Abstract
Background A foremost cause of mortality in intensive care units is sepsis, especially in resource-limited nations such as India, where delays in patient presentations and limited diagnostic facilities pose common challenges. While conventional severity scores such as Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation II (APACHE II) are helpful, the multiple assessments required and labor-intensive calculations create an impractical situation for some clinicians. Thus, an urgent need arises for reliable biomarkers that could be measured at admission for faster and better prognosis. In this context, this study aims to evaluate the combined prognostic potential of three biomarkers, each reflecting a key facet in sepsis pathophysiology: coagulation (prothrombin time-international normalized ratio (PT-INR)), inflammation (interleukin-6 (IL-6)), and immunometabolism (high-density lipoprotein (HDL)). Methodology This prospective, observational study was conducted from July 2023 to March 2025 among 152 adult patients with severe sepsis admitted to Jawaharlal Nehru Hospital in Ajmer, India. Admission levels of PT-INR, IL-6, and HDL cholesterol were measured. Sepsis was diagnosed based on the Sepsis-3 criteria (SOFA score increase of ≥2 points). The prognostic utility of the biomarkers was assessed by correlating their admission levels with clinical severity scores (SOFA and APACHE II) and in-hospital mortality. Statistical analysis included non-parametric tests, receiver operating characteristic (ROC) curve analysis, and binary logistic regression. Results Of the 152 patients enrolled, the in-hospital mortality rate was 38.8%. Non-survivors had significantly higher admission levels of IL-6 (104.87 ± 28.45 pg/mL) and PT-INR (1.97 ± 0.50) and significantly lower levels of HDL (34.89 ± 8.85 mg/dL) compared to survivors. ROC curve analysis demonstrated that IL-6 was an exceptionally strong predictor of mortality, with an area under the curve (AUC) of 0.995, and 100% sensitivity and specificity at a cutoff of 75 pg/mL. PT-INR showed a moderate predictive ability (AUC = 0.606), while HDL was a weaker predictor (AUC = 0.405). A combined logistic regression model incorporating all three biomarkers showed superior prognostic accuracy with an AUC of 0.94. Conclusions The combined use of admission-day biomarkers reflecting inflammation (IL-6), coagulation (PT-INR), and immunometabolism (HDL) provides a powerful and synergistic tool for early risk stratification in severe sepsis. While each marker offers unique prognostic insights, the multi-marker approach demonstrated superior predictive accuracy compared to individual markers alone. Specifically, IL-6 emerged as a remarkably potent and accurate predictor of mortality in this high-acuity patient cohort. This multi-marker strategy is particularly valuable for improving early prognostication and guiding treatment decisions in resource-limited settings.