Knudston 2002.

Study characteristics
Methods	Study design: randomised, double‐blind, placebo‐controlled clinical trial. Intention‐to‐treat: yes Country: Alberta, Canada Setting: outpatient clinic
Participants	Number randomised: N = 84 (EDTA n = 41; placebo n = 43) Exclusions post‐randomisation: none Losses to follow‐up: no dropouts were reported, however there were 4 people unable to complete protocol in the control group and 2 unable to complete protocol in the treatment group. Age (mean years ± SD): EDTA 66 ± 9.1; placebo 65 ± 8.5 Gender (M n (%)): EDTA 33 (85.4%); placebo 32 (83.7) Inclusion criteria: aged 21 years or older with proven coronary artery disease or documented myocardial infarction and stable angina while receiving optimal therapy; 1 mm of horizontal or down sloping ST‐segment depression from the isoelectric line 80 milliseconds after the J point on treadmill test 2 and 14 minutes from the onset of exercise. Exclusion criteria: planned revascularisation, previous chelation therapy, heart failure, inability to walk on the treadmill, resting ECG changes that would interfere with ischaemic assessment, abnormal renal or liver function, untreated lipid abnormality at time of randomisation.
Interventions	Treatment: EDTA weight‐adjusted with a maximum total dose for each treatment of 3 g in 500 mL of 5% dextrose in water x 33 infusions Control: 500 mL of 5% dextrose in water with 20 mL 0.9% sodium chloride x 33 infusions Each 5% dextrose in water solution also contained 750 mg of magnesium sulphate, 5 g of ascorbic acid and 5 g of sodium bicarbonate (titrated to physiologic pH). All participants also received oral multivitamin therapy, 2 tablets 3 times daily except on treatment days. Duration of treatment: 27 weeks total; twice weekly for 15 weeks and once monthly for an additional 3 months. Follow up: 1 year from randomisation
Outcomes	Change in time to reach at least 1mm of ST‐segment depression at the 27‐week evaluation Functional reserve by determination of VO2 max and time to reach anaerobic threshold. Quality of life measured by the Duke Activity Status Index, Health Status Survey Short Form‐36, and Seattle Angina Questionnaire. Other clinical events
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomised in blocks of 10.
Allocation concealment (selection bias)	Low risk	Hospital pharmacist assigned the randomised therapy and prepared 'indistinguishable' solutions.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	As this was a double‐blind study, blinding of participants and personnel to treatment was assured.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinding of outcome assessment was assured.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Six people were unable to complete protocol; four in the placebo group and two in the treatment group; clinical events were presented on an intent‐to‐treat basis. There were apparently no dropouts.
Selective reporting (reporting bias)	Low risk	Only 39 participants per group were analysed with regards to outcomes involving subjective and treadmill data for between‐group comparisons; for clinical events, all participants were included in the between‐group analysis
Other bias	High risk	Despite utilising a randomisation process, there was more multivessel disease, nitrate use and triple therapy at baseline in the placebo group and more past myocardial infarction in the chelation group. These differences may tend to favour chelation and therefore the conclusion of 'no effect' is strengthened.