Olszewer 1990.

Study characteristics
Methods	Study design: randomised, double‐blind for 10 treatments, then completed in an open single treatment fashion. Intention‐to‐treat: not stated Country: Brazil Setting: outpatient clinic
Participants	Number randomised: N = 10 (EDTA n = 5; placebo n = 5) Exclusions post‐randomisation: not stated Losses to follow up: not stated Age (mean years (range)): 47 (41 to 53) Gender: All male Inclusion criteria: peripheral vascular disease (Fontaine Stage II) ‐ intermittent claudication; walking test ‐ claudication between 100 m and 300 m; master exercise test ‐ claudication with < 40 steps; bicycle stress test ‐ claudication before 3 minutes at 50 km/hr Exclusion criteria: pain at rest or at night, gangrene
Interventions	Treatment: EDTA 10 mL (1.5 g) x 20 Control: distilled water 10 mL x 20 (although halfway through all control participants rolled‐over to active treatment for remaining 10 infusions) Duration of treatment: not stated Follow up: not stated
Outcomes	Walking distance measured by Walking Test Number of steps measured by Master Exercise Test Cycling time at 25 km/hr by Bicycle Test (observations made after 10 infusions and after 20 infusions)
Notes	Adequate allocation concealment initially, but the code was broken before end of study.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The study states that the 2 groups were "randomly" and equally divided
Allocation concealment (selection bias)	Unclear risk	The identification of the contents of each vial was in a sealed envelope by the manufacturer; it was not explicitly stated how the dispensing of the treatment was done.
Blinding of participants and personnel (performance bias) All outcomes	High risk	For the first 10 treatments, the blinding of participants and personnel was maintained; in the remaining 10 sessions, only the participants were blinded to treatment assignment and the placebo group was shifted to active treatment.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcome assessment was blinded for the first 10 infusions, however in the remaining 10 infusions the investigators broke the code and the placebo group was shifted to active treatment: Outcome assessment was therefore not blinded in this case.
Incomplete outcome data (attrition bias) All outcomes	High risk	The incompleteness in outcome data is due to the premature stoppage of the placebo arm and then the shifting of this arm to active treatment; the proposed number of sessions for both groups was not followed.
Selective reporting (reporting bias)	Low risk	All outcomes indicated were reported in tables with group differences specified.
Other bias	High risk	The breaking of the code and shifting of the control group into the intervention group because of a seemingly positive response after 10 infusions was premature and perhaps not part of the planned protocol; this casts doubt on the true effect of treatment as accurate assessments with a control group on the prespecified period could not be derived.