Thank you so much for your attention and really thank you to the organizers for the opportunity. I feel like we need a seventh inning stretch or something right now at the end of the day. So thanks for staying and hearing my talk. I'm Deb Siegel. I'm presenting on behalf of the Stratus Trial Investigators, and I just want to say that this work represents an amazing collaboration between hematology, critical care, and laboratory medicine. And the work was coordinated at the Population Health Research Institute, which is a research institute affiliated with McMaster University. These are my disclosures. I received honoraria indirectly to my institution for work that's unrelated to this talk, and also I should disclose I'm a hematologist, but I'd like to say that I'm a critical care adjacent hematologist. So this audience, of course, is familiar with anemia as a common complication of hospitalization, and patients who are admitted to the intensive care unit are particularly high risk for anemia. And you can see here from these data that after a week in the ICU, almost everyone has some degree of anemia. Of course, anemia and critical illness is multifactorial in nature and quite complex, but it does lead to the administration of red blood cell transfusions in up to 40% of patients. And interestingly, half of those transfusions are actually given without evidence of active hemorrhage. We know and we heard from a previous speaker that both anemia and transfusion lead to adverse outcomes, including higher mortality, longer ICU and hospital stay in transfusion reactions. And so iatrogenic blood loss from laboratory testing here represents a potentially modifiable contributor to anemia and transfusion. And we wondered, and others have as well, if we can reduce the volume of blood that's taken for laboratory testing, can we impact anemia and red blood cell transfusion? In the intensive care unit, frequent blood sampling actually does lead to substantial loss of blood. We did an observational cohort study of over 7,000 patients admitted to multiple ICUs, showing volumes over 200 mils per ICU stay of blood that was collected between day two and day seven. And for each additional 150 milliliters that were collected, the risk of red blood cell transfusion was increased over twofold. Another very well-known large observational study in the United States showed collection volumes of 41 milliliters per day, which is like donating a unit of blood every eight days. But what's important to recognize is that the vast majority of blood that is collected into the tubes is actually discarded as waste in the lab. Less than 10% of the blood that's collected is used for the actual testing. So here we have an opportunity. I'm just going to pause here to describe how blood collection tubes work. They are evacuated to create a vacuum inside the tube, which automatically draws blood from your patient, whether it's venipuncture or from a line. Smaller volume tubes, or so-called soft draw or short draw tubes, actually have less vacuum, so they fill to a lower predetermined volume. That's automatic. They have the same cost and the same physical dimensions as the standard volume tubes, and they can be used on the same analyzers. The picture on the left-hand side of the screen shows two citrate tubes. The one on the left with the opaque cap is 2.7 mils, and the one on the right with the translucent cap is 1.8 mils. Given all of these characteristics, we find that these tubes aren't used routinely in adults. What I want to mention as well, because there are sometimes questions about the smallest volume tubes, which are the microtainer tubes, which you see on the lower right-hand side of the screen. Those are capillary tubes, so they fill not by vacuum but by capillary, and we did not use these tubes for our trial because they were not thought to be implementable on a large scale. They have to be manually aliquoted both at the bedside and in the lab. So blood sampling, here we have it. Blood sampling contributes to anemia and transfusion. We know that 90% of the blood that's collected from patients is discarded as waste. We have smaller volume tubes, which are available, compatible, and have the same cost, but they're not used routinely. And so we thought that a randomized control trial evaluating effect on a clinical outcome was needed to change practice. So our hypothesis was that transition to small volume tubes, small volume blood collection tubes will reduce red blood cell transfusion in ICU patients. And in order to test this hypothesis, we conducted a trial. The slide shows the study design. STRATUS was a stepped wedge cluster randomized trial in which the unit of randomization was the intensive care unit. ICUs were eligible if they were adult medical surgical ICUs with 14 or more beds and had the capacity for invasive mechanical ventilation. All the sites had to use standard volume tubes, although the types of tubes and the volumes were variable. And all sites had to be able to collect and transfer electronic data from administrative and laboratory sources because that's the only type of data that we collected for this study. You can see on the left-hand side in the diagram, this is the sort of study diagram showing the randomized control trial. At the beginning of the trial, all of the sites used the standard tubes to which they were already using as part of routine care, and that's shown in white. At six weeks intervals, two ICUs were randomized to switch to the intervention, which was to transition to small volume tubes. I should say that the randomization schedule here was blinded to the study team and also to the investigators. So by the end of the trial, you can see in light blue, all sites had transitioned over to small volume tubes, which is the intervention. The dark blue lines indicate a one-week washout period during which the small tubes were in use, but we excluded data from those weeks. In this trial, all patients admitted to the ICU were included in the study, so we collected data from everybody that was admitted. The study was conducted under a waiver of individual participant consent, and we collected electronic data up to 30 days or until hospital discharge or death. The intervention here was the transition to small volume tubes, which I mentioned. On the left-hand side, you'll see the standard volume tubes ranged in volume from four to six mils, whereas the small volume tubes ranged in volume from 1.8 to three mils, and you can see the types of tubes shown in the colored arrows. I should mention that we did not include blood culture tubes or blood gases for this study. Because the study was blinded, the randomization schedule was blinded, two weeks ahead of the switch date, the site was notified that they were going to switch, and the nursing staff and the laboratory staff received brief targeted educational interventions, which we had tested in a pilot study. And on the date of the switch, all of the standard volume tubes were removed from ICU storage areas in the ICU itself and were replaced with the small volume tubes. So the outcome, our primary outcome for this study was the number of red blood cell units transfused per patient during the ICU admission, and this was among patients who were admitted to the ICU for 48 hours or longer, and we pre-specified this as part of our sample size calculation as this was the population that we felt would be most likely to benefit from this intervention, as they would be most likely to experience the effect of repeated blood loss. Key secondary outcomes shown here, although there were more, included the number of insufficient specimens or specimens which were returned as reported, unable to analyze because of insufficient volume, the proportion of patients who received red blood cell transfusion, a change in hemoglobin from ICU admission to ICU discharge, adjusted for red blood cell transfusion, ICU and hospital length of stay, and ICU and hospital mortality. Our primary analysis included patients admitted to the ICU for 48 hours or longer. What I want to mention here is that during the study in March of 2020, of course, the COVID pandemic started. As a result of the COVID pandemic and changes and disruptions in healthcare and research, we were forced to halt the trial for about five months. At that point, seven sites had not transitioned over to the intervention. So during that period of time, which we call the transition delay, all of the sites continued to use the tubes to which they were allocated at that time, and we continued to collect data during that time period. In August of 2020, we resumed transitions at sites, and all of the sites ended up switching over. So a big challenge for a stepped wedge cluster randomized trial, which it's very important that nothing changes throughout the trial, and this was a big issue for us. So as a result of this, and some concerns about anticipated differences in the patients that were admitted during that time period, and also that the care that they received, we modified the primary analysis to exclude the patients that were admitted during that five-month transition delay during the early or height of the COVID pandemic. And this was, of course, done in a pre-specified manner without reviewing or any data. In fact, we hadn't even received all of the data from the sites at that point. Our analysis, we used a negative binomial mixed model, a hierarchical model, where the study periods or steps were modeled as a fixed effect, and the ICUs were modeled as a random effect, and we included duration of ICU stay as an offset. The analyses were also adjusted for age and sex. We evaluated for the presence of temporal trends, and also for the effect of the COVID pandemic by including interaction terms. So with regards to the results, 58 ICUs were assessed for eligibility. When the eligibility assessment was completed, three sites had already, were using small volume tubes, and were therefore excluded. Twenty-eight sites declined participation for unknown reasons, and two sites had practices that weren't aligned with the protocol. So we included 25 ICUs at varying geographical locations within Canada at both community and academic centres. The average number of patients per ICU was about 1,100, and the average number of patients per period was 1,400. In total, just over 51,000 patients were registered in the trial during the study period. 23,000 of those were admitted to the ICU for less than 48 hours. We excluded an additional 305 patients who were admitted during that washout period, and one patient requested removal of their data. So in the end, we had 27,411 patients who were in the ICU for 48 hours or longer, 6,200 of whom were admitted during that transition delay period. This slide shows the baseline, some selected baseline characteristics of the patients. There were more, obviously, which will be published when the study is published. You can see here in the middle two columns in dark blue, before and after transition, and in the right-hand side in lighter blue, patients admitted during that transition delay and those admitted during other periods. The average age was in the early 60s, and 40% were women. There was a variable diagnosis, according to most responsible diagnosis by ICD codes, that was supplied by hospitals through their administrative databases. There were some imbalances between groups, either in before and after transition, but some imbalance between transition delay and other periods, and that was reflected by standardized difference of greater than 0.1. And we, I should say, did do sensitivity analyses where we looked at the effect and adjusted for these imbalanced covariates. So this slide shows the primary outcome, which was the number of red blood cell units transfused per patient per ICU stay. 30% of the patients in our study who were admitted for 48 hours or longer received at least one red blood cell transfusion. This table shows the least squares mean, which is, of course, the mean from the model, which accounts for the stepped wedge design and also is adjusted for the covariates. You can see that in the dark blue column for all patients, all 27,000 patients, you can see that there was a reduction in the number of red blood cell transfusions after transitioning to small volume tubes, with a mean difference of minus 0.10. So this corresponds to an absolute mean difference of 9.84 units per 100 patients. In the population of patients, or in the analysis of patients where we excluded those admitted during that transition delay, those 6,200 patients that were admitted at the height of COVID, you can see that there was a reduction in the number of red blood cell units transfused after transitioning to small volume tubes, but this was not statistically significant and corresponded to an absolute mean difference of 7.24 red blood cell units per 100 patients. The slide shows the proportion of specimens with insufficient quantity, which we thought was sort of like our primary safety outcome here, and we're showing the results of over 500,000 specimens of EDTA and sodium or lithium heparin tubes. You can see that overall, the proportion of specimens that were reported as insufficient in quantity were very low, like 0.03% or less, and were similar between both groups in both populations. This slide shows the change in hemoglobin from ICU admission to ICU discharge. On the top row, you can see the change in hemoglobin adjusted for red blood cell transfusions, and there was a different, the decrease in hemoglobin from ICU admission to ICU discharge was smaller after transitioning to the small volume tubes compared to the standard volume tubes, and that was in the all-patient population. However, in the population of patients in which the transition delay patients were excluded, that difference was not statistically significant. When we looked at the difference in hemoglobin from admission to discharge without adjusting for red blood cell transfusion, we saw that the decrease in hemoglobin was smaller in both populations, it's smaller with small volume tubes in both analyses. And similarly, we saw the decrease in hemoglobin from admission to discharge was smaller after transitioning to small volume tubes in patients without transfusion. So in conclusion, we showed that transition to small volume tubes for blood collection in ICUs reduced red blood cell transfusion by 10 units per 100 patients. We showed that the transition to small volume tubes lessened ICU-related reductions in hemoglobin, but did not affect lab testing with regards to specimens with insufficient quantity. The small volume tubes were implemented within routine clinical practice within different practice settings in Canada, and these results suggest that the widespread use of small volume tubes, which are a simple, available, cost-neutral intervention, could reduce transfusion in ICUs and support efforts to sustain a blood product supply. I just wanted to end with some discussion points, which I think maybe will come out in the Q&A. This was an intervention that we implemented into routine clinical practice with brief targeted education, which suggests that this is actually scalable. We used pragmatic data collection, which I guess could be in the strengths or limitation section because, you know, obviously there were some limitations in terms of what we could collect. For example, we wouldn't be able to collect quality of life data. We did our trial in community and academic centre, which speaks to the generalizability of our findings, and we showed that there was a small effect, really, at the individual level, but potential for impact at a health system level. An obvious limitation here, of course, was the COVID pandemic, which led to changes in the study conduct, where we had to delay transitions, and also changes in the primary analysis, although all the analyses that I've showed you were pre-specified key secondary analyses. However, we believe that the totality of the findings, particularly the consistency in the magnitude and direction of the effect that we saw with small volume tubes, in addition to the change in hemoglobin or the mitigation of the hemoglobin reduction in ICU, suggests that, in fact, there is an effect of small volume tubes on red blood cell transfusion. And, of course, we only use one brand here of tubes, but they are available worldwide from different manufacturers. So just for illustration, in this trial, which was conducted in 25 ICUs in Canada over almost two years, 36,000 red blood cell transfusions were administered. And if we showed a reduction in red blood cell transfusions of 10 per 100 patients, we actually saved 1,500 units of red blood cells in this trial. So you think about both the cost and the resource implications of that, it's pretty amazing. And then, of course, you know, this is a huge issue, particularly now blood supplies are very tenuous in the U.S. and in the U.K. and in Canada and elsewhere. So this is an important consideration. And I just wanted to acknowledge the amazing team that led to the completion of this trial. And I'd like to always tell the trainees that it takes a village to raise a study. And this was our village, which was amazing. And so at least one of them is in the audience. So thank you to all who participated in our funders. And I'm happy to take questions. Thank you.

Stratus Trial

small volume blood collection tubes

intensive care units

red blood cell transfusions

ICU stay