Delayed catastrophic haemorrhage after transjugular portal venous intervention

Intended learning outcomes

1. Recognise delayed haemorrhagic complications following transjugular portal venous interventions.

2. Understand the principles of anticoagulation reversal in life-threatening bleeding, particularly with unfractionated heparin.

3. Apply viscoelastic haemostatic assays (rotational thromboelastometry/thromboelastography) to guide transfusion during massive haemorrhage.

4. Understand the role of multidisciplinary perioperative teams in recognising and managing catastrophic haemorrhagic complications.

Introduction

Portal vein thrombosis (PVT) is an uncommon vascular disorder characterised by partial or complete obstruction of the portal venous system [1, 2]. It most frequently occurs in association with cirrhosis, hepatobiliary malignancy, abdominal inflammatory conditions or systemic prothrombotic disorders [3]. In contrast, acute non-cirrhotic, non-malignant portal vein thrombosis is a distinct entity occurring in otherwise preserved hepatic parenchyma [1, 4]. Although relatively rare, its reported incidence ranges between approximately 0.7 and 3.7 cases per 100,000 individuals [1]. The clinical presentation is often non-specific, ranging from mild abdominal discomfort to acute abdominal pain associated with nausea, vomiting or features of mesenteric venous congestion. If untreated or progressive, PVT may lead to serious complications including bowel ischaemia, portal hypertension, cavernous transformation of the portal vein and systemic sepsis [2].

The pathogenesis of non-cirrhotic PVT is frequently multifactorial [1, 4]. Identifiable risk factors include inherited thrombophilia, myeloproliferative disorders, local inflammatory processes within the abdomen and transient prothrombotic states [4, 5]. Consequently, evaluation often includes screening for inherited and acquired thrombophilia, such as protein C and protein S deficiency, antithrombin deficiency, factor V Leiden mutation and antiphospholipid antibodies [4]. Nevertheless, even after extensive diagnostic evaluation, up to 40% of cases remain idiopathic [6].

Early anticoagulation remains the cornerstone of treatment and has been shown to improve portal vein recanalisation rates whilst reducing thrombus propagation and associated complications [1]. The European Association for the Study of the Liver (EASL) clinical practice guideline on vascular diseases of the liver [7] generally recommends the initiation of therapeutic anticoagulation in acute non-cirrhotic portal vein thrombosis, with low-molecular-weight heparin (LMWH) suggested as first-line therapy because of its predictable pharmacokinetics and ease of administration [4].

In selected patients with extensive thrombosis, clinical deterioration, or imaging features suggestive of compromised intestinal perfusion, escalation to portal venous interventions, including catheter-directed thrombolysis or transjugular portal venous recanalisation, may be considered [8]. Whilst these procedures may restore portal venous flow and potentially prevent intestinal infarction, they introduce additional procedural and haemostatic risks [9, 10].

In this context, delayed haemorrhagic complications remain poorly characterised. We describe a patient with extensive acute non-cirrhotic PVT who underwent transjugular portal venous intervention after deterioration despite therapeutic anticoagulation and subsequently developed catastrophic delayed haemoperitoneum from hepatic parenchymal rupture.

Report

A 56-year-old man presented with a five-day history of progressively worsening diffuse abdominal pain radiating to the back, associated with vomiting and obstipation. His medical history included long-standing hypertension, type 2 diabetes mellitus, metabolic dysfunction-associated steatotic liver disease (MASLD), chronic alcohol consumption and prior tobacco use. There was no history of cirrhosis, malignancy or known thrombophilia.

On presentation, the patient was haemodynamically stable with left hypochondrial tenderness and no peritoneal signs. Ultrasonography revealed splenomegaly with grade I hepatic steatosis. Contrast-enhanced computed tomography demonstrated partial thrombosis of the main portal vein with extension into both branches, complete thrombosis of the splenic and superior mesenteric veins, and long-segment jejunal wall thickening suggestive of evolving bowel ischaemia.

Therapeutic anticoagulation was initiated with intravenous unfractionated heparin, titrated to maintain an activated partial thromboplastin time between 40 and 60 seconds. Despite escalation of infusion rates, radiological progression and concern for mesenteric ischaemia prompted transfer for transjugular portal venous intervention on hospital day 2. The patient was electively intubated for airway protection and haemodynamic control.

Under fluoroscopic guidance, transjugular intrahepatic portal venous access was achieved and catheter-directed thrombolysis with a 5 mg alteplase bolus was performed, resulting in partial recanalisation of the superior mesenteric and main portal veins. A thrombolysis catheter was left in situ; continuous infusion was deferred because of bleeding risk. The patient was transferred to the critical care unit under mechanical ventilation, and unfractionated heparin was resumed.

On hospital day 5, the patient developed abrupt haemodynamic deterioration with severe hypotension, tachycardia and rapidly progressive abdominal distension. Invasive arterial pressure monitoring and vasopressor support were initiated, and bedside ultrasonography demonstrated significant intraperitoneal free fluid.

Despite transient stabilisation, hypotension recurred. A massive transfusion protocol was activated and guided by rotational thromboelastometry, which demonstrated impaired clot strength. Packed red blood cells, fresh-frozen plasma and platelets were transfused, along with intravenous protamine sulphate for anticoagulation reversal. Haemoglobin decreased from 13 g/dL to 6.4 g/dL within hours (Tables 1 and 2).

Urgent imaging confirmed a large right hepatic intraparenchymal haematoma with capsular rupture and massive haemoperitoneum. Emergency damage-control laparotomy was performed, evacuating intraperitoneal blood and clots, followed by perihepatic packing. No macroscopic bowel ischaemia was identified. Postoperatively, the patient required continued vasopressor support and additional viscoelastic-guided transfusion.

Haemodynamic stability was gradually restored, vasopressors were weaned, and definitive abdominal closure was performed on postoperative day 4. The patient was extubated on postoperative day 5 and transferred to step-down care with progressive recovery.

Question to be answered

What mechanisms may contribute to delayed catastrophic haemorrhage following transjugular portal venous interventions for acute PVT? What perioperative strategies are essential for early recognition and effective multidisciplinary management?

Discussion

This case illustrates an uncommon, but catastrophic, delayed haemorrhagic complication following portal venous intervention. Hepatic parenchymal rupture occurring several days after the procedure suggests a multifactorial mechanism rather than an immediate procedural injury. Several pathophysiological factors may have contributed. First, transhepatic portal venous access inevitably traverses hepatic parenchyma and may create a tract susceptible to delayed bleeding, particularly when anticoagulation is resumed [10, 11]. Second, partial recanalisation of previously thrombosed portal and mesenteric veins can abruptly alter intrahepatic haemodynamics, potentially increasing sinusoidal pressures or redistributing portal flow within fragile hepatic tissue [9, 10, 11]. Third, MASLD present in this patient may have increased hepatic vulnerability to parenchymal injury and capsular rupture. Finally, the resumption of therapeutic unfractionated heparin following vascular manipulation may have amplified bleeding from a previously contained parenchymal injury.

Whilst LMWH is generally recommended as first-line anticoagulation for acute non-cirrhotic PVT, unfractionated heparin was selected in this patient because of the extensive portal–mesenteric venous involvement, suspected mesenteric ischaemia and the potential need for urgent procedural intervention. Its short half-life and reversibility with protamine sulphate allow rapid titration and, as this case demonstrates, facilitate prompt reversal when catastrophic bleeding occurs.

Escalation beyond anticoagulation is sometimes required. Interventional strategies, such as transjugular intrahepatic portal venous access with catheter-directed thrombolysis or mechanical thrombectomy, may be considered in patients with extensive thrombosis, worsening symptoms or imaging features for intestinal ischaemia [8]. Although these procedures can restore portal venous patency and potentially avert bowel infarction, they introduce procedural risks related to transhepatic vascular access, thrombolytic therapy and anticoagulation management [8, 9]. In particular, bleeding complications may arise from hepatic parenchymal injury, capsular disruption or coagulopathy associated with thrombolytic exposure [9].

This case highlights the possibility of delayed haemorrhagic deterioration, even when the immediate post-procedural period appears stable. Portal venous interventions are frequently followed by intensive care monitoring, yet attention often focuses on thrombotic progression or bowel ischaemia rather than delayed intra-abdominal bleeding. Sudden haemodynamic collapse in such patients should therefore prompt rapid evaluation for haemorrhage, including bedside ultrasonography to identify free intraperitoneal fluid. In the present case, the rapid recognition of haemoperitoneum facilitated early activation of a massive transfusion protocol and expedited surgical intervention.

Major bleeding in anticoagulated patients is a common perioperative challenge. When life-threatening haemorrhage occurs, prompt reversal of anticoagulation must be balanced against the thrombotic risk that prompted treatment. In emergency scenarios, protamine administration is typically guided by the estimated amount of circulating heparin and clinical bleeding severity. However, anticoagulation reversal alone is rarely sufficient in the presence of massive haemorrhage, in which dilutional coagulopathy, hypothermia and platelet dysfunction frequently coexist.

For this reason, modern damage-control resuscitation strategies emphasise early haemostatic support using balanced transfusion of red cells, plasma and platelets, supported by evidence from large trauma trials and international bleeding management guidelines [12, 13]. Viscoelastic haemostatic assays such as rotational thromboelastometry (ROTEM) or thromboelastography (TEG) provide real-time functional assessment of clot formation and enable targeted haemostatic therapy, an approach increasingly recommended in major bleeding guidelines and perioperative transfusion protocols [12, 13]. These techniques are now widely integrated into perioperative bleeding algorithms, particularly in cardiac surgery, trauma and liver transplantation. In this case, ROTEM-guided transfusion identified impaired clot strength and directed blood product administration during resuscitation. Such approaches may reduce unnecessary transfusion, whilst improving haemostatic control.

Coordinated multidisciplinary crisis management was central to the successful outcome in this patient’s catastrophic haemorrhage. Following sudden haemodynamic deterioration, rapid bedside assessment enabled the early recognition of haemoperitoneum, prompting immediate activation of the institutional massive transfusion protocol and urgent surgical consultation. Close collaboration between critical care clinicians, anaesthetists, surgeons, interventional radiologists, nursing staff and transfusion services allowed simultaneous haemodynamic resuscitation, reversal of anticoagulation, ROTEM-guided haemostatic therapy and expedited operative source control. This coordinated response enabled effective damage-control management of a massive haemoperitoneum in a time-critical setting.

As a single-patient observation, the findings cannot establish causality between portal venous intervention and hepatic rupture; other contributing factors, such as occult vascular injury or spontaneous hepatic bleeding, cannot be definitively excluded. Histopathological examination of hepatic tissue was not performed, and limits the understanding of potential underlying parenchymal fragility. Additionally, delayed haemorrhagic complications following portal venous interventions remain poorly characterised in the literature, highlighting the need for a larger clinical series to better define incidence, mechanisms and optimal monitoring strategies.

However, this case underscores a core principle of perioperative care: therapeutic interventions that restore vascular patency may also introduce new haemodynamic and haemostatic risks that manifest beyond the immediate procedural period. For patients undergoing portal venous interventions, sustained postoperative vigilance, early recognition of haemorrhagic deterioration, rapid reversal of anticoagulation, and viscoelastic-guided haemostatic resuscitation are critical components of effective perioperative management. A coordinated multidisciplinary response remains central to the successful management of these high-risk scenarios.

Declarations and conflicts of interest

Artificial intelligence declaration

Artificial intelligence tools were used solely for language polishing and formatting. All clinical content, interpretation, references and final revisions were undertaken and verified independently by the authors.