To demonstrate, during a prospective randomized medico-economic study, in morbidly obese patients (BMI > 45 kg/m²) selected for a laparoscopic gastric by-pass that, in comparison to usual care, a temporary intra-gastric balloon (6 months) decreases medical costs and peri-operative morbidity.

On inclusion, comorbidities associated with obesity will be recorded. After recording of the informed consent patients will be randomized in two groups: usual care or intra-gastric balloon. In the latter, a second randomization will be performed between air-filled balloon or water-filled balloon. Six months later, after withdrawal of the balloon (if required), laparoscopic gastric by-pass will be performed. Thereafter, patients will be followed during 6 months.

On each visit, inclusion, 3 months, 6 months, surgery, 1 month and 6 months post surgery, a medical and biological evaluation will be systematically performed.

The main criteria of judgement is the rate of admission in ICU during more than 24 hrs during the 30 days following surgery.

The secondary criteria of judgement is: the differences in pre-operative weight loss between the 2 groups, number of complications occurring during the 30 days after surgery, differences in different scales of quality of life (SF36, IWQOL-lite), differences in operative time, length of hospital stay, readmissions, and medical cots between the two strategies. Efficacy and tolerance of each type of balloon (air-filled or water-filled) will be compared in the intra-gastric balloon group.

Inclusion Criteria:

• Age between 18 and 75 years
• Patients with morbid obesity with BMI > 45 kg/m² requiring a gastric by-pass by laparoscopy.
• No efficacy of previous treatment on obesity
• Firmed informed consent

Exclusion Criteria:

• Patients in whom laparoscopy is contra-indicated
• Indication of gastric by-pass by laparoscopy retained without multidisciplinary advice
• Morbid obesity present since less than 5 years
• Severe associated pathology, non-related to obesity, impairing vital prognosis on a short or mean time
• Past history of gastric surgery, or gastric complication related to gastric lap-band
• Lack of health insurance
• Patient refusing to be followed 6 months before and after surgery
• Drug abuse
• Pregnancy or foreseeable pregnancy during the study
• Patients taking anticoagulant agents or steroids
• Patients with hiatal hernia > 4 cm
• Patients with active ulcerated esophagitis (grade C-D in Los Angeles Classification)
• Patients with duodenal or gastric ulcer
• Patients with gastric or esophageal varices
• Patients with NSAIDs not taking PPI
• Patients with bulimia