To determine factors associated with risk for umbilical catheter-related sepsis, we studied neonates with one or more catheters in place for more than 3 days. Among 225 infants with 357 umbilical catheters, catheter-related sepsis occurred in 14 infants (6%). Catheter-related sepsis occurred in 5% of infants with umbilical arterial catheters and in 3% of infants with umbilical venous catheters. Staphylococcal species accounted for 71% of cases of catheter-related sepsis. Multiple logistic regression analysis revealed that very low birth weight and longer duration of antibiotic therapy were significantly associated with risk for umbilical arterial catheter-related sepsis. Increased risk for umbilical venous catheter-related sepsis was best predicted by the simultaneous occurrence of higher birth weight and infusion of hyperalimentation solution. Catheter duration correlated with duration of antibiotic therapy and with infusion of hyperalimentation solution for both types of catheters; however, in the multivariable analysis, duration of catheterization was not found to be a significant independent predictor of risk for catheter-related sepsis for either type of catheter.