Central venous catheter placement at the time of extracorporeal membrane oxygenation decannulation: is it safe?

PURPOSE: Because of concerns for infectious and hemorrhagic complications, methods of obtaining central venous access after extracorporeal membrane oxygenation (ECMO) vary by institution. For infants requiring ECMO, it has been our practice to exchange the venous cannula for a tunneled central venous catheter (Broviac) at the time of decannulation. The purpose of this study is to compare the incidence of catheter-related complications in these patients to a national registry. METHODS: The medical records of all non-cardiac surgery infants, 12 months or younger, requiring ECMO at our institution from 1993 to 2005 (n = 138) were reviewed. Complete information was available for 134. Center for Disease Control criteria was used to identify cases of catheter-related bloodstream infections (BSIs). Data from the National Nosocomial Infections Surveillance system served as a comparative group. Logistic regression was used to determine risk factors for catheter-related BSI. RESULTS: A total of 134 infants spent a mean of 8.1 +/- 4.3 days (range, 1-21 days) on ECMO. At the time of decannulation, a Broviac catheter was placed in the right internal jugular vein of 95 (71%) and remained in place for a mean of 18.2 +/- 17 days (range, 1-109 days). The incidence of BSI related to these catheters was not significantly different than that reported by the National Nosocomial Infections Surveillance system for all central venous catheters over a similar period (6.4/1000 vs 7.3/1000 catheter days; P = .68). The number of days on ECMO and number of catheter days were independent predictors of catheter-related BSI in both bivariate and multivariate logistic regression models (P