Objectives: The hepatic vasculature is a unique system due to a dual supply that includes the hepatic artery and portal vein, which interact when the liver vascular supply is decreased. Hepatic artery buffer response, an intrinsic regulatory mechanism that compensates for blood supply, maintains increased hepatic artery flow and caliber in response to portal vein failure. Previous studies revealed that portal vein flow showed no alterations to establish adequate blood supply in response to hepatic artery occlusion. Here, we analyzed portal vein flow changes in patients with hepatic artery thrombosis after liver transplant.
Materials and Methods: From December 1988 to October 2017, our center performed 580 liver transplant procedures. Those diagnosed with hepatic artery thrombosis (19 females, 24 males) by Doppler ultrasonography during postoperative week 1 were analyzed. Patients received either surgery or endovascular treatment for hepatic artery thrombosis, with patency confirmed by Doppler ultrasonography. We compared portal vein flow velocity and caliber before and after treatment using Wilcoxon signed rank and Mann Whitney U tests.
Results: Mean patient age was 18.9 ± 21.4 years. Portal vein flow velocity pretreatment (median of 70 cm/s) was significantly higher than posttreatment (median of 52 cm/s) in all patients (P < .001). Median flow velocity decreased significantly after treatment when subgroups were compared, including age (adult vs child), transplant type (orthotopic transplant vs living donor), and treatment (surgery vs endovascular). However, portal vein flow velocity showed a significantly higher decrease in the surgery subgroup than in the endovascular treatment subgroup (P = .018). There was no significant relationship between portal vein calibers before and after treatment (P = .36).
Conclusions: The significant decrease in portal vein flow velocity after successful treatment of hepatic artery thrombosis may represent a compensatory flow change of the portal vein in response to diminished hepatic artery flow.
Key words : Doppler ultrasonography, Hepatic arterial buffer response, Vascular complication
Liver transplantation is an acceptable therapeutic option for end-stage liver diseases. However, complications occurring both early and late posttransplant can affect patient survival and graft viability. The most common complications are vascular complications, which include thrombosis or stenosis, biliary obstruction or leakage, hepatic infarction, and hemorrhage, resulting in rejection early after liver transplant.1 Vascular complications, with an incidence of 9%, may occur in either the arterial or venous system.
Hepatic artery (HA) thrombosis is the most common and severe complication after liver transplant. Thrombosis of the HA occurs in 4% to 12% of adult and 42% of pediatric liver transplant recipients and is associated with a significant mortality rate of 20% to 60%.2,3 Early diagnosis of HA thrombosis is an important issue to prevent the necrosis of liver parenchyma or biliary ischemia and to establish graft viability.
Doppler ultrasonography is considered as the primary imaging modality in the diagnosis of postoperative vascular complications in liver transplant.4 Ultrasonography provides information about the flow pattern, velocity, resistive index, and the caliber of HA. In addition, the hepatic arterial system and the portal venous and hepatic venous systems may be evaluated by Doppler ultrasonography synchronically. The hepatic vasculature is a unique and complicated system due a dual supply that includes the HA and portal vein (PV), which may interact to maintain adequate vascular supply in the liver.5 About 75% of the blood flow entering the liver is supplied by the PV, and the remaining 25% is established by the HA.5
Regulation of liver blood flow is maintained by several mechanisms.5,6 Hepatic arterial buffer response (HABR) is an intrinsic regulation mechanism that is based on the relationship between HA and PV.5-7 According to this mechanism, reduced blood flow in the PV leads to increased blood flow and caliber of the HA due to accumulation of adenosine concentration in the space of Mall.5,6 Similarly, the HA may constrict in response to PV hyperperfusion by the HABR mechanism.7 Although the HA presents compensatory flow changes due to decreased flow in the PV, several studies have revealed that there is no reciprocity of the HABR mechanism in which PV flow did not produce any compensatory response during impaired HA flow.5-9 Here, we analyzed PV flow changes in patients diagnosed with HA thrombosis after liver transplant to determine whether the PV system presents any response to HA thrombosis.
Materials and Methods
Our retrospective study was approved by the Institutional Review Board of Baskent University Hospital (approval number: KA18/77). From December 1988 to October 2017, our center performed 580 liver transplant procedures. Of these, 43 recipients (19 females, 24 males) with HA thrombosis, diagnosed by Doppler ultrasonography during posttransplant week 1, were included in this study. Surgery or endovascular treatment (EVT) was used for treatment of HA thrombosis, with HA patency then confirmed by Doppler ultrasonography. Flow velocity (FV) and caliber measurements of PV before and after treatment were obtained from our database. For analysis, we also divided patients into 3 subgroups: (1) age (child or adult), (2) type of liver transplant (orthotopic liver transplant [OLT] or living-donor liver transplant [LDLT]), and (3) treatment option (surgery or EVT).
Gray-scale, color, and pulsed Doppler ultrasonographs of HA were examined using a 9L4 MHz linear-array transducer system in pediatric recipients and 2 to 5 MHz and 4C1 MHz convex transducer systems in adult recipients (Siemens Sonoline Versa, Siemens Sonoline Elegra, Siemens Sonoline Antares, Munich, Germany) before 2010. After 2010, patients were examined with the Siemens Acuson S2000 ultrasonography system (9L4 MHz linear-array transducer for pediatric recipients and 6C1 MHz convex transducer for adult recipients). Ultrasonography examinations were performed with patients in a supine position both before and after treatment. Diagnosis of HA thrombosis was made by observation of complete absence of HA flow and waveform by color and pulsed Doppler ultrasonography. After treatment with surgery or EVT, HA patency was confirmed by Doppler ultrasonography within 1 hour posttreatment. The treatment was defined as successful if flow and optimal waveform pattern were observed, with FV > 20 cm/s and resistive index > 0.5 in the HA. Caliber and flow pattern of PV were also examined by Doppler ultrasonography at diagnosis of HA thrombosis and after treatment. All measurements of FV and caliber of PV were obtained from our institutional database.
Patients with HA thrombosis were treated with either surgery or EVT. During surgery, laparotomy with an upper abdominal transverse incision was performed, with anastomosed site evaluated carefully to determine the surgical option for effective treatment. First, the HA was washed with heparinized saline solution. When no improvement of blood flow was observed, a thrombectomy procedure was performed by Fogarty catheter, which was inserted into the anastomosed region. If no improvement in blood flow was shown after thrombectomy, reanastomosis or revascularization with vascular conduits was performed.
During EVT, a microcatheter and a 0.016-inch guide wire were positioned within the thrombosed HA via a 5F diagnostic catheter. Angiography was performed to evaluate the patency of distal branches. The microcatheter was advanced distally, and a 0.014-inch guide wire was placed to the distal part of the thrombus. A 6F guiding catheter was then positioned into the origin of the celiac trunk, and a 4F catheter was inserted into the HA for thrombolytic infusion. Infusion was delivered for a few days to treat the residual thrombus. The use of balloon angioplasty or bare or graft-covered stent placement was assessed for treatment of the underlying anatomic defects.
Statistical analyses were performed with SPSS software (SPSS: An IBM Company, version 22.0, IBM Corporation, Armonk, NY, USA). The Kolmogorov-Smirnov test was used to analyze normal distribution of data. All measurements are expressed as means with 2 standard deviations, medians, and minimum and maximum values. We used Wilcoxon signed rank test to compare the pretreatment and posttreatment measurements of FV and caliber of PV in all cases. We also compared pretreatment and posttreatment measurements of FV and caliber of PV within the 3 subgroups (surgery vs EVT treatment, OLT vs LDLT, and pediatric vs adult recipient) by using paired sample t test for normally distributed data and Wilcoxon signed rank test for nonnormally distributed data. We also calculated decreasing degree of FV with the following formula: 1 – FVpost/FVpre, where FVpre is the pretreatment value and FVpost is the posttreatment value. Differences in decreasing degree of FV between the 3 subgroups were analyzed with the Mann-Whitney U test. P values < .05 were statistically significant.
Patient ages ranged from 1 to 63 years (mean age of 18.9 ± 21.4 y). Patients were divided into the 2 age subgroups (< 18 y [pediatric subgroup] and ≥ 18 y [adult subgroup]). In the pediatric subgroup (n = 26), patients ranged from 1 to 18 years old (mean age of 4.3 ± 4.4 y). In the adult subgroup (n = 17), patients ranged from 19 to 63 years old (mean age of 41.3 ± 17.2 y). In the adult subgroup, 11 patients had right lobe graft, 3 patients had left lobe graft, and 3 patients had deceased allograft for liver transplant. In the pediatric subgroup, 23 patients had left lobe lateral segment and 3 patients had left lobe graft for liver transplant.
In the pediatric subgroup, causes of liver diseases included biliary atresia (n = 8), urea cycle disorders (n = 4), Alagille syndrome (n = 2), progressive familial intrahepatic cholestasis (n = 4), lysosomal storage disease (n = 1), acute fulminant toxic hepatitis (n = 1), Budd-Chiari disease (n = 1), primary sclerosing cholangitis (n = 1), Crigler-Najjar syndrome (n = 1), tyrosinemia (n = 1), Wilson disease (n = 1), and cryptogenic cirrhosis (n = 1). In the adult subgroup, causes of liver diseases were cryptogenic cirrhosis (n = 6), Budd-Chiari disease (n = 1), primary biliary cirrhosis (n = 1), Wilson disease (n = 3), autoimmune hepatitis (n = 1), hepatitis B virus-related cirrhosis (n = 4), and hepatitis C virus-related cirrhosis (n = 1).
Eighteen patients required surgery for HA thrombosis treatment (which included washing artery with heparinized saline solution, thrombectomy, revascularization with vascular conduits, and reanastomosis). Thrombectomy by Fogarty catheter was required for 6 patients, and reanastomosis with native HA was required for 3 patients to establish HA patency. In 6 patients, vascular conduits (iliac artery, left gastric artery, splenic artery, ovarian vein, greater saphenous vein, and graft of pig pericardium) were assessed for revascularization. Three patients required only washing the artery with heparinized saline solution in open surgery to establish HA patency.
Twenty-five patients required EVT options for HA thrombosis treatment, which included intra-arterial thrombolytic therapy (IATL), percutaneous transluminal angioplasty (PTA), and stent placement. Patency of HA was established by IATL alone in 10 patients, IATL and PTA in 7 patients, and IATL, PTA, and stenting in 8 patients.
The median overall PV FV value before treatment was 70 cm/s (minimum: 25 cm/s; maximum: 300 cm/s), which was significantly higher than the median overall value posttreatment (52 cm/s; minimum: 20 cm/s; maximum: 135 cm/s) (P < .001). However, the median overall value of PV caliber before treatment (10 mm; minimum: 8 mm; maximum: 18 mm) was not significantly different from the median PV caliber value posttreatment (10 mm; minimum: 7 mm; maximum: 21 mm) (P = .36). When we viewed the results between subgroups, the median values of PV FV were significantly higher before treatment than after treatment in all subgroups (age, treatment option, and transplant type) (Table 1). However, no significant difference was found between pretreatment and posttreatment median values of FV in 3 adult patients with OLT (P = .1) (Table 1). We also found no significant differences between pretreatment and posttreatment PV caliber results in all subgroups (adult vs child, surgery vs EVT, and OLT vs LDLT) (Table 1).
Table 2 shows the decreasing degree in FV of PV after treatment in all subgroups. We found no significant difference in the decreasing degree of FV of PV after treatment between the pediatric and adult subgroups (P = .75). We also observed no significant difference in decreasing degree of FV between the LDLT and OLT subgroups after treatment (P = 0.54). When we compared the decreasing degree of FV of PV after treatment between the surgery and EVT subgroups, we observed a significantly higher decreasing degree of PV FV in the patients who underwent surgical procedures than in those who had EVT (P = .018).
Constancy of liver blood flow is essential to maintain graft viability after liver transplant. Several hepatic hemodynamic mechanisms, including HABR, provide constancy of hepatic blood flow with regard to flow interactions between the HA and PV.5-7 Previous studies have mainly revealed that PV system failure, including stenosis or thrombosis, contributed to the compensatory changes in the HA system (ie, HABR).5-7 However, no reciprocity of the HABR has been reported in the PV system in cases of HA obstruction in previous reports.5-8
Liver transplant is the best treatment option for end-stage liver disease in both pediatric and adult patients. Although the results of liver transplant are more satisfactory in children than in adults, complications, including vascular and biliary complications, more commonly occur in children.10 In pediatric recipients, segmental liver transplant, split deceased-donor transplant, and whole organ deceased-donor transplant are the most preferable techniques for liver transplant. The left lobe and left lobe lateral segment are the major graft types for pediatric patients.11 In our study, left lobe lateral segment transplant was performed in most of the pediatric patients (n = 23), with left lobe transplant performed in only 3 pediatric patients. Most of our adult patients received right lobe LDLT (64.7%).
In both adult and pediatric patients, HA thrombosis features at Doppler ultrasonography include the complete absence of HA and intrahepatic flow of HA.12,13 However, no significant flow changes in the PV system have been previously reported.3,12-14 The PV system cannot regulate its blood flow, and splanchnic vascular flow and intrahepatic resistance affect PV pressure and flow.5 In the early posttransplant period, the PV system generally shows flow changes due to splanchnic flow and postoperative edema or fluid collections. In addition, the FV of PV may range from 15 to 400 cm/s in the early posttransplant period.12 In the present study, we found that the FV of PV ranged from 25 to 200 cm/s in our pediatric subgroup and from 48 to 300 cm/s in our adult subgroup at diagnosis of HA thrombosis. After treatment of HA thrombosis, FV of PV significantly decreased in both pediatric and adult transplant recipients. We also found no differences in the decreased degree of FV between our pediatric and adult subgroups. As mentioned, PV flow depends primarily on changes in the splanchnic and mesenteric vascular system; however, we found that FV of PV decreased significantly after successful treatment of HA thrombosis. This could be because the PV system may present compensatory changes when HA flow decreases to maintain the liver vascular supply. During HABR, besides HA flow changes, dilatation of HA occurs in the obstruction of PV to establish adequate blood supply. However, we observed no significant chang in the caliber of PV after treatment of HA thrombosis in both our adult and pediatric subgroups.
There are 3 main transplant types: OLT, LDLT, and split type. Among those, OLT is the most common type. In our study, most patients in the adult subgroup had LDLT (n = 40), with the remaining patients having OLT (n = 3). In the pediatric subgroup, all patients had LDLT (left lobe lateral segment in 23 patients and left lobe in 3 patients). In LDLT subgroup, FV of PV was significantly decreased after treatment compared with before treatment. However, no significant change in FV of PV was found in the OLT subgroup. This may be related to our small sample size (n = 3). We also observed that PV caliber did not significantly change after treatment in both the LDLT and OLT subgroups.
There are 2 major therapeutic options for HA thrombosis: open surgery and EVT. Open surgery includes thrombectomy, reanastomosis, revascularization with vascular conduits, and retransplant, and EVT includes IATL, PTA, and stent placement.15,16 Ischemic time of liver should be limited to prevent graft loss with HA thrombosis. Although the intervention should be performed in the fastest way possible, during open surgery, adhesion and inflammation in the anastomosis site may sometimes cause difficulty in reaching the anastomosis site and may result in loss of time for effective treatment.15,17 However, after the anastomosis site is reached, hematoma near the anastomosis or kinking (causes of HA obstruction) may be effectively removed by open surgery. If patency is not maintained by those interventions, thrombectomy or reanastomosis should be preferred to provide graft viability. Nevertheless, risks of open surgery, including bleeding or intestinal perforation, may limit the preference of open surgery.15-17
Endovascular treatment options may offer advantages of being easy to perform and allow the thrombosis site in the HA to be quickly reached; this option can also be used during serial interventions for residual and refractory lesions. However, complications include rupture, dissection, distal embolism, and bleeding complications due to the IATL. In pediatric cases, the technique is difficult because the small caliber of the pediatric vessels may restrict performance of EVT procedures.15,18 In our study, the patency of HA was maintained by open surgery in 18 patients and EVT in 25 patients. In our surgery subgroup, 6 cases (33.3%) required vascular conduits, 3 cases (16.6%) required reanastomosis with native HA, 6 cases (33.3%) required thrombectomy, and 3 cases (16.6%) required washing the HA with heparinized saline. For HA thrombosis, revascularization with vascular conduit is the most preferred treatment option due to the advantages of having better arterial flow and lower recurrence risk compared with thrombectomy and reanastomosis of native HA or EVT procedures.15,19,20
In addition to HA flow alterations after successful treatment of HA thrombosis, the PV may also present flow changes during HA failure. We found that FV of PV decreased significantly after treatment in both the surgery and EVT subgroups. Moreover, we observed that the decreasing degree of FV of PV after treatment was significantly higher in patients who were treated with open surgical options versus EVT. The higher ratio of the decreasing degree of FV of PV in the surgery subgroup may be related to a better potential for reestablishment of the arterial inflow during surgery compared with EVT.
Our study has several limitations. First is its retrospective design. Because of its retrospective design, intra- and interobserver variability tests and test-retest reliability were not analyzed. Second, the duration of liver disease, which may affect the hepatic hemodynamic relationship between HA and PV, was not evaluated in our study. Third, we did not analyze the flow volume of PV, which may provide additional information about PV flow changes besides the FV of PV. Finally, our small sample size could decrease the reliability of measurements of PV flow changes and caliber before and after treatment.
Although our results did not show an absolute reciprocity of HABR mechanism in the flow change of PV after HA thrombosis, the change in the FV of PV after successful treatment of HA thrombosis may represent a compensatory response of PV in cases of diminished HA flow. In addition, we observed that changes in FV of PV after treatment of HA thrombosis were significantly higher in patients who were treated with open surgery versus those treated with EVT.
DOI : 10.6002/ect.2018.0128
From the 1Department of Radiology and the 2Department of General Surgery,
Baskent University School of Medicine, Ankara, Turkey
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Sehnaz Tezcan, Guvenevler mah, Kibris sok, 9/17 Kavaklidere, Ankara, Turkey
Phone: +90 533 455 55 33
Table 1. Portal Venous Flow Velocity and Caliber in All Subgroups
Table 2. Decreasing Degree of Portal Venous Flow Velocity After Treatment in All Subgroups
Figure 1. Color Doppler Ultrasonographic Images in a 12-Month-Old Boy With Biliary Atresia Who Underwent Living-Donor Liver Transplant With Left Lobe Lateral Segment
Figure 2. Color Doppler Ultrasonographic Images in a 12-Month-Old Boy With Biliary Atresia Who Underwent Living-Donor Liver Transplant With Left Lobe Lateral Segment