When portal vein (PV) stenosis or thrombosis occurs during the early posttransplantation period, it can be devastating and can result in loss of the graft. Surgical treatments, such as thrombectomy, anastomotic revision, or retransplantation have been considered optimal for treating such complications.1, 2 However, surgical treatments for PV stenosis or thrombosis are limited due to their technical difficulty and are impossible to perform on some patients because of multiple complicating factors.
Balloon angioplasty has been used successfully to treat some patients who experienced early posttransplantation PV stenosis or thrombosis.3–5 However, the number of such patients was limited and some of these patients had to undergo stent placement following balloon angioplasty because of remaining stenosis. To our knowledge, there have been no studies in which the investigators assessed the efficacy of primary stent placement for treating early posttransplantation PV stenosis. In this article, we report the efficacy and safety of percutaneous transhepatic primary stent placement for treating early, i.e., less than 1 month, posttransplantation PV stenosis following living donor liver transplantation.
Between September 1996 and February 2006, 9 (0.8%) of 1,095 patients who had undergone living donor liver transplantation at our institution underwent percutaneous transhepatic primary stent placement in the PV to treat PV stenosis. These patients included 8 males and 1 female, ranging in age from 5 to 61 yr (mean 43 ± 17 yr). In 6 of these 9 patients, reconstruction of the PV during transplantation was performed by standard end-to-end anastomosis of the recipient and donor PV. In 3 adult transplantations, as this standard anastomosis could not be performed, interposition of the cadaveric iliac vein graft between the recipient and donor PV was performed on 1 patient and interposition of the cadaveric iliac vein graft between the recipient left renal vein and the donor PV was performed on 2 patients. The underlying disease and liver graft data for these patients are shown in Table 1.
Table 1. Outcomes of Percutaneous Transhepatic Primary Stent Placement
Type of PV anastomosis
Poststenting follow-up duration (months)
Patency of PV
Abbreviations: No, patient number; POD, postoperative day; PV, portal vein; LC, liver cirrhosis; HBV, hepatitis B virus infection; R; right lobe, L+L, dual left lobes; L, left lobe; LL, left lateral segment; LFT, liver function tests; N/A, not available; F, fail; S, succeed; ICH, intracranial hemorrhage; MOF, multiorgan failure.
Interposed cadaveric iliac vein graft between the recipient left renal vein and the donor portal vein.
The etiology of these patients was verified on poststenting follow-up images.
The degree of stenosis was approximately 70% of donor portal vein diameter.
Pathologic examination of the excised stent during retransplantation showed partial thrombosis.
The degree of stenosis was approximately 80% of donor portal vein diameter.
The initial diagnosis of PV stenosis was based on Doppler ultrasound (US) in 7 patients. In addition, 5 of these patients also underwent computed tomography (CT) or indirect portal venography for further evaluation of the PV stenosis. Stenosis was diagnosed if Doppler US showed no flow in the PV or an acceleration of the flow rate in the poststenotic PV more than 3 times greater than that in the prestenotic PV as well as narrowing of the PV diameter of more than 50% of the main PV in adults or less than 2.5 mm in a child. In the remaining 2 patients, PV stenosis was initially diagnosed on indirect portal venography and magnetic resonance angiography, respectively.
At the time of diagnosis, 7 patients had extrahepatic PV stenosis and 2 had extrahepatic PV occlusion. One patient with stenotic PV and 1 with occluded PV had thrombosis in the PV. Of the 9 study patients, 7 presented with abnormal liver function; 2 of these 7 patients also had pathologically-proven acute graft rejection and 1 had no evidence graft rejection on pathologic examination. In the remaining 2 study patients, the principal clinical manifestations included recurrent melena, hematochezia, and ascites.
Primary Stent Placement
Written informed consent was obtained from each patient or from a legal guardian, and our institutional review board approved this study.
The average interval between living donor liver transplantation and stent placement was 13 ± 10 days (range 2–30 days). The pediatric patient underwent the procedure under general anesthesia; the adults received local anesthesia consisting of an intramuscular injection of lidocaine (Jeil Pharm, Taegu, Korea) as well as intravenous sedation (Demerol; Keukdong Pharm, Seoul, Korea). Percutaneous transhepatic puncture of the intrahepatic PV was performed using a 21-gauge Chiba needle (Cook, Bloomington, IN) under ultrasonographic and fluoroscopic guidance. The needle was exchanged for a 4-French coaxial dilator and a 6–8-French sheath (Cook) over either a 0.018-inch guide wire (Cook) or a 0.035-inch angled hydrophilic guide wire (Terumo, Tokyo, Japan). The 0.035-inch guide wire and a 5-French cobra catheter (Cook) were used to traverse the PV stenosis. Direct main portal venography and a pressure gradient across the stenosis were then obtained. A bolus of heparin (500–5,000 units) was then administered directly into the PV.
Primary stent placement was then undertaken using a Wallstent (Boston Scientific, Natick, MA), a Zilver stent (Cook), or a Symphony stent (Boston Scientific). Stents with the same diameter or with a 1–2-mm larger diameter than that of the nonstenotic extrahepatic PV, were used. Stents 4–8 cm in length were used to cover a stenosis with minimal angulation between the PV and the proximal and distal edges of the deployed stent. Balloon angioplasty following stent placement was not routinely performed; however, it was performed if the deployed stent showed an hourglass deformity of more than 50% of its normal diameter. Angioplasty was carefully performed with a smaller diameter balloon catheter than that of the deployed stent in order to prevent anastomotic disruption during balloon inflation.
After the procedure, a poststenting portal venogram and the pressure gradient were obtained. The percutaneous transhepatic tract was embolized using several coils (Cook). Patients who had normal coagulation function or had PV thrombosis on prestenting US or CT, were given both intravenous heparin for 2–5 days in order to produce an international normalized ratio of 1.5–2.0 as well as oral antiplatelets (aspirin 100 mg/day and/or dipyridamole 75 mg/day) for 3–6 months following the procedure. However, patients who had coagulopathy were given oral antiplatelets when their coagulation function normalized.
The following parameters were documented retrospectively: pressure gradients across a stenosis before and after stent placement; technical success and complications; clinical success; and the patency of the PV inflow. We defined technical success as successful stent placement in the intended location of the PV with subsequently improved PV inflow and less than 30% residual stenosis. We defined clinical success as subsequent improvement of liver function and amelioration of the clinical manifestations relating to portal hypertension. We defined major complications as those necessitating an increased level of care, an additional surgical or interventional manipulation, adverse sequelae, or death. All other complications were defined as minor complications.
Patency of the PV inflow was evaluated by means of Doppler US and CT. Doppler US was routinely performed on days 1, 2, 3, and 7 after the procedure, weekly thereafter until the patient was discharged, and then 1, 6, and 12 months after discharge. CT was also performed at a 7–10-day interval until the patient was discharged and then 1, 3, 6, and 12 months after discharge.
Stents of 10-mm diameter and 4–7 cm in length were used in the pediatric patient and in 4 adults, and stents of 12–14-mm diameter and 4–8 cm in length were used in the remaining 4 adults. As thrombus in the PV was indistinct on direct portal venography, thrombolysis or thrombectomy was not performed.
Technical and clinical success was achieved in 7 (77.8%) of the 9 study patients (Table 1; Fig. 1). Technical failure occurred in 2 patients (Fig. 2). In 1 patient with technical failure (patient #1), the poststenting venogram showed residual stenosis of approximately 70% of the normal extrahepatic PV diameter. Because of the risk of anastomotic rupture during balloon angioplasty, we dilated the stenosis using a balloon catheter with half the diameter of the deployed stent. Although the PV inflow was much-improved following balloon angioplasty, stenosis remained in approximately 50% of the normal extrahepatic PV diameter. Repeat balloon angioplasty using a larger diameter balloon catheter was postponed with the expectation of self-expansion of the deployed stent. Therefore, this case was classified as a technical failure. Nonetheless, the stent had expanded spontaneously up to 80% of its normal diameter with patent PV on the 1-week follow-up CT. The patient's liver function also greatly improved following stent placement; however, the patient died of cerebral and intraventricular hemorrhage 24 days after stent placement. In the other patient (patient #2) who had undergone Y-shaped PV anastomoses with the left renal vein after ligation of the renal vein at the junction of the inferior vena cava, the poststenting venogram showed persistent sluggish PV inflow into the right-sided liver graft but no residual stenosis. The 2-day, poststenting CT suggested patent renal venous outflow into the inferior vena cava; left renal venography through the right femoral vein confirmed patent renal venous outflow into the inferior vena cava caused by a loose ligation. PV inflow into both liver grafts improved following repeat surgical ligation of the left renal vein; however, this patient had to undergo retransplantation due to ischemic hepatic necrosis. Pathologic examination of the excised stent showed partial thrombosis. This patient died of cerebellar hemorrhage 3 weeks after retransplantation.
Except for the above-mentioned 2 patients, follow-up Doppler US revealed patent PV inflow in 6 of 7 study patients. In these patients, liver function normalized gradually and the clinical manifestations related to portal hypertension also ameliorated following stent placement. These patients were still healthy and without recurrence at the time this manuscript was completed, and the last follow-up Doppler US or CT obtained 66.6 ± 16.1 months (range 40.1–89.5 months) after stent placement revealed patent PV inflow. In 1 patient (patient #4), the PV inflow improved slightly following stent placement; however, liver function deteriorated daily. In this patient, the interposed cadaveric iliac vein had been misanastomosed with the ligated splenic artery during transplantation. Therefore, surgical reanastomosis of the interposed iliac vein and the superior mesenteric vein was performed using another iliac vein 2 days following stent placement. The stent was then left in situ. Although PV inflow improved following surgical reanastomosis, this patient died of multiorgan failure 15 days later.
Major complications occurred in 3 patients. A total of 2 patients experienced hemoperitoneum caused by blood oozing from a transhepatic tract of the graft; this complication was controlled by surgical ligation. The other patient suffered a large subcapsular hematoma in the liver 14 days after stent placement. Hepatic arteriography demonstrated an intrahepatic pseudoaneurysm adjacent to a previous transhepatic tract. The pseudoaneurysm was embolized using microcoils, and the patient was discharged without sequelae.
The pre- and poststenting pressure gradients across the stenosis were measured in 7 patients. The mean pre- and poststenting pressure gradients were 8.9 ± 5.8 (range 3–18) mmHg and 2.3 ± 2.0 (range 0–6) mmHg, respectively. The prestenting pressure gradient was greater than 5 mmHg in 4 patients and less than 6 mmHg in 3 patients.
Since the first report of PV angioplasty and stent placement after liver transplantation by Olcott et al.,6 percutaneous transhepatic balloon angioplasty has been considered to be a widely accepted, safe, and effective procedure for treating PV anastomotic stenosis following liver transplantation.3, 7–14 Nonetheless, the reported recurrence rate has been relatively high, i.e., 28.6 to 36.8%, following balloon angioplasty alone.3, 11–14 Also, to our knowledge, most patients who underwent balloon angioplasty did so for delayed onset PV stenosis.
To date, little is known about the role of percutaneous interventional procedures for treating early postoperative PV stenosis or thrombosis. Carnevale et al.4 and Cherukuri et al.5 each reported 1 patient with early posttransplantation (on postoperative days 8 and 10, respectively) PV thrombosis; both of these patients were treated successfully with percutaneous thrombolysis or thrombectomy followed by stent placement. In both of these series, balloon angioplasty was also performed without complications, although stent placement was required due to remaining stenosis following balloon angioplasty.
To minimize the potential necessity for repeat surgery and the risk of anastomotic rupture during balloon angioplasty in the early posttransplantation period (less than 1 month), we attempted primary stent placement using uncovered self-expandable stents. We preferred to perform primary stent placement rather than balloon angioplasty for 2 reasons. First, although anastomotic rupture had not occurred in previous reports,4, 5 we could not exclude the possibility of anastomotic rupture of a fresh anastomotic stenosis. Although to our knowledge there have been no reports regarding the optimal interval for balloon angioplasty following living donor liver transplantation, we believe that there may not be sufficient healing of a vascular anastomosis within a month following living donor liver transplantation as the recipients have also undergone immune suppressive treatment. Second, PV stenosis occurring in the early posttransplantation period is considered to be secondary to technical factors such as a tight suture line, discrepancy of the PV size, tension or twisting of the PV in the area of the anastomosis caused by the orientation of the donor liver in the abdominal cavity, or extrinsic compression of the PV caused by hematoma and reactive edema.1, 4, 8, 10 We presume that stenosis caused by such factors, except for a tight suture line, could not be corrected by balloon angioplasty alone. Following primary stent placement in our study, only 1 patient required balloon angioplasty to relieve residual stenosis. Although the etiology of PV stenosis or occlusion was uncertain in our patients, it suggested that the usual etiology of early posttransplantation PV stenosis is not fibrosis or intimal hyperplasia, which might retain an hourglass deformity on the deployed stent.
However, it was challenging to treat early PV stenosis caused by a tight suture line. Poststenting balloon angioplasty was required in 1 patient who we assume had PV stenosis caused by a tight suture line. In this patient (patient #1), we did not achieve technical success because of the risk of anastomotic rupture during balloon angioplasty. However, the stent was shown to have expanded spontaneously up to 80% of its normal diameter at 1-week follow-up CT. Therefore, we assume that self-expandable stents may be valuable for treating early postoperative PV stenosis caused by a tight suture line.
Stents have usually been used to treat recurrent and elastic PV stenoses following balloon angioplasty, as this procedure has several potential complications.3, 8, 11, 12, 15 First, a stent may interfere with future PV anastomoses if retransplantation becomes necessary. However, a stent can be excised at the time of retransplantation or left in situ with the new anastomosis performed at the level of the superior mesenteric vein. Placement of a jump graft from the superior mesenteric vein to the donor PV is also possible.11, 12 In 2 of our study patients, PV stents were also able to be excised during retransplantation or left in situ during reanastomosis. Second, PV thrombosis or stent-edge stenosis may occur, particularly in patients who have undergone left-lobe transplantation due to an angulated course of the PV.8, 11 Therefore, we deployed a stent of sufficient length to achieve smooth alignment of the PV between the recipient and donor PV. We did not then experience such complications during a mean follow-up period of 66.6 ± 16.1 months. In addition, the reported primary patency rate of the PV after stent placement was excellent (100%).11, 15 Third, there is a risk of functional stenosis of a stent in a child's vessel.11 Nevertheless, our 1 pediatric patient had a sufficient PV diameter to allow stent placement, because the liver graft was provided by an adult donor. In fact, we deployed a 10-mm-diameter stent in this patient. During the 41-month follow-up period, stenosis did not recur in this patient. Funaki et al.11 also reported 100% patency in 12 pediatric patients who had undergone stent placement in the PV using 8-mm-diameter stents during a mean follow-up period of 3 yr. Although further long-term follow-up will be necessary to verify the risk of a functional stenosis in a child's vessel, we assume that functional stenosis in pediatric patients is uncommon when a placed stent is greater than 8 mm in diameter. In addition, a balloon-expandable stent will be useful to prevent functional stenosis in a growing child's vessel.16
In our study, clinical failure occurred in 2 patients and was caused by a loose ligation of normal renal venous outflow and misanastomosis of the interposed iliac vein graft, respectively. In fact, as immediate surgical revision was chosen for these 2 patients, stent placement was not necessary. Except for these patients, our clinical success rate was 100%. Therefore, it is clear that careful review of the surgical and imaging data is important in order to determine the appropriate treatment regimen.
A pressure gradient greater than 5 mmHg across a stenosis has been considered “significant” in some reports7, 10, 12; however, there is no current standard definition regarding a significant pressure gradient. In our 3 study patients, the gradient was less than 6 mmHg. These patients had abnormal liver enzyme values or clinical manifestations related to portal hypertension and showed clinical improvement following stent placement. Several investigators have also reported clinically successful cases following balloon angioplasty or stent placement for treating PV stenosis with a pressure gradient less than 6 mmHg.8, 10, 14 Therefore, we assume that the pressure gradient does not seem to be directly correlated with the clinical results, and treatment is valuable in the early posttransplantation period if patients have symptoms related to PV inflow abnormality or portal hypertension even though the pressure gradient is not significant.
We experienced postprocedural bleeding through a transhepatic tract in 2 of our study patients. Although surgical treatment of such bleeding is easier than revision of the PV anastomosis, further investigation will be necessary in order to avoid the risk of bleeding through a transhepatic tract. The use of a transjugular intrahepatic portosystemic shunt–type approach may minimize the risk of bleeding from the transhepatic tract.5 In addition, embolization of the transhepatic tract using a histoacryl-lipiodol mixture may be effective in preventing bleeding, although there are as yet no clinical studies.17, 18
In summary, although our study had a relatively small number of patients, it achieved acceptable technical and clinical results. Although there were several postprocedural complications, they were cured using surgical ligation or transarterial embolization without sequelae. In addition, the long-term PV patency following stent placement was excellent. Therefore, we assume that percutaneous transhepatic primary stent placement is a useful and generally safe treatment for early posttransplantation PV stenosis. Of course, additional experience and further studies will be needed to verify the success of this technique.
We thank Bonnie Hami, M.A. (USA), for editorial assistance in preparing this manuscript.