Vein graft surveillance: Is graft revision without angiography justified and what criteria should be used?

Purpose: The objective of this study was to assess the accuracy of color-now duplex surveillance parameters to detect infrainguinal vein graft stenoses and to investigate whether graft revision without angiography is justified. Methods: In a prospective study in which three centers participated, the data of graft surveillance in 300 patients were analyzed. for the evaluation of surveillance criteria all patients underwent a digital subtraction angiography if a graft stenosis was suspected. To create a control group, in patients with normal grafts a consented digital subtraction angiography was performed also. From these data the accuracy of seven duplex and three ankle blood pressure-derived variables was assessed. The relation between various surveillance criteria and continued graft patency was determined with life table analysis with the transient state method. Results: The mean follow-up period was 20 months (range, 1 to 40 months). At univariate and multivariate analysis the peak systolic velocity (PSV) ratio provided the best correlation with angiographic stenoses greater than or equal to 70% (PSV ratio cutoff 3.0: sensitivity 80%, specificity 84%). This finding did not differ between the participating centers. With life table methods it was demonstrated that the best combination of efficacy (limitation of the number of unnecessary revisions), safety (minimal number of correctable lesions missed), and reduction of angiograms was obtained by a two-parameter surveillance algorithm. This algorithm included a PSV ratio <2.5 to delineate patients in whom a conservative approach without angiography or revision was appropriate, a PSV ratio greater than or equal to 4.0 to indicate patients in whom vein graft revision without angiography could be scheduled, and a group with PSV ratios between 2.5 and 4.0 in whom angiography was to be performed to determine clinical management on the basis of the stenosis severity. This algorithm had a positive predictive value of 93% and a negative predictive value of 89%. In addition, it resulted in a reduction of the number of angiograms of 49% compared with a policy of angiographies in all patients with a PSV ratio greater than or equal to 2.5. Conclusions: The best criterion to identify a failing graft is the PSV ratio. With a two-parameter algorithm for vein graft surveillance, the incidence of unnecessary revisions and of missed high-grade lesions was acceptably low, whereas the number of angiograms was reduced by one half.