Time series analysis of delta neutrophil index as the predictor of sepsis in patients with acute poisoning

Abstract

Delta neutrophil index (DNI), which reflects the fraction of immature granulocytes, is used to detect infection and sepsis from noninfectious conditions, but few studies have evaluated in the early stage of acute poisoning. This retrospective observational study was performed on acute poisoning patients who visited to the emergency department (ED) and were consecutively admitted in intensive care units over 18-month period. The serial DNI, conventional inflammatory biomarkers, and culture results were obtained in the ED and after admission. The outcomes were the identification of sepsis, bacteremia, and 30-day mortality. Of 166 patients (mean age, 56.0 years) in this cohort, 59 (35.5%) had sepsis and 29 (17.5%) had bacteremia. Initial and peak DNI fractions 24 h after ED admission were strong independent predictors of sepsis development. Analysis of the area under the curve according to multiple receiver operating characteristics showed that DNI had a higher capability to predict sepsis than other parameters (0.815 for DNI, 0.700 for procalcitonin, 0.681 for C-reactive protein, and 0.741 for white blood cell). Using multivariable logistic regression analysis, it was found that DNI was an independent predictor of sepsis (95% confidence interval (CI) of odds: 1.03-1.18) and bacteremia (95% CI: 1.01-1.14). Therefore, initial and serial measurement of DNI may serve as useful risk predictor for development of sepsis or bacteremia in acute poisoning.