Neary 2006.

Methods	RCT, parallel design
Participants	Total number of randomized participants: 38 Inclusion criteria: non‐elective, non‐cardiac emergency surgery, high risk of cardiac complications (previous MI or ischaemia, typical or atypical angina, stroke or transcient cerebral ischaemia) or minor risk (≥ 65 years of age, hypertension, current smoker, serum cholesterol concentration > 6.2 mmol/L, diabetes mellitus) Exclusion criteria: already on beta‐blockers, bradycardia, chronic obstructive airways disease or asthma, history of beta‐blocker intolerance, 2nd‐ or 3rd‐degree heart block, cardiovascular collapse or uncorrected hypovolaemia, anaesthetist does not think that participant should be given beta‐blocker Type of surgery: non‐elective, non‐cardiac emergency surgery (gastrointestinal resection surgery, major limb amputation, arterial reconstruction, orthopaedic procedures) Baseline characteristics not reported Country: UK Setting: single centre, hospital
Interventions	Intervention group (atenolol) Randomized, n = 18; losses = 0; analysed, n = 18 (use of ITT analysis) Details: 1.25 mg atenolol before start of GA, further doses given every 30 min during surgery up to a maximum of 4 doses. Postoperatively, participants were given IV atenolol 5 mg or oral medication (50 mg atenolol) once daily, when oral medication could be tolerated Control group (placebo) Randomized, n = 20; losses = 0; analysed, n = 20 Details: saline placebo, and then placebo oral medication, same as the intervention group
Outcomes	Outcomes measured/reported by study authors: mortality and non‐fatal cardiac events until 30 days after surgery, bradycardia (not defined), hypotension (not defined), MI. Long‐term follow‐up was also performed (2 years) Outcomes relevant to the review: mortality (30 days and at 2 years), bradycardia, hypotension
Notes	Funding/declarations of interest: grant from Gloucester Vascular Research Team Study dates: not reported Notes: early stopping: investigators planned to include 400 participants, but only enrolled 38 participants because of recruiting problems we did not include outcome data for bradycardia and hypotension in analysis because we were uncertain whether these outcomes were measured in both groups; study authors reported that 1 participant in the placebo group had bradycardia and 1 participant in the atenolol group had hypotension
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The study medication was provided in 1 box per participant containing placebo, as well as the active drug, in similar looking batches. The anaesthesiologist randomly chose 1 pack at the induction of anaesthesia. We classed this as a random method
Allocation concealment (selection bias)	Low risk	Quote: "The box also contained a sealed envelope with details of which packs contained the active drug in case of an emergency." Comment: because the sealed envelope was inside the box, we assumed that allocation was adequately concealed
Blinding of participants and personnel (performance bias) All outcomes	High risk	Anaesthetist had access to treatment allocation and was able to decide not to give beta‐blocking agents to a randomized participant
Blinding of outcome assessors (detection bias) All outcomes	Low risk	ECGs and lab reports were evaluated by "blinded cardiology staff"
Incomplete outcome data (attrition bias) All outcomes	Low risk	No apparent losses after randomization
Selective reporting (reporting bias)	Unclear risk	Study authors did not report prospective clinical trial registration or publication of a protocol. It was not feasible to effectively assess risk of reporting bias
Other bias	Low risk	Not detected