The development of bacterial resistance during selective decontamination of the digestive tract (SDD) is controversial. We studied effects on bacterial resistance one year before and during a randomized, placebo-controlled trial of SDD in a surgical intensive care unit. We randomized patients within two different topical regimens (PTA, PCA) or placebo, administered four-times daily to both the oropharynx and gastrointestinal tract. All patients received intravenous ciprofloxacin (200 mg b.d.) for four days. Both SDD regimens successfully reduced aerobic Gram-negative intestinal colonization. There was no increase in resistance of Enterobacteriaceae or Pseudomonas aeruginosa. Acinetobacter calcoaceticus developed multi-resitance over one year, but differences between groups were not significant. We detected a shift towards Gram-positive organisms. Oxacillin-resistant Staphylococcus aureus increased in concert with ciprofloxacin resistance, from 17 to 80·7%, and frequencies of resistance were significantly higher in SDD patients ( P < 0·001). Resistance of coagulase-negative staphylococci (CNS) to oxacillin increased initially (25 to 66·9%), but values returned to baseline in controls. Ciprofloxacin resistance in CNS remained higher ( P < 0·001) in SDD-treated patients (52·5 vs. 23·3%). The incidence of late respiratory tract infections was unaltered by the prophylactic regimen (SDD 35·2%; Placebo 41·2%; n.s.). We cannot recommend SDD as a prophylactic tool in critically ill patients.