Objectives: Our aim was to evaluate the eligibility of hepatic venous drainage technique using circum-ferential fence with peritoneal patch graft and saphenous vein graft in right lobe living-donor liver transplant.
Materials and Methods: Between November 2007 and December 2017, our center performed 1413 right lobe living-donor liver transplants. A circumferential fence was created using the peritoneal patch graft for venous drainage reconstruction in 31 of these patients. We compared data of these 31 patients with data of 62 patients who had circumferential fence created with cryopreserved homologous saphenous vein graft (1:2 ratio). Patients with anastomotic stenosis (n = 10) and without anastomotic stenosis (n = 69) were also compared.
Results: No statistically significant differences were found between the peritoneal patch graft group and the saphenous vein graft group in terms of clinical parameters, circumferential fence diameter, and postoperative anastomotic stenosis (7.1% vs 26.1%; P = .056). Postoperative anastomotic stenosis developed in 10 patients. No statistically significant differences were found between patients with and without anastomotic stenosis in terms of clinical parameters except for diameter of circumferential fence (P = .001). Diameter of circumferential fence less than 18 mm in patients with and without anastomotic stenosis was shown in 80% and 34.6% of patients, respectively (P = .02). Stenosis-free survival in saphenous vein graft group was significantly better than in the peritoneal patch graft group (P = .023). Our correlation analyses showed a strong correlation between diameter of circumferential fence and the stenosis rate in patients (P = .001).
Conclusions: Autologous peritoneal patch graft can be used for creating circumferential fence for hepatic outflow in living-donor liver transplant patients if no better option is available. Despite high anastomotic stenosis rates, peritoneal fencing responded well to radiologic dilatation or stent applications and provided good long-term patency.
Key words : Hepatic venous outflow reconstruction, Peritoneal patch graft, Right lobe liver graft, Saphenous vein graft
Living-donor liver transplant (LDLT) was first performed in 1989; since then, it has been the standard therapy for chronic liver failure and other liver diseases in countries with insufficient organ supply from deceased donors.1 Biliary and vascular reconstruction in LDLT is relatively difficult compared with deceased-donor liver transplant. There are some vital factors that have an impact on satisfactory results in LDLT, which are as follows: (1) proper graft volume, (2) sufficient portal venous flow, (3) safe biliary duct anastomosis, and (4) optimal venous drainage.2
The vascular anatomy of the right lobe (RL) liver graft is relatively complex, and the venous drainage is through the right hepatic vein (RHV), right inferior hepatic vein (RIHV), and middle hepatic vein (MHV) tributaries (segment 5 and 8 veins [V5 and V8]). For this reason, to obtain optimal RL liver graft function and to avoid graft congestion, graft function failure, and even mortality, any venous tributaries greater than 4 to 5 mm should be integrated into the venous drainage system.2-5 Lee and associates have defined many venous drainage techniques used in RL LDLT.2-5 Artificial vascular grafts (polytetrafluo-roethylene [PTFE] and polyethylene terephthalate [Dacron]), cryopreserved homologous grafts (saphenous vein, iliac artery, iliac vein, vena cava inferior, and aorta), and rarely autologous vascular grafts are used for venous drainage reconstruction of the RL liver graft.2-5 In many ways, cryopreserved homologous vascular grafts are ideal materials. However, there are difficulties in the supply of these grafts in countries with insufficient deceased donors.6 Therefore, many centers that perform LDLT use artificial vascular interposition grafts to reconstruct RIHV and MHV tributaries to overcome the shortage of biologic vascular grafts. Homologous saphenous vein grafts (SVG) are used to form the circumferential fence in the last and most important stage of the venous drainage model in RL LDLT patients. According to our knowledge, only one study has reported hepatic venous reconstruction using the round ligament obtained from a living liver donor.6
High-volume LDLT centers such as ours can experience shortages of homologous vascular grafts. For these cases, autologous peritoneal patch graft (PPG) can be used to construct the circumferential fence. In our center, we used autologous PPGs due to obligatory reasons, including when there was insufficient supply of homologous/autologous vascular grafts at the time and patient outcomes were found to be good.7,8 We observed that autologous PPG can be safely used in circumferential fence construction in necessary conditions when vascular grafts are not available. Furthermore, our literature search showed no studies reporting the use of autologous PPG for any stage of hepatic venous reconstruction in LDLT. In the present study, we report the results of 31 patients who had autologous PPG used for circumferential fence construction, which we compared to a selected control group of patients.
Materials and Methods
Between November 2007 and December 2017, our center (Inonu University Institute of Liver Transplantation) performed a total of 2034 liver transplant procedures. Of these, 1653 received LDLT and 381 received deceased-donor liver transplant. A RL liver graft was used in 1413 LDLT patients, and a left or lateral lobe liver graft was used in 240 LDLT patients.
Objectives and study design
The primary objective of our study was to compare 31 patients who had autologous PPG for cir-cumferential fence construction during RL LDLT (PPG group) with 62 RL LDLT patients who had SVG in terms of postoperative anastomotic stenosis and anastomosis diameter. We aimed to test whether PPG could be used in the reconstruction of the circumferential fence in RL LDLT in obligatory situations and to propose an alternative way of venous reconstruction by using autologous PPG.
The 62 patients who received RL LDLT with SVG circumferential fence reconstruction were the control group. The PPG group (n = 31) was matched at random in a 1:2 ratio with the control (SVG) group (n = 62). To minimize the risk of bias, patients in the control group were selected by a surgeon who was not involved in this study. Patient selection criteria and study design are illustrated in Figure 1. Exclusion criteria were as follows: follow-up period less than 30 days, mortality within 30 days, and patients whose radiologic images were unavailable.
The PPG and SVG groups were compared in terms of parameters such as age, male versus female, Model for End-Stage Liver Disease (MELD) score, Child-Pugh score and group (B, C), graft volume, diameter of circumferential fence, radiologic-proven anastomotic stenosis, days to development of anastomotic stenosis, and follow-up time. Patients with anastomotic stenosis (n = 10) and without anastomotic stenosis (n = 69) were also compared according to the parameters mentioned above.
The craniocaudal diameter of the anastomosis formed between the circumferential fence and the recipient hepatic vein stump was measured in the first coronal maximum intensity projection image obtained from the hepatic phase of a multidetector computed tomography (MDCT) scan during the postoperative period. Results were given as diameter in millimeters of circumferential fence. Follow-up time was calculated from date of RL LDLT to the last outpatient clinic visit or mortality date. Duration to development of anastomotic stenosis was calculated as date of RL LDLT to date of hepatic venous outflow obstruction detected with conventional hepatic venography.
Venous drainage reconstruction on back table
Procedures until the preparation of the venous reconstruction after the RL liver graft at the back table are detailed in our previous study.9 In previous years, we used the all-in-one technique to construct the venous outflow of the RL liver graft.9 For this technique, we used the homologous or synthetic vascular grafts to extend RIHV and MHV tributaries (V5, V8) that were larger than 4 to 5 mm to the RHV orifice (Figure 2A). Later, to obtain a single large venous orifice, an SVG was used to construct the circumferential fence. The cryopreserved SVG was split open and circumferentially sutured to the RHV plus combined elongated V8 + V5 + RIHV like a fence (Figure 2B to 2D). However, our experience showed that this technique prolonged the cold ischemia time, and we started a new technique of venous reconstruction for a shorter back-table period. For this technique, all RIHV that are far from RHV are anastomosed to the inferior vena cava lateral wall separately. The MHV tributaries (V5, V8) that are far away from the RHV are reconstructed by artificial vascular grafts (Dacron, PTFE) and anastomosed to the left hepatic vein stump of the recipient separately. The RHV and adjacent venous structures, such as RIHV and V8, are reconstructed into a single orifice by cryopreserved SVG, which is used for circumferential fence construction, and this single orifice is anastomosed to the recipient’s RHV stump. In the beginning of our program, we used cryopreserved homologous vascular grafts and PTFE artificial grafts for venous reconstruction. Currently, we use Dacron to reconstruct the MHV tributaries because of the ease of handling and availability of various sizes and shapes of these grafts.
In our center, we perform over 200 LDLTs annually. Our vascular graft repository consists of vascular grafts obtained from deceased donors (cryopreserved vascular grafts) and SVGs harvested by cardiovascular surgery during coronary artery bypass grafting as well as artificial vascular grafts. Between 2013 and 2014, we performed 250 such procedures. For 31 of these cases, we used PPGs for construction of circumferential fence during venous outflow reconstruction due to lack of homologous vascular graft supply. In this group, all procedures during venous reconstruction were the same as procedures for the SVG group. Briefly, autologous PPGs were removed from the peritoneal layer covering the right upper quadrant of the recipients (Figure 3A). Next, the PPGs were tailored (Figure 3B) and sutured into a circumferential fence to the hepatic vein orifices in the back table (Figure 3C and 3D). After construction of the circumferential fence with PPG, anastomosis was performed between the RL liver graft and the recipient’s inferior vena cava (Figure 4).
Postoperative course and assessment of hepatic venous outflow obstruction
Our standard intraoperative and postoperative medical therapy and follow-up procedures are mentioned in detailed in our previous studies.9,10 Low-molecular-weight heparin was given until the patient became ambulatory. After ambulation, patients were given 100 mg of acetyl salicylic acid and low-molecular-weight heparin was discontinued. Patients were evaluated by Doppler ultrasonography every day for the first 5 postoperative days. In patients with uneventful postoperative follow-up, before discharge (coinciding to postoperative days 9 to 12), MDCT was performed for routine evaluation. In patients who developed postoperative jaundice, ascites, elevated liver function tests, edema in the extremities, or respiratory distress, patients were initially evaluated by Doppler ultrasonography. If a low flow rate of < 10 cm/s or if persistent monophasic waves were determined in the hepatic veins, patients underwent MDCT scans to evaluate the RL liver grafts. Hepatic venography was performed for diagnostic or therapeutic purposes if the MDCT revealed that the hepatic vein was not contrast enhanced, if more than 50% stenosis in the anastomosis was shown, or if low attenuation or a granular pattern was shown in the liver parenchyma, which would suggest hepatic venous outflow problems. Patients were considered to have severe hepatic venous outflow obstruction if venography revealed anastomotic stenosis. After the venography procedure, patients were again evaluated with Doppler ultrasonography, MDCT, and clinical findings, such as remission in ascites and reduction in body weight. Patients with improved clinical findings were followed up every month. In patients with moderate response, repeated balloon angioplasty and/or stent placements were performed.
The statistical analyses were performed using IBM SPSS Statistics version 25.0 (Statistical Package for the Social Sciences, Inc., Chicago, IL, USA). The quantitative variables are expressed as means and standard deviation, median, minimum to maximum (range), and interquartile range. The qualitative variables are reported as numbers and percentages. Shapiro-Wilk and Kolmogorov-Smirnov tests were used to assess normality distribution of quantitative variables. The Mann-Whitney U test was used to compare quantitative variables. Pearson chi-square and Fisher exact tests were used to compare qualitative variables. The point biserial correlation coefficient (Eta statistic) was used to show whether there was an association between diameter of circumferential fence and presence of anastomotic stenosis. Receiver operating characteristic analysis was performed to identify optimum cutoff values of diameter of circumferential fence. The cutoff value for diameter of circumferential fence was then determined to obtain the most ideal sensitivity and specificity. Kaplan-Meier survival estimates were used to determine stenosis-free survival. P values < .05 were considered statistically significant.
Eight patients from the SVG group and 6 patients from the PPG group were excluded because they did not meet the eligibility criteria of the study. Therefore, 79 patients with a median age of 52 years (range, 17-69 y) were evaluated for the study. Median age of the 23 patients in the PPG group (n = 23) and SVG group (n = 56) were 52 years (range, 17-62 y) and 52 years (range, 21-69 y), respectively. We observed no statistically significant differences between the PPG group and the SVG group in terms of age (P = .409), male versus female (P = .270), Child-Pugh score (P = .545), graft volume (P = .117), diameter of circumferential fence (P = .969), duration to development of anastomotic stenosis (P = .240), and follow-up time (P = .197). We observed a significant difference regarding the MELD score between groups (P = .034). Four patients (7.1%) in the SVG group developed postoperative anastomotic stenosis in a median of 487 days (range, 333-807 d) posttransplant, whereas 6 patients (26.1%) in the PPG group developed postoperative anastomotic stenosis in a median of 425 days (range, 163-600 d). Although differences in postoperative anastomotic stenosis rates among the groups did not reach statistical significance, the P value was .056, which is close to the significance margin (Table 1). We found that stenosis-free survival was significantly better in the SVG group (P = .023) (Figure 5).
The patients were grouped according to devel-opment of anastomotic stenosis, with 10 patients in the stenosis group and 69 patients in the nonstenosis group. We observed no statistically significant differences in terms of age (P = .337), male versus female (P = .268), MELD score (P = .929), Child-Pugh score (P = .256), and graft volume (P = .110) among the patients who did and did not develop anastomotic stenosis. However, diameter of the circumferential fence was significantly wider in patients who did not develop anastomotic stenosis (median of 11.6 mm) compared with patients who developed stenosis (median of 20.2 mm; P = .001). Our receiver operating characteristic curve analysis revealed that the cutoff value for circumferential fence diameter for deve-lopment of stenosis was 18 mm (area under the curve of 0.823; P = .001; specificity of 80.0% and sensitivity of 72.1%) (Figure 6). Eight patients (80%) in the stenosis group and 19 patients (34.9%) in the group without stenosis had a circumferential fence diameter of ≤ 18 mm (P = .002) (Table 2). Eta correlation analyses revealed a strong positive relationship between the anastomotic stricture and anastomosis diameter (P = .001, Eta = 0.971). Furthermore, we found that 94.3% of the stenosis were related with anastomosis diameter (Eta-squared = 0.943) (Figure 7).
Right lobe LDLT provides optimal liver volume for recipients during liver transplant; therefore, it is widely used in specialized centers around the world during LDLT. Although RL liver grafts have many advantages for LDLT, variations in venous anatomy have forced researchers to develop venous recon-struction models to overcome venous drainage problems encountered during RL LDLT. To prevent venous stasis in RL liver grafts, proper integration is needed of MHV tributaries (V5, V8) and RIHV to the inferior vena cava drainage system.3
There are 3 strategies to reconstruct the venous drainage of the RL liver graft. The all-in-one technique requires extension of MHV tributaries (V5, V8) and RIHV to RHV orifice by interposition vascular graft, after which the circumferential fence is formed by a SVG and the anastomosis to the recipient’s RHV stump is performed. The second venous reconstruction strategy involves extension of MHV tributaries (V8) and RIHV that are close to RHV to the graft RHV orifice, with the remaining procedures similar to those described above. The venous tributaries that are far from RHV are anastomosed separately to the inferior vena cava. The third strategy is the one that we currently use for reconstruction of MHV tributaries (V5 and V8) with a Dacron Y graft. Anastomosis is performed to the left hepatic vein stump in the recipient, RHV of the RL liver graft is anastomosed to the recipient’s RHV stump, and RIHV is anastomosed separately to the inferior vena cava. All 3 strategies have their advantages and disadvantages. A few centers have proposed Y-graft reconstruction during the implantation of the RL graft to decrease the cold ischemia time.4 However, we chose to perform all venous reconstruction procedures in the back table to decrease the warm ischemia-related injuries in our patients.
The aim of our present study is the endpoint of the above-mentioned procedures, which is the circumferential fence reconstruction that aims to form a single orifice out of the venous tributaries.
We choose to use flexible and easily manipulated materials for reconstruction of the circumferential fence. For this reason, the most frequently used material is cryopreserved SVGs.3-5 However, obtaining a sufficient supply of SVGs in high-volume liver transplant centers is difficult. A few studies have reported the use of the recipient’s own SVG as an alternative.4 The use of the SVGs that remain after the cardiovascular surgical procedures is also an alternative. Remnant vascular grafts in centers with high volumes of coroner artery bypass grafting is another good supply of cryopreserved SVGs. However, we believe that additional surgical procedures to harvest SVGs during a major surgery such as liver transplant should be considered as a last step measure. Therefore, in centers with high-volume LDLTs, alternative strategies for reconstruction of the circumferential fence should be sought. From our point of view, the peritoneum with its single-layer cuboid epithelium or autologous vascular grafts is a good alternative instead of the use of synthetic vascular grafts such as PTFE and Dacron. We used Dacron grafts during circumferential fence construction in a few cases; however, the flexibility and manipulation of these materials were more difficult than the use of SVGs or PPGs.10
We were obligated to use PPGs for circumferential fence reconstruction during a period of low cryopreserved vascular supply. Our experience from this period led us to evaluate the efficacy of PPG in venous reconstruction in RL liver grafts. Because no study so far has investigated the role of PPGs, we were unable to compare our results with other studies. In our present study, the increased hepatic vein outflow obstruction rate that we observed in the PPG group was at the statistically significant margin. We believe this is due to the more flexible or pliable nature of SVGs compared with PPGs. Peritoneal patch grafts are rigid, and a circumferential patch construction leads to a cylindrical orifice. The anastomosis diameter is further narrowed once the wound healing with fibrosis takes place. On the other hand, SVGs are more pliable and anastomosis to the inferior vena cava is wider, leading to a conical shape being formed. Because the connective tissue layer of the SVG is low, the contraction level after healing may be expected to be low.
There are a few limitations of our study. First, our study was designed as a retrospective case control study. There is a risk of bias, as in all retrospective studies, even if we tried to minimize it. In addition, we did not measure the craniocaudal diameter of any circumferential fence that was constructed in the back table. In other words, we could not comment regarding the flexibility of the SVGs and PPGs.
Our present results suggested that a PPG is not flexible enough to absorb the venous pressure observed in vena cava anastomosis; therefore, it should not be used as a first option in vascular reconstruction in LDLT. However, the circumferential fence created with the PPG seems to be a reasonable alternative in obligatory situations when no other graft is available.
DOI : 10.6002/ect.2018.0325
From the 1Department of Surgery and Liver Transplant Institute, and the
2Department of Radiology, Inonu University Faculty of Medicine, Malatya, Turkey
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Sami Akbulut, Department of Surgery and Liver Transplant Institute, Inonu University Faculty of Medicine, Malatya 44280, Turkey
Phone: +90 422 3410660
Figure 1. Patient Selection Flow Chart
Figure 2. Appearance of Some Stages of the Circumferential Fence Model Created With the Homologous Saphenous Vein Graft During Back-Table Reconstruction
Figure 3. Appearance of Some Stages of the Circumferential Fence Model Created With the Autologous Peritoneal Patch Graft During Back-Table Reconstruction
Figure 4. Intraoperative Appearance After Right-Lobe Liver Graft Implantation
Figure 5. Kaplan-Meier Survival Estimate for Stenosis-Free Survival
Figure 6. Calculation of Optimal Cutoff Value for Diameter of Circumferential Fence Using Receiver Operating Characteristic (ROC) Curve Analysis
Figure 7. Demonstration of the Relationship Between Presence of Hepatic Vein (HV) Anastomotic Stenosis and Anastomosis Diameter With Simple Box Plot Graph (P = .001)
Table 1. Comparison of Peritoneal Patch Graft Group and Saphenous Vein Graft Group in Terms of Demographic and Clinic Parameters
Table 2. Comparison of Demographic and Clinical Parameters of Patients With and Without Stenosis