Long-term effects of selective decontamination on antimicrobial resistance

To determine whether selective decontamination of the digestive tract exerts any long-term effects on antimicrobial resistance patterns.

A surveillance and interventional study comparing the antimicrobial sensitivity patterns of clinically important bacterial isolates the year before a 2-yr, double-blind, randomized, controlled study of selective decontamination of the digestive tract, and for the year thereafter when no use of the regimen was made.

A ten-bed respiratory intensive care unit (ICU) in a 1,200-bed teaching hospital.

All 1,528 patients admitted to the ICU over the 4-yr study period were included. There were 406 patients admitted in the year before the study of decontamination of the digestive tract (65% medical, 23% surgical, and 12% trauma), of whom 76% required mechanical ventilation. There were 719 patients admitted during the 2-yr study of selective decontamination (55% medical, 28% surgical, and 17% trauma), of whom 79.6% required mechanical ventilation. There were 403 patients admitted in the subsequent year (61% medical, 25% surgical, and 14% trauma), of whom 76.9% required mechanical ventilation.

We performed daily clinical monitoring to detect nosocomial infection, with microbiological investigation when clinically indicated, as well as twice-weekly routine microbiological surveillance sampling. Antimicrobial susceptibility testing using standard laboratory methods was also performed. Selective decontamination of the digestive tract included parenteral cefotaxime and oral and enteral polymyxin E, amphotericin B, and tobramycin.

The occurrence rate of nosocomial infection was 20.6%, 16.6%, and 25.3%, respectively, in the three study periods. In the year after selective decontamination, there was an increase in the occurrence rate of infection (p = .005), with an-associated increase in infections caused by the Enterobacteriaceae, while a reduction in the level of resistance to the third-generation cephalosporins were found (p = .07). There was a progressive increase in the occurrence rate of infections caused by Acinetobacter species (p = .05). Only 11 infections over the 4 yrs were caused by Enterococcus species. Staphylococcal infections were uncommon (5.7% of admissions), and the level of methicillin resistance did not change. No increase in aminoglycoside resistance occurred.

No long-term effects on antimicrobial resistance or the spectrum of nosocomial pathogens could be attributed to the use of selective decontamination of the digestive tract over a 2-yr period in a respiratory ICU admitting all categories of patients.