Randomized, controlled trial of selective digestive decontamination in 600 mechanically ventilated patients in a multidisciplinary intensive care unit

OBJECTIVE: To evaluate the efficacy of two regimens of selective decontamination of the digestive tract in mechanically ventilated patients. DESIGN: Prospective, randomized, concurrent trial. SETTING: Multidisciplinary intensive care unit (ICU) in a 1,800-bed university hospital. PATIENTS: Consecutive patients (n = 660) who were likely to require mechanical ventilation for at least 48 hrs were randomized to one of three groups: conventional antibiotic regimen (control group A); oral and enteral ofloxacin-amphotericin B (group B); and oral and enteral polymyxin E-tobramycin-amphotericin B (group C). Both treatment groups received systemic antibiotics for 4 days (ofloxacin in group B and cefotaxime in group C). INTERVENTIONS: Patients were randomized to receive standard treatment (control group A, n = 220), selective decontamination regimen B (group B, n = 220), and selective decontamination regimen C (group C, n = 220). After early deaths and exclusions from the study, 185 controls (group A) and 193 (group B)/200 (group C) selective decontamination regimen patients were available for analysis. MEASUREMENTS AND MAIN RESULTS: Measurements included colonization and primary/secondary infection rate, ICU mortality rate, emergence of antibiotic resistance, length of ICU stay, and antimicrobial agent costs. The study duration was 19 months. The patient groups were fully comparable for age, diagnostic category, and severity of illness. One third of patients in each group suffered a nosocomial infection at the time of admission. There was a significant difference between treatment group B and control group A in the number of infected patients (odds ratio of 0.42, 95% confidence interval of 0.27 to 0.64), secondary lower respiratory tract infection (odds ratio of 0.47, 95% confidence interval of 0.26 to 0.82), and urinary tract infection (odds ratio of 0.47, 95% confidence interval of 0.27 to 0.81). Significantly more Gram-positive bacteremias occurred in treatment group C vs. group A (odds ratio of 1.22, 95% confidence interval 0.72 to 2.08). Infection at the time of admission proved to be the most significant risk factor for subsequent infection in control and both treatment groups. ICU mortality rate was almost identical (group A 16.8%, group B 17.6%, and group C 15.5%) and was not significantly related to primary or secondary infection. Increased antimicrobial resistance was recorded in both treatment groups: tobramycin-resistant enterobacteriaceae (group C 48% vs. group A 14%, p < .01), ofloxacin-resistant enterobacteriaceae (group B 50% vs. group A 11%, p < .02), ofloxacin-resistant nonfermenters (group B 81% vs. group A 52%, p < .02), and methicillin-resistant Staphylococcus aureus (group C 83% vs. group A 55%, p < .05). Antimicrobial agent costs were comparable in control and group C patients; one third less was spent for group B patients. CONCLUSIONS: In cases of high colonization and infection rates at the time of ICU admission, the preventive benefit of selective decontamination is highly debatable. Emergence of multiple antibiotic-resistant microorganisms creates a clinical problem and a definite change in the ecology of environmental, colonizing, and infecting bacteria. The selection of multiple antibiotic-resistant Gram-positive cocci is particularly hazardous. No beneficial effect on survival is observed. Moreover, selective decontamination adds substantially to the cost of ICU care.