Completion duplex ultrasound predicts early graft thrombosis after crural bypass in patients with critical limb ischemia

Objective: To determine if intraoperative distal graft end-diastolic velocity (EDV) using completion duplex ultrasound (CDU) predicts patency of crural bypass in patients with critical limb ischemia (CLI). Methods: Records of 116 non-consecutive patients who underwent crural revascularization with vein conduit and CDU between 1998 and 2008 were reviewed. Bypass grafts were performed for rest pain (34%) or tissue loss (66%), while 56% of the reported cases were categorized as "disadvantaged" because of compromised vein quality or diseased arterial outflow. A 10-MHz low-profile transducer was used to image the entire bypass at case completion. Technical adequacy of the grafts was verified by absence of retained valves, arteriovenous fistulas, or localized velocity increases and presence of bypass-dependent distal pulses. Modified Rutherford scores were calculated as surrogate markers of runoff resistance and compared to distal graft EDV. The primary study end point was graft patency during a 1-year posttreatment period. Patency rates were determined using Kaplan-Meier life table methodology and compared using the log-rank test. Predictors of graft patency were determined by Cox proportional hazards. Results: Primary, primary-assisted, and secondary patency for all crural bypasses were 62%, 66%, and 70% at 1 year, respectively. When stratified by tertiles of distal graft EDV (0 - <5 cm/s, 5-15 cm/s, >15 cm/s), 1-year primary patency rates were 32%, 64%, and 84% (P =.001). Low (0 - < 5 cm/s) distal graft EDV (hazard ratio [HR], 3.3 confidence interval [CI], 1.74-6.41; P <.001), poor-quality conduit (HR, 2.5; CI, 1.19-5.21; P =.016), age <70 (HR, 2.08; CI, 1.06-4.00; P =.031), and lack of statin use (HR, 2.04; CI, 1.04-4.00; P =.038) were independent predictors of graft failure. While the modified Rutherford score correlated with distal graft EDV (P =.05), it was not an independent predictor of patency (P =.58). Predictors of low EDV (<5 cm/s) included single-vessel runoff (odds ratio [OR], 3.33; CI, 1.14-9.71; P =.027), poor conduit (OR, 2.94; CI, 1.16-7.41; P = 0.024), and diabetes (OR, 2.86; CI, 1.14-7.21; P =.025). Conclusions: Distal graft EDV predicts crural vein graft patency in patients with CLI. Grafts with EDV <5 cm/s remain at high risk for early failure. The impacts on patency of statins, age, and poor-quality conduit are, again, confirmed. These results highlight the value of EDV using intraoperative CDU for anticipating and, possibly, improving results of open crural revascularization.