Skip to main content

Comment on the article: “Prediction of the development of acute kidney injury following cardiac surgery by machine learning”

To the editor

We read with great interest the article published in Critical Care by Tseng et al. [1] reporting prediction of cardiac surgery-associated acute kidney injury (CSA-AKI) using machine learning (ML). Using SHapley Additive exPlanation (SHAP) analysis, the authors identified a set of intraoperative features attribute to AKI, highlighting the value of intraoperative data in AKI risk prediction.

Among 94 clinical features, intraoperative urine output (IUO) was identified as the most influential feature, which may reflect acute respond for renal function. However, a majority of patients (70.2%) received cardiopulmonary bypass (CPB). In the CPB setting, usually, IUO does not assure actual renal function as a result of cold diuresis and centrally shunted non-pulsatile blood flow [2, 3]. More importantly, intraoperative fluid balance was significantly different in patients with or without AKI (9.8 [7.7–12.8] vs. 11.8 [8.7–15.6] mL/kg/h, P < 0.001) and there was a great individual difference among two cohorts. The authors defined AKI based on the changes in postoperative serum creatinine (Scr) levels, while not adjusting the effect of fluid balance may significantly underestimate the incidence of AKI, as a positive perioperative fluid balance may dilute Scr levels [4]. Taken together, the association between IUO and AKI is complex and needs to be further evaluated after adjusting potential clinical confounders (e.g., intraoperative fluid infusion, conventional ultrafiltration, body temperature).

Typically, the benefits of ML start to become apparent when the sample size exceeds 5000–10,000. The authors validated the ML models in a relatively few patients cohort (202 patients and 49 AKI cases), with the area under the receiver operating characteristic curves (AUCs) ranging 0.781–0.843. In addition to AUC, more interpretable indicators should be introduced into the evaluation of a ML model such as accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and F1 score, as adequate discrimination (AUC > 0.8) may not imply good performance in the sensitivity and positive predictive value of the model. In particular, in the serious clinical condition like CSA-AKI, sensitivity is emphasized over specificity [5]. These parameters can be easily obtained from a confusion matrix. Therefore, if the authors were able to calculate them, we believe that the models will be more credible and will present more practical value.

The rapid development of ML techniques will certainly facilitate the management of AKI patients. The authors made an important contribution in explainable ML technology (SHAP values). We look forward to more valuable researches on CSA-AKI using advanced ML algorithms.

Availability of data and materials

Not applicable.

Abbreviations

CSA-AKI:

Cardiac surgery-associated acute kidney injury

ML:

Machine learning

IUO:

Intraoperative urine output

Scr:

Serum creatinine

AUC:

Area under the receiver operating characteristic curve

References

  1. 1.

    Tseng PY, Chen YT, Wang CH, Chiu KM, Peng YS, Hsu SP, Chen KL, Yang CY, Lee OK. Prediction of the development of acute kidney injury following cardiac surgery by machine learning. Crit Care. 2020;24(1):478.

    Article  Google Scholar 

  2. 2.

    Lannemyr L, Bragadottir G, Krumbholz V, Redfors B, Sellgren J, Ricksten SE. Effects of cardiopulmonary bypass on renal perfusion, filtration, and oxygenation in patients undergoing cardiac surgery. Anesthesiology. 2017;126(2):205–13.

    Article  Google Scholar 

  3. 3.

    Mongero L, Stammers A, Tesdahl E, Stasko A, Weinstein S. The effect of ultrafiltration on end-cardiopulmonary bypass hematocrit during cardiac surgery. Perfusion. 2018;33(5):367–74.

    Article  Google Scholar 

  4. 4.

    Moore E, Tobin A, Reid D, Santamaria J, Paul E, Bellomo R. The impact of fluid balance on the detection, classification and outcome of acute kidney injury after cardiac surgery. J Cardiothorac Vasc Anesth. 2015;29(5):1229–35.

    Article  Google Scholar 

  5. 5.

    Liu X. Classification accuracy and cut point selection. Stat Med. 2012;31(23):2676–86.

    Article  Google Scholar 

Download references

Acknowledgements

None.

Funding

None.

Author information

Affiliations

Authors

Contributions

HZ, YZ, and TTG drafted the letter. All authors read and approved the final version.

Corresponding author

Correspondence to Tongtong Gu.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, H., Zhao, Y. & Gu, T. Comment on the article: “Prediction of the development of acute kidney injury following cardiac surgery by machine learning”. Crit Care 25, 374 (2021). https://doi.org/10.1186/s13054-021-03798-w

Download citation