CVC-related complications were retrospectively analysed for 373 CVCs inserted in 261 hematologic pts consecutively admitted to our Hematology Department between January 2002 and March 2004. Pts diagnosis comprised AML (96 pts, 36.8%), ALL (19 pts, 7.3%), lymphoproliferative disorders (74 pts, 28.3%), MM (51 pts, 19.5%), CML (13 pts, 5%), others (8 pts, 3.1%). The CVCs were polyurethane Plastimed three lumen 7 Fr (193 cases) for chemotherapy and polyurethane Arrow three lumen 12 Fr (173 cases) for chemotherapy and peripheral blood stem cell (PBSC) apheresis. CVCs were inserted, according to physician's judgement, either at bedside (369 cases) or in the operating room (4 cases) and were used for drug infusion in 309 cases (82.8%), for total parenteral nutrition in 19 cases (5.1%), for blood transfusions in 226 cases (60.6%), for PBSC apheresis in 152 cases (40.7%). Fifty five CVCs were inserted in neutropenic pts ( N < 1 x 109/L); severe thrombocytopenia (Plts < 30 x 109/L) was present in 33/373 cases. Antithrombotic prophylaxis with low molecular weight heparin because of previous thrombosis was instituted in 33 cases (8.8%). At univariate and multivariate analysis the following risk factors for catheter-related bloodstream infections were considered: pt age, number of days/catheter, haematological disease, catheter lumen, administration of chemotherapy (standard- vs high-dose), presence of neutropenia. The median duration of CVCs after placement was 22.7 days (range 2–70) for the 7 Fr lumen, and 9 (range 1–39) for the 12 Fr. Major hemorrhagic complications related to the insertion procedure were observed in 4 cases of whom 2 were severely thrombocytopenic. CVC occlusion were observed in 24 cases (6.4%). Thrombotic complications developed in 7 cases (1.87% of inserted CVCs) and in 2/7 were complicated by pulmonary embolism; thrombophilia tests were negative in all 7 pts who developed thrombosis. Among febrile pts the frequency of bacteriemias was 19.6% (73/373 cases) of which 52/373 were CVC-related (13.9%). Approximately 6% of cases (23/373 CVCs) had CVC exit point infection with or without associated bacteriemia. Gram positive bacteria were isolated in 49 cases (67.1%), 38 of which were CVC-related (19 S. epidermidis, 4 S. pneumoniae, 3 S. aureus, 3 Enterococci species, 9 others). Gram negative bacteria were isolated in 23 cases, of which 13 were CVC-related (5 E.coli, 4 E. cloacae, 2 P. aeruginosa, 2 K pneumoniae). Candida was isolated in one case. Overall mortality in our patient population was 4.2% (11/261 pts); one patient only died of infection (P. aeruginosa sepsis) not CVC-related. At univariate and multivariate analysis significant risk factors for infection (p<0.0001) were only the number of days/catheters and duration of neutropenia. Among severely thrombocytopenic pts, haemorrhagic complications at CVC insertion were infrequent (6%). Although pts in our study population frequently displayed prolonged and severe neutropenia, frequency of CVC-related bacteremias was low (13.9%) compared to a general population reported frequency of 5–25%. Moreover, none of the pts died of CVC-related infection.