Arm veins have been used in myocardial revascularisation procedures as a last resort bypass conduit because of their associated low patency. Nevertheless, leg veins and mammary arteries, which are the most commonly used, are sometimes not sufficient, leaving little choice as to the bypass conduit. To assess the properties of arm veins in bypass surgery, we compared a group of 28 patients that underwent an arm vein graft coronary bypass procedure with a matched group of patients in which leg veins were used. In 28 patients, 40 arm vein grafts with 77 distal anastomoses were used (mean 1.9 +/- 0.9; range 1-5). A cerebrovascular accident was the cause of the sole death (2%) during the study period. The mean follow-up was 4.6 years (Standard deviation, SD: 1.5 years). More antianginal medication was used in the arm vein group (P = 0.017). Additionally, the percentage of the expected maximal frequency during exercise testing was lower in the arm vein group as compared to the leg vein group. Digital subtraction angiography showed that the patency of the arm vein bypass grafts was 47% (70% confidence limits, CL: 32%-62%) while the patency of the leg vein grafts was 77% (CL: 64%-87%), which was statistically significant (P = 0.051). Comparison of these figures with the few published reports on arm veins used as coronary bypass grafts reveals similar results. We conclude that the arm vein as a coronary bypass graft is only to be used when mammary arteries and leg veins are not available.