Collateral circulation caused by end-stage hepatic alveolar echinococcosis

The statistical software IBM SPSS 26.0 (IBM Corp, Armonk, NY, USA) was used for data analysis. Normally distributed continuous variables were expressed as Means ± SD and analyzed by t-test; abnormally distributed continuous variables were expressed as median and analyzed by Mann–Whitney u-test. p < 0.05 was considered a statistically significant difference.

In case of PV obstruction, there can be two collateral vascular pathways. one is the portal vein-portal vein pathway, proposed by Balfour and Stewart [5] in 1869, which is known as cavernous degeneration and featured by the formation of tortuous dilated vessels around the first PV, connecting the extrahepatic PV to the intrahepatic PV. The other is the portal—systemic circulation pathway, flowing into the hemi-azygos vein and the superior vena cava [6] through the PV-left esophagogastric vein, esophagogastric fundic plexus and splenic vein and then which is also a common pattern of collateral circulation in portal hypertension caused by cirrhosis.

The HAE patients we observed showed differences in both the incidence and clinical symptoms of PV cavernous degeneration from other diseases. All patients with severe portal trunk occlusion developed cavernous degeneration, with a much higher incidence than reported in thrombosis [6]. 18 patients among the 33 collateral vessels of PV suffered abdominal distension and abdominal pain compared with only 3 with gastrointestinal bleeding. The main symptom of collateral formation caused by other diseases reported is upper gastrointestinal hemorrhage, which is significantly higher than our results [7‐11]. All of this may be related to the fact that AE invasion of the vessels is a chronic process, allowing enough time to gradually establish type I collateral vessels, which relieves the venous pressure in the main PV. In addition, it is noteworthy that we observed both type I and type II collateral vessel morphology, but a separate portal-somatic circulation collateral morphology was absent. Based on the above observations, we hypothesized that the degree and speed of PV trunk obstruction was closely related to the form of PV collateral branch formation. Narrow and mild PV obstruction would allow its surrounding collateral veins to cross the obstruction site and drain the distal blood to the intrahepatic portal branches; that is, the portal vein-portal vein shunt was first formed to maintain normal portal perfusion. As the extent and degree of PV obstruction increased, the portal-portal blood flow was insufficient to relieve portal hypertension, and some of the blood flow was directed to the lower pressure systemic circulation, forming a portal vein- systemic circulation. Khanna and Shiv K. Sarin also made a similar point [11, 12].

The incidence of portal vein cavernous degeneration after portal trunk invasion is higher in HAE patients than in acute portal vein embolism (57.2%), and completely different from those of patients with cirrhosis whose primary manifestation is the establishment of a pure portal-systemic circulation [13, 14]. Therefore, we speculated that when prehepatic portal hypertension occurred and the liver was in good condition without other diseases, type I portal collateral branches served as the main mode of compensatory blood flow. The slower the portal vein was occluded, the higher the incidence of cavernous deformation would be, which coincided with other reports [6]. Moreover, HAE invaded the PV at a low level, between the I-II branches, and was less likely to be combined with umbilical vein dilation. None of the patients showed dilatation of the umbilical vein or abdominal wall vein.

The treatment of such patients was very tricky and liver resection may risk uncontrollable bleeding. Ex vivo liver resection and autotransplantaion (ELRA) enjoyed good intraoperative bleeding control and materialized radical resection. We performed ELRA treatment for end-stage HAE combined with portal vein cavernous degeneration in 9 patients, with initial good results (Fig. 3), hoping to serve as a new alternative treatment.

The Vascular collateralization of the IVC has been reported in details by Kapur [15], but the vertebral and lumbar veins have not been described, which may differ from the specific location of the IVC invaded by HAE. The posterior hepatic segment of IVC between the left atrium and the right renal vein was the main site involved. The collateral circulation of the IVC was thought to have two main pathways: the lumbar venous plexus and the vertebral venous plexus [16] (Fig. 4). Via the four lumbar veins symmetrical to the left and right sides and the lumbar ascending veins, left side of the lumbar venous moves into the hemiazygos vein and the right side enters the superior vena cava and returns to the atrial along the azygos vein. IVC angiography revealed that the collateral vessels of the vertebral venous plexus were dominated by the intraspinal veins, which ascended along the spinal canal. It had been reported to be an important route for communicating the superior and IVC, as well as intra—and extracranial, because of the lack of venous valves and the extensive communicating branches at other venous plexuses (intracranial venous sinus, thoracic, abdominal, pelvic venous plexus), which was of great physiological and clinical importance in the spread of infection, metastasis of tumors, and the formation of collateral circulation [17‐19]. One patient out of 12 showed a metastatic lesion in the brain but not in the lung. Is it possible that this is the route by which neglected hepatic AE metastasize to the brain?

Clinically, IVC obstruction was manifested by ascites and lower limb edema. Most patients with HAE are asymptomatic because of the well-established collateral circulation. The presence or absence of significant symptoms served as an important basis for determining whether the collateral circulation was well established after complete IVC obstruction, and played an important role in deciding which surgical approach to reconstruct the IVC. Our previous study proposed a surgical approach without reconstruction of the IVC in the presence of well-established collateral circulation, yielding good results [20, 21]. But for which careful preoperative evaluation and implementation lay the basis.

It was of concern that the collateral circulation of the inferior vena cava was difficult to observe by CT and was thus usually visualized by angiography. Therefore, once the inferior vena cava was severely obstructed, it was recommended to improve the inferior vena cava angiography to well assess whether there was collateral circulation established.

Collateral circulation of the hepatic vein has rarely been reported in other diseases. It is generally accepted that the short hepatic vein is an important compensatory vessel [22]. When the hepatic vein was obstructed, the short hepatic vein became abnormally thick and compensated for the function of the hepatic vein. In two patients, complete obstruction of the hepatic vein was followed by a dilated traffic vessel between the hepatic vein and the short hepatic vein, in which case blood flowed through the traffic branch into the short hepatic vein and then converged into the IVC (Fig. 5). Restricted by a few reports available in the form of case reports [23‐25], we hypothesized that the appearance of intrahepatic venous plexus resulted from the original presence of fine intervening venous traffic branches in the liver, which gradually expanded and acted as mutual traffic when the main hepatic venous trunk was obstructed. However, if similar traffic vessels were lacking in the liver, the main hepatic venous reflux area would cause congestion, while the uninvolved short venous return area remained unaffected. This may have important implications for ELRA, since both the main hepatic trunk and the short hepatic veins was to be reconstructed when the liver formed thicker short hepatic veins and the remaining liver tissue remained congested. If there was an intrahepatic venous plexus and obstruction of the main hepatic trunk without obvious signs of liver congestion, it was feasible to reconstruct only the main hepatic trunk and suture the short hepatic vein. One case was reconstructed with ELRA and no significant postoperative complications occurred.

The collateral circulation of the hepatic artery in patients with hepatic AE has rarely been reported, but it has been reported in small numbers in patients with HCC because of its particular implications for interventional treatment. We found one case in which the main hepatic artery was invaded, with an artery emanating from the coeliac trunk and running within the hepato-gastric ligament to supply blood to the healthy left liver (Fig. 6). Although it is possible that the artery could be a variant, it does also act as a blood compensation similar to collateral circulation. Phrenic artery is the most common variant of arterial blood supply for liver cancer patient, with equal origin from either the aorta or coeliac axis. It is worth exploring whether direct resection of the originally invaded arterial trunk with reconstruction of only the PV and biliary tract is feasible in patients with end-stage vesicular echinococcosis in whom arterial vascular collaterals are present and the first hepatic hilum invaded.

There are also many shortcomings. Due to the invasive nature of venography, not all cases in this study were observed by angiography. In addition, this study was restricted by a small sample size.