Pearl 1998.

Methods	A prospective randomised controlled study in a single institution. Power calculation was performed a priori. Number of participants randomised: 200 (95 in the early group, 105 in the delayed feeding group) Number of participants analysed: 195 (92 in the early group, 103 in the delayed feeding group); 5 participants were not measurable because of inoperable bowel obstructions and received gastrostomy tubes: (3 participants; 2 in the early group, 1 in the delayed feeding group) or died of multi‐organ system failure within 36 hours of surgery (1 in each group) Analysis: Available case analysis
Participants	Inclusion criteria: All gynaecologic oncology patients undergoing non laparoscopic intra‐abdominal surgery Mean age: 56.5 years (early group), 57.7 years (delayed feeding group) Underlying diagnosis:‐ Early group: cervical cancer 8.7%, ovarian cancer 30.4%, uterine cancer 38.0%, benign 22.8% ‐ Delayed feeding group: cervical cancer 13.6%, ovarian cancer 33.9%, uterine cancer 24.3%, benign 28.1% No significant difference in baseline characteristics between the 2 groups including age, disease and surgical procedure distribution, operating time, and estimated blood loss Location: State University of New York at Stony Brook, Stony Brook, New York, USA Enrolment period: February 1996 to March 1997 Exclusion criteria were not specified.
Interventions	Early group: Participants began a clear liquid diet on the 1st postoperative day and then advanced to a regular diet as tolerated. Delayed feeding group: Participants received nothing by mouth until return of bowel function (defined as the passage of flatus in the absence of vomiting or abdominal distention), then began a clear liquid diet, and advanced to a regular diet as tolerated. All participants had an orogastric tube placed intraoperatively and removed at the completion of surgery. Participants in either group who were unable to tolerate their diet were given nothing by mouth and received intravenous hydration until resolution of their symptoms, at which time they were restarted on a clear liquid diet and advanced as tolerated. A nasogastric tube was placed for intractable nausea, vomiting, or symptomatic abdominal distention. Standard criteria for discharge were used for all study participants.
Outcomes	Gastrointestinal information Morbidity (nausea, vomiting, abdominal distention, nasogastric tube use/duration) Diet tolerance on 1st attempt of clear liquid and regular diet/ time to tolerance Postoperative intervals (presence of bowel sounds, passage of flatus, start of clear liquid diet, start of regular diet) Other morbidities (febrile morbidity, pneumonia, wound complications, atelectasis) Hospital stay
Notes	Funding source: not reported Conflicts of interest: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "patients were randomized using a computer‐generated random number list." Comment: Computer‐generated sequence
Allocation concealment (selection bias)	Low risk	Comment: Sequentially numbered, opaque, sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: Because of the study's context, the participants could not be blinded. The physicians taking care of the participants were aware of their study allocation. It is possible that some outcomes (hospital stay and subjective intestinal morbidities) were influenced by the lack of blinding.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: The outcome assessors were not blinded (information from the study's author).
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "Five patients were nonevaluable, three in the early feeding group and two in the delayed feeding group. Of these, one patient in each group died of multiorgan system failure within 36 hours of surgery. The remaining patients had inoperable bowel obstructions, received gastrostomy tubes, and were placed on hospice care." Comment: Reasons for missing outcome data unlikely to be related to true outcome. Also, the missing outcome data were balanced in numbers across groups, with similar reasons for missing data.
Selective reporting (reporting bias)	Unclear risk	The study protocol was not published in a protocol registry. However, the report included expected outcomes.