Long-term clinical and radiological outcome of endovascular embolization of pancreatitis-related pseudoaneurysms

Introduction

Arterial pseudoaneurysms are pulsatile, blood-filled extraluminal sacs without a clear endothelial layer that communicate with the arterial lumen. Pseudoaneurysms are associated with pancreatitis in approximately 10% of patients (1) and occur due to enzymatic insult to peripancreatic arteries or from erosion of a pseudocyst into adjacent visceral arteries. Although rupture of a pseudoaneurysm is rare (1), it is a serious complication as it can lead to a massive peripancreatic hemorrhage and significant clinical deterioration with a mortality of up to 37% (2). Clinical symptoms of pancreatitis-induced pseudoaneurysms can range from asymptomatic to mild abdominal pain, to melena or major hypotensive shock due to ongoing or intermittent bleeding into the gastrointestinal tract, the peritoneal cavity, or the retroperitoneum. Due to these variable and sometimes non-specific clinical symptoms, imaging studies are important in accurate, timely, and appropriate diagnosis of the pseudoaneurysm.

Conservative management of such active hemorrhages is associated with a high mortality rate of up to 52% (3) and is unrelated to the size of the pseudoaneurysm and the severity of the pancreatitis. Accurate and timely intervention may reduce mortality. Surgical intervention used to be the standard of care, but due to its associated high morbidity and mortality (3), it has now been replaced by minimally invasive treatment techniques in most institutions (4). Endovascular, catheter-directed therapy, which is less invasive and has shown an excellent short-term outcome, is currently the first-choice treatment in patients presenting with pancreatitis-related pseudoaneurysms, despite the limited available data on long-term outcomes after endovascular therapy (5). Most of these data come from case reports to small series with a short follow-up time (2,6–10).

The aim of this study was to assess the safety, short-term and long-term efficacy of transcatheter embolization and the overall survival of patients with an embolized pancreatitis-related pseudoaneurysm.

Material and Methods

Results

Patient demographics and pre-interventional laboratory data

The demographics, medical history, and laboratory findings prior to embolization are listed in Table 1. The majority of included patients were male, suffering from chronic pancreatitis related to alcohol misuse, and presenting with clinical signs of abdominal pain.

OPEN IN VIEWER

Table 1.

Demographics, medical history, and laboratory findings of 34 patients who underwent endovascular interventions for management of pancreatitis-related bleeding.

Number of patients (n)	34
Age: mean, range (years)	55 (35–85)
Male: female ratio	25:9
Underlying disease (n)
Acute pancreatitis	13 (38%)
Chronic pancreatitis	21 (62%)
Cause of pancreatitis (n)
Alcohol misuse	25 (73.53%)
Unknown	6 (17.65%)
Cholecystolithiasis	1 (2.95%)
Post-ERCP	1 (2.95%)
Genetically	1 (2.95%)
Clinical presentation (n)
Acute (or increase in) abdominal pain	18 (52.94%)
Melena or RBPA	5 (14.71%)
Fatigue, minimal symptoms	5 (14.1%)
Hematemesis	3 (26.47%)
Bleeding from abdominal drain	2 (5.88%)
Unclear	1 (2.95%)
Laboratory abnormalities
Hemoglobin level before intervention (g/dL) (n = 34)
Mean	9.56
Maximum	15.1
Minimum	3.0
Amylase before intervention (U/L) (n = 27)
Mean	152
Maximum	1000
Minimum	6
Lipase before intervention (U/L) (n = 31)
Mean	231
Maximum	1731
Minimum	5
Hospital stay (days) (n = 29)
Mean	29
Median	12
Maximum	180
Minimum	3

ERCP, endoscopic retrograde cholangiopancreatography; RBPA, red blood loss per anum.

Angiographic and interventional results

The median and mean time period between non-invasive radiological diagnosis on US/CT/MRI and therapeutic angiography was 0 days (range, 0–67 days) and 4 days, respectively. In two patients who had no clear symptoms, the endovascular embolization procedure was performed 24 and 67 days, respectively, after radiological diagnosis, primarily related to the patient’s preference in regard to undergoing the embolization procedure.

The therapeutic angiographic procedure was performed under local (n = 33; 97%) or general anesthesia (n = 1; 3%).

The location and the presentation of the pseudoaneurysm at catheter-directed angiography is listed in Table 2. One patient presented with two separate pseudoaneurysms, hence the total of 35.

OPEN IN VIEWER

Table 2.

Arterial location of the pseudoaneurysm, n (%).

Splenic artery	12 (34%)
Pancreaticoduodenal artery	9 (26%)
Gastroduodenal artery	6 (17%)
Superior mesenteric artery	3 (8%)
Replaced right hepatic artery	2 (6%)
Left hepatic artery	1 (3%)
Left gastric artery	1 (3%)
Supraduodenal artery	1 (3%)
Angiographic presentation (n = 35)
Pseudoaneurysm without contrast extravasation	30 (86%)
Extravasation of contrast (active bleeding)	5 (14%)

As one patient had two pseudoaneurysms on two different arteries, there were 35 pseudoaneurysms in total, in 34 patients.

Table 3 summarizes the embolic agents used to occlude the pseudoaneurysm using an endovascular approach. The sandwich technique using microcoils was the embolization technique used in the majority of cases (Fig. 1). OPEN IN VIEWER

Fig. 1.

A 46-year-old woman with chronic alcohol-related pancreatitis was diagnosed with an arterial pseudoaneurysm. (a) Angiogram after selective catheterization of the coeliac trunk shows a round opacity in the middle segment of the splenic artery (small arrow), confirming the pseudoaneurysm. (b) Microcatheter-guided coil embolization (small arrow) is successfully performed, as no pseudoaneurysm is seen on the control angiography. Note the hypertrophic gastrosplenic collaterals bridging the occluded main splenic artery. (c) Local splenic infarction is seen as hypodense areas in the spleen (thick arrow). Note the artefact caused by the coils in the splenic artery (small arrow).

OPEN IN VIEWER

Table 3.

Endovascular embolization technique, n (%).

Microcoils (sandwich technique*)	20 (57%)
Liquid agents (glue)	4 (11%)
Microcoils + glue	4 (11%)
Microcoils + microparticles	3 (8.6%)
Microparticles	2 (2.9%)
Microcoils (proximal)	2 (2.9%)
Microcoils + thrombin	1 (3%)

*

Sandwich technique: placement of coils from distal to proximal to the opening of the pseudoaneurysm, to interrupt arterial flow to the pseudoaneurysm. (Figs 1c, 2b, 3c).

Procedure-related complications

Table 4 summarizes the type and frequency of the complications related to the interventional treatment. No major complications, defined as complications of grade II or higher according to the SIR classification were observed. Seventy percent of patients had no procedure-related complications at all. The most frequent minor complication (SIR classification type I) was a partial, focal splenic infarction, which clinically presented itself asymptomatically or with mild fever and pain.

OPEN IN VIEWER

Table 4.

Procedure-related complications, n (%).

No complications	26 (70.30%)
Splenic infarction*	4 (10.80%)
Migration of microcoils †	2 (5.41%)
Postprocedural fever ‡	2 (5.41%)
Subcapsular splenic hematoma §	1 (2.70%)
Radiodermatitis**	1 (2.70%)
Subcutaneous hematoma at the puncture site ††	1 (2.70%)

*

Partial splenic infarction as a result of the diminished arterial flow after splenic artery embolization was shown in four patients on their CT scans after treatment (Fig. 2c). It is listed as a minor complication, as there were no clinical signs and no major treatment is needed. A pneumococcal vaccine was given and needs to be repeated every 5 years.

†

No clinical repercussion.

‡

Two patient experienced postprocedural fever, for which antibiotics were given, with resolution of the fever.

§

No treatment was needed.

**

Most probably related to long angiographic procedure.

††

Treated with ultrasound-guided thrombin injection.

Radiological and clinical outcome

The radiological follow-up period, defined as the time interval between the embolization procedure and the latest follow-up CT, was a mean of 40.5 months (median, 30.5 months; range, 0.1–124 months). The clinical follow-up period, defined as the time interval between the embolization procedure and the end of the study (January 2015) or the patient’s death was a mean of 6.8 years (range, 4 months–16 years).

Four patients (12%) had a recurrent pseudoaneurysm, identified on follow-up CT scan; in three of them, the newly formed pseudoaneurysm arose on a different vessel from the initially embolized pseudoaneurysm 4, 20, and 127 days after initial embolization, respectively. In one patient (3%), the recurrence of the pseudoaneurysm was a re-opening of a previously successfully embolized pseudoaneurysm (Fig. 2) identified 136 days after initial embolization. All four recurring pseudoaneurysms were successfully treated with a catheter-directed approach, using microcoils as embolic material (Fig. 3). No further pseudoaneurysm formation occurred during further follow-up. OPEN IN VIEWER

Fig. 2.

A 69-year-old man, with a history of chronic pancreatitis due to alcohol misuse, underwent a contrast-enhanced CT due to increasing abdominal pain. (a) Axial CT image demonstrates a contrast-enhancing structure at the splenic hilum (arrow), highly suggestive of a pseudoaneurysm. (b) Angiogram confirms a saccular pseudoaneurysm at the distal end of the splenic artery (arrow). (c) Coils are deployed in the splenic artery distal and proximal to the site of the pseudoaneurysm (“sandwich technique”). Control angiogram shows successful embolization (arrow).

OPEN IN VIEWER

Fig. 3.

Four months and 2 weeks later, there was a recurrent increase in epigastric pain, and a contrast-enhanced CT was performed. (a) Axial CT image demonstrates re-opacification of a pseudoaneurysm which was initially successfully embolized (arrow). Note the artifact of the previously placed coils. (small arrow). Note the hypodense areas within the spleen, which represent focal splenic infarcts (arrowhead). (b) Angiogram shows a pseudoaneurysm at the splenic artery (thick arrow), proximal to the previously coiled area (small arrow). (c) Extra microcoils proximal to the initially placed microcoils (thick arrow), were added, resulting in complete exclusion of the pseudoaneurysm. Note the migration of one microcoil to a superior branch of the splenic artery (small arrow).

Five patients (14.7%) died during the mean follow-up of 6.8 years. None of the five patients died as a result of pancreatic-related disease: one patient died of an intracranial bleed, two patients died related to an underlying carcinoma, and two patients died due to liver cirrhosis and terminal liver failure.

Discussion

Our findings in relation to diagnostic methods to identify pancreatitis-related pseudoaneurysms seem to confirm previous findings: contrast-enhanced CT is an accurate, non-invasive radiological modality for the diagnosis of pancreatic inflammatory disease in general and of pancreatitis-induced pseudoaneurysm in particular (2), while ultrasound helps to rule out other pathology but has a low sensitivity for detecting visceral pseudoaneurysms (3). We found a sensitivity of 97% for contrast-enhanced CT, which is nearly identical to the findings of Barge et al. which showed a sensitivity of 95.1% for CT in detection of pancreatitis-related pseudoaneurysms (3).

The splenic artery was the most frequently affected vessel (34%), which corresponds to data from other studies, demonstrating splenic artery involvement in 30%, 34%, and 44%, respectively (7,8,12). This is a logical consequence of the topographic anatomy of the splenic artery and the pancreas.

Five patients (14.7%) had active bleeding at diagnosis, while in 29 patients (85.3%) the diagnosed pseudoaneurysm was not ruptured. None of these five actively bleeding patients experienced failure of the endovascular embolization or mortality associated with the hemorrhagic shock, embolization procedure, or underlying pancreatitis.

When a pancreatitis-related pseudoaneurysm is diagnosed, a conservative “wait-and-see” strategy seems to be unacceptable owing to the unpredictable risk of rupture of the pseudoaneurysm and subsequent massive, life-threatening hemorrhage with mortality in up to 37% of cases (2). We had a mean time interval of 4 days between radiological diagnosis and therapeutic angiographic intervention, which seems to be acceptable for provision of (semi-) urgent treatment.

Different endovascular embolization techniques are described, which are all based on the same principle: to exclude the pseudoaneurysm from the arterial circulation. In this study, microcoil placement from distal to proximal from the pseudoaneurysm (sandwich technique) was the most frequently used technique, which is in line with other studies (2,3,7,12). The use of other embolic agents, however, such as glue, microparticles, or a combination of these embolic agents was also efficient in excluding the pseudoaneurysm. The choice of technique is partly operator-dependent and partly vessel-dependent: glue can be delivered from proximal to distal in narrow or tortuous vessels that would otherwise be difficult to reach with a microcatheter (13); more accurate and controlled embolization can be done using microcoils. Other interventional techniques such as endovascular stent-grafting (7) and ultrasound-guided percutaneous thrombin injection or the use of ethylene vinyl alcohol copolymer (Onyx) (13) have been used successfully in selected cases (14–16), but none of these techniques were used in the patient group that was studied.

Procedure-related complications were categorized as major or minor complications, according to the SIR classification for procedure-related complications (11). Although major complications have been rarely described (12), in this study no major complications were noted. In 30% of patients, however, a minor complication was seen immediately after the embolization procedure, with partial splenic infarction as most frequent complication, which is in line with other published studies showing partial splenic infarction rates in the range of 8–18% (4,7). Endovascular treatment has become the treatment of choice owing to its high efficiency and low invasiveness and associated low morbidity and mortality. Nevertheless, open surgery might still be indicated in these cases of unsuccessful or failed embolization, although it is clear that open repair is associated with higher risks of procedure-related morbidity and mortality (3,8). Additionally, this study underlines that catheter-directed embolotherapy can be used for the definitive treatment of both asymptomatic and symptomatic pancreatitis-related pseudoaneurysms and in particular of bleeding pseudoaneurysms in patients in hemodynamic shock.

The short-term outcome of embolization is well described in the literature, with success rates in the range of 77–100% (3,4,7,9). The overall recurrence rate of pseudoaneurysm was 12%, but only one patient (3%) had a recurrence due to re-opening of the initial pseudoaneurysm that had been successfully embolized. This case showed the need of adequate proximal embolization. All recurrences (12%) occurred within the first 6 months after initial embolization, indicating that this may be a critical time interval, during which clinical and radiological follow-up is strictly necessary. This observation is in line with the findings of Golzarian et al. (9) who showed recurrent pseudoaneurysm formation and bleeding 18 and 28 days after initial embolization, respectively. During further follow-up, no other recurrences occurred, indicating the lasting efficacy of endovascular treatment.

Another finding of this study is the high overall long-term survival of patients who had endovascular treatment for pancreatitis-related pseudoaneurysms (mean follow-up, 6.8 years). The overall survival of the embolized patients, however, relies on multiple factors, mainly on the prognosis of the underlying pancreatitis and other co-morbidities.

There are several limitations to this study. First, it is limited by its retrospective nature. Second, although this study is based on a relatively large patient group, taking into account the rarity of the disease, a total of 34 patients may limit generalization of our findings. A larger group of patients will therefore be needed in the future, in order to reach a higher level of significance.

In conclusion, pancreatitis-related pseudoaneurysm is still a rare disease that should be investigated with contrast-enhanced CT. Interventional, catheter-directed treatment is relatively safe and effective, and can be considered as the first-choice treatment option. Long-term radiological and clinical follow-up of embolized patients shows satisfactory results with durable pseudoaneurysm exclusion and absence of recurrent pseudoaneurysm-related symptoms in the vast majority of patients. Recurrent disease is low and seems to occur within the first 6 months after initially successful embolization.