Iwasaki 2008a.

Methods	Multicenter randomised clinical trial with parallel group design (two interventions groups). Trial duration: 52 weeks.
Participants	Country: Japan. Number of patients randomised: 45, mean age 55 years (82% females). Inclusion criteria: ‐ a medical history and laboratory tests consistent with chronic cholestatic liver disease; ‐ positive antimitochondrial antibody or antipyruvate dehydrogenase complex (PDC);   ‐ serum alkaline phosphatases elevation of at least 1.5 times the upper limit of normal; ‐ the absence of biliary tract obstruction on imaging results; ‐ hyperlipoproteinaemia. Exclusion criteria: ‐ treatment with D‐penicillamine, corticosteroids, colchicine or immunosuppressive agents within 4 weeks; ‐ diagnosis of cirrhosis; ‐ diuretic‐resistant ascites, hepatic encephalopathy, haemorrhage from oesophageal or gastric varices; ‐ hyperbilirubinaemia (greater than 5.0 mg/dL); ‐ serum albumin level less than 3.0 g/dL; ‐ renal insufficiency; ‐ malignancy; ‐ pregnancy; ‐ below 19 years of age.
Interventions	Patients were randomly assigned to receive: Intervention group 1: bezafibrate (400 mg daily orally), n = 20; Intervention group 2: ursodeoxycholic acid (orally at a dose of 600 mg daily), n = 25. All patients had not been treated with ursodeoxycholic acid or bezafibrate within the previous four weeks.
Outcomes	Outcome measure(s): ‐ clinical events; ‐ laboratory data (serum alkaline phosphatases, serum gamma‐glutamyltransferase, serum alanine aminotransferase, IgM, total serum bilirubin, and total cholesterol and triglyceride levels); ‐ adverse events.
Notes	Additional information requested on 14th February 2011 and reply received on 16th February 2011 through personal communication with the principal author Dr. Shinji Iwasaki. Dr. Shinji Iwasaki provided data on the following: ‐ the method of sequence generation; ‐ the number of patients in each intervention group at the end of treatment; ‐ tables with numeric values for biochemical indices; ‐ adverse events; ‐ all‐cause mortality and liver‐related morbidity.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	It was generated by block method using computer‐generated random digits.
Allocation concealment (selection bias)	Low risk	Allocation was controlled by a central and independent randomisation unit, so that intervention allocations could not have been foreseen in advance of, or during enrolment.
Blinding (performance bias and detection bias) All outcomes	High risk	The trial was not blinded, so that the allocation was known during the trial.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The numbers and reasons for dropouts and withdrawals in all intervention groups were described.
Selective reporting (reporting bias)	Low risk	All expected outcomes are reported.
Other bias	Low risk	The trial appears to be free of other components that could put it at risk of bias.