Polymyxin B Hemoperfusion for Septic Shock -- Ready for Primetime? Early Use of Polymyxin B Hemoperfusion in Abdominal Septic Shock: The EUPHAS Randomized Controlled Trial

Background: Treatments for severe sepsis are limited, and its morbidity and mortality, especially from Gramnegative organisms, continues to be significant. Gram-negative organisms release endotoxin, which initiates inflammatory and coagulation cascades, resulting in end-organ dysfunction in patients. Polymyxin B hemoperfusion filters bind endotoxin. Objective: To assess the efficacy of polymyxin B hemoperfusion on improving hemodynamics and 28-day mortality in patients with severe sepsis from an abdominal source. Design: Prospective, randomized, controlled open-label trial conducted in 10 Italian ICUs over 3 years. Participants: 64 patients with severe sepsis who were within 6 hours of emergency surgery for intraabdominal infection. Methods: Primary end points were change in mean arterial pressure (MAP) and vasopressor requirement. Sequential Organ Failure Assessment (SOFA) scores and 28-day mortality were secondary outcomes. The trial was stopped early for efficacy after the first interim analysis. Patients were randomized to receive either conventional medical therapy with (n=34) or without (n=30) 2 polymyxin B hemoperfusion sessions. Results: Patients treated with polymyxin B hemoperfusion had a significant increase in MAP (76 to 84 mm Hg; P =0.001) and reduction in vasopressor requirement (P <0.001) at 72 hours, while those treated with only conventional medical therapy saw no change in these parameters (MAP, 74 to 77 mm Hg; P =0.37). SOFA scores also improved significantly in the polymyxin B group but not in the conventional medical therapy group. Mortality at 28 days was 53% (16 of 30 patients) in the conventional therapy group compared to 32% (11 of 34 patients) in the polymyxin B group (adjusted hazard ratio, 0.36). Conclusions: In highly selected, immediately postoperative patients with intra-abdominal infections, polymyxin B hemoperfusion improved MAP, lowered vasopressor requirements, improved SOFA scores, and reduced 28day mortality. Reviewer's Comments: At first glance, this is an exciting treatment for patients with severe sepsis. However, closer examination dampens the enthusiasm. This is a small study that enrolled 64 patients immediately postoperative for intra-abdominal infections from 10 ICUs over the course of 3 years (an average of 2 patients per ICU per year). As such, the population is highly selected, limiting the ability to generalize the results to other patients with severe sepsis. Although mortality was statistically improved, early stoppage of the trial after only 64 patients likely exaggerates any mortality difference. Furthermore, physiologic outcomes of MAP, vasopressor requirements, and SOFA scores are compared within groups (instead of change compared across groups), while the 28-day mortality is compared between groups. The overall 28-day mortality in the group treated with standard medical therapy (53%) is on the high range for systematic reports of all patients with abdominal sepsis. Finally, these hemoperfusion filters are currently available only for clinical use in a few countries, further limiting the ability to apply this treatment to patients with severe sepsis. (Reviewer-Todd W. Rice, MD, MSc).