ABSTRACT
PURPOSE: This study sought to investigate the screening accuracy of procalcitonin (PCT) for bacteremia, as defined by a positive blood culture, in a South African trauma ICU. METHODS: This was a retrospective chart review study involving 149 patients who were admitted to the ICU of a level-1 trauma center in South Africa between 2016 and 2017. Median PCT levels in patients with and without positive blood cultures were compared. The screening accuracy of PCT for a positive blood culture was summarized as sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Two PCT cut-points were investigated: a general cut-point in the South African context (> 2.0 ng/mL), and a trauma ICU-specific cut-point (prioritizing specificity while optimizing sensitivity) which was determined from a receiver-operator-characteristic curve. RESULTS: Bacteremic patients had higher median PCT levels when compared with non-bacteremic patients (30.5 ng/mL versus 6.6 ng/mL, p = 0.002). The sensitivity, specificity, PPV, and NPV of PCT > 2.0 ng/mL was 86% (95% confidence interval-CI 71-94%), 29% (CI 22-38%), 28% (CI 20-37%), and 87% (CI 73-94%), respectively. The unit-specific cut-point was PCT > 31.0 ng/mL, which had a sensitivity, specificity, PPV, and NPV of 50% (CI 34-66%), 80% (CI 71-86%), 44% (CI 30-59%), and 83% (CI 75-89%), respectively. Unlike PCT > 2.0 ng/mL, PCT > 31.0 ng/mL demonstrated fair-to-good test specificity in a sub-analysis of patients who underwent recent surgery. CONCLUSIONS: Increased PCT levels were associated with bacteremia in this study. PCT > 31.0 ng/mL may be used to rule in suspected bacteremia in this trauma ICU setting.
