Gastric lavage is usually used for gastric preparation before endoscopy in patients with upper gastrointestinal bleeding. However, the benefit-risk balance of putting a nasogastric tube in these patients is not clearly defined. This randomized trial is aimed to determine if the use of erythromycin IV before endoscopy could avoid to put a gastric tube for the management of upper gastrointestinal bleeding.

Acute upper gastrointestinal haemorrhage is one of the main digestive emergencies involving hospital admission. Endoscopic examination plays a key role to determine the cause of the bleeding and to carry out a therapeutic procedure. Endoscopic performance depends on the quality of the examination that may be hampered by residual blood in the gastric cavity. Gastric lavage is usually performed to clear the stomach. However, several teams consider that it is possible to avoid putting a nasogastric tube because it is ineffective in half of the patients, disagreeable in most of them, can induce side effects and need a long time work for nurses. It could be replaced by the use of Erythromycin. There are no official recommendations in that field. Recent studies have shown that Erythromycin, a macrolide antibiotic with gastro kinetic properties can accelerate gastric emptying by inducing gastric contraction. This motilin receptor agonist could improve the gastric cleaning and the quality of endoscopic examination and decrease its duration.

This randomized trial is aimed to determine if the use of erythromycin IV before endoscopy could avoid putting a gastric tube for the management of upper gastrointestinal bleeding. It is a prospective, controlled, randomized, multicentric study with a blind assessment of the main criteria. All patients aged more than 18 years with an acute upper gastrointestinal bleeding, defined by melena or hematemesis, managed by emergency department are enrolled. 270 patients are expected. Informed consent including for endoscopic examination and no contraindication for using Erythromycin (QT enlargement) is needed. Patients are randomized in three groups: Erythromycin alone, nasogastric tube with gastric lavage alone or both Erythromycin and nasogastric tube. Patients are followed-up until first month after bleeding. The main criteria are the visualization of the gastric tract and the other criteria are rebleeding until D 30, transfusion, gastric tube or erythromycin complications.

Inclusion Criteria:

• Hematemesis or melaena
• No QT enlargement

Exclusion Criteria:

• Refusing endoscopy
• Glasgow < 15