Delayed gastric emptying: the state of the highest level of evidence.

Delayed gastric emptying (DGE) has been regarded as the most common complication after pancreaticoduodenectomy (PD). Opinions about DGE and its incidence widely vary between studies and between institutions. To crystallize current concepts of DGE we resorted to a systematic literature search of level I evidence. We found 16 randomized controlled trials (RCTs) where DGE was measured but only 4 of these trials tested methods to influence DGE (erythromycin, enteral nutrition, or antecolic duodenojejunostomy). Constant heterogeneity for the definition of DGE was observed; 13 RCTs used 6 different clinical definitions based on some form of NG tube requirement after surgery, and the 3 remaining RCTs used non-clinical objective criteria. The most common element of the clinical definitions was the need for an NG tube >10 postoperative days. Ten RCTs used some form of this definition and the reported mean incidence of DGE was 17% however the range varied from 5% to 57%. The trials with the least number of cases appeared to have the widest variation in DGE incidence. We concluded after this systematic review that the disparate opinions about DGE could not be mediated with the highest level of evidence. The studies were underpowered or compromised by a lack of homogeneity in definition and design. The incidence of DGE cannot be succinctly measured; therefore the variables that influence DGE are not understood. We can begin to make progress by using the same definition such as the recently published definition provided by the International Study Group of Pancreatic Surgery.