Factors associated with primary vein graft occlusion in a multicenter trial with mandated ultrasound surveillance.

OBJECTIVE: Even in the setting of duplex ultrasound (DUS) surveillance, a significant number of lower extremity vein bypass grafts (LEVBGs) become occluded as a first event. We sought to identify factors that may contribute to these primary occlusions.

METHODS: This was a retrospective analysis of the Project of Ex Vivo Graft Engineering via Transfection III (PREVENT III) multicenter randomized clinical trial, in which 1404 patients with critical limb ischemia (CLI) underwent LEVBG with 1-year follow-up. Subjects were to undergo DUS at regular intervals (1, 3, 6, and 12 months), with reintervention based on prespecified DUS criteria. Patients who had nontechnical graft occlusion as the initial graft-related event were identified, and multivariate analysis was used to determine factors associated with primary graft occlusion.

RESULTS: Primary vein graft occlusion occurred in 200 subjects and accounted for 36% of all primary patency events and 64% of all graft occlusions in the trial. Primary occlusion events were evenly distributed throughout the first postoperative year. Rates of recurrent CLI, loss of secondary patency, and major amputation in those with primary occlusion were 55%, 79%, and 22% respectively as compared to 18%, 10%, and 10% for subjects without primary occlusion (P < .001). On multivariate analysis, African-American race (subdistribution hazard ratio [SHR], 1.50; 95% confidence interval [CI], 1.06-2.12), a graft diameter <3 mm (SHR, 2.31; 95% CI, 1.33-4.01), and nonadherence with ultrasound surveillance (SHR, 1.58; 95% CI, 1.10-2.27) were independently associated with primary graft occlusion. Of the 123 subjects who received their last scheduled surveillance DUS prior to a primary occlusion event, 39 had a critical ultrasound abnormality identified but failed to undergo graft revision, while 84 had no critical ultrasound abnormality identified. Among these 84 subjects, female gender (SHR, 1.65; 95% CI, 1.07-2.54), and graft diameter <3 mm (SHR, 2.12; 95% CI, 1.03-4.37) were independent factors associated with unheralded graft occlusion.

CONCLUSIONS: Among patients undergoing LEVBG for CLI, almost half of primary patency events are occlusions even in the setting of a DUS surveillance protocol. African Americans, patients with smaller-diameter grafts, and those who are nonadherent with surveillance ultrasound are at increased risk. Failure to intervene on critical findings, and lack of sensitivity of DUS threshold criteria to predict thrombosis, are also important contributors. These findings suggest that prevention of vein graft thrombosis requires further improvements in risk stratification, surveillance, and the timing of reinterventions.