Prevention of respiratory tract infections is only possible when the pathogenesis is known. Three types of infection can be distinguished: primary endogenous infections, caused by pathogens carried in the throat at the commencement of mechanical ventilation, generally develop early and can only be prevented by intravenous antibiotics. Secondary endogeneous infections, caused by hospital-acquired pathogens, generally develop later and can be prevented by selective decontamination of the digestive tract (SDD). The GI-tract is decontaminated by oral nonabsorbable antibiotics and for oropharyngeal decontamination a sticky antibiotic ointment is used. To date 16 controlled SDD trials in intensive care have been fully published. In all except one study, the pneumonia rate decreased significantly from 40%-50% in controls to about 10% in SDD-treated patients. All studies showed a consistent reduction of ventilator days, ICU-stay and an improved outcome in SDD-treated patients. However, in only few studies did these differences reach statistical significance. Selection of resistant strains has not been observed during prolonged use of SDD. Sucralfate reduces the pneumonia rate compared to H2-blockers or antacids by not interfering with the gastric barrier. However, gastric colonization is reduced rather than eliminated and sucralfate has almost no effect on oropharyngeal or tracheal colonization. Whether sucralfate is significantly better than a placebo remains to be established. SDD is superior to sucralfate in preventing both colonization and infection.