Abstract: Background – Early and accurate identification of septicemia in neonatal foals will improve survival rates. In human medicine, the neutrophil-to-lymphocyte ratio (NLR), neutrophil-to-monocyte ratio (NMR), monocyte-to-lymphocyte ratio (MLR), and plasma cell-free (cfDNA) concentrations have been used for early identification of bacteremic patients. Hypothesis/Objectives – To determine the utility of the NLR, NMR, MLR, and plasma cfDNA to predict a positive blood culture, and to evaluate the diagnostic performance of these parameters at admission to discriminate between Gram-positive or Gram-negative infection. Animals – 391 hospitalized foals < 5 days old. Inclusion criteria (1) CBC and a blood culture obtained aseptically at admission; (2) documentation of either discharge or non-survival. Methods – Retrospective. Logistic regression analysis generated predictive models and Area Under the Curve (AUC) was calculated to determine the discriminative ability of independent risk factors and composite models. Plasma concentration of cfDNA was measured using fluorometry. Results – Plasma cfDNA, IgG, NLR, and WBC were identified as independent predictors of positive blood culture and used to generate the composite model (AUC: 0.806 95%CI: 0.665-0.826). Plasma cfDNA, IgG, and neutrophil counts were identified as independent predictors of Gram-negative bacteremia and used to generate the composite model (AUC: 0.807 95%CI: 0.669-0.945). Monocyte counts and age were identified as independent predictors of Gram-positive bacteremia and used to generate the composite model (AUC: 0.670 95%CI: 0.57-0.76). Complete results are in Table 1 Conclusions– We generated composite models that could aid in the identification of bacteremic foals at admission. Models are being validated in the clinic.
