Transcatheter arterial embolization for massive hemobilia with N-butyl cyanoacrylate (NBCA) Glubran 2

Background Massive hemobilia is a life-threatening condition and therapeutic challenge. Few studies have demonstrated the use of N-butyl cyanoacrylate (NBCA) for massive hemobilia. Purpose To investigate the efficacy and safety of transcatheter arterial embolization (TAE) using NBCA Glubran 2 for massive hemobilia. Material and Methods Between January 2012 and December 2019, the data of 26 patients (mean age 63.4 ± 12.6 years) with massive hemobilia were retrospectively evaluated for TAE using NBCA. The patients’ baseline characteristics, severities of hemobilia, and imaging findings were collected. Emergent TAE was performed using 1:2–1:4 mixtures of NBCA and ethiodized oil. Technical success, clinical success, procedure-related complications, and follow-up outcomes were assessed. Results Pre-procedure arteriography demonstrated injuries to the right hepatic artery (n = 24) and cystic artery (n = 2). Initial coil embolization distal to the lesions was required in 5 (19.2%) patients to control high blood flow and prevent end-organ damage. After a mean treatment time of 11.2 ± 5.3 min, technical success was achieved in 100% of the patients without non-target embolization and catheter adhesion. Clinical success was achieved in 25 (96.2%) patients. Major complications were noted in 1 (3.8%) patient with gallbladder necrosis. During a median follow-up time of 16.5 months (range 3–24 months), two patients died due to carcinomas, whereas none of the patients experienced recurrent hemobilia, embolic material migration, or post-embolization complications. Conclusion NBCA embolization for massive hemobilia is associated with rapid and effective hemostasis, as well as few major complications. This treatment modality may be a promising alternative to coil embolization.


Introduction
Hemobilia is defined as bleeding into the biliary tree and is an uncommon source of upper gastrointestinal bleeding (1,2). Massive hemobilia with hemodynamic instability is a life-threatening condition and therapeutic challenge (3). Although transcatheter arterial embolization (TAE) using coils is the mainstay treatment for massive hemobilia (1,2), it might fail because of coagulopathy or vascular anatomy complicated by tortuous or narrow feeding arteries (4,5). N-butyl cyanoacrylate (NBCA) is a liquid permanent embolic material that polymerizes in contact with fluids rich in ions.
The polymerization of NBCA does not depend on coagulation parameters. Thus, it provides better hemostasis than that provided by other embolic materials, especially in patients with coagulopathy (6). Moreover, NBCA offers the advantages of a liquid state of matter and low viscosity to penetrate tortuous and narrow involved arteries (6). Glubran 2 (GEM, Viareggio, Italy) is a specific surgical glue in which NBCA is combined with another monomer, metacryloxysulfolane, to produce a more pliable and stable polymer, the milder exothermic reaction (45 C), which results in less inflammation and histotoxicity than those associated with Histoacryl or Trufill (6).
NBCA Glubran 2 has been used for various bleeding conditions (7,8). However, in the setting of massive hemobilia, only a few cases have been reported (9)(10)(11)(12)(13). Although the reported outcomes are encouraging, the use of NBCA for massive hemobilia has not been well evaluated. The aim of the present study was to evaluate the efficacy and safety outcomes of TAE using NBCA Glubran 2 for massive hemobilia.

Study design
Between January 2012 and December 2019, the data of patients with massive hemobilia who underwent TAE using NBCA in a single center were retrospectively reviewed. The present study was approved by the hospital review board. The patients' baseline characteristics including demographics, co-morbidities, manifestations, medications, etiologies, laboratory data on coagulation parameters (international normalized ratio, partial thromboplastin time, platelet count), and imaging findings (upper endoscopy, computed tomography [CT], and arteriography) were collected. The severity of hemobilia, level of hemoglobin, systolic blood pressure, and number of red blood cell (RBC) units transfused before TAE were obtained. Procedurerelated details including arteriography findings, bleeding sites, vessel(s) embolized, embolic materials used, treatment time, and procedure-related complications were obtained. Technical success, clinical success, post-embolization complications, and follow-up outcomes (recurrent hemobilia, embolic material migration, or post-embolization complications) were collected to evaluate the efficacy and safety of NBCA embolization for hemobilia. Data were collected using electronic medical records, including reviews of clinical notes, laboratory values, procedure images, and procedure reports.

Patients and diagnosis of hemobilia
The data of 26 patients (12 men, 14 women; mean age ¼ 63.4 AE 12.6 years; age range ¼ 30-81 years) were reviewed and analyzed. The most common manifestations were melena (53.8%) and right upper quadrant pain (61.5%). Quincke's triad (melena and/or hematemesis, right upper quadrant pain, and obstructive jaundice) was noted in 9 (34.6%) patients. The etiologies of hemobilia were iatrogenic except in 2 (7.7%) patients with gallbladder stones and acute cholecystitis. The co-morbidities, medications, and prior coagulopathies are listed in Table 1. Regarding the severity of hemobilia, the RBC units transfused, level of hemoglobin, and systolic blood pressure before the TAE procedure are listed in Table 2.
The diagnosis of hemobilia was based on the patients' manifestations, recent interventional procedures, and imaging findings. Hemobilia was diagnosed directly in 8 (30.8%) patients with massive bloody output from a percutaneous transhepatic biliary drainage tube (n ¼ 4) or T-tube (n ¼ 4). Pre-procedure upper endoscopy was performed in 16 patients. The most common finding of upper endoscopy was blood in the duodenum, whereas active bleeding from the ampulla was found in only two patients. The most Values are given as n (%) or mean AE SD (range). *Partial thromboplastin time >45 s and/or platelet count <80,000/mm 3 and/or international normalized ratio >1.5. PTBD, percutaneous transhepatic biliary drainage; SMA, superior mesenteric artery.
common pre-procedure CT findings (available for 16 patients) were pseudoaneurysms and contrast extravasation (Table 2). Finally, all patients underwent arteriography to confirm the diagnosis and to identify the bleeding sites.

Management of hemobilia and TAE techniques
All patients experienced hemodynamic instability and received resuscitation therapy with fluid or blood transfusion before TAE. During the procedure, 21 (67.7%) patients still exhibited an unstable hemodynamic status. NBCA, rather than coils, was selected as the primary embolic material for the following reasons: (i) rapid embolisation was required for an unstable hemodynamic status; (ii) we experienced difficulties in using coils successfully; and (iii) we experienced difficulties in accessing the target vessel complicated by extremely tortuous or narrow vascular anatomy. The decision to use NBCA was based predominantly on interventional radiologists' judgement and experience. Coeliac and superior mesenteric arteriographies were performed using a 5-F multipurpose catheter to visualize the arterial anatomy and to identify the source of hemobilia using a transfemoral approach. In addition, the patency of the portal vein was evaluated based on the delayed portal vein phase. A 2.8-F microcatheter (Terumo Corp., Tokyo, Japan) was subsequently introduced coaxially with its tip advanced as close to the bleeding site as possible or into the pseudoaneurysm. After flushing the microcatheter with a 5% dextrose solution, an NBCA-ethiodized oil mixture (Ethiodized Poppyseed Oil injection; Hengrui Medicine, Jiangsu, PR China) (in a ratio in the range of 1:2-1:4) was injected carefully under real-time highresolution fluoroscopic mapping until it reached the pseudoaneurysm or bleeding site. Based on the operators' experience, initial embolization distal to the lesions with pushable 0.018-inch coils (MicroNester Embolization Coil or Hilal Embolization Microcoil; Cook Medical, Bloomington, IN, USA) was performed in some patients to control high blood flow and prevent end-organ damage (14). The ratio, volume, and injection rate of the mixture were based on the size, distance, and target vessel flow. The microcatheter was removed swiftly after injection to avoid catheter adhesion. Completion coeliac and superior mesenteric artery arteriography were performed to confirm the absence of pseudoaneurysms, extravasation, and residual bleeding sites and to evaluate collateral vessels in the embolized hepatic area. In addition, non-target embolization was also assessed by comparing the findings of pre-procedure and completion arteriographies.
All patients underwent close surveillance for postembolization complications and potential aggravation of symptoms and signs. Intravenous antibiotics targeting the biliary microflora were administered in all patients. In addition, all patients underwent post-procedure CT to evaluate the embolization efficacy and potential hepatic infarction or embolic material migration.

Definitions and follow-up
Technical success was defined as complete occlusion of the target vessel or absence of pseudoaneurysms and extravasation on completion arteriography. Clinical success was defined as the cessation of bleeding after TAE with no need for repeat embolization or additional surgery for hemostasis. The treatment time was defined as the duration from the identification of bleeding sites to retraction of the microcatheter. Major complications were defined as unplanned surgery, permanent adverse sequelae, or prolonged hospitalization (15).
During follow-up, CT and/or color Doppler ultrasonography and clinical evaluation were performed on an outpatient basis for all patients (including event-free patients) at one and three months or sooner when clinically indicated. Any instance of recurrent bleeding, hemobilia, embolic material migration, or postembolization complications were recorded.

Results
Before the TAE procedure, patency of the portal vein was observed in all patients. Arteriography demonstrated pseudoaneurysms and/or extravasation of the hepatic or cystic artery in 25 (96.2%) patients  (25) Values are given as n (%), mean AE SD, or median (range). TAE, transcatheter arterial embolization.
( Table 3). One patient underwent upsizing of the drain to control hemobilia, and segmental right hepatic artery transection was the only finding on subsequent arteriography, while the upsized drain was in place. This patient still underwent NBCA embolization of the segmental right hepatic artery for safety. A bleeding site complicated by a tortuous or narrow feeding artery was identified in 8 (30.8%) patients (Fig. 1). NBCA was used as the sole embolic material in 21 (80.8%) patients. Initial coil embolization with subsequent NBCA embolization was performed in the remaining five patients. The embolization site is listed in Table 3.
For the two patients who underwent embolization of the main right hepatic artery, opacification of the distal vessel through intrahepatic collaterals was observed on completion arteriography. The mean injected volume of the mixture was 1.1 AE 0.5 mL (range ¼ 0.5-2.0 mL) and the mean treatment time was 11.2 AE 5.3 min (range ¼ 4-21 min). Technical success was achieved in all patients and clinical success was achieved in 25 (96.2%) patients. No patient underwent secondary NBCA embolization during the same treatment session. Of note, successful NBCA embolization was achieved in a patient who experienced unsuccessful initial coil embolization (Fig. 2). One patient experienced a notable hemoglobin decrease (3 g/dL) three days after NBCA embolization and was considered to have resulted in clinical failure. Unfortunately, this patient refused to undergo further imaging examinations. Post-procedure CT revealed the absence of previously identified pseudoaneurysms and/or contrast extravasation in all patients. Moreover, no patient had imaging-based evidence of hepatic infarction, abscesses, or embolic material migration, whereas gallbladder necrosis was observed in one patient and was considered a major complication. Subsequent cholecystectomy was required for this patient. Six (23.1%) patients developed post-embolization abdominal pain, and the symptoms were dismissed after conservative treatment. No other procedure-related complications were noted, including non-target embolization and catheter adhesion. Two patients died during hospitalization: one patient died due to septicemia and another patient died due to advanced carcinoma. The post-procedure CT and laboratory test findings were evaluated carefully, and neither of the deaths seemed to have a relationship with massive hemobilia or the TAE procedure.

Discussion
The present study revealed that TAE using NBCA Glubran 2 is associated with rapid embolization, a high success rate, and few procedure-related complications. The reported success rate of coil embolization for hemobilia is in the range of 75%-100% (3,(16)(17)(18)(19). The technical and clinical success rates were 100% and 96.8% in the present study, respectively, which are comparable to those obtained in previous studies. No recurrent hemobilia or bleeding was noted during the follow-up, which demonstrated the durability of NBCA embolization.
Surgery and TAE are considered first-line therapies for hemodynamically unstable hemobilia, whereas surgical intervention is infrequently necessary unless other interventions have failed (1,2,20). In the present study, one patient developed massive hemobilia due to gallbladder stones and cholecystitis. Although successful hemostasis was achieved after TAE, the patient developed gallbladder necrosis and underwent subsequent cholecystectomy. Considering the potential risk of gallbladder necrosis secondary to cystic artery embolization (20) and necessity of removing the inflamed gallbladder (1), initial surgical intervention might be considered in such cases. TAE using coils has become the mainstay treatment for massive hemobilia (1,2). Patency of the portal vein should be confirmed before TAE, as performing hepatic artery embolization in the setting of portal vein occlusion may lead to significant hepatic ischemia. According to the available evidence, coil embolization is associated with a high success rate and few complications (3,(16)(17)(18)(19). However, there are several limitations to coil embolization. First, catheterization into the target vessel complicated by extremely tortuous or narrow vascular anatomy may be difficult, resulting in failed coil embolization (4). Second, successful coil embolization is dependent on normal coagulation parameters, and it may fail in patients with coagulopathy (5). These limitations may be overcome by NBCA Glubran 2. The added ethiodized oil provides radiopacity and helps adjust the concentration of NBCA in the mixture and its polymerization time (21), allowing the operator to control the degree of distal vessel penetration (7). In the present study, technical success was achieved in all eight patients with tortuous or narrow feeding arteries. In addition, NBCA polymerizes regardless of the coagulation condition. Thus, it is effective in patients with coagulopathy with a success rate of up to 100% (5,14,22). In the present study, seven patients with coagulopathy were successfully treated with NBCA embolization. Furthermore, NBCA has the advantage of rapid embolization, which is essential for massive hemobilia with hemodynamic instability. Yonemitsu et al. (22) reported a significantly reduced time for TAE using NBCA compared with that using microcoils (9 min vs. 37 min, P < 0.001). In the present study, a similar treatment time of 11.2 min was achieved.
Another potential advantage of NBCA is that it is cost-effective compared with coils (7). One milliliter of NBCA Glubran 2 is comparable in cost to a single conventional pushing coil and is sufficient for successful treatment in most cases (7,23). Similarly, no more than 1 mL of NBCA was required in the present study, which may have led to the low costs. In addition, coil embolization requires a distal-to-proximal fashion to avoid back-bleeding via intrahepatic arterial collaterals (24) which may result in hepatic ischemia and infarction (19), especially when a pseudoaneurysm or extravasation originates from a main artery. However, in this scenario, NBCA embolization may embolize the pseudoaneurysm and preserve the main artery. (25) In the present study, 3 patients successfully underwent direct pseudoaneurysm embolization using the packing technique, and the involved main arteries were preserved. A stent graft or flow-diverting stent may also be used to preserve the main arteries (26,27). However, these devices usually require great costs and a large delivery system, causing difficulties in small target vessel placement. As mentioned earlier, NBCA Glubran 2 has several potential advantages over coils, including embolization of tortuous or narrow target vessels, effectiveness in cases of coagulopathy, rapid embolization, cost-effectiveness, and preservation of the involved main artery. Thus, NBCA embolization is a promising treatment modality for massive hemobilia.
Despite the notable advantages compared with those of coils, the use of NBCA for hemobilia remains limited, perhaps due to the technical difficulties of NBCA embolization and fear of procedure-related complications. (20) The reported complications regarding NBCA include non-target vessel embolization, adhesion, and fracture of the microcatheter (25,28). However, according to our experience, NBCA embolization can be performed safely after a short learning curve. Moreover, Madhusudhan et al. (25) reported a modified injection technique of repeated low-dose NBCA mixture (0.1-0.3 mL) injection flushing in which they achieved satisfactory results. In the present study, no inadvertent distal embolization, non-target vessel embolization due to backflow, or catheter adhesion was observed. This may have been because all the procedures were performed by experienced interventional radiologists, and they were familiar with the characteristics of NBCA.
There are several limitations to the present study. First, this was a retrospective study with a relatively small sample size. Second, a comparison between NBCA Glubran 2 and coils is not available for the present study. Future comparative studies are needed to identify the most appropriate embolic material for massive hemobilia. Third, the decision to use NBCA and coils was made at the discretion of interventional radiologists, which may lead to selection bias. Despite these limitations, to the best of our knowledge, this study is the largest case series regarding TAE using NBCA Glubran 2 for massive hemobilia.
In conclusion, TAE using NBCA Glubran 2 for massive hemobilia is associated with rapid and effective hemostasis, as well as few major complications. This treatment modality may be a promising alternative to coil embolization.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.