Advanced techniques in blood purification

The dysregulated immune response in sepsis

Pathophysiology of the immune response in septic shock is usually described in two phases. During the first phase, a massive dysregulated proinflammatory component, also called a cytokine storm, is associated with tissue damage, organ injuries, and early mortality. During the second phase, the predominant antiinflammatory component, triggering immunoparalysis, is responsible for nosocomial infections, viral reactivations, and late mortality.

During pathogen invasion (either bacteria, viruses, fungi, or parasites), numerous molecular signals are activated, driving the immunoinflammatory response. These alerts are triggered by pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) ^, . PAMPs are described as the “molecular signature” of pathogens. They are not produced by the host, but they are expressed at the pathogen surface and include endotoxins at the surface of gram-negative bacteria, double-stranded ribonucleic acid (RNA) produced by most viruses, and beta-glucans and mannans found in fungi. DAMPs are endogenous molecules, such as high-mobility group box 1 or extranuclear deoxyribonucleic acid (DNA), released during tissue damage. ^, DAMPs are released during sepsis but also during specific conditions such as burns, trauma, surgery, or acute pancreatitis, triggering a sterile inflammatory response. Both PAMPs and DAMPs are recognized by pattern recognition receptors (PRRs), which are ubiquitous on the surface of innate immune cells. Importantly, activation of PRRs by PAMPs or DAMPs leads to the appropriate synthesis of cytokines and the immune response. However, in septic shock, the inflammatory response is dysregulated and leads to a massive release of cytokines (interleukin [IL]-1, IL-17, IL-6, tumor necrosis factor [TNF]-alpha) and an intense activation of complement.

Concomitantly, immunosuppressive patterns also appear, mainly consisting of a reduced antigen-presenting cell function and reduced human leukocyte antigen–DR isotype (HLA-DR) expression on monocytes (mHLDA-DR), which correlates with mortality and lymphocyte apoptosis. Lymphopenia is also a marker of immunosuppression in sepsis and has a poor prognosis. The massive production of the antiinflammatory IL-10 cytokine participates in sepsis-induced immunosuppression as well.

Immunomodulatory therapies and targets for extracorporeal blood purification

Along with antibiotics, controlling the site of infection, and symptomatic treatment, numerous intravenous treatments targeting specific pathways such as IFN-γ, rhIL-7, and TNF-alpha have been tested to modulate the immunoinflammatory cascade in sepsis. Unfortunately, all these treatments have failed to demonstrate any benefits on morbidity and mortality. Again, some authors have highlighted sepsis “heterogeneity” with regard to the site of infection, pathogens, genetic background, and premorbid conditions. Other experts have argued that all these molecules target a specific pathway of the immune response, whereas a more global approach might be more efficient. Based on this second hypothesis, extracorporeal blood purification techniques have been developed to control the dysregulation of the immune system with the removal from the blood a large panel of components that participate in the inflammatory response.

Several extracorporeal techniques have been developed during the last two decades. They target different mediators of the immunoinflammatory response, which can be endotoxins, cytokines, proinflammatory cells, or pathogens such as bacteria and viruses. Most techniques are developed towards one particular target, but some are able to remove two or more mediators (e.g., cytokines and endotoxins) ( Fig. 98.1 ). Contrary to intravenous treatments, blood purification techniques are nonselective.

As this immunoinflammatory response seems to represent a “final common pathway,” these techniques may be proposed in sepsis, but also in trauma, burns, pancreatitis, and other inflammatory situations causing a dysregulated immunoinflammatory response (see Fig 98.1 ).

Endotoxin removal

Two major hemoperfusion devices are available for endotoxin removal. First, the adsorption properties of polymyxin-B have been used in a hemoperfusion cartridge named Toraymyxin (Toray, Tokyo, Japan) since 1993. Polymyxin-B is an antibiotic with great affinity for endotoxin, thanks to ionic and hydrophobic interactions. However, its intravenous use is impossible because of renal and neurologic toxicities. In this hemoperfusion cartridge, polymyxin-B is covalently bound to the fibers. The clinical effects of a polymyxin-B hemoperfusion treatment (typically two 2-hour hemoperfusion sessions 24 hours apart) have been widely studied over the last two decades. Interestingly, this technique represents the current standard of care for abdominal septic shock in Japan. In 2007 Cruz and colleagues reviewed the performance of this hemoperfusion cartridge on septic patients, based on evidence gathered from 1998 to 2006. They found a positive effect on mean arterial pressure (MAP), oxygenation, and a decrease in vasopressor use, to the point that the risk ratio for mortality was 0.53 (95% confidence interval [CI] 0.42–0.65) in the polymyxin-hemoperfusion group. This study was an incentive for further RCTs, mainly focusing on gram-negative abdominal sepsis because of likely high levels of circulating endotoxin in this clinical context. The EUPHAS trial showed encouraging results, consistent with those previously reported by Cruz and colleagues’ meta-analysis. This study was even stopped early because of a significant improvement in 28-day mortality (adjusted hazard ratio [HR], 0.36; 95% CI 0.16–0.80; P = 0.01) after the intermediate analysis. The ABDOMIX trial was conducted subsequently and included a greater number of patients with similar inclusion criteria. However, it did not find such promising outcomes, with, conversely, a nonsignificant increase in 28-day mortality (odds ratio [OR], 1.5872; 95% CI 0.85–2.93; P = 0.14) in the intervention group. Importantly, the mortality rates in the control group greatly differed in these two studies. Lately in 2018, the randomized blinded EUPHRATES trial included a selected population of patients presenting with high levels of circulating endotoxin activity (defined as an endotoxin activity assay [EAA, no unit] ≥0.60) rather than empirically suspected gram-negative sepsis (e.g., of abdominal origin) . The results of this large-scale analysis ( n = 450) failed to demonstrate a reduction in 28-day mortality. It was nevertheless later speculated that patients with EAA ≥0.90 could bear an endotoxin burden exceeding the adsorption capacities of Toraymyxin. A post hoc analysis was performed, focusing on the subpopulation of patients with EAA between 0.6 and 0.9. In this particular subgroup, 28-day mortality was lower in the hemoperfusion group (OR, 0.52; 95 % CI 0.27–0.99; P = 0.047). Evidence was recently found that Toraymyxin hemoperfusion also showed immunomodulation properties, improving mHLA-DR expression in septic patients. This latest input supports the need for further research focusing on a specific selected population and optimal timing for polymyxin-B hemoperfusion initiation.

Another technology aimed at removing circulating endotoxin is the Alteco LPS adsorber (Alteco Medical AB, Lund, Sweden). This cartridge contains a polyethylene matrix covered in a specially designed synthetic peptide engineered for endotoxin adsorption thanks to high affinity for the lipid A moiety of the endotoxin. In a nonrandomized fashion, Adamik and colleagues tested this device on a few selected patients ( n = 18) and were able to demonstrate an improvement from baseline in terms of EAA levels and clinical outcomes but no difference in ICU mortality between groups. The ASSET trial (NCT02335723) designed as a feasibility study of this device was unfortunately stopped because of patient recruitment issues in 2017.