Background: Coronary artery bypass (CABG) grafting is traditionally the gold standard treatment for severe unprotected left main (LMS) coronary artery disease, however percutaneous coronary intervention (PCI) for LMS disease has gained popularity in recent years. This meta-analysis pooled the outcomes of randomised trials comparing PCI and CABG for severe LMS disease. Methods: We searched Medline, Embase and Cochrane from 1980 to 31 October 2016 for randomised trials of PCI versus CABG for severe LMS disease for data extraction and subsequent pooled analyses using random effects models. Results: Amongst 2,275 articles searched, 53 full-texts were reviewed and 6 randomised trials totaling 4,706 patients were included. Composite of death, myocardial infarction (MI) and stroke was similar between PCI and CABG odds ratio 0.98 (95% confidence interval 0.81-1.20, p = 0.86), with no difference in all-cause mortality (p = 0.91), myocardial infarction (p = 0.17) or stroke (p = 0.65). PCI did have higher rates of repeat revascularisation 1.76 (1.02-1.53) and if this is included also higher MACE 1.33 (1.15-1.54, p < 0.001). There were no differences in all of the above individual endpoints at 30 days. Subgroup analyses found no differences in MACE rates between the two treatment modalities for elderlies >65 years of age, either sex, diabetics or all tertiles of syntax scores. Conclusions: PCI had similar rates of death, myocardial infarction, stroke and their composite during follow-up compared to CABG, except for more repeat revascularisation required. PCI therefore provides a viable alternative intervention for severe LMS traditionally treated by CABG.