Gupta 2006

Methods	Triple‐blinded (participants, providers, outcome assessors) randomized placebo‐controlled trial Sequene generation by computer‐generated randomized numbers Follow‐up: 3 months
Participants	Participants: 60 men from a university hospital in Orebro, Sweden Operation: radical retropubic prostatectomy (for prostatic cancer) 2 groups, size: 28/28 (completed) Age (± SD), group 1, 2: 64.5 (± 4.9), 61.1 (± 4.3) All male participants Exclusion criteria: age > 70 Remarks: Gleason score, median (range), group 1, 2: 6 (5‐9), 6 (5‐9)
Interventions	Group 1 (epidural group): on arrival to PACU, ropivacaine‐fentanyl‐adrenaline epidurally at 10 mL/h, IV PCA with 0.9% saline (bolus dose 1 mL, lockout 6 min, used NRS > 3). Group 2 (placebo group): on arrival to PACU, 0.9% saline via epidural at 10 mL/h, IV PCA with 1 mg/mL morphine (bolus dose 1 mg, lockout 6 min, used NRS > 3). In both groups, preoperative anxiolysis with 10 mg diazepam oral 1 h before scheduled surgery and 1 mg‐2 mg midazolam as needed during catheter placement. Standardized placement of epidural at T10 to 12 interspace, tested using 3 mL mepivacaine 2% with adrenaline then bolus dose of 3 mL to 4 mL mepivacaine 2% with adrenaline. Sensory blockade at T8 level. Standardized GA with propofol (participants 1‐55) or thiopentone (participants 56‐60), fentanyl, rocuronium, nitrous oxide in O2, sevoflurane. Intraoperative analgesia with 2% mepivacaine with 2 mL/h‐5 mL/h adrenaline by epidural infusion in all participants. Immediately before transfer to PACU epidural infusion was turned off. In PACU, nurse allowed to administer 1 mg‐2mg morphine bolus as needed if NRS > 5. 1 g paracetamol oral before surgery and every 6 h post‐op during hospitalization. Adjuvants: adrenaline Immediate post‐op pain control: significantly improved
Outcomes	Dichotomous: none Continuous: SF‐36 at 3 months Adverse effects: postoperative nausea, vomiting, sedation and bleeding were reported
Notes	We contacted study author for clarification on attrition, source of funding and conflict of interest but received no response. Funding sources: source of funding not reported. Conflicts of interest: conflict of interest statement not provided.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "computer‐generated randomized numbers", randomized "after successful insertion of the epidural catheter"
Allocation concealment (selection bias)	Low risk	Quote: "every precaution was taken to achieve double blinding...hospital pharmacy sent two double‐blinded bags"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "the patients and surgeons, anaesthesiologists and nurses involved in patient treatment were unaware of method of analgesia and every precaution was taken to achieve double blinding"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "the SF‐36 was given before and 1 and 3 months after the operation to each patient". Participants, as well as providers, were blinded and the participants filled out the questionnaire themselves
Incomplete outcome data (attrition bias) All outcomes	Low risk	60 participants were randomized, 4 participants were excluded after randomization with reasons and group assignments listed and balanced between groups.
Selective reporting (reporting bias)	Low risk	Primary outcomes fully reported
Null bias	Low risk	Quote: "median pain at rest at the incision site was low (< 4) and significantly lower in group E compared with group P at 4–24 h after the operation"