The review, authored by Erika P. Plata-Menchaca of the Universitat Autònoma de Barcelona and Hamad Medical Corporation in Doha, and Enrique Hernández-Jiménez of Loop Diagnostics and the University of Barcelona, takes an unusual angle. Instead of ranking biomarkers by diagnostic accuracy or area under the receiver operating characteristic curve, the authors organize candidate technologies by their regulatory status, intended clinical use, sample type, turnaround time and workflow compatibility. The framing follows the 2026 Surviving Sepsis Campaign guidelines, which emphasize early clinical recognition with validated scores, sensible diagnostic strategies and responsible antimicrobial use and de-escalation. A biomarker, they argue, should be judged by whether it can support a concrete decision, whether to admit, escalate, start or stop antibiotics, or monitor more intensively, within the sepsis pathway.

The starting point is the familiar panel of conventional soluble markers. Procalcitonin, C-reactive protein, interleukin-6 and lactate are widely available, analytically standardized and compatible with routine laboratory workflows, and procalcitonin in particular has accumulated regulatory credentials, with multiple CE-marked assays in Europe and several FDA-cleared platforms in the United States supporting antibiotic initiation or discontinuation claims in specific populations. Lactate, although not specific to sepsis, remains central to severity assessment and shock recognition. The weakness of this entire class is biological non-specificity: CRP and interleukin-6 reflect inflammation in general, and procalcitonin concentrations are confounded by surgery, trauma, renal dysfunction, sampling time and non-infectious inflammation. These markers are useful components of assessment, the authors conclude, but they cannot capture the host-response complexity embedded in the Sepsis-3 definition of the syndrome.

A second generation of soluble and hematological markers aims to sharpen that picture. Presepsin, a soluble fragment of the innate immune receptor CD14, and mid-regional proadrenomedullin, a marker of endothelial dysfunction and vascular stress, both performed well in a meta-analysis of adult sepsis diagnostics, with proadrenomedullin slightly ahead. In practice, presepsin is framed as an infection and sepsis support marker available for near-patient testing in Europe on a dedicated chemiluminescent platform, while proadrenomedullin is more compelling for severity, deterioration risk and level-of-care decisions. Pancreatic stone protein has gone further down the regulatory road: the FDA cleared the IVD CAPSULE PSP assay on the abioSCOPE nanofluidic system in 2024 as an aid in early detection of sepsis in high-risk adult ICU patients, and it is also certified under the EU In Vitro Diagnostic Regulation. Monocyte distribution width, a measure of monocyte shape variability generated automatically by Beckman Coulter hematology analyzers, illustrates yet another implementation model, embedding a sepsis screening signal directly into the routine complete blood count for adult emergency department patients.

The most conceptually significant shift, however, is toward molecular and cellular host-response diagnostics that classify the patient’s biological state rather than detecting a pathogen or a single inflammatory protein. SeptiCyte RAPID, an FDA-cleared RT-qPCR assay running on the Biocartis Idylla cartridge platform, measures a two-gene expression signature involving PLAC8 and PLA2G7 to help distinguish infection-positive sepsis from infection-negative systemic inflammation in ICU patients on their first day of admission. TriVerity, from Inflammatix, extends the logic with a 29-transcript host-response signature read by automated RT-LAMP on PAXgene whole blood, designed to differentiate bacterial from viral from non-infectious illness and to estimate the likelihood of mechanical ventilation, vasopressors or renal replacement therapy within seven days in adult emergency patients. IntelliSep, from Cytovale, takes a biophysical route, using deformability cytometry to measure the mechanical properties of activated leukocytes and stratify emergency department patients by risk of sepsis with organ dysfunction within three days.

A further layer comes from functional immune assays, which measure what immune cells can actually do after stimulation rather than merely how much of a protein they are producing at baseline. Endotoxin tolerance, the blunting of leukocyte responsiveness to a lipopolysaccharide challenge after prior inflammatory stress, is a well-established model of immune reprogramming in sepsis, and studies of monocyte HLA-DR expression and ex vivo tumor necrosis factor-alpha responses support its clinical relevance. In a 2024 study, patients with positive blood cultures produced significantly less TNF-alpha after LPS stimulation than culture-negative patients, linking bacteremia-associated sepsis to functional innate immune suppression. That biology now underpins SeptiLoop, a CE-marked Class C near-patient test under the EU IVDR in which heparinized whole blood is stimulated with LPS and TNF-alpha is detected by lateral flow, identifying patients at risk of induced immunosuppression during suspected bacterial bloodstream infection. The review notes transparently that one of its co-authors is the founder of the company developing this test, and that it was assessed under the same descriptive criteria as every competing technology.

Artificial intelligence adds a complementary strategy: rather than relying on any single analyte, machine-learning models integrate many weakly informative signals, from vital signs and demographics to routine laboratory values and sepsis biomarkers, into risk scores. The regulated exemplar is the Sepsis ImmunoScore, an FDA-authorized software-as-a-medical-device that combines up to 22 electronic health record inputs, including procalcitonin, CRP and lactate, to assess the presence of or progression to sepsis within 24 hours in adults being evaluated for suspected sepsis. The authors caution, however, that AI is not a substitute for validated biological measurement. Its value depends on representative training data, transportability between institutions, calibration, handling of missing data, explainability, avoidance of alert fatigue, and, crucially, prospective evidence that model-guided actions improve outcomes. The most plausible future, they suggest, pairs regulated biomarkers supplying reproducible biological signals with AI integrating them into clinical context.

Underlying the entire review is a persistent translational gap. A molecule can be biologically plausible and statistically associated with sepsis yet still lack a regulated product, a defined intended-use population, validated decision thresholds, workflow compatibility, or any evidence that acting on the result improves outcomes. The regulatory landscape is also tightening and fragmenting. The EU IVDR has raised the evidentiary and quality bar for in vitro diagnostics, particularly higher-risk and near-patient assays, and the first four modules of the EUDAMED database became mandatory on 28 May 2026, increasing transparency across Europe. Regulatory status, the authors stress, must always be reported by jurisdiction and intended use: CE-IVD certification, FDA 510(k) clearance and other national approvals do not imply identical claims, populations or workflows, and the same biological signal may serve very different decisions, from early recognition and prognosis to antimicrobial de-escalation and immune phenotyping.

The authors are candid about their own limitations. The review involved no formal risk-of-bias assessment, evidence grading or quantitative meta-analysis, and comparisons across technologies are complicated by heterogeneous populations, sepsis definitions, reference standards and endpoints. Regulatory status is jurisdiction- and product-specific and changes over time, and many emerging diagnostics and AI tools still lack prospective studies demonstrating an impact on patient outcomes. Even so, the direction of travel is clear. Sepsis diagnostics are moving beyond isolated inflammatory analytes toward integrated, multi-layered host-response testing, combining soluble severity markers, hematology-derived parameters, transcriptomic classifiers, leukocyte biophysical assays and functional immune measurements. The decisive question for each new test is no longer whether its target is associated with sepsis, but whether a certified product, deployed where clinicians actually work, reduces diagnostic uncertainty at the moments when that uncertainty determines treatment, resource use and survival.

Subject of Research: Translation of sepsis biomarkers from discovery into regulated in vitro diagnostic tests

Article Title: From sepsis biomarker discovery to regulated in vitro diagnostics: evidence, certification status and implementation gaps

Article References: From sepsis biomarker discovery to regulated in vitro diagnostics: evidence, certification status and implementation gaps. (n.d.). https://doi.org/10.3389/fmedt.2026.1916860

Image Credits: AI Generated

DOI: 10.3389/fmedt.2026.1916860

Keywords: sepsis, biomarkers, in vitro diagnostics, procalcitonin, host response, FDA clearance, EU IVDR, transcriptomics, immune monitoring, artificial intelligence, antimicrobial stewardship, point-of-care testing