[Prevention of bacterial infection using selective intestinal decontamination in patients with cirrhosis admitted to intensive care. Controlled study in 120 patients].

All cirrhotic patients admitted on a medical intensive care unit, were included in a randomized trial of selective intestinal decontamination provided there was no infection on admission. The selective intestinal decontamination consisted of a regimen of 3 oral, nonabsorbable antibiotics for the 74 first patients (Neomycin 1 gr, Colistin 1.500.000 U, Nystatin 1.000.000 U, every 6 hours), then of norfloxacin, 400 mg BID for the following patients. The duration of treatment was at least 5 days. Of the 120 patients, initially randomized to receive or not the treatment, 26 were ultimately excluded, mainly (18 cases) because of infection present but unrecognized at the time of admission. Ninety four patients were thus compared for the efficiency of the treatment, 45 in the treated group and 49 in the not treated group. The results showed a significant reduction of the episodes of septicemia in the treated group (4 versus 12, P = 0.044). This reduction was evident only for septicemia due to gram negative germs. Mortality was unaffected. When the risk factors were studied, bacterial infection was linked to the degree of hepatic failure. We recommend selective intestinal decontamination for cirrhotic patients admitted on intensive care unit, particularly when hepatic function is poor.