Liberal or restrictive transfusion for VV ECMO – Authors' reply

We read with great interest the comments by Yongnan Li and colleagues on our observational study of the actual level of haemoglobin, and the rate and associated outcomes of packed red blood cell (PRBC) transfusion during venovenous extracorporeal membrane oxygenation (VV ECMO). They raised important concerns about the clinical complexity of patients who undergo ECMO and suggested patient-specific evaluation of oxygenation might be preferable to simply considering haemoglobin levels for a transfusion. Indeed, the final message that emerged from the data of the PROTECMO study is quite simple: the haemoglobin transfusion threshold of 7 g/dL, which is currently recommended for non-bleeding patients who are critically ill, is also applicable for patients receiving VV ECMO, without increasing mortality.1Martucci G Schmidt M Agerstrand C et al.Transfusion practice in patients receiving VV ECMO (PROTECMO): a prospective, multicentre, observational study.Lancet Respir Med. 2023; 11: e245-e255Summary Full Text Full Text PDF Scopus (9) Google Scholar However, at the patient level, there are several important issues that must be considered. Regarding the points raised by Li and colleagues, we agree that haemolysis is a serious risk during ECMO, as shown by our 28-day mortality analysis. When blood changes from laminar to turbulent flow, erythrocytes become prone to ellipsoidal deformation, ultimately leading to cell wall destruction (ie, haemolysis). This process might adversely affect blood flow through microcirculation and the ECMO oxygenator, and there are doubts regarding the effective ability to increase oxygen delivery by PRBC transfusions, particularly when they are stored for extended periods.2Sakota D Sakamoto R Sobajima H et al.Mechanical damage of red blood cells by rotary blood pumps: selective destruction of aged red blood cells and subhemolytic trauma.Artif Organs. 2008; 32: 785-791Crossref PubMed Scopus (59) Google Scholar Transfusions might worsen haemorheological abnormalities when haemolysis has occurred and, therefore, constant monitoring and timely changing of the circuit might be more efficient. Furthermore, we appreciate the opportunity to discuss the feasibility of using red cell mass (RCM), rather than haemoglobin alone, to determine the need for transfusion. RCM could be a future target for decision making regarding transfusions under a steady condition, but it is difficult to estimate in patients undergoing ECMO, with their complex fluid shifts and multiple transfusions. Therefore, we initially calculated the RCM at baseline but, since RCM is calculated from haematocrit and BMI, there is a high level of correlation with haemoglobin. More importantly, the primary aim of the study was to study the current practice of transfusions during ECMO, and haemoglobin is the parameter used worldwide to guide decision making regarding transfusions. Nonetheless, the correlation between RCM and haemoglobin, although strong, has yet to be fully elucidated, and shows a variability that might warrant further detailed studies (appendix). Interestingly, this variability was supported when we analysed patients who were transfused (or not transfused) in different haemoglobin classes—patients transfused in the lower haemoglobin class (<7 g/dL) had a lower RCM (table). However, this finding requires more detailed confirmation, with direct measurement of RCM (even if the methods based on optimised carbon monoxide rebreathing are not validated in endotracheally intubated patients receiving ECMO). In any event, for practical reasons, RCM was not used as a criterion to evaluate PRBC transfusion in our study since there are no cutoffs validated by extensive evidence or use.TableDifferences in daily red cell mass among daily haemoglobin classesNo PRBC transfusionPRBC transfusion<7·0 g/dL13·37% (3·10)12·31% (2·10)7·0–7·9 g/dL13·91% (2·10)14·13% (2·23)8·0–9·9 g/dL16·65% (2·46)16·42% (2·22)≥10·0 g/dL20·26% (3·10)20·42% (2·90)Data are mean (SD). PRBC=packed red blood cell. Open table in a new tab Data are mean (SD). PRBC=packed red blood cell. Finally, we agree that haemoglobin is only one component of oxygen delivery, and physiologically relevant parameters must be incorporated into the latter assessment before transfusion (although we have only surrogate measures of tissue dysoxia clinically available at the bedside). Near-infrared spectroscopy and central venous oxygen saturation measurement have been considered as potential parameters to measure oxygen delivery, but there is a paucity of data that support their superiority over a haemoglobin-based approach.3Engoren M Brown RR Dubovoy A A retrospective analysis of the effect of blood transfusion on cerebral oximetry entropy and acute kidney injury.Perfusion. 2017; 32: 35-43Crossref PubMed Scopus (6) Google Scholar, 4Fischer M-O Guinot P-G Debroczi S et al.Individualised or liberal red blood cell transfusion after cardiac surgery: a randomised controlled trial.Br J Anaest. 2022; 128: 37-44Summary Full Text Full Text PDF PubMed Scopus (12) Google Scholar Moreover, considering VV ECMO physiology, we do recognise that, in a system where haemoglobin concentration is the oxygen carrier, we should principally balance the extracorporeal blood flow rate to patient's cardiac output. In conclusion, the results of the PROTECMO study reflect the current practice patterns and caution against liberal transfusions to target haemoglobin higher than 7 g/dL. In the continued search for potential ways to better assess tissue oxygenation, the future challenge will be to combine such parameters with haemoglobin to guide PRBC transfusions. Author declarations remain the same as in the original Article. Download .pdf (.31 MB) Help with pdf files Supplementary appendix Transfusion practice in patients receiving VV ECMO (PROTECMO): a prospective, multicentre, observational studyDuring VV ECMO, there was no universally accepted threshold for transfusion, but PRBC transfusion was invariably associated with lower mortality only when done with haemoglobin concentration of less than 7 g/dL. Full-Text PDF