BACKGROUND Invasive coronary angiography is the gold standard means of imaging bypass vessels and carries a small but potentially serious risk of local vascular complications, including myocardial infarction, stroke and death. We evaluated computed tomography as a non-invasive means of assessing graft patency. METHODS Fifty patients with previous coronary artery bypass surgery who were listed for diagnostic coronary angiography underwent contrast enhanced computed tomography angiography using a 16-slice computed tomography scanner. Images were retrospectively gated to the electrocardiogram and two dimensional axial, multiplanar and three dimensional reconstructions acquired. Sensitivity, specificity, positive and negative predictive value, accuracy and level of agreement for detection of graft patency by multidetector computed tomography. RESULTS A total of 116 grafts were suitable for analysis. The specificity of CT for the detection of graft patency was 100%, with a sensitivity of 92.8%, positive predictive value 100%, negative predictive value 85.8% and an accuracy of 94.8%. The kappa value of agreement between the two means of measuring graft patency was 0.9. Mean radiation dose was 9.0 +/- 7.2 mSv for coronary angiography and 18.5 +/- 4 mSv for computed tomography. Pooled analysis of eight studies, incorporating 932 grafts, confirmed a 97% accuracy for the detection of graft patency by multidetector computed tomography. CONCLUSIONS Computed tomography is an accurate, rapid and non-invasive method of assessing coronary artery bypass graft patency. However, this was achieved at the expense of an increase in radiation dose.