Steed 2002.

Methods	A prospective randomised controlled study in a single institution. Participants were randomised by the clinic nurses according to a computer‐generated random number list. Power calculation was performed a priori. Number of participants randomised: 107 Number of participants analysed: 96 (47 in the early group, 49 in the delayed feeding group); 7 women were excluded because of intraoperative injury of the gastrointestinal tract, and 4 were excluded because of self withdrawal. Analysis: Available case analysis
Participants	Inclusion criteria: Gynaecologic oncology and uro‐gynaecology patients undergoing laparotomy; mean age: 50 years (early group), 52 years (delayed feeding group) Underlying diagnosis of malignancy: 63.0% (early group), 54.0% (delayed feeding group) No significant difference in baseline characteristics between the 2 groups including age, body mass index (BMI), race, malignancy, bowel preparation, procedure, type of incision, operating time, blood loss, need for blood transfusion, and placement of a suprapubic catheter. However, there were significantly more women in early group who received epidural analgesia for pain treatment (55% vs 37%). Location: Royal Alexandra hospital, Edmonton, Alberta, Canada Enrolment period: October 2000 to June 2001 Exclusion criteria: Pregnancy, postoperative intensive care unit admission, endotracheal or nasogastric intubations in the immediate postoperative period, perioperative hyperalimentation, history of gastrointestinal surgery (excluding appendectomy) or bowel obstruction, history of pelvic or abdominal radiation, and history of peritonitis
Interventions	Early group: Participants received clear fluids on the 1st postoperative day. After 500 ml of clear fluids was tolerated, a regular solid diet was given. Delayed feeding group: Participants received nothing by mouth until at least 2 of the following signs of return of bowel function were present: 1) presence of bowel sounds; 2) passage of stool or flatus; 3) subjective hunger. They were then started on clear fluids, and advanced to a regular diet in a stepwise fashion. If there was evidence of ileus (defined as > 2 episodes of emesis of at least 100 ml each within a 24‐hour time period, with associated abdominal distention and no bowel sounds), participants were treated with restriction of oral intake, intravenous fluids, and nasogastric suction, if necessary. Discharge criteria included the tolerance of a solid diet, passing flatus, and the discontinuance of intravenous fluids and medications. Participants were not required to have had a bowel movement.
Outcomes	Hospital stay Gastrointestinal information Morbidity (incidence and episodes of postoperative ileus) Time to tolerance of a solid diet Other morbidities (wound infection, deep venous thrombosis, urinary tract infection, pneumonia, pulmonary oedema)
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 prospectively randomized with a computer‐generated random number list." Comment: Computer‐generated sequence
Allocation concealment (selection bias)	High risk	Comment: Open random number list
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	Low risk	Comment: The outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "Seven women were excluded because of intraoperative injury of the gastrointestinal tract, and 4 patients were excluded because of self‐withdrawal." Comment: Reasons for missing outcome data unlikely to be related to true outcome.
Selective reporting (reporting bias)	Unclear risk	The study protocol was not published in a protocol registry. However, the report included expected outcomes.