To our knowledge we conduct the largest retrospective interventional radiological cohort study of cases with secondary UAFs from a single center.
The main symptom in our UAF patients was hematuria in all cases ranging from non-life-threatening transient hemorrhage to hypovolemic shock (in 41.2% of the cases). The need for blood transfusions and/or inotropic therapy is common. Flank pain and fever are further occurring as well (Heers et al., 2018).
Our diagnostic work up showed that in the minority of cases an active bleeding site could be located via CECT (2/10 cases; 20.0%) or angiography (3/17 cases; 17.7%). This confirms the findings of Guntau et al. with two detected UAFs in eight patients by CECT alone (Guntau et al., 2017). However, a negative CECT or angiogram does not rule out the diagnosis, therefore cystoscopy is used for confirmation of UAF and detecting the site of the fistula (Krambeck et al., 2005). Nevertheless, CECT proves to be crucial for planning the interventional approach, to locate the ureteral crossing and to rule out any other bleeding sites. Provocative angiography as an invasive imaging tool has been described as an effective procedure to demonstrate the active hemorrhage (Das et al., 2016; Fox et al., 2011; Guntau et al., 2017; van den Bergh et al., 2009). However, the risk of triggering extensive bleeding or to re-open a currently clotted fistula seems unnecessarily high. Positive CECT and angiography is shown in Fig. 1.
In the therapeutic management, there has been a paradigm shift from a surgical approach, which was still the standard of care in 2004, toward interventional angiography (Bergqvist et al., 2001; Fox et al., 2011; Madoff et al., 2004). Patients typically carry a high risk for complications due to previous extensive surgery and radiation therapy which lead to adhesions, ଁbrosis, and frail tissue (Heers et al., 2018). In patients with UAF to the IIA only, arterial coil embolization can be considered. Most of the UAF involve EIA or CIA, so covered stent treatment is necessary (Muraoka et al., 2006; van den Bergh et al., 2009). Stent graft treatment is less invasive and offers rapid bleeding control (Patel et al., 2014).
Endovascular therapy offers high technical and early clinical success rates (100% and 82.4% in our cohort). Severe complications like stent occlusions or stent graft infections are rare (Malgor et al., 2012; Okada et al., 2013). Stent thrombosis is one important postinterventional complication: In our experience major risk factors for developing stent thrombosis were underlying vasculopathy or advanced tumor disease with extrinsic stenosis of iliac vessels. Too maintain the risk for stent occlusions as low as possible patients should be put on single anti platelet therapy after the procedure. Periprocedural complications can be decreased by using equipment with high personal experience level. When using large sheaths it is often necessary to use a closure device. In one case device failure led to prolonged manual compression and stent occlusion the day after. Literature already disclosed controversies after use of Prostar XL (Maniotis et al., 2017; Power et al., 2019).
In spite of a high number of reported urinary tract infections which sometimes lead to septicemia in these patients, the incidence of secondary stent graft infections seems to be overrated (Darcy, 2009). Stent graft placement in an infectious site is critical. Nevertheless, UAF typically requires urgent treatment. It is therefore recommended that these patients must receive perioperative antibiotic treatment (Hong et al., 2016). In the current study none of the patients suffered from morbidity due to stent graft infections during their postinterventional course. In the so far largest published multi-center study by Heers et al. including 24 patients only one patient developed stent infection. This matches previously published experience (Malgor et al., 2012).
In the current series, recurrent bleedings developed after insufficient coverage of the fistula during the initial intervention and were manageable by a secondary treatment. None of these patients received secondary surgical treatment for recurrent hemorrhage. Additional risk factors included complex underlying situations with severe vasculopathy or stenosing tumor burden. Following our experience, we suggest to cover the ureteral crossing from the CIA to the EIA by a stent graft with prior coil embolization of the IIA. IIA coil embolization prevents retrograde perfusion of the UAF through the gluteal arteries, but also lowers the risk of recurrent bleedings – eg., in cases of fistulas from the proximal IIA. Embolization of the IIA alone would not be sufficient for UAF treatment (Massmann et al., 2020). It is crucial to detect fistula recurrence. In our series recurrent hematuria happened in 4/17 cases (23.5%), one of them died due to recurrence of the UAF (5.9%). Three recurrent fistulas could successfully be treated in a revisional endovascular procedure. Guntau et al. identified recurrent hematuria in one of eight patients. This patient could also successfully be treated with a secondary endovascular treatment, a combination of mild dilatation of the native iliac artery and a short distal landing zone seemed to be the cause (Guntau et al., 2017). Van den Bergh et al. found in their review of various procedures for the treatment of UAF a recurrent fistula related mortality of 18 out of 139 patients (13%) (van den Bergh et al., 2009). In our cohort one patient died consecutively due to recurrent hemorrhage 229 days after initial treatment. This patient had a large pelvic tumor mass which led to external compression and consecutive stenosis of the iliac vessels. The inserted stent graft unfolded incompletely during the initial procedure and stent thrombosis occurred during the postinterventional course. This led to a secondary procedure to treat acute lower limb ischemia, which resulted in reopening of the UAF with consecutive bleeding (Fig. 2).
Our study is limited by its retrospective nature and despite being the largest single center endovascular experience so far to our knowledge, by the small number of cases. In addition, follow-up periods varied substantially.