Evaluation of biomArkers for Quick SEpsis Diagnosis - ErASED Study

Since the "Sepsis-3" consensus statement in 2016, sepsis has been defined as a life-threatening organ dysfunction caused by a dysregulated host response to infection. Recognition of sepsis is mostly based on clinical criteria. To aid diagnosis, a wide range of biomarkers has been identified, however, with few exceptions such as procalcitonin, none is part of the routine assessment of a septic patient. Proadrenomedullin, serum calprotectin and a score of combined values of IL-6 + IL-8 + IL-10 + MCP-1 have been proposed as biomarkers to aid diagnosis and predict prognosis in sepsis, but data about their kinetics and their correlation to mortality, organ dysfunction and microbiological diagnosis is still lacking. The aim is at prospectically studying their kinetics in clinically septic patients during the first hours of their presentation in our Emergency Department, with the aim of assessing their ability to distinguish sepsis from other causes of life-threatening organ dysfunction and disease severity. Patients will be thus divided in cases (culture-confirmed infection) and controls (patient without demonstrated cause of infection). Secondary objectives will evaluate the correlation between the biomarkers' values, mortality, admission to Intensive Care Unit and organ dysfunction. In patients with confirmed infection, moreover, we will correlate biomarkers with the etiological diagnosis. The expectetion is to be able to enrol at least 120 patients in 12 months. With this number of subjects, it will be possible to detect significant differences in the mean values of the biomarkers with a power of 90% and a type I error of 5%. Analyses will produce summary indicators to synthesize the kinetics of the biomarkers including area under the curve, percentage of time spent over a critical threshold (e.g., the 75th percentile of the values), and variability indicators such as standard deviation and difference between the last and the first value. Time series among groups will be analysed with standard statistical techniques such as repeated ANOVA and advanced temporal clustering techniques. For the secondary objectives will be used the Wilcoxon test with suitable post hoc strategies to correct for multiple comparisons.