In this Original Research Yang and coworkers aimed to evaluate retrospectively the relationship between blood eosinophils concentrations in the early stage and mortality in critical ill patients hospitalized for AECOPD. The aim of this study is clear, introduction is well written, with updated references. Study methods are correctly explained and even if there's some limitations, these are clearly highlighted in the discussion. Sample size were good, statistical analysis in my opinion (as a clinician) is appropriate and results of this study provide new concepts in this field. However, I have a few concerns that might affect study results.

Major comments:

1) My major concern is about how COPD diagnosis was made. Considering that data come from a database, I think authors can't say precisely how COPD diagnosis was made, and in the light of the well-known problem of overdiagnosis in COPD, this could affect results. Considering that patients arrived with and acute exacerbations of course they can’t do a spirometry in that moment, but it would be interesting to know how many of these had a previous properly diagnosed COPD according to GOLD document, and how many were followed by a pulmonologist, because too often COPD is diagnosed, especially for non-pulmonologist, only considering the smoking history. Moreover congestive heart failure may mimic AECOPD or overlap it, so a previous correct diagnosis of COPD would be fundamental.

2) Table 1 shows that almost half of the subjects have pneumonia. They developed it after admission, they were admitted for it? COPD exacerbations have different definitions, in GOLD 2021 document for defining an exacerbations pneumonia should be ruled out, so this point need a clarification, and if these subjects were admitted for pneumonia another analysis on blood eosinophils considering only “pure” AECOPD should be made.

3) Only 5 subjects in the survivor category and 0 in the non-survivor had asthma (Table 1). Were it current asthma or also history of asthma in youth? I don’t think that 5 subjects may affect results, but it would be more precise to exclude those subjects. Instead I would be concerned if more subjects had history of asthma or other comorbidities that may affect eosinophils count, in this case that has to be considered precisely.

4) Authors clearly highlighted in the discussion the limitation about systemic corticosteroids. However this however this limitation could be partially overcome with some expedients: a) How many subjects were on chronic corticosteroids therapy before admission? Excluding them (theoretically they would not be a lot considering that chronic OCS is not a recommend therapy for COPD) results could be more strong. b) Considering that a single dose of corticosteroids might cause a fall in blood eosinophils count in few hours (less than 24 h) [1] and that usually in Emergency Department (ED) corticosteroids are administered early in this patients, it would be interesting to evaluate this results considering only patients with blood eosinophils measured in ED or 24 hours before admission, if possible excluding those who underwent corticosteroids at home before ED admission.

5) Blood eosinophils expressed in absolute numbers would be more precise because in COPD, especially in the last years, absolute number is considered more reliable than percentage [2], however I can’t suggest a solution for this limitation, that has been clearly highlighted in the discussion.

6) As shown in table 1, blood lymphocytes were also lower in non-survivor. It would be interesting to include that in one of the regression models.

Minor comments:

1) In table 1 legend is not clear how data are presented (number and %, median and interquartile range, etc.)

2) It would be nice to see the complete regression analysis including other variables included in the model to see the specific impact of every variable on outcomes. Obviously this would result in a large table, that might be inserted in an online supplement if possible.

Comments to the Editor: thanks for giving me the opportunity to revise this interesting article, that add important findings to the field. In my opinion it has to be considered for publication, but only after major revision, especially regarding major comments 1) and 2).

References:

1. Schleimer RP, Bochner BS. The effects of glucocorticoids on human eosinophils. J Allergy Clin Immunol. 1994 Dec;94(6 Pt 2):1202-13

2. Wedzicha JA. Eosinophils as Biomarkers of Chronic Obstructive Pulmonary Disease Exacerbation Risk. Maybe Just for Some? Am J Respir Crit Care Med. 2016 May 1;193(9):937-8.