Original extrahilar approach for hepatic portal revascularization and relief of extrahepatic portal hypertension related to later portal vein thrombosis after pediatric liver transplantation. Long term results.

Abstract

* Abbreviation: PVT, portal vein thrombosis. Thrombosis of the portal trunk is not infrequent after liver transplantation (1-3). It has been reported either with full size (0-10%) (4, 5 or reduced grafts (0.8-11%) 1-3), and is much more frequent in small children with portal vein hypoplasia(2-4). When it occurs early, it has been identified as a cause of graft failure. Although late portal vein thrombosis can be extremely well tolerated, with cavernous transformation of the portal vein and the formation of hepatopetal collaterals which deliver blood to the liver, extrahepatic portal hypertension develops eventually with its related complications. Little data are available in the literature concerning its treatment at this stage, which usually consists of the creation of portosystemic shunts. Since 1990, we have used an original extrahilar approach to treat seven children who received transplants and who developed late portal vein thrombosis and symptomatic extrahepatic portal hypertension with normal hepatic function and structure. By using a vein graft interposed between the superior mesenteric vein and the left intrahepatic portal vein, the venous splanchnic flow was rerouted to the liver. The low resistance of the normal liver allowed the successful relief of the portal hypertension. METHODS Once portal hypertension has been confirmed by direct pressure gradient measurement (atrium/mesenteric vein), the umbilical vein is canalized to allow confirmation of a normal pressure gradient between intrahepatic portal branches and the right atrium (to exclude an intrahepatic block). Venography through the recanalized umbilical vein also allows confirmation of the patency of the left portal vein (Fig. 1). The umbilical remnant is then further dissected, as well as the distal part of the left portal vein(Rex recessus) and its branches for segments III and IV. Careful dissection provides excellent exposure of the vein for a 3- to 4-cm length without any damage to biliary or arterial branches. A small laterally positioned vascular clamp provides complete control of the vein and all its branches. The umbilical remnant is divided at its junction with the left portal vein, and this venotomy is then longitudinally extended on its ventral aspect. A venous graft (patient's own jugular vein or iliac vein allograft) is anastomosed end-to-side to this termino-lateral opening. The venous graft is brought in front of the duodenum and through the transverse mesocolon to reach the superior mesenteric vein, to which it is connected in an end-to-side fashion(Fig. 2). Access to the umbilical part of the left portal vein is reduced when a parenchymal bridge is present; in these cases, the presence of the bridge carries the risk of direct compression of the shunt when the abdomen is closed and a segment of the parenchyma should be resected. RESULTS Of the 388 children who received liver transplants between March 1984 and December 1993, eight children developed late portal vein thrombosis (PVT*). One asymptomatic patient was successfully treated conservatively and the other seven patients were treated with the extrahilar approach. All the latter liver grafts were first transplants. Indication for liver replacement was biliary atresia cirrhosis in six children (one with hepatopulmonary syndrome) and cryptogenetic cirrhosis in the remaining one patient. All grafts were ABO compatible (six identical), five were full-size livers (four donors less than 20 kg), and two were partial liver grafts. At the time of transplantation, all recipients except one weighed less than 12 kg, and five had hypoplasia of the portal vein (less than 5 mm). In all cases, normal liver histology was confirmed by preoperative biopsy and patency of intrahepatic portal branches proven by Doppler ultrasound. In the first patient, the PVT was diagnosed early, during the fifth postoperative week, during investigations aimed at finding the cause of refractory ascites. A superior mesenteric-left intrahepatic portal vein bypass(MRS) was performed successfully and the ascites resolved completely(6). In the other six patients, the diagnosis was made later in the postoperative course when signs of portal hypertension developed clinically: none had previously presented with symptoms suggesting the acute phase of a thrombotic process. Surgical correction was performed between 6 and 96 months after transplantation. All six patients presented with splenomegaly and moderate hypersplenism (platelet count between 60 and 125×103), and all had esophageal varices except for one who presented with massive bleeding from the Roux-en-Y loop. Three patients had normal liver tests; in the others, there was a slight elevation of liver enzymes with normal bilirubin. These abnormalities were related to biliary stenosis in two patients and hepatitis C virus hepatitis in one patient. The indication for surgery in four patients was proven gastrointestinal hemorrhage and/or chronic anemia. One other patient with esophageal varices was treated conservatively for 4 years without any bleeding episode; with time, however, the risk of hemorrhage increased (grade IV esophageal varices with red spots and gastric varices). The patient then developed a lymphoproliferative disease in the form of a solitary nodule in segment II of the liver; a left lobectomy (left lateral segment) was performed combined with portal revascularization. In the last patient it was considered that the recurrence of hepatopulmonary syndrome was related to the PVT (unpublished personal observation); hepatic revascularization achieved regression of hypoxia. In three patients, another procedure was performed at the same time as the shunt (left lobectomy in one patient and biliary tract revision in two patients). Extrinsic compression (due to insufficient excision of the parenchymal bridge) caused complete blockage of flow in one patient and was diagnosed immediately after operation: it was corrected successfully at relaparotomy. In the two patients in whom a venous allograft from an unrelated ABO-identical cadaveric donor was used, progressive sclerosis of the venous allograft led to obstruction of the bypass. In the first patient, the bypass was patent at 3 weeks but thrombosis occurred during the third month. In the other patient, a long irregular stenosis of the allograft was seen by Doppler scan and angiography, and was corrected successfully using balloon percutaneous angioplasty. Over the next 4 months, repeat Doppler scans were normal, but then thrombosis occurred. In both instances, the obstruction was corrected successfully using a jugular autograft. Finally, percutaneous transhepatic transluminal dilatation of a stenosed distal anastomosis had to be performed in one patient 7 months after surgery. The stenosis has not recurred and the patient remains well 4 years later. The effectiveness of this innovative decompressive technique was assessed perioperatively and confirmed by the subsequent clinical course(Table 1). During surgery, there was a drop in mesenterico-atrial pressure gradient (median before and after bypass: 14 and 7 mmHg, respectively); this was confirmed by intraoperative angiography or flow measurement in four patients. In three patients, esophagoscopy was performed both before and after surgery. In two patients it confirmed a significant reduction in the size of the varices within 2 weeks of the procedure (from grade II or III, with red spots, to flat varices without red spots); in one patient, endoscopy was repeated 3 years after surgery and confirmed the disappearance of the varices. In the third patient, who presented before surgery with grade III esophageal varices with red spots and gastric varices, a repeat examination at 6 months showed almost complete resolution of the varices. A fourth patient had only endoscopy 22 months after the surgery that confirmed that the esophagus was normal. All seven patients are alive and well with functioning bypass grafts(Fig. 3). In four patients, the follow-up is currently more than 4 years and they have permanent relief of clinical symptoms and bypass patency confirmed by Doppler ultrasound. The other three patients are clinically well with a patent shunt 10-14 months after operation. There has been no further gastrointestinal bleeding and, in two patients, the splenomegaly has already clinically regressed. DISCUSSION In case of late PVT after orthotopic liver transplantation, recurrent chronic portal hypertension is the main clinical complication. Splenectomy, sclerotherapy, and portosystemic shunts have all been used as methods of treatment (7-9). The largest experience was reported at a meeting of the European Society for Organ Transplantation in 1993 (K. Tepetes, Pittsburgh, PA) and concerned seven portosystemic shunt operations in six patients: none of these patients required retransplantation and all were doing well (mean follow-up, 30 months). Another series of five patients was reported in 1994 by Knechtle et al. (10); all patients are alive but two have thrombosed their shunt, with one developing recurrent bleeding. Reoperation for portal thrombectomy is feasible only when the thrombosis is diagnosed immediately. However, if the diagnosis is delayed, thrombectomy is unsuccessful and furthermore contraindicated because of the risk of bleeding from the cavernomatous transformation of the liver hilum. It also requires a direct anterior access to the portal trunk, which in children is very difficult because the biliary reconstruction is usually performed using a Roux-en- Y loop lying in front of the protal trunk. However, the extrahilar approach in the umbilical fissure allows dissection in a territory free of portal hypertension or cavernomatous transformation, and at some distance from the main arterial and biliary structures. In two patients, revision of the choledocho-jejunal anastomosis was performed without significant blood loss. The formation of new collateral venous pathways that repatients some portal flow to the liver has been described in patients presenting with PVT some time after transplantation (9, 11, 12). However, no data concerning intrahepatic portal hemodynamics after PVT are available. In our experience, in all acute or chronic cases except one (presenting with extensive venous thrombosis involving caval and portal systems), the intrahepatic portal branches remained free of thrombus (Fig. 1) and portal flow was maintained probably by partial rerouting of arterial hepatic inflow, and its subsequent redistribution by reversed portal flow in several segmental branches (Doppler: personal observations), as also suggested by Burke et al. (13). A detailed Doppler investigation should be performed in all posttransplant PVT cases, including the intrahepatic portal branches, before portosystemic surgical decompression is considered. Patency allows the splanchnic flow to be rerouted to the intrahepatic portal venous system. In our experience, only one of the cases considered for this bypass method had significant contraindications. This was a patient in 1995 who had extensive thrombosis of the portal and splanchnic veins, probably due to hypercoagulable state. No portosystemic shunts have been used to treat PVT in either adult or pediatric patients. One adult patient is currently being considered for mesenteric-Rex shunt. Moreover, increased confidence with this technique led us to perform it successfully in patients with PVT due to idiopathic causes or following pancreas surgery. The long-term clinical course was excellent in all our patients, who are alive and well with good graft function (follow-up after transplantation: 4-9 years). Liver function is normal in all patients except for one, who has a residual biliary tract problem. These results are at least as good as those reported after transplantation and portosystemic shunt operations. However, the extrahilar approach to the left portal vein allows the portal blood to be rediverted to the liver (Fig. 3) and is thus more physiologically correct than diversion away from the liver. It simultaneously cures the portal hypertension and revascularizes the liver. Revascularization, at least theoretically, must benefit the liver physiologically; however, since the liver is “normal” before surgery in patients with PVT(histological structure and function, according to the usual tests), it is difficult to quantify shortterm postoperative benefit. Comparison with the results of other groups, long-term studies, or more sensitive measurements of preoperative liver function are necessary. The influence of hepatotrophic factors could be assessed by pre- and postoperative liver volume measurement, as we already observed an increase in size after this bypass in one transplant recipient and two patients with idiopathic PVT. As a result of the high success rate using jugular autografts and because growth must be considered when the technique is used in children, we believe the jugular vein of the patient should be the first choice for use in the bypass. In two patients we implanted, some time after transplantation (6 and 33 months) mismatched venous allografts procured from organ donors. Allografts used at the time of transplantation itself for other types of reconstruction achieve good results in our experience, but in these patients both grafts thrombosed. This thrombosis was probably not related to the technique itself, but to the use of a mismatched vein graft some time after transplantation, when the level of immunosuppression had been reduced. This resulted in a rejection process with chronic fibrosis and subsequent thrombosis. In both patients a second operation using the jugular vein was successful. From this series, we can conclude that late portal vein thrombosis is often well tolerated, but symptomatic chronic extrahepatic portal hypertension develops with time. If the liver parenchyma is normal and the intrahepatic portal branches remain patent, extrahilar mesenteric-left intrahepatic portal vein bypass is a good alternative to the conventional portosystemic shunt or long-term sclerotherapy. We strongly recommend the use of a jugular vein autograft for cases where the immunosuppression therapy has been reduced to maintenance levels.Figure 1: Perioperative opacification of the intrahepatic portal vein in a patient with a thrombosed portal vein: the intrahepatic portal system is patent (R and L, right and left portal veins; X, Rex recessus).Figure 2: Schematic and general aspect of the mesenteric-left portal vein(Mes-L.P.V.) shunt (Spl.V., splenic vein; S.M.V., superior mesenteric vein; IMV, inferior mesenteric vein).Figure 3: Portography (percutaneous transhepatic transanastomotic catheterization): the splanchnic flow is totally rerouted into the liver (R and L, right and left portal veins; G, bypass graft; S, superior mesenteric vein).