EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Key words : Biliary anastomotic stricture, Hepatic artery pseudoaneurysm, Hepatic artery thrombosis

Living donor liver transplantation (LDLT) is commonly used in countries with limited availability of deceased donor organs, such as Japan.¹ In Japan, over the past 3 decades, LDLT for pediatric patient programs has started and expanded to include adults.^2-5 Grafts in adult LDLT mainly include right lobe (RL) and left lobe (LL) grafts. If no candidate satisfies the RL or LL graft selection criteria and anatomic eligibility, the right posterior segment (RPS) graft is occasionally selected.^6,7 The common situation favoring the RPS graft selection involves the following: (1) the RPS graft can be resected with independent vascular and biliary components that do not affect the residual liver, (2) the conventional full RL graft is expected to be too large, with a residual liver volume that is too small, and (3) the conventional LL graft is not large enough for the recipient, and the RPS graft is expected to be large enough for the recipient.

The RPS graft is a partial liver graft of Couinaud segments VI and VII with the right hepatic vein (RHV).⁶ The complicated vascular and biliary anatomy of the RPS graft is often associated with technical difficulties, and this procedure, which is predominantly performed at experienced high-volume centers, accounts for only 2.2% of the LDLT procedures performed in Japan.¹ Previous reports concerning the long-term outcomes of LDLT with the RPS graft are thus limited.^8-10

In this study, we examined the feasibility of using the RPS graft in adult LDLT through safety of donor operations and efficacy for recipient outcomes compared with other conventional grafts and anatomic points of the surgical technique, including long-term outcomes.

Materials and Methods

Between August 2000 and March 2017, 296 adult LDLTs procedures were performed at Kumamoto University Hospital. Two cases that had a left lateral segment graft and 4 cases in which donors were <20 years old at the time of donor surgery were excluded. The types of graft used during this period were the RL graft in 162 cases, LL graft in 119 cases, and RPS graft in 9 cases. The first LDLT procedure using an RPS graft was performed in 2006. This study was conducted in accordance with the 2013 Declaration of Helsinki and 2018 Declaration of Istanbul. Ethical approval was obtained from the institutional review board of Kumamoto University (approval No. 1826). Written informed consent was obtained from all donors.

Donor and graft selection

At our institution, LDLT donors were selected from spouses or relatives up to the third degree of kinship (eg, parents, offspring, siblings, grandparents, uncles, aunts). Age of LDLT donors was limited to <70 years old. If the RL graft was procured, donor age was limited to ?65 years old. Routine preoperative work-up and imaging studies, including dynamic computed tomography (drip infusion cholecystocholangiography), were performed. The preoperative estimated ratio of the liver remnant needed to be ?30%, and the estimated graft-to-recipient weight ratio (GRWR) needed to be ?0.7%, as measured by a 3-dimensional image analysis system (SYNAPSE VINCENT; Fujifilm Corp). The first choice of graft was the LL graft when the estimated GRWR was ?0.7%. The RL graft was used when the preoperative estimated GRWR in the LL graft was ?0.7% with the estimated GRWR of the RL graft ?0.7% and estimated ratio of the liver remnant after the RL graft procurement was ?30%. An RPS graft was considered when estimated ratio of the liver remnant after the RL graft procurement was ?30%.

The vascular anatomy and biliary anatomy were precisely reevaluated by multislice dynamic computed tomography for feasibility of resection of the posterior segment with remnants of vascular and biliary orifices that may be reconstructed in the recipient operation.

Surgical procedure

The basic donor hepatectomy procedure has been previously described.^11,12 For procurement of the RPS graft, after isolation of the right posterior hepatic artery (HA) and the right posterior portal vein (PV), the demarcation line was examined by a selective clamping test of the right posterior HA and PV. Transection of the hepatic parenchyma was performed based on the demarcation line and RHV under intraoperative ultrasonographic guidance. During transection of the hepatic parenchyma, the posterior hepatic duct was divided based on the image obtained from the intraoperative cholangiography. After the posterior HA, PV, and RHV were divided, the RPS graft was excised.

Statistical analyses

We presented quantitative data as medians and interquartile ranges and qualitative data as frequencies and rates. We analyzed categorical variables by using ?2 and Fisher tests. We used Bonferroni correction for multiple comparison analysis. We analyzed continuous variables by using the Kruskal-Wallis and Mann-Whitney U tests. We performed Cox proportional hazards analysis with variables having P < .40 identified by a univariate analysis. We set statistical significance at P < .05. We used IBM SPSS software program (version 25.) for all calculations.

With respect to donor safety, we compared selected variables among the 3 graft type groups. For pre-operative variables, we included age, sex, and body mass index (BMI). For operative and postoperative variables, we included operative time, blood loss, length of postoperative hospital stay, and laboratory data (total bilirubin, alanine aminotransferase [ALT], and international normalized ratio of prothrombin time [PT-INR]). Postoperative complications were categorized in accordance with the Clavien-Dindo classification.¹³ Complications, such as pleural effusion not requiring treatment and temporal hyperbilirubinemia without biliary complications, were excluded.

With respect to recipient outcomes, we compared the following variables among the 3 graft type groups. For preoperative variables, we included age, sex, BMI, Model for End-Stage Liver Disease score (MELD), the product of donor age and preoperative MELD,¹⁴ and ABO blood type compatibility. For operative and postoperative variables, we included operative time, blood loss, cold ischemic time, warm ischemic time, and GRWR.

Donor characteristics according to type of graft

The donor characteristics of the 3 groups are shown in Table 1. The RPS graft group had a higher proportion of men than the RL graft group (P = .02). Although the donor age among the 3 graft groups was significantly different as shown by Kruskal-Wallis test, post hoc analyses showed no significant difference. No significant differences were shown in BMI, operative time, blood loss, postoperative hospital stay, or incidence of complications, other than Clavien-Dindo grade I, among the 3 groups. Postoperative recovery of liver function among the 3 groups, as estimated by laboratory parameters (total bilirubin, ALT, and PT-INR), is shown in Figure 1. In the RPS graft group, the total bilirubin level was significantly higher than that in the LL group on postoperative days (POD) 1, 3, 6, and 8 (P = .003, P = .006, P = .03, and P = .009, respectively), the ALT level was significantly higher than that in the LL group on POD1 and POD3 (P = .01 and P = .02, respectively) and the RL group on POD1, POD3, POD6, and POD8 (P < .001, P < .001, P = .004, P = .01, and P = .04, respectively); the PT-INR level was significantly lower than that in the RL group on POD1 and POD3 (P = .04 and P = .005, respectively). No significant differences were shown among the 3 groups in terms of postoperative recovery of the liver function at approximately 2 weeks postoperatively.

Recipient characteristics according to type of graft

The recipient characteristics of the 3 groups are shown in Table 2. The RPS graft group had a higher proportion of women (P = .03), shorter cold ischemic time (P = .007), and lower GRWR (P = .04) than the RL graft group. No significant differences were shown in recipient age, MELD score, the product of donor age and preoperative MELD score, ABO compatibility, operative time, blood loss, 6-month graft mortality, or 5-year graft survival among the 3 groups. Furthermore, Cox proportional hazards analysis revealed that no factors were significantly associated with graft survival, including graft type (Table 3).

Details of living donor liver transplant using the right posterior segment graft

Details of LDLTs with RPS grafts are listed in Table 4. The branching patterns of the PV, bile duct, and HA in the studied 9 cases were categorized as follows. In PV type I, the right posterior PV and the right anterior PV showed branching from the right PV. In PV type II, the right posterior PV showed branching from the portal trunk separately (Figure 2). In bile duct type I, the right hepatic duct was formed by fusion of the right posterior and anterior segmental ducts. Bile duct type II showed triple confluence of the right anterior segmental duct, right posterior segmental duct, and left hepatic duct into the common hepatic duct. Bile duct type III showed drainage of the right posterior segmental duct into the left hepatic duct (Figure 3). The branching patterns of HA in all cases showed the typical anatomy in which the right posterior HA and the right anterior HA branched from the right HA.

The number of bile duct stumps in the RPS graft was more than 2 in 4 cases; the number of PV stumps was 2 in 2 cases. The type I PV branching pattern was prone to result in several stumps on the graft. In case 5, because the PV for segment 5 protruded over the ventral side of the RHV exposed to the initial transection line of the hepatic parenchyma, the transection line changed (Figure 4). The median actual GRWR was 0.82% (range, 0.52%-1.22%), and 3 cases showed a small graft size with level below 0.7% as our institutional standard. No severe postoperative complications (Clavien-Dindo grade ?IIIa) among RPS donors were observed.

In recipients, vessel reconstruction was performed with various innovations, especially in the matching of each vasculature. In case 4, the PV for segment 6 and the PV for segment 7 of the graft were anastomosed with the right and left PV of the recipient, respectively, and the biliary tract of segment 6 and the biliary tract of segment 7 of the graft were collectively anastomosed with the left hepatic duct of the recipient. In case 5, the PV for segment 6 and segment 7 of the graft were anastomosed with an interposition using the recipient’s left PV vascular graft excised from the native liver on the back table; subsequently, the left PV vascular graft was anastomosed with the left PV of the recipient in situ. Several stumps of the biliary tract of segment 6 and the biliary tract of segment 7 of the graft were anastomosed with the right hepatic duct, cystic duct, and left hepatic duct, respectively.

Hepatic artery thrombosis and rupture of HA pseudoaneurysms occurred in recipients as early complications. Severe acute cellular rejection and sepsis were the causes of death within the first year after LDLT at our institution. As previously reported, case 3 succumbed to sepsis caused by multiple liver abscesses due to rupture of the HA pseudoaneurysm.¹⁵ Although biliary anastomotic stricture occurred in 4 recipients as a late complication, all cases were treated with transpapillary endoscopic management. In case 6, endoscopic placement of a biliary stent for biliary anastomotic stricture was performed 1 year postoperatively; subsequently, the patient has been free of a biliary stent at ~5 years postoperatively because the biliary anastomotic stricture had improved. Whether allograft failure after liver transplant was connected to the history of biliary anastomotic stricture was unclear; however, allograft failure advanced gradually, and the patient eventually died ~8 years postoperatively. In case 7, the patient underwent retransplant from a deceased donor at ~4 years postoperatively because of allograft failure after liver transplant as a result of refractory chronic rejection. The other 5 patients had good long-term survival, although biliary anastomotic stricture or thrombotic microangiopathy occurred as postoperative complications.

In our institution, the RPS graft may be selected when the preoperative estimated GRWR in the LL graft is ?0.7%, the estimated ratio of the liver remnant after the RL graft procurement is ?30%, and no other suitable donor candidate is available. However, controversy remains regarding the feasibility of LDLT using the RPS graft because of the complicated vascular and biliary anatomy, technical difficulties, and limited experience with this graft.^8-10

In RPS donors, previous studies reported that the overall complication rate and severe complication rate (Clavien-Dindo grade ?IIIa) were 34.6% and 5.6%, respectively.⁷ Our study also showed an overall complication rate of 33.3%, although no severe complications occurred. Although Clavien-Dindo grade I complications were slightly more frequent in RPS donors than in RL and LL donors, no significant differences in overall morbidity rates were shown among the 3 groups. Although the sample number of RPS donors in this study was small, the outcome appears to be acceptable when the RPS graft is compared with other types of grafts. Furthermore, we found no significant differences among the 3 groups in postoperative recovery of the liver function from approximately 2 weeks postoperatively.

Among RPS recipients, previous work reported that the incidence of each complication was 18.8% for biliary anastomotic stricture, 8.7% for HA thrombosis, and 8.7% for bile leakage, and 5-year patient survival rates were shown to range from 50% to 80.8%.⁷ Our study showed similar findings: the incidence of each complication was 44.4% for biliary anastomotic stricture, 11.1% for HA thrombosis, 11.1% for HA pseudoaneurysm, and 11.1% for thrombotic microangiopathy, and the 5- and 10-year graft/patient survival rates were 66.7%/77.8% and 50.0%/62.2%, respectively (data not shown). Biliary and arterial complications are the most common and lethal complications because of technical difficulties associated with anastomosis of small vessels. Although a previous study reported that the endoscopic approach to bile duct stenosis was intractable because of the relatively sharp angle of the bile duct,⁸ all cases with biliary anastomotic stricture in our study were able to be treated by transpapillary endoscopic management.

In RPS graft procurement, a preoperative anatomic evaluation of the branching pattern and running course of the vessels is more important than in other types of grafts. If there are 3 or more PVs or HAs supplying the RPS or multiple small bile ducts draining the segment, procurement of the graft will not be possible because the reconstruction in the recipient will be too difficult. Fortunately, we had such an experience during the study period. The RPS graft may be an unavoidable last choice when conventional grafts cannot be used. However, to assess the feasibility of RPS grafts, it is necessary to assess the superiority and inferiority of other types based on them. In our surgical procedure for procurement of the RPS graft, we performed transection of the hepatic parenchyma based on the demarcation line and the RHV for donor safety and minimization of the ischemic area of the RPS graft. In contrast, a previous report suggested that an extended RPS graft that adds the RHV drainage area in the right anterior sector to the RPS graft is recommended for easier parenchymal transection.⁸ The meaning of the ischemic area of the RPS graft has yet to be determined.

In a situation where deceased organ donations are limited in number, as in Japan, an RPS graft is one alternative way to proceed with LDLT. However, the procedure is complicated even after meticulous simulation before surgery, and the incidence of surgical complications, such as biliary stenosis, can occur. Another alternative method is the use of a dual graft.¹⁶ However, this is naturally controversial because the method requires 2 living donors, so donor risk is doubled.

In recipient operations using the RPS graft, several technical points must be considered in the reconstruction of small vessels. In the reconstruction of the PV, for example, the anastomosis orifice faces ventrally more than other types of grafts. Therefore, other measures are needed, such as sufficiently dissecting the proximal portal trunk and using a patch graft at the anterior wall to prevent kinking at the anastomosis site. Furthermore, the diameters of the segmental branches of the PV may be small (sometimes 2 stumps) in the RPS graft. The unification of the 2 stumps with an interposition vascular graft on the back table may be useful in some cases, such as in case 5.

Our study had several limitations. First, our study was retrospective and conducted at a single institution. Second, the sample size was too small to provide an accurate assessment of LDLT using the RPS graft. Further experience and continuous improvement are necessary.

The RPS graft could be an effective alternative in terms of donor safety and recipient long-term outcomes when no eligible candidates are available among families and relatives. However, RPS graft procurement and recipient operation have several technical points, particularly in the reconstruction of vessels. Therefore, careful preoperative anatomic evaluation and simulation are important.

Volume: 22 Issue: 10 October 2024

FULL TEXT