Maximilian Dietrich

Extracorporeal liver-support therapies remain controversial in critically ill patients, as most studies have failed to show an improvement in outcomes. However, heterogeneous timing and inclusion criteria, an insufficient number of treatments, and the lack of a situation-dependent selection of available liver-support modalities may have contributed to negative study results. We retrospectively investigated the procedural characteristics and safety of the three liver-support therapies CytoSorb, Molecular Adsorbent Recirculating System (MARS) and therapeutic plasma exchange (TPE). Whereas TPE had its strengths in a shorter treatment duration, in clearing larger molecules, affecting platelet numbers less, and improving systemic coagulation and hemodynamics, CytoSorb and MARS were associated with a superior reduction in particularly small protein-bound and water-soluble substances. The clearance magnitude was concentration-dependent for all three therapies, but additionally related to the molecular weight for CytoSorb and MARS therapy. Severe complications did not appear. In conclusion, a better characterization of disease-driving as well as beneficial molecules in critically ill patients with acute liver dysfunction is crucial to improve the use of liver-support therapy in critically ill patients. TPE may be beneficial in patients at high risk for bleeding complications and impaired liver synthesis and hemodynamics, while CytoSorb and MARS may be considered for patients in whom the elimination of smaller toxic compounds is a primary objective.

Sepsis is defined as organ failure caused by dysregulated host response to infection. While early antibiotic treatment in patients with acute infection is essential, treating non-infectious patients must be avoided. Current guidelines recommend procalcitonin (PCT) to guide discontinuation of antibiotic treatment. For initiation of therapy, there is currently no recommended biomarker. In this study, we evaluated Host-Derived Delta-like Canonical Notch Ligand 1 (DLL1), a monocyte membrane ligand that has shown promising results in differentiating infectious from non-infectious critically ill patients. Soluble DLL1 levels were measured in plasma samples of six different cohorts. The six cohorts comprise two cohorts with non-infectious inflammatory auto-immune diseases (Hidradenitis Suppurativa, Inflammatory Bowel Disease), one cohort of bacterial skin infection, and three cohorts of suspected systemic infection or sepsis. In total, soluble DLL1 plasma levels of 405 patients were analyzed. Patients were divided into three groups: inflammatory disease, infection, and sepsis (defined according to the Sepsis-3 definition), followed by the evaluation of its diagnostic performance via Area Under the Receiver Operating Characteristics (AUROC) analyses. Patients of the sepsis group showed significantly elevated plasma DLL1 levels compared to patients with uncomplicated infections and sterile inflammation. However, patients with infections had significantly higher DLL1 levels than patients with inflammatory diseases. Diagnostic performance was evaluated and showed better performance for DLL1 for the recognition of sepsis (AUC: 0.823; CI 0.731–0.914) than C-reactive protein (AUC 0.758; CI 0.658–0.857), PCT (AUC 0.593; CI 0.474–0.711) and White Blood Cell count (AUC 0.577; CI 0.46–0.694). DLL1 demonstrated promising results for diagnosing sepsis and was able to differentiate sepsis from other infectious and inflammatory diseases.