Gallus 1985.

Study characteristics
Methods	Study design: double‐blind, randomised. Cross‐over after 8 weeks; no washout period Country: Australia Setting: single centre Intention‐to‐treat: not mentioned
Participants	Number of participants randomly assigned: 47 Number of participants analysed: 38 (19 in each group) Exclusions postrandomisation: 9 Losses to follow‐up: 0 Age (mean): group A: 68 years, group B: 66 years Sex: group A: 17 males, 2 females, group B: 14 males, 5 females Inclusion criteria: stable claudication distance > 6 months; presence of peripheral vascular disease documented through clinical examination by vascular surgeon and supplemented by angiography or non‐invasive testing; aged > 50 years; pledge not to change smoking habits during trial; informed consent Exclusion criteria: vascular surgery or sympathectomy within previous 6 months; ischaemic leg ulcer or rest pain; exercise tolerance limited by conditions other than peripheral vascular disease; treatment with lipid‐lowering or antiplatelet drugs
Interventions	Treatment: pentoxifylline 400 mg bid for 1 week, then 400 mg tid for 7 weeks Control: placebo Duration: 8 weeks, then cross‐over for another 8 weeks; no washout phase (group A: placebo followed by pentoxifylline; group B: pentoxifylline followed by placebo)
Outcomes	Primary: geometric mean TWD and PFWD Secondary: ABI
Notes	Treadmill protocol: 4 km/h at 10% inclination Geometric mean PFWD and TWD expressed in metres only
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "A random number sequence was used to form the two treatment groups."
Allocation concealment (selection bias)	Unclear risk	Not mentioned.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and personnel blinded from allocation and held by hospital pharmacy.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Results were withheld from investigators during the study."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No incomplete outcome data.
Selective reporting (reporting bias)	Unclear risk	Protocol not available; insufficient information available to permit judgement.
Other bias	Low risk	Study appeared free of other bias.