Red blood cell transfusion and outcome in cardiac surgery: the songs remain the same
European Journal of Cardio-Thoracic Surgery · 2023
DOI: 10.1093/ejcts/ezad119 ↗Kurz gefasst
Kommentar in der Zeitschrift European Journal of Cardio-Thoracic Surgery, 2023. Verfasst von Marco Ranucci. Beteiligte Einrichtung: Department of Cardiovascular Anesthesia and Intensive Care, IRCCS Policlinico San Donato, San Donato Milanese, Milan, Italy. Hauptthemen laut MEDLINE: Erythrocyte Transfusion, Cardiac Surgical Procedures. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 15 Literaturangaben.
Abstract (Originalsprache)
In this issue of the European Journal of Cardio-Thoracic Surgery, Tang et al. [1] provided an interesting contribution to the controversial issue of intraoperative red blood cell (RBC) transfusions as determinants of bad short and long-term outcomes after cardiac operations. There are many points of strength in this retrospective study. There is a long period of data collection (20 years) and follow-up (up to 20 years). Data are retrieved from a large national registry, whose compilation is mandatory in Denmark; additionally, they are coming from multiple institutions. This results in a very large patient population, with a minimal risk of selection bias and high reliability of the parameters collected. The outcome measure (mortality from any cause), albeit not specifically related to the cardiovascular consequences of the cardiac operation, is simple and objective. The results demonstrate a higher long-term mortality in patients receiving RBC transfusions, with a dose-dependent effect in the whole patient population, excluding patients at high-risk (age >84 years and/or EuroSCORE >8). Many previous retrospective studies found similar results [2–5]. As for any study retrospectively comparing 2 groups exposed or not exposed to a perioperative condition, the main problem remains how to deal with confounding factors that may be differently distributed between groups. This specially applies to therapeutic interventions, as RBC transfusions are intended to be, since it is quite obvious that those requiring a therapeutic intervention are in a condition of greater severity. Basically, there are 2 techniques to address this problem. The first (partially followed by Tang and associates) is to develop multivariable logistic regression models, inclusive of the main independent variable and the potential confounders. This however applies to outcome measures at a fixed interval of time (i.e. 30-day mortality). If the outcome is measured at different intervals of time, the Kaplan–Meier estimate should be corrected for confounders using Cox-regression models. The second, generally considered superior, is to develop propensity-matched groups, avoiding any other adjustment for confounders. Tang and associates did not use a propensity-score matching, nor adjusted their survival curves for the potential confounders using a Cox-regression analysis, and this, in our opinion, is an important point of weakness. Regardless of the technique used to account for risk factor differences between transfused and non-transfused patients, it is out of question that the statisticians can adjust only for data that are retrievable from databases. And the main problem is that even the most complete databases do not contain the very relevant information related to clinical judgement. It is daily clinical practice that the decision to increase the oxygen content of arterial blood and consequently the oxygen delivery by transfusing RBC is not only related to measurable factors, but often dependent on the clinical aspect of the patient (i.e. heart rate, tachypnoea, poor exercise tolerance). The main risk factor for receiving RBC during and after major surgery is preoperative anaemia or, in general, the preoperative haemoglobin level. This specially applies to cardiac surgery with cardiopulmonary bypass, where the inevitable haemodilution may decrease the haemoglobin value to the level that is considered a reasonable trigger for transfusions in each institution. And here is the major limitation of the study of Tang and associates: a number of risk factors was considered, but not preoperative anaemia. The risk assessment using the EuroSCORE does not overcome this problem, since both the EuroSCORE I and II are not inclusive of anaemia as a risk factor (that is conversely included in the ACEF II) [6]. The absence of this adjustment is disappointing, since anaemia, beyond being a risk factor for RBC transfusions, is associated with morbidity and mortality following cardiac operations [7–10] and therefore may be a major player even in deteriorating the long-term outcome. Apart from the above considerations, what remains on the ground, and what is not clarified by the study of Tang and associates, is the contrasting results between retrospective and prospective, randomized controlled trials (RCTs). If we want to shift from association to causation, the RCT is the correct way. All the existing RCTs in this field were based on different trigger values for RBC transfusion (restrictive vs liberal strategy). Regardless of the trigger values applied, in all the studies the liberal strategy group received significantly more RBC units; if RBC transfusions are determinants of morbidity and mortality, one would expect a difference in favour of the restrictive group. Conversely, no study could demonstrate an inferiority of the liberal group, actually even suggesting a trend towards more complications in the restrictive group [11–15]. There are certainly many reasons to limit the RBC transfusions during and after cardiac surgery and other major surgeries. Shortage of blood products is a direct consequence of the ageing population, with a progressive decrease of the ratio between potential donors and potential recipients; RBC transfusions are an expensive and often inappropriate therapeutic measure; and the non-inferiority of restrictive transfusion strategies justifies their application. But, at the light of the present evidence, the causative effect of RBC transfusions in terms of morbidity and mortality in the cardiac surgery scenario is far from being demonstrated.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Cardiovascular Anesthesia and Intensive Care, IRCCS Policlinico San Donato, San Donato Milanese, Milan, Italy.
ORCID-Profile: Marco Ranucci
Zitieren
APA
Ranucci, M. (2023). Red blood cell transfusion and outcome in cardiac surgery: the songs remain the same. European Journal of Cardio-Thoracic Surgery. https://doi.org/10.1093/ejcts/ezad119
Vancouver
Ranucci M. Red blood cell transfusion and outcome in cardiac surgery: the songs remain the same. European Journal of Cardio-Thoracic Surgery. 2023. doi:10.1093/ejcts/ezad119
BibTeX
@article{Ranucci2023,
author = {Ranucci, Marco},
title = {Red blood cell transfusion and outcome in cardiac surgery: the songs remain the same},
journal = {European Journal of Cardio-Thoracic Surgery},
year = {2023},
doi = {10.1093/ejcts/ezad119}
}
EUVIMED