Central venous blood culture: a useful test for catheter colonisation?

Abstract

To evaluate the widespread practice of using qualitative culture of venous blood taken through central venous catheters (CVCs) as a means of diagnosing catheter colonisation in situ. A prospective clinical study. Three hundred fifty-one CVCs were inserted into 228 critically ill patients. Prior to CVC removal, blood was taken for qualitative culture from the CVC and a peripheral site. All catheter tips underwent semi-quantitative analysis of bacterial colony count [1]. One hundred eighteen (33.6%) CVCs were found to be colonised on removal. A positive central blood culture had a sensitivity and specificity of 50.8% and 78.9% when compared with the 'gold standard' of catheter tip culture. This gives a positive predictive value of 47.7% and a negative predictive value of 76.5%. The sensitivity and specificity of a positive peripheral blood culture were 41.5% and 77.7% with positive and negative predictive values of 48.8% and 72.9%, respectively. When only those catheters removed because of systemic sepsis (n = 139) were considered, a positive central blood culture had a sensitivity of 58.8% and a specificity of 69.3%. Our results indicate that the use of central blood culture confers a small advantage in sensitivity compared with peripheral blood culture. This advantage was further improved by only considering the catheters removed because of systemic sepsis but at the cost of a loss of specificity. Qualitative blood culture is a poor tool for the diagnosis of in-situ CVC colonisation.