The study, published in Clinical Research in Cardiology, followed 113 patients treated with an Impella microaxial flow pump between January 2021 and December 2023. All patients had cardiogenic shock classified as stage C or higher on the Society for Cardiovascular Angiography and Interventions scale, meaning their circulatory failure was severe enough to require mechanical assistance. The researchers measured four cytokines, interleukin-1β, interleukin-6, interleukin-8, and tumor necrosis factor-α, at five predefined time points: at device implantation and then at 24, 72, 120, and 168 hours afterwards. Serum samples were analyzed using chemiluminescent immunoassays on a Siemens IMMULITE 2000 XPi system, allowing precise quantification of even low cytokine concentrations.

The biological rationale for the study rests on a growing recognition that cardiogenic shock is more than a plumbing failure. Myocardial ischemia and reperfusion release damage-associated molecular patterns that activate innate immune pathways, triggering a flood of proinflammatory mediators. These cytokines, including interleukin-6 and tumor necrosis factor-α, are associated with myocardial depression, vasoplegia, endothelial dysfunction, and microcirculatory impairment, all of which are closely linked to organ failure and death. The problem is that current shock classification frameworks, such as the SCAI staging system and guideline-based strategies from the American and European cardiology societies, rely almost exclusively on hemodynamic and metabolic parameters. They capture how badly the circulation is failing but say little about the underlying inflammatory injury that may be driving the patient toward irreversible decline.

The results revealed a clear temporal pattern. At the moment of implantation, tumor necrosis factor-α was the only cytokine significantly elevated in patients who would later die. By 24 hours, both interleukin-6 and tumor necrosis factor-α were higher in non-survivors. The divergence peaked at 72 hours, driven overwhelmingly by interleukin-6, which showed the most pronounced, sustained, and temporally consistent separation of any molecule measured. At 120 hours, interleukin-6 remained the only cytokine still discriminating between outcomes, and by 168 hours no cytokine showed significant differences. Interleukin-1β never differed between groups at any time point, likely because it is released locally and transiently, whereas interleukin-6 acts as a downstream amplifier that better reflects sustained systemic inflammation.

The statistical analysis strengthened the clinical significance of this signal. Using time-point-specific multivariable logistic regression models designed to avoid collinearity between repeated measurements, the researchers found that log-transformed interleukin-6 was already an independent predictor of 30-day mortality at 24 hours, with an odds ratio of 3.07. At 72 hours, the model was most robust: interleukin-6 carried an odds ratio of 4.68 and lactate an odds ratio of 2.74, while C-reactive protein and procalcitonin lost their significance. Receiver operating characteristic analysis at 72 hours yielded an area under the curve of 0.745 and identified an optimal cut-off of 164.5 picograms per milliliter, with 68.0 percent sensitivity and 82.1 percent specificity. In exploratory landmark-type Kaplan–Meier analysis, patients above this threshold had markedly lower 30-day survival, with a log-rank p value below 0.001.

Perhaps most importantly, interleukin-6 added prognostic value beyond the markers clinicians already use. The researchers compared a base model containing age, lactate, and the vasoactive-inotropic score, a composite measure of pharmacological hemodynamic support, against an extended model that also included interleukin-6. Adding the cytokine improved the area under the curve from 0.79 to 0.82 and significantly improved overall model fit according to the likelihood-ratio test. Notably, the vasoactive-inotropic score was no longer statistically significant in the fully adjusted model, suggesting that the inflammatory signal captured information about patient trajectory that the intensity of drug support alone could not. Interaction analyses showed that the association between interleukin-6 and mortality was not modified by hemodynamic severity or by blood culture positivity.

That last finding addresses one of the thorniest questions in shock medicine: is the inflammation driven by the shock itself, or by a hidden infection? Blood cultures were obtained in 90 of the 113 patients, and 16 of those cultures were positive, predominantly with gram-positive organisms such as Staphylococcus epidermidis and Cutibacterium acnes. There was no difference in culture positivity between survivors and non-survivors, and interleukin-6 concentrations showed no consistent association with blood culture status. The authors are careful, however, not to overinterpret this. Blood cultures are an imperfect marker of infection, and the retrospective design meant that culture-negative infections, contamination, ventilator-associated infections, and the effects of antibiotic exposure could not be systematically excluded. Persistent interleukin-6 elevation, they conclude, should be read as a marker of ongoing inflammatory activation that may reflect sterile shock inflammation, unrecognized infection, or vasoplegic physiology.

The study also offers a window into how the devices themselves might interact with the immune system. Non-survivors required higher pump performance levels at the first follow-up assessments, likely reflecting higher rotational speeds, and non-physiological shear stress and hemolysis are known to modulate inflammatory signaling. Yet because every patient in the cohort received a microaxial pump and there was no control group, the study cannot determine whether the support itself attenuates, amplifies, or merely accompanies the inflammatory response. The cohort also had a comparatively low 30-day mortality rate of 26.5 percent, which the authors attribute to selection: patients on veno-arterial ECMO, those supported for less than 24 hours, and the most refractory stage E patients were all excluded. Non-survivors in the study were older, had lower ejection fractions, higher SOFA scores, and more often required continuous hemodiafiltration, consistent with greater baseline vulnerability.

The findings arrive at a moment when the field is recalibrating its expectations for mechanical circulatory support. The randomized DanGer Shock trial demonstrated that microaxial pump support can improve outcomes in selected patients with infarct-related shock, but benefit depends on timing, etiology, and overall severity, while the CHIP-BCIS3 trial of elective unloading during complex percutaneous coronary intervention did not reduce major adverse events. Against that backdrop, the Marburg results suggest that biology, not just hemodynamics, may determine who benefits. The authors emphasize that interleukin-6 at 72 hours should not be used as the sole criterion for starting, withholding, or modifying support. Instead, a persistently elevated value should trigger structured clinical reassessment, including evaluation of support adequacy, catecholamine exposure, residual ischemia, infection, and device complications.

The study’s limitations are acknowledged candidly. It was single-center and retrospective, exact intervals from symptom onset to implantation were unavailable, and attrition reduced statistical power at later time points. Only preselected cytokines were measured, leaving out potentially relevant mediators such as interleukin-10, and serial cardiac biomarkers were not standardized across all time points. The 164.5 picogram per milliliter threshold requires external validation and assay standardization before it can enter routine practice. Still, the central message is compelling: in patients whose hearts are being mechanically supported, the immune system’s trajectory over the first three days tells a story that blood pressure and lactate alone cannot. Serial interleukin-6 measurement, the authors conclude, is a hypothesis-generating tool that may finally give clinicians a biological window into which shock patients are recovering and which are quietly sliding toward a fate that no pump, however sophisticated, can reverse on its own.

Subject of Research: Cytokine trajectories and inflammatory risk stratification in cardiogenic shock patients supported by microaxial flow pumps

Article Title: Temporal cytokine trajectories during microaxial flow pump support identify a high-risk inflammatory phenotype in cardiogenic shock

Article References: Kreutz, J., Amoah, J., Mihali, K., Harbaum, L., Chatzis, G., Patsalis, N., Syntila, S., Schieffer, B., & Markus, B. (2026). Temporal cytokine trajectories during microaxial flow pump support identify a high-risk inflammatory phenotype in cardiogenic shock. Clinical Research in Cardiology. https://doi.org/10.1007/s00392-026-03040-z

Image Credits: AI Generated

DOI: 10.1007/s00392-026-03040-z

Keywords: cardiogenic shock, interleukin-6, microaxial flow pump, mechanical circulatory support, cytokines, inflammation, prognostic markers, 30-day mortality, Impella, biomarkers, SCAI classification, lactate