Factors Associated with Mortality in Coronavirus-Associated Mucormycosis: Results from Mycotic Infections in COVID-19 (MUNCO) Online Registry

Round 1

Reviewer 1 Report

Introduction
- Several pathogens act by evading the host immune system and survive within the human body through bacterial internalization mechanisms. Staphylococcus aureus has demonstrated the ability to interact and infect different cell strains, such as osteoblasts, causing osteomyelitis and bone and joint infections, while becoming increasingly resistant to antibiotic therapy and a reservoir of bacteria that can make infection difficult to treat.
- The 2019 coronavirus pandemic is a rapidly evolving global emergency that continues to challenge healthcare systems. Emerging research describes a plethora of patient factors, including demographic, clinical, immunologic, hematologic, biochemical, and radiographic findings, that could be useful to clinicians in predicting the severity and mortality of COVID-19. The current literature regarding predictive factors for the clinical course and outcomes of COVID-19 describes findings associated with increased disease severity and/or mortality, including age > 55 years, multiple preexisting comorbidities, hypoxia, specific computed tomography findings indicative of extensive pulmonary involvement, various laboratory test abnormalities, and biomarkers of end-organ dysfunction.
- There is a need to better understand the new coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), for which coronavirus disease 2019 (COVID-19) continues to cause significant morbidity and mortality worldwide. COVID-19 is comparable to other respiratory infectious diseases such as influenza A (H7N9) and influenza A (H1N1) virus infections of avian origin. Patients hospitalized with laboratory-confirmed infection with SARS-CoV-2 (n = 83), H7N9 (n = 36), and H1N1 (n = 44) viruses were analyzed.
Both COVID-19 and H7N9 patients had a longer lengths of hospitalization than H1N1 patients (P < 0.01), higher complication rates, and more severe cases than H1N1 patients. H7N9 patients had a higher hospitalization-to-fatality ratio than COVID-19 patients (P = 0.01). H7N9 patients had similar patterns of lymphopenia, neutrophilia, elevated alanine aminotransferase, C-reactive protein, lactate dehydrogenase and those observed in H1N1 patients, all of which were significantly different from COVID-19 patients (P < 0.01). H7N9 or H1N1 patients had more prominent symptoms, such as fever, fatigue, yellow sputum, and myalgia, than COVID-19 patients (P < 0.01). The mean duration of viral shedding was 9.5 days for SARS-CoV-2 versus 9.9 days for H7N9 (P = 0.78). For severe cases, the elapsed time from disease onset to severity was 8.0 days for COVID-19 vs 5.2 days for H7N9 (P < 0.01); the comorbidity of chronic heart disease was more common in COVID-19 patients than in H7N9 (P = 0.02). Multivariate analysis showed that chronic heart disease was a possible risk factor (OR > 1) for COVID-19 compared with H1N1 and H7N9.
- A calixarene derivative (1), with four α-l-C-fucosyl units linked by a flexible spacer, and a monomeric analog (2) with a single fucose moiety were synthesized. Compounds 1 and 2 were tested for antibiofilm activity against Pseudomonas aeruginosa (Gram-) and Staphylococcus epidermidis (Gram+). Macrocyclic compound 1 showed a very high percentage of biofilm inhibition against two different bacterial strains, while compound 2, which does not possess a macrocyclic structure, showed only moderate biofilm inhibition against P. aeruginosa and no biofilm inhibition against S. epidermidis. The multivalent fucose derivative could be a new broad-spectrum antibiofilm agent.

Author Response

We are unclear if these comments are for our manuscript.

Reviewer 2 Report

The research article entitled “Factors Associated with Mortality in Coronavirus Associated 2 Mucormycosis- Results from Mycotic Infections in COVID-19 3 (MUNCO) Online Registry” is interesting, but it needs a major revision. Below are the points which I observed while reviewing it.

1.      Table 2 is shown before table 1. Please adjust the sequence.

2.      Table 2 should be under the result section, followed by table 1 in sequences.

3.      Under table 1a following labs, there is no unit, which is creating confusion in getting the data.  

4.      % Calculation in the table is not the right way. Please consider your p-value and %. Suppose your table 1b, where “Sinus 307 (90%) 239 (92.6%) 68 (81.9%) 0.005” and “Cerebral 52 (15.2%) 17 (6.6%) 35 (42.2%) 0.001”. Please take the overall patient in one group as 100% and then calculate the % for recovery, death, and p-value or take overall 341 as 100% and then calculate the rest. It is very confusing and not calculated the right way. Please redo all calculations again.

5.      Results sections are not properly explained.

6.      Conclusion needs to elaborate.

7.      There are a lot of articles published in this area; what is new in your article?

Author Response

Response to Reviewer 2

1. Table 2 is shown before table 1. Please adjust the sequence.

Response: Thank you for letting us know, this has been edited.

2. Table 2 should be under the result section, followed by table 1 in sequences.

Response: Thank you for this comment, the table sequence has been edited.

3. Under table 1a following labs, there is no unit, which is creating confusion in getting the data.  

Response: These are now included.

4. % Calculation in the table is not the right way. Please consider your p-value and %. Suppose your table 1b, where “Sinus 307 (90%) 239 (92.6%) 68 (81.9%) 0.005” and “Cerebral 52 (15.2%) 17 (6.6%) 35 (42.2%) 0.001”. Please take the overall patient in one group as 100% and then calculate the % for recovery, death, and p-value or take overall 341 as 100% and then calculate the rest. It is very confusing and not calculated the right way. Please redo all calculations again.

Response:  We appreciate the comments, Overall, in each group is 100% and this is now reflected in the table

5. Results sections are not properly explained.

Response: Thank you for your comment. We have edited the results section to clearly display and explain the tables in the proper format.

6. Conclusion needs to elaborate.

Response: Thank you for this comment. We have edited this.

7. There are a lot of articles published in this area; what is new in your article?

Response: MUNCO dataset is a large dataset spanning 4 countries and including 341 patients in the final analysis. Findings are more generalizable. This dataset is the first ones to establish low BMI as a risk factor for mortality and perhaps more importantly refuting zinc and azithromycin as associated with increased mortality. The latter has been been suggested in earlier hypothesis.

Reviewer 3 Report

An important topic is adressed in this paper. Results are based on multicenter study and respectful data base of patients with mucormycosis.

The results are not clearly presented and need major revision. 

1. Table 2 should be placed after tables 1a to 1c. Data presented at the footnote of the table should be presented in the text of Results section.

2. The model seems overcapacitated. Tests for the evaluation of the model (does the model fit?) are nor presented as well as parameters showing the explanatory value of the model.

Table 1. Data are directly transfered from SAS output making the table difficult to read and confusing.

Column "Group" is not necessary since the types of the data are obvious. Other data should be put as a subheading of a baseline characteristic.

Remove raw No. Number associated with the characteristic suggest the presence (yes) of the characteristic.   

What does sub-speciality means? If it is ward where patient was treated the variable is not necessary.

Within one characteristics like amphotericin B regimen multiple comparisons were made and that need adjustment of p- value. The results as presented might lead to false conclusions. 

Steroid treatments are presented on tables 1a and 1c. which is confusing.

Doxycycline and other treatments are presented as subheading of corticosteroids. Multiple comparisons are repeated and the results are completely inconclusive. 

Overall, statistical analysis needs to be repeated and the results should be more focused on the research question and more judiciously presented according to good statistical practice. 

Tables should clearly presented avoiding unnecessary shading.

Author Response

Response to reviewer 3

An important topic is addressed in this paper. Results are based on multicenter study and respectful data base of patients with mucormycosis.

The results are not clearly presented and need major revision. 

1. Table 2 should be placed after tables 1a to 1c. Data presented at the footnote of the table should be presented in the text of Results section.

 

Response: Thank you for letting us know, this has been edited.

 

2. The model seems over capacitated. Tests for the evaluation of the model (does the model fit?) are nor presented as well as parameters showing the explanatory value of the model.

 

Response: We appreciate the reviewer's comment.  We agree that with 11 parameters estimated and 83 deaths, the model may be at risk for being overfit, however the usual 10 events per variable estimated rule-of-thumb may be overly conservative in some settings [1].  Our a priorimodel consisted of several important a priori confounders, along with several main exposures of interest.  We used logistic regression with Firth's correction to try to reduce overoptimism in estimated regression coefficients and compared this a priori approach with both a standard logistic regression model (with no shrinkage) and a more parsimonious model with variables selected based on univariate statistical significance.  There was no meaningful difference, and certainly no qualitative differences in our results or conclusions, therefore we presented the results from our original approach.

 

[1]  van Smeden, M., de Groot, J.A., Moons, K.G. et al. No rationale for 1 variable per 10 events criterion for binary logistic regression analysis. BMC Med Res Methodol 16, 163 (2016). https://doi.org/10.1186/s12874-016-0267-3

 

 

 

3. Table 1. Data are directly transfered from SAS output making the table difficult to read and confusing.

Response: The table is edited.

4. Column "Group" is not necessary since the types of the data are obvious. Other data should be put as a subheading of a baseline characteristic.

Response: Thank you. The tables are edited.

5. Remove raw No. Number associated with the characteristic suggest the presence (yes) of the characteristic.   

 

Response: this was edited

 

6. What does sub-speciality means? If it is ward where patient was treated the variable is not necessary.

Response: Thank you. We have deleted this.

7. Within one characteristics like amphotericin B regimen multiple comparisons were made and that need adjustment of p- value. The results as presented might lead to false conclusions. 

Response:  94.1% of overall patients got amBisome and this was proportionally higher amongst those who recovered. This was statistically significant. The number of patients getting other forms of amphotericin B were very small to deduce any meaningful results.

8. Steroid treatments are presented on tables 1a and 1c. which is confusing.

Response: Thank you. Steroid treatments have been deleted from table 1c.

9. Doxycycline and other treatments are presented as subheading of corticosteroids. Multiple comparisons are repeated and the results are completely inconclusive. 

Response: Thank you for this comment. The tables are now edited.

10. Overall, statistical analysis needs to be repeated and the results should be more focused on the research question and more judiciously presented according to good statistical practice. 

Response: we have edited the way data is now presented

11. Tables should clearly presented avoiding unnecessary shading.

            Thank you. We have edited the tables.

Round 2

Reviewer 2 Report

The author revised the manuscript but my 4th concern is not corrected yet.

The author's way of calculating percentages is not correct. Everything can't be 100%. If he is taking the overall patient 100%, then among that recovery will be 75.65 %, and the death of 83 patients will be 24.34%, not 100%. 

Similarly, in the same table for sinus patients, the number is 307, which is 90% of overall patients. While calculating the percentage of recovery among these patients, it will be 77.85% (Using 307 as 100%), not 92.6%. How can the recovery be higher than overall patients?  Please recalculate all percentages in the table.  

Author Response

1. The author revised the manuscript but my 4th concern is not corrected yet.

Response: The 4^th concern mentioned listed the percentage calculation on the tables to be incorrect. We have now revised the calculations on all tables to reflect the previous comment more accurately.

2. The author's way of calculating percentages is not correct. Everything can't be 100%. If he is taking the overall patient 100%, then among that recovery will be 75.65 %, and the death of 83 patients will be 24.34%, not 100%. 

Response: Thank you for bringing this to our attention. We have revised the tables and recalucated the percentages for all tables.

3. Similarly, in the same table for sinus patients, the number is 307, which is 90% of overall patients. While calculating the percentage of recovery among these patients, it will be 77.85% (Using 307 as 100%), not 92.6%. How can the recovery be higher than overall patients?  Please recalculate all percentages in the table.  

Response: Thank you. We have recalculated these values

Reviewer 3 Report

It is not multivariable analysis but MULTIVARIATE analysis.

Author Response

Two statistical terms, multivariate and multivariable, are repeatedly and interchangeably used in the literature, when in fact they stand for two distinct methodological approaches.¹ While the multivariable model is used for the analysis with one outcome (dependent) and multiple independent (a.k.a., predictor or explanatory) variables,^2,3 multivariate is used for the analysis with more than 1 outcomes (eg, repeated measures) and multiple independent variables. For this reason, we list multivariable analysis.