Kauer 1990.

Methods	Design: randomised double‐blinded No. eligible: 100 No. randomised: 78 No. analysed: 68 Drop‐outs/withdrawals: 10 (2 in treatment group 1, 4 in each of treatment groups 2 and 3: 5 had asymptomatic bacteriuria, 3 were given an incorrect antibiotic, 2 had abdominal not vaginal surgery) Years of recruitment: not reported Setting: Roman Catholic Hospital, Groningen, The Netherlands
Participants	Inclusion criteria: women ≥ 20 years of age having vaginal hysterectomy Exclusion criteria: allergy to study drugs, antibiotics within 48 hours of surgery, preexisting infection Age: mean 55 to 60 years Type of hysterectomy: vaginal
Interventions	Treatment 1: 1500 mg cefuroxime (second‐generation cephalosporin) Treatment 2: 500 mg metronidazole (antiprotozoal) Treatment 3: 1500 mg cefuroxime + 500 mg metronidazole Route: IV Single/multiple doses: single Timing of doses: 15 minutes preoperatively
Outcomes	Postoperative infection, early Urinary tract infection Pelvic infection Need for therapeutic antibiotics Adverse effects (narrative data only) Hospital length of stay Follow‐up: duration not clearly stated
Funding	Not stated
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	States that sequence was generated through "table of random numbers"
Allocation concealment (selection bias)	Low risk	States that "patients were assigned by the hospital pharmacist..."
Blinding (performance bias and detection bias) All outcomes	Low risk	States "vial and colour of the solution being indistinguishable...the observer was unaware of the antibiotics used..."
Incomplete outcome data (attrition bias) All outcomes	Low risk	Proportions of withdrawals and reasons for withdrawals fairly balanced across treatment groups
Selective reporting (reporting bias)	Unclear risk	Insufficient information to make a conclusive judgement
Other bias	Low risk	Baseline demographic characteristics similar between treatment groups