Margarit 2005.

Methods	Trial design: randomised, single‐centre clinical trial Mean follow‐up: 44 months (range: 3 to 60) Study duration: randomisation from October 1998 to September 2000, 5 years from randomisation Language: English Type of information: journal article Judgement on quality: high risk
Participants	Setting: Liver Transplantation Unit, Hospital General Vall Hebron, Barcelona, Spain Allocation of participants: 63 participants, 33 allocated to Intervention A, 30 allocated to Intervention B Sex ratio: total: 43 (71.7%) males, 17 (28.3%) females Intervention A: 25 (78.1%) males, 7 (21.9%) females Intervention B: 18 (64.3%) males, 10 (35.7%) females Mean age: total: not reported Intervention A: 56 ± 8 Intervention B: 57 ± 7 Indication (no. (%)): HCV cirrhosis: total: 35 (58.3%), Intervention A: 15 (46.9%), Intervention B: 20 (71.4%) Alcoholic cirrhosis: total: 16 (26.7%), Intervention A: 11 (34.4%), Intervention B: 5 (17.9%) HBV cirrhosis: total: 5 (8.3%), Intervention A: 2 (6.3%), Intervention B: 3 (10.7%) Cryptogenic cirrhosis: total: 2 (3.3%), Intervention A: 2 (6.3%), Intervention B: 0 (0%) Haemochromatosis: total: 2 (3.3%), Intervention A: 2 (6.3%), Intervention B: 0 (0%) Type of donor: not reported Inclusion criteria: first elective liver transplant, informed consent Exclusion criteria: renal failure, preoperative steroid consumption
Interventions	Intervention A: no intervention Intervention B: methylprednisolone: 100 mg twice daily tapered to 20 mg/day by day 6 and tapered to complete stop at 3 months if possible Concomitant immunosuppression: Tacrolimus: 0.05 mg/kg twice daily aiming for trough levels of 10 to 15 ng/mL for "a few weeks" and 8 to 12 ng/mL thereafter
Outcomes	Mortality, infection, bacterial infection, viral infection, fungal infection, toxicity, HCV recurrence, severity of recurrent hepatitis C, renal insufficiency, de novo hypertension, de novo diabetes mellitus, dyslipidaemia, neurological complications, diarrhoea
Notes	Cross‐over between intervention arms: no Sample size calculation: not reported Sources of funding: Fujisawa GM
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: Generation of randomisation sequence not described
Allocation concealment (selection bias)	Low risk	Quote from email correspondence with author (Bilbao, I) in response to the question "How was the randomisations sequence generated?": "68 closed envelopes numbered 1 to 68 respectively, were prepared before starting the trial. The envelopes were opened consequently when the surgery started." Comment: Sealed, opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: Blinding of participants and medical staff not performed
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote from email correspondence with author (Bilbao, I) in response to question "Were the outcome assessors blinded to the patient's treatment?": "NO". Quote from email correspondence with author (Bilbao, I) in response to question "Were the pathologists confirming rejection blinded to the patient's treatment?": "YES". Comment: Blinding of pathologists performed; blinding of all other outcome assessors not performed
Incomplete outcome data (attrition bias)	Unclear risk	Quote from the publication: "Three patients were excluded after randomization because of perioperative death (n = 2) and positive cross‐match (n = 1)." Quote from email correspondence with author (Bilbao, I) in response to question "Of the three patients who died after randomization, had any of the patients begun treatment? If this is the case which group were the two deaths part of and which was the patient with the positive cross‐match part of?": "One death belonged to Tacro group and the other to the Tacro + Prednisone. The patient with positive crossmatch belonged to Tacro group. The death in Tacro group did not begin treatment because intraoperative problems and at the end of the surgery received standard immunosuppression at that time (Tacro + steroids). The one in Tacro + steroids died at 3 rd day post‐op and began treatment. The patient with positive crossmatch began treatment for 36 hours, but then was changed to standard triple therapy" Comment: Three participants removed from analysis following randomisation following data‐dependent processes not described in exclusion criteria. It was possible to incorporate some of the data on mortality from these participants based on email correspondence with the author. Otherwise, the number of withdrawals and reasons for withdrawal were not reported
Selective reporting (reporting bias)	Low risk	Comment: All outcomes appear to have been reported
Other bias	Unclear risk	Comment: No sample size calculation reported
Free of early stopping?	Low risk	Comment: Study not stopped early
Free of baseline imbalance?	High risk	Quote from the publication: "No differences were found between treatment groups except for a higher incidence of graft steatosis in TACRO + ST and of HCV‐positive patients in the TACRO group. Primary graft dysfunction secondary to ischaemic–reperfusion injury that could affect tacrolimus metabolism and pharmacokinetic parameters was similar in both groups." Comment: Baseline imbalance observed in recipient HCV cirrhosis and donor graft steatosis