SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

Aortohepatic conduits provide a vital alternative for graft arterialization during liver transplantation. Conflicting results exist with respect to the rates of comorbidities, and long-term survival data on primary grafts are lacking. To identify the complications associated with aortohepatic conduits in primary liver transplantation and their impact on survival, we conducted a single-center, retrospective cohort analysis of all consecutive adult (n = 1379) and pediatric primary liver transplants (n = 188) from 1998 to 2009. The outcomes of aortohepatic conduits were compared to those of standard arterial revascularization. Adults with a conduit (n = 267) demonstrated, in comparison with adults with standard arterialization (n = 1112), an increased incidence of late (>1 month after transplantation) hepatic artery thrombosis (HAT; 4.1% versus 0.7%, P < 0.001) and ischemic cholangiopathy (7.5% versus 2.7%, P < 0.001) and a lower 5-year graft survival rate (61% versus 70%, P = 0.01). The adjusted hazard ratio (HR) for graft loss in the conduit group was 1.38 [95% confidence interval (CI) = 1.03-1.85, P = 0.03]. Notably, the use of conduits (HR = 4.91, 95% CI = 1.92-12.58) and a warm ischemia time > 60 minutes (HR = 11.12, 95% CI = 3.06-40.45) were independent risk factors for late HAT. Among children, the complication profiles were similar for the conduit group (n = 81) and the standard group (n = 107). In the pediatric cohort, although the 5-year graft survival rate for the conduit group (69%) was significantly impaired in comparison with the rate for the standard group (81%, P = 0.03), the use of aortohepatic conduits did not emerge as an independent predictor of diminished graft survival via a multivariate analysis. In conclusion, in adult primary liver transplantation, the placement of an aortohepatic conduit should be strictly limited because of the greater complication rates (notably late HAT) and impaired graft survival; for children, its judicious use may be acceptable. Liver Transpl 19:916–925, 2013. © 2013 AASLD.

Abbreviations
CI

confidence interval

HAT

hepatic artery thrombosis

HR

hazard ratio

MELD

Model for End-Stage Liver Disease

NS

not significant

PELD

Pediatric End-Stage Liver Disease

POM

postoperative month

Aortohepatic conduits are a vital alternative in liver transplantation. They are more commonly used in demanding cases when a standard anastomosis to the native hepatic artery fails to provide sufficient inflow because of conditions such as severe atherosclerosis and intimal dissection. Many centers have reported the use of conduits in primary grafts and regrafts with conflicting results [eg, a risk factor for hepatic artery thrombosis (HAT) and graft loss[1-8] or a promising alternative, particularly for recipients at high risk for arterial thrombosis[9-13]. Furthermore, there is only 1 publication describing long-term results for adult primary liver transplantation, and aortohepatic conduits had no negative impact on complications or survival.[14] Remote outcomes for children are currently unavailable. In an era of increasingly older patients on liver transplant waiting lists and split liver transplantation advocacy, aortohepatic conduits could potentially serve as an important source of arterial inflow.[9, 13-16]

To broaden our perspective on the pros and cons of aortohepatic conduits and to define the sequelae of their usage and patient selection criteria, this study focused exclusively on primary liver transplantation. For the first time, the impact on postoperative comorbidities and long-term survival was analyzed separately for adults and children.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

We retrospectively analyzed all adults (≥18 years) and children (<18 years) undergoing primary liver transplantation from 1998 to 2009 at the University of Miami/Jackson Memorial Hospital. This study conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was conducted with approval from the institutional review board of the University of Miami. The data that were collected and analyzed included recipient and donor demographics, intraoperative data, the type of liver allograft, the arterial anatomy of the recipient and donor, the indication for the placement of an aortohepatic conduit and the type of arterial graft, posttransplant complications and their treatment, retransplantation information, causes of graft loss and patient death, and survival.

Orthotopic liver transplantation was performed with standard techniques, with the recipient's inferior vena cava preserved in most cases (91.0% for adults and 95.2% for children).[17, 18] For arterial anastomoses, extra attention was always paid to their alignment to keep the vessels straight and prevent a redundant and tortuous course. Typically, arterial revascularization of a liver allograft involved an end-to-end anastomosis of the donor's celiac trunk to the recipient's common hepatic artery. If this was not readily achievable because of the vessel quality, a size discrepancy, or other reasons, the recipient's splenic artery or celiac trunk was used as an alternative inflow, and these cases formed the standard group. When all these attempts failed to establish sufficient arterial inflow, aortohepatic conduits were placed between the infrarenal aorta and hepatic hilum through the antepancreatic route in most cases (ie, the conduit group).[19] If access to the infrarenal aorta was considered dangerous because of severe atherosclerosis, extensive adhesions, or varices, the common iliac arteries or the supraceliac aorta were used as alternative sources of arterial inflow.[20] Patient background information, postoperative complication profiles, and patient and graft survival rates were compared between the 2 groups.

The patency of an arterial anastomosis was assessed serially with intraoperative and postoperative Doppler ultrasonography. Since September 1999, protocol Doppler ultrasonography has been performed for all pediatric recipients every 12 hours in the first week[21] and on a daily basis for the first 3 days in adults. After this time frame and beyond the initial hospital stay, Doppler ultrasonography was indicated whenever patients showed an elevation of liver enzyme levels or signs of hepatic dysfunction. If there was clinical suspicion of compromised arterial flow, computed tomography, formal angiography, or surgical exploration was performed according to each patient's condition. A diagnosis of HAT was determined by the absence of hepatic arterial enhancement on an angiogram or by the identification of a complete occlusion of the artery during surgical exploration. The onset of HAT within 1 month after transplantation was defined as early, and the onset of HAT after that was defined as late. Hepatic artery stenosis (≥50% vessel narrowing) was identified with computed tomography or angiography. Biliary complications included anastomotic leaks or strictures, ischemic cholangiopathy (nonanastomotic biliary strictures and abscesses), and other complications (sludges, stones, and kinking). Postoperative hemorrhaging necessitating surgical re-exploration was labeled as bleeding. Patients requiring temporary renal replacement therapy and those on mechanical ventilation for more than 1 week after surgery were designated with hemodialysis and prolonged ventilation, respectively.

Acetylsalicylic acid has been prophylactically prescribed to patients (both adults and children) with an aortohepatic conduit at our center since the late 1990s. It is started at least 24 hours after transplantation in patients with a platelet count > 50,000/μL with no signs of surgical bleeding. In high-risk individuals (eg, patients with hypercoagulable disorders or a previous history of venous thromboembolism), clopidogrel bisulfate is additionally administered.

The mainstay of maintenance immunosuppression was tacrolimus (Prograf, Astellas Pharma US, Inc.) as part of a dual- or triple-agent regimen with prednisone and mycophenolate mofetil (CellCept, Genentech USA, Inc.).

Statistical Analysis

Patients were censored if they were alive at the time of last follow-up. A graft was considered to be lost when a patient died or underwent retransplantation. Survival rates were calculated as the interval between transplantation and patient death or graft loss. Unadjusted survival rates were estimated with the Kaplan-Meier method, and the log-rank test was applied to compare the survival distributions.

To assess the independent impact of the use of aortohepatic conduits on patient and graft survival, multivariate Cox regression analyses using the forced entry method were performed. The recipient variables were as follows: age, sex, race, body weight (children only), malignancy in the explanted liver (adults only), hepatitis C positivity (adults only), use of aortohepatic conduits, existence of portal vein thrombosis at the time of transplantation (adults only), estimated blood loss, number of units of packed red blood cells and fresh frozen plasma transfused during transplantation, time spent in the operating room, posttransplant hospital stay, and era (the first era was from 1998 to 2003, and the second era was from 2004 to 2009). Moreover, the following donor variables were included: age, sex, type of liver graft, cold and warm ischemia times, and presence of aberrant hepatic arteries requiring multiple anastomoses. To check for consistency, Cox proportional hazard models were also generated with the stepwise backward elimination method, and the log-likelihood ratio was applied to determine the goodness of fit.

To elucidate the factors contributing to late HAT, a multivariate Cox regression analysis involving methods of forced entry and stepwise backward elimination using the log-likelihood ratio was performed, and it included the following variables: recipient age, sex, and race; use of aortohepatic conduits; pretransplant transarterial chemoembolization; hepatitis C positivity; donor age and sex; cold and warm ischemia times; multiple hepatic arteries requiring complex reconstruction; and donor/recipient cytomegalovirus infection status.

The Mann-Whitney test or the Student t test was used to compare continuous variables, and the χ2 test or Fisher's exact probability test was used to compare categorical variables. Data are shown as medians and interquartile ranges, means and standard deviations, or numbers and percentages as appropriate. Statistical significance was defined as P < 0.05. All statistical analyses were performed with IBM SPSS Statistics 20 (SPSS, Inc.).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

Patient Demographics

From January 1998 to December 2009, 1379 adult primary liver transplants and 188 pediatric primary liver transplants were performed at our institution. The patients were divided into 2 groups: a conduit group (n = 267 for adults and n = 81 for children) and a standard group (n = 1112 for adults and n = 107 for children). Patients' background data are outlined in Table 1. The median estimated blood loss and transfusion requirements were greater and duct-to-duct biliary reconstruction was less frequently employed for adults in the conduit group versus adults in the standard group. Meanwhile, children in the conduit group were younger and smaller, were more likely to receive a split liver (either deceased or living donor) or a reduced size liver, and had longer cold ischemia times, and duct-to-duct biliary reconstruction was less common for them versus children in the standard group. They were also sicker: they more often had either a Model for End-Stage Liver Disease (MELD) score greater than 25 or a Pediatric End-Stage Liver Disease (PELD) score greater than 15 or met status 1A or 1B criteria under Organ Procurement and Transplantation Network policy (69% for the conduit group versus 48% for the standard group, P = 0.03). Overall, the time spent in the operating room and the length of the hospital stay after transplantation were longer for patients (both adults and children) in the conduit group versus those in the standard group. Liver allografts from donation after cardiac death were transplanted into 2.3% of adults (32/1379; 3.0% for the conduit group versus 2.2% for the standard group, P = 0.41), but they were never used in children. Other variables were comparable between the 2 groups (Table 1).

Table 1. Patient Characteristics
FactorAdult PatientsPediatric Patients
Conduit (n = 267)Standard (n = 1112)P ValueConduit (n = 81)Standard (n = 107)P Value
  1. NOTE: The bolded values are significant.

  2. a

    The data are presented as medians and interquartile ranges.

  3. b

    Multiple hepatic arteries requiring complex reconstruction.

Age (years)a54 (47–61)53 (47–60)0.521.0 (0.5–4.1)4.9 (0.9–12.3)<0.001
Female sex [n (%)]102 (38.2)368 (33.1)0.1143 (53.1)64 (59.8)0.36
Body weight (kg)a76 (66–90)79 (68–90)0.338.1 (5.7–13.9)18.9 (8.6–45.1)<0.001
MELD scorea22 (16–25)21 (16–24)0.4022 (—)17 (9–18)0.19
PELD scorea19 (10–25)16 (1–23)0.16
Pretransplant transarterial chemoembolization [n (%)]7 (2.6)32 (2.9)0.820 (0)0 (0)
Donor age (years)a44 (27–57)42 (26–55)0.089 (2–17)9 (2–17)0.69
Graft [n (%)]  0.24  <0.001
Whole258 (96.6)1071 (96.3) 52 (64.2)87 (81.3) 
Reduced size0 (0)0 (0) 15 (18.5)2 (1.9) 
Split7 (2.6)19 (1.7) 13 (16.0)10 (9.3) 
Living donor2 (0.7)22 (2.0) 1 (1.2)8 (7.0) 
Operation      
Cold ischemia time (minutes)a417 (354–497)419 (351–493)0.91447 (374–515)395 (346–452)0.01
Estimated blood loss (L)a7.1 (3.3–12.0)5.0 (2.5–8.5)<0.0010.8 (0.4–1.5)0.7 (0.3–1.5)0.31
Fresh frozen plasma (U)a20 (11–34)16 (8–27)<0.0012.0 (0.6–4.0)2.0 (0.0–5.0)0.52
Packed red blood cells (U)a13 (8–22)10 (6–15)<0.0012.5 (1.5–4.0)2.0 (1.0–4.0)0.54
Multiple arteries [n (%)]b33 (12.4)128 (11.5)0.703 (3.7)4 (3.7)>0.99
Duct to duct [n (%)]69 (25.8)510 (45.9)<0.0018 (9.9)24 (22.4)0.04
Time in operating room (hours)a13.5 (11.6–15.1)12.1 (10.5–13.6)<0.00111.6 (10.1–13.1)10.0 (8.7–11.8)<0.001
Length of stay (days)a13 (9–21)11 (8–17)<0.00115 (11–23)11 (8–19)0.047

The indications and the types of aortohepatic conduits are listed in Table 2. The majority of the cases required an aortohepatic conduit for poor arterial perfusion, which was determined by the loss or severe attenuation of pulsatile blood flow under palpation of the recipient hepatic arteries or by a dampened flow and tardus parvus waveform of intrahepatic arteries after reconstruction via intraoperative Doppler ultrasonography. Other reasons for using aortohepatic conduits in adults included intimal dissection in the recipient or donor hepatic arteries, a complex anatomy of the arterial blood supply to the liver precluding standard hepatic arterial reconstruction, and difficulty in exposing recipient arteries because of severe adhesions or extensive varices in the hepatic hilum or the suprapancreatic area. In children, anatomical incongruity was the second most common indication. Most of the conduits were anastomosed to the infrarenal abdominal aorta with an iliac arterial graft from the donor via the antepancreatic route (94.4% for adults and 95% for children). The donor aorta was used as a conduit in several cases as previously described.[22] In adults, the recipient saphenous vein, the preserved pediatric donor aorta, and various combinations of donor iliac, carotid, and subclavian arteries were also used as conduits. Direct anastomosis of the donor celiac trunk to the recipient infrarenal aorta was performed in 2 children. Other sources of arterial inflow included the supraceliac aorta or the right common iliac artery (the latter was applied only to adult patients) whenever the infrarenal aorta could not be safely accessed because of atherosclerosis or adhesions.

Table 2. Indications and Types of Aortohepatic Conduits
FactorAdult Patients (n = 267)Pediatric Patients (n = 81)
Indication [n (%)]  
Poor arterial perfusion144 (53.9)62 (76.5)
Intimal dissection59 (22.1)1 (1.2)
Anatomy36 (13.5)13 (16.0)
Adhesions/varices24 (9.0)3 (3.7)
Other4 (1.5)2 (2.5)
Recipient arterial inflow [n (%)]  
Infrarenal abdominal aorta252 (94.4)77 (95.1)
Supraceliac abdominal aorta10 (3.7)4 (4.9)
Right common iliac artery5 (1.9)0 (0)
Conduit [n (%)]  
Iliac artery244 (91.4)63 (77.8)
Carotid artery13 (4.9)13 (16.1)
Thoracic aorta1 (0.4)3 (3.7)
Other9 (3.4)2 (2.5)

Postoperative Complications (Table 3)

Adults in the conduit group had a significantly increased incidence of HAT (6.4%) in comparison with adults in the standard group (2.4%, P = 0.001). When the 2 groups were divided according to early-onset HAT (≤1 month after transplantation) and late-onset HAT (>1 month), the adults in the conduit group were found to be at higher risk for late onset than the adults in the standard group (4.1% versus 0.7%, P < 0.001), but the rates of early thrombosis were identical (2.2% versus 1.7%, P = 0.55). The type of conduit used and the recipient arterial inflow did not affect the occurrence of HAT (P = 0.61 and P = 0.38, respectively). Preoperative transarterial chemoembolization (n = 39) was associated with an increased risk of arterial complications approaching 13% (5/39), whereas the rate was 3.7% for patients with no chemoembolization (50/1340, P = 0.004). The presence of multiple hepatic arteries requiring complex reconstruction did not correlate with arterial complications (3.1% for multiple arteries versus 4.1% for a single artery, P = 0.54). In the standard group, the use of the recipient splenic artery (n = 15) or celiac trunk (n = 21) did not increase the risk of arterial complications in comparison with the use of the proper or common hepatic arteries (n = 1041; 0% versus 3.3%, P = 0.63). Ischemic cholangiopathy was more frequently encountered in the conduit group versus the standard group (7.5% versus 2.7%, P < 0.001), but the incidences of other types of biliary complications were comparable. Notably, late HAT strongly correlated with ischemic cholangiopathy (P < 0.001). Furthermore, adults with an aortohepatic conduit were more likely to develop small bowel obstructions, need hemodialysis, and require retransplantation in comparison with adults in the standard group (all P < 0.05).

Table 3. Postoperative Complications
FactorAdult PatientsPediatric Patients
Conduit (n = 267)Standard (n = 1112)P ValueConduit (n = 81)Standard (n = 107)P Value
  1. NOTE: The bolded values are significant.

Arterial complications, total [n (%)]20 (7.5)35 (3.1)0.0015 (6.2)12 (11.2)0.23
HAT17 (6.4)27 (2.4)0.0015 (6.2)9 (8.4)0.56
Early (≤POM 1)6 (2.2)19 (1.7)0.554 (4.9)8 (7.5)0.56
Late (>POM 1)11 (4.1)8 (0.7)<0.0011 (1.2)1 (0.9)>0.99
Hepatic artery stenosis3 (1.1)8 (0.7)0.510 (0)3 (2.8)0.26
Biliary complications, total [n (%)]47 (17.6)132 (11.9)0.0112 (14.8)14 (13.1)0.73
Anastomotic leak/stricture27 (10.1)94 (8.5)0.3910 (12.3)9 (8.4)0.38
Ischemic cholangiopathy20 (7.5)30 (2.7)<0.0011 (1.2)5 (4.7)0.24
Other0 (0.0)8 (7.2)0.371 (1.2)0 (0)0.43
Small bowel obstruction [n (%)]12 (4.5)24 (2.2)0.035 (6.2)1 (0.9)0.09
Bleeding [n (%)]21 (7.9)62 (5.6)0.163 (3.7)6 (5.6)0.73
Hemodialysis [n (%)]64 (24.0)200 (18.0)0.033 (3.7)6 (5.6)0.73
Prolonged ventilation [n (%)]47 (17.6)192 (17.3)0.9016 (19.8)8 (7.0)0.01
Retransplant [n (%)]41 (15.4)91 (8.2)<0.00112 (14.8)10 (9.3)0.25

The onset of late HAT (n = 19) ranged from 36 days to 4.6 years (median = 101 days) after transplantation. A multivariate Cox regression analysis revealed that the use of aortohepatic conduits [hazard ratio (HR) = 4.91, 95% confidence interval (CI) = 1.92-12.58, P = 0.001] and a warm ischemia time > 60 minutes (HR = 11.12, 95% CI = 3.06-40.45, P < 0.001) were independent factors associated with late HAT. Salvage transplantation was performed for 17 patients (89%). One patient succumbed to sudden cardiac arrest during retransplantation, and a histopathological examination of the explanted liver graft was available for the remaining 16 cases. The most common finding was ischemic cholangiopathy (n = 13), which was followed by central perivenulitis (n = 7). Six of the 16 patients (38%) had histologically confirmed chronic rejection, and 2 (12%) demonstrated atherosclerosis of the thrombosed arteries. All 4 patients with a hepatitis C virus infection showed recurrent disease in the liver graft.

Among children, the complication profiles were similar for the 2 groups. The only exception was the higher incidence of prolonged ventilation for children in the conduit group versus children in the standard group (20% versus 8%, P = 0.01).

In the entire cohort, arterial complications (n = 72) were ultimately treated with retransplantation for 42 patients (58%), surgical revision of the arterial anastomosis for 13 (18%), and angioplasty for 12 (17%). Among the remaining 5 patients (7%), 1 had stenosis that was repaired with adhesiolysis near the arterial anastomosis, 2 were critically ill and died soon after transplantation, 1 showed adequate collateral flow, and 1 had a severely stenotic artery that was too small for angioplasty and therefore not amenable to definitive treatment. As for biliary complications (n = 205), surgical revision was required for 80 patients (39.0%), percutaneous transhepatic cholangioplasty/drainage was required for 73 (35.6%), endoscopic retrograde cholangioplasty was required for 39 (19.0%), and retransplantation was required for 13 (6.3%; adults only). Among the 166 adult patients who did not undergo retransplantation for biliary complications, complete resolution was not achieved for 9 patients (5.4%) despite extensive interventions, and they ultimately died of sepsis. In contrast, all biliary complications in the pediatric population (n = 26) were successfully treated.

Patient and Graft Survival

With a median follow-up of 89 months (range = 17-168 months), the 5-year overall patient survival rates for adults were similar in the 2 groups (69% ± 3% for the conduit group versus 74% ± 1% for the standard group, P = 0.15; Fig. 1A). However, the conduit group demonstrated a significantly diminished 5-year graft survival rate in comparison with the standard group (61% ± 3% versus 70% ± 1%, P = 0.01; Fig. 1B). The adjusted HR for graft loss in the conduit group versus the standard group, calculated with a multivariate Cox regression analysis, was 1.38 (95% CI = 1.03-1.85, P = 0.03). During a 5-year period after transplantation, graft loss was encountered in 100 of the 267 patients (37.5%) in the conduit group and in 321 of the 1112 patients (28.9%) in the standard group (P = 0.006; Table 4). The conduit group was more susceptible to graft loss from HAT (P = 0.001), acute rejection (P = 0.004), and infection/sepsis (P = 0.02; Table 4).

image

Figure 1. (A) Overall patient and (B) graft survival rates for adult primary liver transplantation with standard anastomoses and aortohepatic conduits. The patient survival curves are similar for the 2 groups; however, the conduit group (n = 267) had a decreased 5-year graft survival rate in comparison with the standard group (n = 1112, P = 0.01).

Download figure to PowerPoint

Table 4. Causes of Graft Loss
EtiologyAdult PatientsPediatric Patients
Conduit (n = 267)Standard (n = 1112)P ValueConduit (n = 81)Standard (n = 107)P Value
  1. NOTE: The bolded values are significant.

Graft failure, all causes [n (%)]50 (18.7)161 (14.5)0.0815 (18.5)11 (10.3)0.11
Recurrent hepatitis C9 (3.4)54 (4.9)0.300 (0)1 (0.9)>0.99
Primary nonfunction9 (3.4)33 (3.0)0.733 (3.7)0 (0)0.08
HAT16 (6.0)23 (2.1)0.0013 (3.7)5 (4.7)>0.99
Chronic rejection3 (1.1)19 (1.7)0.797 (8.6)3 (2.8)0.10
Acute rejection8 (3.0)9 (0.8)0.0041 (1.2)0 (0)0.43
Other5 (1.9)23 (2.1)0.841 (1.2)2 (1.9)>0.99
Patient death, all causes [n (%)]50 (18.7)160 (14.4)0.0810 (12.3)8 (7.0)0.26
Infection/sepsis21 (7.9)49 (4.4)0.023 (3.7)3 (2.8)>0.99
Cardiac/respiratory11 (4.1)36 (3.2)0.480 (0)0 (0)
Recurrent hepatocellular carcinoma5 (1.9)27 (2.4)0.590 (0)0 (0)
Cerebral4 (1.5)13 (1.2)0.762 (2.5)0 (0)0.18
De novo malignancy3 (1.1)6 (0.5)0.393 (3.7)1 (0.9)0.32
Other6 (2.2)29 (2.6)0.742 (2.5)4 (3.7)0.70

The mid- and long-term outcomes for children in the conduit group were significantly impaired in comparison with those for children in the standard group with respect to patient survival rates (82% ± 4% versus 92% ± 3% at 3 years, P = 0.04; Fig. 2A) and graft survival rates (73% ± 5% versus 86% ± 3% at 3 years, P = 0.009; 69% ± 5% versus 81% ± 4% at 5 years, P = 0.03; Fig. 2B). In the pediatric population, the use of aortohepatic conduits did not emerge as an independent predictor in a multivariate Cox regression analysis for either patient survival (HR = 1.31, 95% CI = 0.23-7.56, P =0.76) or graft survival (HR = 1.53, 95% CI = 0.49-4.81, P = 0.47). Nonetheless, when intraoperative and postoperative variables (blood loss, transfusion requirements, operative time, and length of stay) were removed from the Cox regression analysis, the adjusted HRs were 2.48 for patient death (95% CI = 1.04-5.90, P = 0.04) and 2.05 for graft loss (95% CI = 1.03-4.08, P = 0.04) in the conduit group with the standard group as a reference. In contrast to the adult cohort, no striking differences in the causes of graft loss after transplantation were observed between children in the conduit and standard groups (Table 4).

image

Figure 2. (A) Overall patient and (B) graft survival rates for pediatric primary liver transplantation with standard anastomoses and aortohepatic conduits. The conduit group (n = 81) showed worse patient (*P = 0.04 at 3 years) and graft survival rates (**P = 0.009 at 3 years and P = 0.03 at 5 years) than the standard group (n = 107).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

There is no doubt that aortohepatic conduits in liver transplantation provide a lifesaving solution for many patients with excellent prognoses. Table 5 summarizes previous studies describing the use of aortohepatic conduits in more than 90 cases. Although several groups have shown an association between conduits and HAT via multivariate analyses, only the odds ratio of a logistic regression analysis was reported for a time-to-event outcome.[4, 7, 23] Moreover, none of these groups calculated risk-adjusted survival, and the true consequences of conduits remained ambiguous. Therefore, this series is the first to clearly delineate the long-term deleterious effects (increased complication rates and diminished graft survival) of aortohepatic conduits in adult primary transplantation. Representing one of the largest studies to date, our results offer revealing insights into the indications for aortohepatic conduits. In adult primary transplantation, the placement of a conduit should be strictly limited to recipients whose hepatic and splenic arteries or celiac trunks are not suitable for direct anastomosis despite exhaustive measures to isolate the vessels. We reluctantly accept the shortcomings of conduits only for patients who can otherwise receive no arterial inflow.

Table 5. Reports on the Use of Aortohepatic Conduits in Liver Transplantation
StudyGraftsConduits (n)Incidence of HAT (%)5-Year Graft Survival (%)
ConduitsNonconduitsP ValueConduitsNonconduitsP Value
  1. NOTE: The bolded values are significant.

  2. a

    The incidence of HAT for supraceliac conduits was 10.4% (relative risk = 5.76).

  3. b

    The rate for early HAT-related mortality (≤1 month after transplantation) was 13.3%.

  4. c

    Late HAT (>1 month after transplantation).

  5. d

    The odds ratio for late HAT (>1 month after transplantation) was 10.69.

  6. e

    Early HAT (≤1 month after transplantation).

  7. f

    The odds ratio for early HAT (≤1 month after transplantation) was 3.13.

  8. g

    The odds ratio for HAT was 2.99.

  9. h

    The HR for late HAT (>1 month after transplantation) was 4.91.

Muiesan et al.[13] (1998)Primary and retransplant119 (adult)4.22.90.55
99 (pediatric)4130.06
Zamboni et al.[9] (2002)Primary and retransplant955.33.2NS67.768.6NS
Stange et al.[2] (2003)Primary and retransplant95 (adult)ab
Vivarelli et al.[23] (2004)Primary and retransplant108 (adult)6.5cd 1.0cd0.001
Del Gaudio et al.[4] (2005)Primary and retransplant101 (adult)12.9ef4.7ef0.00253.469.20.01
Nikitin et al.[14] (2008)Primary only149 (adult)4.72.90.2159670.09
Duffy et al.[7] (2009)Primary and retransplant3328.4g4.7g0.0149
Warner et al.[8] (2011)Primary and retransplant117 (adult)11.1e4.4e0.02
This study (2013)Primary only267 (adult)4.1ch0.7ch<0.00161700.01
81 (pediatric)680.5669810.03

In adult primary transplantation, a significantly increased incidence of late HAT associated with aortohepatic conduits has rarely been described.[23] We also found a strong association between late HAT and ischemic cholangiopathy or retransplantation, and this was in accordance with previous studies.[24-27] Consequently, the conduit group had an increased risk for graft loss up to 5 years after transplantation due to HAT and infection/sepsis, the latter contributing to some extent to the higher number of acute rejection episodes as a counteraction of decreased immunosuppression.

The histopathology of the explanted liver graft due to late HAT has not been investigated previously. Against expectations, only 12% had atherosclerosis. Instead, there was a higher occurrence of central perivenulitis (up to 44%) in the conduit group, whereas the prevalence in adults was 28% in a recent study.[28] Ischemia secondary to the occlusion of large and medium arteries has been proposed as one of the mechanisms of centrilobular injury in the liver graft, but there is increasing evidence that central perivenulitis is related to rejection (38% had chronic rejection in our series).[29] Late HAT could be a manifestation of late rejection, and the dilemma of cause and effect remains to be solved.

Children with a conduit also demonstrated inferior patient/graft survival rates according to log-rank tests. Nevertheless, a multivariate Cox regression analysis showed discordant results depending on whether intraoperative and postoperative variables were included in the model. Two factors likely contributed to this inconsistency: (1) blood loss, transfusion requirements, operative times, and lengths of stay reflected the magnitude of the surgery and confounded the placement of conduits, and (2) the sample size of the pediatric patient population was small. Intriguingly, in children with an age < 12 months or a body weight ≤ 6 kg, the 5-year graft survival rates were similar for the conduit and standard groups (data not shown). If standard arterial revascularization is not readily achievable in these subsets, more liberal use of aortohepatic conduits may be indicated. Our observation is in agreement with only a few other centers describing identical or even superior outcomes with conduits up to 4 years after transplantation despite the reservation of conduits for more demanding circumstances (ie, small children with small arteries).[10-13, 30]

Our findings underscore crucial steps in the perioperative period for successful arterialization of the liver allograft and patient recovery. Pretransplant recipient imaging (triple-phase computed tomography unless contraindicated) for detecting anomalies of the hepatic arteries and checking the takeoffs of the renal arteries is invaluable for building surgical strategies and preventing unintentional injuries. Donor-recipient size matching and delicate and meticulous surgical maneuvers at organ procurement (iliac and carotid arteries should be kept as long as possible with minimal manipulation at all times) and during transplantation (gentle dissection and careful clamping of the recipient hepatic artery) are paramount in preserving normal arterial anatomy and facilitating standard anastomoses.[7] In complex cases such as patients with multiple previous surgeries or embolization procedures and portomesenteric thrombosis, mass clamping of the hepatic hilum is an efficacious alternative for reducing undesired arterial injury.[31] If aortohepatic conduits are inevitable, the alignment (geometry and angle) of the arterial graft is the key to minimizing the wall shear stress gradient, which triggers vascular endothelial responses resulting in intimal hyperplasia, atherosclerosis, and thrombosis.[32, 33] More attention should be paid to the northbound route (against gravity) of the infrarenal aortohepatic conduit because its use was abandoned in 1 center on account of failure.[34] To maintain graft patency, a natural curve from the infrarenal aorta to the hepatic hilum is critical. The clothoid curve (also known as the Euler or Cornu spiral), used in railway, highway, and roller coaster designs, is inspiring because it produces an ideal shape for a transition (or easement) curve connecting a straight section with a section of a given curvature.[35, 36] Retroperitonealization of the aortohepatic conduit is recommended for preventing an internal hernia.[37] Recently, the phenomenon of the neovascularized liver was investigated in detail, and late HAT and Roux-en-Y anastomosis were associated with its occurrence.[38] It is an alluring hypothesis that hepaticojejunostomy may serve as a fail-safe mechanism in patients with an aortohepatic conduit, yet clinical evidence is insufficient to draw a definitive conclusion. Lastly, we concur with other centers advocating posttransplant antiplatelet prophylaxis.[2, 5, 8, 14, 39]

The limitations of this study include the following: (1) its retrospective design; (2) the selection bias for which patient received a conduit; (3) the era bias with patient data collected during a 12-year period; (4) potential confounders [ie, recipient and donor background information (age, sex, race, cold and warm ischemia times, and other factors) and intraoperative factors (estimated blood loss, transfusion requirements, and operative times) that may affect survival outcomes]; and (5) the diagnostic bias, which arose from the lack of routine screening for HAT and stenosis beyond the initial hospital stay. To control and minimize the first 4 drawbacks, we restricted our investigation to primary transplantation, and patients were further divided into adults and children (stratification). Furthermore, a multivariate analysis was applied to control multiple risk factors. The issue related to the fifth limitation, the diagnostic bias, is that asymptomatic hepatic artery problems that could have occurred late in the posttransplant course may have been undetected, and their true incidence could have been higher than that reported in this study. However, all patients who presented with clinically significant episodes (an elevation of liver enzyme levels and/or signs of hepatic dysfunction) were subjected to a detailed workup and medical/surgical interventions if indicated. Therefore, our results should remain valid.

In conclusion, the use of aortohepatic conduits in adult primary liver transplantation was associated with an increased risk of complications (notably late HAT). The adjusted 5-year graft survival rate for patients with a conduit was impaired in comparison with the rate for patients with standard reconstruction, and this highlights the importance of strict patient selection. In children, conduits may be placed judiciously.

ACKNOWLEDGMENT

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

The authors thank Debbie Weppler, Mary Murtha, and Yoshiki Hiraki for their contributions to the data collection; Dr. Koji Okabayashi for his critical review of the manuscript; and Dr. Sergio Santiago for his outstanding technical assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES