Aims The CARDia trial randomised diabetic patients to coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI) and concluded that PCI is a potentially safe and feasible alternative to CABG in selected patients with diabetes mellitus (DM) and multivessel coronary artery disease. The impact of insulin treatment on clinical outcomes after revascularisation is unclear. The present study is a sub group analysis of the CARDia trial comparing the cardiovascular outcomes at 12 months following revascularisation between the insulin treated (IT) and non-insulin treated (NIT) group.

Methods 508 patients with an established diagnosis of DM and de novo coronary artery disease were identified and randomised to CABG or PCI. Of those, 316 patients were treated with oral antidiabetic medication and the rest were treated with additional subcutaneous insulin injections. Demographics, clinical presentation, history, haemodynamic parameters, anti diabetic therapy, concomitant medications, duration of DM and HBA1C were documented. Death, stroke and myocardial infarction were classified as the primary outcome events. The secondary outcome events included death, MI, Stroke, repeat revascularisation and TIMI major bleed. The clinical results of patients in the IT and NIT groups were compared.

Results There were 192 patients in the IT group (37.8%). Asian patients constituted one fifth of the total population with a slightly higher representation (24.5% vs 21.6%) in the NIT. The clinical severity of dyspnoea, heart rate, systolic and diastolic BP, body mass index, risk factors for coronary artery disease appeared similar in the IT and NIT groups, but more patients in the IT group had a prior MI (30.7% vs 19.6%, p=0.004) and duration of diabetes was longer in the IT group (14 vs 6 yrs, p<0.001). For the comparison of CABG vs PCI for the primary outcome events the HR and 95% CI in the IT and NIT groups respectively were 1.66 (0.76 to 3.76) and 1.01 (0.51 to 2.01). For death, MI, stroke, repeat revascularisation they were 2.47 (1.18 to 5.20) in the IT and 1.41 (0.71 to 2.57) in the NIT group. The results suggest that IT patients may have a worse outcome with PCI compared to CABG, whereas no difference was found for NIT patients.

Conclusion Our data suggest that insulin treatment is a marker for higher risk for PCI when compared with CABG. Treatment with insulin rather than diabetic status alone should be considered when choosing the mode of revascularisation.