Color-flow duplex scanning of infrainguinal vein bypasses was used to identify failing grafts. Several duplex parameters were compared to determine their value in identifying and quantifying the degree of stenosis. Intraarterial digital subtraction angiography was used as the "gold standard" to determine the severity of the stenosis. The goal of this study was to identify specific color-flow duplex criteria for grading stenotic lesions. After a retrospective analysis some of these parameters were prospectively validated. The surveillance protocol required a color-flow duplex scan every 3 months for the first year and every 6 months during the second year. One hundred sixteen vein grafts in 112 patients were studied. Forty-three stenoses were identified and classified into categories from 30% to 49%, 50% to 69%, and 70% to 99% diameter reduction. These stenoses were identified in either the bypass graft or adjacent inflow or outflow arteries. Failing grafts were evaluated further by intraarterial digital subtraction angiography. Patients with normal appearing bypasses (without suspected stenotic lesions) had intravenous digital subtraction angiography. The five duplex parameters that were studied included the following: (1) graft peak systolic velocity (PSV-graft), (2) the maximum peak systolic velocity (at the site of a stenosis or in normal grafts at the narrowest segment of the bypass) (PSV-max), (3) the ratio between PSV-graft and PSV-max, (PSV-index), (4) end-diastolic velocity (EDV) at a stenosis or from narrowest graft segment, (5) color-flow image diameter measurements. For discrimination of different degrees of stenosis, threshold values of these parameters were calculated by receiver operating characteristic analysis. Diameter reduction measured by color-flow imaging was best to identify all stenotic lesions greater than 29% (sensitivity 88%, specificity 99%). Peak systolic velocity-index proved optimal identification of stenoses greater than 49% (sensitivity 89%, specificity 92%), and 70% to 99% stenoses were associated with increased EDV (sensitivity 91%, specificity 100%). The PSV-index criteria were then validated prospectively in a separate group of vein grafts. The data support the value of surveillance of femorodistal vein grafts and demonstrate that calculation of the degree of graft stenosis is feasible.