Review

Portal vein thrombosis: diagnosis, management, and

endpoints for future clinical studies
Laure Elkrief, Virginia Hernandez-Gea, Marco Senzolo, Agustin Albillos, Anna Baiges, Annalisa Berzigotti, Christophe Bureau,
Sarwa Darwish Murad, Andrea De Gottardi, François Durand, Juan-Carlos Garcia-Pagan, Ton Lisman, Mattias Mandorfer, Valérie McLin,
Lucile Moga, Filipe Nery, Patrick Northup, Alexandre Nuzzo, Valérie Paradis, David Patch, Audrey Payancé, Vincent Plaforet, Aurélie Plessier,
Johanne Poisson, Lara Roberts, Riad Salem, Shiv Sarin, Akash Shukla, Christian Toso, Dhiraj Tripathi, Dominique Valla, Maxime Ronot,
Pierre-Emmanuel Rautou, on behalf of ERN RARE-LIVER and VALDIG, an EASL consortium

Portal vein thrombosis (PVT) refers to the development of a non-malignant obstruction of the portal vein, its branches, Lancet Gastroenterol Hepatol
its radicles, or a combination. This Review first provides a comprehensive overview of all aspects of PVT, namely the 2024

specifics of the portal venous system, the risk factors for PVT, the pathophysiology of portal hypertension in PVT, the Published Online
July 9, 2024
interest in non-invasive tests, as well as therapeutic approaches including the effect of treating risk factors for PVT or
https://doi.org/10.1016/
cause of cirrhosis, anticoagulation, portal vein recanalisation by interventional radiology, and prevention and S2468-1253(24)00155-9
management of variceal bleeding in patients with PVT. Specific issues are also addressed including portal Investigators listed in the
cholangiopathy, mesenteric ischaemia and intestinal necrosis, quality of life, fertility, contraception and pregnancy, appendix (p 1)
and PVT in children. This Review will then present endpoints for future clinical studies in PVT, both in patients with Faculté de médecine de Tours,
and without cirrhosis, agreed by a large panel of experts through a Delphi consensus process. These endpoints et service d’hépato-
include classification of portal vein thrombus extension, classification of PVT evolution, timing of assessment of PVT, gastroentérologie, Le Centre
Hospitalier Régional
and global endpoints for studies on PVT including clinical outcomes. These endpoints will help homogenise studies Universitaire de Tours, Tours,
on PVT and thus facilitate reporting, comparison between studies, and validation of future studies and trials on PVT. France (Prof L Elkrief); Centre de
recherche sur l’inflammation,
Introduction absence of the portal vein (also known as Abernethy Université Paris-Cité, Paris,
France (Prof L Elkrief,
The term portal vein thrombosis (PVT) refers to the malformation), to be excluded. When performed by Prof F Durand, L Moga,
development of a non-malignant obstruction in the experienced radiologists, reported accuracy for the A Payancé, A Plessier, J Poisson,
portal vein, its branches, its radicles, or a combination. diagnosis of PVT at imaging ranges from 88–98% and Prof D Valla, Prof M Ronot,
Prof P-E Rautou); Barcelona
Tumorous occlusion of the portal vein—erroneously with sensitivity and specificity values of 80–100%.3
Hepatic Hemodynamic
referred to as malignant thrombosis—is characterised by Diagnostic accuracy might be lower when the radiologist Laboratory, Liver Unit, Hospital
the intra­luminal development of malignant tissue and is not aware of the suspicion of PVT, or when the quality Clínic de Barcelona, Institut de
will not be discussed here. PVT can develop in patients of the images are suboptimal.4 Investigacions Biomèdiques
August Pi i Sunyer, Barcelona,
without underlying liver disease or affect patients with In patients without cirrhosis, complete obstruction of
Spain (V Hernandez-Gea,
chronic liver disease, mostly those with cirrhosis and the main portal vein, or obstruction of both of its two A Baiges, J-C Garcia-Pagan);
more rarely porto–sinusoidal vascular disorder or Budd– branches, is the predominant form (>80% of the Centro de Investigación
Chiari syndrome. patients).5 Conversely, in patients with cirrhosis, non- Biomédica en Red
Enfermedades Hepáticas y
Recent PVT refers to the formation of a thrombus in occlusive PVT (ie, when the lumen is only partially
Digestivas, Madrid, Spain
the portal vein, its branches, its radicles, or a combination occluded by the thrombus) accounts for around (V Hernandez-Gea, A Baiges,
within the last 6 months. Ultrasound coupled with 70% of the cases.6 PVT, in the absence of cirrhosis, is a J-C Garcia-Pagan, Prof A Albillos);
Doppler is usually the first-line diagnostic method, rare disease. However, collaborative efforts conducted Departament de Medicina i
Ciències de la Salut, Universitat
allowing the direct detection of the thrombus in the over the last two decades have yielded new knowledge de Barcelona, Barcelona, Spain
portal vein and the absence of flow in case of complete on PVT pathophysiology and management. PVT (V Hernandez-Gea, A Baiges,
PVT. Contrast-enhanced CT is usually preferred to MRI developing in patients with cirrhosis is much more J-C Garcia-Pagan); Department
to support the diagnosis of PVT, and is likely to show a common and has also been the subject of numerous of Surgery, Oncology and
Gastroenterology, University
hyperattenuating (hyperintense) thrombus on the recent studies. of Padova, Padova, Italy
unenhanced phase and an absence of enhancement of (M Senzolo); Departamento de
the lumen in the contrast-enhanced portal venous phase. Methods Gastroenterología y
Enlargement of the portal vein can be observed when VALDIG (Vascular Liver Disease Group) is an independent Hepatología, Instituto Ramón
y Cajal de Investigación
PVT is complete and recent.1 Chronic PVT refers to network of researchers with a common interest in vascular Sanitaria, Hospital
either portal cavernoma or incomplete resolution of the liver diseases endorsed by the European Association for Universitario Ramon y Cajal,
portal vein obstruction 6 months after a recent PVT. the Study of the Liver (EASL) that aims to foster research Madrid, Spain (Prof A Albillos);
Diagnosis of chronic PVT is based on contrast-enhanced in this field. In November, 2022, the VALDIG steering Department of Visceral Surgery
and Medicine, Inselspital, Bern
CT or MRI. Typical features include an absence of committee held an International Expert Conference on University Hospital, Bern,
visualisation of the portal vein, usually associated with a PVT in Paris, France, and included a faculty of Switzerland (Prof A Berzigotti);
cavernoma (ie, tortuous porto-portal collaterals bypassing 35 international experts. Service d’Hépatologie Hôpital
the occluded portion of the portal vein), enhancing after The first part of this Review consists of an extensive Rangueil, Université
Paul Sabatier, Toulouse, France
contrast injection.2 This imaging also allows congenital review of the current knowledge on PVT, summarising (Prof C Bureau); Department of
malformation of the portal vein, such as congenital the lectures given during the meeting. Faculty members Gastroenterology and

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 1

 Review

Hepatology, Erasmus MC of the Paris PVT meeting wrote a section of this review fibroblasts (figure 1). These changes are more
University Medical Center, and reviewed the whole manuscript. pronounced in patients with cirrhosis and PVT than in
Rotterdam, Netherlands
(S D Murad); Gastroenterology
The second part of the Review is dedicated to endpoints those without.16 Understanding why these changes occur
and Hepatology Department, for future clinical studies on PVT. These endpoints are and their contribution to portal vein thrombosis is an
Ente Ospedaliero Cantonale the result of extensive discussions that took place in attractive field that requires further research.
Faculty of Biomedical Sciences 2022 and 2023 during and after the Paris PVT meeting.
of Università della Svizzera
Italiana, Lugano, Switzerland
These endpoints were proposed by the organisers of Haemostasis in the portal circulation
(Prof A De Gottardi); Service the conference to a panel of 57 experts and were voted In patients with cirrhosis, there are well-documented
d’Hépatologie (Prof F Durand, on in January, 2024, according to a Delphi method changes in the haemostatic system in the systemic
L Moga, A Payancé, A Plessier, (appendix p 31–34). circulation. These include thrombocytopenia and
Prof D Valla, Prof P-E Rautou),
Department of Pathology
decreased plasma levels of coagulation factors, inhibitors
(Prof V Paradis), and Service de The specifics of the portal venous system of coagulation, and proteins involved in clot break­down.17
Radiologie (V Plaforet, Characteristics of the portal vein The net result of all haemostatic alterations is a haemo­
Prof M Ronot), AP-HP Hôpital From an embryological point of view, the portal venous static system that seems to be in fragile balance, with
Beaujon, Clichy, France;
Department of Surgery,
system develops from the vitelline venous system.7 Since both hypocoagulable and hypercoagulable features.
University Medical Center adult blood cells also derive from the vitelline venous Different groups have investigated haemostatic changes
Groningen, Groningen, system,8 this might explain the strong association in the portal circulation in patients with cirrhosis to
Netherlands (Prof T Lisman);
between myeloproliferative neoplasms and splanchnic understand better their potential role in PVT development
Vienna Hepatic Hemodynamic
Lab, Division of vein thrombosis, although evidence to support this (figure 1). The hypothesis was that local inflammatory
Gastroenterology and hypothesis is absent. responses, for example, related to bacterial translocation
Hepatology, Department of Physiologically, the portal venous system has some products18 lead to endothelial activation with the release
Medicine III, Medical University
specific characteristics: it does not drain into the heart; it of prohaemostatic proteins such as von Willebrand factor
of Vienna, Vienna, Austria
(M Mandorfer); Swiss Pediatric does not have venous valves; it drains capillaries from the and factor VIII,19 and to local activation of platelets and
Liver Center, Department of abdominal organs and thereafter ramifies into an coagulation (appendix p 3).20–22 However, a study by
Pediatrics, Gynecology and extensive network of liver sinusoids; and it is a low- Driever and colleagues23 has questioned the existence of
Obstetrics, University of
pressure system. The portal venous system is a high- a relative hypercoagulable state in the portal circulation
Geneva, Geneva, Switzerland
(Prof V McLin); Immuno- compliance, low-resistance system that can accommodate having analysed markers of inflammation and
Physiology and Pharmacology a large blood volume, as occurs after a meal, without coagulation activation in samples taken from the portal
Department, School of substantially increasing portal pressure. Histologically, vein, the hepatic vein, and the jugular vein in patients
Medicine and Biomedical
the portal vein wall is composed of tunica intima with with cirrhosis. Although markers of activation of endo­
Sciences, University of Porto,
Portugal (F Nery); Transplant endothelial cells and subendothelial connective tissue. It thelial cells, platelets, and coagulation were higher in the
Institute and Division of contains scattered fibroblasts and is bounded externally portal vein than in the hepatic vein, the difference in
Gastroenterology, NYU by a network of elastic fibres, similar to veins from the levels of these markers between the portal vein and the
Langone, New York, NY, USA
(Prof P Northup); Intestinal
systemic venous system. Specifics of the phenotype of jugular vein was, at most, modest. The authors concluded
Stroke Center, Department of endothelial cells lining the portal venous system, that comparison of markers from the portal vein with
Gastroenterology, IBD and especially the extrahepatic system where PVT usually that of the hepatic vein might erroneously lead to the
Intestinal Failure, AP-HP occurs, are not entirely known. Several single-cell conclusion that the portal circulation in patients with
Hôpital Beaujon, Clichy, France
(A Nuzzo); Laboratory for
transcriptomic analyses performed in healthy individuals cirrhosis is hypercoagulable, as the difference between
Vascular and Translational showed that portal endothelial cells have a distinct levels of haemostatic markers between the portal and
Science, INSERM UMR 1148, phenotype from other liver endothelial cells,9–12 but a hepatic vein is explained mainly by hepatic clearance of
Paris, France (A Nuzzo); comparison with other vascular beds has not been these markers. Likewise, Rahr and colleagues24 observed
Department of Hepatology and
Liver Transplantation, Royal
performed. Performing such comparisons might uncover that in patients without liver disease undergoing gastric
Free Hospital, London, UK mechanisms explaining the pathogenesis of PVT. surgery for benign disease, one marker of coagulation
(Prof D Patch); Service de In cirrhosis, the portal venous system has modifications activation (fibrinopeptide A) was higher in portal blood
Gériatrie, Hôpital Corentin due to the increase in portal venous pressure (portal than in systemic blood, although other markers did not
Celton (AP-HP), Paris, France
(J Poisson); Department of
hypertension): the calibre of the portal venous system differ. Although the study by Driever and colleagues25 has
Haematological Medicine, increases to accommodate the increase in portal venous different conclusions from previous works and has been
King’s College Hospital NHS flow and the thickness of the vein increases due to criticised, other observations argue against the concept
Foundation Trust, London, UK hypertrophy of its walls, particularly in the intima layer.13 of the cirrhotic portal vein as an inflammatory and
(L Roberts); Northwestern
Memorial Hospital,
Indeed, data from preclinical models, both in vitro and hypercoagulable entity. First, in a prospective study by
Northwestern University, in vivo, revealed that pressure-induced mechanical stress Turon and colleagues26 assessing risk factors for the
Chicago, IL, USA (Prof R Salem); causes an increase in the size and the number of smooth development of PVT in patients with cirrhosis, markers
Institute of Liver and Biliary
muscle cells, leading to a thickening of the vessel wall.14,15 of haemostasis and inflammation were not identified as
Sciences, New Delhi, India
(Prof S Sarin); Department of Patients with cirrhosis and severe portal hypertension predictors for PVT development. Instead, markers of
Gastroenterology, Seth GS requiring liver transplant have a variable degree of portal hypertension fully explained the risk for PVT
Medical College and KEM intimal thickening affecting the portal vein wall focally or development. Second, the Lisman group showed that
Hospital, Mumbai, India
circumferentially with an increase in the number of portal vein thrombi are often missing classical

2 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

haemostatic components such as platelets and fibrin.16 Healthy

Finally, portal vein thrombi are characterised by intimal Normal haemostasis
hyperplasia of the portal vein wall, which can result from Procoagulants Anticoagulants
changes in portal vein pressure rather than from media
activation of coagulation or the evolution of a former
100 µm
thrombus.

PVT in patients without cirrhosis Cirrhosis

Rebalanced haemostasis
Causes and underlying mechanisms with prothrombotic features
The prevalence of acquired and inherited risk factors for Procoagulants Anticoagulants
intima
PVT in the absence of cirrhosis was reviewed in 2019 by
Hernández-Gea and colleagues27 and is summarised in 100 µm
the appendix (p 23). Multiple risk factors were found in
Similar levels of markers of
15–36% of the patients with PVT, suggesting that inflammation and activation of
investigators should systematically perform compre­ coagulation in the systemic
circulation and in the portal vein
hensive screening for risk factors for thrombosis.27 In up
to 30% of the patients with PVT, no risk factor for Cirrhosis with portal vein thrombosis
thrombosis is identified. Data from the last 10 years have Rebalanced haemostasis
clarified the role of some new risk factors, including with enhanced prothrombotic
features
central obesity,28 SARS-CoV-2 infection or vaccination
(rarely),29–31 cytomegalovirus infection,32 and clonal intima 500 µm
haematopoiesis.33 Obesity is present in 25% of patients
with one or more risk factors for thrombosis and Differences between systemic
circulation and portal vein
45% with idiopathic PVT in the absence of cirrhosis.28 In unexplored
multivariate analysis, an increased waist circumference
was the strongest variable associated with idiopathic Figure 1: Haemostasis and the portal vein wall in healthy individuals, patients with cirrhosis, and patients
PVT.28 Cytomegalovirus disease can be associated with with cirrhosis and PVT
PVT with similar severity and evolution as patients PVT=portal vein thrombosis.
without cytomegalovirus disease, but with much higher
association with the prothrombin G20210A gene accumulate with age and favour arterial and venous (Prof A Shukla); Service de
variant (22%) than without (4%), suggesting that these cardiovascular comp­ lications and the development of Chirurgie Viscérale, Hôpitaux
Universitaires de Genève,
entities act synergistically to promote thrombosis.32 haematological malignancies.33 A first study, using Geneva, Switzerland
Although myeloproliferative neoplasm is the most molecular profiling in patients with myeloproliferative (Prof C Toso); Department of
common risk factor for PVT, the underlying mech­ neoplasm and PVT in the absence of cirrhosis, showed Liver and Hepato-Pancreato-
anisms are unknown. Recent data have clarified the that the presence of TP53 variants, a high JAK2 allele Biliary Unit, Queen Elizabeth
Hospital, University Hospitals
link between JAK2V⁶¹⁷ F mutation—the most commonly burden (>50%), or both is predictive of worse haemato­ Birmingham, Birmingham, UK
found mutation in patients with both PVT and myelo­ logical prognosis (myelo­ fibrosis, acute leukaemia, or (Prof D Tripathi); Institute of
proliferative neoplasm—and thrombosis. Both endo­ death).36 In a second study, next-generation sequencing Immunology and
thelial cells and endothelial progenitors can harbour profiling in 55 patients with idiopathic PVT revealed that Immunotherapy, University of
Birmingham, Birmingham, UK
JAK2V⁶¹⁷ F mutation; JAK2V⁶¹⁷ F mutated endothelial cells 25 had high molecular-risk variants. These variants were (Prof D Tripathi)
display increased P-selectin and von Willebrand factor associated with an increased risk for recurrent throm­ Correspondence to:
expression, and favour leucocyte adhesion and venous bosis.37 Thus, molecular profiling by next generation Prof Pierre-Emmanuel Rautou,
thrombi formation.34 Mathematical models from a sequencing in patients with PVT in the absence of Service d’Hépatologie, AP-HP
2022 paper showed that JAK2V⁶¹⁷ F mutation occurs early cirrhosis seems helpful for identifying underlying causes Hôpital Beaujon, 92110 Clichy,
France
in life, explaining the young age of patients with and adjusting treatment strategies based on prognostic Pierre-emmanuel.rautou@
myeloproliferative neoplasm and PVT in the absence of stratification. inserm.fr
cirrhosis.35 Other complex mechanisms contributing to See Online for appendix
the increased risk of thrombosis in myeloproliferative Pathophysiology of portal hypertension
neoplasm, involving all blood cell types and plasmatic Increased resistance to portal blood flow is the main
factors, have been reviewed by Guy and colleagues.34 patho­physiological mechanism leading to the develop­
Next-generation sequencing data have highlighted clonal ment of portal hypertension in patients with cirrhosis or
haematopoiesis of indeter­minate potential (CHIP) as a in those with PVT in the absence of cirrhosis. The
risk factor for PVT in the absence of cirrhosis. CHIP is pathophysiology of portal hypertension in cirrhosis has
due to the presence of clonally expanded hematopoietic been extensively reviewed and discussed (appendix p 24).38
stem cells caused by high molecular-risk variants (TET2, The increase in resistance is mainly located at the
DNMT3A, ASXL1, TP53, etc) in individuals without intrahepatic level and has two main components:
evidence of haemato­logical malignancy. These variants structural and dynamic (due to an increase in intrahepatic

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 3

 Review

vascular tone). Once portal hypertension is present, showed that despite some patients having hyperdynamic
portosystemic collaterals, trying to decompress the portal circulation, the mean cardiac index and systemic vascular
venous system, develop. However, the concomitant resistance values of the two cohorts of patients were
development of splanchnic and systemic vasodilation, by within the range of values observed in healthy people.43
increasing portal blood flow, maintains or even aggravates This finding is in line with another well-described
portal hypertension despite the formation of these finding in patients with cirrhosis,44 showing that a hyper­
collaterals. Vasodilation of the systemic and splanchnic dynamic circulation appears and progressively aggravates
beds promotes an effective hypovolaemia that triggers the in relation to the severity of portal hyper­tension and liver
activation of vasoactive systems (eg, renin, angio­tensin, dysfunction. Currently, there is an absence of clinical
aldosterone) that, by promoting sodium and water data to determine whether vasoactive systems are
retention, expands plasma volume trying to refill the activated or if there is plasma volume expansion in
vascular bed. Expansion of plasma volume facilitates patients with PVT in the absence of cirrhosis. Further­
the maintenance of an increased portal blood flow. The more, the potential implications of these changes in the
activation in intrahepatic vasoactive systems further context of developing portal hypertension-related comp­
increases intrahepatic vascular tone, closing a vicious lica­tions remain unexplored.
circle and maintaining portal hypertension (appendix p 24). A pioneering study in a small group of patients with
In patients with PVT in the absence of underlying liver PVT in the absence of cirrhosis showed that propranolol
disease, the increased resistance to portal blood flow is is, as it is shown in patients with cirrhosis, able to
exerted by the occlusion of the portal venous system by reduce azygos blood flow.45 Because portal pressure is
the thrombus (appendix p 24). Thrombosis does not challenging to measure in patients with PVT in the
require the occlusion of the entire vessel to produce a absence of cirrhosis, reduction in azygos blood flow—
marked increase in resistance and portal hypertension. an index of porto-collateral blood flow—is considered a
Thus, according to Poiseuille’s law, a reduction of just surrogate of the potential use of propranolol in these
50% of the vessel lumen already leads to a marked patients. However, as previously stated, hyperdynamic
increased pressure gradient.27 circulation is not a universal finding in patients with
Similar to patients with cirrhosis and portal PVT in the absence of cirrhosis. The absence of a
hypertension, patients with PVT and no underlying hyperdynamic circulation could affect the capacity of
chronic liver disease can develop portosystemic non-selective beta blockers such as propranolol to
collaterals. However, these patients also develop porto- reduce portal hypertension (not shown in PVT but
portal collaterals that bypass the obstruction and reach shown in cirrhosis).46 Of note, in patients with cirrhosis,
the intrahepatic portal circulation, helping to decompress carvedilol is more effective in decreasing the hepatic
the system and perfuse the liver. These porto-portal vein pressure gradient than propranolol,47 supposedly
collaterals are very rare in patients with portal hyper­ because carvedilol reduces portal blood flow and
tension secondary to increased resistance at the intra­ intrahepatic resistance. Because the increase in
hepatic level (eg, cirrhosis or porto-sinusoidal vascular intrahepatic resistance in PVT is due to the occlusion at
disorder). The potential effect of the presence and the portal vein level, there is no clear rationale
magnitude of these porto-portal collaterals perfusing the supporting a higher efficacy of carvedilol over
liver on the natural history of PVT in the absence of propranolol in reducing portal pressure in patients
cirrhosis is unknown. with PVT.
Unfortunately, there is a scarcity of good experimental
models of chronic PVT. The most studied experimental Non-invasive tests versus liver biopsy
model of prehepatic portal hypertension is the partial In clinical practice, the main challenge is to identify
portal vein ligated rat or mouse model. Both of these patients with cirrhosis from those without to enable
models develop an extensive network of portosystemic cirrhosis-specific management (particularly regarding
collaterals (more than 90% portal blood flow shunted a surveillance for hepatocellular carcinoma [HCC]). The
few days after the portal vein ligation procedure) that is workup for evaluating underlying liver disease at PVT
associated with rapid development of splanchnic and diagnosis should include a CT scan and liver stiffness
systemic vaso­dilation and sodium and water retention.39,40 measurement, to determine cirrhosis. A liver biopsy is
However, contrary to what happens in patients with not necessary when the suspicion of cirrhosis is high,
chronic PVT, no or only minor porto-portal collaterals namely in patients with at least one risk factor for
develop. cirrhosis and either segment IV atrophy together with a
Some small preliminary studies suggested that patients nodular liver surface or liver stiffness measurement
with PVT in the absence of cirrhosis had similar systemic greater than 20 kilopascals (kPa). A liver biopsy is
vasodilation as patients with cirrhosis.41,42 However, a generally unnecessary when there are no indicators of
study including 39 patients with PVT in the absence of underlying liver disease (normal liver blood test, normal
cirrhosis and 39 with compensated cirrhosis, matched on liver morphology, and normal liver stiffness measure­
the severity of portal hypertension and liver dysfunction, ment). However, liver biopsy should be considered when

4 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

porto-sinusoidal vascular disorder or other chronic liver with vitamin K agonists or direct oral anticoagulants was
diseases are suspected (appendix p 25). more effective in preventing recurrence of venous
In patients with PVT in the absence of cirrhosis, liver thromboembolism than vitamin K agonists or direct oral
stiffness measurement obtained through either anticoagulants alone.70 In patients with myeloproliferative
transient43,48,49 or two-dimensional shear wave elasto­ neoplasm and splanchnic vein thrombosis, the effect of
graphy,50 while generally elevated compared to healthy cytoreductive treatment on the risk of recurrence of PVT
volunteers, is usually less than 10 kPa. An animal study is unclear. Hydroxyurea did not significantly decrease
showed that 50%, 80%, and 100% portal vein occlusion the rate of recurrent thrombosis in a subgroup analysis
(generated by insertion and inflation of a balloon of a systematic review of 218 patients, including
catheter) lead to a decrease in liver stiffness (measured 180 with PVT.69 A small prospective study evaluated the
using magnetic resonance elastography) of 0·8%, effect of pegylated interferon alfa-2a in 20 patients with
7·7%, and 12·3%, respectively.51 These findings suggest myeloproliferative neoplasm and PVT.71 PVT did not
that PVT does not markedly affect liver stiffness recur in any of the patients during a median follow-up of
measurement values. In case-control studies comparing 2·2 years. Another small retrospective case series,
patients with cirrhosis with patients with either PVT in including 18 patients with myeloproliferative neoplasm
the absence of cirrhosis43,48,49 or porto-sinusoidal vascular and PVT, found no recurrence of thrombosis after 2 years
disorder,43,52,53 a liver stiffness measurement of more than of treatment with ruxolitinib.72 In summary, the benefit
20 kPa was specific to patients with cirrhosis. In addition, of adding hydroxy­urea to oral anticoagulation to prevent
although liver stiffness measurement is higher in thrombosis recurrence remains uncertain, particularly in
patients with porto-sinusoidal vascular disorder than in patients with PVT. Preliminary studies suggest that
those with PVT in the absence of cirrhosis,43 it cannot pegylated interferon and ruxolitinib could effectively
discriminate between the two conditions. prevent thrombosis recurrence in patients with
Liver morphological changes can help discriminate myeloproliferative neoplasm and PVT.
cirrhosis from PVT in the absence of cirrhosis or porto- The effect of eculizumab on thrombosis recurrence in
sinusoidal vascular disorder. Cirrhosis is associated with splanchnic vein thrombosis caused by paroxysmal
caudate lobe hypertrophy, segment IV atrophy, and a nocturnal haemoglobinuria has been evaluated. In
nodular liver surface.54,55 In chronic PVT, liver morphology one observational study, including 62 patients, 42 were
is normal at early stages, while peripheral atrophy, treated with eculizumab and 20 were not treated with
segment IV hypertrophy, or both, can appear later.56 In eculizumab. A new thrombosis in the splanchnic or extra
long-standing PVT in the absence of cirrhosis, the liver splanchnic territory occurred less frequently in patients
surface can be nodular in 20% of patients.57 PVT can treated with eculizumab than those without.73
occur in patients with porto-sinusoidal vascular disorder.58 In inflammatory bowel diseases, the inflammatory
However, no CT scan feature can help discriminate response leads to a hypercoagulable state, substantially
patients with PVT and porto-sinusoidal vascular disorder increasing the risk of venous thromboembolism. It has
from those with PVT and a normal liver.59–62 been suggested that 5-aminosalicylic acid and infliximab
The usefulness of serum biomarkers of fibrosis has not could reduce the risk of venous thrombosis (outside the
been evaluated in PVT. splanchnic territory),74 indicating that bowel disease
control is associated with a decreased risk of thrombosis.
Effect of aetiological therapy Whether these results also apply to PVT recurrence
In patients with PVT in the absence of cirrhosis, data on remains unknown.
controlling the identified risk factor for thrombosis Regarding Behçet’s disease, it has been suggested that
recurrence are scarce. In myeloproliferative neoplasm, treatment with steroids or immunosuppressive agents
several studies investigated the effect of controlling decreases the recurrence of thrombosis and liver-related
the haematological disease on venous thrombosis events in patients with Budd–Chiari syndrome.75,76 Data
recurrence, mainly focusing on patients with a history of specifically on PVT in Behçet’s disease are lacking
thrombosis outside the splanchnic vascular bed.63–67 In Altogether, these data support either the treatment or
these patients, studies evaluating the effect of cyto­ the removal of associated risk factors when identified.
reductive treatment on thrombosis recurrence gave However, data are still lacking to support anti­coag­ulation
inconsistent results.68 A 2018 paper presented a pooled interruption when the cause of PVT is well controlled.
analysis of 565 patients with myeloproliferative neoplasm-
related venous thromboembolism and explored the role Anticoagulation in patients with recent and chronic
of hydroxyurea (hydroxycarbamide) in combination PVT in the absence of cirrhosis
with aspirin or oral anticoagulants. Oral anticoagulation A multicentric European cohort study of people with
was independently associated with lower thrombosis recent PVT showed that anticoagulation with low
recurrence, whereas hydroxyurea was not.69 By contrast, molecular weight heparin (LMWH) followed by
in a 2021 systematic review of 1235 patients with myelo­ vitamin K agonists led a low rate of mesenteric
proliferative neoplasm, the combination of cytoreduction infarction (2%) and recanalisation of the portal vein in

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 5

 Review

40% of patients.5 Baveno VII recommendations indicate Spleen stiffness measurement seems promising, but
that anticoagulation should be started at a therapeutic further studies are needed.48,50
dose immediately after diagnosis of recent PVT.77 LMWH In patients with chronic PVT in the absence of
is widely accepted in this situation. Unfractionated cirrhosis, bleeding is the most frequent complication of
heparin should be used cautiously, as a high prevalence portal hypertension. In a study of 178 patients with PVT
of heparin-induced thrombocytopenia has been reported in the absence of cirrhosis, bleeding was the first
in this population.78 Although data are limited, direct oral manifestation of PVT in 15% of patients, and occurred in
anticoagulants can be considered a primary option in 16% of patients during the follow-up period.83 As seen in
patients without so-called triple positive antiphospholipid patients with cirrhosis, the presence of large oesophageal
syndrome or features of intestinal ischaemia.77 A retro­ varices is an independent predictor of variceal bleeding.83
spective study in 330 patients with recent PVT in the Despite secondary prophylaxis, rebleeding occurs in up
absence of cirrhosis showed that direct oral anticoagulants to 47% of patients within 5 years of the first bleeding
were associated with similar rates of recanalisation as episode.83,84 Sarcopenia was associated with a higher risk
LMWH and higher rates than warfarin derivatives.79 of refractory variceal bleeding in a study of 51 patients
Moreover, the risk of major bleeding was significantly with chronic PVT in the absence of cirrhosis.85 Conversely,
lower with direct oral anticoagulants than with warfarin. anticoagulant therapy was not associated with a higher
In this study, most patients had a local risk factor for risk of variceal bleeding.80,83 Endoscopic band ligation
thrombosis, and only 11% had an myeloproliferative seems safe in patients with chronic PVT in the absence
neoplasm. of cirrhosis without interrupting vitamin K agonists,
In chronic PVT in the absence of cirrhosis, the based on an assessment of 471 endoscopies in that
approach varies according to associated risk factors for setting.86 Given these findings, recommendations for
thrombosis (appendix p 26). In patients with at least one patients with cirrhosis could be applied to patients with
strong risk factor for thrombosis recurrence, namely chronic PVT in the absence of cirrhosis with acute
anti­
phospholipid syndrome, myelo­ proliferative neo­ variceal bleeding and for primary or secondary
plasm, or personal or first-degree history of spontaneous prophylaxis of variceal bleeding.77
venous thromboembolism, long-term anticoagulation is Portal vein recanalisation with or without a transjugular
recommended. In patients without a strong risk factor intrahepatic portosystemic shunt (TIPS) procedure has
for thrombosis recurrence, a randomised open-label been performed in patients with portal-hypertension
controlled trial showed that rivaroxaban 15 mg once daily related bleeding not controlled with endoscopic
reduced the recurrence of thromboembolic events or therapy.87–91 The rebleeding rate in patients with patent
death without increasing the occurrence of major stents is less than 10%.89–91
bleeding.80 In patients in whom anticoagulation was
discontinued, a D-dimer concent­ ration of less than Portal cholangiopathy
500 ng per mL (measured with the Innovance technique, Portal cholangiopathy is broadly defined as biliary ductal
Siemens [Munich, Germany]) 1 month after anti­ and gallbladder wall abnormalities in patients with PVT in
coagulation interruption predicted a low risk of the absence of cirrhosis and portal cavernoma.92 Portal
recurrence.80 On the other hand, an observational study cholangiopathy develops due to collateral formation
gathering 134 patients with PVT in the absence of around the biliary system, which causes compression of
cirrhosis and no strong risk factor for thrombosis, bile ducts, ischaemic bile duct injury, chronic inflam­
factor VIII concentration equal to or greater than 150% mation, and fibrosis, which can in turn lead to stricture
was predictive of recurrent thrombosis.81 Thus, although formation.93 Despite an 80–100% prevalence of portal
long-term anticoagulation is broadly recommended in cholangiopathy at magnetic resonance cholangio­
patients with chronic PVT in the absence of cirrhosis, the graphy,93,94 only 5–15% of patients develop symptoms after
challenge is identifying populations that do not need a median duration from diagnosis of PVT of 7 years (range
such long-term treatment. Next-generation sequencing, 0–24 years).93–96 Whether portal cholangiopathy can
D-dimers, and factor VIII could be helpful in this progress over time remains debated.94,97 The clinical stages
situation.37,80 of portal cholangiopathy have been conventionally divided
into asymptomatic and symptomatic, including abdominal
Variceal bleeding in patients with PVT pain, jaundice, and cholangitis.96,98 The cholangiographic
In patients with recent PVT, gastro-oesophageal varices findings include extrinsic impressions or indentations,
can develop as soon as PVT occurs and can progress shallow impressions or indentations, irregular ductal
during the first year of follow-up.82 Screening for varices contour, stricture, up­stream dilatation, filling defects, bile
should thus be performed in patients with recent PVT, duct angulation, and ectasia.92,94,96 Classification of portal
and endoscopy should be repeated 1 year after PVT cholangiopathy is based on the severity of biliary changes:
diagnosis in the absence of recanalisation.77 Non-invasive no abnormalities (grade 0); irregularities or angulations of
methods, including liver stiffness measurement, are not the biliary tree (grade 1); indentations or strictures without
accurate enough to replace endoscopy in this setting.48 dilatation (grade 2); strictures with dilatation (extrahepatic

6 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

duct >7 mm, intrahepatic duct >4 mm; grade 3). Currently, However, these factors are more accurate for patients with
intervention is not based on these biliary changes and is arterial mesenteric ischaemia than in those with venous
restricted to symptomatic patients only. Following acute mesenteric ischaemia.108 Thrombolytic therapies
multidisciplinary discus­sion, treatment options include have only been evaluated in small case series (reviewed in
ursodeoxycholic acid, endo­ scopic biliary drainage, a recent meta-analysis114 with a moderate rate [around 20%]
radiological portal vein recanalisation, and surgery, either of vein recanalisation but a high risk [around 90%] of
alone or combined. Most patients with symptomatic bleeding).104,115–118 Based on these pre­ liminary data,
portal cholangiopathy are managed endo­ scopically to thrombolytic therapy can be carefully evaluated in
remove biliary stones, with good long-term outcomes.96,99 experienced centres for selected patients who are clinically
Sphincterotomy and ductal clearance (balloon or basket) deteriorating despite medical therapy to avoid laparotomy
are feasible, although haemobilia can occur.100 Endoscopic and intestinal resection. Although the evidence is scarce,
stent placement (plastic or self-expandable metal stent) is patients with a history of acute mesenteric ischaemia are
usually needed for stenosis.99–101 Radiological portal vein considered to have an indication for long-term
recanalisation is increasingly used to treat patients with anticoagulation therapy.77
symptomatic portal cholangio­ pathy, with promising
results when performed in expert centres.87,90 Biliary Chronic abdominal pain
dilatation can be treated with endoscopic retrograde Patients with chronic PVT might have chronic or
cholangiopancreato­ graphy before recanalisation to recurrent abdominal pain. Contrast-enhanced CT should
decrease the risk of haemobilia. Surgical approaches be performed to rule out the recurrence of thrombosis,
have also been proposed, including portosystemic shunt particularly in the small collaterals or tributaries. In rare
followed by hepatico-jejunostomy,100,102,103 but are cases, mesenteric vein thrombosis has been associated
increasingly replaced by interventional radiology with intestinal stricture because of previous intestinal
techniques. ischaemia.119 Diagnosis is based on symptomatology of
small bowel occlusion, and visualisation of benign small
Mesenteric ischaemia bowel stenosis at imaging (CT enteroclysis). Finally,
PVT can extend to the mesenteric vein and lead to acute chronic abdominal pain has been reportedly associated
mesenteric ischaemia in 4–58% of the patients, with a with portal cavernoma alone, in the absence of obvious
mortality ranging from 10–45%.5,104–106 Factors associated recurrent thrombosis. The effect of portal vein recanal­
with intestinal resection in patients with acute mesenteric isation on these symptoms is debated.87,90 Exceptional
vein thrombosis include diabetes mellitus (suggesting cases of protein-losing enteropathy (via portal hyper­
an arteriosclerosis-related component) and thrombus tension related impaired lymphatic drainage or intestinal
extension to the small venules.105 As there is no accurate hyperpermeability120,121) have been reported in patients
clinical or laboratory sign, diagnosis of acute mesenteric with portal cavernoma.83
ischaemia relies on CT, particularly signs of intestinal
injury including wall thickening, decreased or absent Interventional radiology
contrast enhancement, bowel dilatation, pneumatosis Despite the rapid introduction of anticoagulation in
intestinalis, or portal venous gas.4 A multiphasic CT scan patients with recent PVT, complete recanalisation occurs
protocol, including unenhanced, arterial, and venous in 40% of the patients, and the development of portal
phase images, without a positive oral contrast agent, and hypertension is likely. Moreover, data show that the
excellent CT image quality improves the interobserver outcome of patients with PVT associated with JAK2V⁶¹⁷ F
agreement of imaging features of acute mesenteric mutation is poor despite anticoagulant treatment,
ischaemia and diagnostic accuracy.4,107,108 No randomised suggesting that additional treatment options are
clinical trial has ever been performed in acute mesenteric needed.117,122 Accordingly, portal vein recanalisation using
ischaemia, and treatment is based on preclinical and interventional radiology has been proposed in patients
observational studies. First-line medical treatments with PVT whose clinical features suggest a low probability
provided in expert centres includes early anticoagulation, of recanalisation, a high risk of mesenteric ischaemia, or
bowel rest, proton pump inhibitors, and broad-spectrum both. Results reported so far show that this approach is
oral antibiotics, with reported low rates of intestinal feasible and that higher recanalisation rates than
necrosis and death.109–113 Intestinal surgical resection anticoagulation alone can be achieved.117,123–132 It might
followed by delayed anastomosis is indicated in trans­ lead in the long term to lower incidence of portal
mural necrosis.104,109 In a prospective study of 67 patients hypertension-related compli­ cations. However, severe
with arterial and venous acute mesenteric ischaemia, procedure-related morbidity and fatalities have been
organ failure, plasma lactate levels greater than reported when using throm­bolytics (appendix p 4).117,123–132
2 mmol per L, and small bowel dilatation greater than Large prospective multi­ centre studies are needed to
2·5 cm on CT scan were independent predictors of determine the benefit–risk ratio of this approach.
trans­mural intestinal necrosis requiring surgical resection In patients with chronic PVT in the absence of
and could help select patients requiring surgery.108 cirrhosis, portal vein recanalisation with or without TIPS

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 7

 Review

has been used exponentially over the last decade. surveys, focusing on disease-related variables without
Modern imaging modalities delineating the venous addressing the social effects and adopting a patient-
anatomy and improved technical approaches to the centred perspective. However, in the field of PVT, patients
thrombosed venous system have revolutionised the face specific problems that are not well reflected in these
inter­ventional radiology approaches to portal cavernoma generic questionnaires.
and opened the possibility of providing definitive portal One study evaluated HRQoL in PVT (either with or in
decompressive treatments.133,134 Recognised indications the absence of cirrhosis) and found that PVT was
for this approach include variceal bleeding recurring associated with new onset depression and anxiety.144
despite non-selective beta blockers and band ligation Future studies on HRQoL in PVT should identify
and symptomatic portal cholangiopathy. Cases of meaningful PROs that can be combined with clinical
patients with difficult to treat ascites, chronic abdominal markers for a comprehensive assessment of treatment
pain, enteropathy treated with portal vein recanal­ effectiveness that in turn can be a direct measure of
isation, or a combination, have also been reported success for high-quality patient-centred care. Identifying
(appendix p 7).87,89–91,116,125,135–138 High technical success rates relevant PROs is an unmet need where the involvement
have been reported by expert teams, associated with good of patient associations will be crucial.
control of the symptoms, with evidence indicating that
recanalisation might also improve muscle mass.87 Fertility, contraception, and pregnancy
Patients are placed under general anaesthesia Fertility, contraception, and pregnancy are frequent and
(appendix p 27), then the venous system can be accessed relevant issues among patients with PVT as women of
via transjugular, percutaneous transhepatic, transplenic, childbearing age account for 20% of the PVT patient
or transmesenteric routes depending on the morphology population and their 5-year survival rate is above 90%.5
of the occlusion, the quality of the inflow and outflow, The prevalence of infertility in patients with PVT is
and the anticipated need to place a TIPS. Contrast unknown. In the general population, contraception and
injection of the splenomesenteric system identifies the pregnancy are associated with a hyper­ coagulable state
remnant portal vein cord and shows the extent of the with increased activity of procoagulant factors, a decrease
occlusion to the portal tributaries and intrahepatic in particular natural anticoagulant factors, and
branches. The occlusion is crossed with catheters and fibrinolysis.145,146 Accordingly, women taking oral contra­
guidewires, with the help of balloons. Often, a snare is ception containing oestrogens are at increased risk of
advanced and used to facilitate through-and-through venous thromboembolism outside the splanchnic
access. From there, liberal use of portal, splenic, or territory (three-fold higher risk during the first year of
mesenteric stents is performed along with completion use).146–148 There is little evidence for the role of oral
TIPS if needed, ensuring prompt inflow and outflow. contraception in PVT development. Indeed, the proportion
Varices might be embolised during the procedure to of women with a recent diagnosis of PVT taking oral
minimise future risk of bleeding and maximise contraception varies from 5–50% and is frequently
hepatopetal flow.87,90 After the procedure, the maintenance associated with other prothrombotic conditions.5,149 Never­
and duration of anticoagulant therapy are decided on a theless, in women with known PVT, contraception
case-by-case indication. containing oestrogen should be avoided and progestogen-
only oral contraception or an intra­uterine device should
Specific issues be preferred. In the general population, during pregnancy
Effect on quality of life and postpartum there is a five-fold increased risk of
Patients with PVT in the absence of cirrhosis are usually venous thromboembolism.146–148 Whether pregnancy also
young and economically active people who could favours PVT is unclear, as only rare cases of pregnancy
experience health selection effects and social inequalities revealing PVT have been reported.150
due to their PVT. Moreover, PVT diagnosis can be The outcome of pregnancy in women with PVT is
associated with diagnostic conundrum due to the rarity typically good when the disease is treated and well-
of the disease, and recognising associated underlying controlled.151,152 Pregnancy should be planned, and a
thrombophilic dis­orders can be challenging, leading to preconception visit is recommended to inform the
diagnostic delays and uncertainty. The development of patient about the risks of pregnancy with PVT, ensure
portal hypertension brings anxiety,139 work-related portal hypertension and underlying prothrombotic
disability due to the necessary multiple hospital visits factors are controlled before conception, and adjust drug
and interventions, fear of complications, and social prescription.152,153 Five studies assessed the outcome of
stigmatisation.140–143 Under­standing the effect of PVT on pregnancy in a total of 260 women with known and
health-related quality of life (HRQoL) will allow the controlled PVT.149,151,154–156 There was no maternal death,
identification of relevant patient-related outcomes and the rate of live births was high (about 85%,
(PROs) and the incorporation of patient-centred care into comparable to that in a general population). However,
clinical practice. The most widely used generic HRQoL the rates of prematurity (11%) and fetal death (2%)
questionnaires are the 12 item and 36 item short form appeared to be higher than in the general population.

8 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

Concerning maternal outcomes, no splanchnic vein transplantation can be considered in rare circumstances
rethrombosis was observed, but the rate of pre-eclampsia and on a case-by-case basis if life-threatening extrahepatic
was high (4%), supposedly explained by thrombotic complications such as pulmonary vascular disease
occlusion of the placenta circulation, especially in women mandate restoration of portal flow (panel 1).
with underlying prothrombotic disorders. Bleeding In children with recent PVT, in case of partial or
related to portal hypertension is a rare complication complete isolated left portal vein thrombosis, watchful
during pregnancy when adequate pro­phylaxis has been
provided (3% haemorrhages, 50% without adequate Panel 1: Management of portal vein thrombosis in children
prophylaxis).157 Overall, pregnancy should not be contra­
indicated if the PVT and underlying thrombotic factors Risk factors
are controlled. • Umbilical venous catheter
• Omphalitis or sepsis of abdominal origin
PVT in children: specificities as compared with adults • Exchange transfusion
PVT is the most common cause of portal hypertension in • Dehydration
children and differs from PVT in adults in its causes, • Thrombophilia (rare)
clinical presentation, and management. In paediatrics, Clinical presentation: diagnostic delay is frequent
chronic PVT is the most common presentation. Recent • Splenomegaly or hypersplenism
PVT is rarely diagnosed, but diagnoses are typically in • Haematemesis
premature infants in neonatal ICUs where Doppler • Haemorrhoids
ultrasound is often performed routinely after umbilical • Liver nodules
venous catheter placement. Growth retardation is an • Growth retardation
important and common manifestation of PVT in
children.158 The most common risk factor for the Diagnostic imaging
development of chronic PVT in children is an umbilical • Ultrasound coupled with Doppler
venous catheter placed in the neonatal period.159 When • Contrast-enhanced CT
there is no history of such a catheter, the cause usually Pre-therapeutic imaging
remains unclear. History might reveal an episode of • Retrograde portal venogram
profound dehydration, exchange transfusion, or sepsis. • Patent or dominant main portal vein
Thrombo­philia workup usually does not reveal a general • Rex recesses permeability (for surgical planning)
risk factor for thrombosis. Rarely, in children, recent • Left-right portal flow
splenectomy could be a cause of acute PVT. When an • Direct portal venogram (rarely)
umbilical venous catheter is incriminated, the left portal • Patent or dominant main portal vein
vein is most often involved, with the main portal vein • Angioplasty access and feasibility (in experienced
spared.160 When a thrombus in the main portal vein goes centres)
unnoticed, the patient has subclinical, cavernous
transformation of the portal vein that ultimately can Management of chronic PVT
present as portal hyper­tension between 2 and 17 years • Watch for consequences of portal hypertension and
with one or more of the following symptoms: haema­ portosystemic bypass
temesis, splenomegaly, thrombo­cytopenia, or, rarely, liver • If Rex permeable: consider pre-emptive reportalisation
nodules in the absence of parenchymal liver disease.159,161 according to local resources
Management of a child with PVT depends on their age • If Rex not permeable: watch and wait for complications
at presentation. In case of neonatal evidence of main • If life-threatening portal hypertension-related
portal obstruction, LMWH is considered but should be complications: perform surgical shunt (distal splenorenal
balanced against the risk of intracranial bleeding. Catheter- preferred)
directed thrombolysis for recent PVT has been success­ Long-term follow-up
fully attempted in children aged 3–17 years.161 In the setting • Rex shunt performed
of chronic PVT, the preferred method is to perform a • Follow annually for signs of portal hypertension
meso-Rex shunt (appendix p 28) because it restores the (platelet count, Doppler)
blood flow to the liver and should be considered pre- • Invasive imaging and angioplasty in the case of
emptively.162,163 This approach is only possible if the Rex suspected stenosis
recessus is permeable and wide enough, associated with • Cavernoma—no surgical shunt
left–right portal flow documented on retrograde portal • Monitor for complications of portal hypertension
venogram.164 When the Rex recessus is not permeable • Other surgical shunts
and the patient presents with complications of portal • Monitor for complications of portal hypertension and
hypertension, splenorenal or mesocaval shunts are portosystemic bypass
palliative, but expose the patient to the long-term • Monitor for growth-related changes in the shunt
complications of portosystemic bypass.165,166 Liver

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 9

 Review

management is usually preferred. If the thrombus can be considered. The hepatic venous pressure gradient
extends to both portal branches or the main portal vein (HVPG) decreases in most patients who have an SVR,183
systemic heparin-based anticoagulation is considered, which could primarily be explained by a decline in
weighing the benefits against the risks of intracranial intrahepatic vascular resistance. In contrast, data on the
bleeding in this at-risk population.160 In children with effect of SVR on portal blood flow velocity in cirrhosis
recent PVT following a splenectomy, directed throm­ are limited and conflicting. Although a small paired
bolysis has shown promising results in experienced phase-contrast MRI study reported that portal blood flow
centres.161 There is no evidence or consensus on primary velocity is unaffected by SVR,184 a Doppler ultrasound-
or secondary prophylaxis of oesophageal varices in based study found an increase from 12·44 to 14·14 cm
children. per s.185 HCV cure ameliorates abnormalities in routine
coagulation tests and laboratory biomarkers suggestive
PVT in patients with cirrhosis of hypercoag­ulability, possibly leading to a more stable
Epidemiology and risk factors haemostatic equil­ibrium.186,187 Finally, although HCV
A meta-analysis revealed a pooled prevalence of PVT of eradication has been shown to improve systemic
14% in patients with cirrhosis and a pooled incidence of endothelial dysfunction, this effect seemed to be less
4·6 % per 100 patient-years.167 Prevalence increases with marked or even absent in patients with cirrhosis188–192 and
the severity of cirrhosis.6 50% of cases of PVT are it is still unknown if or how endothelial dysfunction in
diagnosed at the time of liver transplantation.168 PVT in the portal and splanchnic veins is modified. Clinical
patients with cirrhosis has been consistently associated observation showed that the risk of PVT is not
with features reflecting the severity of portal hypertension, abolished and does not seem to relevantly decrease in
namely high-risk oesoph­ageal varices, history of variceal the first years following HCV cure.193,194 This apparent
bleeding, presence of ascites, and lower platelet contradiction between patho­ physiological and clinical
count.26,167,169,170 The association of PVT with the following data is probably explained by the modest mag­nitude of
features remains debated: a decrease in portal blood flow the changes mentioned above on Virchow’s triad
velocity,26,169,170 non-selective beta-blockers,26,171,172 plasma components. There is a knowledge gap about the effect
D-dimer concentrations,26 inflammation reflected by of HBV suppression, abstinence from alcohol, and
plasma interleukin (IL-6) con­cent­ration, or neutrophil-to- weight loss on PVT incidence. Although it seems likely
lymphocyte and platelet-to-lymphocyte ratios,173–175 features that removal or suppression of the primary causal factor
of the metabolic syndrome including higher body mass during long-term follow-up will reduce PVT risk,
index, obesity, and diabetes.26,171,176–178 Although suggested surveillance for PVT should be continued regardless of
by previous retro­spective case series,179 PVT in cirrhosis curing the underlying issue. Long-term studies are
has not been associated with the ABO blood group180,181 nor therefore needed to define the effect of aetiological
inherited thrombophilia26,169,182 in large prospective cohorts therapies on PVT risk. Until this knowledge gap is filled,
(appendix p 29). surveillance for PVT should be continued regardless of
aetiological cure.
Effect of aetiological treatment
Removal or suppression of the primary causal factor in Variceal bleeding
cirrhosis includes hepatitis C virus (HCV) cure, In patients with PVT, HVPG is not accurate to estimate
hepatitis B virus (HBV) suppression in the absence of portosystemic gradient because of the presinusoidal
hepatitis D virus (HDV) infection, and abstinence from block. However, clinical retrospective observations
alcohol. Besides improving hepatic function, these indirectly suggest that PVT aggravates portal hyper­
measures ameliorate portal hypertension, thereby tension in patients with cirrhosis. First, after variceal
reducing the risk of decompensation of cirrhosis.77 The bleeding, PVT has been independently associated with
definition and effect of the removal or suppression of the inability to control bleeding, 14-day and 6-week
primary causal factor in patients with cirrhosis unrelated rebleeding195 and 6-week mortality.196 In addition, PVT
to viral hepatitis or alcohol is less well established, which has been associated with a longer time to achieve
does not necessarily imply that the respective therapies eradication of varices in patients undergoing endoscopic
(eg, phlebotomy for haemo­chromatosis) are less effective band ligation.197 However, the rebleeding rate is similar
in modifying the disease course. in patients with or without PVT (45% at 1 year) treated
Sustained virological response (SVR) following HCV with endoscopic band ligation and propranolol.198,199
direct-acting antiviral therapy provided an unprecedented These considerations do not translate into practical
opportunity to study the effect of removing the primary specificities since managing variceal bleeding in patients
causal factor on PVT development in patients with with cirrhosis and PVT does not differ from those
cirrhosis. In this context, the effect of SVR on without PVT.77 The only difference might be that TIPS
Virchow’s triad of components that contribute to should be systematically considered in patients with
development of thrombosis, namely haemodynamic cirrhosis, PVT, and variceal bleeding, as TIPS prevents
changes, endothelial dysfunction, and hypercoagulability, the recurrence of variceal bleeding and ascites and

10 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

positively affects PVT by restoring portal blood flow.198 in patients receiving anticoagulation (9·7% vs 1·7%,
Initial concerns about anticoagulation and variceal p<0·001). In agreement with this finding, aggregated
bleeding in patients with cirrhosis and PVT have been data meta-analyses of anticoagulation in patients with
waived, as anticoagulation is even associated with lower cirrhosis and PVT also show that anti­ coagulation
rates of variceal bleeding.200 Moreover, once variceal improves survival.208,209 An important conclusion drawn
bleeding occurs, anticoagulation is not associated with from these studies is that the beneficial effects of
poorer outcomes.201 anticoagulation on cirrhosis progression is independent
of the drug used.
Rationale for anticoagulation Preclinical studies investigated the beneficial effect of
Once diagnosed PVT can spontaneously regress in up to anticoagulation in cirrhosis. Liver inflammation activates
two-thirds of patients, which seems more frequent in thrombin and protease-activated receptors, promoting
those with partial than complete PVT and those for hepatic stellate cell activation and thus supporting the
whom liver function improves.202 By contrast, the rate of hypothesis that anticoagulation could reduce fibrogenesis
spontaneous recanalisation is very low in patients with and cirrhosis progression. This hypothesis is supported
occlusive PVT.177 Anticoagulation in patients with by the following evidence: thrombin generation is main­
cirrhosis and PVT was studied in a meta-analysis of tained or increases throughout cirrhosis progression;210
predominantly retrospective studies. Anticoagulation fibrogenesis is increased in mice with a mutation to
resulted in higher (partial or complete) recanalisation factor V Leiden and CCl4-induced cirrhosis and decreased
(71% vs 42%), complete recanalisation (53% vs 33%), and in transgenic mice overexpressing tissue factor pathway
less progression of PVT (9% vs 33%), without an increase inhibitor with inferior vena cava ligation;211,212 and chronic
in any bleeding episodes.200 Anticoagulation is most treatment with enoxaparin or rivaroxaban decreases
effective if started within 6 months of diagnosis of PVT.203 collagen and fibrin deposition, deactivates hepatic stellate
Anticoagulation has long been proposed mainly for those cells, and reduces portal pressure in different models of
for whom liver transplantation was considered to rats with cirrhosis.213,214 Proposed indications for anti­
facilitate porto-portal anastomosis. coagulation are summarised in panel 2.
A growing body of evidence indicates that the beneficial
effect of long-term anticoagulation in patients with What type of anticoagulant?
cirrhosis goes beyond its effect on PVT. The first evidence The choice of anticoagulant is based on the type of
came from a randomised controlled trial in patients thrombosis (recent vs chronic PVT), the context
with Child-Pugh B7 to C10 cirrhosis that evaluated the (mesenteric ischaemia, listing for liver trans­ plant­
efficacy of LMWH in preventing PVT.204 In this study, ation, etc), and the patient, especially the severity of
enoxaparin (40 mg subcutaneously daily) reduced cirrhosis and comorbidities. The characteristics and
incident PVT by 90% and death and hepatic decomp­ limitations of the most used anticoagulants are
ensation by nearly 70%.204 Enoxaparin treatment, but not summarised in the appendix (p 11).215 In patients with
PVT, was independently associated with preventing recent thrombosis, especially those presenting with (or
decomp­ensation and death. Similar results were found at risk of) mesenteric ischaemia, LMWH is preferred
in a randomised controlled trial performed in 90 patients
with Child-Pugh B7 to C10 cirrhosis.205 In this study, the
Panel 2: Indications for anticoagulation in patients with
incidence of portal hypertension-related complications
cirrhosis and PVT
(mostly ascites) was lower in patients who received
rivaroxaban 10 mg per day than in those who received Non-candidates for liver transplantation
placebo, without an increase in major bleeding events. Usual indications
Next, in a well-matched cohort of US veterans with • Recent (<6 months) completely or partially occlusive
atrial fibrillation and cirrhosis (mostly Child-Pugh A), (>50%) thrombosis of the main portal vein
warfarin and direct oral anticoagulants were associated • Symptomatic PVT (independent of the PVT extension)
with lower all-cause mortality and hepatic decomp­ • Progression of PVT without anticoagulation
ensation compared with no anticoagulation.206 Finally, • Extension to the superior mesenteric vein
an individual participant data meta-analysis of Debated indications
five non-randomised studies containing 500 patients • Minimally occlusive (<50%) thrombosis of the main
that compared anticoag­ulation with LMWH or vitamin K portal vein
agonists with no anticoagulation in patients with
cirrhosis and PVT. LMWH or vitamin K agonists Potential candidates for liver transplantation
improved overall survival after adjustment for con­ Usual indications
founders, including Child-Pugh score.207 The survival • Any PVT (independent of the degree of occlusion and
benefit of anticoagulation was independent of PVT extension)
recanalisation. However, it is worth noting that there was PVT=portal vein thrombosis.
increased risk of portal hypertension unrelated bleeding

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 11

 Review

due to their short half-life and predictable series report safe reversal of dabigatran in the
pharmacokinetics. LMWH has been suggested to be peritransplant period.230,231 Therefore, there is a role for
more effective than warfarin concerning complete PVT direct oral anticoagulants in Child-Pugh class A cirrhosis
resolution, although this has not been prospectively and with caution in Child-Pugh class B cirrhosis, but
evaluated.200 The anti-Xa assay can lead to excessive direct oral anticoagulants should generally be avoided
dosing due to reduced anti­thrombin levels, especially in in patients with Child-Pugh class C cirrhosis. Further
more advanced cirrhosis.216 Fondaparinux has the prospective studies, including pharmacokinetic assays,
advantage of once-fixed daily dosing with a reduced are needed to optimise patient selection and choice of
risk of heparin-induced thrombo­cytopenia, potentially direct oral anticoagulants. Duration of anti­coagulation
higher efficacy than LMWH, but with an increased risk remains uncertain, as 30–40% of the patients develop
of bleeding (albeit minor).217 recurrent thrombosis between 2–5 months after
For longer-term management, vitamin K agonists have cessation of anticoagulation.218,232,233
historically been used in patients with cirrhosis and have
the advantage of once-daily oral therapy. International Interventional radiology
normalised ratio (INR) monitoring can be challenging Having previously been considered a contraindication
due to baseline elevated INR. A platelet count of less than to a TIPS procedure, PVT is increasingly seen as an
50 × 10⁹ per L with a vitamin K antagonist is associated indication. In candidates for liver transplantation, TIPS
with an increased risk of bleeding.218 An advantage of is an option to achieve recanalisation in patients with
vitamin K agonists over direct oral anticoagulants or obliterative PVT and an alternative to anticoagulation in
LMWH is the availability and familiarity with reversal patients with partial PVT. The objectives are to recanalise
agents. Thus, vitamin K agonists remain preferred for the portal vein trunk and restore portal blood flow to
patients on the liver transplantation waiting list, provided allow end-to-end portal anastomosis during the trans­
the baseline INR is not substantially elevated. In most plant procedure. Even if TIPS is often effective at
Western countries, it should be noted that donor bridging patients with PVT to transplantation, this
allocation policy is based on the patient’s MELD score. A technique has some limitations. PVT is more common
pitfall of vitamin K agonists in candidates for trans­ in patients with end-stage cirrhosis and a high MELD
plantation is that the MELD score includes INR which is score. However, a high MELD score (>18) might represent
artificially increased by vitamin K antagonists, with a a contraindication to TIPS due to the risk of further
corresponding increase in MELD. Therefore, caution deterioration in liver function and a high rate of hepatic
should be taken to avoid prioritising patients receiving encephalopathy.234,235 In addition, misplacement of the
vitamin K antagonists for organ allocation. TIPS can be a source of major technical difficulties
Patients with cirrhosis were systematically excluded during the transplant procedure, especially when the
from the seminal studies of direct oral anticoagulants in upper part of the stent is placed in the right atrium, or
the management of venous thromboembolism and the lower part is placed in the mesenteric vein. For these
stroke prevention in atrial fibrillation. However, there reasons, stenting in PVT before transplantation should
has been much interest in the off-label use of direct maintain 3–4 cm of unstented main portal vein, with the
oral anticoagulants in PVT, due to the ease of oral cranial end of the stent landing slightly short of the
administration without the requirement for (INR) hepatic vein-caval confluence. In this patient population,
monitoring.219–223 Of note, direct oral anticoagulants also stenting into the mesenteric system should never be
lead to an increase in INR, and thus could lead to the performed. The second indication concerns patients with
prioritising of patients on the waiting list for liver cirrhosis and PVT progressing despite anticoagulation
transplantation.224 The potential for direct oral anti­coag­ when drug compliance is confirmed (eg, based on
ulant accumulation in patients with decompensated therapeutic drug monitoring) and HCC is excluded, even
cirrhosis remains a concern, although the metabolism of without a history of bleeding or ascites. No data are
apixaban is not appreciably affected by chronic liver available to determine if TIPS is superior to anti­
disease.225 However, available data from several meta- coagulation in this context, and the choice between these
analyses suggest that the use of direct oral anticoagulants two options should be discussed with expert centres.
in patients with cirrhosis carry a risk of bleeding The complexity of TIPS placement depends on the age
(including intracerebral haemor­rhage or gastrointestinal) of the thrombus, the degree of occlusion, and its extent
similar to or probably lower than vitamin K (portal vein alone or involving additional splanchnic
antagonists.203,226–228 In patients with cirrhosis, a 2023 meta- veins). The technical elements that allow a successful
analysis of 29 studies showed that compared with TIPS include the patency of intrahepatic portal
conventional anticoagulants, direct oral anticoagulants branches,116 and a visible (even though occluded) portal
were associated with lower incidences of all bleeding vein on ultrasound when the clot is recent. However,
events (OR 0·71, 95% CI 0·52–0·98) and major bleeding there is an increase in reports of successful TIPS in
events (OR 0·55, 95% CI 0·37–0·83), even in patients patients with cavernoma, and no visible intrahepatic
receiving direct oral anticoagulants for PVT.229 Small case portal vein.236 In this setting, this approach uses a

12 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

transplenic or transjugular rendezvous approach that 27 patients receiving anticoagulation, more than
despite being technically challenging, is rapidly gaining 50% had an improvement in PVT, while only
adoption and use in multiple centres to permit 7% had progression.249 The presence of PVT complicates
socialisation of the procedure. It remains at the HCC treatment by increasing the morbidity and
technically challenging end of the TIPS spectrum. mortality of hepatic resection and comp­licating trans­
In patients without occlusive PVT, the technical success arterial chemoembolisation. Accordingly, when PVT is
rate of TIPS is 75–98% leading to recanalisation in complete or progressive, HCC and liver-related mortality
84% of patients (including complete recanalisation in is increased.249
73% of patients).237 After complete recanalisation, long- In patients with HCC without PVT, the question of
term patency is maintained so anticoagulation might using prophylaxis to prevent PVT is unanswered. The
not be required.237,238 In patients with occlusive number needed to treat to prevent one PVT with
PVT—even those with cavernous transformation or anticoagulation is about 6.6 Although waiting for
involvement of the superior mesenteric vein—several dedicated prospective studies, anticoagulation could be
case series have reported that recanalisation can occur in considered to prevent PVT in patients with large
65–100% of cases.126,236,238,239 The overall complication rate tumour burden and concomitant portal hypertension.
in expert centres is around 10%, and is even higher Anticoagulation should also be considered in patients
following thrombectomy or thrombolysis.237 with partial PVT and not delayed until progression since
it seems to affect prognosis. Since response to
PVT and hepatocellular carcinoma anticoagulation seems lower in patients with PVT and
In patients with cirrhosis, HCC is a specific risk factor HCC than in those with PVT without HCC,249 full-dose
for tumoral portal vein obstruction and PVT. anticoagulation should be considered as soon as PVT
Differentiating tumoral portal vein obstruction from occurs in patients with HCC.LMWH remains the
PVT is essential. The presence of more than three standard treat­ment for cancer-associated thrombosis in
non-invasive A-VENA criteria (comprising thrombus patients without cirrhosis, and LMWH has been shown
enhance­ ment, enlargement of the portal vein, neo­ to be superior to vitamin K antagonists in terms of
vascularity, proximity to HCC or prior treatment site, and efficacy and safety in many clinical trials.250 Caution
α-fetoprotein [AFP] greater than 1000 ng/dL), have must be used in patients treated with tyrosine kinase
100% sensitivity, 93·6% specificity, 80% positive inhibitors since these drugs interfere with direct oral
predictive value, and 100% negative predictive value for anti­coagulants and LMWH.251,252 Dose adjustment might
tumoral portal vein obstruction.240 be needed.253
In patients awaiting liver transplantation, the
prevalence of PVT is higher in patients with HCC PVT and liver transplantation
(41% vs 27%) than in those without.241–244 Two prospective Detailed imaging using CT or MRI should be systematic
studies showed that HCC is an independent risk factor during the evaluation process for eligibility for liver
for PVT, with a 1-year incidence of 25%, even in patients transplantation, with special attention given to portal and
with compensated cirrhosis.245,246 Patients particularly at mesenteric vein patency. In addition, due to the 5–10%
risk are those with multinodular HCC, a total tumour chance of new PVT occurring after registration on the
volume greater than 3 cm³, or both.245–247 This increased waiting list for liver transplantation, periodic screening
thrombotic risk in HCC could be explained by multiple with Doppler ultrasound is strongly recommended for all
prothrombotic alterations of haemostasis driven by patients awaiting transplantation.
HCC, illustrated by the increased maximum clot During transplantation, one key challenge is restoring
firmness at thromboelastography observed in patients portal blood flow to decrease the risk of delayed graft
with HCC and PVT.245 Mechanisms underlying this function and prevent post-transplant thrombosis.254 Every
prothrombotic state include thrombocytosis (or a normal technical effort should be made to avoid stricture and
platelet count despite clinically significant portal hyper­ excessive length of the anastomosed veins. However, a
tension), increased platelet activation and function, pre-existing PVT remains one of the main risk factors for
increased plasma fibrinogen, inhibition activation of post-operative complications. The intraoperative strategy
fibrinolysis, higher levels of prothrombotic circulating should be defined based on carefully assessing the pre­
microvesicles, and increased neutrophil extracellular transplantation images of the thrombus and the
traps and contact activation system.248 portosystemic collaterals.255 A physiological liver blood
The natural history of PVT in patients with HCC and inflow draining the mesentery should be restored when­
its effect on mortality has not yet been investigated. Only ever possible. In patients with PVT affecting only part of
one single-centre study explored the evolution of PVT the spleno–mesenteric veins (Yerdel 1–3), strategies
in patients with HCC. In 75 patients not receiving include thrombectomy, jump graft from the mesenteric
anticoagulation, PVT progressed in 63%, improved in 7%, vein, reno–portal anastomosis, or anasto­mosis between a
and remained stable in 30%.249 Progression of PVT was large varix and the donor portal vein.255–257 Procedures not
associated with non-response to HCC treatment. Of respecting the physiological portal inflow (typically

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 13

 Review

considered in some patients. The need for postoperative

Panel 3: Consensus recommendations for endpoints for studies on portal vein anticoagulation remains debated,259,260 and should be
thrombosis defined together with the surgeon, who can best
Classification of PVT extension appreciate the quality of the portal inflow. Long-term
• For studies on the natural history or medical treatment of PVT, the classification anti­
coagulation can be advised in patients with a
proposed in the AASLD and Baveno VII guidelines should be used (degree of hypercoagulable state, non-physiological portal vein
agreement 96%). reconstruction (or both), or interposition graft.255
• For studies related to liver transplantation, Yerdel’s classification should be used
(degree of agreement 80%). Endpoint for clinical studies on PVT
• For studies on interventional radiology (including portal vein recanalisation, TIPS), Classification of PVT extension
Marot’s classification should be used (degree of agreement 77%). The heterogeneity of PVT location, extent, and degree of
occlusion calls for standardised descriptive terminology
Classification of PVT evolution to improve the reporting, comparison, and validation of
• Measurement of the degree of occlusion of PVT should be assessed on contrast- future studies and trials. Many classifications have been
enhanced cross-sectional imaging with the portal venous phase on a plane strictly proposed to date that take into consideration various
perpen­dicular to the main portal vein (degree of agreement 96%). anatomical landmarks, various degrees of occlusion, and
• Minimal thrombosis or occlusion corresponds to a %RL equal to or greater than 50%, the coexistence of collaterals or an underlying hepatic
partial thrombosis or occlusion corresponds to a %RL less than 50%, and complete disease. The main published PVT classification systems
thrombosis or occlusion corresponds to a %RL of 0% (degree of agreement 96%). are summarised in the appendix (p 14). This section
• Upstream and downstream veins, namely the superior mesenteric vein, splenic vein, details the results of the Delphi expert consensus process,
and right and left portal branches should be visually assessed using a semi- which aimed to determine optimal endpoints for clinical
quantitative scale, as follows: no thrombosis=no thrombus or occlusion; incomplete studies of PVT (appendix p 31–34).
thrombosis=any amount of thrombus or occlusion but visible remnant lumen; Several classifications differentiate the involvement of
complete thrombosis=no remnant lumen (degree of agreement 94%). segmental intrahepatic branches, left and right portal
• The VALDIG PVT criteria evaluating PVT evolution should be based on changes in the vein branches, and the main portal vein. Some also
three following criteria: thrombus of the main portal vein, thrombus of the upstream include a possible extension to portal vein tributaries,
and downstream veins, and occurrence of new thrombus in veins initially devoid of such as the splenic and superior mesenteric veins. The
thrombus. These criteria allow the categorisation of PVT into four categories: 2021 practice guidelines from the American Association
complete resolution, improvement, stability, and progression of PVT (degree of for the Study of Liver Diseases (AASLD) suggest using
agreement 93%). three categories: intrahepatic, main portal vein, and
Timing of assessment
portal vein tributaries,215 but neither the recent update
• For patients in a stable clinical condition endpoints should be assessed on imaging at
from the Baveno consensus77 nor the Chinese guide­
6 and 12 months (degree of agreement 88%).
lines261 on PVT in cirrhosis mention any anatomical
• For patients at particularly high risk of recurrent PVT or progressive PVT (eg, patients
classification. The assessment of the residual portal
on the orthotopic liver transplant waiting list who cannot receive anticoagulation)
venous lumen is also essential. Although some studies
endpoints should be assessed on imaging at 4–6 weeks intervals (degree of
differentiate complete from partial occlusion, others
agreement 85%).
favour using thresholds. For instance, the Baveno group,
• For patients with worsening signs of portal hypertension or liver function, endpoints
in line with Yerdel and colleagues,256 recommends
should be assessed immediately, to identify progression and proceed to TIPS or other
separating complete occlusions (no persistent lumen)
treatment options (degree of agreement 98%).
from partially occlusive thrombosis (obstructing
more than 50% of the original vessel lumen) and mini­
Global endpoints for studies on PVT mally occlusive thrombosis (obstructing less than 50% of
• Global endpoints for studies on PVT should be different for patients with PVT in the the original vessel lumen).77 More granular systems exist,
absence versus in the presence of cirrhosis as well as for those patients with recent such as the one proposed by Bauer and colleagues262 in
versus chronic PVT (degree of agreement 98%). patients undergoing TIPS and endovascular procedures.
AASLD=American Association for the Study of the Liver. PVT=portal vein thrombosis. TIPS=transjugular intrahepatic
Although no consensual definition exists, the presence
portosystemic shunt. %RL=percentage of the remnant lumen. of cavernoma should be reported. Finally, the length of
thrombosis or the clot burden could represent promising
endpoints similar to what has been validated in brain
reno–portal anastomosis in the absence of a splenorenal ischaemia or pulmonary embolism.263 Dedicated tech­no­
shunt or a cavoportal anastomosis) are associated with logies are needed to translate this into clinical practice.
poorer post-transplant outcomes.256,258 In patients with Most classifications were developed in patients
diffuse spleno–mesenteric thrombosis (Yerdel 4) and no undergoing liver transplantation and focused on short-
dominant collateral shunting, the risk to benefit profile term clinical outcomes. A few studies have assessed the
should be carefully balanced. Pre­ transplant portal relationship between the location of PVT and clinical
vein recanalisation, a non-physiological cavo–portal outcomes in patients with cirrhosis who were not
anasto­mosis, or a multivisceral trans­plant­ation can be candidates for liver transplantation,262,264 or for patients

14 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

with PVT in the absence of cirrhosis undergoing patients with PVT. Although contrast-enhanced CT and
endo­vascular portal vein recanalisation,88 but long-term MRI are the recommended imaging techniques, no head-
outcome data are scarce. Onda and colleagues265 also to-head comparison exists to determine the best one.
proposed a classification of PVT occurring after hepatic Studies that categorise the progression of PVT as either
resection according to the site and grade of the thrombus. part of the natural history or as a response to treatment
In another context, several classifications of portal vein rarely assess the volume or anatomical location of the clot
tumour invasion have been published. Although they do nor the clinical consequences of such progression.200
not deal with the same patients, they are worth The criteria used in clinical trials or proposed in
mentioning because they propose to label the different guidelines261 to classify PVT evolution are summarised in
sites of portal vein occlusion, which facilitates the appendix (p 19).
comparison between studies. The Vp classification We thus decided to conduct a pilot study including
system from the Japan Liver Cancer Study Group is the 44 patients with PVT (32 with cirrhosis and 10 with porto-
most commonly used.266 The occlusions are classified sinusoidal vascular disorder) to determine the repro­
from Vp1 (thrombus located beyond second-order ducibility of assessment of PVT.267 We found that the
branches) to Vp4 (thrombus located in the main portal remnant lumen, as a percentage of remnant portal lumen
vein),266 which can also be applied to patients with PVT largest diameter (%RL), was the most reproducible feature
without tumour invasion. (ie, better intraclass correlation coefficient, better Pearson
Accordingly, the classifications of PVT extension correlation coefficient, narrower limit of agreement on
depend on the clinical situation (appendix p 30). For the Bland Altman plot). We also found that the limits of
studies on the natural history or medical treatment of agreement of %RL were plus or minus 22%; therefore, we
PVT, the classification proposed in the AASLD215 and proposed to round this to 25% for convenience. These
Baveno VII guidelines77 should be used (panel 3, degree results allowed us to propose standardised criteria to
of agreement 96%). For studies related to liver trans­ categorise main portal vein thrombus, and its progression
plantation, Yerdel’s classification256 should be used (decrease of more than 25% of the %RL) or improvement
(panel 3, degree of agreement 80%). For studies on inter­ (increase of more than 25% of the %RL). These criteria
ventional radiology (including portal vein recanalisation, are summarised in table 1 and figure 2.
TIPS), Marot’s classification88 should be used (panel 3, Based on these findings, we proposed the following
degree of agreement 77%). statements, each reported by the degree of agreement.
Measurement of the degree of occlusion of PVT should
Classification of PVT evolution be assessed on contrast-enhanced cross-sectional imaging
Evolution of PVT encompasses the occurrence of PVT in with the portal venous phase on a plane strictly perpen­
a patient without previous PVT, or recanalisation or dicular to the main portal vein (panel 3, degree of
extension of PVT in a patient with previous PVT. There are agreement 96%).
many unanswered questions regarding clot surveillance Studies on PVT should focus on the patency of the
and techniques for monitoring therapeutic efficacy in main portal vein, given its frequent involvement and that

Main portal vein thrombus Thrombus of the upstream and downstream New thrombus Overall assessment Degree of
veins* site agreement
Complete resolution† Complete resolution† No Complete resolution 97%
Complete resolution† No complete resolution or no unequivocal No Improvement 98%
progression‡
Complete occlusion that becomes partial; Complete resolution; no unequivocal progression‡ No Improvement 94%
increase by more than 25% of the %RL§
Increase or decrease of 25% or less of the %RL§ No complete resolution or no unequivocal No Stability 96%
progression‡
Partial occlusion becoming complete; decrease Any Any Progression 96%
by more than 25% of the %RL§
Any Unequivocal progression‡ of one or more Any Progression 96%
thrombosis site
Any Any Yes Progression 93%
The proposed criteria follow the structure of the RECIST criteria used in oncology.268 The core concepts are: to assess the thrombus of the main portal vein following the
methodology presented; to assess the thrombus of the upstream and downstream veins following the methodology presented above; to assess the occurrence of new
thrombus in veins initially free from thrombus; and to combine the above assessments in order to categorize the patients into four categories: complete resolution,
improvement, stability, and progression of the thrombus. *Upstream veins refer to splenic vein, superior mesenteric vein, and inferior mesenteric vein. Downstream vein
refers to right and left portal branches. †Complete disappearance of any thrombus in the considered vein. Synonym of complete recanalisation. ‡Unequivocal decrease of the
remnant lumen in the considered vein. This feature is visually assessed. Equivocal progressions should be considered as non-progression. §%RL, percentage of the remnant
lumen according to the described methodology.

Table 1: The VALDIG PVT criteria

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 15

 Review

Diameter of the main portal vein

Diameter of the
thrombus or occlusion

(Main portal vein–thrombus)

% of remnant lumen (%RL)=
Main portal vein

Strictly perpendicular
to the main portal vein

Intrahepatic portal branches, the superior mesenteric vein, and splenic

vein are assessed separately and semi-quantitatively (no occlusion, partial Minimal Partial Complete
occlusion, or complete occlusion) %RL ≥50% %RL <50 %RL=0%

Figure 2: Methods to be used to assess portal vein thrombus on CT scan

For the portal vein: measurements are to be performed on a plane strictly perpendicular to the main portal vein on portal venous phase cross-sectional imaging;
select the image visually displaying the smaller remnant portal vein lumen; take measurement of the largest diameter of the main portal vein of this image (mm);
take measurement of the largest thrombus on the same image (mm). Compute the percentage of remnant lumen of the main portal vein (%RL), further categorised
as: minimal thrombus (%RL ≥50%), partial thrombus (%RL <50%), and complete thrombus (%RL=0%). Ideally, the same imaging modality should be used to monitor
the evolution of the thrombus. For up and downstream veins the superior mesenteric vein, splenic vein, and right and left branches are visually assessed using a
semiquantitative scale: no thrombosis (no thrombus, complete patency), incomplete thrombus (any amount of thrombus but visible remnant lumen), and complete
thrombus (no remnant lumen).

it has the most tangible haemo­ dynamic and clinical 3D portal vein reconstruction, clot volume quantification,
conseq­ uences. Measure­ ments should be performed or both, and set thresholds for changes in clot volume
after selecting the image visually displaying the smallest that should be considered clinically relevant. Examples
remnant portal vein lumen and include: the measure­ of this tech­ nology as applied to PVT are published
ment of the largest diameter of the main portal vein on elsewhere.269,270 Crucially, establishing a correlation
this image (mm), and the measurement of the largest between thrombus-based surrogate markers and
thrombus on the same image (mm), in order to compute validated clinical endpoints is needed to guide clinicians
the percentage of remnant portal lumen (%RL; figure 2). on the relevance of interventions in PVT for improved
Minimal throm­bosis or occlusion corresponds to a clinical outcomes in affected patients.
%RL equal to or greater than 50%, partial thrombosis or
occlusion corres­ponds to a %RL less than 50%, and Timing of assessment
complete thrombosis or occlusion corresponds to a In the context of PVT, repeated standardised imaging
%RL of 0% (panel 3, degree of agreement 96%). should: inform the progression or spontaneous regres­
Upstream and downstream veins, namely the superior sion of thrombosis in patients in whom a watchful waiting
mesenteric vein, splenic vein, and right and left portal approach in the context of cirrhosis has been chosen;
branches should be visually assessed using a semi- inform the response to anticoagulation (improve­ ment,
quantitative scale, as follows: no thrombosis: no throm­ stability, progression of PVT); and provide anatomical
bus or occlusion; incomplete thrombosis: any amount of details to select an appropriate treatment in case of
thrombus or occlusion but visible remnant lumen; complications (eg, variceal bleeding). Imaging protocols
complete thrombosis: no remnant lumen (panel 3, should be guided by safety and reliability. High-quality
degree of agreement 94%). data regarding the optimal timing for assessment of PVT
The VALDIG PVT criteria evaluating PVT evolution evolution in patients with PVT with or without cirrhosis
should be based on changes in the three following are lacking, and data are mainly derived from systematic
criteria: thrombus of the main portal vein, thrombus reviews of retrospective data and expert opinion (figure 3).77
of the upstream and downstream veins, and occur­
rence of new thrombus in veins initially devoid of PVT with cirrhosis
thrombus. These criteria allow for the categorisation of In patients with minimally occlusive PVT (<50%) not
the PVT into four categories: complete resolution, undergoing anticoagulation, ultrasound is usually
improvement, stability, and progression of PVT (table 1, sufficient to identify progression or regression.169 Follow-
panel 3; degree of agreement 93%). up ultrasound every 4–6 weeks in the first 3 months and
Future clinical trials should strive to standardise subsequently every 3 months within the first year might
and quantify clot volume and treatment responses, be proposed. Follow-up with contrast-enhanced imaging
specifically using modern imaging technologies such as might be considered after 6 months. Patients not

16 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

Non-cirrhotic PVT Cirrhosis

Repeat imaging if:

• Abdominal pain or diarrhoea
• Worsening of portal
hypertension or liver function

Any PVT Partially (≥50%) Minimally (<50%)

occlusive PVT occlusive
If anticoagulation
is stopped

Anticoagulation Watch and wait

At 4–6 weeks
At 3 months
At 6 months
(At 3 months) At 4–6 weeks
Every 6 months afterwards
At 6 months At 3 months
Every 6 months At 6 months
Every 4–6 weeks in candidates for liver
afterwards Every 6 months afterwards,
transplantation in whom anticoagulation has
during HCC screening
to be stopped

Figure 3: Timing of assessment in studies on PVT in patients with and without cirrhosis
After the diagnosis of recent PVT, imaging should be repeated at 6 months (with contrast-enhanced CT) and 12 months for patients in good clinical conditions,
with an additional control at 3 months commonly performed. In patients with PVT treated with anticoagulation, a control should be performed at 6 and 12 months
after starting treatment if the patient is stable. Additional imaging is commonly planned 3 months after treatment to analyse early response to anticoagulation. If a
decision to stop anticoagulation is taken, imaging controls every 3 months (every 4–6 weeks on the waiting list for liver transplantation) seem reasonable. In case of
clinical worsening, imaging should be repeated irrespective of the timing. PVT=portal vein thrombosis

progressing after 12 months might be kept under 6–12 months and a contrast-enhanced CT or MRI every
standard imaging surveillance every 6 months. 2–4 years seem reasonable.81
In patients with PVT in whom anticoagulation has To summarise, results of the Delphi consensus
been decided, follow-up imaging should be performed indicated that both for patients with and without
6 and 12 months after starting treatment if the patient is cirrhosis, endpoints for clinical studies on PVT should
in a stable condition. Additional control is commonly be assessed at imaging: for patients in a stable clinical
planned 3 months after treatment to depict early response condition: at 6 months and 12 months (panel 3, degree of
to anticoagulation. In patients with worsening clinical agreement 88%); for patients at particularly high risk of
condition (eg, worsening ascites) who are potentially recurrent PVT or progressive PVT (eg, patients on the
candidates for TIPS, imaging should be repeated orthotopic liver transplant waiting list who cannot
irrespective of the timing mentioned to identify (or receive anticoagulation): at 4–6 weeks intervals (panel 3,
exclude) thrombosis progression to the superior degree of agreement 85%); and for patients with
mesenteric vein and splenic vein. If a decision to stop worsening of signs of portal hypertension or liver
anticoagulation is taken, throm­ bosis recurrence might function: immediately, to identify progression and
occur in the first 2–5 months. Hence, imaging follow-up proceed to TIPS or other treatment options (panel 3,
every 3 months seem reasonable. In patients on the degree of agreement 98%).
waiting list for liver transplantation, given the importance
of recurrent PVT in this setting, the interval can be Global endpoints for studies on PVT
reduced to every 4–6 weeks. In patients with PVT in the absence of cirrhosis, the
main objective is to prevent the extension of the
PVT without cirrhosis thrombus to the superior mesenteric vein and,
The proposed timing does not differ from that suggested accordingly, prevent intestinal ischaemia.271 After PVT
for PVT in cirrhosis. Namely, after the diagnosis of recent occurrence, portal recanalisation is the most important
PVT, imaging should be repeated at 6 months (with thrombus-related outcome. Clinical complications,
contrast-enhanced CT) and 12 months for patients in good namely portal hypertension and portal cholangiopathy,
clinical condition, with an additional control at 3 months do not occur in patients who have recanalisation.271
commonly performed.77 In case of clinical worsening, However, recanalisation cannot be the only outcome of
imaging should be repeated irrespective of the timing. interest. First, recanalisation only occurs in around
For patients with a chronic PVT with cavernous 30–40% of patients with recent PVT in the absence of
transformation, the 12-month CT control might coincide cirrhosis treated with anticoagulation.5,82 Second, after
with an assessment of portal cholangio­ pathy using recanalisation, there is an 18–38% risk of recurrent
magnetic resonance cholangiography. In patients who thrombosis.80,272,273 Third, patients with PVT in the absence
improved or are stable on treatment, an ultrasound every of cirrhosis are at an increased risk of developing other

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 17

 Review

Thrombus outcome Degree of Clinical events Degree of

agreement agreement
Recent PVT in the absence Avoid progression and ideally promote thrombus regression 93% Portal hypertension related events; mesenteric ischaemia; major 93%
of cirrhosis and prevent extra-splanchnic venous thrombosis bleeding not caused by portal hypertension (ie, safety of
anticoagulation)
Chronic PVT in the Avoid progression and prevent extra-splanchnic venous 96% Portal hypertension related events; mesenteric ischaemia; portal 95%
absence of cirrhosis thrombosis cholangiopathy complications; health-related quality of life; major
bleeding not caused by portal hypertension (ie, safety of
anticoagulation)
Recent PVT in patients Avoid progression and ideally promote thrombus regression 96% Decompensation of cirrhosis; mesenteric ischaemia; major 95%
with cirrhosis and enable a physiological portal anastomosis at time of liver bleeding not caused by portal hypertension (ie, safety of
transplant surgery in current or future liver transplant anticoagulation)
candidates
Chronic PVT in patients Avoid progression and ideally promote thrombus regression. 96% Decompensation of cirrhosis; major bleeding not caused by portal 95%
with cirrhosis Enable a physiological portal anastomosis at time of liver hypertension (ie, safety of anticoagulation); post-transplant
transplant surgery in current or future liver transplant mortality and graft failure; health-related quality of life
candidates
PVT=portal vein thrombosis.

Table 2: Relevant endpoints for clinical studies on PVT

extrasplanchnic venous thrombosis regardless of portal To summarise, global endpoints for studies on PVT
recanalisation, especially in those with a strong risk should be different for patients with PVT in the absence
factor for thrombosis recurrence.80 versus in the presence of cirrhosis as well as for those
In patients with cirrhosis and PVT, the main objective patients with recent versus chronic PVT (panel 3, degree
of treating PVT is to recanalise the portal vein trunk and of agreement 98%). Similarly, global endpoints for
restore portal blood flow to allow end-to-end portal studies on PVT can focus on the thrombus outcome, on
anastomosis during the transplantation procedure, and clinical events, or both (table 2).
therefore prevent postoperative complications and early
graft failure.258 As PVT aggravates portal hyper­tension, Conclusion
another objective would be to decrease the incidence The growing interest in PVT within the medical
of decompensation of cirrhosis, and consequently community has led to substantial advances in knowledge,
transplant-free survival. However, whether recanalisation primarily through observational studies. These studies
directly influences liver-related outcomes has not been have shown a correlation between PVT and an elevated
shown.207 risk of complications related to portal hypertension, both
For studies involving patients with PVT, we believe in patients with and without cirrhosis. Currently,
that clinical endpoints should be: clinically relevant; anticoagulation stands as the first-line treat­ment, with
associated with thrombus regression or progression; an increasing trend toward portal vein recanalisation
meaningful to the patient; and take into account the using interventional radiology. However, the hetero­
safety of the treatment. Certainly, survival is another geneous nature of existing studies hampers the
relevant clinical outcome that can be considered in development of standardised management of patients
clinical studies on PVT in the absence of cirrhosis; with PVT. In the coming years, randomised controlled
however, due to the very low mortality rate, powering trials are expected to provide clarity on the role of
studies based solely on this outcome alone is not anticoagulation and interventional radiology in PVT
feasible. management. These studies will necessitate different­
iation between patients with cirrhosis and those without,
considering thrombus outcomes and relevant clinical
Search strategy and selection criteria events. Lastly, future clinical trials should assess the
References for this Review were identified through PubMed interest of quantifying the degree of obstruction and use
searches using the search terms “recent portal vein contemporary imaging technology to refine the
thrombosis”, “chronic portal vein thrombosis”, “portal evaluation of the degree of obstruction and treatment
cavernoma”, “extra-hepatic portal vein obstruction”, responses.
and “portal vein thrombosis and cirrhosis” from database Contributors
Conceptualisation of the manuscript: LE, PER. Data curation and
inception until Dec 31, 2023. Articles were also identified
analysis of original data in the pilot study: LE, LM, VP, MR. All authors
through searches of the authors’ own files. Only papers contributed to the writing, review, and editing of the initial draft.
published in English and French were reviewed. The final
Acknowledgments
reference list was generated based on originality and The Paris PVT meeting was supported by FILFOIE, ERN-RARE Liver,
relevance to the scope of this Review. Société Nationale Française de Gastroentérologie, and Association

18 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

Française pour l’étude du Foie. This Review was written on behalf of 17 Lisman T, Porte RJ. Pathogenesis, prevention, and management
the ERN RARE-LIVER group, and was a study supported by VALDIG, of bleeding and thrombosis in patients with liver diseases.
an EASL consortium. Res Pract Thromb Haemost 2017; 1: 150–61.
18 Violi F, Pignatelli P, Castellani V, Carnevale R, Cammisotto V. Gut
Declaration of interests dysbiosis, endotoxemia and clotting activation: a dangerous trio for
VHG has received speaker fees from Gore Medical and Cook Medical. portal vein thrombosis in cirrhosis. Blood Rev 2023; 57: 100998.
ADG has received consulting fees from Astrazeneca and Swedish 19 Praktiknjo M, Trebicka J, Carnevale R, et al. Von Willebrand
Orphan Biovitrum. MR serves as consultant from Quantum Surgical and Factor VIII portosystemic circulation gradient in cirrhosis:
(fees paid to institution) and has received speaker fees from Servier, implications for portal vein thrombosis. Clin Transl Gastroenterol
Guerbet, GE Healthcare, Ipsen, Angiodynamics, and AstraZeneca. 2020; 11: e00123.
FN has received speaker fees from Advanz and Permanyer. VML serves 20 Violi F, Ferro D, Basili S, et al. Ongoing prothrombotic state in
as consultant from AstraZaneca and has received speaker fees and travel the portal circulation of cirrhotic patients. Thromb Haemost 1997;
support from Albireo Ipsen. AB serves as consultant for Boehringer 77: 44–47.
Ingelheim and has received speaker fees from GE Healthcare and 21 Delahousse B, Labat-Debelleix V, Decalonne L, d’Alteroche L,
Hologic. AA has served as a lecturer for Boehringer-Ingelheim and Perarnau J-M, Gruel Y. Comparative study of coagulation and
Gore Medical. AN has received research funding from MSD-Avenir and thrombin generation in the portal and jugular plasma of patients
consulting fees from Abbvie and Janssen. MM serves as a lecturer for with cirrhosis. Thromb Haemost 2010; 104: 741–49.
AbbVie, Ipsen, Echosens, Gilead Sciences, and Gore Medical, and has 22 Queck A, Carnevale R, Uschner FE, et al. Role of portal venous
received travel support from AbbVie and Gilead Sciences. DP has served platelet activation in patients with decompensated cirrhosis
as lecturer for Boehringer-Ingelheim and Gore Medical. DT has received and TIPS. Gut 2020; 69: 1535–36.
research grants from Boehringer-Ingelheim and Novartis and has served 23 Driever EG, Magaz M, Adelmeijer J, et al. The portal vein in
as lecturer for Gore Medical. JCGP has received speaker fees from Gore patients with cirrhosis is not an excessively inflammatory or
Medical and research grants from Mallinkrodt and AstraZeneca. All other hypercoagulable vascular bed, a prospective cohort study.
J Thromb Haemost 2022; 20: 2075–82.
authors declare no competing interests.
24 Rahr HB, Sørensen JV, Larsen JF, Jensen FS, Bredahl C. Markers of
References coagulation and fibrinolysis in portal blood from patients with and
1 Berzigotti A, García-Criado A, Darnell A, García-Pagán J-C. without gastric malignancy. Scand J Gastroenterol 1994; 29: 516–21.
Imaging in clinical decision-making for portal vein thrombosis. 25 Violi F, Carnevale R, Pignatelli P. “The portal vein in patients with
Nat Rev Gastroenterol Hepatol 2014; 11: 308–16. cirrhosis is not an excessively inflammatory or hypercoagulable
2 De Gaetano AM, Lafortune M, Patriquin H, De Franco A, Aubin B, vascular bed, a prospective cohort study”: comment from Violi et al.
Paradis K. Cavernous transformation of the portal vein: patterns of J Thromb Haemost 2023; 21: 186–87.
intrahepatic and splanchnic collateral circulation detected with 26 Turon F, Driever EG, Baiges A, et al. Predicting portal thrombosis
Doppler sonography. AJR Am J Roentgenol 1995; 165: 1151–55. in cirrhosis: a prospective study of clinical, ultrasonographic
3 Jha RC, Khera SS, Kalaria AD. Portal vein thrombosis: imaging and hemostatic factors. J Hepatol 2021; 75: 1367–76.
the spectrum of disease with an emphasis on MRI features. 27 Hernández-Gea V, De Gottardi A, Leebeek FWG, Rautou P-E,
AJR Am J Roentgenol 2018; 211: 14–24. Salem R, Garcia-Pagan JC. Current knowledge in pathophysiology
4 Copin P, Ronot M, Nuzzo A, et al. Inter-reader agreement of and management of Budd-Chiari syndrome and non-cirrhotic
CT features of acute mesenteric ischemia. Eur J Radiol 2018; non-tumoral splanchnic vein thrombosis. J Hepatol 2019; 71: 175–99.
105: 87–95. 28 Bureau C, Laurent J, Robic MA, et al. Central obesity is associated
5 Plessier A, Darwish-Murad S, Hernandez-Guerra M, et al. Acute with non-cirrhotic portal vein thrombosis. J Hepatol 2016;
portal vein thrombosis unrelated to cirrhosis: a prospective 64: 427–32.
multicenter follow-up study. Hepatology 2010; 51: 210–18. 29 Zheng X, Gao F, Wang L, Meng Y, Ageno W, Qi X. Incidence
6 Senzolo M, Garcia-Tsao G, García-Pagán JC. Current knowledge and outcomes of splanchnic vein thrombosis after diagnosis of
and management of portal vein thrombosis in cirrhosis. J Hepatol COVID-19 or COVID-19 vaccination: a systematic review and
2021; 75: 442–53. meta-analysis. J Thromb Thrombolysis 2023; 55: 18–31.
7 Carneiro C, Brito J, Bilreiro C, et al. All about portal vein: a pictorial 30 Deltenre P, Payancé A, Elkrief L, et al. Splanchnic vein thrombosis
display to anatomy, variants and physiopathology. Insights Imaging associated with SARS-CoV-2 infection: a VALDIG case-control
2019; 10: 38. study. JHEP Rep Innov Hepatol 2023; 5: 100894.
8 Vink CS, Mariani SA, Dzierzak E. Embryonic origins of the 31 Maan R, Lauw MN, China L, et al. Extensive splanchnic vein
hematopoietic system: hierarchies and heterogeneity. HemaSphere thrombosis after SARS-CoV-2 vaccination: a Vascular Liver Disease
2022; 6: e737. Group (VALDIG) initiative. Hepatology 2024; published online
9 Su T, Yang Y, Lai S, et al. Single-cell transcriptomics reveals zone- Feb 15. https://doi.org/10.1097/HEP.0000000000000787.
specific alterations of liver sinusoidal endothelial cells in cirrhosis. 32 De Broucker C, Plessier A, Ollivier-Hourmand I, et al. Multicenter
Cell Mol Gastroenterol Hepatol 2021; 11: 1139–61. study on recent portal venous system thrombosis associated with
10 Zhang H, Ma Y, Cheng X, et al. Targeting epigenetically maladapted cytomegalovirus disease. J Hepatol 2022; 76: 115–22.
vascular niche alleviates liver fibrosis in nonalcoholic steatohepatitis. 33 Khetarpal SA, Qamar A, Bick AG, et al. Clonal hematopoiesis of
Sci Transl Med 2021; 13: eabd1206. indeterminate potential reshapes age-related CVD: JACC review
11 Bondareva O, Rodríguez-Aguilera JR, Oliveira F, et al. Single-cell topic of the week. J Am Coll Cardiol 2019; 74: 578–86.
profiling of vascular endothelial cells reveals progressive organ- 34 Guy A, Poisson J, James C. Pathogenesis of cardiovascular events in
specific vulnerabilities during obesity. Nat Metab 2022; 4: 1591–610. BCR-ABL1-negative myeloproliferative neoplasms. Leukemia 2021;
12 Trimm E, Red-Horse K. Vascular endothelial cell development and 35: 935–55.
diversity. Nat Rev Cardiol 2023; 20: 197–210. 35 Hermange G, Rakotonirainy A, Bentriou M, et al. Inferring the
13 Wilson JB. Changes in the portal and splenic veins in portal initiation and development of myeloproliferative neoplasms.
hypertension and their relation to splenomegaly. Gut 1961; 2: 310–15. Proc Natl Acad Sci USA 2022; 119: e2120374119.
14 Malmqvist U, Arner A. Isoform distribution and tissue contents 36 Debureaux P-E, Cassinat B, Soret-Dulphy J, et al. Molecular
of contractile and cytoskeletal proteins in hypertrophied smooth profiling and risk classification of patients with myeloproliferative
muscle from rat portal vein. Circ Res 1990; 66: 832–45. neoplasms and splanchnic vein thromboses. Blood Adv 2020;
15 Albinsson S, Nordström I, Hellstrand P. Stretch of the vascular 4: 3708–15.
wall induces smooth muscle differentiation by promoting actin 37 Magaz M, Alvarez-Larrán A, Colomer D, et al. Next-generation
polymerization. J Biol Chem 2004; 279: 34849–55. sequencing in the diagnosis of non-cirrhotic splanchnic vein
16 Driever EG, von Meijenfeldt FA, Adelmeijer J, et al. Nonmalignant thrombosis. J Hepatol 2021; 74: 89–95.
portal vein thrombi in patients with cirrhosis consist of intimal 38 Iwakiri Y, Trebicka J. Portal hypertension in cirrhosis:
fibrosis with or without a fibrin-rich thrombus. Hepatology 2022; pathophysiological mechanisms and therapy. JHEP Rep Innov Hepatol
75: 898–911. 2021; 3: 100316.

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 19

 Review

39 Albillos A, Colombato LA, Groszmann RJ. Vasodilatation and 64 Fruchtman SM, Mack K, Kaplan ME, Peterson P, Berk PD,
sodium retention in prehepatic portal hypertension. Gastroenterology Wasserman LR. From efficacy to safety: a Polycythemia Vera Study
1992; 102: 931–35. group report on hydroxyurea in patients with polycythemia vera.
40 Sikuler E, Kravetz D, Groszmann RJ. Evolution of portal hypertension Semin Hematol 1997; 34: 17–23.
and mechanisms involved in its maintenance in a rat model. 65 Marchioli R, Finazzi G, Specchia G, et al. Cardiovascular events
Am J Physiol 1985; 248: G618–25. and intensity of treatment in polycythemia vera. N Engl J Med 2013;
41 Lebrec D, Bataille C, Bercoff E, Valla D. Hemodynamic changes in 368: 22–33.
patients with portal venous obstruction. Hepatology 1983; 3: 550–53. 66 Masciulli A, Ferrari A, Carobbio A, Ghirardi A, Barbui T. Ruxolitinib
42 Braillon A, Moreau R, Hadengue A, Roulot D, Sayegh R, Lebrec D. for the prevention of thrombosis in polycythemia vera: a systematic
Hyperkinetic circulatory syndrome in patients with presinusoidal review and meta-analysis. Blood Adv 2020; 4: 380–86.
portal hypertension. Effect of propranolol. J Hepatol 1989; 9: 312–18. 67 Cortelazzo S, Finazzi G, Ruggeri M, et al. Hydroxyurea for patients
43 Seijo S, Reverter E, Miquel R, et al. Role of hepatic vein with essential thrombocythemia and a high risk of thrombosis.
catheterisation and transient elastography in the diagnosis of N Engl J Med 1995; 332: 1132–36.
idiopathic portal hypertension. Dig Liver Dis 2012; 44: 855–60. 68 Tremblay D, Kosiorek HE, Dueck AC, Hoffman R. Evaluation of
44 Alvarado E, Garcia-Guix M, Mirabet S, Villanueva C. The relationship therapeutic strategies to reduce the number of thrombotic events in
of hyperdynamic circulation and cardiodynamic states in cirrhosis. patients with polycythemia vera and essential thrombocythemia.
J Hepatol 2018; 69: 746–47. Front Oncol 2021; 10: 636675.
45 Feu F, Bordas JM, Luca A, et al. Reduction of variceal pressure by 69 De Stefano V, Rossi E, Carobbio A, et al. Hydroxyurea prevents
propranolol: comparison of the effects on portal pressure and arterial and late venous thrombotic recurrences in patients with
azygos blood flow in patients with cirrhosis. Hepatology 1993; myeloproliferative neoplasms but fails in the splanchnic venous
18: 1082–89. district. Pooled analysis of 1500 cases. Blood Cancer J 2018; 8: 112.
46 Villanueva C, Albillos A, Genescà J, et al. Development of 70 Hamulyák EN, Daams JG, Leebeek FWG, et al. A systematic review
hyperdynamic circulation and response to β-blockers in of antithrombotic treatment of venous thromboembolism in
compensated cirrhosis with portal hypertension. Hepatology 2016; patients with myeloproliferative neoplasms. Blood Adv 2021;
63: 197–206. 5: 113–21.
47 Bañares R, Moitinho E, Piqueras B, et al. Carvedilol, a new 71 Mascarenhas J, Kosiorek H, Prchal J, et al. A prospective evaluation
nonselective beta-blocker with intrinsic anti-alpha1-adrenergic of pegylated interferon alfa-2a therapy in patients with polycythemia
activity, has a greater portal hypotensive effect than propranolol in vera and essential thrombocythemia with a prior splanchnic vein
patients with cirrhosis. Hepatology 1999; 30: 79–83. thrombosis. Leukemia 2019; 33: 2974–78.
48 Sharma P, Mishra SR, Kumar M, Sharma BC, Sarin SK. Liver 72 Pieri L, Paoli C, Arena U, et al. Safety and efficacy of ruxolitinib in
and spleen stiffness in patients with extrahepatic portal vein splanchnic vein thrombosis associated with myeloproliferative
obstruction. Radiology 2012; 263: 893–99. neoplasms. Am J Hematol 2017; 92: 187–95.
49 Shen Y, Ma W, Hang Y, Liu L-L, Jiang W, Wu S-D. Clinical 73 Plessier A, Esposito-Farèse M, Baiges A, et al. Paroxysmal nocturnal
application of liver stiffness measurement in patients with hemoglobinuria and vascular liver disease: eculizumab therapy
cavernous transformation of portal vein. Exp Ther Med 2021; 21: 442. decreases mortality and thrombotic complications. Am J Hematol
50 Madhusudhan KS, Kilambi R, Shalimar, et al. Measurement of 2022; 97: 431–39.
splenic stiffness by 2D-shear wave elastography in patients with 74 Gala D, Newsome T, Roberson N, et al. Thromboembolic events in
extrahepatic portal vein obstruction. Br J Radiol 2018; 91: 20180401. patients with inflammatory bowel disease: a comprehensive
51 Chatelin S, Pop R, Giraudeau C, et al. Influence of portal vein overview. Diseases 2022; 10: 73.
occlusion on portal flow and liver elasticity in an animal model. 75 Alibaz-Oner F, Karadeniz A, Ylmaz S, et al. Behçet disease with
NMR Biomed 2021; 34: e4498. vascular involvement: effects of different therapeutic regimens on
52 Sharma P, Agarwal R, Dhawan S, et al. Transient elastography the incidence of new relapses. Medicine 2015; 94: e494.
(fibroscan) in patients with non-cirrhotic portal fibrosis. 76 Desbois AC, Rautou PE, Biard L, et al. Behcet’s disease in
J Clin Exp Hepatol 2017; 7: 230–34. Budd-Chiari syndrome. Orphanet J Rare Dis 2014; 9: 104.
53 Elkrief L, Lazareth M, Chevret S, et al. Liver stiffness by transient 77 de Franchis R, Bosch J, Garcia-Tsao G, Reiberger T, Ripoll C,
elastography to detect porto-sinusoidal vascular liver disease with Baveno V. II Faculty. Baveno VII—renewing consensus in portal
portal hypertension. Hepatology 2021; 74: 364–78. hypertension. J Hepatol 2022; 76: 959–74.
54 Ito K, Mitchell DG, Siegelman ES. Cirrhosis: MR imaging features. 78 Randi ML, Tezza F, Scapin M, et al. Heparin-induced
Magn Reson Imaging Clin N Am 2002; 10: 75–92. thrombocytopenia in patients with Philadelphia-negative
55 Vilgrain V, Lagadec M, Ronot M. Pitfalls in liver imaging. Radiology myeloproliferative disorders and unusual splanchnic or cerebral
2016; 278: 34–51. vein thrombosis. Acta Haematol 2010; 123: 140–45.
56 Vilgrain V, Condat B, Bureau C, et al. Atrophy-hypertrophy complex 79 Naymagon L, Tremblay D, Zubizarreta N, et al. The efficacy and
in patients with cavernous transformation of the portal vein: safety of direct oral anticoagulants in noncirrhotic portal vein
CT evaluation. Radiology 2006; 241: 149–55. thrombosis. Blood Adv 2020; 4: 655–66.
57 Gupta P, Kalra N, Gulati A, et al. Changes in liver morphology in 80 Plessier A, Goria O, Cervoni JP, et al. Rivaroxaban prophylaxis in
patients with extrahepatic portal venous obstruction: a retrospective noncirrhotic portal vein thrombosis. NEJM Evid 2022; 1: a2200104.
magnetic resonance imaging study. Br J Radiol 2019; 92: 20180890. 81 Baiges A, Procopet B, Silva-Junior G, et al. Incidence and factors
58 De Gottardi A, Sempoux C, Berzigotti A. Porto-sinusoidal vascular predictive of recurrent thrombosis in people with non-cirrhotic
disorder. J Hepatol 2022; 77: 1124–35. portal vein thrombosis. J Hepatol 2023; 78: 114–22.
59 Glatard A-S, Hillaire S, d’Assignies G, et al. Obliterative portal 82 Turnes J, García-Pagán JC, González M, et al. Portal hypertension-
venopathy: findings at CT imaging. Radiology 2012; 263: 741–50. related complications after acute portal vein thrombosis: impact
of early anticoagulation. Clin Gastroenterol Hepatol 2008;
60 Kang JH, Kim DH, Kim SY, et al. Porto-sinusoidal vascular disease
6: 1412–17.
with portal hypertension versus liver cirrhosis: differences in
imaging features on CT and hepatobiliary contrast-enhanced MRI. 83 Noronha Ferreira C, Seijo S, Plessier A, et al. Natural history and
Abdom Radiol 2021; 46: 1891–903. management of esophagogastric varices in chronic noncirrhotic,
nontumoral portal vein thrombosis. Hepatology 2016; 63: 1640–50.
61 Lampichler K, Semmler G, Wöran K, et al. Imaging features
facilitate diagnosis of porto-sinusoidal vascular disorder. Eur Radiol 84 Spaander MCW, Hoekstra J, Hansen BE, Van Buuren HR,
2023; 33: 1422–32. Leebeek FWG, Janssen HLA. Anticoagulant therapy in patients
with non-cirrhotic portal vein thrombosis: effect on new thrombotic
62 Valainathan SR, Sartoris R, Elkrief L, et al. Contrast-enhanced CT
events and gastrointestinal bleeding. J Thromb Haemost 2013;
and liver surface nodularity for the diagnosis of porto-sinusoidal
11: 452–59.
vascular disorder: a case-control study. Hepatology 2022; 76: 418–28.
85 Lattanzi B, Gioia S, Di Cola S, et al. Prevalence and impact of
63 Barbui T, Vannucchi AM, Finazzi G, et al. A reappraisal of the
sarcopenia in non-cirrhotic portal hypertension. Liver Int 2019;
benefit-risk profile of hydroxyurea in polycythemia vera:
39: 1937–42.
a propensity-matched study. Am J Hematol 2017; 92: 1131–36.

20 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

86 Guillaume M, Christol C, Plessier A, et al. Bleeding risk of variceal 109 Bala M, Kashuk J, Moore EE, et al. Acute mesenteric ischemia:
band ligation in extrahepatic portal vein obstruction is not increased guidelines of the World Society of Emergency Surgery.
by oral anticoagulation. Eur J Gastroenterol Hepatol 2018; 30: 563–68. World J Emerg Surg 2017; 12: 38.
87 Artru F, Vietti-Violi N, Sempoux C, et al. Portal vein recanalisation 110 Corcos O, Castier Y, Sibert A, et al. Effects of a multimodal
alone to treat severe portal hypertension in non-cirrhotic patients management strategy for acute mesenteric ischemia on survival
with chronic extrahepatic portal vein obstruction. and intestinal failure. Clin Gastroenterol Hepatol 2013; 11: 158–65.
JHEP Rep Innov Hepatol 2022; 4: 100511. 111 Clair DG, Beach JM. Mesenteric ischemia. N Engl J Med 2016;
88 Marot A, Barbosa JV, Duran R, Deltenre P, Denys A. Percutaneous 374: 959–68.
portal vein recanalization using self-expandable nitinol stents in 112 Nuzzo A, Maggiori L, Ronot M, et al. Predictive factors of intestinal
patients with non-cirrhotic non-tumoral portal vein occlusion. necrosis in acute mesenteric ischemia: prospective study from an
Diagn Interv Imaging 2019; 100: 147–56. intestinal stroke center. Am J Gastroenterol 2017; 112: 597–605.
89 Fanelli F, Angeloni S, Salvatori FM, et al. Transjugular intrahepatic 113 Nuzzo A, Maggiori L, Paugam-Burtz C, et al. Oral antibiotics reduce
portosystemic shunt with expanded-polytetrafuoroethylene-covered intestinal necrosis in acute mesenteric ischemia: a prospective
stents in non-cirrhotic patients with portal cavernoma. Dig Liver Dis cohort study. Am J Gastroenterol 2019; 114: 348–51.
2011; 43: 78–84. 114 Gao F, Wang L, Pan J, et al. Efficacy and safety of thrombolytic
90 Knight GM, Clark J, Boike JR, et al. TIPS for adults without therapy for portal venous system thrombosis: a systematic review
cirrhosis with chronic mesenteric venous thrombosis and and meta-analysis. J Intern Med 2023; 293: 212–27.
EHPVO refractory to standard-of-care therapy. Hepatology 2021; 115 Di Nisio M, Valeriani E, Riva N, Schulman S, Beyer-Westendorf J,
74: 2735–44. Ageno W. Anticoagulant therapy for splanchnic vein thrombosis:
91 Wei B, Huang Z, Wu H, et al. Portal vein recanalization for ISTH SSC subcommittee control of anticoagulation.
noncirrhotic portal vein cavernous transformation: transjugular J Thromb Haemost 2020; 18: 1562–68.
intrahepatic portosystemic shunt creation versus portal vein stent 116 Senzolo M, Tibbals J, Cholongitas E, Triantos CK, Burroughs AK,
placement. J Vasc Interv Radiol 2023; 34: 187–94. Patch D. Transjugular intrahepatic portosystemic shunt for portal
92 Chandra R, Kapoor D, Tharakan A, Chaudhary A, Sarin SK. vein thrombosis with and without cavernous transformation.
Portal biliopathy. J Gastroenterol Hepatol 2001; 16: 1086–92. Aliment Pharmacol Ther 2006; 23: 767–75.
93 Dilawari JB, Chawla YK. Pseudosclerosing cholangitis in 117 Rössle M, Bettinger D, Trebicka J, et al. A prospective, multicentre
extrahepatic portal venous obstruction. Gut 1992; 33: 272–76. study in acute non-cirrhotic, non-malignant portal vein thrombosis:
94 Llop E, de Juan C, Seijo S, et al. Portal cholangiopathy: radiological comparison of medical and interventional treatment.
classification and natural history. Gut 2011; 60: 853–60. Aliment Pharmacol Ther 2020; 52: 329–39.
95 Condat B, Vilgrain V, Asselah T, et al. Portal cavernoma-associated 118 Benmassaoud A, AlRubaiy L, Yu D, et al. A stepwise thrombolysis
cholangiopathy: a clinical and MR cholangiography coupled with regimen in the management of acute portal vein thrombosis in
MR portography imaging study. Hepatology 2003; 37: 1302–08. patients with evidence of intestinal ischaemia.
96 Shukla A, Gupte A, Karvir V, Dhore P, Bhatia S. Long term Aliment Pharmacol Ther 2019; 50: 1049–58.
outcomes of patients with significant biliary obstruction due to 119 Eugène C, Valla D, Wesenfelder L, et al. Small intestinal stricture
portal cavernoma cholangiopathy and extra-hepatic portal vein complicating superior mesenteric vein thrombosis. A study of
obstruction (EHPVO) with no shuntable veins. J Clin Exp Hepatol three cases. Gut 1995; 37: 292–95.
2017; 7: 328–33. 120 Ribera J, Pauta M, Melgar-Lesmes P, et al. Increased nitric oxide
97 Bhatia V. Endoscopic retrograde cholangiography in portal production in lymphatic endothelial cells causes impairment of
cavernoma cholangiopathy—results from different studies lymphatic drainage in cirrhotic rats. Gut 2013; 62: 138–45.
and proposal for uniform terminology. J Clin Exp Hepatol 2014; 121 Milazzo L, Peri AM, Lodi L, Gubertini G, Ridolfo AL, Antinori S.
4 (suppl 1): s37–43. Intestinal lymphangiectasia and reversible high liver stiffness.
98 Kumar M, Saraswat VA. Natural history of portal cavernoma Hepatology 2014; 60: 759–61.
cholangiopathy. J Clin Exp Hepatol 2014; 4 (suppl 1): s62–66. 122 Naymagon L, Tremblay D, Zubizarreta N, Moshier E, Schiano T,
99 Oo YH, Olliff S, Haydon G, Thorburn D. Symptomatic portal Mascarenhas J. Portal vein thrombosis patients harboring
biliopathy: a single centre experience from the UK. JAK2V617F have poor long-term outcomes despite anticoagulation.
Eur J Gastroenterol Hepatol 2009; 21: 206–13. J Thromb Thrombolysis 2020; 50: 652–60.
100 Khare R, Sikora SS, Srikanth G, et al. Extrahepatic portal venous 123 Hollingshead M, Burke CT, Mauro MA, Weeks SM, Dixon RG,
obstruction and obstructive jaundice: approach to management. Jaques PF. Transcatheter thrombolytic therapy for acute mesenteric
J Gastroenterol Hepatol 2005; 20: 56–61. and portal vein thrombosis. J Vasc Interv Radiol 2005; 16: 651–61.
101 Costamagna G, Pandolfi M, Mutignani M, Spada C, Perri V. 124 Smalberg JH, Spaander MVMCW, Jie K-SG, et al. Risks and
Long-term results of endoscopic management of postoperative benefits of transcatheter thrombolytic therapy in patients with
bile duct strictures with increasing numbers of stents. splanchnic venous thrombosis. Thromb Haemost 2008;
Gastrointest Endosc 2001; 54: 162–68. 100: 1084–88.
102 Chattopadhyay S, Govindasamy M, Singla P, et al. Portal biliopathy 125 Cao G, Ko G-Y, Sung K-B, Yoon H-K, Gwon DI, Kim J-H. Treatment
in patients with non-cirrhotic portal hypertension: does the type of of postoperative main portal vein and superior mesenteric vein
surgery affect outcome? HPB 2012; 14: 441–47. thrombosis with balloon angioplasty and/or stent placement.
103 Agarwal AK, Sharma D, Singh S, Agarwal S, Girish SP. Portal Acta Radiol 2013; 54: 526–32.
biliopathy: a study of 39 surgically treated patients. HPB 2011; 126 Rosenqvist K, Eriksson L-G, Rorsman F, Sangfelt P, Nyman R.
13: 33–39. Endovascular treatment of acute and chronic portal vein thrombosis
104 Acosta S. Epidemiology of mesenteric vascular disease: clinical in patients with cirrhotic and non-cirrhotic liver. Acta Radiol 2016;
implications. Semin Vasc Surg 2010; 23: 4–8. 57: 572–79.
105 Elkrief L, Corcos O, Bruno O, et al. Type 2 diabetes mellitus as a 127 Klinger C, Riecken B, Schmidt A, et al. Transjugular portal vein
risk factor for intestinal resection in patients with superior recanalization with creation of intrahepatic portosystemic shunt
mesenteric vein thrombosis. Liver Int 2014; 34: 1314–21. (PVR-TIPS) in patients with chronic non-cirrhotic, non-malignant
106 Schoots IG, Koffeman GI, Legemate DA, Levi M, van Gulik TM. portal vein thrombosis. Z Gastroenterol 2018; 56: 221–37.
Systematic review of survival after acute mesenteric ischaemia 128 Wolter K, Decker G, Kuetting D, et al. Interventional treatment of
according to disease aetiology. Br J Surg 2004; 91: 17–27. acute portal vein thrombosis. Rofo 2018; 190: 740–46.
107 Wadman M, Block T, Ekberg O, Syk I, Elmståhl S, Acosta S. Impact 129 Gerwing M, Wilms C, Heinzow H, et al. Escalating interventional
of MDCT with intravenous contrast on the survival in patients with recanalization therapy in non-cirrhotic, non-malignant acute
acute superior mesenteric artery occlusion. Emerg Radiol 2010; portal vein thrombosis. Eur J Gastroenterol Hepatol 2019; 31: 1584–91.
17: 171–78. 130 Mansour N, Öcal O, Gerwing M, et al. Interventional recanalization
108 Nuzzo A, Joly F, Ronot M, et al. Normal lactate and unenhanced therapy in patients with non-cirrhotic, non-malignant portal vein
CT-scan result in delayed diagnosis of acute mesenteric ischemia. thrombosis: comparison between transjugular versus transhepatic
Am J Gastroenterol 2020; 115: 1902–05. access. Abdom Radiol (NY) 2022; 47: 1177–86.

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 21

 Review

131 Saito H, Sugihara F, Ueda T, et al. Efficacy of endovascular 154 Hoekstra J, Seijo S, Rautou PE, et al. Pregnancy in women with
treatment for completely occlusive acute-subacute portal and portal vein thrombosis: results of a multicentric European study on
mesenteric vein thrombosis with severe complications in patients maternal and fetal management and outcome. J Hepatol 2012;
without cirrhosis. Jpn J Radiol 2023; 41: 541–50. 57: 1214–19.
132 Luo J, Li M, Zhang Y, et al. Percutaneous transhepatic intrahepatic 155 Aggarwal N, Chopra S, Raveendran A, Suri V, Dhiman RK,
portosystemic shunt for variceal bleeding with chronic portal vein Chawla YK. Extra hepatic portal vein obstruction and pregnancy
occlusion after splenectomy. Eur Radiol 2018; 28: 3661–68. outcome: largest reported experience. J Obstet Gynaecol Res 2011;
133 Entezari P, Riaz A, Thornburg B, Salem R. Percutaneous 37: 575–80.
ultrasound-guided superior and inferior mesenteric vein access for 156 Shukla A, Singh A, Saxena A, et al. Navigating portal hypertension:
portal vein recanalization-transjugular intrahepatic portosystemic unlocking safe passage to healthy pregnancy in EHPVO. Liver Int
shunt: a case series. Cardiovasc Intervent Radiol 2021; 44: 496–99. 2024; 44: 454–59.
134 Habib A, Desai K, Hickey R, et al. Portal vein recanalization- 157 Payancé A, Ceccaldi PF, De Raucourt E, et al. Pregnancy and
transjugularintrahepatic portosystemic shunt using the vascular liver diseases: vascular liver diseases: position papers from
transsplenic approach to achieve transplant candidacy in patients the francophone network for vascular liver diseases, the French
with chronic portal vein thrombosis. J Vasc Interv Radiol 2015; Association for the Study of the Liver (AFEF), and ERN-rare liver.
26: 499–506. Clin Res Hepatol Gastroenterol 2020; 44: 433–37.
135 Semiz-Oysu A, Keussen I, Cwikiel W. Interventional radiological 158 Sarin SK, Bansal A, Sasan S, Nigam A. Portal-vein obstruction in
management of prehepatic obstruction of [corrected] the splanchnic children leads to growth retardation. Hepatology 1992; 15: 229–33.
venous system. Cardiovasc Intervent Radiol 2007; 30: 688–95. 159 Bersani I, Piersigilli F, Iacona G, et al. Incidence of umbilical vein
136 Woodrum DA, Bjarnason H, Andrews JC. Portal vein venoplasty catheter-associated thrombosis of the portal system: a systematic
and stent placement in the nontransplant population. review and meta-analysis. World J Hepatol 2021; 13: 1802–15.
J Vasc Interv Radiol 2009; 20: 593–99. 160 Williams S, Chan AKC. Neonatal portal vein thrombosis: diagnosis
137 Liu J, Zhou C, Shi Q, et al. Exploration of interventional therapy and management. Semin Fetal Neonatal Med 2011; 16: 329–39.
strategy for portal vein occlusion: a case series study. 161 Koo KSH, Lamar DL, Shaw DWW, Monroe EJ, Shivaram GM.
Eur J Gastroenterol Hepatol 2020; 32: 507–16. Catheter-directed thrombolysis for portal vein thrombosis in
138 Kobe A, Puippe G, Müllhaupt B, Pfammatter T. Recanalization children: a case series. J Vasc Interv Radiol 2018; 29: 1578–83.
of chronic noncirrhotic, nonmalignant splanchnic thromboses 162 Superina RA, de Ville de Goyet J. Pre-emptive Meso-Rex bypass for
is feasible: a transsplenic assisted patient-tailored approach. children with idiopathic prehepatic portal hypertension: trick or
J Vasc Interv Radiol 2021; 32: 1377–85. treat? J Pediatr Gastroenterol Nutr 2014; 58: e41.
139 Smith ML, Wade JB, Wolstenholme J, Bajaj JS. Gut microbiome- 163 Mack CL, Zelko FA, Lokar J, et al. Surgically restoring portal blood
brain-cirrhosis axis. Hepatology 2023; published online March 6. flow to the liver in children with primary extrahepatic portal vein
https://doi.org/10.1097/HEP.0000000000000344. thrombosis improves fluid neurocognitive ability. Pediatrics 2006;
140 Carol M, Pérez-Guasch M, Solà E, et al. Stigmatization is common 117: e405–12.
in patients with non-alcoholic fatty liver disease and correlates with 164 Bertocchini A, Falappa P, Grimaldi C, Bolla G, Monti L,
quality of life. PLoS One 2022; 17: e0265153. de Ville de Goyet J. Intrahepatic portal venous systems in children
141 Schomerus G, Leonhard A, Manthey J, et al. The stigma of alcohol- with noncirrhotic prehepatic portal hypertension: anatomy and
related liver disease and its impact on healthcare. J Hepatol 2022; clinical relevance. J Pediatr Surg 2014; 49: 1268–75.
77: 516–24. 165 Superina RA, Alonso EM. Medical and surgical management of
142 Shiha G, Korenjak M, Casanovas T, et al. MAFLD 2022: portal hypertension in children. Curr Treat Options Gastroenterol
an ELPA/ALPA/EASO-ECPO joint statement on disease stigma. 2006; 9: 432–43.
J Hepatol 2022; 77: 1717–19. 166 Chiu B, Superina RA. Encephalopathy caused by a splenorenal
143 Tu T. Stigma: a major barrier to hepatitis B elimination. shunt can be reversed by performing a mesenteric-to-left portal vein
Nat Rev Gastroenterol Hepatol 2022; 19: 622. bypass. J Pediatr Surg 2006; 41: 1177–79.
144 Gairing SJ, Galle PR, Schattenberg JM, Kostev K, Labenz C. Portal 167 Pan J, Wang L, Gao F, et al. Epidemiology of portal vein thrombosis
vein thrombosis is associated with an increased incidence of in liver cirrhosis: a systematic review and meta-analysis.
depression and anxiety disorders. J Clin Med 2021; 10: 5689. Eur J Intern Med 2022; 104: 21–32.
145 Battaglioli T, Martinelli I. Hormone therapy and thromboembolic 168 Bert J, Geerts A, Vanlander A, et al. Up to 50% of portal vein
disease. Curr Opin Hematol 2007; 14: 488–93. thrombosis remains undiagnosed until liver transplantation.
146 Marik PE, Plante LA. Venous thromboembolic disease and Clin Transplant 2020; 34: e14107.
pregnancy. N Engl J Med 2008; 359: 2025–33. 169 Nery F, Chevret S, Condat B, et al. Causes and consequences of
147 Pabinger I, Grafenhofer H. Thrombosis during pregnancy: risk portal vein thrombosis in 1243 patients with cirrhosis: results
factors, diagnosis and treatment. Pathophysiol Haemost Thromb of a longitudinal study. Hepatology 2015; 61: 660–67.
2002; 32: 322–24. 170 Zocco MA, Di Stasio E, De Cristofaro R, et al. Thrombotic risk
148 Bates SM, Greer IA, Middeldorp S, Veenstra DL, Prabulos A-M, factors in patients with liver cirrhosis: correlation with MELD
Vandvik PO. VTE, thrombophilia, antithrombotic therapy, and scoring system and portal vein thrombosis development. J Hepatol
pregnancy: antithrombotic therapy and prevention of thrombosis, 2009; 51: 682–89.
9th ed: American College of Chest Physicians evidence-based 171 Nery F, Correia S, Macedo C, et al. Nonselective beta-blockers and
clinical practice guidelines. Chest 2012; 141 (suppl): e691S–e736S. the risk of portal vein thrombosis in patients with cirrhosis: results
149 Wiegers H, Hamulyák EN, Damhuis SE, et al. Pregnancy of a prospective longitudinal study. Aliment Pharmacol Ther 2019;
outcomes in women with Budd–Chiari syndrome or portal vein 49: 582–88.
thrombosis—a multicentre retrospective cohort study. BJOG 2022; 172 Xu X, Guo X, De Stefano V, et al. Nonselective beta-blockers
129: 608–17. and development of portal vein thrombosis in liver cirrhosis:
150 Janssen HL, Wijnhoud A, Haagsma EB, et al. Extrahepatic portal a systematic review and meta-analysis. Hepatol Int 2019; 13: 468–81.
vein thrombosis: aetiology and determinants of survival. Gut 2001; 173 Xing Y, Tian Z, Jiang Y, et al. A practical nomogram based on
49: 720–24. systemic inflammatory markers for predicting portal vein
151 Mandal D, Dattaray C, Sarkar R, Mandal S, Choudhary A, Maity TK. thrombosis in patients with liver cirrhosis. Ann Med 2022;
Is pregnancy safe with extrahepatic portal vein obstruction? 54: 302–09.
An analysis. Singapore Med J 2012; 53: 676–80. 174 Nery F, Carneiro P, Correia S, et al. Systemic inflammation as a risk
152 Shukla A, Sadalage A, Gupta D, et al. Pregnancy outcomes in factor for portal vein thrombosis in cirrhosis: a prospective
women with Budd–Chiari syndrome before onset of symptoms longitudinal study. Eur J Gastroenterol Hepatol 2021;
and after treatment. Liver Int 2018; 38: 754–59. 33 (suppl 1): e108–13.
153 Ramani N, Kawli K, Karad A, et al. Gonadal dysfunction in male 175 Huang X, Fan X, Zhang R, Jiang S, Yang K, Chen S. Systemic
patients with Budd–Chiari syndrome and its reversibility with inflammation and portal vein thrombosis in cirrhotic patients with
treatment. Hepatol Int 2022; 16: 640–48. gastroesophageal varices. Eur J Gastroenterol Hepatol 2020; 32: 401–05.

22 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

176 Qi X, Guo X, Yoshida EM, et al. Transient portal vein thrombosis in 197 Dell’Era A, Iannuzzi F, Fabris FM, et al. Impact of portal vein
liver cirrhosis. BMC Med 2018; 16: 83. thrombosis on the efficacy of endoscopic variceal band ligation.
177 Francoz C, Belghiti J, Vilgrain V, et al. Splanchnic vein thrombosis Dig Liver Dis 2014; 46: 152–56.
in candidates for liver transplantation: usefulness of screening and 198 Lv Y, Qi X, He C, et al. Covered TIPS versus endoscopic band
anticoagulation. Gut 2005; 54: 691–97. ligation plus propranolol for the prevention of variceal rebleeding
178 Molinari M, Fernandez-Carrillo C, Dai D, et al. Portal vein in cirrhotic patients with portal vein thrombosis: a randomised
thrombosis and renal dysfunction: a national comparative study of controlled trial. Gut 2018; 67: 2156–68.
liver transplant recipients for NAFLD versus alcoholic cirrhosis. 199 Nicoară-Farcău O, Han G, Rudler M, et al. Effects of early
Transpl Int 2021; 34: 1105–22. placement of transjugular portosystemic shunts in patients with
179 Qi X, Ren W, De Stefano V, Fan D. Associations of coagulation factor high-risk acute variceal bleeding: a meta-analysis of individual
V Leiden and prothrombin G20210A mutations with Budd–Chiari patient data. Gastroenterology 2021; 160: 193–205.
syndrome and portal vein thrombosis: a systematic review and 200 Loffredo L, Pastori D, Farcomeni A, Violi F. Effects of
meta-analysis. Clin Gastroenterol Hepatol 2014; 12: 1801–12. anticoagulants in patients with cirrhosis and portal vein thrombosis:
180 Scheiner B, Northup PG, Gruber AB, et al. The impact of ABO a systematic review and meta-analysis. Gastroenterology 2017;
blood type on the prevalence of portal vein thrombosis in patients 153: 480–87.
with advanced chronic liver disease. Liver Int 2020; 40: 1415–26. 201 Cerini F, Gonzalez JM, Torres F, et al. Impact of anticoagulation on
181 Ollivier-Hourmand I, Repesse Y, Nahon P, et al. ABO blood upper-gastrointestinal bleeding in cirrhosis. A retrospective
group does not influence Child-Pugh A cirrhosis outcome: multicenter study. Hepatology 2015; 62: 575–83.
an observational study from CIRRAL and ANRS CO12 CIRVIR 202 Xu S, Guo X, Xu X, et al. Natural history and predictors
cohorts. Liver Int 2022; 42: 1386–400. associated with the evolution of portal venous system thrombosis
182 Fortea JI, Carrera IG, Puente Á, et al. Portal thrombosis in cirrhosis: in liver cirrhosis. Eur J Gastroenterol Hepatol 2021;
role of thrombophilic disorders. J Clin Med 2020; 9: 2822. 33 (suppl 1): e423–30.
183 Semmler G, Lens S, Meyer EL, et al. Non-invasive tests for clinically 203 Senzolo M, M Sartori T, Rossetto V, et al. Prospective evaluation of
significant portal hypertension after HCV cure. J Hepatol 2022; anticoagulation and transjugular intrahepatic portosystemic shunt
77: 1573–85. for the management of portal vein thrombosis in cirrhosis. Liver Int
184 Bradley C, Scott RA, Cox E, et al. Short-term changes observed in 2012; 32: 919–27.
multiparametric liver MRI following therapy with direct-acting 204 Villa E, Cammà C, Marietta M, Luongo M, Critelli R, Colopi S, et al.
antivirals in chronic hepatitis C virus patients. Eur Radiol 2019; Enoxaparin prevents portal vein thrombosis and liver
29: 3100–07. decompensation in patients with advanced cirrhosis.
185 Faheem HA, Abdulmeged NA, Hussein HA, Elmoursi AA, Gastroenterology 2012; 143: 1253–60.
Elnaggar HTMY, Ghait RS. Impact of sustained virologic response 205 Sanchez AP, Turon F, Martinez J, et al. Rivaroxaban improves
in regression of portal hypertension in Egyptian patients with survival and decompensation in cirrhotic patients with moderate
hepatitis C virus-associated cirrhosis and portal hypertension. liver dysfunction. Double-blind, placebo-controlled trial. J Hepatol
Egypt Liver J 2022; 12: 27. 2023; 78: S2–3.
186 Tripodi A, D’Ambrosio R, Padovan L, et al. Evaluation of coagulation 206 Serper M, Weinberg EM, Cohen JB, Reese PP, Taddei TH,
during treatment with directly acting antivirals in patients with Kaplan DE. Mortality and hepatic decompensation in patients
hepatitis C virus related cirrhosis. Liver Int 2017; 37: 1295–303. with cirrhosis and atrial fibrillation treated with anticoagulation.
187 Russo FP, Zanetto A, Campello E, et al. Reversal of Hepatology 2021; 73: 219–32.
hypercoagulability in patients with HCV-related cirrhosis after 207 Guerrero A, Campo LD, Piscaglia F, et al. Anticoagulation improves
treatment with direct-acting antivirals. Liver Int 2018; 38: 2210–18. survival in patients with cirrhosis and portal vein thrombosis:
188 Muñoz-Hernández R, Ampuero J, Millán R, et al. Hepatitis C virus the IMPORTAL competing-risk meta-analysis. J Hepatol 2023;
clearance by direct-acting antivirals agents improves endothelial 79: 69–78.
dysfunction and subclinical atherosclerosis: HEPCAR study. 208 Valeriani E, Di Nisio M, Riva N, et al. Anticoagulant treatment for
Clin Transl Gastroenterol 2020; 11: e00203. splanchnic vein thrombosis in liver cirrhosis: a systematic review
189 Di Minno MND, Ambrosino P, Buonomo AR, et al. Direct-acting and meta-analysis. Thromb Haemost 2021; 121: 867–76.
antivirals improve endothelial function in patients with chronic 209 Wang L, Guo X, Xu X, et al. Anticoagulation favors thrombus
hepatitis: a prospective cohort study. Intern Emerg Med 2020; recanalization and survival in patients with liver cirrhosis and
15: 263–71. portal vein thrombosis: results of a meta-analysis. Adv Ther 2021;
190 Álvarez B, Restrepo C, García M, et al. Liver stiffness hinders 38: 495–520.
normalization of systemic inflammation and endothelial activation 210 Tripodi A, Salerno F, Chantarangkul V, et al. Evidence of normal
after hepatitis C virus (HCV) eradication in HIV/HCV coinfected thrombin generation in cirrhosis despite abnormal conventional
patients. Vaccines (Basel) 2020; 8: 323. coagulation tests. Hepatology 2005; 41: 553–58.
191 Caballero-Marcos A, Romero-Cristóbal M, Puerto M, et al. 211 Anstee QM, Goldin RD, Wright M, Martinelli A, Cox R, Thursz MR.
HCV eradication in recurrent hepatitis C after liver transplantation Coagulation status modulates murine hepatic fibrogenesis:
normalizes enhanced endothelial activation. Transpl Int 2021; implications for the development of novel therapies.
34: 2214–25. J Thromb Haemost 2008; 6: 1336–43.
192 Bauer D, Kozbial K, Schwabl P, et al. Angiopoietin 2 levels decrease 212 Simonetto DA, Yang HY, Yin M, et al. Chronic passive venous
after HCV-cure and reflect the evolution of portal hypertension. congestion drives hepatic fibrogenesis via sinusoidal thrombosis
Dig Liver Dis 2022; 54: 1222–29. and mechanical forces. Hepatology 2015; 61: 648–59.
193 Mandorfer M, Turon F, Lens S, et al. Risk of non-tumoural portal 213 Cerini F, Vilaseca M, Lafoz E, et al. Enoxaparin reduces hepatic
vein thrombosis in patients with HCV-induced cirrhosis after vascular resistance and portal pressure in cirrhotic rats. J Hepatol
sustained virological response. Liver Int 2021; 41: 2954–64. 2016; 64: 834–42.
194 Kondili LA, Zanetto A, Quaranta MG, et al. Predicting de-novo 214 Vilaseca M, García-Calderó H, Lafoz E, et al. The anticoagulant
portal vein thrombosis after HCV eradication: a long-term rivaroxaban lowers portal hypertension in cirrhotic rats mainly by
competing risk analysis in the ongoing PITER cohort. deactivating hepatic stellate cells. Hepatology 2017; 65: 2031–44.
United European Gastroenterol J 2024; 12: 352–63. 215 Northup PG, Garcia-Pagan JC, Garcia-Tsao G, et al. Vascular liver
195 Gao Z, Zhao J, Liu X, Li S, Wang M, Gao Y. Portal vein thrombosis disorders, portal vein thrombosis, and procedural bleeding in
associated with high 14-day and 6-week rebleeding in patients patients with liver disease: 2020 practice guidance by the American
after oesophageal variceal band ligation: a retrospective, Association for the Study of Liver Diseases. Hepatology 2021;
multicentre, nested case-control study. Hepatol Int 2021; 15: 1183–95. 73: 366–413.
196 Augustin S, Altamirano J, González A, et al. Effectiveness of 216 Senzolo M, Rodriguez-Castro KI, Rossetto V, et al. Increased
combined pharmacologic and ligation therapy in high-risk patients anticoagulant response to low-molecular-weight heparin in plasma
with acute esophageal variceal bleeding. Am J Gastroenterol 2011; from patients with advanced cirrhosis. J Thromb Haemost 2012;
106: 1787–95. 10: 1823–29.

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 23

 Review

217 Senzolo M, Piano S, Shalaby S, et al. Comparison of fondaparinux 238 Wang Z, Jiang M-S, Zhang H-L, et al. Is post-TIPS anticoagulation
and low-molecular-weight heparin in the treatment of portal vein therapy necessary in patients with cirrhosis and portal vein
thrombosis in cirrhosis. Am J Med 2021; 134: 1278–85. thrombosis? A randomized controlled trial. Radiology 2016;
218 Delgado MG, Seijo S, Yepes I, et al. Efficacy and safety of 279: 943–51.
anticoagulation on patients with cirrhosis and portal vein 239 Luca A, Miraglia R, Caruso S, et al. Short- and long-term effects of
thrombosis. Clin Gastroenterol Hepatol 2012; 10: 776–83. the transjugular intrahepatic portosystemic shunt on portal vein
219 Dong S, Qi H, Li Y, et al. A systematic review and meta-analysis of thrombosis in patients with cirrhosis. Gut 2011; 60: 846–52.
anticoagulation therapy for portal vein thrombosis in patients with 240 Sherman CB, Behr S, Dodge JL, Roberts JP, Yao FY, Mehta N.
cirrhosis: to treat or not to treat? Hepatol Int 2021; 15: 1356–75. Distinguishing tumor from bland portal vein thrombus in liver
220 Gupta S, Hidalgo J, Singh B, et al. Usage of direct acting oral transplant candidates with hepatocellular carcinoma: the A-VENA
anticoagulants in cirrhotic and non-cirrhotic portal vein thrombosis: criteria. Liver Transpl 2019; 25: 207–16.
a systematic review. Cureus 2021; 13: e16922. 241 Rodríguez-Castro KI, Porte RJ, Nadal E, Germani G, Burra P,
221 Koh JH, Liew ZH, Ng GK, et al. Efficacy and safety of direct oral Senzolo M. Management of nonneoplastic portal vein thrombosis
anticoagulants versus vitamin K antagonist for portal vein in the setting of liver transplantation: a systematic review.
thrombosis in cirrhosis: a systematic review and meta-analysis. Transplantation 2012; 94: 1145–53.
Dig Liver Dis 2022; 54: 56–62. 242 Davidson BR, Gibson M, Dick R, Burroughs A, Rolles K.
222 Ng CH, Tan DJH, Nistala KRY, et al. A network meta-analysis of Incidence, risk factors, management, and outcome of portal vein
direct oral anticoagulants for portal vein thrombosis in cirrhosis. abnormalities at orthotopic liver transplantation. Transplantation
Hepatol Int 2021; 15: 1196–206. 1994; 57: 1174–77.
223 De Gottardi A, Trebicka J, Klinger C, et al. Antithrombotic treatment 243 Ravaioli M, Zanello M, Grazi GL, et al. Portal vein thrombosis
with direct-acting oral anticoagulants in patients with splanchnic and liver transplantation: evolution during 10 years of experience
vein thrombosis and cirrhosis. Liver Int 2017; 37: 694–99. at the University of Bologna. Ann Surg 2011; 253: 378–84.
224 Lisman T, Bernal W, Adelmeijer J, Kamphuisen P-W, Bos S, 244 Faccia M, Santopaolo F, Gasbarrini A, Pompili M, Zocco MA,
Porte RJ. Clinically relevant increases in the international Ponziani FR. Risk factors for portal vein thrombosis or venous
normalized ratio and model of end-stage liver disease score by thromboembolism in a large cohort of hospitalized cirrhotic
therapeutic doses of direct oral anticoagulants in patients with patients. Intern Emerg Med 2022; 17: 1327–34.
cirrhosis. Res Pract Thromb Haemost 2023; 7: 100052. 245 Zanetto A, Senzolo M, Vitale A, et al. Thromboelastometry
225 Semmler G, Pomej K, Bauer DJM, et al. Safety of direct oral hypercoagulable profiles and portal vein thrombosis in cirrhotic
anticoagulants in patients with advanced liver disease. Liver Int patients with hepatocellular carcinoma. Dig Liver Dis 2017;
2021; 41: 2159–70. 49: 440–45.
226 Violi F, Vestri A, Menichelli D, Di Rocco A, Pastori D, Pignatelli P. 246 Mahmoudi TM, Marquez V, Kayal A, Carvalho R, Weiss AA.
Direct oral anticoagulants in patients with atrial fibrillation and HCC complicated by PVT: outcome and the role of anticoagulation
advanced liver disease: an exploratory meta-analysis. therapy. Can Liver J 2019; 2: 121–26.
Hepatol Commun 2020; 4: 1034–40. 247 Grasso M, Vitale A, Pizzirani E, et al. Analysis of clinical course of
227 Nisly SA, Mihm AE, Gillette C, Davis KA, Tillett J. Safety of direct portal vein thrombosis in cirrhosis with hepatocellular carcinoma.
oral anticoagulants in patients with mild to moderate cirrhosis: Dig Liver Dis 2021; 53: S43.
a systematic review and meta-analysis. J Thromb Thrombolysis 2021; 248 Zanetto A, Campello E, Pelizzaro F, et al. Haemostatic alterations
52: 817–27. in patients with cirrhosis and hepatocellular carcinoma:
228 Intagliata NM, Davitkov P, Allen AM, Falck-Ytter YT, Stine JG. laboratory evidence and clinical implications. Liver Int 2022;
AGA technical review on coagulation in cirrhosis. Gastroenterology 42: 1229–40.
2021; 161: 1630–56. 249 Senzolo M, Shalaby S, Grasso M, et al. Role of non-neoplastic portal
229 Li Z, Xu W, Wang L, et al. Risk of bleeding in liver cirrhosis vein thrombosis in the natural history of patients with cirrhosis
receiving direct oral anticoagulants: a systematic review and and first diagnosis of hepatocellular carcinoma. Hepatology 2024;
meta-analysis. Thromb Haemost 2023; 123: 1072–88. 79: 355–67.
230 Williams C, Stewart E, Conzen KD, Wolf S, Tran TT. Dabigatran 250 Falanga A, Brenner B, Khorana AA, Francis CW. Thrombotic
reversal with idarucizumab in 2 patients with portal vein complications in patients with cancer: advances in pathogenesis,
thrombosis undergoing orthotopic liver transplantation. prevention, and treatment—A report from ICTHIC 2021.
Semin Cardiothorac Vasc Anesth 2021; 25: 200–07. Res Pract Thromb Haemost 2022; 6: e12744.
231 Intagliata NM, Maitland H, Pellitier S, Caldwell SH. Reversal of 251 Haouala A, Widmer N, Duchosal MA, Montemurro M, Buclin T,
direct oral anticoagulants for liver transplantation in cirrhosis: Decosterd LA. Drug interactions with the tyrosine kinase
a step forward. Liver Transpl 2017; 23: 396–97. inhibitors imatinib, dasatinib, and nilotinib. Blood 2011;
232 Amitrano L, Guardascione MA, Menchise A, et al. Safety and 117: e75–87.
efficacy of anticoagulation therapy with low molecular weight 252 Lafaie L, Hodin S, Saïb S, Bin V, Bertoletti L, Delavenne X. Tyrosine
heparin for portal vein thrombosis in patients with liver cirrhosis. kinase inhibitors and direct oral anticoagulants: in vitro evaluation
J Clin Gastroenterol 2010; 44: 448–51. of drug-drug interaction mediated by P-glycoprotein.
233 Rodríguez-Castro KI, Senzolo M, Sartori MT. Optimal length of Fundam Clin Pharmacol 2022; 36: 860–68.
anticoagulant therapy in cirrhotic patients with portal vein 253 Patel SH, George TL, Wang T-F, et al. Increased bleeding risk
thrombosis. Clin Gastroenterol Hepatol 2012; 10: 820–21. associated with concurrent vascular endothelial growth factor
234 Ferral H, Gamboa P, Postoak DW, et al. Survival after elective receptor tyrosine kinase inhibitors and low-molecular-weight
transjugular intrahepatic portosystemic shunt creation: prediction heparin. Cancer 2021; 127: 938–45.
with model for end-stage liver disease score. Radiology 2004; 254 Kelly DM, Shiba H, Nakagawa S, et al. Hepatic blood flow plays
231: 231–36. an important role in ischemia-reperfusion injury. Liver Transpl 2011;
235 Horhat A, Bureau C, Thabut D, Rudler M. Transjugular intrahepatic 17: 1448–56.
portosystemic shunt in patients with cirrhosis: indications and 255 Bhangui P, Lim C, Levesque E, et al. Novel classification of
posttransjugular intrahepatic portosystemic shunt complications in non-malignant portal vein thrombosis: a guide to surgical
2020. United European Gastroenterol J 2021; 9: 203–08. decision-making during liver transplantation. J Hepatol 2019;
236 Thornburg B, Desai K, Hickey R, et al. Pretransplantation portal 71: 1038–50.
vein recanalization and transjugular intrahepatic portosystemic 256 Yerdel MA, Gunson B, Mirza D, et al. Portal vein thrombosis in
shunt creation for chronic portal vein thrombosis: final analysis adults undergoing liver transplantation: risk factors, screening,
of a 61-patient cohort. J Vasc Interv Radiol 2017; 28: 1714–21. management, and outcome. Transplantation 2000; 69: 1873–81.
237 Rodrigues SG, Sixt S, Abraldes JG, et al. Systematic review with 257 Azoulay D, Quintini C, Rayar M, et al. Renoportal anastomosis
meta-analysis: portal vein recanalisation and transjugular during liver transplantation in patients with portal vein thrombosis:
intrahepatic portosystemic shunt for portal vein thrombosis. first long-term results from a multicenter study. Ann Surg 2022;
Aliment Pharmacol Ther 2019; 49: 20–30. 276: e825–33.

24 www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9

 Review

258 Ghabril M, Agarwal S, Lacerda M, Chalasani N, Kwo P, Tector AJ. 266 Ikai I, Kudo M, Arii S, et al. Report of the 18th follow-up survey of
Portal vein thrombosis is a risk factor for poor early outcomes after primary liver cancer in Japan. Hepatol Res 2010; 40: 1043–59.
liver transplantation: analysis of risk factors and outcomes for 267 Plaforet V, Moga L, Elkrief L, Rautou PE, Ronot M. Inter-reader
portal vein thrombosis in waitlisted patients. Transplantation 2016; reliability for the assessment of portal vein thrombosis. Insights
100: 126–33. Imaging 2024; 15 (suppl 1): 123.
259 Sanchez-Ocaña R, Tejedor-Tejada J, Cimavilla-Roman M, et al. 268 Eisenhauer EA, Therasse P, Bogaerts J, et al. New response
Utility of oral anticoagulants as prophylaxis of recurrent portal evaluation criteria in solid tumours: revised RECIST guideline
thrombosis after liver transplantation. Transplant Proc 2019; (version 1·1). Eur J Cancer 2009; 45: 228–47.
51: 83–86. 269 Luca A, Caruso S, Milazzo M, et al. Natural course of extrahepatic
260 Bos I, Blondeau M, Wouters D, et al. Therapeutic anticoagulation non-malignant partial portal vein thrombosis in patients with
after liver transplantation is not useful among patients with cirrhosis. Radiology 2012; 265: 124–32.
pre-transplant Yerdel-grade I/II portal vein thrombosis: a two-center 270 Suda T, Takatori H, Hayashi T, et al. Investigation of thrombosis
retrospective study. J Thromb Haemost 2021; 19: 2760–71. volume, anticoagulants, and recurrence factors in portal vein
261 Hepatobiliary Disease Study Group, Chinese Society of thrombosis with cirrhosis. Life 2020; 10: 177.
Gastroenterology, Chinese Medical Association. Consensus for 271 Condat B, Pessione F, Helene Denninger M, Hillaire S, Valla D.
management of portal vein thrombosis in liver cirrhosis (2020, Recent portal or mesenteric venous thrombosis: increased
Shanghai). J Dig Dis 2021; 22: 176–86. recognition and frequent recanalization on anticoagulant therapy.
262 Bauer J, Johnson S, Durham J, et al. The role of TIPS for portal vein Hepatology 2000; 32: 466–70.
patency in liver transplant patients with portal vein thrombosis. 272 Amitrano L, Guardascione MA, Scaglione M, et al. Prognostic
Liver Transpl 2006; 12: 1544–51. factors in non-cirrhotic patients with splanchnic vein thromboses.
263 Fink MA, Mayer VL, Schneider T, et al. CT angiography clot burden Am J Gastroenterol 2007; 102: 2464–70.
score from data mining of structured reports for pulmonary 273 Condat B, Pessione F, Hillaire S, et al. Current outcome of portal
embolism. Radiology 2022; 302: 175–84. vein thrombosis in adults: risk and benefit of anticoagulant therapy.
264 Ma J, Yan Z, Luo J, Liu Q, Wang J, Qiu S. Rational classification of Gastroenterology 2001; 120: 490–97.
portal vein thrombosis and its clinical significance. PLoS One 2014;
9: e112501.
265 Onda S, Furukawa K, Shirai Y, et al. New classification-oriented
treatment strategy for portal vein thrombosis after hepatectomy.
Ann Gastroenterol Surg 2020; 4: 701–09.

www.thelancet.com/gastrohep Published online July 9, 2024 https://doi.org/10.1016/S2468-1253(24)00155-9 25