»» Thank you for that introduction. My name is Sylvia Stephanos and I'm currently a PGY-2 Critical Care Pharmacy resident at the University of Colorado. I'm originally from Houston, Texas where I graduated from the University of Houston College of Pharmacy. My training includes completing a PGY-1 at Methodist University Hospital in Memphis before pursuing a PGY-2 in Critical Care. I have no disclosures or conflicts of interest to report for this presentation. Today I'll be presenting our research which was a systematic review using trial sequential analysis to evaluate the use of restrictive resuscitation in patients with sepsis. Our study objective was to systematically assess the effects of a restrictive fluid resuscitation approach in the septic patient both during and after the initial resuscitation period typically defined as 30 milliliters per kilogram of IV crystalloid fluid. As many of us know, the Surviving Sepsis Campaign Guidelines state that for patients with sepsis-induced hyperperfusion or septic shock, they suggest that at least 30 milliliters per kilogram of IV crystalloid fluid be given within the first 3 hours of resuscitation. However, due to the lack of prospective intervention studies guiding directly or comparing different fluid volumes for resuscitation in sepsis, the most recent 2021 Guidelines downgraded this recommendation from a strong to a weak recommendation due to the low quality of evidence. Historically we've always utilized an aggressive fluid resuscitation approach of at least 30 milliliters per kilogram of IV crystalloid fluid. However, this recommendation is largely based on not only a historical precedent but also the early EGDT study that utilized 30 milliliters per kilogram within the first 6 hours of the onset of sepsis. However, that mortality benefit seen in that study was largely driven by the administration of early antibiotics rather than that fluid dose specifically. And so the recommendation to reduce that time frame down to 3 hours is based on the following studies including GRIVRS, PROCESS, ARISE, and PROMISE, which on average administrate about 30 milliliters per kilogram before randomization. And the average time to randomization was 3 hours or less in all those studies. And so that's where we get that recommendation from. Once again, no randomized controlled trials guiding that or comparing different fluid volumes. Just like any other medication, too much fluid can be associated with an increased risk of harm. And previous observational studies have suggested that increased fluid overload after the initial resuscitation period in sepsis can be associated with an increased risk of mortality. So these graphs here show that patients who had lower fluid balances in quartiles 1 and 2, shown in the green brackets, had higher rates of survival compared to those who had larger fluid balances in quartiles 3 and 4, not only during the initial 12 hours after fluid resuscitation but even up to 4 days. A previous meta-analysis was done to compare the use of restrictive versus a standard of care fluid resuscitation approach. It included six randomized controlled trials that across the board showed no difference in mortality as shown by the diamond at the bottom touching the null line. However, all these randomized trials still utilized an early goal-directed therapy of at least 25 to 30 milliliters per kilogram given before randomization. And overall, the study was under power to detect a mortality difference. We then conducted a systematic review using Cochrane-based methodology using two different search engines including MBASE and Medline given the more recent publications of larger randomized controlled trials that use effective fluid separation strategies. We evaluated the PICO question, does restrictive resuscitation or fluid restriction reduce mortality compared with the standard fluid approach in the septic patient? Our study selection used the following inclusion-exclusion criteria. We included prospective randomized trials that looked at sepsis or septic shock studies after the original 2001 EGDT study. And they had to have used a restrictive resuscitation approach that led to effective fluid differences between both arms. And for us, that was defined as at least a 500-milliliter difference or a statistically significant difference between fluid volumes. We excluded studies that had no comparator group, studies that only focused on feasibility or hemodynamic monitoring, studies that showed no difference in fluid volume between the arms of the study or if they included pediatric patients in the population. Our primary outcome was the incidence of mortality between the restrictive and standard fluid resuscitation approaches. And other secondary outcomes were mechanical ventilator days, rates of AKI or need for renal replacement therapy, ICU and hospital length of stay, duration of vasopressors, and the incidence of digital or limb ischemia. A unique aspect of our study is that we use trial sequential analysis, or TSA, as part of our statistics. And so TSA is a cumulative meta-analysis method that is used to estimate if the study effect is large enough to be unaffected by further studies. Other benefits of TSA are that it accounts for the influence of prior studies on subsequent results. It allows for multiple comparisons similar to an interim analysis of a randomized controlled trial and it also reduces the risk of Type 1 and Type 2 errors. So before we proceed, I'll briefly go over how to interpret a figure of a TSA graph. So the first line is this blue line, which is the Z-curve. And this is a cumulative effect of the studies published over time. And so each data point represents a study in chronological order and the study effect that it has cumulatively. The second are these red floating lines at the top and the bottom. And these are the boundaries for statistical significance. That means if the Z-curve crosses one of these boundaries, it suggests either a significant benefit or a significant harm. Third is this vertical line that's on the right side. And this is the required information size, meaning this is the required number of patients needed to meet power for that specific outcome. So if the Z-curve crosses the vertical line, we have achieved power to be able to assess the outcome. And lastly is the red wedge, which is the non-inferiority wedge, meaning that if the Z-curve is trending or is within this wedge, it suggests statistical futility. So if we increase the sample size or add future studies, it is unlikely to change the outcome that we're already seeing. After a systematic review of the literature, we included eight randomized controlled trials that met our inclusion criteria with the 2020 FRESH and the 2022 CLASSIC studies being the most recent additions after the previous meta-analysis that was conducted. Across the board there were only two studies that had a low risk of bias. And similar to previous randomized controlled trials in the previous meta-analysis, all studies still utilized an early goal-directed therapy approach of giving at least 30 milliliters per kilogram before randomization. And therefore, our evaluation of the use of restrictive fluids has only been able to be assessed after that initial resuscitation period. However, for all studies there was an effective fluid separation between both arms as shown in the bold values. This is the Forest Plot for a Primary Outcome of Mortality across all eight randomized controlled trials. In total there was no significant difference in mortality between a restrictive approach and the usual standard of care approach. However the weight of this outcome was largely driven by the 2022 CLASSIC study. However there was still low heterogeneity between all studies. The outcomes of the meta-analysis for mortality were confirmed by our TSA findings for this outcome. As you can see, the Z-curve does not cross any boundary for significance, meaning that there is no significant difference in mortality. However while it does not cross that vertical line, meaning we have not achieved power to assess this outcome, the addition of the 2022 CLASSIC study shows a trend towards the non-inferiority wedge, meaning that it's likely the addition of future studies or increasing sample size of these studies is unlikely to make a difference in the mortality outcome as we've seen with the recent publication of the CLOVER study. This is the TSA for the Secondary Outcome of Mechanical Ventilator Duration, which shows that we saw a significant reduction in mechanical ventilator duration with the use of restrictive fluid approach. And this was achieved with the addition of the 2020 CLASSIC study as the Z-curve crosses the boundary for significance there. It also crosses the boundary for the vertical line, meaning that we have achieved power to be able to assess this outcome. We saw a reduction in 1.25 days of mechanical ventilator duration in 7 studies that assessed this outcome. There were no significant differences in other secondary outcomes including ICU and hospital length of stay, rates of acute kidney injury or need for renal replacement therapy, duration of vasopressors or incidence of limb ischemia. There are several limitations to our systematic review and meta-analysis. The first is that a majority of the weight, nearly 50% of the mortality outcome was driven by a single larger randomized controlled trial, being the 2022 CLASSIC study. Additionally, the timing of the fluid restriction approach occurred after the initial 6 hours of early goal-directed therapy. Therefore we're unable to assess the Surviving Substance Campaign guidelines and recommendations for giving 30 milliliters per kilogram within 3 hours. Lastly, there was significant heterogeneity in the approach to fluid restriction and selection with understanding that fluid selection may play a role in other outcomes such as acute kidney injury and mortality. Strengths of our study include the focused approach on randomized trials with an effective fluid separation between the intervention and control arms. We also included studies that had protocolized fluid restriction strategies which allows for replication of these methods to be able to assess which patients are fluid responsive. And lastly, we evaluated clinically important outcomes beyond mortality such as the duration of mechanical ventilation which was significantly shorter in the fluid restriction arm of this meta-analysis. So in conclusion, our results suggest that a restrictive approach to fluid management after the initial resuscitation in the subject patient does not increase mortality but can be associated with a decreased duration of mechanical ventilation. Additionally, further study is needed regarding the restrictive approach during the initial 3-6 hours of the resuscitation period in sepsis. And lastly, the trend towards non-inferiority for mortality suggests that future studies should shift primary outcomes toward focused on other outcomes such as mechanical ventilation, rates of acute kidney injury, or renal replacement therapy. Lastly, I would like to thank major contributors to the study including Drs. Robert McLaren and Paul Reynolds for their brilliant minds and their guidance on this project. Thank you for your attention.

restrictive fluid resuscitation

sepsis

mortality

mechanical ventilation duration

acute kidney injury