Cavernous transformation of the portal vein (CTPV) is usually secondary to long-standing portal vein thrombosis (PVT) or portal vein obstruction. As a sequela of portal vein obstruction, especially complete extrahepatic portal vein obstruction (EHPVO), fibroblasts transform the clot into a firm, collagenous plug in, which tortuous venous channels develop. On this basis, portal hypertension caused by PVT may promote the formation of periportal or intrahepatic venous collateral circulation, resulting in CTPV, bypassing extrahepatic portal venous occlusion over time, and interrupting portal inflow (1–5). CTPV may occur as early as 6–20 days after EHPVO, with an average time of approximately 5 weeks (3, 5, 6). It was first reported in 1869 by Balfour and Stewart and is also referred to as a portal cavernoma due to the sponge-like appearance of the portal vein (7).

Cavernous transformation of the portal vein may remain insidious for long-term presentation. It often manifests as gastroesophageal variceal bleeding, splenomegaly, and thrombocytopenia. The biliary tree may undergo morphological and functional changes due to CTPV, resulting in obstructive jaundice. It has recently been termed “portal biliopathy” (8). The mortality rate due to variceal hemorrhage is 5%, and the overall mortality rate is 10% in both adults and children with CTPV (9, 10). Even worse, more than 43% of obstetric patients with non-cirrhotic prehepatic portal hypertension and the development of esophageal varices will suffer significant variceal bleeding during pregnancy. This potentially catastrophic complication is associated with a 33% perinatal mortality rate (11). Recently, it was suggested that the presence of ascites may be of great importance to predict the incidence of death in patients with CTPV, mainly attributed to its close correlation with the deterioration of liver function (12–14). CTPV and chronic PVT remain a challenge at the time of transplant. They are relative contraindications to liver transplantation at many centers because the non-physiological portal flow may increase perioperative and postoperative risks associated with surgical techniques.

The hemodynamics of the portal venous system are characterized by low pressure, slow flow, and high volume. Similar to other venous thromboses, the formation of PVT is also multifactorial due to reduced blood flow, endothelial injury, and hypercoagulability. PVT is caused by a combination of local and risk factors (Figure 1).

The causes of CTPV and acute PVT in adults are similar. The commonly affected population may be divided into cirrhotic and non-cirrhotic patients. The composition ratio of cirrhotic and non-cirrhotic cases varies by region. A study of 254 autopsies from Sweden reported that 28% of PVT cases were cirrhotic, whereas non-cirrhotic PVT accounted for most of the rest (15). PVT occurs in ~20% of cirrhotic patients, who are a great body of inpatients in China. The development of cirrhotic PVT is closely associated with static portal blood flow due to portal hypertension and endothelial injury due to intestinal infection and therapeutic inflammation. In non-cirrhotic patients, a systemic hypercoagulable state is often implicated in PVT (Figure 1). The occurrence of vascular malformation in children suggests that a congenital defect is often a contributing factor, such as prior umbilical cannulation and infection. Inherited and acquired prothrombotic disorders (e.g., latent myeloproliferative disorder, protein C or protein S deficiency, and antiphospholipid syndrome) and thrombotic stimuli, such as pregnancy or oral contraceptives, intra-abdominal infection, and surgical procedures, have been observed in adults, while repeated abdominal infections, sepsis, abdominal invasive procedures, trauma, and congenital anomalies, with or without a prothrombotic state, have been alleged to lead to CTPV in children (16–19). After the exclusion of the aforementioned causes, the etiology of EHPVO remains obscure in up to 50% of patients (20).

The prevalence of PVT in the general population ranges from 0.7 to 1/10⁵ (21). It is increased ~1,000-fold in cirrhotic patients, with a range of 0.6%−50%, increasing proportionally with liver cirrhosis severity (22, 23). Epidemiological data of non-cirrhotic PVT in the general population are limited by its infrequency. The prevalence of EHPVO was estimated to be as high as 1.0% in an autopsy study in Sweden (15). However, it was much lower (3.7 per 100,000 population) in another Swedish study based on hospital discharge diagnoses (21), suggesting that EHPVO commonly develops at a late stage of some diseases. CTPV among adults is quite rare, with an incidence of 15.6% among EHPVO (24).

The concept of CTPV as a relatively rare disease is mainly based on clinical series and case reports. The literature on CTPV is scant. Therefore, there is considerable heterogeneity in its treatment. Accordingly, the current therapies are mainly extrapolated from cohort studies and/or based on clinician expertise. Optimal treatment for patients with CTPV becomes important with increasing identification by imaging modalities.

Anticoagulation is the cornerstone of treatment for acute PVT without malignancy and should be initiated at diagnosis. Anticoagulant treatment must be considered in cirrhotic patients with PVT following the implementation of adequate prophylaxis for gastrointestinal bleeding (25). However, the main portal vein of CTPV is commonly considered to be unable to be recanalized with anticoagulation because it is a complete obstruction of the mesentericoportal axis. There is only one case reported in which the portal cavernoma was reversed by long-term (5 years) anticoagulation (26). Anticoagulation may lessen the degree of bile duct obstruction in certain patients with CTPV with cholangiopathy, probably by maintaining the patency of periportal or intrahepatic venous collaterals and reducing the compression of the bile duct (27). It is not beneficial for children to take long-term anticoagulation since the procoagulant state is an occasional cause of chronic non-cirrhotic EHPVO (28, 29).

Endoscopic therapy cannot reduce portal hypertension and is mainly used for temporary hemostasis in acute variceal bleeding. Endoscopic variceal ligation (EVL) is recommended for the management of active esophageal variceal bleeding (12, 30, 31). When acute bleeding from isolated gastric varices and gastroesophageal varices type 2 (GOV2) extends beyond the cardia, sclerotherapy or endoscopic therapy with tissue adhesive should be taken into account. Either ligation or endoscopic therapy tissue adhesive can be used for GOV1 bleeding. Primary prophylaxis with endoscopic treatment has been recommended for patients with cirrhosis with EHPVO by the Baveno VI consensus (31). Recently, endoscopic ultrasound-guided coil (EUS-coil) therapy has emerged as a promising option due to its therapeutic superiority over endoscopic glue injection. It is believed that EUS-coil therapy is the next intervention for primary and secondary hemostases of gastric variceal bleeding and will result in a paradigm shift (32, 33).

Multiple endoscopic procedures may become a local risk for PVT due to traumatic inflammation in the portal system. When portal-systemic collaterals are blocked by endoscopic procedures without being diverted, the increased portal pressure carries an increased risk of biliary complications, mostly from the suppression of periportal collateral vessels. Thus, portal decompression should be performed early, even if the risk of bleeding is not high (34–37).

Gastroesophageal devascularization between the portal and azygos veins alone without a shunt has been less performed recently due to its higher rebleeding rate and lower 5-year survival rate (38, 39). There are several surgical approaches by which portacaval shunts can be established to decompress portal hypertension (40–42). The experience from adults indicated that non-selective surgical shunts such as side-to-side splenorenal or meso-caval shunts can totally divert the portal flow toward the systemic circulation but apparently potentiate ischemia-perfusion to the liver at the same time (40, 43). In this regard, non-selective shunts and their derivative techniques are taken into account only in patients with refractory life-threatening bleeding (44–46). Selective shunts, such as the distal splenorenal shunt developed by Warren et al., have been considered as effective as non-selective shunts in controlling variceal bleeding. It preserves a portion of portal perfusion to the liver and is better in preventing portosystemic encephalopathy (40, 47). However, widespread use of distal splenorenal shunts has been limited in terms of technical requirements, particularly in adults (38).

Meso-Rex shunting is a recently established surgical technique. The patient's internal jugular vein is used as an autograft by which the superior mesenteric vein blood flow is diverted into the umbilical portion of the left portal vein. This technique sufficiently restores a substantial portion of physiological portal blood flow to the liver and avoids the long-term adverse consequences of portosystemic shunting, such as liver atrophy and growth retardation (16, 40, 45). This portal-flow-preserving shunt has been widely performed in children with CTPV. The bypass patency rates might reach 100% 1 year after the operation and remained at 78% during the 10-year follow-up. Of the 490 reported cases, only three deaths occurred (Table 1) (40, 45, 48–65).

Later, several modifications of meso-Rex bypass using alternate sources of venous inflow and graft conduits in children when standard meso-Rex shunting cannot be achieved (66, 67). It has been recommended by the Baveno consensus that meso-Rex bypass, the most physiological shunting and the only “curative surgery,” should be performed for patients in the early stage of EHPVO (31, 44, 68, 69). However, the controversy concerning the legitimacy of its utilization in an asymptomatic child is still unresolved. Available data show an inverse correlation between the restoration of appropriate portal flow to the liver following meso-Rex shunt and the age of the patient (49, 69). It has been proposed that the assessment of surgical feasibility should be performed in all children with CTPV in a more prophylactic manner (70–72), while in fact, many medical centers routinely defer preemptive meso-Rex bypass until intractable and refractory symptoms are established.

Inspired by the success of meso-Rex bypass in children with CTPV, a modified meso-Rex bypass (splenic vein to cystic part of the portal vein) was successfully performed in an adult patient after liver transplantation (73). Recently, a novel meso-Rex bypass with umbilical vein recanalization and intraoperative stenting was reported (58). It may reduce the risk of intravascular thrombosis and bypass vessel occlusion compared with traditional meso-Rex shunting. The limited data on surgical shunting in adults with CTPV will encourage more studies for adult patients.

Cavernous transformation of the portal vein was once considered an absolute contraindication for liver transplantation due to unsatisfactory portal flow to the graft (74, 75). However, incremental refinements of the surgical technique have allowed recipients with CTPV to undergo liver transplantation using a prominent collateral vein, dilated coronary vein, interposition graft, portal vein arterialization, or some anastomotic methods, including renoportal or cavoportal anastomosis, for inflow reconstruction. Occasionally, multivisceral transplantation, such as liver-small intestinal transplantation, may be lifesaving and should be considered for patients with diffuse mesenteric venous thrombosis (74, 76–83).

Surgical challenges have promoted the application of interventional technology in patients with CTPV. The multitude of complications due to CTPV can be effectively addressed by various interventional vascular therapies.

Patients with CTPV should be treated in regional central hospital with strong surgical and interventional teams. Meso-Rex bypass is recommended for children with CTPV. Incremental refinements of the interventional techniques, such as mTIPS creation and PVS, can greatly benefit adult patients with CTPV by reducing the need for liver transplantation or transforming contraindications of liver transplantation into indications. The optimal treatment for patients is a personal interventional strategy on the basis of their heterogeneity in mesentericoportal obstruction.

CT contributed to conception and design of the study. ZH analyzed the literature and drew the tables. BW wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

This study was supported by the National Natural Science Fund of China (82170625, U1702281), the National Key R&D Program of China (2017YFA0205404), and the 135 projects for disciplines of excellence of West China Hospital, Sichuan University (ZYGD18004).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.