Every year, sepsis kills roughly 350,000 adults in the United States alone, according to the Centers for Disease Control and Prevention. Antibiotics and aggressive supportive care remain the backbone of treatment, but when the body’s own inflammatory response turns destructive, clinicians have few tools to rein it in. A growing body of research is now testing a different tactic: routing a patient’s blood through a specialized filter that physically strips out the proteins fueling organ failure, then returning the cleansed blood to the body. As of May 2026, no large randomized human trial has confirmed the strategy works, but early results from animal studies and small clinical cohorts have been striking enough to push several devices toward more rigorous testing.
The strongest evidence so far
The most compelling preclinical data come from a randomized, sham-controlled experiment in baboons infected with Streptococcus pneumoniae pneumonia. Led by researchers at Wake Forest University School of Medicine, the study was published in the American Journal of Respiratory and Critical Care Medicine and tested an extracorporeal blood filter coated with surface-immobilized heparin sulfate. The team ran the device for four hours alongside timed antibiotic delivery, then measured markers of organ injury. Animals that received the active filter showed reduced organ damage compared with sham-treated controls. Because the trial used a primate model and a randomized design, it carries more translational weight than typical rodent work, though it involved a small cohort and a single bacterial species.
“The heparin-coated device was able to attenuate the host response to pneumococcal pneumonia even when started after the onset of organ dysfunction,” the study’s authors wrote, noting that the filter reduced circulating markers of endothelial injury and coagulation activation alongside standard antibiotic therapy.
On the human side, an observational cohort published in Experimental and Therapeutic Medicine in 2017 found that initiating blood purification at the moment of sepsis diagnosis reduced circulating levels of HMGB1 and appeared to improve patient prognosis. HMGB1, or high-mobility group box 1, is a protein released by damaged and dying cells that amplifies inflammation during sepsis. By clearing it from circulation early, the filtration protocol appeared to blunt the inflammatory cascade before it could overwhelm organs. The study documented protocol variables such as timing from diagnosis, filter configuration, and flow rates, giving other teams a reproducible framework. Patient-outcome signals were encouraging, though the lack of randomization means confounding factors cannot be ruled out.
Separate preclinical work in rats has reinforced a broader principle: targeting even a single circulating protein can meaningfully shift outcomes. A study published in Critical Care (Remy et al., 2018) showed that binding cell-free hemoglobin, a molecule released when red blood cells break apart, with the scavenger protein haptoglobin improved survival in experimental sepsis. Taken together, these studies suggest that certain blood-borne proteins are not passive bystanders during sepsis but active drivers of organ damage, and that intercepting them is biologically plausible.
What remains uncertain
No completed large-scale randomized controlled trial in humans has confirmed whether heparin-coated or similar blood filtration devices reduce mortality in septic patients. The baboon study, while well-designed, involved six treated and six control animals and a single bacterial pathogen. Whether the results hold across polymicrobial infections, fungal sepsis, or patients with widely varying immune profiles is unknown.
A peer-reviewed analysis examining the role of extracorporeal purification in septic hyperinflammation flagged several recurring trial pitfalls. Timing is a persistent challenge: start the filter too late and the inflammatory cascade has already caused irreversible damage; start it too early and the device may strip away beneficial immune mediators along with harmful ones. The review also noted that unintended removal of protective substances remains poorly characterized, raising the possibility that aggressive filtration could blunt the body’s own recovery mechanisms.
Device-specific questions add another layer of complexity. The Seraph-100 hemoperfusion cartridge, developed by ExThera Medical, is designed to capture pathogens directly from the bloodstream. A clinical report documented bacterial capture on the device surface via microscopy and reported improved hemodynamic endpoints in septic patients. A formal randomized trial of Seraph-100 in septic shock is registered on ClinicalTrials.gov (NCT05011656) and, as of April 2026, is listed with a status of recruiting. The trial includes defined primary and secondary endpoints. Until it reports results, the device’s net benefit in a controlled setting remains unconfirmed. Separately, translational research published in Frontiers in Immunology analyzed used hemoadsorption filters from septic shock patients and identified captured endothelial injury-related markers, confirming that devices do bind clinically relevant proteins. But whether removing those specific molecules translates into better patient outcomes has not been established in outcome-driven trials.
How to read the evidence
Three tiers of evidence are emerging in this field, and distinguishing among them matters.
The first tier consists of primary experimental studies with controlled designs, such as the baboon heparin-filter trial and the rat haptoglobin study. These provide the most direct evidence that targeting specific blood proteins can reduce organ injury or death, but they are limited by species differences and small sample sizes. The second tier includes human observational data, like the HMGB1 cohort, which offers real-world clinical signals but cannot establish causation. The third tier is translational and review literature, which synthesizes what is known and identifies gaps but does not generate new patient-level evidence.
One practical distinction is worth tracking: the difference between removing a pathogen and removing a host protein. Some devices, like the Seraph-100 cartridge, are engineered primarily to capture bacteria and other pathogens directly from the blood. Others, like the heparin-coated filter tested in baboons, target the body’s own inflammatory molecules. Both strategies aim to reduce sepsis severity, but they carry different risk profiles. Pathogen-capture devices risk missing organisms that have already seeded tissues. Protein-removal devices risk depleting molecules the immune system needs. The optimal approach may eventually combine both functions, but no trial has tested that combination head-to-head against standard care.
There is also an important gap between biochemical success and clinical benefit. Many early studies report impressive reductions in circulating markers such as HMGB1, cell-free hemoglobin, or endothelial injury proteins. Yet lowering a blood marker does not automatically translate into fewer days on a ventilator, less kidney failure, or improved survival. Regulators and guideline committees will ultimately require hard outcomes, including mortality, intensive care length of stay, and long-term functional recovery, before endorsing widespread use.
Where patients and families stand right now
For clinicians and families watching this space, the essential picture is this: blood filtration for sepsis has moved beyond theoretical plausibility into early-stage testing with measurable biological endpoints. The current evidence supports the idea that selectively removing damaging proteins, or even circulating pathogens, can alter the course of experimental sepsis and may improve some clinical signals in humans. But the absence of large, well-controlled trials means these devices should still be viewed as investigational, not established standard of care.
Patients who encounter these technologies today are most likely to do so through clinical trials or carefully considered compassionate-use decisions. In those settings, questions worth raising with the medical team include: Which molecules does the device target? How soon after diagnosis will treatment begin? What outcomes is the team tracking? And how does the intervention fit alongside antibiotics, fluids, and organ-support therapies?
As results from ongoing randomized studies become available, the picture should sharpen. If trials confirm that early, targeted blood purification meaningfully reduces organ failure or death, extracorporeal filtration could join antibiotics and source control as a core pillar of sepsis management. If benefits prove narrow or inconsistent, these devices may find a niche role in specific subgroups, such as patients with extreme inflammatory responses or unusually high levels of particular circulating proteins. For now, the science supports cautious optimism: the bloodstream is emerging not just as a conduit for infection but as a therapeutic target in its own right, and researchers are learning how to edit its contents in real time to tilt the balance toward recovery.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
