ABSTRACT: BACKGROUND: In predicted severe acute pancreatitis, infections have a negative effect on clinical outcome. A start of enteral nutrition (EN) within 24 hours of onset may reduce the number of infections as compared to the current practice of starting an oral diet and EN if necessary at 3-4 days after admission. METHODS/DESIGN: The PYTHON trial is a randomised controlled, parallel-group, superiority multicenter trial. Patients with predicted severe acute pancreatitis (Imrie-score >/= 3 or APACHE-II score >/= 8 or CRP > 150 mg/L) will be randomised to EN within 24 hours or an oral diet and EN if necessary, after 72 hours after hospital admission.During a 3-year period, 208 patients will be enrolled from 20 hospitals of the Dutch Pancreatitis Study Group. The primary endpoint is a composite of mortality or infections (bacteraemia, infected pancreatic or peripancreatic necrosis, pneumonia) during hospital stay or within 6 months following randomisation. Secondary endpoints include other major morbidity (e.g. new onset organ failure, need for intervention), intolerance of enteral feeding and total costs from a societal perspective. DISCUSSION: The PYTHON trial is designed to show that a very early (< 24 h) start of EN reduces the combined endpoint of mortality or infections as compared to the current practice of an oral diet and EN if necessary at around 72 hours after admission for predicted severe acute pancreatitis. TRIAL REGISTRATION: ISRCTN: ISRCTN18170985.