Welcome everyone. My name is Jorien de Raesveld and I'm an MD and PhD candidate at the Amsterdam UMC in Amsterdam, the Netherlands. My research focuses on transfusion thresholds and behavior in high-risk patients. I have no conflicts of interest to declare. Today I will talk to you about our study on red blood cell transfusion in veno-venous extracorporeal membrane oxygenation, a multicenter cohort study which will appear online on the website of CCM during this congress. We feel honored to be selected as a late-breaker presentation for the 2022 SCCM congress. Veno-venous extracorporeal membrane oxygenation, or VVACMO, is an important supportive therapy in severe respiratory failure. By securing extracorporeal decarboxylation and oxygenation, it provides a supportive function when other conventional therapies such as mechanical ventilation are insufficient. Although patient outcomes have been improving, mortality and complication rates in this patient population remain high. One possibly contributing factor is anemia, which can be caused by different patient, disease, and ECMO-derived factors. Currently, guidelines on hemoglobin thresholds for red blood cell transfusion in patients on veno-venous ECMO are scarce. The International Extracorporeal Life Support Organization, or ELSO, advises based on expert opinion to maintain a nematogrid level of over 40%, equivalent to a hemoglobin of 13 grams per deciliter. This is in sharp contrast with a threshold of 7 grams per deciliter, which is currently recommended in most other critically ill patients. Moreover, the European Society of Intensive Care Medicine, the ESSICOM, concludes in their guidelines that no recommendation on the optimal hemoglobin threshold in veno-venous ECMO can be made. As a consequence, thresholds applied are generally relatively liberal and show great variety, as you can see in this figure in the middle. As a result, red blood cell transfusion during veno-venous ECMO is common, with observational studies showing an occurrence rate of 67 to 100 percent and 0.3 to 2 units per day administered. Although transfusion can be lifesaving, it is also a risk-bearing intervention with substantial risks for mobility and mortality in this critically ill patient population. Thus far, only observational studies have been performed on red blood cell transfusion in veno-venous ECMO. However, extrapolation is difficult due to single-center design and small sample sizes. Therefore, our aim was to create an overview of red blood cell transfusion in patients on veno-venous ECMO by describing the proportion receiving red blood cells, the amount of red blood cells transfused, and the center's transfusion regimen. This international mixed method study was performed in 16 centers in six countries, the Netherlands, Sweden, Belgium, Australia, Croatia, and Italy. The study consisted of a combined method of a survey and a retrospective observational study. All adult patients that received veno-venous ECMO between the 1st of January 2018 and the 1st of July 2019 were included. Exclusion criteria were if the applied mode was extracorporeal carbon dioxide removal or if the support was less than 12 hours. First, to determine the transfusion practice in the different centers, a survey was developed focusing on local hemoglobin thresholds for red blood cell transfusion, rational laboratory units, and anti-coagulation strategy. Second, data was collected retrospectively using patient charts. Here you can see the data that was collected as part of the retrospective observational study. Data was collected before, during, and after ECMO and includes data on patient demographics, daily data during ECMO on laboratory values and red blood cell transfusion up to a maximum of 28 days or until decanalation, whatever occurred first. Data on ECMO characteristics and outcome parameters including successful weaning and 28-day mortality. Non-parametric data were presented as a median with interquartile range and compared using a Mann-Whitney U-test. Categorical variables were presented as percentages and frequencies and analyzed using the Chi-square or Fisher's exact test. Based on the survey, different transfusion practices were defined as the following three parts, restrictive, intermediate, and liberal. Restrictive threshold was defined as an hemoglobin level below 7.5 grams per deciliter, intermediate as between 7.5 and 9 grams per deciliter, and the liberal threshold was defined as above 9 grams per deciliter. The transfusion behavior was compared using a Crescal-Wallis test and post-HUC testing with DUN tests for multiple comparison of groups. Adjusted p-values were considered significant if they were below 0.05. This leads us to our results. In this flowchart, you can see that 230 patients received ECMO during our study period. 19 of them were excluded due to a run duration of less than 12 hours or too much missing data, and three patients were excluded since they received extracorporeal carbon dioxide removal as a mode for ECMO. In the remaining 208 patients, 88% received one or more red blood cell transfusions. On the median total duration of ECMO of nine days, transfusion was given on a median of three days, which adds up to a third of all days on phenovenous ECMO. Total median amount of red blood cells received was six units. The occurrence rate of red blood cell transfusion did not differ between patients with a restrictive and a liberal threshold. The median delta hemoglobin, defined as a difference in hemoglobin on a transfusion day minus the predefined threshold, was 0.1 grams per deciliter below the protocol's predefined threshold, therefore implying protocol adherence. However, in centers with a liberal threshold, in 78%, a transfusion was given on a day that the nadir hemoglobin level was actually higher than a protocol's threshold. Patients with a liberal transfusion threshold had more days when they received a red blood cell transfusion and received a higher amount in total per day on ECMO and per transfusion day. In the figure on the left, you see the amount of red blood cell transfused in the liberal and restrictive group per transfusion day. On the right, you see the amount of red blood cells transfused per day on ECMO. You see that as well in per day as per transfusion day, the amount is significantly higher in the liberal group. In this figure, you see the daily nadir hemoglobin level plotted over time. It is divided by if a transfusion occurred, as shown as the black dashed line, or if not, as shown as the gray line. Here, you can see that in the first week, the lowest hemoglobin level was significantly lower on the days a patient received a transfusion. After day 12, the significant difference disappears. During ECMO, 71% suffered from one or more complications. Of the 81 patients suffering a hemorrhagic complication, 98% received red blood cells. However, in the non-bleeding patients, the transfusion rate was still remarkably high, with 81% of the patients still receiving red blood cells during their ECMO support. 28-day mortality was 28%. No differences in survival were found between the transfused and non-transfused patients, and also between the different transfusion thresholds applied. We present the first international multicenter data on transfusion of red blood cells in patients on venovenous ECMO. The main finding of our study is a transfusion of red blood cells in patients on venovenous ECMO is very common, with almost 9 out of 10 patients receiving red blood cells, and the amounts received during ECMO are considerably high. Moreover, this frequency and amount of red blood cell transfusion is also found in the absence of bleeding. Lastly, variance in the center thresholds is high, although no differences were found in the complication rate and survival between the regimen. The occurrence rate and amount of red blood cells received by patients on venovenous ECMO is remarkably high in comparison with the general ICU population, where one third receives a red blood cell transfusion during their ICU stay. One explanation for this difference can be the use of quite liberal transfusion thresholds in ECMO, in comparison to the other critically ill patient groups. An existing hypothesis for the use of liberal thresholds in ECMO states that in respiratory failure, decreased oxygen diffusion and thereby decreased uptake can be expected, resulting in tissue hypoxemia. By providing a larger amoglobin buffer, it is assumed that the delivery of oxygen will be preserved and the incidence of hypoxemia will be reduced. However, this has not been demonstrated in clinical studies thus far. The present study raises the question if transfusion thresholds should be reconsidered in patients on venovenous ECMO. No interventional studies have been performed examining the effects of implementing a more restrictive transfusion threshold for red blood cell transfusion in patients on ECMO. However, in similar patient populations, such as septic shock, cardiac surgery and even acute myocardial infarction, large randomized controlled trials have shown that a restrictive transfusion threshold is safe. Moreover, during the past decades, it has become clear that transfusion carries a substantial risk for mobility and mortality. Examples are reactions such as the transfusion-related acute lung injury, TRALI, and the transfusion-associated circulatory overload, TACO, for which the population on venovenous ECMO may be extra vulnerable to develop due to the presence of several risk factors. Therefore, it might be extra important that unnecessary transfusion of blood in patients on venovenous ECMO is avoided. Beside the risk associated with transfusion exposure, blood products are also expensive, adding costs to the growing national healthcare expenses. Lastly, blood products are becoming more and more scarce, so indications for transfusion should be informed by high-quality data where possible. This study has several strengths. First, to our knowledge, this is the first multi-center and international retrospective study regarding transfusion of red blood cells in patients on ECMO. Second, a mixed-method approach was used to combine observational data with center-specific protocols. Third, it gives a complete overview by not only reporting on red blood cell transfusion, but also the daily hemoglobin level and including the threshold applied. Some limitations should, however, be recognized. A major limitation is that the chronology between the transfusion time and the corresponding laboratory values cannot be ascertained. Therefore, the direct effect of transfusion could not be evaluated. Moreover, the indications for transfusion were not recorded. And lastly, despite being a large international multi-center study, the sample size is still relatively small due to the specific patient population. Therefore, no multivariate model for the relation between transfusion and outcome could be performed. In conclusion, the occurrence rate of red blood cell transfusion in patients on venovenous ECMO is very high, even in the absence of bleeding. Although transfusion practice was usually liberal, the high variation we found in thresholds reflect the lack of evidence. No differences in survival and complications were found between the different transfusion thresholds applied. Our data supports the conduct of an additional clinical randomized control trials to determine indications for and optimal transfusion thresholds in the specific patient population. Thank you for your attention. And I would like to give a special thanks to all the co-authors stated on this slide for their help in the conduct of the study. Thank you for listening, and have a great day.

red blood cell transfusion

venovenous extracorporeal membrane oxygenation

transfusion practices

patient population

transfusion thresholds