Prospective, randomized trial of two different modalities of flushing central venous catheters in pediatric patients with cancer.

Abstract

There are limited prospective data on whether the method of flushing affects the complication rate of tunnelled central venous catheters (CVCs). During a 25-month period, 203 pediatric patients who had newly placed Broviac-Hickman CVCs were randomly assigned to standard flushing with heparin solution or to experimental flushing with normal saline via a positive-pressure cap. Two hundred twenty-one complications were recorded among 75,249 CVC-days (2.94 per 1,000 CVC-days). A higher incidence of CVC occlusion (83 v 41 episodes; P = .0002) and bacteremia (24 v 9; P = .01) were found in the experimental arm. The cumulative probability of developing at least one CVC complication was higher in the experimental arm than in the standard arm (65.1% [95% CI, 55% to 75%] v 43.8% [95% CI, 34% to 54%], respectively; P = .01). No difference was found in either the cause or the frequency of premature removal of CVCs between the two study arms. After a median follow-up of 360 days (range, 4 to 1,073), CVC survival was similar: 77% (95% CI, 66% to 84%) for the experimental arm and 69% (95% CI, 53% to 80%) for the standard arm (P = .7). The factors associated with the occurrence of CVC complication were a diagnosis of leukemia/lymphoma, double-lumen CVC, and experimental flushing. The only factor significantly associated with premature removal of a CVC was a diagnosis of leukemia/lymphoma (hazard rate, 2.3; 95% CI, 1.1 to 4.7). An increased complication rate was found with normal saline flushing, but additional investigation is warranted to clarify whether it is related to saline use or to once-a-week flushing.