Brain Fluid Protein Hemopexin Sparks Scientific Debate Over Sepsis Detection After Stroke

A quiet but consequential scientific exchange is unfolding in the pages of the journal Neurocritical Care, and it centers on one of the most urgent unanswered questions in modern intensive care: how can doctors know, early and reliably, that a patient battling a brain hemorrhage is sliding into sepsis? The dispute concerns hemopexin, a heme-scavenging protein found in cerebrospinal fluid, which a team of Chinese researchers has proposed as a candidate biomarker for secondary sepsis and poor in-hospital outcomes after spontaneous intracerebral hemorrhage. The exchange, published as a formal response by Tuxiu Xie, Weixiao Feng, Yu He, Yuhang Cao, and Shuixiang Deng of Huashan Hospital, Fudan University, in Shanghai, illustrates how translational proteomics findings are stress-tested by the scientific community before they can move toward the clinic.

The original study, published in August 2026, was an exploratory translational proteomics investigation. Rather than examining blood, the researchers turned to cerebrospinal fluid, the clear liquid that bathes the brain and spinal cord and that often reflects biochemical events happening inside the skull far more faithfully than peripheral blood does. Using proteomic analysis, they identified hemopexin as a molecule of interest, reporting that its levels in cerebrospinal fluid appeared to track with the later development of secondary sepsis and with in-hospital outcomes in patients who had suffered a spontaneous intracerebral hemorrhage, a devastating form of stroke caused by a ruptured blood vessel inside the brain.

Hemopexin is not an arbitrary candidate. It is the body’s highest-affinity binder of free heme, the iron-containing molecule at the heart of hemoglobin. When blood cells rupture, as they do in large quantities after an intracerebral hemorrhage, heme is released into surrounding tissues and fluids. Free heme is a double-edged molecule: it is essential for life, but in its unbound form it catalyzes the formation of reactive oxygen species, damages cell membranes, and amplifies inflammation. Hemopexin acts as a molecular sponge, sequestering heme and escorting it to receptors that clear it from the extracellular space. In the brain, where hemorrhage floods tissue with hemoglobin breakdown products, hemopexin sits at the front line of the body’s attempt to contain the damage.

That biology is precisely why the protein has attracted attention before. A 2018 study published in the Journal of Cerebral Blood Flow and Metabolism by Leclerc and colleagues reported that increased brain hemopexin levels improved outcomes after intracerebral hemorrhage in experimental models, supporting the idea that the protein is protective rather than merely a passive marker of bleeding. Against that backdrop, the new proteomic finding that cerebrospinal fluid hemopexin might signal something as seemingly distant as sepsis, a systemic infection-driven syndrome, was provocative. It suggested that the biochemical storm unleashed by brain injury and the immune collapse of sepsis might be connected through measurable molecular threads.

The letter to the editor, authored by Ahmed H and Niazi M and entitled Cerebrospinal Fluid Hemopexin as a Sepsis-Specific Signal After Intracerebral Hemorrhage, challenged that interpretation. Letters of this kind are a standard mechanism of scientific self-correction: they allow outside experts to interrogate a study’s assumptions, methods, and conclusions in public. The correspondents appear to have questioned whether the hemopexin signal truly represents a sepsis-specific phenomenon, or whether it might instead reflect more general processes, such as the sheer volume of bleeding, blood-brain barrier disruption, or the intensity of the inflammatory response to the hemorrhage itself. That distinction matters enormously, because a biomarker that merely tracks injury severity would be far less useful for clinical decision-making than one that specifically predicts a treatable complication like infection.

In their formal response, published on October 6, 2026, Xie and colleagues defended their work while engaging with the critique. The response, prepared by the same team from the Department of Critical Care Medicine at Huashan Hospital, addressed the methodological concerns raised by the correspondents. Among the references cited in the exchange is a landmark methodological paper by Riley and colleagues in the British Medical Journal on calculating the sample size required for developing a clinical prediction model, a detail that signals the debate touched on the statistical foundations of biomarker research. Prediction models, unlike simple association studies, are intended to generate risk estimates for individual patients, and the field has learned through painful experience that models built on small samples with too many variables tend to fail when tested on new patients.

That concern is especially acute in sepsis research after intracerebral hemorrhage. Patients with brain bleeds are among the most vulnerable in the intensive care unit: they may be immobile, intubated, unable to protect their airways, and fitted with invasive devices, all of which raise infection risk. Detecting sepsis early in such patients is notoriously difficult because the classic signs, fever, elevated heart rate, abnormal white blood cell counts, can be produced by the brain injury itself. A 2025 multicenter retrospective study published in the Journal of Medical Internet Research by Liu and colleagues attempted to build an explainable prediction model for sepsis in patients with intracerebral hemorrhage, underscoring both the clinical demand for such tools and the methodological rigor required to build them credibly. Against this landscape, any proposed biomarker, particularly one drawn from an invasive fluid like cerebrospinal fluid, must clear a high evidentiary bar.

The Shanghai team’s response also reflects the realities of exploratory translational science. The original study was explicitly framed as exploratory, a word that carries weight in the biomarker literature. Exploratory proteomics studies generate hypotheses; they do not, on their own, establish clinical validity. The researchers acknowledged the preliminary nature of their findings while maintaining that hemopexin deserves continued investigation as a candidate marker. Their response, like the original study, was published without dedicated funding, and the authors declared no conflicts of interest. They also disclosed that ChatGPT was used solely for English language editing, with all authors reviewing and verifying the final content, a level of transparency about artificial intelligence use that has become increasingly expected in biomedical publishing.

What makes this exchange worth watching is what it reveals about the path from a proteomic discovery to a bedside test. For cerebrospinal fluid hemopexin to become a genuine clinical tool, several hurdles remain. The finding would need independent replication in larger, ideally multicenter cohorts. Researchers would need to demonstrate that hemopexin adds predictive value beyond existing clinical variables such as hemorrhage volume, location, Glasgow Coma Scale score, and standard inflammatory markers. They would need to establish whether serial measurements, rather than a single time point, improve prediction. And because lumbar puncture is invasive and not without risk in patients with mass lesions in the brain, the practical case for a cerebrospinal fluid marker would need to be compelling, or the signal would need to be mirrored in blood.

The deeper scientific question raised by the letter, whether hemopexin is a sepsis-specific signal or a general readout of injury and inflammation, is exactly the kind of question that rigorous debate is designed to resolve. If hemopexin turns out to be a broad marker of heme burden and barrier disruption, it may still be valuable, perhaps as an indicator of secondary brain injury. If it is genuinely specific to the transition toward systemic infection, it could open a window into the poorly understood mechanisms by which catastrophic brain injury destabilizes systemic immunity. Either answer would advance the field; the current exchange ensures the question is asked properly. For now, the verdict of the scientific process is still pending, and patients with intracerebral hemorrhage, who face sepsis as one of the most dangerous complications of their condition, await the larger studies that will determine whether this heme-scavenging protein can deliver on its early promise.

Subject of Research: Cerebrospinal fluid hemopexin as a candidate biomarker for secondary sepsis after spontaneous intracerebral hemorrhage

Article Title: Response to the Letter to the Editor Entitled “Cerebrospinal Fluid Hemopexin as a Sepsis-Specific Signal After Intracerebral Hemorrhage” Regarding the Study “Cerebrospinal Fluid Hemopexin as a Candidate Biomarker for Secondary Sepsis and In-hospital Outcome After Spontaneous Intracerebral Hemorrhage: An Exploratory Translational Proteomics Study”

Article References: Xie, T., Feng, W., He, Y., Cao, Y., & Deng, S. (2026). Response to the Letter to the Editor Entitled “Cerebrospinal Fluid Hemopexin as a Sepsis-Specific Signal After Intracerebral Hemorrhage” Regarding the Study “Cerebrospinal Fluid Hemopexin as a Candidate Biomarker for Secondary Sepsis and In-hospital Outcome After Spontaneous Intracerebral Hemorrhage: An Exploratory Translational Proteomics Study”. Neurocritical Care. https://doi.org/10.1007/s12028-026-02663-4

Image Credits: AI Generated

DOI: 10.1007/s12028-026-02663-4

Keywords: hemopexin, cerebrospinal fluid, sepsis, intracerebral hemorrhage, biomarkers, proteomics, neurocritical care, heme, prediction models, inflammation, stroke, critical care