POBBLE 2005.

Methods	RCT, parallel design
Participants	Total number of randomized participants: 103 Inclusion criteria: scheduled for infrarenal vascular surgery under GA Exclusion criteria: already taking beta‐blockers, if giving beta‐blockers would be dangerous (e.g. because of intolerance), receiving treatment for asthma, aortic stenosis, bradycardia, hypotension, perioperative beta‐blockade had already been shown to be beneficial, unstable angina or angina with a positive dobutamine stress test Type of surgery: elective infrarenal vascular surgery Baseline characteristics Intervention group (metoprolol) Age, median (IQR): 73 (61‐79) years Gender, M/F: 40/13 Control group (placebo) Age, median (IQR): 74 (66‐76) years Gender, M/F: 35/9 Country: UK Setting: multi‐centre; 4 hospitals
Interventions	Intervention group (metoprolol) Randomized, n = 55; losses = 5 (2 surgery cancelled; 3 death); analysed, n = 53 (for mortality), 50 (for other outcomes) (use of ITT analysis) Details: test dose given (25 mg or 50 mg metoprolol depending on weight) on day before surgery. If tolerated, then participants given a minimum of 2 additional oral doses of trial drug each morning and evening up to start of surgery. Then, 2 mg‐4 mg metoprolol in slow IV injection over 5‐10 min before induction of GA. Followed by oral treatment twice a day for 7 days. Extended for another 7 days if additional surgery was required within a week Control group (placebo) Randomized, n = 48; losses = 5 (4 surgery cancelled; 1 death); analysed, n = 44 (for mortality), 43 (for other outcomes) Details: placebo agent, given same as the intervention group
Outcomes	Outcomes measured/reported by study authors: fatal and non‐fatal cardiovascular events (MI, unstable angina, ventricular tachycardia, or stroke) within 30 days of operation, length of hospital stay, hypotension, bradycardia Outcomes relevant to the review: fatal and non‐fatal cardiovascular events (MI, ventricular tachycardia, or stroke), length of hospital stay (see notes below, mortality (at 30 days), hypotension (decrease in SBP > 25%), bradycardia (HR < 50 bpm)
Notes	Funding/declarations of interest: supported by The British Heart Foundation Study dates: July 2001‐March 2004 Note: we did not combine data in analysis for length of hospital stay, because data were reported as median values. Accounting for death, study authors reported a median stay (95% CI) of 9 ( 8‐12) days in the intervention group, and 12 (9‐19) days in the placebo group data for mortality at 2 years were not reported we contacted Prof Dr G Hamilton (University College London Medical School) by e‐mail, who kindly provided information about a trial listed on www.controlled‐trials.com/mrct (ISRCTN13072628), which is the POBBLE trial.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Central randomization via Web page
Allocation concealment (selection bias)	Low risk	See above
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "Anesthesiologists were unblinded to the drug allocation because those involved in the trial at participating centers would have immediately identified the active treatment and, for safety reasons, refused to collaborate in a blinded fashion"
Blinding of outcome assessors (detection bias) All outcomes	Low risk	Holter‐ECG results were coded centrally
Incomplete outcome data (attrition bias) All outcomes	Low risk	Small loss of participants (< 10%) because of cancellation of surgery, and death
Selective reporting (reporting bias)	Unclear risk	Study was retrospectively registered with a clinical trial register (ISRCTN13072628). It was not feasible to effectively assess risk of reporting bias from this report
Other bias	Low risk	Not detected