Hospital-wide Surveillance of Healthcare-associated Bloodstream Infections, 1999-2007: Finnish Experiences

Background:

Surveillance of healthcare-associated laboratory-confirmed bloodstream infection (LCBI) has been conducted as a part of Finnish Hospital Infection Program (SIRO) since 1999, using the CDC definition for LCBI. The definition was revised in 2008 by excluding the infections caused by single blood cultures positive for common skin contaminants.

Objective:

We analyzed the SIRO surveillance data from years 1999-2007 to describe the epidemiology, microbiology and outcome of LCBI. We also evaluated the effect of the change of the CDC definition on LCBI rates.

Methods:

Hospital-wide prospective incidence surveillance for LCBI was conducted by local infection control nurses who regularly reviewed positive blood culture results. Clinical and microbiological data were recorded on standardized form, including the CDC criteria for LCBI: recognized pathogen isolated from blood culture, common skin contaminant isolated from two blood cultures or one blood culture from patient with intravascular access device and physician institutes antimicrobial therapy. Patient-days were obtained from the hospital’s information technology department and outcomes at 7 and 28 days from the population registry.

Results:

A total of 7916 LCBIs were identified in 6669 patients; 2104 (27%) were related to intensive care. The overall rate was 0.8 LCBIs per 1000 patient-days (range by hospitals, 0.2-0.9; range by specialties, 0.1-1.4). The rates were highest in pediatrics (1.4) and internal medicine (1.3). Hematology contributed to the high rate in internal medicine and neonatal intensive care in pediatrics. Of the LCBIs, 6144 (78%) were primary, 46% of which were associated with central lines. The urinary tract and surgical site were the most common sources of secondary LCBI. The most common pathogens were coagulase-negative staphylococci (28%), Staphylococcus aureus (13%), Escherichia coli (11%), and enterococci (9%), however, the distribution varied significantly by patient groups. Methicillin-resistance was detected in 4% of S. aureus isolates and vancomycin-resistance in 1% of enterococci. Overall case-fatality was 8% within 7 days (range by pathogens, 4-20%), highest for Candida albicans (20%) and Pseudomonas aeruginosa (18%). Of all LCBIs, 1287 (16%) were caused by skin contaminants isolated from one blood culture, after excluding them the overall rate was 0.7 LCBIs per 1000 patient-days.

Conclusions:

LCBIs may cause serious outcome. Despite the considerable workload for participating hospitals, hospital-wide surveillance on LCBI provides useful data on causative organisms and their antibiotic resistance patterns to give feedback to different specialties. On national level, the same denominator data could be utilized for laboratory-based Clostridium difficile infection surveillance.