The international, prospective COSMOS (CytOSorb® TreatMent Of Critically Ill PatientS) Registry: results from the first 300 patients.

Abstract

Introduction: Blood purification techniques are being investigated as adjunctive options in critically ill patients not only to treat severe inflammation but also to remove harmful substances such as myoglobin in rhabdomyolysis. Yet, the available evidence is limited, and further research is needed to clarify their clinical benefits.

Methods: The international prospective COSMOS Registry (NCT05146336, 23 Nov 2021) tracks CytoSorb® (CS) utilization patterns and outcomes in critical care settings. Clinical assessment was performed before, during, and after CS treatment, with a 90-day follow-up. Device-related adverse effects were reported by investigators as the safety evaluation. Data were analyzed according to a pre-specified statistical plan using descriptive statistics and paired tests to compare pre- and post-treatment values, with subgroup and safety analyses performed.

Results: A total of 300 adult patients (30.3% female, mean age 59 ± 15 years) from 22 sites were included in this analysis. The most common indications for CS therapy (multiple indications possible per patient) were septic shock (48.3%), rhabdomyolysis (12.8%), cardiogenic shock (11.5%), liver failure (11.5%), and acute respiratory distress syndrome (ARDS; 5.0%). On average, each patient received 3.3 ± 3.3 adsorbers, with 27.9% of patients receiving 4 or more adsorbers. CS was integrated in conjunction with kidney replacement therapy (75.6%), standalone hemoperfusion (7.1%), intermittent hemodialysis (IHD; 10.6%), extracorporeal membrane oxygenation (ECMO; 3.9%), and sustained low-efficiency daily dialysis (SLEDD; 4.9%). At baseline, median (interquartile range, IQR) APACHE II and SOFA scores were 24 [18, 30] and 12 [9, 15], respectively. Fluid balance improved from +1675 [141, 3348] mL pre-CS to +115 [-1100, 1495] mL post-CS, and norepinephrine requirements decreased from 0.21 [0.09, 0.40] µg/kg/min to 0.08 [0.02, 0.22] µg/kg/min (p < 0.0001 for both). Ratio of partial pressure of oxygen in arterial blood to the fraction of inspiratory oxygen concentration (P/F ratio) improved from 120 [72, 208] to 176 [115, 255] (p < 0.0001). Platelet counts decreased from 123 [76, 185] to 72 [42, 118] × 10⁹/L (p < 0.0001), while albumin levels remained stable from 2.6 [2.3, 3.1] to 2.5 [2.3, 3.0] g/dL (p = 0.112). ICU mortality was 33.1%, which was lower than mortality estimates historically associated with comparable APACHE II and SOFA scores. No serious adverse effects related to the device or device deficiencies were reported.

Conclusions: Real-world CytoSorb® use as part of standard care in critically ill patients was associated with improvements in several clinical and laboratory parameters; however, these findings should be interpreted cautiously given the observational design and absence of a control group. Observed mortality was lower than mortality estimates historically associated with established severity scores.