Thank you. No financial disclosures. I want to begin by acknowledging that this work was led by Kusa Menon and Lauren Source, who were the first and senior authors on the paper that was published in Critical Care Medicine in 2022, describing the results of the review. So why did we conduct a systematic review? Scott laid out some of the reasons, but variables identifying children with sepsis and their outcomes had not yet been described in this way. This was preparation for the definition's development following the approach of the Adult Sepsis 3.0 approach, and we wanted to summarize not only what was known about identifying sepsis in children, but also what seemed to work and look at what predictors were readily available and used around the world. There were two aims of this work. The first was to identify variables associated with sepsis, severe sepsis, or septic shock in children with infection, and secondly, to find variables associated with multiple organ dysfunction or death in children with sepsis, severe sepsis, or septic shock. We would look at the following domains, demographic, clinical, laboratory, organ dysfunction, and illness severity variables. We followed standard methods for systematic reviews. We looked for studies with sepsis, septic shock, or septicemia in the title or abstract that were published between 2004 and 2020. There had to be an outcome of sepsis, septic shock, septicemia, new or progressive multi-organ dysfunction, or mortality. It had to be a case-controlled cohort or randomized trial, including children from greater than 27 weeks to less than 18 years. Admission criteria included papers that did not have values within the first 24 hours after admission, no comparator group, or sepsis criteria not specified. We used standard databases and the tool InsideScope. Screening of the title and articles and data abstraction was all done by two independent reviewers and co-authors. Any discrepancies were resolved by a third reviewer. And for papers where we needed more data, the corresponding authors were contacted twice. A quality and prognostic studies tool was used to assess for quality and bias. And because we wanted to understand the settings where this work was done, we looked at whether the countries were classified as lower income, lower middle income, upper middle income, or high income, following the World Bank classification system. To pool data for inclusion in the meta-analysis, it had to be reported in more than one study and reported in a similar way, allowing for pooling of data. Applying these criteria yielded 106 published papers. Of these, 25 had data that couldn't be pooled or had a variable that just appeared in one study. So that left 81 for the full systematic review. I've highlighted here in yellow the column describing the number of studies in the analysis and in orange the number of patients. You will note that despite applying these criteria to studies around the world, in the end in the existing literature, there were only 18 studies representing the lower middle income or lower income countries, comprising 1.8% of patients in the data set. The ICU made up the majority of study settings, 84%, and 88% of patients. Thus there were few studies that allowed us to evaluate that first aim, looking for variables among febrile children associated with the outcome of sepsis. Most of these were described in the narrative review and didn't make it into the pooled meta-analysis. Of those that did, two factors emerged as significant, decreased level of consciousness and PRISM score. As we switch to that second aim of what factors predict mortality among children with sepsis, I want to draw your attention to the mortality that we extracted from these studies. As you look on the left side of each of these figures are the high income countries. On the right side are the lower middle income countries. On the left side of your screen is the data for sepsis, the right is septic shock. And the y-axis is the proportion of children in that study who experienced mortality. You will notice not only does mortality increase as we move to the lower income settings, the confidence bars widen as well, indicating less data available. So not only do we have fewer children or more children experiencing mortality, we have less data from these settings. In this data set we identified factors associated with mortality. There were demographic variables including malnutrition, chronic conditions and oncologic conditions. Clinical variables, these shown are from the cardiovascular group, mechanical ventilation and decreased level of consciousness or GCS. Among the lab variables, acid base variables were significant, renal and electrolyte variables, hematologic, albumin, procalcitonin and ALT. Now we'll move to the organ dysfunction variables which were also significant. On each of these forest plots you'll see on the left side survivors, on the right side non-survivors where our confidence interval isn't crossing one, we have significance. So both multi-organ dysfunction syndrome and number of organ dysfunctions was associated with mortality. Similar findings for the scores we looked at including PIM 3 here and PRISM. Now I think I'm not telling you a surprising finding that markers of organ dysfunction were associated with death in this data set. But this was an important step to take. Most of these scores were derived in general ICU settings and here we took a global sepsis population across settings and found that these markers of organ dysfunction were still relevant and predicted poor outcomes. So when we look at our two aims and what the systematic review produced, we found there were few criteria associated with sepsis among children with infection that we could identify in the existing literature applying these search strategies. This highlights two things. One it highlights that although there was preceding febrile illness risk stratification research and as someone here from the emergency medicine setting this has been the major project of our work to define outcome and predict risk in febrile illness. The outcomes and definitions of sepsis had rarely been used in that work and thus wouldn't make it into this meta analysis. Additionally we found that criteria associated with mortality among children with sepsis were related to organ dysfunction both concepts and scores and some organ dysfunctions were more ominous than others. So this allowed us to identify variables commonly measured that were associated with mortality across settings and various sepsis definitions. It importantly identified differences in country income that contributed to mortality differences and differences in representation of patients in the published literature. Not only did this allow a comprehensive search for variables that would be important to include but really highlighted the limitations of the existing literature in terms of representation of both lower income settings, non-ICU settings that would be really important to address in the next phases of our work. I'd like to conclude by acknowledging all the co-authors of this paper and the task force. Thank you.

pediatric sepsis

mortality predictors

socioeconomic disparities

PRISM scores

systematic review