The systemic immune-inflammation index in predicting sepsis mortality

Abstract

Background The systemic immune-inflammation index (SII) is a novel parameter and its role in the prognosis of sepsis has never been explored previously. Methods We retrospectively assessed 267 patients with blood-culture confirmed sepsis. Clinical and laboratory data recorded at intensive care unit (ICU) admission were analyzed. Outcomes of interest included in-hospital mortality and length-of-stay (LOS) in the ICU. Sequential Organ Failure Assessment (SOFA) scores, SII, neutrophil-lymphocyte ratio (NLR), and platelet-lymphocyte ratio (PLR) were calculated. Multivariable regression analysis was used to identify independent predictors of LOS and mortality. Area under receiver operator characteristic (AUROC) curves were used to determine optimum cut-offs, and the incremental effect of SII on the SOFA score was assessed using model discrimination and calibration properties. Results There were 76 (28.5%) non-survivors. SII, NLR, and PLR were independent predictors of sepsis mortality, with adjusted odds ratios of 1.51 (1.24-1.84), 1.67 (1.30-2.13) and 1.24 (1.11-1.39). SII and SOFA score were independent predictors of LOS. SII had an AUROC of 0.848, and the optimum cut-off was 564 with a sensitivity and specificity of 85.5% and 71.2%. The addition of SII to the model had a significant incremental effect on the predictive ability of SOFA score (Net Reclassification Index = 0.084, P = 0.025; Integrated Discrimination Index = 0.056, P = 0.001). Conclusion The SII is an inexpensive parameter that can be used in addition to clinical sepsis scores to improve the accuracy of patient assessment.