BACKGROUND Left internal thoracic artery (LITA) grafts to the left anterior descending coronary artery (LAD) during coronary bypass surgery (CABG) have greater patency rates than saphenous vein grafts and reduce long-term cardiac morbidity and mortality rates. The benefits of multiple versus single arterial grafts and the role of different arterial conduits with respect to short- and medium-term outcome remains controversial. The purpose of this study was to compare the perioperative and intermediate-term results of: (1) patients receiving 2 arterial grafts versus 1 arterial graft and (2) patients receiving a right internal thoracic artery (RITA) versus a radial artery (RA) as the second arterial graft. RESULTS Retrospective analysis of prospectively gathered data on consecutive patients undergoing isolated CABG at our institution between 1989 and 1996 was conducted. The first section of the study compared outcomes for 1 arterial graft (LITA to LAD, n = 2333) versus 2 arterial grafts (LITA + RA or LITA + RITA, n = 378). The second section of the study compared outcomes for the RITA (n = 132) versus the RA (n = 171) as second arterial grafts since 1992, when the radial series was initiated. Part I: By multivariable stepwise logistic regression, the use of 1 arterial graft was associated with an increased incidence of perioperative cardiac morbidity and mortality (odds ratio 2.2, 95% confidence interval 1.4 to 3.3), with the use of our current patient selection criteria. Double-arterial graft patients had a nonsignificant trend toward increased intermediate-term actuarial survival (P = 0.12) and cardiac event-free survival (P = 0.09). Part II: Comparison of preoperative demographics revealed a higher incidence of diabetes (27% vs 11%, P < 0.001), peripheral vascular disease (16% vs 8%, P = 0.03), and elderly age (13% vs 2%, P = 0.001) in patients receiving an RA versus those receiving a RITA as the second arterial graft. Perioperative outcome analysis revealed a decreased intensive care unit stay in the RA versus RITA group (median 30.4 vs 36.2 hours, respectively, P = 0.005) but no significant difference in hospital length of stay. There was no significant difference in perioperative mortality or cardiac morbidity rates. RITA patients had a higher incidence of sternal wound infection (5.3% vs 0.6%, P = 0.01), however, and tended to have increased blood product transfusion rates (51% vs 40%, P = 0.06). CONCLUSIONS The use of 2 arterial grafts is safe, with a reduction in perioperative cardiac morbidity or mortality rates compared with 1 arterial graft after adjustment for other risk variables. When comparing RITA to RA as second arterial grafts, patients receiving an RA have a lower incidence of sternal wound infection and decreased transfusion requirements, with no difference in perioperative or intermediate-term cardiac morbidity or mortality rates.