Revascularisation meta‐analysis and randomisation

Biccard and Rodseth [1] are to be congratulated on their detailed analysis of an important clinical question, namely the benefit of different types of coronary revascularisation before high-risk non-cardiac surgery. However, I would like to point out an important bias to their results. Although the authors have been careful to select only randomised controlled trials, what they are comparing (percutaneous coronary interventions vs coronary artery bypass grafts) was not actually randomised in the studies they have analysed. That is, the two studies randomised patients to receive pre-operative revascularisation or not; but whether those randomised to revascularisation received percutaneous coronary interventions or coronary artery bypass grafts was decided by the treating physicians. So it is possible, likely even, that patients with comorbidities deemed unsuitable for coronary artery bypass grafts, were offered percutaneous coronary interventions; having survived the latter, they then did poorly after their high-risk non-cardiac surgery. Patients who underwent coronary artery bypass grafts may have done better simply because they were (non-randomly) better selected for ability to survive major surgery. This by no means detracts from an excellent meta-analysis; but merely re-emphasises the importance of interpreting the conclusions in the light of each individual patient’s risk and circumstances.