Hub colonization and subsequent intraluminal progression due to frequent opening and manipulation of intravenous systems is the cause of many catheter-related infections (CRI). A prospective, comparative, randomized study was performed to assess a new closed-needleless hub device (CLAVE) compared with conventional open systems (COS). End-points were hub and skin colonization, catheter tip colonization, catheter-related bloodstream infection (CRBSI) and number of accidental needlesticks. All cultures were processed following standard semiquantitative microbiological techniques. The study involved patients who underwent heart surgery over an 11-month period in a post-surgical ICU. During the study period, 352 patients underwent major heart surgery and 1774 catheters were inserted. Overall, 865 catheters in 178 patients were allocated to the CLAVE system and 909 catheters in 174 patients to COS. The groups were similar regarding underlying conditions and risk factors for infection. Comparison of endpoint results in CLAVE and COS groups was as follows: incidence density per 1000 catheter-days of tip colonization: 59.2 versus 83.6 (P=0.003); of hub colonization: 7.56 versus 24.66 (P=0.0017); of skin colonization: 41.5 versus 58.9 (P=0.038); and of CRBSI 3.78 versus 5.89 (P=0.4). There was one accidental needlestick and one catheter-related prosthetic endocarditis in the COS group. Multivariate analysis showed that CLAVE use was an independent protective factor for tip colonization. CLAVE offered significant protection from catheter-tip and hub colonization.