Effectiveness and outcome of endovascular therapy for late-onset postpancreatectomy hemorrhage using covered stents and embolization

Effectiveness and outcome of endovascular therapy for late-onset postpancreatectomy hemorrhage using covered stents and embolization Nicole Hassold, MD,a Franziska Wolfschmidt, MD,a Alexander Dierks, MD,a Ingo Klein, MD,b Thorsten Bley, MD,a and Ralph Kickuth, MD,a Wuerzburg, Germany

Objective: The purpose of this study was to evaluate the clinical and long-term outcome of patients who underwent covered stent treatment because of late-onset postpancreatectomy hemorrhage in a greater number of patients. A secondary study goal was to compare embolization techniques with covered stents regarding differences in early and late clinical outcome, rebleeding, and vessel patency. Methods: Between December 2008 and June 2015, 27 consecutive patients suffering from major hemorrhage after pancreatic surgery underwent either covered stent placement or embolization of the affected visceral artery. The patients’ medical reports and radiologic images were retrospectively reviewed. The main study end point was technical and clinical success, including survival and complications; the secondary end points were perfusion distal to the target vessel and, for covered stent placement, patency of the affected artery. Results: Covered stent placement was successful in 14 of 16 patients (88%); embolization was successful in 10 of 11 (91%) patients. For the embolization group, the overall 30-day and 1-year survival rate was 70%, and the 1- and 2-year survival rate was 56%; for the covered stent group, these rates were 81% and 74%, respectively. The 30-day patency of the covered stent was 84%, and 1-year patency was 42%; clinically relevant ischemia was observed in two patients. Infarction distal to the embolized vessel occurred in 6 of 11 patients (55%). Conclusions: Endovascular treatment using either covered stents or embolization techniques is an effective and safe emergency therapy for life-threatening postpancreatectomy hemorrhage with good clinical success rates and long-term results. Covered stent placement preserving vessel patency in the early postoperative phase should be preferred to embolization if it is technically feasible. (J Vasc Surg 2016;-:1-11.)

Postpancreatectomy hemorrhage (PPH) is one of the most severe complications after pancreatic surgery, accounting for 11% to 38% of overall mortality.1 Whereas early PPH occurring within 24 hours of surgery caused by technical failure or coagulopathy requires repeated laparotomy and surgical hemostasis, late hemorrhage occurring >24 hours after pancreatic surgery1 results from vessel erosion caused by pancreatic leakage, leakage of the biliodigestive anastomosis, infection with vessel From the Institute of Diagnostic and Interventional Radiologya and Department of General, Visceral, Vascular and Pediatric Surgery,b University of Wuerzburg. Author conflict of interest: R.K. is a consultant for Abbott Vascular. T.B. is a consultant for GSK and MSD; is on the speakers bureaus of Bayer, Bracco, GE, Guerbet, and HeartFlow; and has received Siemens research funding (Deutsche Forschungsgesellschaft DFG: BL 1132/1-2) and research cooperation from Siemens, Noras, and Rapid. Parts of the data have been presented as a poster at the annual European Congress of Radiology, Vienna, Austria, March 4-8, 2015. Correspondence: Nicole Hassold, MD, Institute of Diagnostic and Interventional Radiology, University of Würzburg, Oberdürrbacher Straße 6, Würzburg 97080, Germany (e-mail: [email protected]).

involvement, or vascular injury during resection and pseudoaneurysm formation in the postoperative course.2 According to current literature, interventional techniques tend to be the first-line treatment of late PPH.3-7 Selective embolization and covered stents have been widely used for arterial injury repair, and several published studies give proof of effectiveness.2,5,8-12 To our knowledge, to date, there are only few case series with low numbers of patients and several individual case reports presenting the use of covered stents for PPH. Data presented lack in comparing stent graft implantation with selective embolization, especially in the setting of clinical fate. In addition, patency rates after covered stent implantation have not yet been evaluated. The purpose of this study was to evaluate the clinical and long-term outcome of patients who underwent covered stent treatment because of PPH in a greater number of patients and to compare embolization techniques with stent grafts regarding differences in early and late clinical outcome, rebleeding, and vessel patency.

The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any

METHODS

manuscript for which they may have a conflict of interest.

Study sample Data were obtained from a retrospective clinical database and reviewed for patients undergoing angiography for suspected late PPH between December 2008 and

0741-5214 Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.05.071

1

2

Journal of Vascular Surgery

Hassold et al

---

2016

Table I. Demographic data and clinical, technical, and follow-up information of patients treated with covered stents

Patient No./age, years/sex

Diagnosis

Surgery

1/63/M

Pancreatic carcinoma

Left pancreatic resection with splenectomy

2/54/F

Intraductal papillary mucinous neoplasm

Segmental pancreas resection with pancreaticojejunostomy

3/70/F 4/63/M

Onset of bleeding, postoperative day

Type and location of hemorrhage

210

Erosion SMA

18

Erosion CHA

Cholangiocellular carcinoma PPPD

13

Erosion CHA

Pancreatic gastrointestinal stromal tumor and neuroendocrine tumor

PPPD

21

Pseudoaneurysm SA

5/74/M

Duodenal adenoma

PPPD

21

Erosion PHA and stump of GDA

6/39/M

Papillary carcinoma

PPPD

23

Erosion stump of GDA

7/48/M

Chronic pancreatitis

PPPD

12

Erosion PHA

8/59/M

Cholangiocellular carcinoma

PPPD

8

9/62/M

Pancreatic carcinoma

PPPD

18

Erosion CHA

10/52/M

Cardiac and pancreatic carcinoma

Gastrectomy and Whipple

4

Erosion CHA

11/70/M

Papillary carcinoma

PPPD

240

12/76/M

Papillary carcinoma

PPPD

8

Erosion PHA and branch of SMA

13/52/M

Intraductal papillary mucinous neoplasm

Partial pancreatic resection

13

Erosion CHA

14/61/M

Cholangiocellular carcinoma PPPD

15

Ruptured pseudoaneurysm stump of GDA

15/73/M

Neuroendocrine tumor of pancreas

PPPD

9

Erosion CHA with massive free bleeding

16/69/M

Pancreatic carcinoma

PPPD

11

Erosion CHA and SA

Erosion CHA and PHA

Pseudoaneurysm CHA

CHA, Common hepatic artery; GDA, gastroduodenal artery; PHA, proper hepatic artery; PPPD, pylorus-preserving pancreaticoduodenectomy; SA, splenic artery; SMA, superior mesenteric artery.

June 2015. Institutional Review Board approval was waived for this retrospective study. Because of the retrospective nature of this study, no patient informed consent was required. There were 27 patients (23 male, 4 female; mean age, 60.8 6 11.2 years; range, 39-85 years) identified who had undergone diagnostic angiography and covered stent implantation or embolization because of PPH. Hemorrhage was suspected when sentinel bleeding (fresh blood loss through abdominal drains or nasogastric tubes, hematemesis, or melena), unexplained hypotension or tachycardia, or decrease in hemoglobin concentration occurred.

Late PPH was evaluated according to the International Study Group of Pancreatic Surgery guidelines13: B, Mild: small- or medium-volume blood loss (decrease in hemoglobin level <3 g/dL), mild clinical impairment, transfusion of 1 to 3 units of packed cells C, Severe: large-volume blood loss (decrease in hemoglobin level $3 g/dL), clinically significant impairment Grade A patients suffering from minor episodes of hemorrhage without clinical impact were not included in the study. Computed tomography (CT) was used as the

Journal of Vascular Surgery Volume

-,

Number

Hassold et al

3

-

Table I. Continued.

Grade of hemorrhage

Hemoglobin level before intervention, g/dL

Material

C

2.1

Advanta V12 5/22 mm

C

7.1

Advanta V12 8/38 mm

C

7.6

Advanta V12 5/22 mm

C

7.5

Advanta V12 5/16 mm, coils

C

3.8

Advanta V12 7/38 mm

B

7.6

Advanta V12 7/22 mm

C

9.0

B

Complication Intestinal ischemia after stent thrombosis, multiorgan failure

Patency or occlusion, day

Outcome

Occlusion (55)

Died (70 days)

Patent

Alive

Stent migration after 2 years

Occlusion (78)

Alive

Rupture of pseudoaneurysm during intervention

Not applicable

Alive

Patent

Alive

Occlusion (571)

Alive

Advanta V12 6/38 mm

Occlusion (230)

Alive

8.6

Advanta V12 7/38 mm, 7/22 mm

Occlusion (59)

Alive

C

10.9

Advanta V12 5/22 mm

Multiorgan failure

C

8.0

Advanta V12 7/38 mm

Rebleeding after 22 days: liver abscess after iatrogenic stent occlusion and newly developed stenosis of portal vein

C

6.6

C

6.3

C

Rebleeding after 10 weeks treated by covered stent

Died (1 day) Occlusion (22)

Alive

Advanta V12 5/22 mm, 5/16 mm

Patent

Alive

BeGraft 4/21 mm (hepatic artery), coils (SMA)

Patent

Alive

5.1

Advanta V12 7/38 mm, 7/22 mm

Patent

Alive

C

8.0

BeGraft 4/18 mm

Partial thrombosis (1)

Alive

C

8.0

Advanta V12 7/38 mm

Multiorgan failure

Patent

Died (2 days)

C

7.3

Advanta V12 5/12 mm, BeGraft 6/28 and 18 mm, coils, 2 Amplatz vascular plugs

Cardiopulmonary resuscitation during intervention, multiorgan failure

primary imaging modality in 18 patients; 9 patients were directly referred for emergency angiography. Interventional procedures Diagnostic. Unless the patient was already intubated, interventions were performed under local anesthesia. Retrograde common femoral artery access using a long 6F or 7F sheath (Destination; Terumo, Tokyo, Japan) was performed. Vital parameters were monitored continuously. After initial angiography of the abdominal aorta, angiograms of the celiac trunk and the superior mesenteric artery (SMA) using a 5F diagnostic catheter

Died (1 day)

(C1, C2dcobra shaped, sidewinder shaped; Cook, Bjaeverskov, Denmark) were obtained. Covered stent implantation. After localization of the extravasation site, a 0.035-inch guidewire was passed beyond the bleeding site. In coaxial technique, the long sheath was advanced over the guidewire and the diagnostic catheter distal to the site of extravasation. A balloon-mounted covered stent was placed at the site of extravasation, and the sheath was then drawn back to allow initial exposure of the device. The following covered stents were used in this study: Advanta V12 (Atrium Medical, Hudson, NH) and BeGraft (Bentley InnoMed

4

Journal of Vascular Surgery

Hassold et al

---

2016

Table II. Demographic data and clinical, technical, and follow-up information of patients treated with embolization techniques Patient No./ age, years/ sex

Diagnosis

Surgery

Onset of bleeding, postoperative day

Type and location of hemorrhage

Hemoglobin level before Grade of intervention, hemorrhage g/dL

Material

Complication

Outcome

17/44/M

Duodenal carcinoma

PPPD

11

Erosion CHA

C

8.0

Amplatzer Hepatic plug abscess

Alive

18/69/M

Pancreatic carcinoma

PPPD

16

Erosion stump of GDA

C

5.6

Coils

Died (6 days)

19/41/M

Chronic pancreatitis

PPPD

48

Erosion SA

C

8.1

Amplatzer Spleen Alive plug/coils infarctions

20/72/M

Pancreatic carcinoma

Left pancreatic resection

16

Erosion right gastroepiploic artery

C

7.0

Coils

Died (tumor progress)

21/50/F

Pancreatic metastasis

PPPD

86

Erosion PHA and SA

C

6.4

Coils

Died (tumor progress)

22/85/M

Pancreatic carcinoma

PPPD

63

Pseudoaneurysm of left hepatic artery

B

8.0

Coils

Alive

23/62/F

Pancreatic carcinoma

PPPD

21

Erosion CHA

B

6.1

Coils

24/61/M

Papillary carcinoma

PPPD

19

Pseudoaneurysm SMA

C

9.9

Coils

Alive

25/58/M

Pancreatic carcinoma

PPPD

26

Pseudoaneurysm CHA

C

6.1

Coils

Alive

26/55/M

Pancreatic carcinoma

PPPD

8

Long erosion of CHA and GDA stump

C

8.4

Coils

Hepatic failure

Died (3 days)

27/60/M

Pancreatic carcinoma

PPPD

23

Erosion stump of GDA/CHA

B

8.4

Coils

Hepatic failure

Died (2 days)

Hepatic failure

Hepatic abscess

Alive

CHA, Common hepatic artery; GDA, gastroduodenal artery; PHA, proper hepatic artery; PPPD, pylorus-preserving pancreaticoduodenectomy; SA, splenic artery; SMA, superior mesenteric artery.

GmbH, Hechingen, Germany). Completion angiograms were obtained in all patients. Anticoagulation therapy was not administered. Embolization. Coaxial microcoil embolization was performed after superselective catheterization with a 2.7F microcatheter (Progreat, Terumo) if stent implantation was not possible because of the anatomic situation of the extravasation site. Embolization with Amplatzer vascular plugs (St. Jude Medical, St. Paul, Minn) was directly performed through a long 7F sheath. Interventions were performed by interventional radiologists with 4 to 18 years of experience. End point definition The primary end points were technical success and clinical success. Technical success was defined as complete cessation of bleeding in control angiography; clinical success was defined as absence of symptoms of acute hemorrhage (hemodynamic stability, absence of recurrent decrease of hemoglobin level by >1.5 g/dL, no need for packed cells, absence of hemorrhagic secretion through drainage catheter, and no repeated angiographic or surgical intervention). The secondary end point was end-organ perfusion; for covered stent placement, a secondary end point was the patency of the affected artery. Both were documented by CT angiography or digital subtraction

angiography. The period of hospitalization was another minor study end point. Complications of endovascular treatments were classified according to the reporting standards of the Society of Interventional Radiology, as follows14: minor complicationsd requiring no or only nominal complication-associated therapy, and no further consequence including overnight admission for observation only; major complicationsd requiring therapy, hospitalization, unplanned increase in level of care; those resulting in permanent adverse sequelae or death. Clinical follow-up Patients were examined for clinical symptoms of ongoing or recurrent hemorrhage. The patient’s clinical condition and laboratory test results (hemoglobin concentration, lactate level, liver enzymes) were documented until discharge or death and at the appointments at the outpatient clinic up to the recorded period. Complications and mortality during the course of the study were documented. Recurrent bleeding and ischemic effects were assessed by followup CT scans or, in case of clinical symptoms of acute rebleeding, by angiography. The mean imaging follow-up was 378 6 406 days (range, 1-1165 days). Mean clinical follow-up time was 749 6 632 days (range, 1-1917 days).

Journal of Vascular Surgery Volume

-,

Number

Hassold et al

5

-

Fig 1. Pseudoaneurysm of the common hepatic artery (arrows) detected 8 months after pancreatic resection (A). After the pseudoaneurysm was confirmed by selective angiography (B), a covered stent was advanced to the extravasation site (arrow) while being protected by a long sheath (C). Control angiogram reveals complete exclusion of the pseudoaneurysm and patency of the hepatic arteries (D). Contrast-enhanced control computed tomography (CT) scan 6 days after intervention confirms patency (E and F; arrows).

Statistics Descriptive statistics were calculated. Continuous data are presented as mean 6 standard deviation and range if appropriate; categorical data are presented as counts and percentages. Mortality was determined at 30 days, 1 year, and 2 years. Kaplan-Meier estimates were generated for stent patency and mortality. Analyses were performed using GraphPad Prism 5 (GraphPad Software Inc, La Jolla, Calif) and Microsoft Excel 2010 (Microsoft, Redmond, Wash).

RESULTS Patients. Between December 2008 and June 2015, 366 pancreatic resections were performed at our hospital. Of

these patients, 27 were referred to the angiography department because of late PPH. Indications for pancreatic resection were malignant lesions in 25 patients and benign lesions in 5 patients. Operative procedures and underlying disease, hemoglobin level at the time of reference to the angiography unit, and severity of hemorrhage are shown in Tables I and II. Average onset of bleeding was 36 6 57 days (range, 4-240 days) after surgery; 22 patients had grade C and 5 patients had grade B hemorrhage. Sentinel bleeding was identified in nine patients. The common hepatic artery (CHA) was affected in 11 patients; the stump of the gastroduodenal artery, in 5 patients; the proper hepatic artery (PHA), in 4 patients; the splenic artery (SA), in 5 patients; the SMA, in 3

6

Hassold et al

Journal of Vascular Surgery ---

Fig 2. Kaplan-Meier graph of stent patency. Solid line, Patency rate; dotted line, 95% confidence interval. Below, Patients at risk.

patients; and the right gastroepiploic artery, in 1 patient. In four patients, erosion of two vessels was identified. In all patients, arterial injury was effectively treated either by covered stent placement or by embolization. Tables I and II contain details about interventional procedures and results. Covered stents. In 14 of 16 patients (88%), covered stent placement successfully ceased extravasation (Fig 1). Mean total procedure time of covered stent placement was 75 6 23 minutes (range, 44-123 minutes). After stent deployment, patency was confirmed for all patients. After intervention, all but one patient (No. 16) were hemodynamically stable. Secondary coil embolization of the SA was necessary in one patient because the pseudoaneurysm rupture occurred immediately after stent implantation. In one patient (No. 16), conversion to embolization with Amplatzer vascular plugs was necessary because acutely life-threatening bleeding of the proximal CHA and the SA occurred during intervention. Despite successful embolization and cardiopulmonary resuscitation, the patient died the next day. Two patients (No. 9 and No.15) were hemodynamically stable under circulatory supportive medication but died 1 day after intervention of multiorgan failure due to prolonged hypovolemic shock. In both patients, control CT showed absence of ongoing bleeding and patency of the treated vessel. Therefore, overall clinical success rate was 81%. Follow-up by contrast-enhanced CT or by angiography detected vessel occlusion in 6 of 16 patients (37.5%) 171 6 208 days (range, 22-571 days) after intervention (Figs 2 and 3), resulting in a 30-day patency of 84% and a 1-year patency of 42%. Four of these patients underwent stenting of the hepatic artery. In one patient, the stent could no longer be detected after 3 months, most

2016

likely because of migration (Fig 4). No morphologic or laboratory signs of ischemia were found in these patients. CT scans revealed no signs of infection of the occluded grafts. There were neither clinical signs of infection nor elevations of inflammatory markers in those patients. In one patient, partial thrombosis of the stent was detected 1 day after implantation. The 1-year follow-up CT scan revealed remaining perfusion. Patient No. 1 developed intestinal ischemia 7 weeks after stent implantation into the SMA. Although rotational thrombectomy was successfully accomplished, mesenteric ischemia was not completely reversible. The patient died of multiorgan failure 2 weeks later. Rebleeding occurred in one patient 78 days after stent implantation into the hepatic artery because of a newly developed pseudoaneurysm adjacent to the stent. After stent elongation, no further bleeding occurred. One patient with known thrombosis of the portal vein underwent repeated angiography 22 days after stent deployment into the CHA with suspected erosion distal to the stent. Unintended stent occlusion occurred during positioning of a guidewire; bleeding ceased, but hepatic abscesses developed because of necrosis. Six of 11 patients with successfully implanted stent grafts in the hepatic artery had transient moderate (up to fivefold) elevations of liver enzymes, whereas patency of all stent grafts was apparent on CT. The patient with partial thrombosis of the stent graft had approximately 10-fold elevated liver enzymes (aspartate transaminase, 603 IU/L; alanine transaminase, 590 IU/L). Complete acute thrombosis occurred only in one patient. This patient developed hepatic abscesses and had massively elevated liver enzymes (aspartate transaminase, 1359 IU/ L; alanine transaminase, 779 IU/L). In all patients, liver enzymes normalized within 1 week. Liver enzymes were not elevated in any patient with late thrombosis at the time of detection. In summary, major complications were observed in seven patients (44%), in three of whom stent-related complications occurred directly. Minor complications were observed in four patients (Table III). Embolization. In all 11 patients, primary embolization was technically successful (Fig 5). Mean total procedure time was 99 6 31 minutes (range, 50-171 minutes) for embolization, which was significantly longer compared with stent graft placement (P ¼ .033, unpaired two-tailed t-test). In patient No. 17, an Amplatzer vascular plug was successfully used as a primary embolic agent in a condition of acutely life-threatening massive bleeding of the CHA. Second-look laparotomy ruled out backdoor bleeding in this patient. Asymptomatic segmental spleen infarction was detected by CT in patient No. 19 after embolization of the SA. Embolization of the CHA was performed in three patients; embolization of the PHA was performed in

Journal of Vascular Surgery Volume

-,

Number

Hassold et al

7

-

Fig 3. A, Selective angiogram of the celiac trunk showing acute bleeding arising from the proper hepatic artery (PHA) that was successfully treated by stent implantation. B, Control angiogram confirms patency of the PHA. C, Follow-up computed tomography (CT) scan 230 days after intervention reveals stent occlusion and maintained perfusion of intrahepatic arteries through collaterals.

Fig 4. A, Contrast-enhanced computed tomography (CT) scan showing patent hepatic artery directly after stent (arrow) placement in the common hepatic artery (CHA). B, Follow-up CT scan after 2 years showing occlusion of the CHA and absence of the implanted stent due to migration.

three patients. Clinically relevant liver infarction occurred in all patients after embolization of the PHA. One patient died 6 days after intervention of acute liver failure, and two patients developed hepatic abscesses but recovered fully. Two patients who underwent embolization of the CHA because stent implantation was not possible owing to tortuous vessel anatomy in one and long erosion and therefore insufficient anchoring zones in the other died of hepatic failure 2 days and 3 days, respectively, after intervention. Liver function of the third patient remained unimpaired because sufficient collaterals through the SMA were present. Hepatic perfusion remained normal in those two patients who underwent superselective embolization of intrahepatic arteries. Liver function normalized in all surviving patients. In summary, major complications were observed in six patients (54%) after

embolization (Table III). Duration of hospitalization tended to be longer in patients after embolization (37 6 12 days) compared with patients who were treated with covered stent implantation (27 6 6 days; P ¼ .393, two-tailed t-test). Outcome. Three patients (21%) died after successful stent implantation and one patient with conversion to embolization, corresponding to an overall 30-day survival rate of 81% and a 1- and 2-year survival rate of 74% (Fig 6). No deaths occurred after week 9. For the embolization group, overall 30-day survival was 70%; 1- and 2year survival rate was 56%. Survival rates of the stent and embolization group did not differ significantly (logrank-test, P ¼ .523; Fig 4). Survival of patients with a benign lesion was 100%, in both the stent graft group

8

Journal of Vascular Surgery

Hassold et al

---

Table III. Major and minor complications of stent placement and embolization Stent (n ¼ 16), patient No.

Embolization (n ¼ 11), patient No.

Major complications Intestinal ischemia

1

Hepatic failure Multiorgan failure

18, 26, 27 1, 9, 15, 16

Spleen infarction Rebleeding Rupture of pseudoaneurysm Hepatic abscess

19 6, 10 4 10

23, 17

Minor complications Stent occlusion without signs of ischemia Stent migration

3, 6, 7, 8, 14 3

(n ¼ 4) and the embolization group (n ¼ 1). Two patients died in the follow-up period because of tumor progression, resulting in a 1- and 2-year survival of 72% (all patients) and 65% (cancer patients) in the remaining patients.

DISCUSSION Although PPH is rare with a prevalence of 2.5% to 9.3%, it is still one of the most feared complications of pancreatic surgery.13 Whereas early bleeding occurring within 24 hours of surgery is often caused by stump insufficiency of the gastroduodenal artery due to technical failure and can be safely handled with repeated laparotomy, late PPH occurring beyond 24 hours of surgery is caused by inflammatory vascular erosion arising from pancreatic leakage or anastomotic insufficiency. Access to the bleeding site after extensive surgery and inflammation is often difficult for the surgeon. Operative management of acute bleeding after pancreatectomy is related to high morbidity and mortality in the range of 13% to 60%.15 Therefore, the endovascular approach is widely accepted as the first-line therapy option for PPH, whereas surgery is reserved for patients for whom an angiographic approach fails or is not feasible. Embolization is an endovascular treatment widely used for arterial injuries, and several published studies give proof of its effectiveness with success rates of 80% to 100%.3,11,13 The liver can tolerate embolization of the hepatic arteries because of its dual blood supply from arterial and portal circulation supply and collaterals. However, rare complications, such as postembolization syndrome, liver abscesses, and organ failure, occur if embolization of the hepatic artery is necessary. Covered stent placement may be a favorable alternative to embolization in terms of preserving patency of the affected vessel and thus preventing infarction or ischemia distal

2016

to the site of intervention.7 All of our patients with embolization of the PHA developed hepatic infarctions. In the current literature, the incidence of hepatic ischemia after extrahepatic embolization was reported at 29% to 70% and that of hepatic infarction at 40% to 45%.5 This high rate of infarction after embolization might also be explained by hematomas compromising the portal vein and therefore a reduction of both arterial and portal venous blood supply. Although Mine et al16 observed sufficient collateral development in 72% of all patients after embolization of the hepatic artery, it is still difficult to predict sufficient collateral circulation from preinterventional CT scans or angiograms. Stent graft implantation is favorable for patients dependent on hepatic artery patency because of portal vein occlusion or early after liver transplantation. One patient initially treated by covered stent implantation developed portal vein thrombosis postoperatively. Hepatic abscesses developed when unintended stent occlusion occurred during repeated angiography. Being the faster technique is another advantage of stent graft implantation, especially in an emergency situation, when time until hemostasis is crucial; the mean procedural time for stent implantation was significantly less (P ¼ .033) at 75 6 23 minutes compared with 99 6 31 minutes for embolization techniques. Current literature suggests that covered stent placement is superior to embolization for bleeding control.1,7,9 Rebleeding due to covered stent dislodgment has been observed in individual cases.8 Selection of proper stent size, both diameter and length, can be challenging. Because the diameter of the affected vessel is often decreased as a result of circulatory instability and vascular spasms, it is important to avoid undersizing to provide appropriate sealing and to avoid endoleaks and stent migration.11,17 Oversizing not only can result in vessel rupture, but it is thought to be a possible cause of stent thrombosis.5 Determination of the correct length is hampered by the fact that vascular spasm might disguise the entire extent of vessel erosion. Especially in patients with ongoing vessel erosion due to pancreatic or anastomotic leakage, proper stent length with safe proximal and distal landing zones is essential to avoid rebleeding. So far, few data are available regarding correct stent sizing, although there is broad agreement that stent sizing should be based on angiographic findings and the preinterventional CT scans.5,17,18 In fact, in our clinical practice, all patients with significant spasm of an affected vessel are treated by embolization, as deployment of a covered stent may be extremely hazardous in terms of rupture. A tortuous way to the extravasation site and vessels eroded over a long distance and therefore very fragile vessels with unsafe anchoring zones are circumstances in which we favor embolization techniques. Stenosis of the afferent vessel is another anatomic obstacle to stent graft placement.

Journal of Vascular Surgery Volume

-,

Number

Hassold et al

9

-

Fig 5. A, Contrast-enhanced computed tomography (CT) scan showing perihepatic abscess formation (asterisk) and erosion of the hepatic artery (arrow). B, Selective angiogram of the celiac trunk showing severe erosion of the hepatic artery. Because of the severely affected vessel, the patient was primarily treated with coil embolization. C, Control angiogram after embolization of the hepatic artery showing arterial liver perfusion through collaterals. D, Follow-up contrast-enhanced CT scan after 3 years showing arterial perfusion of the liver with segmental hyperperfusion.

Fig 6. Survival proportions of patients treated with covered stents (dashed line) and embolization (solid line), including patients at risk after stent implantation (x) and embolization (dots).

Last, successful implantation depends heavily on the interventional radiologist’s experience and skills, as this technique inevitably has a learning curve. Preinterventional CT scans are helpful to choose the best treatment option because not only the extent of

the vascular injury and vessel anatomy but also underlying causes like pancreatic fistulas, anastomotic leakage, and intra-abdominal inflammation are detected. To reduce the risk of recurrent bleeding, adequate treatment of those underlying causes is essential. In our study cohort, rebleeding occurred in two patients, which was caused by insufficiency of the biliodigestive anastomosis and peritonitis in one patient and pancreatic fistula with fluid collection in the direct vicinity of the hepatic artery in the other patient. Progressive occlusion of the stent was observed in several patients and even stent loss in one. Potentially because of the gradual occlusion process, formation of collateral circulation led to sufficient perfusion without any clinical symptoms of ischemia in all patients after stent implantation into the hepatic artery, whereas embolization of the hepatic artery or SA led to transient ischemia and to at least partial organ infarction in most cases. Ischemia occurred in one patient suffering from acute stent thrombosis in the SMA and in another patient after unintended iatrogenic stent occlusion. We observed a high number of midterm occlusions detected 1 month to 1 year after intervention and one late occlusion after 571 days. In the current literature, there are only limited data on midterm and long-term stent patency. Because of longer follow-up periods, we could observe that the

10

Journal of Vascular Surgery

Hassold et al

---

development of covered stent thrombosis is an ongoing process that can be observed not only acutely after intervention. Lim et al,18 covering a follow-up period closest to ours (mean, 356 days), reported 1-year patency rates of 69% of hepatic stents compared with 47% (only hepatic artery) in our cohort of patients. To date, definite guidelines with special regard to the anticoagulation regimen after implantation of covered stents in arterial injuries do not exist and clinical experience is limited. We desist from periprocedural anticoagulation in actively bleeding patients,8 whereas patients without active extravasation will generally receive acetylsalicylic acid 100 mg daily on a lifelong basis. We used balloon-expandable stent grafts as we and others (Stampfl et al) had good results using balloonmounted grafts that were primarily designed for coronary arteries.8,10 Boufi et al and Lim et al prefer self-expanding stents (Viabahn [W. L. Gore & Associates, Flagstaff, Ariz]; NitiComVi [TaeWoong Medical, Gyeonggi-do, Republic of Korea]).18,19 Gwon et al and Heiss et al used both balloon-expandable stent grafts (Jostent [Abbott Vascular, Abbott Park, Ill]; Advanta V12) and selfexpandable stent grafts (ComVi [TaeWoong Medical]; Symbiot [Boston Scientific, Marlborough, Mass]).5,12 Technical success rates are within the same range in these groups: Stampfl et al, 87%; Boufi et al, 80%; Lim et al, 94%; Gwon et al, 100%; Heiss et al, 100%. We have a broad stock of different sizes of our standard covered stents (Advanta V12, BeGraft). Mainly for economic reasons, we restrain from stockpiling both self-expanding and balloon-expandable stent grafts. Balloon-expandable stents have to be used carefully in eroded, fragile vessels with special focus on proper diameter selection, but this holds true for self-expanding stents as well. As a matter of fact, peri-interventional rupture of a pseudoaneurysm of the SA occurred in one patient immediately after stent placement. Compared with self-expanding stents, postdilation of the balloon-expanded stent we use is possible to a greater extent, which might be advantageous in the case of unintended undersizing and subsequent endoleakage. There are several limitations to this study. First, the sample size is relatively small. Second, this study is retrospective and lacks randomization. Although prospective randomization would have been helpful to assess the outcome of covered stent placement in comparison with embolization, it is difficult from an ethical point of view to perform a randomized controlled study in lifethreatening situations. Third, follow-up periods differ widely in our study population. A multicenter trial could therefore be beneficial to underline the data presented in our study.

CONCLUSIONS Covered stent implantation and embolization techniques are highly effective, safe interventional

2016

procedures to manage life-threatening PPH with good technical and clinical results. We consider covered stent implantation the first-line treatment of late-onset PPH, but embolization techniques remain important for all situations when covered stent implantation is not possible because of anatomic or technical difficulties.

AUTHOR CONTRIBUTIONS Conception and design: NH, FW, AD, IK, TB, RK Analysis and interpretation: NH, IK, TB, RK Data collection: NH, FW, AD, RK Writing the article: NH, TB, RK Critical revision of the article: NH, FW, AD, IK, TB, RK Final approval of the article: NH, FW, AD, IK, TB, RK Statistical analysis: NH Obtained funding: Not applicable Overall responsibility: RK

REFERENCES 1. Wente MN, Veit JA, Bassi C, Dervenis C, Fingerhut A, Gouma DJ, et al. Postpancreatectomy hemorrhage (PPH): an International Study Group of Pancreatic Surgery (ISGPS) definition. Surgery 2007;142:20-5. 2. Stampfl U, Hackert T, Sommer CM, Klauss M, Bellemann N, Siebert S, et al. Superselective embolization for the management of postpancreatectomy hemorrhage: a singlecenter experience in 25 patients. J Vasc Interv Radiol 2012;23:504-10. 3. Schafer M, Heinrich S, Pfammatter T, Clavien PA. Management of delayed major visceral arterial bleeding after pancreatic surgery. HPB (Oxford) 2011;13:132-8. 4. Darnis B, Lebeau R, Chopin-Laly X, Adham M. Postpancreatectomy hemorrhage (PPH): predictors and management from a prospective database. Langenbecks Arch Surg 2013;398:441-8. 5. Gwon DI, Ko GY, Sung KB, Shin JH, Kim JH, Yoon HK. Endovascular management of extrahepatic artery hemorrhage after pancreatobiliary surgery: clinical features and outcomes of transcatheter arterial embolization and stentgraft placement. AJR Am J Roentgenol 2011;196:W627-34. 6. Puppala S, Patel J, McPherson S, Nicholson A, Kessel D. Hemorrhagic complications after Whipple surgery: imaging and radiologic intervention. AJR Am J Roentgenol 2011;196: 192-7. 7. Sanjay P, Kellner M, Tait IS. The role of interventional radiology in the management of surgical complications after pancreatoduodenectomy. HPB (Oxford) 2012;14:812-7. 8. Goltz JP, Basturk P, Hoppe H, Triller J, Kickuth R. Emergency and elective implantation of covered stent systems in iatrogenic arterial injuries. Rofo 2011;183:618-30. 9. Stoupis C, Ludwig K, Inderbitzin D, Do DD, Triller J. Stent grafting of acute hepatic artery bleeding following pancreatic head resection. Eur Radiol 2007;17:401-8. 10. Stampfl U, Sommer CM, Bellemann N, Weitz J, Bockler D, Richter GM, et al. The use of balloon-expandable stent grafts for the management of acute arterial bleeding. J Vasc Interv Radiol 2012;23:331-7. 11. Rossi M, Rebonato A, Greco L, Citone M, David V. Endovascular exclusion of visceral artery aneurysms with stent-grafts: technique and long-term follow-up. Cardiovasc Intervent Radiol 2008;31:36-42. 12. Heiss P, Bachthaler M, Hamer O, Piso P, Herold T, Schlitt HJ, et al. Delayed visceral arterial hemorrhage following

Journal of Vascular Surgery Volume

13.

14.

15.

16.

-,

Number

Hassold et al

11

-

Whipple’s procedure: minimally invasive treatment with covered stents. Ann Surg Oncol 2008;15:824-32. Grutzmann R, Ruckert F, Hippe-Davies N, Distler M, Saeger HD. Evaluation of the International Study Group of Pancreatic Surgery definition of post-pancreatectomy hemorrhage in a high-volume center. Surgery 2012;151:612-20. Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003;14(Pt 2):S199-202. Standop J, Glowka T, Schmitz V, Schafer N, Overhaus M, Hirner A, et al. Operative re-intervention following pancreatic head resection: indications and outcome. J Gastrointest Surg 2009;13:1503-9. Mine T, Murata S, Ueda T, Takeda M, Onozawa S, Yamaguchi H, et al. Contribution of extrahepatic collaterals

to liver parenchymal circulation after proper hepatic artery embolization. J Gastroenterol Hepatol 2014;29:1515-21. 17. Bellemann N, Sommer CM, Mokry T, Kortes N, Gnutzmann D, Gockner T, et al. Hepatic artery stent-grafts for the emergency treatment of acute bleeding. Eur J Radiol 2014;83:1799-803. 18. Lim SJ, Park KB, Hyun DH, Do YS, Park HS, Shin SW, et al. Stent graft placement for postsurgical hemorrhage from the hepatic artery: clinical outcome and CT findings. J Vasc Interv Radiol 2014;25:1539-48. 19. Boufi M, Belmir H, Hartung O, Ramis O, Beyer L, Alimi YS. Emergency stent graft implantation for ruptured visceral artery pseudoaneurysm. J Vasc Surg 2011;53:1625-31.

Submitted Mar 1, 2016; accepted May 28, 2016.