PREVENTION OF INFECTION IN CRITICALLY ILL PATIENTS BY SELECTIVE DECONTAMINATION OF THE DIGESTIVE-TRACT

Objective: To determine whether selective decontamination of the digestive tract using oral and nonabsorbable antimicrobial agents and parenteral cefotaxime prevents infection in critically ill patients. Design: Randomized, controlled trial without blinding. Setting: Surgical trauma and medical intensive care units in a tertiary referral hospital. Patients: One hundred fifty patients admitted to surgical trauma and medical intensive care units during a 3-year interval, whose condition suggested a prolonged stay (> 3 days). Intervention: Patients were randomly allocated to an experimental group (n = 75) that received cefotaxime, 1 g intravenously every 8 hours for the first 3 days only, and oral, nonabsorbable antibiotics (gentamicin, polymyxin, and nystatin by oral paste and oral liquid) for the entire stay in the intensive care unit. Control patients (n = 75) received usual care. Measurements: The number of infections, total hospital days, and deaths, as well as the number of days in intensive care unit, were recorded. Results: Control patients experienced more infections (36 compared with 12, P = 0.04), including bacteremias (14 compared with 4, P = 0.05) and pulmonary infections (14 compared with 4, P = 0.03). Although total hospital days, days in intensive care, and the overall death rate all were lower in the treatment group, these differences were not statistically significant. Clinically important complications of selective decontamination of the digestive tract were not encountered. Conclusions: Selective decontamination of the digestive tract decreases subsequent infection rates, especially by gram-negative bacilli, in selected patients during long-term stays in the intensive care unit.