Evaluation of Different Control Algorithms for Carbon Dioxide Removal with Membrane Oxygenators

Round 1

Reviewer 1 Report

Dear authors,

I have read with interest your paper regarding different control algorithms for CO₂ removal with membrane oxygenator.

As a clinician, I found the manuscript very original and innovative, even if quite far from clinical practice. I summarize below the aspects that I believe could be improved.

I think that the introduction, in lines 56 to 65, does not do justice to the role of the Perfusionist, which is not limited to checking a couple of parameters, and who is presented as error-prone. Furthermore, human physiology is more complex than you report and requires several trained figures at the bed side. I can see that you attempted to simplify the extracorporeal system in order to present the algorithms you tested, but several factors which can influence gas exchange through the membrane lung have not been taken into account in your manuscript. The factors that first come to mind are Temperature, Hematocrit, Couagulation and Recirculation in veno-venous systems. The choice (and possibly its reasons) of not considering these issues in the model you built should be made explicit in the study limitations.

Modeling and control section is scientifically sound and full of details. I do not feel I have the appropriate knowledge to comment on the mathematical formulas presented, but the explanations of the three control models are clear. I would like, on the other hand, to point out that this section starts with the assumption that the compartments are homogenous at all times (line 90), which not always true in the clinics.

Results and discussion section. In this section all results are well described and the choice of the references has been done carefully. The limits of a simplified model are thus partially overcome by the support given by previous publications. Yet, the true nature of the analysed biological system, which is complex, is often overlooked. The choice of setting blood flow rate at 60 ml/min (line 429) needs further justification. In clinical practice, higher flow rates are used for CO2 removal and such a slow flow poses an increased risk of coagulation. According to which reference did you choose this value?  Moreover, a sudden change in blood flow rate modifies the pressure inside the circuit, becoming dangerous both for patient circulation and for the hollow fibers of the membrane oxygenator. Your algorithms modify blood flow, but do not take into account the pressure variations derived from the modifications in blood flow inside the circuit. Can you please explain why? Could this be one of the study limitations?

The conclusion is balanced in its recognition of some of the physiological challenges that may arise while reading the paper.

Minor revision:

-Line 67: “PI” is not written in extended form for the first time;

-Figure 3 GFR (gas flow rate) is reported as SFR in the picture, moreover I suggest a different colour for Disturbance (S), orange is too similar to red.

-Line 369 Hill et al. citation number is missing and Others has to be written as others.

Author Response

Thank you again very much for your thorough review. Please see the attachment for our answers.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear authors,

The manuscript should be accepted for publication after the authors follow the reviewer’s comments and revise the manuscript.

CO2 pharmacokinetics analysis using the 3-pool model for ECMO support is useful and will contribute to the evolution of ECMO systems in the future. However, the specific control method and parameters are subject to further investigation. In addition, there are concerns that what kind of method will be used to verify the function and safety in vitro, and then in-vivo setting, and how the safety can be guaranteed.

Best wishes

Author Response

Thank you very much for your review. Please see the attachment for our response.

Author Response File: Author Response.docx

Reviewer 3 Report

“Evaluation of different control algorithms for carbon dioxide removal with membrane oxygenators” provides a computational study comparing 3 different controllers. The authors present an adjusted model of CO2/O2 transfer in blood oxygenation/CO2 removal systems. This model is used to provide a state of this system for specific controller inputs. A description of the computational part itself and its outputs are provided. Although the motivation and need for such a controller in clinical practice to improve the monitoring of such a system is clear, it is not easy to understand the selected parameters for the study and the impact/solution for the proposed practical problem. Moreover, it is not clear from the conclusions which controller would have the best potential to be implemented with the respect to the realization. I would recommend the manuscript to be published after major revisions according to the following suggestions:

 

1.       From the introduction, the authors use volume as a measure of the gas amount. They should define from the beginning the state of gas (p,T).

2.       Line 73: “All the above-mentioned studies use the sweep flow rate as the manipulated variable”. Why do the authors believe that using both sweep flow and blood flow rates is advantageous? What are the benefits/risks of such an approach?

3.       Please provide a list of symbols and/or thoroughly describe them in the text.

4.       Fig. 1: PI regulator is presented but in the whole study PID regulator is used. Should be changed.

5.       pCO2b and pCO2pl are used (e.g. Figure 3, Line 423) but seems to refer to the same quantity. Please check and correct.

6.       There is not clear from the study what is the conclusion. Which of the controller is the most suitable for the application? Why? Can it be practically implemented or is it limited by hardware, etc.?

7.       Line 47: Are porous or rather asymmetric fibers used more often?

8.       Line 61: “serious injuries to tissues and organs” Is this valid conclusion in general or only relevant for pediatric usage as referred in the cited reference? What is the root cause for these injuries – e.g., too high/low CO2 blood saturation, …. ?

9.       Authors used the same flow ranges for blood/gas control. Nevertheless, typically for CO2 removal systems from blood, the gas flow rates are significantly higher than blood flow rates. Could authors comment on why they decided on these?

10.   Authors should move the section with their model to the initial part of the manuscript (from L369) since it does not belong to the results and discussion section.

11.   What are the risks associated with blood/gas flow rate controls? Are the outputs and regulation times relevant to the practical aspect of the problem? E.g. is it possible to regulate and safe to control the blood pump in the respect to introducing hemodynamic effects, hemolysis, …? (Line 328: “high variations in Qb might be unpleasant for the patient”, Are the authors sure that it is only unpleasant and not rather causing direct potential harm?)

12.   How relevant are the RMSE outputs from regulators in comparison to typical measurement errors of target blood saturation as a controlled parameter?

13.   Is there a possibility to use such a regulation in the current clinical practices? Is the sampling/monitoring protocol of CO2 saturation in blood suitable?

14.   More clear way of expression should be considered within a discussion section, currently it is more descriptive of the results found.

15.   In the Conclusions, there is mainly a summary of performed work, motivation, and outlook but the conclusions from work are missing. Should be rewritten.

16.   Self-citation [36] in the Conclusion is irrelevant to the concluded work and should be removed

17.   Table A2: - could authors provide more details about why the selected parameters for gas/blood volume were selected?

 

 

Author Response

Thank you very much for your review. It was not easy addressing all your points, but they were all valid and the process made the paper much stronger. I really wish more reviewers were as engaged as you are. Please see the attachment for our answers.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Dear Authors,

Thank you very much for addressing my doubts thoroughly and for stressing in the discussion the open questions as limitation of your study.

In my opinion the aim of your study: "evaluation of different control algorithms for carbon dioxide removal with membrane oxygenator" is now presented reporting a clinical part that widens the pool of readers and increases the curiosity.

In my opinion the manuscript is ready for the publication.

Author Response

Thank you so much!