UNLABELLED: For many years, coronary artery by-pass graft (CABG) remained the only effective treatment of multivessel disease compared to medical treatment. The first technical revolution was in 1977 when Gruentzig introduced balloon percutaneous transluminal coronary angioplasty (PTCA), the 2nd in the 90's with the developments of stents and antiaggregant protocols. The equipment for PTCA became more and more sophisticated, and the skill of cardiologists greater. In the 90's, interventional cardiology played a predominant role in revascularization as the number of CABG decreased at the same time, and emergency CABG for bail out almost disappeared (0% to 0.5%). Systematic stenting decreased the need for repeat revascularization to about 18-20% nowadays in the majority of centers, except in diabetic patients. Despite this fact restenosis remains the pitfall of angioplasty, mostly in diabetic patients presenting multivessel disease in which surgery with "all arterial grafts" gives good long term RESULTS: The first studies comparing PTCA and CABG are favourable to surgery (BARI), then late ones using stents (ARTS, ERACI 2) showed that stenting was at least equivalent to CABG, in terms of mortality or serious complications (major acute coronary events, MACE), despite a higher target vessel revascularisation (TVR) mainly due to restenosis in the angioplasty cohort. The same results are observed by stenting a high risk lesion as the unprotected left main stenosis can be, until then treated surgically. However, high volume centers studies treating by PTCA+stent the unprotected left main artery (LMA) shows that the 1 year survival rate is similar to surgery, but always related to a restenosis rate of 20% at 6 months in the stent group, which represents the only significant difference in terms of MACE; the new drug eluting stents lead us to expect, according to SIRIUS and TAXUS II studies, to reduce the restenosis rate, and by the way, the MACE could be dramatically lowered from 50% to 60%. Randomised studies would be necessary, but the extrapolation of the actual data, more particularly results of subgroups with a high risk of restenosis, diabetic patients and small vessels, lead us to think that stenting could come in first intention before surgery if TVR is significantly reduced. A complex anatomy, failed attempted chronic occlusion, several lesions on tortuous vessels, would remain the last surgical indication if CABG provides a more complete revascularization. The impact of these new drugs seems promising. However, we should await early results of studies in diabetic patients and bifurcations. But in high volume experienced centers, CABG indications would be reduced in the future to the technical pitfalls of stenting (complex or tortuous anatomy, chronic occlusions) or to the adverse additional cost of this device, unless reduction of restenosis or TVR could also cancel this extra cost. We expect randomised studies CABG versus stented angioplasty using drug eluting stents to confirm these preliminary data.