Left-sided (sinistral) portal hypertension (LPH) is characterized by the development of gastric varices secondary to obstruction of the splenic vein, most commonly as a result of pancreatic pathology such as pancreatitis and pancreatic malignancy.^1–3 The most common cause of the occlusion is thrombosis of the splenic vein or compression by an adjacent mass.¹ In the setting of splenic vein occlusion, venous outflow from the spleen travels towards the stomach via the short gastric veins and then towards the portal vein.^3,4 Collateral circulation from the gastric collateral veins leads to the dilation of gastric submucosal vessels and subsequently increased venous flow and pressure, resulting in isolated gastric varices, which pose a risk of massive hemorrhage if ruptured.⁵ However, there is no consensus on the treatment of acute bleeding in LPH. Traditionally, splenectomy has been considered a treatment of choice for bleeding gastric varices secondary to LPH.⁶ However, the risks associated with emergent splenectomy and post-operative recovery are not neglectable because of inflammation, adhesion, and bleeding tendency. Furthermore, there is a lack of evidence regarding the long-term efficacy of splenic artery embolization in the setting of acute gastric variceal hemorrhage, as well as the optimal technique for performing the embolization. In this study, we present a retrospective analysis of patients who received splenic artery embolization for gastric variceal bleeding secondary to LPH, focusing on the embolization techniques and outcomes.

Patients

The institutional review board of the University of North Carolina at Chapel Hill approved this retrospective study (IRB No. 17-2445) with a waiver of informed consent. From April 2008 to March 2020, all consecutive patients diagnosed with LPH complicated by gastric variceal bleeding and had undergone splenic artery embolization were identified in an administrative database. Patients who received splenectomy after splenic artery embolization were excluded. During the study period, 22 patients received splenic artery embolization for LPH. Three patients who received splenectomy after embolization (one for retroperitoneal mass, two for scheduled splenectomy) were excluded. Nineteen patients (median age: 44.5 years, range: 27–83 years, 13 male and 6 female) finally enrolled in this study. Patient demographics, presenting symptoms, embolization reports, and follow-up clinical reports were reviewed in the electronic medical record. Picture archiving and communication system was reviewed for preprocedural imaging, embolization procedures, and follow-up imaging.

Diagnosis

The diagnosis of LPH was based on the results of computed tomography (CT) or magnetic resonance imaging (MRI), clinical manifestation, endoscopy, and history. Imaging studies provided detailed images of the liver, pancreas, and spleen, and the abdominal vessels such as splenic arteries, splenic veins, portal veins, and collateral veins. Splenic vein occlusion was determined by contrast-enhanced CT/MRI findings, including splenic vein thrombosis, extrinsic compression by adjacent mass, and splenic vein stricture accompanied by gastric collateral vein development. On endoscopic findings, patients with peptic ulcer were excluded from this study.

Splenic artery embolization procedure

All embolization procedures were performed in an interventional radiology suite under general anesthesia or conscious sedation. Celiac and splenic artery access was made via the right femoral or left radial route using standard angiographic techniques. Celiac arteriography and selective splenic arterial angiography were performed with a 5-Fr diagnostic catheter and a 2.4-Fr microcatheter. Delayed splenic venous drainage was evaluated during the splenic angiography to determine embolization territory or level (proximal or distal). Splenic artery embolization techniques included: i) proximal splenic artery embolization, ii) partial parenchymal embolization, and iii) total parenchymal embolization. For the proximal embolizations, pushable platinum coils (Nester and Tornado; Cook Medical, Bloomington, IN, USA) or detachable vascular plugs (Amplatzer Vascular Plug IV; Abbott Medical, Plymouth, MN, USA) were used. For distal splenic parenchymal embolization, Gelfoam slurry (Gelfoam; Pfizer, Kalamazoo, MI, USA) or non-absorbable microspheres (500–700, 700–900 μm) (Embosphere microspheres; BioSphere Medical, Rockland, MA, USA) were used. Embolizations were considered complete after confirming the resolution of venous return to the gastric varices (Fig. 1). Patients were observed in the intensive care unit or general inpatient ward to manage the post-embolization syndrome and discharged after follow-up endoscopy with no evidence of infection or other adverse events.

Figure 1. Splenic arteriograms of partial parenchymal embolization in a patient with splenic vein occlusion secondary to necrotizing pancreatitis leading to left-sided portal hypertension related hemorrhage. (A) Arterial phase of splenic angiogram. (B) Venous phase of splenic angiogram with occluded splenic vein (dashed arrows), outflow via a gastric varix (white arrows), and gastroepiploic vein (black arrows). (C) After upper partial parenchymal embolization of the spleen. There is no outflow via the gastric varix.

Measured outcomes and clinical follow-up

The outcome measures of this study included the technical success of the procedures, the clinical success of bleeding control, the adverse events, and the recurrence of variceal bleeding rate. All patients were followed after discharge in the gastroenterology clinic with endoscopic follow-up of gastric varices performed 1–3 months after discharge and regular follow-up endoscopy thereafter. Follow-up CT/MRI was performed when deemed clinically appropriate by the primary physician. Technical success was defined as immediate occlusion of the target vessel as demonstrated by completion angiography. The clinical success of bleeding control was assessed by the cessation of gastric variceal bleeding in the first 30 days after embolization. Rebleeding was identified by clinical signs of bleeding, such as hematemesis or melena, accompanied by laboratory hemoglobin reduction requiring blood product transfusion. Recurrence of variceal bleeding was defined as rebleeding greater than 30 days after embolization confirmed by endoscopy. Strategies for managing rebleeding in patients who underwent prior splenic artery embolization were reviewed.

Etiologies of splenic vein thrombosis were pancreatitis (n = 9), chronic liver disease (n = 6), and hematologic abnormalities (n = 4). Splenic vein thrombosis or chronic occlusion was identified in all patients. Preprocedural endoscopy demonstrated gastric varices with recent bleeding stigmata (n = 17) or active bleeding (n = 2).

Splenic artery embolization was successfully performed in all patients without immediate adverse events. Embolization level and material were as follows; Proximal splenic artery embolization with coils or vascular plugs (n = 9), partial distal parenchymal embolization with particles (n = 7), and total parenchymal embolization with particles (n = 3). One patient developed a left pleural effusion that required thoracentesis, and one patient had ileus with leukocytosis and fever, requiring prolonged hospitalization (13 days after embolization). The remaining 17 patients were discharged without procedure-related adverse events. Common post-embolization syndrome, such as fever and abdominal pain, was managed conservatively. The median hospital stay after embolization was 3 days (range, 1–13 days).

The median follow-up period was 55 months (range, 7–165 months). Follow-up imaging was performed in 15 of 19 patients (78.9%, CT [n = 14] and MRI [n = 1]), and all patients showed residual spleen even after total parenchymal embolization (Fig. 2).

Figure 2. Computed tomography images of total splenic artery embolization. (A) One week after embolization, more than 80% of splenic infarction is noted. (B) Five years later image showing regressed infarcted splenic parenchyma and slightly increased size of viable spleen.

All patients had no early rebleeding within 30 days after embolization. Clinical follow-up was performed in all patients. Two of 19 patients (10.5%) had recurrent variceal bleeding. One patient who received partial splenic parenchymal embolization showed recurrent variceal bleeding at 29 months after primary embolization and was treated with repeated partial splenic artery embolization with particles. The other patient who had received proximal splenic artery embolization with coils had recurrent variceal bleeding 111 months after the initial embolization. In this patient, transhepatic variceal embolization was performed. The remaining 17 patients had no further bleeding during clinical follow-up.

In this retrospective study, splenic artery embolization for gastric variceal bleeding secondary to LPH achieved an 89% success rate. Only two patients required additional treatment for post-embolization sequela, namely, a left pleural effusion requiring thoracentesis (n = 1) and ileus with leukocytosis requiring prolonged hospitalization (n = 1). During the follow-up period, two (10.5%) delayed recurrent variceal bleeding events occurred (at 2 years and 9 years post-embolization), and each was successfully treated with percutaneous embolization.

Splenic vein occlusion caused by splenic vein thrombosis or compression from surrounding structures has been recognized as the pathophysiology of LPH.^3,7,8 Gastric variceal bleeding is one of the most serious adverse events of LPH.^1,9–11 Splenectomy has been proposed as a treatment for LPH because splenectomy completely blocks the venous outflow from the spleen.¹² However, since the most common cause of LPH is an acute or chronic inflammation of the pancreas, there are concerns about the surgical difficulty of splenectomy and prolonged recovery after surgery.^12,13 Splenic artery embolization has emerged as an alternative treatment option that offers several advantages over surgery, such as the ability to be safely performed in an emergency setting with lower morbidity and shorter recovery time.¹³ Moreover, it has been shown that splenic immune function is preserved in patients undergoing total or partial splenic artery embolization.¹⁴

There have been sparse case reports and clinical studies on the appropriate method of splenic artery embolization for LPH. Still there are controversies regarding optimal embolization level and extent due to the limited number of cases and retrospective studies on the topic. ^9–11,13,15

Wang et al¹⁵ reviewed 14 patients who had bleeding with LPH. They suggested a staged two-step embolization technique which included distal splenic parenchymal embolization and proximal splenic artery coil embolization in the same procedure for acute hemorrhage. After the patient became clinically stable, interval staged distal and proximal embolization was performed. Although they did not provide follow-up periods, the proposed two-step approach showed no recurrent bleeding.

In the present study, three embolization techniques for splenic artery embolization were used.

Proximal splenic artery embolization without parenchymal embolization

There is a lack of evidence supporting proximal splenic artery embolization in the setting of LPH. Previous studies of traumatic splenic injury have shown different results regarding the efficacy of preventing rebleeding using this technique.^16–18 However, LPH is a unique pathophysiology distinct from blunt trauma, with LPH-related bleeding resulting from increased localized (left-sided) venous hypertension. That said, proximal splenic artery embolization can decrease the overall splenic arterial perfusion, which can, in turn, reduce left-sided venous hypertension.

Partial splenic parenchymal embolization

Partial parenchymal embolization with particles is another technique for LPH-related bleeding.^19,20 Particles are usually directed toward the upper pole of the spleen, where the venous outflow is directed towards gastric varices in patients with LPH. A microcatheter is used to select upper pole splenic artery branches, which are then embolized. Following embolization, delayed angiography is performed to evaluate for persistent drainage via the gastric varices. If residual venous drainage via the varices is identified, additional embolization is performed.

Total splenic parenchymal embolization

Particle embolization of the entire splenic parenchyma was performed in 3 patients. Although this technique often results in severe post-embolization syndrome such as abdominal pain, fever, or pleural effusion, in our limited sample, this technique could be the most effective at preventing rebleeding from gastric varices compared with the two aforementioned techniques because total parenchymal embolization could block entire venous return from splenic parenchyma. In our study, all patients had the post-embolization syndrome, and one showed left pleural effusion requiring thoracentesis. The pain was manageable by intravenous or oral pain medicines.

Rebleeding rates after splenic artery embolization in LPH are not clear.¹ In the present study, there were two delayed rebleeding after splenic artery embolization. One patient showed rebleeding after 111 months after proximal splenic artery embolization with coils. The treatment of rebleeding after proximal splenic artery embolization is technically challenging. As in our case, occlusion of the main splenic artery may render transarterial embolization infeasible. Therefore, other percutaneous methods, such as transhepatic and transportal variceal embolization may be required. The other patient received partial parenchymal splenic embolization, and rebleeding occurred 29 months after initial treatment. Rebleeding was successfully controlled by additional partial splenic artery embolization with particles.

This study has several limitations. First, while relatively large for an analysis of its type, the total sample size remains small, and the generalizability of our techniques and outcomes may be difficult. Gastric variceal bleeding from LPH is a rare condition, and the most extensive case series reported included only 14 patients.¹⁵ Further, there have been no reports on the long-term efficacy of splenic artery embolization and recurrence rates following embolization. Although clinical follow-up was performed in all patients, cross-sectional imaging with CT or MRI was not performed in all patients. To identify rebleeding rates and establish optimal embolization strategies, further large-scale prospective studies would be required. However, a prospective randomized trial of various techniques for LPH-related variceal hemorrhage is essentially impractical given the relatively low incidence of the disease. Therefore, retrospective case series though inherently limited, will likely be used to guide treatment decisions for this rare but potentially fatal disease.

In conclusion, splenic artery embolization secondary to splenic vein occlusion showed a 100% short-term bleeding control rate. Two of 19 (10.5%) patients experienced delayed rebleeding, and they were successfully treated by percutaneous embolization. Further studies should be conducted on the safety and long-term efficacy of each embolization technique.

None.

No potential conflict of interest relevant to this article was reported.