Cytokine Patterns in COVID-19 Patients: Which Cytokines Predict Mortality and Which Protect Against?
Round 1
Reviewer 1 Report
Even if I'm not able to see the issues from the previous reviewers, looking to the unpublished material, seem that the authors solved all the previous concerns. If so, I believe that it’s suitable for publication after minor revisions.
Below, the minor points to solve prior to accept for pubblication:
- increase the quality of the table 3 panel C and of the figure 1 (resize fig.1 panel B and use the same font/style for axes beetween panels A and B)
- use the same characters (font, size and style) in the tables (eg. in table 1 see Age line, Lab Findings, D-dimer etc etc)
- expand the Limitation of the study section (add some lines about treatments, unbalanced study cohort - 34 VS 6 patients)
- Inclede some lines in the study populations related to the timing of the infection, all patients in the same time frame of one months? two? please specify in order to support also the thesy that al infection are due to Delta variant..
Author Response
We would to thank the reviewer for his promise comments.
The minor points to solve:
- increase the quality of the table 3 panel C and of the figure 1 (resize fig.1 panel B and use the same font/style for axes beetween panels A and B), Done
- use the same characters (font, size and style) in the tables (eg. in table 1 see Age line, Lab Findings, D-dimer etc etc): Done
- expand the Limitation of the study section (add some lines about treatments, unbalanced study cohort - 34 VS 6 patients). Done
- Include some lines in the study populations related to the timing of the infection, all patients in the same time frame of one months? two? please specify in order to support also the thesthat al infection are due to Delta variant. Done in lines 68-70.
Reviewer 2 Report
Review of the Article:
cimb-1936563: Cytokine Patterns in COVID-19 Patients: Which Cytokines Predict Mortality and Which Protect Against?
Minor corrections:
Abstract:
The following phrase: However certain cytokines have protective effects and higher levels of these cytokines increase survival levels and lower lung damage” or something like that.
In my opinion, the quotation and the "something like that" diminish the seriousness of the article, so I strongly encourage you to rewrite the paragraph.
Materials and methods:
It is important to determine the COVID-19 strains from all the patients, since this is key to determine that the differences are not due to different strains, but differences in the patients’ responses.
It is indicated that the patients were diagnosed via PCR assays in both Study population and study design. Try not to repeat information.
In line 84 indicate what the acronym “CRP” stands for
Results:
Table 1 contains different font types
In line 135: I think you are describing 40 patients, not 40 patient groups. Remove the word groups.
Tables 2A and 2B: There are several values in bold but their metabolites are not statistically significant. The bolded letters are not showing anything in particular, I recommend leaving only the significant metabolites bolded. Additionally, there are different fonts in the data.
The legend of the figures 2A,2B,2C, requires more information. Figure 2C needs to be explained more, what does 0 and 1 mean. Also, you mention “SE are the standard errors of the regression coefficients” But there is no SE anywhere in the tables.
Table 3: The correlation coefficient was not reported. Here it is necessary to add the correlation matrix to understand the correlation patterns of the dataset. Also in this table, “SE are the standard errors of the regression coefficients” But there is no SE anywhere in the table.
Table 3C is not a table, it is a figure.
Table 4: Different font types here as well. Correct the SE legend.
Major revision:
Significant amounts of changes are necessary.
All data must be presented, even if the results are not statistically significant. For instance, table 2B cannot contain only the significant variables. It is also necessary to report the other variables as well.
Figure 2 and the related statement are very daring and depend on the model selected. The R2 value of that model must be reported and I bet it is very small. The amount of data is very low and based on just one point at the left, you can’t conclude that IGF1 is high when % of lung injury is low, because in the following set of points, you have IGF1 values lower than 40 pg/mL (the lower value of the entire dataset).
It has been proven that the sex of the individuals is associated with a particular type of response. In particular, females present a more active antiviral response, while males have higher basal levels of pro-inflammatory cytokines.
I recommend reading and including this article:
(https://pubmed.ncbi.nlm.nih.gov/33271804/)
In conclusion, the sex of the patients is a co-variate that can’t be ignored, and stratifying your data based on that will provide you with deeper conclusions.
A very significant factor to take into account is the COVID-19 strain. This cannot be ignored because, each strain provokes different symptoms and reactions in the patients’ bodies.
Unfortunately, I wouldn’t recommend this article for publication since it requires a lot of work and deeper data analysis.
Author Response
We would to thank the reviewer for his comments.
Minor corrections:
Abstract:
The following phrase: However certain cytokines have protective effects and higher levels of these cytokines increase survival levels and lower lung damage” or something like that.
The paragraph was rewritten.
Materials and methods:
It is important to determine the COVID-19 strains from all the patients, since this is key to determine that the differences are not due to different strains, but differences in the patients’ responses.
The strain of the virus is probably the Delta one. All patient hospitalized within one month. Most probably they infected with same strain of the virus. Lines 68-70.
It is indicated that the patients were diagnosed via PCR assays in both Study population and study design. Try not to repeat information.
DONE, the point was deleted in the study design
In line 84 indicate what the acronym “CRP” stands for: DONE
Results:
Table 1 contains NOW the same font types.
In line 135: The word groups was removed
Tables 2A and 2B: the significant metabolites are bolded. And the same font was done.
The legend of the figures 2A,2B,2C now are detailed better. Figure 2C now is more clear Also, you mention “SE are the standard errors of the regression coefficients is for part A&B. The legend was corrected, lines 140-147.
Table 3: The correlation coefficient was not reported. Here it is necessary to add the correlation matrix to understand the correlation patterns of the dataset. Also in this table, “SE are the standard errors of the regression coefficients” But there is no SE anywhere in the table.
Table 3C is now considered as figure 1.
Table 4: The same font type, done. The SE legend was corrected.
Major revision:
- All data must be presented, even if the results are not statistically significant. For instance, table 2B cannot contain only the significant variables. It is also necessary to report the other variables as well
****Multivariate analysis is used to describe analyses of data where there are multiple variables or observations for each unit or individual. The analysis of these parameters present just the significant parameters which are correlated with mortality. Those which are not, the SIGMA STAT statistical program didn’t show.
- Figure 2 and the related statement are very daring and depend on the model selected. The R2 value of that model must be reported and I bet it is very small. The amount of data is very low and based on just one point at the left, you can’t conclude that IGF1 is high when % of lung injury is low, because in the following set of points, you have IGF1 values lower than 40 pg/mL (the lower value of the entire dataset).
Figure 2 (which is now figure 3) was corrected. R=0.46
- It has been proven that the sex of the individuals is associated with a particular type of response. In particular, females present a more active antiviral response, while males have higher basal levels of pro-inflammatory cytokines.
I recommend reading and including this article:
(https://pubmed.ncbi.nlm.nih.gov/33271804/)
In conclusion, the sex of the patients is a co-variate that can’t be ignored, and stratifying your data based on that will provide you with deeper conclusions.
**Data on the correlation between gender and mortality was added. Lines 200-212 in the result session. New Figure 4 discuss this issue and in the discussion session in lines 306-313.
- A very significant factor to take into account is the COVID-19 strain. This cannot be ignored because, each strain provokes different symptoms and reactions in the patients’ bodies.
The strain of the virus is probably the Delta one. All patient hospitalized within one month. Most probably they infected with same strain of the virus. Lines 68-70
Round 2
Reviewer 2 Report
Revision 2:
My answer:
The article improved remarkably, and one of the main suggestions, to include sex components in the analysis, was incorporated into the article.
On the other hand, The fact that the COVID-19 strain wasn’t identified in all samples is an important flaw. I believe this aspect should be thoroughly discussed so that the readers may take it into consideration.
Again, I don’t agree with the following statement for it is to daring:
“IGF-1 didn’t significantly affect the severity or mortality in COVID-19 patients. However, low concentrations of IGF-1 were correlated with high lung involvement and conversely, a high concentration of IGF-1 was seen in patients with little lung involvement”.
Low concentrations of IGF-1 are not correlated with high lung involvement, because your model R^2 value is 0.2116. This is a very low correlation between your model and your data. This has to be rephrased.
Minor revision:
Fix legend of figure 4, it says “orang”
Author Response
We would to thank the reviewer again for his promise comments.
Response to comments:
- The strain of the corona virus is detailed in lines 64-70, "Electronic medical record (EMR) data from 40 patients diagnosed with COVID-19 from June 2021 to August 2021, in the Galilee Medical Center’s COVID-19 Department, Nahariya, Israel, were used as the database. The patients were diagnosis based on a positive polymerase chain reaction (PCR) assay for the SARS-CoV-2 virus. All of these patients were not vaccinated. The strain of the virus is probably the Delta one. Most probably they infected with same strain of the virus.
- The statement about IGF-1 was rearranged in lines 192-195. " IGF-1 didn’t significantly affect the severity or mortality in COVID-19 patients. However, low concentrations of IGF-1 were correlated weakly (R=0.46) with high lung involvement and conversely (Fig. 3).
Minor revision:
The legend of figure 4 was corrected
Round 3
Reviewer 2 Report
After a very thorough review, I have no further questions or concerns.
