Artificial nutrition is an inherent part of management in acute pancreatitis. However, there is no consensus regarding the optimal time of the commencement of feeding in these patients. Our aim was to compare the effect of enteral v. parenteral nutrition with regard to the time points when they were administered in the randomised controlled trials. The search was undertaken in the Cochrane Central Register of Controlled Trials, MEDLINE and Science Citation Index as well as in the proceedings of major gastroenterology meetings. The summary estimate of the effect associated with artificial nutrition was calculated using a random-effects model and presented as a risk ratio (RR) and 95 % CI. A total of eleven randomised controlled trials were included. When started within 48 h of admission, enteral nutrition, in comparison with parenteral nutrition, resulted in a statistically significant reduction in the risks of multiple organ failure (RR 0·44; 95 % CI 0·23, 0·84), pancreatic infectious complications (RR 0·46; 95 % CI 0·27, 0·77) and mortality (RR 0·46; 95 % CI 0·20, 0·99). After 48 h of admission, enteral nutrition, in comparison with parenteral nutrition, did not result in a statistically significant reduction in the risks of multiple organ failure (RR 0·73; 95 % CI 0·33, 1·63), pancreatic infectious complications (RR 0·31; 95 % CI 0·07, 1·34) and mortality (RR 0·67; 95 % CI 0·22, 2·10). Enteral nutrition is more effective than parenteral nutrition in reducing the risk of multiple organ failure, pancreatic infectious complications and mortality in patients with acute pancreatitis. The magnitude of these benefits may depend on the timing of the commencement of nutrition.