Improved Survival beyond 28 Days up to 1 Year after CytoSorb Treatment for Refractory Septic Shock: A...

It is currently unknown whether CytoSorb treatment for septic shock improves long-term survival beyond 28 days from intensive care unit (ICU) admission and which factors determine outcome.

This was a long-term follow-up retrospective analysis of patients with septic shock who were treated with continuous renal replacement therapy (CRRT) + CytoSorb (n = 67) or CRRT alone (n = 49). These patients were previously analyzed for 28-day mortality. The primary outcome was the time to long-term all-cause mortality. Factors associated with time to event were analyzed both weighted by stabilized inverse probability of treatment weights (sIPTW) as well as unweighted stratified by therapy received.

The median follow-up for the total cohort was 30 days (interquartile range [IQR]: 5-334, maximum 1,059 days) after ICU admission and 333 days (IQR: 170-583) for those who survived beyond 28 days (n = 59). Survival beyond 28 days was sustained up to 1 year after ICU admission for both treatment regimens: 80% (standard error [SE] 7%) vs. 87% (SE 7%), for CytoSorb vs. CRRT, respectively, p = 0.853. By sIPTW, CytoSorb was significantly associated with long-term outcome compared to CRRT (adjusted hazard ratio [aHR] 0.59, 95% confidence interval [CI] 0.37-0.93, p = 0.025). Independent factors associated with long-term outcome in CytoSorb-treated patients were baseline log10 lactate levels (aHR 5.1, p = 0.002), age in the presence of comorbidity (aHR 2.60, p = 0.013), and presence of abdominal sepsis (aHR 0.34, p = 0.004). A lactate level above 6.0 mmol/L at the start of CytoSorb therapy had a positive predictive value of 79% for mortality (p = 0.013).

Survival is achieved with CytoSorb and CRRT for patients with septic shock beyond 28 days from ICU admission and may be improved for CytoSorb treatment. Lactate levels above 6.0 mmol/L at the start of CytoSorb therapy are predictive of worse outcome with high specificity and positive predictive value.