Impact of routine multiple site decontamination in intubated patients on ICU-acquired infections

Introduction

We have shown that a the combination of selective digestive decontamination with topical antibiotics (SDD) and a decontamination regimen using nasal mupirocin with chlorhexidine bodywashing (M/C) markedly reduced acquired infections (AI) in intubated patients as compared with SDD alone, M/C alone or none [1]. We report the surveillance of AI in our ICU before and after the implementation of multiple site decontamination (MSD) as a routine prevention procedure.

Methods

MSD was implemented in June 2007. We compared the incidence rates of AI (expressed per 1,000 patient-days (‰) or per 1,000 device-days for device-related AI) between the last 1-year period before (period A, 7,723 patient-days) and the first 1-year period after (period B, 7,646 patient-days) MSD use was started.

Results

In period B (MSD) versus period A (no MSD), there was a reduction in the rates of total AI (9.4‰ vs 23.6‰), bloodstream infections (1.0‰ vs 4.0‰), ventilator-acquired pneumonia (5.6‰ vs 19.1‰) (all P < 0.01), and to a less extent catheter-related urinary tract infection (1.0‰ vs 2.2‰, P = 0.11). Multivariate analysis with the Cox regression model showed that period A (OR 2.34 (1.64 to 3.20)) and the presence of a central venous catheter (OR 2.07 (1.22 to 3.57)) were the two independent risk factors for AI. In period B, there was a statistically significant reduction in the rates of AI involving S. aureus (1.0‰ vs 3.0‰), coagulase-negative staphylococci (0.4‰ vs 1.3‰), Candida (0.4‰ vs 1.3‰) and aerobic Gram-negative bacilli (5.0‰ vs 15.8‰), especially Enterobacteriaceae (3.4‰ vs 9.4‰), Pseudomonas aeruginosa (1.4‰ vs 3.9‰), S. maltophilia (0 vs 1.7‰) and Acinetobacter sp. (0 vs 0.8‰) (all P ≤0.05). MRSA AI rates were also lower (0.3‰ vs 0.9‰, NS) and no AI due to VRE occurred during both period. The rates of AI involving antimicrobial-resistant Enterobacteriaceae were lower in period B for 12 antimicrobials tested, especially for ticarcillin (2.1‰ vs 6.5‰), ticarcillin/clavulanate (0.8‰ vs 4.0‰), cefotaxime (0.3‰ vs 2.2‰) and colistin (0.5‰ vs 1.8‰) (all P ≤0.02). Similar findings were observed for antimicrobial-resistant P. aeruginosa although differences were not significant. Overall, the rate of AI involving organisms defined as multidrug-resistant (MDROs) markedly decreased in period B (2.5‰ vs 9.1‰, P < 0.001).

Conclusions

Routine use of MDS was associated with a strong reduction in ICU AI. The reduction was consistently observed for all principal classes of pathogens, including MDROs. No new MDRO emerged.