Portobiliary fistula after liver biopsy in a young transplant patient presenting with GI bleeding and hemorrhagic shock managed by embolization: a case report

Abstract

Introduction:

Gastrointestinal (GI) bleeding in liver transplant recipients is diagnostically and therapeutically challenging, particularly when routine evaluations fail to identify a source. Rare vascular causes must be considered when standard workups are inconclusive.

Presentation of case:

A 23-year-old male with a complex history of four liver transplants over several years for chronic rejection related to medication noncompliance experienced recurrent episodes of obscure GI bleeding over months. Extensive evaluation – including multiple upper and lower endoscopies, nuclear tagged-red blood cell scans, and CT angiography – was repeatedly negative. Ultimately, provocative mesenteric angiography revealed a portobiliary fistula, felt to be iatrogenic following a percutaneous liver biopsy. The lesion was successfully treated with coil and gelfoam embolization, with complete resolution of bleeding and subsequent normalization of liver function tests.

Discussion:

This case highlights an uncommon vascular etiology of obscure GI bleeding in the transplant population. Portobiliary fistula/hemobilia should be considered in patients with recurrent, unexplained GI hemorrhage after liver biopsy or invasive hepatobiliary procedures. When intermittent bleeding eludes conventional imaging, provocative angiography can be diagnostic by unmasking occult, intermittent extravasation. Timely endovascular therapy can achieve durable hemostasis while preserving graft function.

Conclusion:

Clinicians should maintain a high index of suspicion for portobiliary fistula in post-biopsy transplant patients presenting with hemobilia or unexplained GI bleeding. Provocative angiography is a valuable tool for identifying occult bleeding sources, and endovascular embolization offers an effective, organ-sparing treatment. The case also underscores the critical need for strict adherence to immunosuppressive therapy to prevent graft loss and repeated transplantation.

Introduction

Gastrointestinal (GI) bleeding can originate anywhere from the esophagus to the rectum and is generally classified as upper, proximal to the ligament of Treitz, or lower, distal to ligament of Treitz, GI bleeding. Upper GI bleeding is far more common and serious, with an incidence of 80–150 cases per 100 000 population. However, lower GI bleeding occurs in about 20%–30% of acute GI bleeds^[1^]. In solid-organ transplant recipients, including liver transplant (LT) patients, GI bleeding was reported in roughly 4%–9% of orthotopic liver transplant patients, where they experience significant GI hemorrhage postoperatively. The causes in this population mirror general causes such as ulcers, gastritis, and varices, but can also include transplant-specific issues such as anastomotic ulcers or leaks, portal hypertensive lesions from pre-existing disease, or opportunistic infections like cytomegalovirus (CMV) colitis^[2,4^].

Liver biopsy is routinely used in transplant management to diagnose rejection, recurrent disease, or graft injury. Techniques include percutaneous ultrasound-guided and trans-jugular biopsy. It has a good safety profile, as in large series report major complication rates of only ~0.5%–2% and mortality around 0.01%–0.2%^[5,6^]. Hemorrhage is the most feared complication, with pain being the most common and considered minor complications in about 9%–13% of cases and usually self-limited^[6^]. However, iatrogenic creation of abnormal communications, fistulas, is very rare. For example, bile leaks occur in <0.001% of biopsies^[7^]. Hemobilia, which is bleeding into biliary tree, from biopsy is also rare; one series found a post-biopsy hemobilia incidence of 0.05%^[8^]. Even rarer are portobiliary fistulae: isolated case reports describe arteriobiliary or portobiliary fistulas after interventions such as endoscopic retrograde cholangiopancreatography (ERCP) or transjugular intrahepatic portosystemic shunt^[8,9^], but none previously after liver biopsy in a transplant recipient.

Mechanistically, the condition results from traumatic communication between a biliary radical and an adjacent portal venous branch, a tract that may persist in the presence of local ischemia, infection, prior biliary manipulation, or tissue fragility^[10,11^]. In transplant recipients, additional contributors such as chronic rejection, ischemic cholangiopathy, repeated biliary interventions, and altered hilar anatomy may further increase susceptibility to biopsy-related portobiliary communication, emphasizing the need for heightened procedural caution in this population^[11,12^].

These facts underscore the novelty of our case: an intrahepatic portal-to-bile duct fistula after biopsy, manifesting as life-threatening GI bleeding.

Case presentation

A 23-year-old male with end-stage liver disease, initially of unknown etiology, underwent four sequential LTs over several years. His first transplant occurred at age 15 – a living-donor segmental graft from his father for acute liver failure – but that graft eventually failed from chronic rejection despite therapy. He received a second orthotopic transplant several years later but admitted poor adherence to immunosuppressive medications, and that graft likewise developed rejection. A third transplant was performed a few years after that and was complicated intraoperatively by early hepatic artery thrombosis, leading to graft loss. He then underwent a fourth transplant shortly thereafter. After the fourth transplant, he was discharged on tacrolimus, prednisone, and mycophenolate with only mildly abnormal liver chemistries (Aspartate aminotransferase (AST)/Alanine transaminase (ALT) ~ 40–60 U/L, alkaline phosphatase 150–200 U/L, total bilirubin 1.5–2.0 mg/dL, and INR 1.2), indicating generally preserved graft function. A liver biopsy was performed later in the course showed no histologic evidence of rejection.

Several months after discharge and 1 week after normal outpatient labs, the patient developed fatigue with dark, tarry stools, without abdominal pain. Vital signs were stable, but hemoglobin was 8.2 g/dL (baseline ~12.0); stool guaiac testing was positive. He was admitted for acute GI bleeding. Urgent esophagogastroduodenoscopy (EGD) and flexible sigmoidoscopy within 24 hours showed no active bleeding and only minor gastritis. A technetium-tagged red blood cell (RBC) scan during prolonged observation showed no definite bleeding site, and contrast-enhanced CT angiography of the abdomen likewise revealed no contrast extravasation. He remained hemodynamically stable; intravenous proton-pump inhibitors and transfusions were given as needed. After 3 days, the bleeding stopped spontaneously (Hb stabilized at 10.5 g/dL), and he was discharged with outpatient follow-up.

Over the next several weeks, he had recurrent similar episodes. On the second presentation, he again had melena with hemoglobin falling to 7.9 g/dL. Repeat EGD and full colonoscopy were negative for an upper or lower source. Video capsule endoscopy was considered but deferred because of altered anatomy and ongoing blood loss; repeat mesenteric CT angiography and another tagged RBC scan again showed no active bleeding. Supportive care (IV fluids, Proton pump inhibitors) was provided, and the bleeding abated over 48 hours.

HIGHLIGHTS

• Portobiliary fistula identified as a rare iatrogenic cause of hemobilia and obscure GI bleeding after percutaneous liver biopsy in a transplant recipient.

• Routine investigations – including multiple endoscopies, nuclear tagged-red blood cell scans, and CT angiography – do not always reveal intermittent bleeding sources, underscoring diagnostic limitations.

• Provocative mesenteric angiography can unmask occult hemorrhage but is a high-risk procedure that should be performed only in well-resourced centers by experienced interventional teams.

• Portobiliary fistula is a rare but important potential complication of percutaneous liver biopsy and other hepatobiliary interventions.

• Prompt recognition and timely endovascular management (e.g., coil and gelfoam embolization) can achieve durable hemostasis, preserve graft function, and prevent further morbidity.

A few months later, he experienced a third significant episode and presented more unstable with fresh hematochezia and shock; systolic BP was 89 mmHg, and pulse was 133. He required transfusion of three units of packed RBCs and was re-admitted. Laboratory tests at that time showed Hgb 6.8 g/dL, platelets 120 000/µL, INR 1.3, ALT/AST 70/65 U/L, ALP 230 U/L, and TBili 3.0 mg/dL. Repeat EGD and colonoscopy again failed to identify a source. Interventional radiology performed a mesenteric angiogram of the celiac and superior mesenteric circulations, which initially showed no extravasation. Given high suspicion for intermittent hemorrhage, provocative angiography (systemic heparin and nitroglycerin infusion) was attempted; this revealed subtle contrast extravasation into the left intrahepatic biliary tree from a branch of the left portal vein, with a small fistulous tract between a portal venous branch and an adjacent bile duct. Coil embolization and deployment of gelfoam slurry across the fistula tract achieved immediate cessation of contrast flow into the bile ducts; the procedure was otherwise uneventful (Fig. 1).

Figure 1.

Fluoroscopic angiography image obtained during interventional radiology demonstrates selective catheterization of the hepatic vasculature. Contrast injection reveals pathological opacification within the liver parenchyma, consistent with an intrahepatic bilio-portal fistula. The image shows contrast extravasation and pooling into the biliary system and portal venous structures, confirming the abnormal communication responsible for recurrent occult gastrointestinal bleeding.

Over the following week the patient’s bleeding ceased permanently. Hemoglobin recovered to 11.0 g/dL without further transfusions. Follow-up labs normalized (AST/ALT ~30 U/L, ALP 120 U/L, and TBili 1.1 mg/dL) and repeat ultrasound showed no hematoma. The patient was counseled on strict compliance with his immunosuppressive regimen; at 3-month follow-up, he remained asymptomatic, with stable graft function and no recurrence of bleeding. Table 1 summarizes the events post transplantation until full recovery and follow-up.

Table 1.

Timeline of post-liver transplant events and interventions.

Timeline	Summary of the events and interventions
Day 0 (fourth transplant discharge)	Stable graft function on tacrolimus, prednisone, and mycophenolate. Baseline labs were AST/ALT 40–60 U/L, ALP 150–200 U/L, TBili 1.5–2.0 mg/dL, and INR 1.2. Liver biopsy showed no rejection.
Week 12 (episode 1)	Patient developed fatigue and melena, with hemoglobin dropping to 8.2 g/dL. Esophagogastroduodenoscopy and sigmoidoscopy revealed only minor gastritis. Tagged RBC scan and CT angiography did not localize the bleeding. The patient was transfused and given IV proton-pump inhibitors; bleeding stabilized spontaneously and hemoglobin recovered to 10.5 g/dL.
Week 16 (episode 2)	Melena recurred with hemoglobin 7.9 g/dL. Repeat EGD and colonoscopy were normal. Mesenteric CT angiography and tagged RBC scan again failed to identify a source. Supportive care was provided, and bleeding resolved without intervention.
Month 6–7 post-transplant (episode 3)	Patient came with fresh hematochezia, hypotension (BP 89 mmHg), and tachycardia. Hemoglobin fell to 6.8 g/dL. EGD and colonoscopy remained negative. Mesenteric angiography initially showed no active bleeding. Provocative angiography with heparin and nitroglycerin revealed a portobiliary fistula from a branch of the left portal vein to an intrahepatic bile duct. Coil and gelfoam embolization was done.
Week 1 post-embolization	Hemoglobin improved to 11.0 g/dL, and liver function tests normalized (AST/ALT ~30 U/L, ALP 120 U/L, and TBili 1.1 mg/dL). Ultrasound showed no hematoma.
3-month follow-up	Patient remained asymptomatic, with stable graft function and no recurrence of bleeding. Adherence to immunosuppressive therapy was reinforced.

Throughout these episodes, the grafts’ synthetic function remained relatively preserved; mild cholestatic trends (elevated ALP and bilirubin) were noted only during active bleeding and rapidly improved after fistula closure. The patient acknowledged medication nonadherence as a contributor to rejection and loss of prior grafts. No other sources of bleeding (e.g., varices, peptic ulcer disease, or a primary coagulopathy) were identified.

Discussion

GI bleeding in a patient with multiple LTs is an especially challenging scenario. In a transplanted liver, the usual landmarks for GI sources may not apply straightforwardly, for example, altered anatomy from hepatico-jejunostomy or Roux-en-Y biliary reconstruction can confound endoscopic access. In our patient, the bleeds were painless and intermittent, consistent with bleeding from the upper GI tract that eventually ended up in the intestine, to present with melena or hematochezia. The differential for recurrent obscure bleeding after transplant would include peptic ulcer, heightened by steroids, variceal hemorrhage, in cases of persistent portal hypertension, anastomotic ulcer or leak, and immunosuppression-related gastritis or infection, such as CMV colitis^[2,3^]. However, despite thorough bidirectional endoscopy and imaging workup, the bleeding source eluded detection on routine studies. This pattern of obscure GI bleeding is well recognized: when endoscopies and imaging (CT and angiography) fail to identify the source, specialized studies such as capsule endoscopy or provocative angiography are needed. Indeed, after traditional studies were negative, provocative mesenteric angiography was performed in episode 4. Provocative angiography, by giving heparin or vasodilators during catheter angiography, has been shown to localize bleeding in roughly half of occult cases^[13,14^]. In this patient, systemic heparin and nitroglycerin unmasked flow through the portobiliary fistula, visualizing contrast leaking into the biliary radicles.

The key underlying pathology in this case was an iatrogenic bilio-portal fistula. In transplant recipients, altered hilar anatomy, repeated biliary manipulation, episodes of cholangitis, and ischemic or rejection-related bile duct injury can increase susceptibility to a portobiliary communication after even minor iatrogenic trauma^[11^]. Needle traversal of closely apposed portal venules and bile ducts can establish a tract that may fail to seal when local healing is impaired by inflammation, infection, or ischemia^[10,15^]. In this context, blood may enter the biliary system during anticoagulation, transient portal pressure elevation, or direct provocation. These transplant-specific vulnerabilities may partly explain why portobiliary fistulas, although exceedingly rare, are disproportionately reported in patients with prior graft injury or biliary interventions^[11,12^].

In general, arteriobiliary fistula, causing hemobilia, is more commonly reported, and hepatic artery injury (traumatic and iatrogenic) can cause frank hemobilia leading to GI bleeding. By contrast, a portal vein–bile duct fistula is exceedingly uncommon. In principle, because portal venous pressure (~10 mmHg) is usually lower than biliary tract pressure (~10–15 mmHg), bile should flow into the portal system rather than blood into the bile duct, a situation termed “bilhemia” when bile enters the veins. However, if a true communication is created, for example, by a needle tract from a biopsy puncturing both portal and biliary channels, blood can be forced into the bile ducts under certain conditions, like during anticoagulation or increased portal pressure and then drain into the duodenum via the papilla^[5,16^]. Our patient’s behavior (noncompliance with anticoagulation or steroids aside from immunosuppression) and the angiographic provocation likely allowed detection of otherwise intermittent bleeding.

Considering the literature, several case reports have documented the occurrence of portobiliary or bilio-venous fistulas, though such cases remain extremely rare. Reported scenarios include fistula formation following endoscopic retrograde cholangiopancreatography for cirrhotic patient^[9^], post-transplant portal vein to bile duct fistulas diagnosed via ERCP and treated endoscopically^[17^], as well as cases arising in a transplanted liver post hepaticojejunostomy^[11^]. Another case involved mispositioned biliary metal stent perforating into the portal vein in a patient with pancreatic head mass^[18^]. These cases are summarized in Table 2.

Table 2.

Summary of portobiliary fistulas cases.

Authors	Demorphic and clinical data for the patients	Etiology	Management
Saldana-Ruiz et al (2021)	70-year-old male with cryptogenic cirrhosis	Post-ERCP	Fully covered self-expandable metal stent via ERCP; subsequent liver transplantation
Murray et al (2019)	60-year-old woman with hepatitis C-related cirrhosis complicated by recurrent hepatic encephalopathies and chronic portal vein thrombosis	Post-ERCP	Endoscopic deployment of fully covered self-expandable metal stent
Kumar et al (2024)	60-year-old male patient with nonalcoholic steatohepatitis (NASH) related cirrhosis	Manipulation of hepaticojejunostomy stricture	Portal vein stent graft plus coil embolization and percutaneous biliary drainage
Chatterjee et al (2023)	64-year-old man with a mixed-density pancreatic head mass causing obstructive jaundice	Mispositioned biliary metal stent perforating PV	Whipple procedure with portal vein resection, thrombectomy, and reconstruction

In transplant recipients, percutaneous biopsy is frequently performed with careful precautions (e.g., ultrasound guidance and avoiding coagulopathy). Nevertheless, any biopsy carries risk. Studies show that using ultrasound guidance and limiting needle passes keeps serious complication rates low. One large audit found an overall complication rate of ~19.8%, mostly pain or mild bleeding, with only 0.9% requiring hospitalization for a hematoma^[19^]. Another review reported major complications, such as hemorrhage requiring intervention, organ perforation, infection, and death in only 2%–4% of biopsies. Specifically, hemobilia occurs in about 0.02%–0.3% of all percutaneous biopsies, accounting for roughly 3% of major complications^[6,20^].

The delayed nature of bleeding, demonstrated by its late appearance for several weeks after biopsy, is consistent with formation of a tract that became clinically significant only under provocation. Doppler ultrasound can sometimes detect high-velocity flow in arterioportal fistulas, but small portobiliary tracts are usually only seen on angiography or dedicated cross-sectional imaging^[21^]. In this case, digital subtraction angiography during provocative testing was diagnostic. Once identified, the fistula was successfully treated by transcatheter embolization, and coils and gelfoam were deployed in the tract to occlude flow. This approach is analogous to embolization of hepatic artery pseudoaneurysms; reported success rates for controlling hemobilia by arterial embolization exceed 95%^[22^]. Endoscopic stenting has also been used in portobiliary fistulas involving biliary anastomoses, but our case was intrahepatic and best managed by interventional radiology. The prompt cessation of bleeding after embolization confirms the fistula as the culprit.

Beyond the immediate emergency, this case underscores important general principles. First, in cases of obscure GI bleeding, especially when multiple standard tests are negative, one must consider rare vascular causes and use advanced diagnostics, such as capsule endoscopy and provocative angiography. Second, any history of recent liver biopsy or intervention should raise suspicion for vascular-biliary fistula if GI bleeding and jaundice occur together, as the classic Quincke triad of hemobilia consisting of pain, jaundice, and GI bleeding. Finally, this case highlights that non-adherence to immunotherapy can lead to repeated graft loss, compounding morbidity; the patient’s multiple re-transplants, four transplants within 7 years, were ultimately due to chronic rejection secondary to medication noncompliance. This is a known risk in young transplant recipients.

This report is limited by its nature as a single case, which restricts generalizability to other transplant populations. Certain diagnostic decisions, such as deferring video capsule endoscopy due to altered anatomy, may have influenced the timing of fistula detection. Additionally, the patient’s history of multiple grafts and noncompliance complicates interpretation of causality and may not reflect typical post-biopsy outcomes. Nonetheless, the case highlights important diagnostic and management considerations for rare portobiliary fistulas.

Conclusion

In summary, we report a life-threatening GI hemorrhage in an LT patient that was eventually traced to a biopsy-induced portobiliary fistula. While GI bleeding and biopsy complications are well described separately, the combination is extraordinarily rare. We emphasize that interventional angiography was key to diagnosis and treatment. Awareness of such unusual causes is crucial in complex transplant patients with recurrent bleeding.

Methods

This work has been reported in line with the SCARE criteria^[23^].

Ethical approval

Ethical approval was not applicable for this study, as our institution’s IRB committee does not necessitate approval for publishing individual case reports or case series. Written informed consent for the publication of this case report was obtained from the patient’s guardian in compliance with SCARE guidelines. Ethical approval was not required for this case report as per the policy of the hospital.

Consent

Written informed consent was obtained from the patient’s parents/legal guardian for publication and any accompanying images.

Sources of funding

No funding was received for this study.

Author contributions

F.K. and A.G.: study concept or design and editor. M.M., O.E., N.R., and M.A.A.M.: data collection, data analysis or interpretation, and writing the paper. M.M.: editor and reviewer. O.E.: corresponding author

Conflicts of interest disclosure

The authors declare no conflicts of interest.

Research registration unique identifying number (UIN)

Not applicable.

Guarantor

Osama Ewidat.

Provenance and Peer Review

Not commissioned; externally peer reviewed.

Data availability statement

Data are available from the corresponding author upon request.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT for language and grammar polishing in order to improve readability. After using this tool, the authors reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.