Modeling variability in the inferior vena cava into fenestrated endografts for retrohepatic caval injuries

Objective: Injury of the retrohepatic inferior vena cava (IVC) is rare, but extremely fatal. Open repair of these injuries is challenging. Various maneuvers, including atriocaval shunting and total vascular isolation, have been described, but are poorly tolerated in the severely injured patient. Endovascular repair is an attractive alternative strategy, but effective hemostasis of complex injuries requires an endograft that excludes the injury while permitting flow from the hepatic veins. Unfortunately, IVC and hepatic vein anatomy is highly variable and has not been clearly described in injured patients. Our purpose was to characterize critical human IVC morphology in trauma patients, and develop the design parameters of an off-the-shelf fenestrated endograft intended for caval deployment. Methods: One hundred consecutive adult trauma patients with an admission computed tomography scan including a portal venous phase of the abdomen were reviewed. Specific anatomic measurements including segmental IVC lengths and diameters were obtained. Multiple theoretical endografts were modeled to optimize caval coverage in the retrohepatic segment, assuming 10% to 40% oversizing for seal. Results: This sample population had a mean age of 50 years, height of 173 cm, and weight 84 kg. Seventy-one percent were male and 89% had a blunt mechanism of injury. The median caval length from the renal veins to right atrium was 111 mm (interquartile range [IQR], 102-120 mm), diameter was 22 mm (IQR, 19-26 mm), and hepatic venous orifice area was 336 mm(2) (IQR, 267-432 mm(2)). All patients had a landing zone of at least 12 mm in the suprahepatic and 10 mm in the suprarenal segments. Three models of graft length were developed to accommodate patients with segmental and overall dimensions in the smallest half, third quartile, and fourth quartile. These could provide 95% of patients with coverage of the retrohepatic segment without risk of hepatic or renal vein occlusion. Four graft diameters were developed for cross-sectional fit. Graft diameters of 20, 24, 30, and 38 mm could provide adequate coverage in, respectively, 11%, 35%, 49%, and 16% of patients. These combinations of graft length and diameter would accommodate 93% of patients. Conclusions: We defined human IVC morphology essential for endovascular therapy and developed parameters for fenestrated IVC endografts to address retrohepatic caval injuries in trauma patients. Although additional study and testing are required, this proof-of-concept study supports the hypothesis that exclusion of the most devastating retrohepatic IVC injuries can be achieved with a reasonable number of off-the-shelf fenestrated endografts. These findings form the basis for additional research toward the development of novel devices for endovascular therapy of these often lethal injuries.