IL-6 and lactate better discriminated 28-day mortality than CRP, procalcitonin and NLR in retrospective ICU sepsis cohort

Interleukin-6 (IL-6) and venous lactate showed greater discrimination of 28-day mortality than C-reactive protein (CRP), procalcitonin (PCT) and the neutrophil-to-lymphocyte ratio (NLR) in a single-centre retrospective cohort of 310 adults with Sepsis-3 sepsis requiring intensive care. The study abstract reports that 28-day mortality was 52.6%, while 82.3% of patients received vasopressors and 65.2% met a prespecified septic-shock approximation. Thus, the findings describe a predominantly high-acuity cohort rather than all patients with sepsis.

Patients had to remain in the intensive care unit for at least 24 hours. The investigators analysed each marker using the first value obtained after ICU admission, recorded at a median of 0.6 hours with an interquartile range of 0.5 to 4.5 hours, and, in parallel, the worst value within the first 24 hours. They compared discrimination using area under the receiver-operating-characteristic curve (AUC) and the DeLong test, overall and within severity strata and by admission source. The septic-shock approximation was based on vasopressor use and a first post-ICU venous lactate above 2 mmol/L.

With first measurements, CRP had an AUC of 0.504 (95% confidence interval, 0.439 to 0.567) and PCT had an AUC of 0.549 (0.486 to 0.614), both at or near chance-level discrimination. The NLR point estimate was below chance at 0.438 (0.371 to 0.501), although its confidence interval included 0.50. IL-6 had an AUC of 0.694 (0.631 to 0.756), and venous lactate had an AUC of 0.682 (0.622 to 0.742). Results were similar when the worst values within 24 hours were used, with all paired DeLong p values of at least 0.19.

Adding first IL-6 and lactate values to a partially window-aligned clinical model increased AUC from 0.782 to 0.808. The likelihood-ratio p value was 0.0004, while the DeLong comparison yielded p = 0.057. Performance was weaker among the 92 patients who did not meet the septic-shock approximation: AUC was 0.578 for IL-6 and 0.502 for lactate. The authors said this attenuation meant that restriction to a narrow, high-severity case mix could not be excluded as an explanation. By contrast, assay censoring appeared insufficient to explain the limited prognostic information associated with CRP and neutrophil-based indices: CRP and NLR were not significantly correlated with reconstructed SOFA, and only 1.3% of first CRP values reached the assay ceiling. These findings come from a retrospective analysis summarized in the abstract; they do not establish prospective performance beyond the included high-acuity ICU population, and the authors called for prospective evaluation.

Subject of Research: Retrospective comparison of inflammatory markers, IL-6 and lactate for predicting 28-day mortality in high-acuity ICU sepsis

Article Title: Head-to-head comparison of routinely ordered inflammatory markers, interleukin-6 and lactate for 28-day mortality in high-acuity ICU sepsis: a predominantly vasopressor-treated cohort

Article References: Feng, L., Xie, Q., Wang, Y., & Li, K. (2026). Head-to-head comparison of routinely ordered inflammatory markers, interleukin-6 and lactate for 28-day mortality in high-acuity ICU sepsis: a predominantly vasopressor-treated cohort. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14472-3

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14472-3

Keywords: sepsis, intensive care, 28-day mortality, interleukin-6, venous lactate, C-reactive protein, procalcitonin, neutrophil-to-lymphocyte ratio, vasopressors, retrospective cohort, Head-to-head, comparison