In the early days of intensive care rates of overt gastrointestinal bleeding were high and it soon became widely accepted that some form of prophylaxis was required. This view was supported by randomised trials indicating that the incidence of clinically important bleeding can be reduced by administration of either H₂receptor antagonists or antacids.1 Observational studies, however, suggested that the higher gastric pH consequent on such treatment was associated with bacterial overgrowth in the stomach, tracheobronchial colonisation and nosocomial pneumonia. The publication of a randomised trial indicating that the cytoprotective agent sucralfate was associated with a trend toward a lower incidence of ventilator associated pneumonia (VAP)2 had a considerable influence on clinical practice, as did studies documenting that with modern intensive care the incidence of clinically important bleeding is low, especially in those receiving enteral nutrition. Intensive care clinicians have therefore been faced with two fundamental questions: which patients, if any, should receive prophylaxis against gastrointestinal bleeding and which agent should be used? Given this uncertainty the publication of the large (1200 patients), prospective, randomised, controlled trial by Cook et alcomparing the …