Nosocomial Infection Surveillance in French Intensive Care Units: 2007 Results of the REA-RAISIN National Network and Temporal Trends

Background: Nosocomial infection (NI) surveillance in intensive care units (ICU) is a priority as patients are at higher risk of infection due to their critical status and invasive device exposure.
Since 2004, surveillance networks implemented by the 5 regional coordinating centers (CCLIN) are monitored at national level through the NI alert, investigation and surveillance network (RAISIN) collaborating with the national public health institute (InVS).

Objective: To assess and compare NI rates over time and amongst units, providing relevant information to target infection control policies.

Methods: Six month a year (Jan-June) on a voluntary basis, ICU collected data for each patient hospitalized more than 2 days. This patient-based surveillance focused on device-related NI: ventilator-associated pneumonia (PNE), urinary tract infection (UTI) associated with indwelling urinary catheter (UC), central venous catheter (CVC) colonization and catheter-associated bacteraemia (COL/CRB) and blood stream infection (BSI).

Results: In 2007, 165 ICU included 22,927 patients.
Patients characteristics were as follows: mean age 61.4 years, M/F sex-ratio 1.6, admission type was medical 68%, scheduled surgery 14% and emergency surgery 18%; patient origin was community 55%, acute care 36%, long term care 5% and other ICU 3%; 10% were trauma patients, 13% had impaired immunity; 55% received antibiotic treatment at admission, mean SAPS II severity score was 41.7 and average length of stay in ICU 11.2 days. Device exposure was frequent: intubation 64%, CVC 60% and UC 84% with device utilization ratio (DUR) being 61.0, 63.2 and 81.6 respectively.
Among 22,927 patients, 3,298 (14.4%) had at least one infection; 7.6% of CVC were colonized or infected. The most frequent micro-organisms were P. aeruginosa (15.5%), E. coli (13.9%), S. aureus (11.9%), S. epidermidis (5.5%) and Candida albicans (5.3%).
Overall incidence rates were: 15.48 PNE /1000 intubation-days, 6.01 COL and 0.97 CRB /1000 CVC-days, 6.47 UTI /1000 UC-days and 3.63 BSI /1000 ICU-days.
Patient's characteristics and NI rates greatly varied between ICU and patient-based data allowed better unit comparisons through risk-adjusted rates distributions and standardized infection ratio.

From 2004 to 2007, changes in patient characteristics mainly concerned SAPSII (from 39.4 to 41.7), antibiotic treatment at admission (+13.1%) and device exposure (DUR for intubation +8.7%, CVC +4.8%, UC +4.5%). Incidence rates decreased for UTI (‑22.2%), and PNE (-4.8%) but increased for BSI (+9.7%) and COL (+4.3%). Percentage of MRSA within species decreased from 48.7 to 35.6%.

Conclusions: These data from a large sample of French hospitals serve as a national reference to better document infectious risk in ICU, providing the participants with an evidence-based approach for improving care practices and reducing NI.