R. Howard1, S. Kurz1, M. Sherman1, J. Underhill1, J. L. Eliason1, D. Coleman1 1University Of Michigan,Department Of Surgery/Division Of Vascular Surgery,Ann Arbor, MI, USA
Introduction: Secondary aortoduodenal fistula (SADF) is a rare but life-threatening complication of abdominal aortic reconstruction, with mortality rates of 60%. Clinical presentation varies but often includes gastrointestinal (GI) bleed. Treatment typically requires complex surgical repair that has historically been associated with considerable morbidity and mortality. This retrospective study examines the contemporary management of SADF at a tertiary vascular surgical practice.
Methods: Thirteen patients were managed for SADF between 2004-2014. Vascular reconstructions included graft explantation with extra-anatomic bypass (N=10) and endovascular repair followed by definitive reconstruction (N=3). Duodenal reconstruction included primary repair (N=7), resection (N=3), exclusion (N=2), and no repair (N=1). Primary endpoints included bile leak, major complication, and mortality. Student’s t-test was used for data comparisons.
Results: Of the 13 patients with SADF, six presented with GI bleed and seven without. During a mean follow-up of 631 dy the rate of major complication was high (77%). Five patients (38%) developed duodenal leak. All leaks occurred following graft explantation with extra-anatomic bypass, and the majority of these patients (80%) had no prior history of GI bleed. Factors that trended toward increased risk of bile leak included female gender (67% v. 30%, p=0.252) and method of duodenal repair (67% resection v. 43% primary repair, p=0.401). There were no leaks identified following duodenal exclusion with gastrojejunostomy. Patients that developed duodenal leak had longer mean ICU stay (7.0 v. 2.3 dy, p=0.004), longer mean overall hospital stay (36.6 v. 18.5 dy, p=0.012), and greater late mortality (40% v. 13%) (Figure 1). There were two SADF-related deaths. One patient expired POD 1 secondary to hemorrhage, and another patient expired seven weeks post-operatively from persistent sepsis. Overall mortality trended higher in females (67% v. 20%, p=0.125) and those presenting without GI bleed (43% v. 17%, p=0.308).
Conclusion: While the overall mortality of patients treated at our institution for SADF represents an improvement from historical rates, surgical reconstruction results in major morbidity. Interestingly, those patients presenting with acute GI bleed tended to have better outcomes than those without. Duodenal leak remains a serious complication of SADF repair. In cases where duodenal resection is being entertained over primary repair, duodenal exclusion may represent a more appropriate conservative approach. Finally, the greater incidence of morbidity and mortality in females with SADF is consistent with overall disparities in gendered outcomes in vascular surgery and warrants further investigation.
