Immune Markers in Septic Shock

This single-study explainer examines whether routinely available immune-cell measures could help distinguish septic shock patients with model-estimated reduced ICU mortality under corticosteroid treatment. The investigators evaluated lymphocyte count, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and neutrophil count in a retrospective target trial emulation using the Medical Information Mart for Intensive Care IV database. (pmid:42809156)

Evidence

The study included adult ICU patients meeting the Sepsis-3 definition of septic shock. It identified 5,533 septic shock stays and compared patients who received corticosteroids with those who did not. The reported primary outcome was ICU mortality; reported secondary outcomes included in-hospital mortality, ICU and hospital length of stay, renal replacement therapy, and vasopressor duration. (pmid:42809156)

Investigators extracted baseline NLR, PLR, neutrophil count, and lymphocyte count. They used an Augmented Inverse Probability Weighting doubly robust estimator to evaluate corticosteroid effects and univariate area under the receiver-operating-characteristic curve (AUROC) analyses to compare the markers’ ability to identify stays with model-estimated corticosteroid benefit. The corticosteroid group had higher Sequential Organ Failure Assessment scores and greater vasopressor requirements than the control group before adjustment. (pmid:42809156)

Lower baseline lymphocyte count showed the highest reported univariate performance among the evaluated hematologic measures for identifying cases with model-estimated corticosteroid benefit regarding ICU mortality, with an AUROC of 0.65. Higher NLR and higher PLR also had modest predictive ability, and lower lymphocyte count, higher NLR, and higher PLR were associated with a greater likelihood of model-estimated benefit. (pmid:42809156)

The authors also reported that the proportion of stays predicted to benefit increased as NLR increased. After adjustment for disease severity, lymphocyte count remained predictive of stays likely to benefit from steroid use. The study characterized the evaluated markers’ predictive power as modest, including for the secondary outcomes. (pmid:42809156)

Analysis — Biomarker-guided corticosteroid targeting

The central result is a stratification finding, not a demonstrated treatment rule. The analysis asks whether baseline hematologic measures can identify a subgroup for which the study’s model estimates lower ICU mortality with corticosteroids. That framing matters because the reported marker associations describe model-estimated heterogeneity of benefit within an observational analysis rather than a randomized comparison of a biomarker-guided treatment strategy. (pmid:42809156)

Lymphocyte count was the strongest of the reported evaluated measures for the primary outcome, while NLR and PLR moved in directions consistent with an inflammatory or immune-profile signal. These measures are attractive for research because they were extracted from routinely available hematologic data, but the study’s own description of performance as modest means the findings do not establish that any individual marker can accurately direct corticosteroid use. Statistical discrimination and a clinically validated decision threshold are different questions. (pmid:42809156)

The comparison is also shaped by baseline illness severity. Patients receiving corticosteroids were reported to be sicker than controls, and the investigators applied AIPW adjustment to address confounding. The persistence of lymphocyte-count prediction after adjustment for disease severity supports the reported association within this model, but it does not by itself establish that the marker identifies a causal biological response to corticosteroids. Independent evaluation of the proposed stratification approach would be needed before interpreting it as a treatment-selection method. (pmid:42809156)

Limitations

This was a retrospective analysis of observational database data. The authors state that its retrospective nature precludes definitive causal inference and that residual or unmeasured confounding may remain despite use of the doubly robust AIPW estimator. Consequently, differences between treated and untreated stays may not be fully resolved by the analytic adjustment. (pmid:42809156)

The outcome language is specifically model-estimated benefit, and the reported discrimination was modest. The findings therefore support further research on simple hematologic markers as candidate tools for septic-shock stratification, while leaving uncertainty about whether they would improve outcomes when used to guide care in a prospective setting. (pmid:42809156)