EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Liver transplant (LT) is an effective treatment for end-stage liver disease, but it is a complex procedure associated with significant vascular and biliary complications. Hepatic artery thrombosis (HAT) has been reported in 1.4% to 8% of adult living donor LT (LDLT), with outcomes continuing to improve.¹ Arterial anastomosis is a crucial part of the transplant procedure. Many suggestions and techniques continue to emerge that are proposed to improve outcomes. However, only a few of these techniques persist in common use, and the specific choice of technique is most often based on the surgeon’s preference. Another major factor is the use of surgical loupes, which are known to shorten the duration of the procedure while maintaining and improving outcomes.^2,3 In LDLT, the graft hepatic artery is often short in length and may have a small caliber. The size discrepancy between the right hepatic artery (RHA) of the graft and the recipient’s dilated RHA presents a technical challenge to the surgeon who performs the anastomosis. Techniques to overcome the mismatch issue, such as the fish-mouth technique, end-to-side suturing, and oblique division of the graft artery, have been previously described.⁴ Millis and colleagues suggested an interposition graft to mitigate the size difference between the graft and recipient arteries.⁵ Although these techniques have been effective, some details, such as the longer duration of the procedure, the longer anastomotic lines, and the additional manipulation required for the procedure, should be considered and compared with an equally effective alternative such as the left hepatic artery (LHA) for the anastomosis, which has not been well studied. All these techniques have a common goal, ie, matching the conduit size to best facilitate successful perfusion.

We routinely anastomose the graft RHA to the recipient LHA, as the size match is better, more convenient, and provides a better alignment of the arteries after the anastomosis. This prospective observational study was conducted to assess whether anastomosis with the recipient LHA had a comparable outcome to anastomosis with the recipient RHA and whether dissection of the RHA had any effect on the biliary complications in these patients.

Materials and Methods

Our study focused on patients who had LDLT performed at the Centre for Liver and Biliary Sciences, Max Super Specialty Hospital, New Delhi, from December 2013 to February 2015. The patients were followed up through May 2019. A total of 200 adult patients (all were recipients of adult-to-adult LDLTs) consented to the use of their data for this study and were prospectively observed on the basis of the recipient artery that was chosen for arterial anastomosis. There were 100 consecutive patients who underwent anastomosis to the recipient LHA (the LHA group) and 100 consecutive patients who underwent anastomosis to the recipient RHA (the RHA group). All the transplants were LDLT from donors who were either a first-degree or second-degree relative and were approved by the National Organ and Tissue Transplant Organization. This study was approved by our institutional ethics committee. Written informed consent was obtained before the surgery from both the donors and the recipients.

Inclusion criteria were (1) right lobe adult-to-adult ABO-compatible LDLT from related donor; (2) patients scheduled for duct-to-duct anastomosis; and (3) single artery in donor graft for anastomosis on preoperative imaging. Exclusion criteria were (1) retransplant; (2) patients with sclerosing cholangitis or widely separated ducts that required 2 biliary anastomoses or bilioenteric anastomosis as suggested by preoperative imaging; (3) thrombosed/dissected artery; (4) a history of transarterial radioembolization or transarterial chemoembolization; (5) multiple arteries; and (6) pediatric LT.

Liver graft implants were executed by a single surgeon (SG) with experience in more than 1500 related LDLT procedures. Arterial anastomosis was performed with the aid of loupes (with a magnification of ×4.5), and Prolene 8-0 sutures were placed in an interrupted fashion. Biliary anastomosis was performed in an end-to-end fashion with polydioxanone 6-0 sutures in an interrupted fashion.

Routine Doppler ultrasonography was performed at the end of the procedure and confirmed the adequacy of flow in all the grafts, and this was repeated in the postoperative period for up to 7 days. Normal hepatic artery flow on Doppler image is characterized by a hepatopetal pattern, with a systolic velocity of >30 cm/s, acceleration time (end diastole to first systolic peak) <100 milliseconds, and resistive index between 0.5 and 0.7.6 Patients for whom the Doppler imaging showed evidence of decreased arterial flow (resistive index <0.5, acceleration time >100 milliseconds, or blunted waveform) were evaluated further by computed tomographic angiography to document HAT.

Postoperative outcomes related to arterial and biliary complications in the 2 groups were analyzed with appropriate statistical tests. Primary endpoints were development of HAT and death in the postoperative period (within 90 days). Secondary endpoints were development of bile leak, development of biliary stricture or cholangitis requiring intervention, biliary sepsis, and death due to biliary complications/aneurysmal bleed.

Statistical analyses
Statistical analyses were performed with GraphPad Prism 7. Comparison of means was performed by 2-tailed t test. The Fisher exact test was used to evaluate categorical data. P < .05 was considered significant.

All patients enrolled in the study had undergone a right lobe LDLT. Average age of the recipients was 48 years in both the LHA and RHA groups. The patients who required a transplant for end-stage liver disease were most often male. The mean Child-Turcotte-Pugh (CTP) score was above 10 in both groups with no significant difference, and more than 60% of patients were CTP category C. Among the total study group, the most common cause of end-stage liver disease was hepatitis C (~50%). The average donor age was 28.9 years in the LHA group and 30.9 years in the RHA group. Most of the donors (60%) were female (Table 1).

A modified right lobe (MRL) graft was retrieved in ~80% patients. The duct size, cold ischemia time (CIT), warm ischemia time (WIT), and graft-to-recipient weight ratio (GRWR) were comparable in both groups. Ductoplasty was required in 19 patients in the LHA group and 16 patients in the RHA group. The arterial diameters were calculated from preoperative computed tomography results. When we compared the arterial lumen of the recipient and donor arteries, size mismatch was significantly more in the RHA group (15% vs 72%; P = .03) (Table 2).

Comparison of outcomes showed that the early morbidity and mortality was similar in both groups. Overall, there was 10.5% mortality in the early postoperative period. Major causes included poor graft function, sepsis with multiorgan failure, and secondary hemorrhage. Three patients developed HAT. For 1 patient in the RHA group, the graft was saved by immediate surgery to refashion the anastomosis. The other 2 patients who developed HAT died soon after the event. Bile leak with sepsis was the cause of death in 7 patients.

Late strictures among survivors were more common in the RHA group compared with the LHA group (29.7% vs 22.7%) during follow-up, although this difference was not statistically significant. Most (~80%) of the strictures were managed effectively by endoscopic stenting. Two patients in each group required Roux-en-Y hepaticojejunostomy (Table 3). To mitigate the common confounding factors related to the quality of the graft, subgroup analysis was performed among the patients who received MRL grafts with a single duct and GRWR ?0.8. The CIT and WIT were comparable in both groups, and there was no significant difference in the duct size. Bile leak, mortality, and incidence of late-onset biliary strictures were less common in the LHA group, although the difference was not statistically significant (Table 4).

This is one of the few prospective studies to evaluate the outcomes of the LHA and RHA of LDLT recipients for arterial anastomosis. Arterial reconstruction in LDLT is more complicated and challenging than in the deceased donor LT, and size mismatch is a key factor to be considered before reconstruction.⁷ The hypothesis of our study was that the recipient LHA was a better size match than the RHA for arterial anastomosis in right lobe liver grafts to achieve a satisfactory end-to-end anastomosis because (1) the orientation of the LHA in relation to the RHA of the graft allows a linear alignment of the arteries and (2) the size of the LHA in patients with liver disease usually matched that of the RHA of the liver graft. All of these aspects of the LHA procedure serve to reduce the risk of arterial complications, reduce turbulence across the anastomosis, and thereby improve the graft perfusion. Arterial anastomosis has been extensively studied, as it is a key factor for a successful transplant outcome. Researchers worldwide have suggested several techniques to match the size of the vessels in LT, especially in children where the left lobe and segmental⁴ grafts are used.⁸ Haberal and colleagues have demonstrated that smaller arteries or multiple arteries may not increase the complication rates, and they have suggested various techniques to enable a better match of the arterial lumen, especially the smaller vessels.⁹ There are few reports on the use of the splenic artery¹⁰ in children where the graft artery is larger than the recipient hepatic artery. However, instead of a graft or other techniques to match the size of the arteries, the option to consider the LHA as the first choice may avoid unnecessary manipulation of the arteries during an interposition graft procedure or any other size-matching techniques. The LHA procedure is also a worthy option if other techniques fail. Although there are other factors associated with the development of vascular complications (eg, a female recipient or metabolic liver disease as the cause of liver failure), we did not find any such association in our series.¹¹

Olthoff and colleagues suggested that technical expertise is an important factor to reduce arterial thrombosis, graft failure, and death after LDLT, which is the case for our study because all the anastomoses were performed by the same surgeon. The rate of HAT was significantly higher in centers that perform fewer than 4 adult LDLTs, which supports the importance of technical expertise to perform a successful anastomosis.¹² The reported incidence rate of early HAT varies from 1.4% to 8% after LT.^13-18 In LDLT, the incidence of HAT ranges from 3% to 5%.^19-21 The incidence of HAT in our study was 1.5% for both groups combined, which is comparable to most other centers, and there is a high volume of LDLTs (more than 250 per year) performed at our center. Early HAT was associated with a rise in liver enzymes, bile leak, sepsis, and death in more than 50% of our patients. Of 3 patients diagnosed with HAT, only 1 patient’s graft could be saved by revision of the anastomosis. We observed no significant difference between LHA and RHA groups with regard to HAT.

The biliary anastomosis remains the Achilles’ heel of LDLT.²² Graft bile ducts are exclusively dependent on the arterial perfusion. Although there are several factors for biliary complications, arterial perfusion is a determinant factor that is often overlooked, especially with regard to the orientation of the vessels and complications due to decreased blood flow to the biliary complex. Few studies have focused on both immediate and long-term biliary complications. Overall incidence varies from 15% to 60%.23,24 Postoperative bile leak and biloma are known causes of morbidity and mortality. In our study, the biliary anastomoses were all performed by the same surgeon; thus the anastomosis technique was consistent for all patients in both study groups. Overall, 26 patients (13%) among both groups developed bile leak; for 7 of those 26 patients (27%), this progressed to sepsis and early postoperative death, whereas the remaining patients were treated effectively with radiological intervention, intensive care, and/or biliary stenting based on the severity of symptoms and the time of diagnosis.

Shah and colleagues from Toronto and Morioka and colleagues from Kyoto University have documented that the improved techniques and expertise have reduced the incidence rate over time of bile leaks as a complication of LT; however, the stricture rate in follow-up has remained at 20% to 25%.25,26 These details indicate that a meticulous biliary anastomosis may not be independently effective to reduce biliary complications. Other than surgical expertise, factors that affect the healing of the bile ducts include donor age, preoperative Model for End-Stage Liver Disease score, CIT, duct diameter, variations in ductal anatomy, cellular rejection, hilar dissection, and, most importantly, arterial thrombosis. All of these factors affect the ductal perfusion and result in ischemia. Hilar dissection and dissection of the artery is often performed closely adjacent to the bile duct in an effort to obtain a sufficient length for vessel anastomosis, and this can contribute to duct ischemia and subsequent anastomotic and nonanastomotic lengthy strictures, not susceptible to endoscopic management. In our study, biliary complications (biloma, bile leak, biliary sepsis/death) were comparable in the 2 study groups during the early period. However, strictures during follow-up were more common in the RHA group, although the difference was not statistically significant. Immediate postoperative outcomes are implicated by the anastomotic technique, graft quality, technical expertise, and vascular supply. Late strictures were identified in 20 in the LHA group and 27 in the RHA group. This is high compared with results from other series, and the cause may be the interrupted technique of anastomosis, whereas most other series were performed with a combination of duct-to-duct and bilioenteric anastomosis. The follow-up period in most of the studies is around 2 years, whereas the mean follow-up period in our study was more than 60 months (Table 5). The average CTP score before transplant was more than 10 in more than 60% of the recipients. At our center, duct-to-duct anastomosis is preferred over Roux-en-Y hepaticojejunostomy even in the presence of multiple ducts. The duct-to-duct anastomosis technique allows easy access in case of biliary obstruction or bile leak. Strictures in the follow-up period could be effectively managed by endoscopic balloon dilatation or repeated biliary stenting in 90% of patients. This may indicate that most strictures were short in length and amenable to successful endoscopic management. Nonanastomotic strictures are usually long and often require surgical intervention. Stent exchange in biliary strictures requires multiple procedures at 3 to 4 monthly intervals over a year. Three patients in each group did not respond to endoscopic treatment and had undergone a percutaneous transhepatic biliary drainage or surgical revision of anastomosis (Table 3).

The overall mortality rate in our series was 10.5%. Cause of death did not differ between the 2 groups. Toshima and colleagues documented an incidence rate of 10.1% early postoperative deaths after related LDLT in adults.³¹ Similarly, Stey and colleagues (in 2013) and Xiao and colleagues (in 2009) reported a 17% incidence rate for 3-month mortality and 24% for 1-month mortality.^32,33 Important causes of death in our study included sepsis with multiorgan failure and early graft dysfunction. This may be the result of the large number of patients among both groups (two-thirds) with CTP class C liver failure, which is associated with higher risk of postoperative complications. Sepsis (from arterial thrombosis or bile leak) was the cause of 7 of 21 deaths (33%) in our present series. There was no difference between the 2 groups in this regard, which emphasizes that the use of the recipient LHA for anastomosis is equal to the safety of the RHA .anastomosis

To eliminate the effect of important confounding factors related to the graft quality, we performed a subgroup analysis of outcomes in patients GRWR ?0.8 with a single duct and did not require ductoplasty in the MRL graft (Table 4). No vascular complication in the early postoperative period was documented in either of the subgroups. Early biliary complications, as well as late strictures, were more common in the RHA group compared with the LHA group, although the difference was not statistically significant. This lower incidence rate of biliary complications in the LHA group could be the result of better perfusion of the biliary system in the LHA group, possibly from the better alignment and reduced turbulence to the blood flow. The better size match of the LHA with the graft RHA could also be explained by the lack of size-matching techniques required. The LHA anastomosis has an extra advantage to preserve the RHA as secondary option if LHA proves overly difficult or as a back-up or for a possible future early revision of the arterial anastomosis. The use of the LHA should also help to avoid the dissection along the common bile duct to expose the RHA and thereby preserve the arterial supply to the recipient common bile duct from the lateral and ascending branches, which could help to reduce the duct ischemia and subsequent biliary complications (long anastomotic and nonanastomotic strictures) in the long-term.

We found that anastomosis of the graft artery in a right lobe liver graft with the recipient LHA was feasible in most right lobe LDLTs and provided a better size match compared with the recipient RHA. The use of the recipient LHA could reduce the biliary complications in long-term outcomes compared with the use of the recipient RHA.

Volume: 20 Issue: 2 February 2022

FULL TEXT