The results of a simplified quantitative broth dilution quantitative tip culture (QTC) of 331 central venous catheters were compared with clinical data prospectively recorded in critically III patients to diagnose bacteremic or nonbacteremic catheter-related sepsis (CRS) (36 catheters), as opposed to contamination (42 catheters) or simple colonization from a distant septic focus (seven catheters). Thirty-five of 36 catheters associated with CRS yielded 10(3) colony-forming units per milliliter (CFU/mL) or more, and 3.8 X 10(2) Candida organisms grew from one. In contrast, 5 X 10(2) CFU/mL or less grew from 37 of 42 contaminated catheters. A QTC of 10(3) CFU/mL or more was 97.5% sensitive and 88% specific for the diagnosis of CRS. The QTC appeared especially useful for the diagnosis of CRS secondary to blood-borne seeding of catheters, or associated with coagulase-negative staphylococci.