1016

Abstract

Listen

Translate article

Introduction: Clinical signs of infection with elevated WBC or fever while on extracorporeal membranous oxygenation (ECMO) are difficult for the clinician to discern systemic infective processes from systemic inflammatory reactions induced by the ECMO circuit. This false information triggers the clinician to initiate possibly unwarranted or prolonged antibiotic therapy and/or diagnostic investigations. Utilizing procalcitonin (PCT) as a bio-marker of bacterial infection, we evaluated the effectiveness of PCT measurement in patients on ECMO with clinical signs of infection. Methods: An algorithmic protocol was utilized to measure PCT as part of a clinical workup for infectious processes. This protocol was utilized on all ECMO patients with elevated WBC (14.1 ± 5.3 B/L) and clinical signs of infection. Antibiotics (Abx) treatment was considered when PCT was ≥ 2 ng/mL. After initiating Abx treatment, PCT was trended until it decreases below 2 ng/mL with negative cultures; discontinuation of the antibiotics was done. Retrospective data was accumulated for the patients on ECMO from Jan/2012 to Jun/2013 and analyzed to validate our protocol. Results: During this study period, a total of 20 patients with 21 episodes of suspected infection were identified. 48% of patients had infections based on positive cultures (5 blood sepsis; 4 pneumonia). PCT levels of patients with and without infection were 23.4 ± 31.8 ng/ml and 1.78 ± 1.74 ng/dl respectively. PCT was able to predict infection with sensitivity of 90%, specificity of 79%; positive predictive value of 82% and negative predictive value of 90%. All of patients in true positive group, defined as having both elevated PCT levels and subsequent positive culture findings, received antibiotics treatment. 89% of patients in true negative group, defined as having no elevation in PCT and subsequent negative culture results, avoided unnecessary treatment. Conclusions: PCT is an effective test for differentiating infection from inflammation in patients on ECMO utilizing PCT ≥ 2 ng/ml as diagnosis of infection. Using our protocol, empiric antibiotics treatment could be minimized, and PCT trending could reflect the response to treatment or possible recurrence in an infected patient.