Karanikolas 2006.

Methods	Double‐blind (participants, outcome assessor) placebo‐controlled, randomized clinical trial Sequence generation was randomized Follow‐up: 6 months
Participants	Participants: 65 adults in a university setting in Patras, Greece Operation: lower limb amputation with pain score > 60/100 VAS 48 h prior to amputation 5 groups, group size: 13 Age: group means ranging 69.2 to 74.3 with largest SD 13 Men/women: 35/53
Interventions	Group 1 (Epi/Epi/Epi): preop: lumbar epidural analgesia bupivacaine (0.2%, fentanyl 2 µg/mL at 4 mL/h to 8 mL/h) for 48 h, GA preincision: epidural bupivacaine (0.5% 10 mL to 15 mL, fentanyl 100 µg), post‐op epidural bupivacaine (0.2% fentanyl 2 µg/mL at 4 mL/h to 8 mL/h) Group 2 (PCA/Epi/Epi): preop: PCA fentanyl (IV, demand 25 µg, lockout 20 min), preincision: epidural bupivacaine (0.5% 10 mL to 15 mL, fentanyl 100 µg), post‐op epidural bupivacaine (0.2%, fentanyl 2 µg/mL at 4 mL/h to 8 mL/h) Group 3 (PCA/Epi/PCA): preop: PCA fentanyl (IV, demand 25 µg, lockout 20 min), preincision: epidural bupivacaine (0.5% 10 mL to 15 mL, fentanyl 100 µg), post‐op PCA fentanyl (IV, demand 25 µg, lockout 20 min) Group 4 (PCA/GA/PCA): preop: PCA fentanyl (IV, demand 25 µg, lockout 20 min), general anaesthesia with LMA, sevoflurane and remifentanil infusion, post‐op PCA fentanyl (IV, demand 25 µg, lockout 20 min) Group 5 (control/GA/control): preop: meperidine (50 mg 4‐6 x/d IM) paracetamol/codeine 30/500 mg orally plus as‐needed IV paracetamol 650 mg 3 x/d and parecoxib 40 mg 2 x/d, GA with LMA, sevoflurane and remifentanil infusion, post‐op: meperidine (IM) paracetamol/codeine 30/500 mg orally plus as‐needed IV paracetamol 650 mg 3 x/d and parecoxib 40 mg 2 x/d Immediate pain control: significantly improved preop and post‐op
Outcomes	Dichotomous: phantom limb pain at 6 months Continuous: VAS and McGill pain questionnaire and phantom limb pain frequency scores for phantom and stump pain at 6 months Effective regional anaesthesia not reported, but interventions reduced the severity of pain pre‐ and postoperatively.
Notes	There are minor discrepancies regarding the dosing described between the preliminary report of the ongoing registered trial (Karanikolas 2006) and the final report. We reported the treatment according to the latest publication. We contacted the study author for confirmation and additional information, but received no response. Hence, we could only use the data extracted from the publications and the information provided on clinicaltrials.gov/ct2/show/NCT00443404. Funding sources: "support was provided solely from institutional and/or departmental sources." Conflicts of interest: no conflict of interest statement was provided
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Described as "prospective, randomized, clinical trial", with “computer generated blocks with five treatment groups and 13 patients per group.”
Allocation concealment (selection bias)	Low risk	Quote: “sequentially numbered sealed envelope... concealed until after consent was obtain.” Recruitment, outcome assessment and protocol management clearly separated.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The trial is described as "double‐blind" in the title. Detailed description of blinding procedures. Quote: “control group patients had an epidural catheter placed subcutaneously.” D.A. i.e. the person “responsible for adjusting the epidural...” may not have been blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Detailed description of blinding procedures. Quote: “a second blinded investigator interviewed all participants.” “A third blinded investigator conducted all interviews during the analgesic protocol.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Only minor attrition is reported, and attributed to groups. Seemingly, attrition affected mainly the control groups. ITT analysis is reported. Per protocol or ITT analysis did not change results.
Selective reporting (reporting bias)	Low risk	Protocol review and primary outcomes fully reported on
Null bias	Low risk	Quote: "all patients had severe ischemic pain before analgesia started, but pain scores improved markedly and were significantly lower in all intervention groups compared with control at all times while the protocol was in effect"