Comparison of Surgical Ligation and Endovascular Embolization in the Management of Type II Congenital Extrahepatic Portosystemic Shunt: A Retrospective Clinical Study

Background: Surgical ligation and endovascular embolization have been recommended for type II congenital extrahepatic portosystemic shunt (CEPS); however, no consensus has been reached. This study was designed to compare the safety and ecacy of surgical ligation and endovascular embolization for the treatment of type II CEPS. Methods: In this retrospective study, 23 consecutive patients diagnosed with type II CEPS between March 2011 and April 2019 were divided into either a surgical group (n=13; 41.5±19.9years) or the interventional group (n =10; 44.9±19.7years). The surgical group underwent laparoscopic surgical ligation of the shunt alone or ligation of the shunt and splenic artery and/or vein. The interventional group underwent endovascular embolization using microcoils, detachable coils and vascular plug. Results: All 23 patients received a one-step shunt closure, and their clinical symptoms were signicantly improved within 3 months post-procedure and without recurrence during follow-up. The serum ammonia levels in both groups decreased after the procedure and dropped to normal level at 6 to 12 months post-procedure. Compared with baseline, the portal vein diameter in interventional group increased signicantly at 3-, 6-, 12-, and 36-month post-occlusion (P=0.01 for all). The procedure time was shorter in interventional group (127.0±43.2minutes) than surgical group (219.8±56.7minutes; P <0.001). The intraoperative blood loss in interventional group (32.0±62.5mL) was less than that in surgical group (238.5±396.9mL; P=0.001). Conclusion: Both surgical ligation and endovascular embolization are effective in the treatment of type II CEPS. Endovascular embolization has the advantages of shorter procedure time, and less intraoperative blood loss. The ligation of the portosystemic shunt and splenic artery and vein is feasible with apparent safety, and it could avoid a second surgical treatment.


Background
A congenital extrahepatic portosystemic shunt (CEPS), also known as an Abernethy malformation, is classi ed based on the extent of intact hypoplastic portal veins. Type I CEPS is characterized with an endto-side portocaval shunt and the absence of intrahepatic portal vein branches, while type II CEPS has some hypoplastic intrahepatic portal veins preserved with a side-to-side portocaval shunt diverting the portal vein blood to the inferior vena cava (IVC) [1][2][3][4][5]. Although CEPS is a rare clinical disease with less than 400 cases reported, it may develop severe complications, such as hepatic encephalopathy, pulmonary arterial hypertension, hepatopulmonary syndrome, or nodular liver lesions, among others [6].
To date, most of the published articles regarding CEPS are case reports, with a few small-sample clinical studies. Therefore, no consensus is achieved in the treatment of CEPS [6]. The recommended strategies for asymptomatic patients with CEPS are keeping them under careful observation or prophylactic interventional therapy [6]. Medical treatment is primarily used to improve symptoms caused by CEPS or its complications. While liver transplantation is the only curative treatment for type I CEPS, shunt occlusion by surgical ligation or endovascular embolization are the most common treatments for type II CEPS. The use of shunt occlusion depends on the expertise of local surgeons and the length and caliber of the shunts; however, a comparison of the two occlusion modalities in the management of type II CEPS has not been reported. We therefore retrospectively compared the safety and e cacy of surgical ligation and endovascular embolization in the treatment of type II CEPS from two medical centers.

Study design and patient population
This was a retrospective clinical study performed at two medical centers in accordance with the Declaration of Helsinki and approved by the respective institutional review boards. All patients provided written informed consent prior to treatment. All patients in this study have not been reported. This retrospective study included consecutive patients diagnosed with CEPS through abdominal ultrasound, computed tomographic angiography (CTA) and/or magnetic resonance imaging (MRI) between March 2011 and April 2019 in two hospitals. Study patient selection was performed in conjunction with radiologists, hepatobiliary surgeons, and interventional radiologists. Patients diagnosed with type II CEPS who presented with one malformation and at least one apparent CEPS manifestation or complication, including hepatic encephalopathy, gastrointestinal bleeding, pulmonary hypertension, hepatic myelopathy, hepatopulmonary syndrome, hepatic cirrhosis, or hepatic adenoma were included in the study. Patients with type I CEPS, asymptomatic patients, multiple malformations, hepatocellular carcinoma, coagulopathies, active infection, chronic renal failure, or contraindications to angiography and surgery were excluded. Initially, portal venography by way of systemic venous access and traversal of the portosystemic shunt and a 15-min shunt balloon occlusion test were performed. Patients with a large and short shunt, which may have high risk of embolic materials migration, or portal venous pressure (PVP) > 25 mmHg after test occlusion, underwent surgical ligation (surgical group), with all others undergoing endovascular embolization (interventional group).

Portographyand shunt balloon occlusion test
To evaluate hemodynamics, measure PVP, and assess whether the shunt vessel could be closed, Portal venography and shunt balloon occlusion test were conducted. However, considering the heterogeneous morphology of CEPS, a standard portogram does not exist. Each angiographic examination was tailored to the patients according to the prior image ndings of the suspected CEPS.
In general, the right femoral artery and femoral vein were cannulated using a 4-F vascular sheath (Radifocus; Terumo, Tokyo, Japan) and 6 to 10-F vascular sheath (Radifocus Introducer II; Terumo, Tokyo, Japan) using the Seldinger technique under local anesthesia or a transjugular or percutaneous transhepatic approach if the transfemoral vein approach failed. An indirect portal venography was then performed via selective angiography of the celiac artery, splenic artery, and superior mesenteric artery using a 4-F hepatic artery catheter (Cordis, Miami Lakes, FL, USA) inserted through the arterial vascular sheath. Next, direct venography of the inferior mesenteric vein, splenic vein, and portal vein were performed using a 4-F Cobra (Cordis) catheter inserted through the venous vascular sheath and catheterized into the portal vein. Lastly, a balloon occlusion test was performed (Fogarty, Thru Embolectomy Catheter, Edwards Lifesciences LLC, Irvine, CA, USA; Boston Scienti c, Watertown, Mass) in the infrahepatic segment of IVC or the shunt to detect the intrahepatic portal branches and measure PVP before and 15 min after balloon occlusion. If the shunt was suitable for endovascular embolization and endovascular occlusion of the shunt was performed at the same time.

Laparoscopic surgical ligation
Under general anesthesia, a catheter was placed in the portal vein via the shunt through the femoral vein to monitor PVP. A 12 mm port for optics was placed in the umbilical area and three 5 mm working ports were placed in upper abdomen, right lower abdomen, and right lateral abdomen.
If PVP was < 25 mmHg after 15 min of occlusion, without signs of intestinal wall edema and redness examined at laparoscopy, the shunt was ligated alone. If PVP was > 25 mmHg and/or there were signs of intestinal wall edema and redness, the splenic artery and vein were dissected and test ligated if PVP was < 25 mmHg without signs of intestinal wall edema and redness after 15 min, the splenic artery and/or vein were also taped. Otherwise, partial occlusion of the shunt was performed and completely occluded during the follow-up, completing a so-called "two-step closure."

Endovascular shunt occlusion
The embolization materials used were either one, or a combination of 0.018 inch pushable micro coils (MicroNester; Cook Medical, Bloomington, USA), a 0.018 inch detachable coils (Interlock; Boston Scienti c, Marlborough, Massachusetts, USA), and an Amplatzer vascular plug (16-mm, 18 mm , 20 mm , or 22 mm AVP IIA; Abbott Park, Illinois USA). After selective shunt catheterization, the microcoils and interlock coil were inserted through a coaxial 2.7-F microcatheter (Progreat; Terumo, Tokyo, Japan) , the plug was inserted through a 6 to 10-F long sheath (Cook medical, USA). Indirect portography through SMA to detect whether there were residual shunts.

Postprocedural treatment
To prevent portal and mesenteric thrombosis, 30 mg of subcutaneous enoxaparin was administered every 12 h, beginning 12 h after the procedure and throughout 7 days of in-hospital observation. Patients were then discharged if no complications occurred and 10 mg of oral rivaroxaban was administered once daily for 30 days.
Abdominal ultrasound and contrast-enhanced CT performed at 1 week and 1 month after the procedure determined whether portal vein system thrombosis had occurred. If there was an absence of thrombosis 1 month after the procedure, rivaroxaban was stopped.

Follow-up
Follow-up at 3, 6, and 12 months after the procedure, and every 12 months thereafter included clinical status, routine blood tests, serum ammonia, liver function, coagulation, and abdominal contrastenhanced CT. The diameter of the main portal vein was measured by two radiologists independently, with the mean value used as the nal result.

Statistical analysis
Continuous variables were recorded as means ± standard deviations, and categorical data as numbers and percentages. All data were tested for normality using a Shapiro-Wilk test. Differences in PVP, hemoglobin, serum ammonia, portal vein diameter, and Child-Pugh score pre-and post-procedure within each group were compared with paired-Student t test (normally distributed) or paired-Wilcoxon rank-sum test (not normally distributed). Differences between groups were tested using a Student t test or Wilcoxon rank-sum test. Differences in the main symptoms, comorbidities, and location of shunt vessels between groups were tested using the two-tailed Fisher exact test. P<0.05 indicated a statistically signi cant difference (SPSS Statistics for Windows, version 20.0; IBM, Armonk, NY).

Patient characteristics
Baseline demographic and clinical data of the study patients are presented in Table 1. A total of 27 patients were recruited, four cases were excluded due to type I CEPS (n = 1), heptocellular carcinoma (n = 1), and incomplete data (n = 2). Twenty-three patients were included in the nal analysis, 13 patients ( ve male, eight female; mean age, 40.7 years ± 20.4; age range, 3-67 years) in the surgical group and 10 patients (eight male, two female; mean age, 44.9 years ± 19.7; age range, 17-69 years) in the interventional group (Fig. 1). The main manifestations of CEPS in this cohort were hepatic encephalopathy (six in the surgical group and ve in the interventional group) and gastrointestinal bleeding (four in the surgical group and two in the interventional group) with no signi cant difference between groups (P > 0.99 and P = 0.66, respectively). Other manifestations included dyspnea, abdominal pain, hepatic myelopathy, hemoptysis, and fatigue, with no signi cant differences between groups (P > 0.99 for all).
Of the 13 patients in the surgical group, 10 experienced hepatic cirrhosis, two experienced hepatic adenoma, one experienced pulmonary hypertension, and one experienced hypersplenism. Similarly, of the 10 patients in the interventional group, ve experienced hepatic cirrhosis, one experienced pulmonary hypertension, and one experienced hypersplenism.

Intraoperative Parameters
All 23 patients received a one-step closure. For patients in the surgical group, three underwent shunt ligation alone, six underwent shunt and splenic artery ligation, and four patients underwent ligation of the shunt, and splenic artery and vein. The post-procedure PVP (17.5 ± 3.8 mmHg; range: 8.1-22.1 mmHg) was signi cantly greater than pre-procedure (12.9 ± 3.7 mmHg; range: 5.9-17.6 mmHg; P < 0.001) in surgical group, but it was lower than 25 mmHg. In interventional group, the shunts were embolized using an Amplatzer vascular plug in four patients, with the remaining six cases using a combination of detachable and pushable coils (Figs. 2-3  The procedure time was signi cantly shorter in the interventional group (127.0 ± 43.2 min) than that in surgical group (219.8 ± 56.7 min; P < 0.001). The intraoperative blood loss in the interventional group (32.0 ± 62.5 mL) was signi cantly less than that of the surgical group (238.5 ± 396.9 mL; P = 0.001). Additionally, the treatment expense of interventional group (46331.5 ± 18839.1 Yuan) was lower than that of surgical group (59561.3 ± 21450.1 Yuan) but without statistical signi cance (P = 0.20) (Table 3).

Clinical Outcomes
The mean follow-up period was 26.1 ± 17.6 months (range, 6-60 months). Clinical symptoms of all patients signi cantly improved within 3 months of their procedure without recurrence during follow-up.
Serum ammonia levels decreased after both procedures to normal levels (9-30 µmol/L) at 6 to 12 months post-procedure. The symptoms of hepatic encephalopathy and hepatic myelopathy associated with hyperammonemia were improved within 1 month of the procedure, and remained improved during the follow-up period (Table 4). Patients with gastrointestinal bleeding (four in the surgical group and two in the interventional group) stopped bleeding after their procedure. One patient in the surgical group suffered hematemesis 6 months after occlusion. Gastroscopy revealed hemorrhage of a gastric ulcer, and bleeding stopped after endoscopic and acid suppression therapy. One patient with hemoptysis and one with dyspnea related to pulmonary hypertension achieved symptomatic relief at 1 week and 3 weeks after the procedure, respectively. Finally, one patient with hepatic adenoma underwent simultaneous resection.
Portal vein thrombosis was detected in one patient in the interventional group using contrast-enhanced CT scan at 1 month after the procedure. Rivaroxaban was discontinued, intravenous heparin sodium (40 U/kg to control the activated partial thromboplastin time between 40 and 60 s) and urokinase (4000 U/kg, twice daily) were administered. Seven days later, the portal vein thrombosis disappeared. Once daily oral rivaroxaban (10 mg) resumed for 4 months without thrombosis (Fig. 3).

Complications
No signs of portal hypertension and splenic abscess have been observed, and no other complications occurred.

Discussion
CEPS is a congenital malformation created by an abnormal shunt between the portal vein and vena cava or azygos/hemiazygos system due to abnormal development of the umbilical vein and yolk vein during the embryonic period. In combination with other congenital malformations, patients with CEPS present with hypergalactosemia, high bile acid, high serum ammonia, and hepatic encephalopathy caused by the portal and vena cava shunt. It has been reported that 66-100% of the patients with CEPS have hyperammonemia and 17-30% have hepatic encephalopathy [7][8][9].
Currently, liver transplantation is the only curative treatment for type I CEPS. Surgical ligation and endovascular embolization have been recommended to treat symptomatic type II CEPS; however, no consensus has been achieved and no comparison between the two procedures has been reported. Except for patients who are not suitable for endovascular occlusion due to shunt anatomy, no research has compared the more advantageous modality for patients suitable for both the treatment procedures so far.
Our study found that both of the procedures were safe and effective in patients with type II CEPS. However, compared with laparoscopic surgical ligation, the endovascular approach has the advantages of shorter procedure time, less intraoperative blood loss, and lower treatment expense.
Angiography is the gold standard for diagnosing CEPS and identifying the type of CEPS present [10,11]. While shunt anatomy and ow dynamics can be determined by angiography and/or direct venography, balloon shunt occlusion venography is critical for detecting intrahepatic portal vein branches. This branch identi cation could also determine the type of CEPS, and the pressure gradients pre-and postocclusion that is crucial for assessing the risk of portal hypertension after shunt occlusion [11][12][13][14].
In previous reports, 25 mmHg was used as a threshold to monitor PVP. If PVP was less than 25 mmHg after 15-20 min of occlusion, it indicated that the risk of portal hypertension was insigni cant after shunt occlusion. Otherwise, a two-step closure is recommended with partial occlusion to initially acclimatize the intrahepatic portal system to increase ow and then completely occlude in 6 to 12 months [11,[15][16][17][18].
In order to avoid portal hypertension post-occlusion for patients with PVP of more than 25 mmHg, we performed surgical ligation of the shunt combined with splenic artery and/or vein and thereby avoided two surgical procedures for shunt closing. Based on the abundant communication among the splenic artery, splenic vein, and splanchnic vessels, ligation of the splenic artery and splenic vein will not typically cause splenic necrosis or abscess. In this cohort, six patients underwent shunt and splenic artery ligation and four underwent additional splenic vein ligation. No portal hypertension or splenic abscess occurred and a second surgical treatment was avoided.
Portal vein and mesenteric venous thrombosis is a common complication after shunt ligation that can lead to portal hypertension; therefore, postprocedural anticoagulation therapy is very important [6,10].
Heparin and warfarin have been used to prevent thrombosis in previous studies; however, the activated partial thromboplastin time and international normalized ratio should be closely monitored to adjust drug dosage [16,19]. In our practice, enoxaparin and rivaroxaban have been used sequentially for anticoagulation after occlusion, and do not require monitoring indicators of coagulation and is more convenient to use. In the follow-up period, only one patient (1/23) suffered portal vein thrombosis, which disappeared after 7 days of conservative treatment.
Theoretically, the hypoplastic intrahepatic portal vein should dilate after the perfusion of portal vein blood increased due to shunt ligation. In this cohort, patients after endovascular embolization achieved signi cant increase in portal vein diameter, however, there was no signi cant difference in surgical ligation. This maybe attribute to the limited number of cases and the well-grow intrahepatic portal vein branches in several patients before occlusion.
Our study had some limitations. First, this study included only a small number of patients primarily due to the low incidence of CEPS. Second, the follow-up period was short, and due to the retrospective characteristics of the study, the follow-up intervals were not completely consistent and partial follow-up data was missing. Thirdly, the portosystemic gradient gradient was not measured.

Conclusions
Both surgical ligation and endovascular embolization are effective in the treatment of type II CEPS. The endovascular approach has the advantages of shorter procedure time, less intraoperative blood loss, and lower treatment expense. Surgical ligation is recommended for patients with a high risk of embolic materials migration due to short and large extrahepatic portosystemic shunt, and patients with PVP more than 25 mmHg after test occlusion, which could lead to portal hypertension. Ligation of the portosystemic shunt and the splenic artery and vein is feasible with apparent safety, and it could avoid a second surgical treatment.