Biological Effects of Hemoadsorption in Septic Shock (Bio-HA380)

2. Background and rationale 2.1. The clinical problem: septic shock and its refractory form Septic shock is the most severe form of sepsis and continues to carry an in-hospital mortality of between 30% and 50% despite advances in compliance with the Surviving Sepsis Campaign care bundles1. Within the spectrum of septic shock, this protocol distinguishes two clinically and biologically relevant severity strata.

The first stratum comprises patients with established septic shock as per Sepsis-3 criteria who, despite requiring vasopressor support and showing hyperlactatemia, do not meet the thresholds of refractoriness. This intermediate-severity population, proposed for inclusion to characterize whether the magnitude of the biological effect of hemoadsorption depends on the baseline inflammatory burden, is expected to present a lower circulating cytokine load at baseline than the refractory stratum.

The second stratum comprises refractory septic shock, a particularly severe subgroup in which standard resuscitation measures-guided fluid therapy, vasopressors, source control, early antibiotic therapy, and hydrocortisone-are insufficient to reverse tissue hypoperfusion and progressive organ dysfunction.

Until 2026 the definition of refractory septic shock was variable and operationally heterogeneous2, which hampered comparison across studies. Two consensus statements published in 2026 now provide a reference for defining it reproducibly:

• A recent international Delphi consensus from the SCCM/ESICM societies3 defines refractory septic shock as persistently elevated lactate concentrations and/or prolonged capillary refill time in a fluid-unresponsive septic shock patient requiring a norepinephrine equivalent ≥ 0.5 µg/kg/min and in whom clinical ultrasound has ruled out a mixed-shock component.
• The Delphi consensus of the Spanish Society of Intensive, Critical Care Medicine and Coronary Units (SEMICYUC)4 concludes that refractory septic shock is a clinical entity with persistent hypotension and signs of global hypoperfusion for more than one hour despite optimized initial treatment-including the use of hydrocortisone-with elevated lactate as a marker of hypoperfusion, and that requires advanced hemodynamic monitoring and echocardiography.

The operational integration of both definitions for the refractory stratum, together with the operational definition of the non-refractory septic shock stratum, is detailed in the study population section.

2.2. Mechanisms of action of hemoadsorption (HA) in septic shock The pathophysiology of septic shock is dominated by an uncontrolled immuno-inflammatory response with massive release of mediators, pro- and anti-inflammatory cytokines5, DAMPs (damage-associated molecular patterns) and PAMPs (pathogen-associated molecular patterns), producing vasoplegia, endothelial dysfunction, glycocalyx damage and, in many patients, a subsequent immunoparalysis phase that increases the risk of nosocomial infections and late mortality.

Endothelial damage and glycocalyx degradation are central elements in the pathophysiology of septic shock. The endothelial glycocalyx, a layer of proteoglycans and glycosaminoglycans approximately 0.5 µm thick on the luminal surface of the endothelium, regulates vascular permeability, leukocyte adhesion and the inflammatory response6, 7. During sepsis, the release of metalloproteinases, heparanase and other inflammatory mediators causes the shedding of syndecan-1, heparan sulfate and other glycocalyx molecules into the circulation7-10. This process is associated with increased capillary permeability, interstitial edema, third-space fluid leakage, and progression to multiorgan failure8, 11. Elevated plasma syndecan-1 levels correlate with greater severity of septic shock, development of acute respiratory distress syndrome (ARDS), extrapulmonary organ dysfunction, and mortality11-13. In parallel, the release of angiopoietin-2 by activated endothelial cells antagonizes Tie2 signaling, destabilizes the endothelial barrier and amplifies vascular dysfunction12, 14.

Selective modulation of the immune response through extracorporeal adsorption of medium-sized mediators (5-60 kDa) is an adjunctive strategy whose biological rationale is well established and whose hemodynamic effect has been described by multiple authors. The HA380 HA cartridge (Jafron Biomedical), specifically, uses a synthetic neutral macroporous polymer resin with high affinity for cytokines in the 10-60 kDa range, especially IL-6, TNF-α, IL-8 and IL-1015. The device is connected to an extracorporeal therapy circuit (in this protocol, always integrated into a continuous renal replacement therapy [CRRT] circuit) and operates for 4-6 hours per cartridge. Removal of proinflammatory cytokines (IL-6, TNF-α, IL-8) with HA380 also aims to reduce their effect on endothelial damage, and recent studies with other cytokine adsorbents have demonstrated the ability to remove circulating angiopoietin-212, 16. Although the specific literature on the effect of HA380 on recovery of glycocalyx integrity is limited, reducing the burden of cytokines and endotoxic mediators could attenuate the glycocalyx degradation cascade and contribute to the hemodynamic stabilization observed in clinical studies17, 18. This mechanism provides an additional biological rationale for assessing biomarkers of endothelial dysfunction (angiopoietin-2, syndecan-1, soluble thrombomodulin) as secondary variables in the present study.

The inclusion of a non-refractory septic shock stratum is mechanistically motivated: because cytokine adsorption is a mass-transfer process driven by the concentration gradient across the resin, the absolute and relative effect of HA380 may differ between a high-burden refractory phenotype and an intermediate-burden phenotype. Characterizing both strata in parallel allows this dependence to be described as an exploratory signal.

Unlike endotoxin-specific adsorbents (Toraymyxin system with immobilized polymyxin B19, oXiris with grafted heparin), HA380 is not designed for selective removal of bacterial lipopolysaccharide, since its mechanism relies on hydrophobic interactions and size exclusion of medium-range molecules. Therefore, the expected effect on circulating endotoxic activity is minimal, which precisely defines its clinical niche and guides the interpretation of the study results.

2.3. Study rationale The available literature on HA with HA38018, 20 shows hemodynamic improvement in patients with refractory septic shock, expressed as a reduction in vasopressor doses and a decrease in the Vasopressor-Inotropic Score21. However, evidence linking the hemodynamic effect with the biological and/or prognostic effect remains limited, representing a substantial gap in relevant data prior to incorporating this device into routine clinical practice.

This study is designed as a proof of concept to quantify the biological effect of two consecutive HA sessions with HA380 on cytokines, immune function, and endothelial integrity, compared with standard of care. By recruiting two severity strata in parallel-septic shock and refractory septic shock-each with its own concurrent control, the study additionally explores whether the magnitude of the biological effect depends on baseline severity, a question relevant to defining the population in which a future confirmatory trial should be conducted.

2.4. HA380-specific experimental evidence and study positioning The capacity of HA380 to adsorb mid-range cytokines has been confirmed in vitro. In a circuit comparing HA380 with CytoSorb 300 mL, both devices removed IL-6, IL-10, TNF-α, and MCP-1, although CytoSorb did so faster and to a greater extent, concentrating the bulk of adsorption within the first 120 minutes22. This finding has two implications for the present protocol: it confirms the biological plausibility of an IL-6 decrease as the primary endpoint, although the magnitude of the HA380 effect may be smaller than that described for other adsorbents, which reinforces its proof-of-concept nature and the value of evaluating an expanded panel of mediators.

In vivo and in vitro studies with HA380 mini modules show an early adsorption profile subject to progressive saturation. The 4-hour extraction rate of meropenem falls from 95% at 10 minutes to less than 20%, and that of piperacillin from 98% to 37%23. An equivalent pattern is described for vancomycin and gentamicin, with an initial extraction above 90% falling to 28% at 4 hours24, 25. Analysis of the vancomycin mass-transfer zone in cartridges with styrene-divinylbenzene sorbent confirms that adsorptive capacity is concentrated in a front that advances and is exhausted over the course of the session26, which supports this kinetic interpretation. These saturation kinetics and the rebound phenomenon described in Section 8.3 justify the protocol of using two consecutive cartridges, given that the first cartridge exhausts much of its adsorptive capacity within the first hour. The HA session scheduled at 8-12 hours aims to consolidate the biological effect and to take advantage of the per-cartridge service life set in the protocol at 4-6 hours.

Adsorption by the HA380 cartridge resin of antimicrobials commonly used in septic shock, such as vancomycin, gentamicin, meropenem, and piperacillin23-25, carries the risk of transient subtherapeutic concentrations during the cartridge's peak-uptake phase. This uptake capacity is not limited to antimicrobials: the removal of other drugs such as ticagrelor by the HA380 cartridge has also been documented27, which illustrates the nonspecific nature of adsorption and the need to monitor the concentrations of concomitant treatments. Although all available evidence comes from animal or in vitro models and has not been confirmed in patients, increasing antibiotic doses by 15 to 35% during HA is recommended23, 24, together with recording the timing of each dose relative to the HA session. Where therapeutic drug monitoring (TDM) of vancomycin or aminoglycosides is feasible, its incorporation into data collection would help document this effect. This aspect constitutes one of the future lines of research arising from this project.

Whether the biological effect translates into clinical benefit remains to be confirmed. In a retrospective observational study, HA380 incorporated into the cardiopulmonary bypass (CPB) circuit during type A aortic dissection surgery was associated with a slower rise in IL-6 (146 vs. 206 pg/mL) and a lower incidence of acute kidney injury (25.4% vs. 44.6%) and severe ARDS28. In a randomized trial in the same population, a non-significant reduction (p = 0.093) in plasma free hemoglobin during CPB was observed29. In the pediatric population, only case series and isolated case reports are available18. The present study aims to address the lack of a direct and consistent relationship between biological signal and clinical benefit.

Ongoing prospective studies with resin cartridges (HA330/HA380) in sepsis and septic shock pursue clinical or hemodynamic endpoints, for example in norepinephrine-resistant septic shock (NCT05136183) or the combination of HA380 with the oXiris membrane (HEMOX-HDF, NCT04997421). Few characterize, prospectively and simultaneously, the effect on cytokines, immune function (mHLA-DR), and endothelial integrity across severity strata of septic shock. That is the specific niche of this study. Direct visualization of the microcirculation during extracorporeal blood purification, whose impact remains uncertain30, would allow the relationship between the biological effect and tissue perfusion to be observed.

3. Hypotheses Primary hypothesis Within each severity stratum (septic shock and refractory septic shock), treatment with two consecutive HA380 hemoadsorption sessions (two cartridges 8-12 h apart, always integrated into a CRRT circuit) produces, relative to its concurrent standard-of-care control, a more desirable global outcome on a hierarchical ordinal Desirability Of Outcome Ranking (DOOR) that combines 30-day survival with the biological and clinical response at 72 h (T6), expressed as a DOOR probability greater than 0.5 in favor of hemoadsorption.

Secondary hypotheses Within each stratum, the experimental group shows a greater percentage reduction in plasma IL-6 at 24 h (T4) and 72 h (T6) than its concurrent control (the biological signal that constituted the primary endpoint in Version 2.0).

• The biological effect is accompanied by recovery of monocytic HLA-DR expression (mHLA-DR)32, 33 and by a decrease in markers of endothelial dysfunction (angiopoietin-2, syndecan-1, soluble thrombomodulin)8, 12, 14, 34, 35.
• The biological modulation translates into a more favorable hemodynamic trajectory during the first 72 h (decrease in VIS, norepinephrine dose, and lactate clearance); beyond their role as components of the primary DOOR endpoint, these individual trajectories are described in an exploratory manner with no confirmatory intent.

The magnitude of the biological effect of HA380 differs between the two severity strata, reflecting their different baseline inflammatory burden; this between-stratum comparison is exploratory and descriptive, with no confirmatory intent and no formal power to estimate an interaction.

In survivors, recovery of monocytic HLA-DR expression, as the latest immunological event, persists at the late visit (day 90 or hospital discharge, whichever occurs first), with the exact day of sampling recorded and modeled as a covariate.

4. Objectives Primary objective To estimate the overall desirability of the outcome of HA380 hemoadsorption compared with concurrent standard of care, within each severity stratum, using a pre-specified hierarchical ordinal Desirability Of Outcome Ranking (DOOR) that ranks each patient by 30-day survival and, among day-30 survivors, by the number of four pre-defined components (plasma IL-6, Vasopressor-Inotropic Score [VIS], SOFA-2 and serum lactate) that show a clinically relevant decrease from baseline (T0) to 72 h (T6).47, 48 Secondary objectives

• To quantify the biological effect of HA380 on plasma IL-6, measured as the percentage decrease between T0 and T4 (24 h) and between T0 and T6 (72 h), estimated separately within each severity stratum against its concurrent control (the former primary endpoint).
• To characterize the effect of HA on an expanded panel of pro- and anti-inflammatory cytokines (IL-8, IL-10, TNF-α, IL-1β) and on the IL-6/IL-10 ratio31, 36.
• To assess the modulation of the immune response through monocytic HLA-DR expression32, 33.
• To quantify the effect on markers of endothelial dysfunction and glycocalyx damage (angiopoietin-2, syndecan-1, soluble thrombomodulin)8, 9, 12, 14, 34, 35.
• To describe the course of prognostic biomarkers (MR-proADM, presepsin, procalcitonin)37-39.
• To describe the early hemodynamic and organ-dysfunction trajectory (VIS, norepinephrine, lactate40, SOFA-2) during the first 72 h.
• To describe, as an exploratory and descriptive objective, whether the magnitude of the biological effect differs between the two severity strata.
• To describe late immune recovery (mHLA-DR) and vital status at day 90 or hospital discharge, whichever occurs first.
• To document the safety and feasibility of the procedure. 5. Study design This protocol comprises two parallel pilot clinical trials run under a single shared protocol: one in patients with septic shock (SS) and one in patients with refractory septic shock (RSS). The design is single-center, open-label, parallel-group, and proof-of-concept. Severity stratum is a clinical stratification criterion, not a randomized factor: it is not a factorial design, and no formal estimate of a severity × intervention interaction is sought. Each stratum constitutes a self-contained randomized controlled comparison whose primary inference is drawn within the stratum.

Throughout this protocol, organ dysfunction is quantified with the updated Sequential Organ Failure Assessment (SOFA-2) score; every reference to SOFA in this document denotes SOFA-2.46 Sixteen patients will be enrolled: eight per stratum, allocated 1:1 to the experimental group or the control group (4 + 4 per stratum). Randomization is performed independently within each stratum using block randomization in blocks of 2 with sealed opaque envelopes or a simple centralized system. Owing to the visible nature of the intervention (extracorporeal circuit), the study is open label. The laboratory processing the biomarkers will be blinded to group allocation and to stratum.

The between-stratum comparison of the magnitude of the biological effect is exploratory and descriptive only, consistent with the proof-of-concept nature and the sample size.