL-6 values are presented. What about CRP?

We have added a panel for the CRP values.

8. Figure 3: What about IL-6? For these investigations with 2 factors, 2 way ANOVAs would be useful. The term "post-hoc subgroup analysis" is misleading. For instance, figure 3a: Why do you compare the subgroups "PET / no PET" only for infants with "no suspected prenatal infection" and why do you compare the Chorio-subgroups (yes / no) only in the population of infants with suspected prenatal infection.

We have added a panel for the IL-6 values. One-way ANOVA was utilized because we were comparing subcategories within the larger categories, e.g., “PET” or “No PET” are the two subgroups within “No Suspected Prenatal Infection”. We agree “post-hoc” is misleading and changed it to “unplanned” on line 193. We clarified, “unplanned subgroup analyses were performed by subcategorizing those without suspected prenatal infection based on their PET status and subcategorizing those with suspected prenatal infection based on their Chorio status.”

9. Two-way ANOVA: Has the interaction term been investigated? I think this would be informative.

Thank-you for the suggestion. We have added on line 215, “Overall, there was not a statistically significant interaction between postnatal age and infection status for IL-6 (P = 0.08), but a significant interaction was present for CRP (P < 0.001).”

10. Line 288: "It is notable that ..the directionality ..paralleled the significant increase in MCP-1 ..". It would be helpful for the reader to refer on the corresponding figures 3a and 3b. Furthermore, exact p values should be given.

We have updated the manuscript as suggested.

Reviewer 3 Report

In this study Ewald et al looks at some biomarkers in very preterm infants, and looks how they are altered in different pre/postnatal clinical settings. While there are couple of notable differences in MCP1, IL6 and CRP between different groups, I think there is no sufficient data for a strong association or application of any of the tested readouts as a biomarker. I think either further data to increase strength of the associations or more mechanistic understanding of why the levels of these cytokines and chemokine change between different groups is necessary, or testing of more comprehensive panel of cytokines and chemokines to look for differences is required to publish this study. 

 

Other minor comments: 

lines 65-69: It would increase general clarity if the definition includes how the gestation time is different for moderately preterm (born at 32 to 34 weeks of pregnancy) and extremely preterm (born before 32 weeks of pregnancy) infants. 

Figure 4 legend: Why are results described within the figure legend? Figure legend is only supposed to provide information regarding the contents in the figure.

Author Response

We appreciate your comments and improved the manuscript by completing the following specific updates.

1. I think there is not sufficient data for application of any of the tested readouts as a biomarker. I think either further data to increase strength of the associations or more mechanistic understanding of why the levels of these cytokines and chemokine change between different groups is necessary.

We agree that our study, even in the context of other studies, is not sufficient to modify current clinical management, but we do consider our study a novel and important contribution to the field. Within the conclusion we highlight the need for further studies to strengthen the associations and enhance our mechanistic understanding.

2. Line 69: It would increase general clarity if the definition includes how the gestation time is different for moderately preterm (born at 32 to 34 weeks of pregnancy) and extremely preterm (born before 32 weeks of pregnancy) infants.

We apologize for the oversight and clarify that, as recommended.

3. Figure 4: Why are results described within the figure legend?

We have removed a description of the results from the figure legend.

Reviewer 4 Report

The authors present an research article where they analyze diffrerent biomarkers associated with neonatal infection in the context of preeclampsia.

1. Although the authors summarize the rationale behind this study in the introduction. I feel a further, in detail explanation on why they selected these biomarkers will be useful. Also, I ask them to go into extra details about the role of this biomarkers in inflammation/regulation of inflammation.

2. Did the authors calculated a sample size in advance, or did they just use a number of infants available? Please, add this information.

3.  In figure 3, did the authors calculate statistical significance for the days graphed or just showed the values per day?

4. Please for Figure 2 and 4 add figure legends next to the graphs instead of only in the figure description.

5. In the discussion, please explain in further detail the lack of immune response associated with the chorioamnionitis.

6. I feel the results after figure 1 and during discussion are heavily descriptive but lack certain biological interpretations which make it difficult for the readers to understand and relate to, please try to correct this by revising the writing of some of these sections.

7. The references are mostly from before 2020, with some exceptions, please revise the text as mentioned in the previous points and, if available, include updated references.

8. Add a conclusion section.

9. The authors mentioned postnatal infection of 44%, but only had 3 positive cultures. As they mention in their methods, 7 days of antibiotics are considered postnatal infection. Is this a standard? This categorization seems to induce some great bias in the definition of infection. Please explain in detail this decision and support it based on previous studies.

 

Author Response

We appreciate your comments and improved the manuscript by completing the following specific updates.

1 and 7. Explain why the biomarkers were selected, including their role in inflammation / regulation of inflammation. The references are mostly from before 2020, with some exceptions, please include updated references.

We updated the second paragraph of the introduction, including several post-2020 references.

2. Please specify if the authors calculated a sample size or used a number of infants available.

We clarified on line 85 that all infants were eligible during the 20 month-long recruitment phase.

3. In Figure 1, did the authors calculate statistical significance for the days graphed or just show the values per day?

We did not assess statistical significance in that Figure, it sets the stage for the analyses in the subsequent Figures.

4. For Figure 2 and 4 add figure legends next to the graphs instead of only in the figure description.

We have incorporated this recommendation.

5 and 6. Results and discussion are descriptive but lack certain biological interpretations which make it difficult for the readers to understand. In the discussion, explain the lack of immune response associated with the chorioamnionitis.

Beginning on line 264, we have added, “The etiology of the hypo-responsive transcriptional phenotype of perinatal monocytes has not been fully elucidated [24], but ovine studies suggest the development of endotoxin cross-tolerance following repeated exposure to intra-amniotic cytokines, as elaborated during the evolution of chorioamnionitis [25].”

8. Add a conclusion section.

We have added that section.

9. The authors mention in their methods, 7 days of antibiotics are considered postnatal infection. Is this a standard? Please support it based on previous studies.

That is the standard for our institution, and we clarified that and added a reference from a second institution / publication on lines 112-118.

Reviewer 5 Report

The report is attached 

Comments for author File: Comments.pdf

Minor

Author Response

We appreciate your comments and improved the manuscript by completing the following specific updates.

1. Overall, the findings are interesting and would provide essential information for biomarkers screening and relevant diagnosis in infants.

Thank-you!

2. Detailed information about the measurement of the biomarkers is needed to ensure results are reproducible and applicable (for ex. used kits, used protocol) or at least cite one paper to ensure that their results are reproducible.

We apologize for the oversight and added the reference [19] that describes our assay technique. The methods section on lines 90-98 now states, “MCP-1 and IL-6 were measured on 200 microliters of blood collected into EDTA tubes daily for 7 days and then weekly until achievement of either discharge or maturation to 36 weeks postmenstrual age. Samples were obtained during previously scheduled morning lab draws and processed within 4 hours with plasma was stored at −80 degrees Celsius. The plasma samples were analyzed using a customized magnetic bead assay (Millipore Sigma, Burlington, MA, USA) on BioPlex 200 with BioPlex manager 6.1 software (Bio-Rad, Hercules, CA, USA) [19]. All samples were run in duplicate, alongside 7 standards and a quality control, with our intra-assay correlation coefficient > 0.9.”

3. It is not clear if other infections, rather than the tested E. coli, S. epidermidis, and K. pneumonia, was ruled out? Were the levels of infections monitored overtime as the measured biomarkers to ensure that such change in biomarkers is solely attributed to the examined microorganism and not as a general host response?

The discussion now includes the limitation on line 322 that, “We do not routinely screen for viral infections beyond the meningitis panel included in cerebrospinal fluid testing, but the aerobic blood cultures that were obtained do detect a wide variety of bacteria and fungi, and all the positive results are reported.”

4. Line 123: One-way ANOVA, is there any correction test has been tested here?

We now include that the one-way ANOVA analyses did not incorporate any correction for multiple comparisons.

5. Line 156: relatively stable values, relative to what? What the authors mean here?

We changed to “the levels then persisted relatively unchanged” to clarify.

6. Line 164: for the infant with S. epidermidis sepsis, why were the biomarker levels low despite of the infection?

We added, “The low biomarker levels in the presence of positive cultures for S. epidermidis is consistent with the findings of others, and the relatively low virulence nature of that infection [21].”

7. Figure 1: can the authors use the same log scale in both panels (MCP-1 and IL-6)?

We are now using the same log scale for both panels.

8. Line 169: acute or chronic (is a typo here or?)

We clarified it was an acute exacerbation of chronic multi-organ failure.

9. Line 286: both infectious and non- infections (also another typo?)

Thank-you for noticing that “non-infections” should have been “non-infectious”

Round 2

Reviewer 2 Report

Tthe manuscript has been sufficiently improved. In my opinion, there is nothing to be said against publication.

Author Response

Thank-you for reviewing our manuscript!

Reviewer 3 Report

I do feel while the authors points are valid, current data alone still doesn't make a strong case for a conclusive paper. 

The specific comments are those which I already mentioned in my first review report and authors are already aware of. Such as providing more mechanistic data or comprehensive panels. Please refer to my first review report for further clarification.

Author Response

We understand and appreciate your review of our manuscript. Although additional mechanistic or clinical data are not available, the data we provide can advance research and patient care once it is disseminated, replicated and expanded. We have revised the concluding paragraph of the introduction and the conclusion itself to better highlight these considerations.

Reviewer 4 Report

The authors have addressed all the issues presented in the previous round of reviews.

I ask the authors to improve the quality of their images because some of them are hard to read.

I also ask the authors to rephrase the conclusion and the final paragraph of the introduction to highlight the importance of this study. 

Author Response

Thank-you for reviewing our manuscript. We improved the readability of each figure panel. We also rephrased the conclusion and introduction. The introduction now highlights our main conclusions as suggested in the journal’s instructions.