Glue Embolus-Induced Portosplenic Thrombosis in Transjugular Intrahepatic Portosystemic Shunt: A Novel Finding During Liver Transplantation

Introduction

Portal and splenomesenteric vein thrombosis complicates 10% to 25% of liver transplant candidates and lengthens operative time, increases transfusion needs, and worsens early graft survival [1–3]. When occlusion extends through a thrombosed transjugular intrahepatic portosystemic shunt (TIPS) into the inferior vena cava (IVC) – Yerdel grade III or IV – the risk of biliary complications and mortality rises sharply, yet such anatomy is encountered with growing frequency as transplant programs accept sicker patients [4].

These operations simultaneously expose a distinct coagulation paradox. Patients arrive anticoagulated to limit further thrombus propagation, yet they undergo one of the highest blood-loss surgical procedures [5–7]. Reversal of vitamin K antagonists with 4-factor prothrombin complex concentrate (4F-PCC) offers near-instant correction without the fluid load of plasma, an advantage in cirrhotic physiology, but safety data in the presence of active splanchnic thrombosis remain limited [8–10]. Conversely, while not always required if complete thrombectomy is achieved, aggressive intra- and postoperative anticoagulation can be essential to maintain portal inflow, particularly when the clot burden is significant or its removal is technically challenging [11]. How these emerging tools can be combined with flexible, imaging-guided anticoagulation therapy after reperfusion has received little attention.

Finally, endoscopic or endovascular treatment of gastric varices can involve liquid embolic agents (“glues”) that can cast within the left-gastric/portal venous system. When present, this nonthrombotic material can perpetuate segmental obstruction and is less likely to respond to anticoagulation, instead requiring mechanical removal or segmental exclusion. The histologic appearance of such material within portal thrombus has not been well described; our case provides illustrative pathology.

This report describes the case of a 62-year-old woman with liver cirrhosis and portal systemic thrombosis with TIPS prior to OLT and with portal vein glue embolization in the explanted liver. The case offers practical insight into integrating mechanical debulking and low-volume coagulation factor replacement for one of the most challenging scenarios in transplant surgery. Written informed consent was obtained from the patient for publication.

Case Report

A 62-year-old White woman with end-stage liver disease from primary biliary cholangitis (Model of End-Stage Liver Disease score of 31, body mass index of 18.5 kg/m2) was listed for orthotopic liver transplantation (OLT). She had no history of tobacco use, consumed alcohol only at social events, and denied recreational drug use.

She was diagnosed with cirrhosis 2 years prior to OLT, complicated by gastric varices. She subsequently underwent glue embolization of varices and TIPS placement 1 year prior to OLT. Four months prior to transplant, she was admitted to the hospital with a concern for sepsis. Blood cultures were positive for Escherichia coli. As part of the work up, she was evaluated for TIPS as a source of potential infection. CTA revealed stent thrombosis extending to the cavo-atrial junction. A portal vein thrombus was also seen, extending into the splenic vein (Figure 1). As treatment, she underwent TIPS thrombectomy during the same hospitalization. CTA after thrombectomy revealed a patent TIPS stent with a small amount of residual thrombus within the stent near the junction with the inferior vena cavae. She was discharged from the hospital on warfarin 2.5 mg once daily with a target international normalized ration (INR) of 2 to 3 (normal range 0.8 to 1.1) and 3 to 6 weeks of antibiotic therapy

She was then admitted to our institution for liver transplantation. Admission lab work was significant with an INR of 3.5 and hemoglobin 8.9 g/dL (normal value 12 to 15.5 g/dL).

After induction of general anesthesia, an arterial and central venous catheter was placed. A transesophageal echocardiography (TEE) probe was inserted for hemodynamic monitoring. After an extensive discussion between the surgical and anesthetic teams, it was determined that the treatment could be completed without the need for veno-venous bypass, and initial coagulopathy would be corrected with a single dose of 4-factor prothrombin complex concentrate (25 IU/kg, 1011 IU total).

During the dissection phase, the TIPS was removed from the middle hepatic vein. After clamping the portal vein, its lumen was assessed. There was chronic partial thrombosis extending into the superior mesenteric vein and splenic vein. These clots were surgically removed, and partial flow was confirmed with Doppler assessment. There was a small remnant clot burden in the splenic vein, which was left in place. Portal vein graft anastomosis was completed in the standard end-to-end fashion and excellent flow was verified. Then, the liver was reperfused, with minimal reperfusion syndrome, which was medically corrected.

From an anesthesia standpoint, the challenge was adequate resuscitation while keeping in mind the tendency to clot. Viscoelastic testing guided fresh-frozen plasma and PCC resuscitation. TEE imaging guided real-time assessment of heart function and major vasculature for presence of clots.

She required 4 units of packed red blood cells, 4 units of fresh-frozen plasma, and 450 mL of cell-salvaged blood. Immediate postreperfusion transesophageal echocardiography demonstrated normal right-heart size and function with tricuspid annular plane systolic excursion of 1.88 cm and no echocardiographic evidence of pulmonary embolism (Figure 2). The patient was extubated in the operating suite and postoperatively started on unfractionated heparin (UFH) at 12 IU/kg−1/h−1, titrated to an anti-Xa range of 0.3 to 0.7 IU/mL−1.

Ultrasound on postoperative day (POD) 0 showed echogenic material in the main portal vein, and CT confirmed worsening thrombus extending toward the main portal vein confluences. On POD 1 she underwent re-exploration and further surgical thrombectomy. A third laparotomy on POD 3 evacuated a 12-cm perihepatic hematoma. Anticoagulation was paused until POD 11, when UFH was restarted at a flat rate of 500 IU/h−1.

Upper gastrointestinal hemorrhage from gastric varices occurred on POD 21 and POD 31. Management included octreotide infusion, endoscopic band ligation, and proximal splenic artery embolization. Unfractionated heparin was withheld for 72 h during each bleeding episode, then resumed at low intensity and gradually escalated.

Triphasic CT on POD 43 showed the portal vein thrombus had regressed to a 1.4×0.2 cm eccentric rind; the superior mesenteric vein was patent (Figure 3). Histopathology of the explant specimen labeled “portal vein clot” showed serpiginous, empty (optically clear) channels coursing through organized thrombus (Figure 4). The morphology was consistent with dissolved polymerized embolic material introduced during prior variceal/venous glue therapy, supporting a diagnosis of glue embolus incorporated within a chronic portal thrombus. Unfractionated heparin was replaced with apixaban 5 mg twice daily. Liver biochemistry normalized, creatinine remained 0.9–1.0 mg/dL−1 (normal range 0.7 to 1.3 mg/dL), and the patient was discharged to inpatient rehabilitation on POD 47. Duplex ultrasound at discharge confirmed brisk portal flow. A follow-up CT angiogram 6 months after transplant showed a patent portal vein along with expected postsurgical changes from liver transplantation.

Discussion

Portal systemic thrombosis once excluded candidates from OLT, yet accumulating evidence shows that acceptable outcomes can still be achieved provided portal flow is restored during the operation [12–14]. One study reported that 5-year survival fell significantly from Yerdel grade I thrombosis to grade IV, largely because higher grades carry a greater risk of early graft dysfunction [4,15]. Treatment is individualized and can include management with thrombectomy, surgical debulking, and postoperative anticoagulation.

TIPS is an established percutaneous method for the management of portal hypertension. As our case demonstrates, it is not without risk. Major TIPS-related complications occur in 3% to 5% of cases necessitating intervention [16]. Portal vein puncture can be encountered during stent placement. Hepatic encephalopathy is a physiologic consequence of TIPS placement. Hypercoagulability seen in end-stage liver disease patients contributes to the development of stent thrombosis. While these risks are present, the procedure overall has a high success rate and helps decrease portal venous pressure [16].

Previous studies have looked at the risk of thrombosis in cirrhotic patients undergoing TIPS procedures. Partial venous thrombosis is common in cirrhotic patients listed for liver transplantation, but the presence of a thrombus in TIPS extending into the portosystemic venous system is more unique. Our literature search found no specific case report documenting glue embolus material as a source of portosystemic thrombus during OLT.

Our case of a fully occluded TIPS and chronic spleno-portal thrombus illustrates management principles that may help improve outcomes. Successful transplantation requires coordinated, time-sensitive use of mechanical thrombectomy, targeted warfarin reversal with prothrombin complex concentrate, point-of-care imaging, and viscoelastic testing to guide correction of coagulopathy (Table 1), following multidisciplinary discussion between transplant surgery, anesthesia, interventional radiology, and hepatology.

Suction thrombectomy, performed 3 months before transplantation, reestablished stent patency and reduced the clot burden. Early-series large-bore aspiration catheters in the splanchnic circulation demonstrate shorter procedure times and fewer complications than traditional rheolytic devices [17,18]. Our experience supports using aggressive mechanical debulking to simplify portal reconstruction and avoid complex vascular reconstruction with jump grafts at the time of transplant.

Bleeding risk at incision was mitigated with a single weight-based dose of 4F-PCC. Randomized data are lacking, but observational reports show 4F-PCC reduces exposure to plasma and is not linked to excess hepatic artery or portal vein thrombosis when used with real-time imaging surveillance and viscoelastic testing to guide therapy [19–23]. In the present case, no new macrovascular clot developed, and anticoagulation could be restarted within hours of reperfusion, supporting 4F-PCC as a practical alternative to large-volume fresh-frozen plasma administration.

Balancing re-thrombosis against bleeding required repeated adjustment of UFH. We used a stepwise approach: full-dose infusion, pause for major hemorrhage, resume at a fixed low rate once hemoglobin stabilized, then escalate to therapeutic anti-Xa targets as Doppler confirmed flow. Similar flexible algorithms have been advocated in recent guideline updates, which emphasize bedside ultrasound and thromboelastography over isolated laboratory thresholds [24–27].

A unique aspect of this case was the intraoperative discovery and subsequent pathological verification of embolic glue material within the organized portal vein thrombus (Figure 4). Cyanoacrylate-based glues are widely used for the endoscopic obliteration of gastric varices [28,29]. As the embolization procedure was performed at another facility, the exact nature of the embolization material was unable to be verified. While systemic embolization is a known, albeit rare, complication, the histological appearance of this material within an organized splanchnic thrombus removed during transplant has not been well described [30]. In our patient’s specimen, the material appeared as amorphous, serpentine, optically clear spaces, likely because of dissolution of the polymerized embolic agent during standard tissue processing, leaving a characteristic negative imprint within organized thrombus [31]. This pattern, in the appropriate clinical context, may serve as a pathologic signature of prior glue embolization. Moreover, these findings show that extensive thrombosis in patients with a history of glue embolization may be composed not just of a blood clot but can also consist of a complex amalgam of thrombus and foreign material, potentially making mechanical and surgical thrombectomy more challenging. It also explains the persistent occlusion of the left-gastric venous system despite anticoagulation and the technical difficulty in achieving complete luminal clearance by thrombectomy alone. Importantly, nonthrombotic cast material will not regress with anticoagulation, requiring mechanical debulking and segmental exclusion rather than pharmacologic management [32,33].

Preoperative recognition is challenging. CT may depict this as high-attenuation/heterogeneous intraluminal material, but differentiation from organized thrombus is imperfect [34]. When a history of variceal glue therapy exists, surgeons and anesthesiologists should anticipate mixed lesions (glue+thrombus), plan for meticulous thrombectomy, and avoid overinterpreting early postoperative “residual filling defects” as purely thrombotic. To the best of our knowledge, the histologic appearance of such glue within portal venous thrombus during OLT has not been previously reported.

Limitations of this report are that it was a single-patient observation, and causality between each intervention and the favorable outcome could not be proved. Anti-Xa levels were not recorded consistently, and follow-up was limited to 6 weeks. Long-term patency of the portal reconstruction and the durability of direct oral anticoagulant therapy in this setting remain unknown.

Future research directions should focus on prospective registries that capture device type, clot morphology, and graft outcomes after splanchnic suction thrombectomy. Further investigation into the incidence and management of retained embolic material in chronic portal vein thrombosis is warranted. Such registries should also capture prior embolic agent exposure and correlate radiologic, operative, and pathologic appearance with outcomes. Randomized trials comparing PCC with viscoelastic-guided plasma transfusion during liver transplantation are also needed. Finally, the safety and optimal dosing of direct oral anticoagulants for long-term prophylaxis after complex portal reconstructions warrant systematic study.

Conclusions

Portosystemic clot burden, while not an uncommon finding in end-stage liver disease, can pose increased complexity during liver transplantation. When a thrombus encapsulates the TIPS, it warrants close cooperation between surgical and anesthesia teams to ensure a safe surgical course. Cirrhotic patients are predisposed to bleeding and simultaneously thrombosis because of their underlying pathology, which adds to the challenge of safe management in an already complex procedure.

In this case report we demonstrate the successful management of portosystemic TIPS-associated thrombus. Explant histopathology of specimens revealed prior variceal glue embolization material as a source of the thrombus, which is a novel finding. Intraoperative mechanical debulking and postoperative anticoagulation led to a favorable outcome. Close multidisciplinary cooperation between the anesthesia and surgical teams can ensure a successful transplant in such complex cases.