Procalcitonin for the early diagnosis of sepsis in burn patients: A retrospective study

Background: The gold standard for sepsis diagnosis in burn patient still relies on microbiological cultures, which take 48-72h to provide results, delaying the start of antimicrobial therapy. Thus, biomarkers allowing an earlier sepsis diagnosis in burn patients are needed. Methods: This retrospective observational study included 150 burn patients with total burned surface area > 15%. Clinical diagnosis of sepsis among these patients was done according to the American Burn Association criteria. Biomarker (procalcitonin, white blood cells and platelet countings, prothrombinemia, D-dimers, C-reactive protein, blood lactate and temperature) values were available for 48 patients without sepsis (2767 timepoints) and 102 patients with sepsis (652 timepoints). Quantitative variables were compared with Mann-Whitney tests and qualitative variables were compared with Pearson chi-square test. Effect size was measured by the probability of superiority. Receiver operating characteristic (ROC) curves evaluate capacity for sepsis diagnosis. Sensitivity, specificity, positive and negative predictive values were calculated for some cut-off values, including the best cut-off defined by the maximum of Youden index. Results: Statistically significant differences between the groups of septic and non-septic patients, with medium to large effect size, were detected for all the biomarkers considered, except temperature. PCT was the biomarker with the largest AUC and effect size (AUC=0.71). Analysis of the PCT ROC curve showed that 0.5ng/mL cut-off presented highest sensitivity and lowest specificity, whereas 1.5ng/mL cut-off was associated with lowest sensitivity and highest specificity. Conclusion: Procalcitonin showed to be the best of the biomarkers studied for an early diagnosis of sepsis. Its use should be considered in antimicrobial stewardship programs in Burn Units. (C) 2017 Elsevier Ltd and ISBI. All rights reserved.