Baseline Characteristics Associated with Good Collateral Status Using Hypoperfusion Index as an Outcome

Round 1

Reviewer 1 Report

This was an interesting manuscript which may improve our knowledge in this research area! However, I have some remarks that has to be addressed before I think it is suitable to publish it in this journal.

1.     MT (Mechanical tromboectomy) may be used after pharmacological trombolysis with alteplas/Actilyse. Was this the case or not? Whether this was done and in which cases should be stated in the methods/baseline table, since it may be of significance for the outcome such as hemorrhagic transformation etc.

2.      The timeframe of receiving the MT is not stated!?! MT within how many minutes/hours? This may be of great importance for the outcome as it may also differ between the groups analyzed in the tables.

3.      Regarding the statistical analyses exhibited in Table 1-3: there were a lot of analyses of covariates between the groups, the small groups that were divided into arterial territories (ICA/M1/proximal M2). With dozens of analyses and ”All p-values were two sided and the p value of ≤0.05 was considered to be statistically significant ”, there is a risk of type 1 errors (false positive finding). How do you address this? Executing Post Hoc Multiple Comparisons? Your p-values that are stated as significant are not very low, in fact between 0.019-0.045, therefore I am concerned that this may be a false positive finding!

4.      ”Figure 1. Demonstrates the relation between collaterals status and BMI values.” Figure 1 is in fact illustrative, but you have to exhibit it better. It would be great if you could exhibit it a bit larger and less blurry (perhaps without yellow color?) and the legend should be in the correct place.

5.     In line 134-136 and the figure, you have to improve it!! Which number does the figure have? 1 or 4? You state different numbers for this figure. Probably it should be Figure 2?!

6.      Please do a minor language/spell check, e.g. A) line 77 ”rate, and blood oxygen level measure with Sp02 at admission”, oxygen ”02” should be O₂. B) line 128-130 lacks a verb "Right side proximal M2 occlusion significantly more common in patients with poor collateral status [63.6% (7/11) vs. 11.1% 129 (1/9); p=0.028] than those with good collateral status.”

7.      The discussion is written quite clear and sound. Expand it a bit, the obesity paradox is of great interest to many doctors/researchers! The limitations stated in the discussion perhaps have to be expanded if you fail to address my remarks number 1-3 in a very good way.

Author Response

Please see the attached file responding to Reviewer 1. Thank you for your consideration. 

Author Response File: Author Response.docx

Reviewer 2 Report

 

In this study, the authors investigated the association between collateral status and various baseline characteristics in patients suffering from acute ischemic stroke secondary to anterior circulation large vessel occlusion. The design of the study is appropriate. The description of the methods requires clarifications. The results are generally clear.

The discussion is informative although possibly would require further details on potential biological reasons underlying the finding of higher BMI associated with better collateral status.

 

My specific comments are below.

 

1.     It would be good to add in the abstract the full-spelled version of AIS.

2.     “HI is defined as the Tmax > 10 sec lesion volume (Tmax>10) divided by the Tmax > 49 6 sec lesion volume (Tmax>6)”. For reproducibility purposes, more details need to be provided to describe the hypoperfusion index. How is it calculated? What does Tmax stand for and mean?

3.     The abbreviation CTA needs to be defined.

4.     Please clarify which test was used in the analysis in table 1.

5.     A quite high number of tests have been performed, and no multiple comparison correction has been applied.

6.     It would be good to include the individual datapoints in figure 1.

7.     It is unclear why for the ROC curve the BMI cutoff of 35 was chosen, while for previous analyses the authors used the BMI cutoff of 30. How many subjects in this study have BMI>35?

8.     “BMI ≤35.0 kg/m2 was identified to predict poor collateral due to high sensitivity and negative predictive value at this cut-off point (Figure 4).”. According to figure 4, the performance of the BMI cutoff used by the authors does not seem good, as most of the points lie along the diagonal of the curve, suggesting a random-guessing behavior.

9.     “Patients with higher BMI (>35 kg/m2) and older age can have a poor collateral status 180 as predicted by the HI” The conclusion is opposite to what reported in the results section, where higher BMI is associated with better collateral status.

Author Response

Please see the attached file responding to Reviewer 2's comments. Thank you for your consideration. 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

 

I thank the authors for satisfactorily addressing most of my concerns.

I have one comment related to the Bonferroni correction. The authors claim to have applied the Bonferroni correction, but the Bonferroni-adjusted significance level is not reported. Therefore, it is not possible to evaluate which p-values are below the Bonferroni-adjusted significance level.  Based on the number of tests performed, it seems evident that the three statistically significant results of this paper (i.e., BMI in “poor” versus “good collateral”, hematocrit level and the hemorrhagic transformation in the three groups defined by the affected arterial territory) will not be significant anymore with the new Bonferroni-adjusted significance level. This should be clarified in the manuscript.

Author Response

I have one comment related to the Bonferroni correction. The authors claim to have applied the Bonferroni correction, but the Bonferroni-adjusted significance level is not reported. Therefore, it is not possible to evaluate which p-values are below the Bonferroni-adjusted significance level.  Based on the number of tests performed, it seems evident that the three statistically significant results of this paper (i.e., BMI in “poor” versus “good collateral”, hematocrit level and the hemorrhagic transformation in the three groups defined by the affected arterial territory) will not be significant anymore with the new Bonferroni-adjusted significance level. This should be clarified in the manuscript.

 

We thank the reviewer for these critiques. Per the reviewer’s recommendation, we have added the Bonferroni corrected p values in Table 1 under the hematocrit and hemorrhagic transformation within 48 hrs sections. The vessel-based subgroup Bonferroni correct p values still showing significance with Hct when comparing ICA to M2 occlusions as well as M1 to M2 occlusions. Moreover, there was a statistically significant different in hemorrhagic transformation within 48 hrs between ICA and M2 occlusions while approaching significance between M1 and M2 occlusions. We have also clarified these adjusted p values within the discussion within the first paragraph (lines 162-174).

 

Regarding the BMI, Table 2 is a comparison of the baseline characteristics between the poor and good CS groups. In this instance, based on the statistical method, post hoc Bonferroni corrections were not needed after consulting statistics experts. Post hoc Bonferroni corrections were applied to Tables 1 and 3.

 

Thank you for your consideration of the manuscript.