Diagnosis of central venous catheter-related bloodstream infection without catheter removal: A prospective observational study

Abstract

BackgroundCatheter-related bloodstream infections (CRBSI) resulting from bacterial colonisation of an intravascular catheter are the leading cause of nosocomially acquired sepsis contributing significantly towards in-hospital morbidity and mortality. Suspicion of central venous CRBSI leads frequently to catheter withdrawal but not all infection requires the catheter to be withdrawn; therefore, diagnosis of central venous CRBSI without catheter withdrawal is a necessity. MethodsThe study was prospectively performed in a cohort of adult patients who had short term central venous catheter use. The samples collected from each patients included, skin swab from insertion site, swab from catheter hub, paired blood samples from catheter and from the peripheral vein for quantitative blood culture collected within 15 min of each other and catheter-tip sample by cutting off the tip (distal 5-cm segment). All samples were processed immediately. Results50 episodes of clinical sepsis involving 100 patients occurred in the study population. 28 of the episodes were confirmed as CR-BSI (56%). Blood culture from the central venous catheter had the highest sensitivity (71.43%) and the greatest negative predictive value (86.67%). However, the peripheral blood culture was most specific and had the highest positive predictive value (specificity75%; positive predictive value 50%). The most accurate technique was differential quantitative blood cultures (accuracy 72%), followed by semiquantitative superficial cultures (accuracy 68%), although there were no statistically significant differences between values. ConclusionWe recommend combining semiquantitative cultures and peripheral blood cultures to screen for CR-BSI, leaving differential quantitative blood cultures as a confirmatory and more specific technique.