Surveillance of healthcare-associated infections (HAIs) in intensive care units is a priority, as patients are at increased risk of infection due to their critical condition and the invasive devices to which they are exposed. Since 2004, surveillance coordinated by the Network for Alert, Investigation, and Surveillance of HAI (Raisin) in intensive care has targeted infections associated with invasive devices for which a prevention strategy is essential: pneumonia (PNE), colonization (COL) of central venous catheters (CVC), and associated infection or bacteremia (ILC/BLC), urinary tract infection (UTI), and bacteremia (BAC). Each year, participating departments collect data for six months on all patients hospitalized for more than two days (d). From January to June 2006, 158 departments included 22,090 patients: their mean age was 61 years and the male-to-female ratio was 1.6. At admission, 68% of patients were in medical care, 18% in emergency surgery, and 14% in elective surgery; 55% of patients were admitted from outside the hospital, 37% from short-stay units, 5% from medium- or long-stay units, and 3% from another intensive care unit; 10% of patients were trauma patients and 12% were immunocompromised; 51% were receiving antibiotic therapy at admission. Their mean IGS II score was 40, and the mean length of stay was 11 days. Exposure to invasive devices was common: intubation (61%), central venous catheter (59%), and urinary catheter (81%). Among 22,090 patients, 3,113 (14.1%) developed at least one infection, totaling 5,284 events including CLIs. The most frequently isolated microorganisms were Pseudomonas aeruginosa (15.0%), Escherichia coli (14.8%), Staphylococcus aureus (14.0%), Candida albicans (5.7%), and S. epidermidis (5.5%); 39.5% of S. aureus strains were methicillin-resistant (48.7% in 2004). The observed incidence rates were 16.17 PNE per 1,000 days of intubation, 4.87 COL (and 0.83 BLC) per 1,000 days of CVC, 7.94 URI per 1,000 days of catheterization, and 3.27 BAC per 1,000 days of hospitalization. Patient characteristics and incidence rates varied significantly from one unit to another. From 2004 to 2006, a decrease in incidence was observed for CLIs (-16.9%), URIs (-5.9%), PNE (-5.9%), and BAC (-1.5%). These data serve as a national benchmark, enabling a better understanding of healthcare-associated infections in intensive care and improving their control through the feedback of results to participating units.