Extensive Use of Split Liver for Pediatric Liver Transplantation: A Single-Center Experience

Transcription

1 ORIGINAL ARTICLES Extensive Use of Split Liver for Pediatric Liver Transplantation: A Single-Center Experience Marco Spada, Bruno Gridelli, Michele Colledan, Andrea Segalin, Alessandro Lucianetti, Wanda Petz, Silvia Riva, and Giuliano Torre The results of the extensive use of in situ liver splitting in a pediatric liver transplant program are presented. All referred donors were considered for split liver, and when the donor-recipient body weight ratio (DRWR) was greater than 2, the grafts were split. A modified split-liver technique was adopted when the DRWR was 2 or less. Eighty liver procurements were attempted and 72 (90%) were performed, enabling 65 children to receive 42 split, 22 whole, and 8 reduced-size livers. The right portions of the grafts were transplanted by other centers into adults. Median patient waiting time was 22 days, with no mortality on the waiting list. After a median follow-up of 14 months, overall patient and graft survival rates were 85% and 81%, respectively. Fifty-eight children received a single allograft, whereas 7 children required retransplantation. Two-year actuarial survival rates were 85% for split-liver recipients, 84% for whole-liver recipients, and 67% for reduced-size liver recipients. Vascular complications developed in 18% of the patients, with no difference among the 3 groups with different technique. Biliary complications developed in 25% of the children, mainly in reduced-size and split-liver recipients. Patient and graft survival rates for right split-liver grafts were 84% and 79%, respectively. Adopting a liberal policy of liver splitting provides allografts of optimal quality for pediatric transplantation, allowing a dramatic decrease in the waiting list time. The in situ split-liver technique should be considered the method of choice for expanding the cadaveric liver donor pool. (Liver Transpl 2000;6: ) T he majority of Italian children with end-stage liver disease, and their families, have had to go to foreign transplant centers for the last 15 years to receive care. One of the objectives we pursued, first at the Maggiore Hospital of Milan and then at the Riuniti Hospital of Bergamo, was to make pediatric liver transplantation a readily available therapy in Italy. We started a pediatric liver transplant program in Milan in 1988, and for the first 6 years, the annual number of transplantations was less than 10. In 1994, for a number of reasons, including increased donor availability and referral of candidates from pediatric centers, the transplant activity increased significantly (Fig. 1). To better meet the needs of the children under our care, we moved to the Bergamo hospital in October 1997, where we expanded the use of split-liver transplantation. Because the number of pediatric transplant donors is less than the number of pediatric transplant recipients, and the majority of the recipients are aged younger than 2 years, it has been common practice to use reduced-size livers from adult donors, discarding the excess liver parenchyma, particularly in the 1980s and early 1990s. 1-4 This practice shifted the imbalance between need and availability of liver grafts from the pediatric to the adult population of transplant candidates. The incidence of end-stage liver failure requiring transplantation in the pediatric age group is 2 in 10,000 live births 5 ; therefore, approximately 100 of the 500,000 children born each year in Italy will require liver transplantation. In Italy, there are currently 3 organ procurement transplant agencies that allocate cadaver donor organs. Our center is linked to North Italy Transplant (NITp), which coordinates the activity of 7 liver transplant centers. Six of these centers provide care mainly to adults, whereas the Bergamo center mainly focuses on pediatric recipients. In 1998, the NITp organized the organ procurement of approximately 300 cadaver organ donors. 6 We therefore reasoned that the extensive use of the liver-splitting procedure would generate enough liver grafts for children without interfering with the waiting lists of adult patients. We report the impact of the adoption of a liberal policy of in situ liver splitting on our pediatric liver transplant program. Patients and Methods Study Population Between October 1997 and October 1999, a total of 72 pediatric liver transplantations were performed in 65 chil- From the Liver Transplantation Center, Ospedali Riuniti di Bergamo, Bergamo, Italy. Address reprint requests to Marco Spada, MD, PhD, Centro Trapianti di Fegato, Chirurgia III, Ospedali Riunuti, Largo Barozzi 1, Bergamo, Italy. Telephone: ; FAX: ; mspada@ospedaliriuniti.bergamo.it Copyright 2000 by the American Association for the Study of Liver Diseases /00/ $3.00/0 doi: /jlts Liver Transplantation, Vol 6, No 4 (July), 2000: pp

2 416 Spada et al Figure 1. Number of pediatric liver transplantations performed in Milan between 1988 and dren at the Liver Transplantation Unit of the Ospedali Riuniti of Bergamo, Italy. The transplant recipients were 32 boys and 33 girls, ranging in age from 0.1 to 21 years (mean SD, years; median, 1.6 years). Body weight ranged from 2.8 to 62 kg (mean, kg; median, 10.5 kg). Figure 2 shows transplant recipient distribution by age: 39% of the transplant recipients were aged younger than 1 year, and 66% were aged younger than 3 years. Two patients aged 19 and 21 years were included on the study because they were referred to our center for liver failure secondary to extrahepatic biliary atresia. The original liver diseases of the patients are listed in Table 1. Sixty-three children underwent primary liver transplantation, whereas 2 patients who previously underwent transplantation elsewhere received a second graft because of chronic rejection of the first transplanted liver. Of 65 transplant recipients, 4 patients (6%) were United Network for Sharing Organs (UNOS) status 1, 6 patients (9%) were UNOS status 2, and 55 patients (85%) were UNOS status 3. Splitting Criteria and Donor Procedures Since the beginning of the Bergamo pediatric liver transplantation program, we adopted a liberal policy of liver splitting. The decision of whether to split a graft was based mainly on recipient rather than donor criteria. We did not exclude children who required retransplantation or who had fulminant hepatic failure. In the NITp area, cadaver liver donors are allocated on a rotation basis to the 7 transplant centers. The liver of every donor assigned to our center was evaluated for elective transplantation when at least 1 blood group type A, type B, type O (ABO)-compatible recipient with a donor-recipient body weight ratio (DRWR) of 12 or less was wait listed. Gross pathological findings at the procurement were the main criteria for organ acceptance or refusal. Donor evaluation did not require special or additional invasive or noninvasive tests. Particular care was applied to the evaluation of liver function, hemodynamic and metabolic status, and liver macroscopic appearance of donors aged older than 50 years, but age alone was not an exclusion criterion. Every accepted Figure 2. Recipient distribution by age.

3 Split Liver for Pediatric Liver Transplantation 417 Table 1. Original Liver Disease of the Patients Diagnoses No. of Patients % Extrahepatic biliary atresia Intrahepatic cholestatic disease* 5 8 Secondary cirrhosis 5 8 Tumor 4 6 Fulminant hepatitis 2 3 Chronic rejection 2 3 Metabolic disease 2 3 Others 2 3 *Alagille syndrome, Byler disease. Cryptogenetic cirrhosis, Langerhans cells histiocytosis, autoimmune hepatitis. Hepatocarcinoma, hemangiosarcoma, rhabdomyosarcoma. Patients who previously underwent transplantation elsewhere. Crigler-Najjar disease, perinatal hemochromatosis. Budd-Chiari syndrome, arteriohepatic fistula. graft was attributed to the most urgent recipient on the waiting list. Time on the waiting list was an allocation criterion for patients with the same UNOS status. The liver was procured and transplanted as a whole liver when the DRWR was 2 or less. Donor procedures were performed using the standard technique of rapid multiorgan procurement. When the DRWR was between 2 and 12, the organ was considered for split liver. Of the 2 grafts obtainable with the splitting procedure, as a rule, the right was offered to another center, with the agreement of a restitution of the split liver at the first opportunity. The allocation of the left grafts received from other centers according to this agreement was performed using the same criteria previously described. Standard surgical facilities for multiorgan procurement were used, and no special equipment was required from the donor hospital. When possible, the splitting procedure was performed in situ as described by Rogiers et al 7 by a surgical team composed of members of both centers involved in the transplantation. Briefly, the procedure consists of sectioning the liver along the falciform ligament, dividing the left lateral segment ([LLS] segments II and III of Couinaud) from segments IV to VIII and the caudate lobe. The left graft included the left hepatic vein, left branch of the portal vein, and left branch of the hepatic artery, along with the common hepatic artery and celiac axis. The right graft included the vena cava, right branch of the hepatic artery, and portal vein. During procurement, the following variations in hepatic artery anatomy were encountered: left hepatic artery coming from the left gastric artery with no left branch originating from the proper hepatic artery (2 cases); left hepatic artery coming from the left gastric artery and right hepatic artery originating from the superior mesenteric artery, with absence of the proper hepatic artery (1 case); accessory left hepatic artery coming from the left gastric artery (1 case); anomalous left gastric artery originating from the left hepatic artery (1 case); and right hepatic artery originating from the superior mesenteric artery, with no right branch coming from the proper hepatic artery (3 cases). We recently used and described an alternative in situ splitting technique that results in left grafts substantially larger than those obtained with the standard technique. 8 This technique consists of sectioning the liver along a plane directed from the right border of the gallbladder fossa to the division of the portal vein and to the right margin of the median hepatic vein. The left graft (segments I to IV) includes the vena cava with the left and median hepatic veins, portal vein, and left branch of the hepatic artery along with the common hepatic artery and celiac axis. The right graft (segments V to VIII) includes the right hepatic vein with a patch of vena cava, right branch of the hepatic artery, and right branch of the portal vein. This alternative splitting technique was recently adopted for patients with a DRWR of 2 or less and a graft-recipient weight ratio of 0.8 or greater. We estimated the left graft volume (segments I to IV) using a formula for calculating the standard liver volume in the donors from their body surface area 9 and assuming that the left lobe volume was 40% of the donor s estimated liver mass. 10 The splitting procedure was performed ex situ only when required for hemodynamic instability of the donor or logistic reasons. In all cases, an effort was made to limit ischemic time as much as possible. When a center with a suitable recipient for the right graft could not be found, the liver was reduced on the back table using the standard technique. 3 Organ procurements were performed in 72 donors ranging in age from 0.1 to 65 years (mean SD, years; median, 17.5 years). Body weight ranged from 3.8 to 105 kg (mean, kg; median, 60 kg). Recipient Procedures Split and reduced grafts were transplanted using the piggyback technique, with retention of the recipient vena cava in all but 1 case. In 1 child with hepatic hemangiosarcoma, the inferior vena cava was removed with the liver and a neocava was created on the back table using the donor iliac vein. During split-liver transplantation, attention was given to prevent venous outflow obstruction, leaving the left hepatic vein short and using the triangulation technique described by Emond et al. 11 Direct end-to-end anastomosis of the portal vein was performed in 38 cases (90%). In 4 cases (10%) of biliary atresia, when the recipient portal vein was severely hypoplastic, the portal vein of the graft was anastomosed to the confluence of the splenic vein and superior mesenteric vein of the recipient with an interposition graft. As previously described, the entire hepatic artery and celiac axis were kept with the LLS. Arterial reconstruction was performed by anastomosing the celiac axis or the common hepatic artery of the graft to the celiac axis, common hepatic artery, or aorta of the recipient. In 4 cases, arterial reconstruction was accomplished by means of an arterial interposition graft. When a

4 418 Spada et al unique or accessory anomalous left hepatic artery originating from the left gastric artery was present, no particular bench reconstruction was required because the entire celiac axis was with the LLS graft. Details of the arterial reconstruction are listed in Table 2. Left graft biliary reconstruction was always with a Roux-en-Y hepaticojejunostomy. When possible, small bile ducts were enlarged by anterior incision. Inclusion of more than 1 ductal orifice in the biliary enterostomy was required in 13 cases. All vascular and biliary reconstructions were performed using surgical loupe magnification (original magnification 3.5). Whole livers were transplanted with the standard technique, without preservation of the vena cava. Postoperative Management No special posttransplantation care was required for recipients of split grafts. All patients were managed according to the established protocol for pediatric liver transplantation, including Doppler ultrasound performed daily for the first postoperative week to evaluate hepatic artery, portal vein, and hepatic vein flow. Immunosuppression consisted of oral cyclosporine microemulsion and steroids in 41 patients and oral cyclosporine microemulsion, steroids, and azathioprine in 4 children. Thirteen children were administered oral cyclosporine, basiliximab, and steroids, and another 7 patients were administered tacrolimus and steroids. Statistical Analysis Values are expressed as mean SD or median and range. For statistical comparison, log-rank test for Kaplan-Mayer survival analysis, Mann-Whitney test, or Kruskal-Wallis test for continuous data, Chi-squared test, and/or Fisher s exact probability test for categorical data were used. P less than.05 is considered statistically significant. Results Donor Procedure Eighty liver procurements were begun: in 4 cases, the liver was not used because of steatosis. In 4 donors considered for split liver, the procedure could not be completed because the donor became hemodynamically unstable (n 1) or because of anatomic and/or dimensional reasons (n 3). In these cases, the whole liver was procured and transplanted at another center. Of 72 procurement procedures, 42 procedures (58%) were split livers. In all but 2 cases, the splitting procedures were performed using the in situ technique, and in 36 cases, consisted of the procurement of the LLS, in 3 of the left lobes (segments I to IV), and in 1 case, of the right segments I and IV to VIII. In 8 cases (11%), a transplant center with a suitable recipient for the right graft could not be found, and the liver, procured as a whole, was reduced on the back table. In 22 donors (31%), the liver was procured and transplanted as a whole organ. No procurement procedures were abandoned because of intraoperative technical complications, and no extrahepatic organs were jeopardized, allowing the regular procurement of kidneys, pancreases, hearts, and lungs. Blood transfusion was needed in some cases during the splitting, mainly in donors with low preoperative hemoglobin levels. The split-liver procedures added minutes to the procurement when the LLS was procured, whereas using the alternative technique required more minutes. In the case of procurement of the right segments I and IV to VIII, the LLS was used by our center for a combined split-liver and small-bowel transplantation, performed in a 1.3-year-old patient. The demographics of the 72 donors, classified by the adopted procurement technique, are listed in Table 3. Table 2. Type of Arterial Reconstruction in the 42 Split-Liver Transplantations Split Liver Celiac axis Common hepatic artery Arterial Anasomosis Recipient Celiac axis Common hepatic artery No. of Cases 2 15 No. of Grafts 0 2 Aorta 5 0 Celiac axis 1 0 Common hepatic 17 0 artery Aorta 2 2 Patient and Graft Survival Patient median time on the waiting list was 22 days, with a significant decrease from 60 days (October through December 1997) to 7 days (August through October 1999). No patient died on the waiting list. Of the 65 transplant recipients, 39 patients (60%) received a split liver; 35 patients, an LLS; 3 patients, a left lobe; and 1 patient, a right graft (segments I and IV to VIII). Six children (9%) received a reduced-size liver (LLS), and the remaining 20 patients (31%) received a whole organ. Patient demographics classified by graft type are listed in Table 4. At a median follow-up of 17 months (range, 2 to 28

5 Split Liver for Pediatric Liver Transplantation 419 Table 3. Profiles of Donors Classified by the Adopted Procurement Technique Surgical Technique Whole Reduced II-III I-IV P No. of donors Age (yr) * (0.1-59) (6-62) (3-65) (20-33) Weight (kg) (3.8-90) (20-64) (30-105) (60-85) Ischemia (min) * ( ) ( ) ( ) ( ) NOTE. Data reported as mean SD (range). *Whole and reduced liver v split II-III. Whole and reduced liver v split II-III; reduced v split I-IV. Split months), 55 patients (85%) are currently alive. Overall 1-year and 2-year actuarial patient survival rates are 84% and 84%, respectively (Fig. 3). The 1-year and 2-year actuarial graft survival rates are 78% and 78%, respectively (Fig. 3). Fifty-eight patients (89%) received a single allograft and 7 children (11%) required retransplantation, one from an ABO-incompatible donor. Three patients were recipients of a whole liver (15%), whereas 4 children were in the split-liver group (10%). Data for patients who underwent retransplantation are listed in Table 5. The 1-year and 2-year actuarial patient survival rates of patients who underwent transplantation with a split liver are 85% and 85%, respectively; for patients who received a reduced-size liver, 67% and 67%, respectively; and for patients who received a whole liver, 85% and 85%, respectively (Fig. 4). The actuarial allograft survival rates at 1 and 2 years are 76% and 76% for split livers, 63% and 63% for reduced livers, and 78% and 78% for whole livers, respectively. For patients who received a split liver, actuarial survival rate based on UNOS status at the time of transplantation was determined. The 1-year and 2-year survival rates of patients with UNOS status 2 and 3 were 100% and 100% and 90% and 90%, respectively (Fig. 5). Of the 4 patients with UNOS status 1, 2 patients died at 9 months and at 17 days after transplantation, whereas 2 patients are alive. Table 4. Profiles of Recipients Classified by Graft Type Surgical Technique Whole Reduced II-III I-IV P No. of recipients Age (yr) * (0.5-19) ( ) (0.1-7) (10-21) Weight (kg) (6-62) (4-11) (3-20) (38-55) DRWR ( ) ( ) ( ) ( ) RBC (ml/kg) NS ( ) (5-63.1) (2-88.1) Abbreviations: RBC, red blood cells; NS, not significant. *Whole versus reduced and split liver II-III; split II-III versus split I-IV. Whole versus reduced, split liver II-III, and split I-IV; reduced versus split I-IV; split II-III versus split I-IV. Each group versus the others. Intraoperative RBC transfusion. Split

6 420 Spada et al Figure 3. Kaplan-Meier survival curve showing 1- and 2-year actuarial overall patient (X) and graft ( ) survival rates. Recipient Mortality and Morbidity Perioperative mortality occurred in 8 patients (12%) after transplantation. Four children underwent retransplantation for technical complications and eventually died, whereas the procedure was successful in 3 patients (Table 5). Four patients died within 1 month of the primary transplantation. Data concerning these latter patients and 2 other children who died 39 and 287 days after transplantation are listed in Table 6. Of the 2 patients who developed primary nonfunction (PNF), 1 patient was a 12-year-old boy with Budd-Chiari syndrome and protein C, protein S, and antithrombin III deficiency who received a whole liver and died 3 days after transplantation with disseminated intrahepatic thrombosis, possibly related to his original disease. Postmortem examination showed disseminated intravascular coagulation. The second child was a 14-month-old girl with hepatoportal sclerosis and Figure 4. Kaplan-Meier survival curve showing 1- and 2-year actuarial patient survival rates for recipients of split-liver (X) allografts, reduced-size liver ( ) allografts, and whole livers ( ). multiple arteroportal intrahepatic fistulas who underwent transplantation with an LLS from a split liver with a DRWR of 7.78 and died 2 days after surgery. Her liver graft failed possibly because of insufficient portal inflow. Three other children died after primary split-liver transplantation. The first, a 2-month-old boy, received an urgent LLS for acute hepatic failure of unknown origin. Allograft function in the posttransplantation period was excellent, but the patient did not show a neurologic recovery, and a brain computed tomographic scan and magnetic resonance imaging showed diffuse cerebral edema with multiple intracerebral collections. The boy eventually died 9 months after transplantation with normal liver function. The second patient, a 2-month-old girl weighing 3.4 kg, received an urgent split-liver transplant for acute hepatic failure of unknown origin from a donor Table 5. Profiles of Transplant Recipients Who Required Retransplantation Patient No. Age (yr) 1st TX Technique DRWR Cause of Failure 2nd TX Technique DRWR Status 1 12 Whole 1.3 HA thrombosis Whole 0.5 Alive 2 1 Reduced 4.4 PV HV thrombosis Split (II-III) 10.0 Dead Whole 2.7 HA PV thrombosis Reduced 7.1 Dead Split (II-III) 7.6 HV thrombosis Reduced 3.2 Alive Whole 0.6 HA PV thrombosis Split (II-III) 9.7 Dead 6 1 Split (right) 1.7 HA thrombosis Split (II-III) 6.5 Dead 7 21 Split (I-IV) 1.5 HA thrombosis Whole 1.6 Alive Abbreviations: HA, hepatic artery; PV, portal vein; HV, hepatic vein; TX, transplantation.

7 Split Liver for Pediatric Liver Transplantation 421 Figure 5. Kaplan-Meier survival curve showing 1- and 2-year actuarial split-liver recipient survival rates for patients with UNOS status 1 ( ), UNOS status 2 (X), and UNOS status 3 ( ). There is a statistically significant difference (P F.05) between UNOS status 2 and 3 patients compared with UNOS status 1 patients. with a DRWR of The oversized graft precluded direct closure of the abdomen, and a polypropylene prosthesis was used. The patient developed respiratory failure requiring prolonged mechanical ventilation and sepsis, and on day 17, portal vein thrombosis was diagnosed. Revision of the anastomosis was attempted, but during surgery, the child became hemodynamically unstable, had a cardiac arrest with resuscitation, and at the end of the procedure was transferred to the pediatric intensive care unit (ICU) in unstable condition and died soon after. A 16-month-old girl who received an LLS for extrahepatic biliary atresia developed acute and fatal gastrointestinal bleeding after a liver biopsy performed on postoperative day 35 to confirm clinical suspicion of acute rejection. The last patient who died after primary liver transplantation was a 5-month-old boy who received a reduced liver for extrahepatic biliary atresia. Allograft function in the immediate postoperative period was good; however, on day 11 he developed aspiration pneumonia and eventually had a fatal cardiac arrest. Postmortem examination showed nonspecific liver damage related to preservation injury and bilateral severe pneumonia. Four patients experienced early vascular complications and underwent unsuccessful retransplantation. Two girls aged 20 months and 6 months underwent whole-liver transplantation for extrahepatic biliary atresia. Both grafts were from pediatric donors aged 8 years and 10 days, weighing 24 and 3.8 kg, respectively. The first patient developed acute intraoperative thrombosis of the interposition graft, which was placed between the recipient aorta and the common hepatic artery of the graft, with acute graft nonfunction. The child underwent retransplantation with a reduced graft on day 1. She was transferred to the pediatric ICU in stable condition, became hypotensive a few hours later, and had a fatal cardiocirculatory arrest. Postmortem examination showed hepatic necrosis with thrombosis of the portal vein, dissection of the hepatic artery, necrosis of the right kidney, and pulmonary edema. The second girl developed acute hepatic artery and portal vein thrombosis on day 3 and underwent revision of the anastomosis with the use of an interposition vein graft and subsequent split-liver retransplantation. Liver recovery was regular, but the patient developed sepsis and died on day 9 after primary transplantation. Postmortem examination showed bacterial bilateral pneumonia and meningitis. A 1-year-old girl who underwent transplantation for Byler s disease with a reduced-size liver graft from a pediatric donor (age, 9 years; weight, 35 kg) developed acute thrombosis of the left hepatic vein and portal vein and underwent urgent retransplantation with a Table 6. Profiles of Transplant Recipients Who Died After Primary Transplantation Patient No. Age (yr) Transplant Technique UNOS Status DRWR Cause of Death Reduced Pneumonia Whole PNF Split (II-III) PNF Split (II-III) Neurological complication Split (II-III) GI hemorrhage Split (II-III) Sepsis 17 Abbreviations: GI, gastrointestinal; PO, postoperative. PO Day

8 422 Spada et al split-liver graft. During the procedure, the hemodynamically unstable patient experienced a fatal cardiocirculatory arrest. Finally, a 20-month-old boy with extrahepatic biliary atresia underwent transplantation with a right graft (segments I and IV to VIII) from a pediatric donor (age, 2 years; weight, 14 kg). Acute hepatic artery thrombosis was diagnosed on day 3, and an urgent retransplantation was performed the next day with an LLS. The liver showed a normal recovery of function; however, the boy developed progressive brain damage and died on day 7 after primary transplantation. Vascular complications occurred in 13 patients (20%), with hepatic artery thrombosis developing in 6 patients (9%), portal vein thrombosis in 7 patients (11%), and venous outflow obstruction in 4 patients (6%). Complications were equally distributed in wholeliver (20%), split-liver (21%), and reduced-liver (17%) recipients. Arterial occlusion was diagnosed in 2 splitliver recipients (5%), 1 who received a right graft and the other a left lobe, and in 4 whole-liver recipients (20%). Portal vein complications occurred in 4 splitliver recipients (10%), 1 reduced-size liver recipient (17%), and 2 whole-liver recipients (10%). Two children (5%) who received split livers experienced venous outflow obstruction compared with 1 patient with a reduced-size liver (17%). Details regarding vascular complications and their treatment are listed in Table 7. Biliary complications developed in 16 children (25%), primarily in patients who received a split liver (11 of 39 patients) or reduced liver (4 of 6 patients) compared with the whole-liver recipients (1 of 20 patients). Biliary complications included anastomotic stricture (n 7), anastomotic leak (n 3), or cutedge leak (n 5). Stenoses were successfully treated in all but 1 patient with percutaneous transhepatic cholangiography, drainage, and balloon dilatation. One patient required revision of the anastomosis. Cut-edge leaks were treated with percutaneous drainage in 4 patients and surgical revision in 1 patient. Anastomotic leaks always required surgical intervention on the anastomosis. A 10-month-old boy who underwent transplantation with an LLS for extrahepatic biliary atresia developed progressive intrahepatic biliary tree distension 3 months after transplantation. Percutaneous transhepatic cholangiography showed a patent hepaticojejunostomy with an intestinal stricture a few centimeters below the bilioenterostomy, where an intestinal valve had been created at the time of Kasai portoenterostomy. The child underwent laparotomy with resection of the intestinal loop and redo of the hepaticojejunostomy. No biliary complications directly Table 7. Profiles of Transplant Recipients With Vascular Complications Patient No. Age (yr) Recipient Weight (kg) Age (yr) Donor Weight (kg) Transplant Technique Complication Treatment Status Split (II-III) Venous outflow Balloon dilatation Alive obstruction Whole HA thrombosis Retransplantation Alive Split (II-III) PV thrombosis Observation Alive Reduced LHV PV thrombosis Retransplantation Dead Whole HA thrombosis* Retransplantation Dead Whole LHV HA PV Revision anastomoses Dead thrombosis Split (II-III) Venous outflow Retransplantation Alive obstruction Split (II-III) PV thrombosis Revision anastomosis Alive Whole HA PV thrombosis Retransplantation Dead Split (II-III) PV thrombosis* Observation Alive Split (II-III) PV thrombosis* Revision anastomosis Dead Split (right) HA thrombosis Retransplantation Dead Split (I-IV) HA thrombosis Retransplantation Alive Abbreviations: HA, hepatic artery; PV, portal vein; LHV, left hepatic vein. *Anastomosis performed with an interposition graft.

9 Split Liver for Pediatric Liver Transplantation 423 resulted in graft or patient loss. Details for biliary complication are listed in Table 8. Eight children required reexploration for intestinal perforation, 4 children among the split-liver recipients (10%), 1 child in the reduced-size liver group (17%), and 3 children among the whole-liver recipients (15%). Two patients, 1 in the split-liver group and the other in the reduced-liver group, required reexploration for intra-abdominal hemorrhage. All cases of intestinal perforation and hemoperitoneum were diagnosed early, and none led to graft loss or death. Early Postoperative Course Forty-eight patients required intraoperative blood transfusion, 17 patients (85%) in the whole-liver group, 6 patients (100%) in the reduced-size liver group, and 25 patients (64%) in the split-liver group. Overall intraoperative blood transfusion requirements were ml of red blood cells per kilogram of recipient s body weight (median, 16.6 ml; range, 2 to 88 ml). No statistically significant differences in terms of blood transfusion were observed among the different transplantation techniques, although reduced-size liver recipients required more transfusions than the other patients (Table 4). Posttransplantation graft function recovery tended to be slower in split-liver recipients compared with children who received a whole liver. Table 9 lists the chronologic changes in values for serum alanine aminotransferase, serum total bilirubin, and prothrombin international normalized ratio according to the 2 types of grafts transplanted. Overall median hospitalization was 22 days, with a Table 8. Profiles of Transplant Recipients With Biliary Complications Recipient Cold Ischemia (min) No. of Ductal Orifices Patient No. Age (yr) Weight (kg) DRWR Transplant Technique Complication PO Day Treatment Hospital Stay (d) Whole 476 Stenosis 639 PTC dilatation and 14 stenting Split (II-III) Cut-edge leak 12 Percutaneous drainage Split (II-III) Stenosis 78 PTC dilatation and 30 stenting Reduced Stenosis 21 PTC dilatation and 58 stenting Split (II-III) Stenosis 251 Revision of anastomosis Reduced Stenosis 34 Revision of anastomosis Split (II-III) Roux limb stenosis 115 Revision of anastomosis Split (I-IV) Cut-edge leak 8 Surgical treatment Split (II-III) Anastomotic leak 23 Revision of anastomosis Split (II-III) Stenosis 148 PTC dilatation and 34 stenting Split (I-IV) Cut-edge leak and stenosis Percutaneous drainage, PTC dilatation, and stenting Reduced Anastomotic leak 8 Revision of anastomosis Split (II-III) Cut-edge leak 8 Percutaneous drainage Split (II-III) Stenosis 36 Revision of anastomosis Split (II-III) Cut-edge leak 16 Percutaneous drainage Reduced Anastomotic leak 5 Revision of anastomosis 21 Abbreviations: PO, postoperative; PTC, percutaneous transhepatic cholangiography.

10 424 Spada et al Table 9. Changes in Serum ALT, Serum Bilirubin, and INR Values in Split-Liver and Whole-Liver Transplant Recipients Time After TX (d) ALT (IU/L) Split Liver (n 39) Whole Liver (n 20) Bilirubin (mg/dl) INR ALT (IU/L) Bilirubin (mg/dl) ,050 1,289* * * * * * 3 1,296 1,509* * * * * * * * * * NOTE. Data reported as mean SD. Abbreviations: TX, transplantation; ALT, alanine aminotransferase; INR, international normalized ratio. *P.05, split-liver recipients versus whole-liver recipients. INR median ICU stay of 7 days. Patients who underwent transplantation with a whole liver had a median hospital stay of 15 days; those who underwent transplantation with a reduced-size liver, 21 days; and those who received a split liver, a median of 24 days. Right Grafts Of the right grafts obtained by in situ split liver (segments I and IV to VIII in 37 cases and segments V to VIII in 3 cases), 1 graft was transplanted by our center into a pediatric recipient and is included in our study, whereas the others were transplanted by other centers, as well as the 2 right grafts obtained by ex situ split liver. Nine liver transplant centers collaborated with us, sharing the split-liver grafts. Seven centers were in the NITp area, 1 center was an Italian center affiliated with a different organ procurement agency, and 1 center was a European center in Germany. Data concerning the right lobe grafts from in situ split livers were presented by us at the 1999 Meeting of the Nord Italia Transplant. 12 Briefly, 36 grafts were used for primary liver transplantation, whereas 3 right livers were used for retransplantation of patients who previously underwent transplantation with whole livers (n 2) or right split livers (n 1). The patient population included 24 men and 15 women with a median age of 52 years (range, 19 to 63 years). The most common causes of end-stage liver disease were chronic active hepatitis C (n 12) and B (n 7). At the time of primary transplantation, 9 patients (25%) were UNOS status 3, 26 patients (72.2%) were UNOS status 2, and 1 patient (2.8%) was UNOS status 1. Overall actual patient and graft survival rates after primary right graft in situ split-liver transplantation were 84% and 79%, respectively. One graft (3%) was lost because of PNF. Discussion The use of a liberal policy of liver splitting, made possible by strict collaboration with other transplant centers, allowed us to meet the needs of all the pediatric patients referred to our center, with no mortality on the waiting list. The consensus-based policy adopted by several liver transplant centers in the NITp area to split all suitable donor livers had a significant impact on reducing waiting time for pediatric patients. In a few months, the waiting list time for our pediatric patients decreased from 60 to 7 days. We started to use the split-liver technique on a regular basis at the end of In 1998, this policy had a significant impact on the liver transplant activity in the NITp area: although the increase in the number of donors per million of inhabitants from 1997 to 1998 was 11% (from 16.5 to 18.3 donors/million inhabitants), the number of liver transplantations had a 17% increase (from 243 to 284 transplantations). 6 The formula used allowed us to share grafts with collaborating adult transplant programs located not only in our areas, but also in other areas of the country and in Europe. The collaborating centers obtained good results in terms of patient and graft survival rates, similar to the ones reported by the European Split Liver Registry for adult patients. 13 No definitive study comparing the advantages of in situ versus ex situ techniques have been performed.

11 Split Liver for Pediatric Liver Transplantation 425 Although excellent results have been reported with the ex situ technique, 14 we share the view of the inventors of the in situ technique that this method has several advantages, 15,16 and in particular, it facilitates the sharing of split grafts between collaborating centers. The overall patient and graft survival rates of our series of in situ split-liver transplantations showed an improvement over the ex situ pediatric split-liver transplantation experience. 5,17-23 By in situ splitting of the liver, we were able to have short ischemia times, resulting in good liver allograft function and a low incidence of PNF. The sharing of the 2 split grafts between 2 centers allowed the transplantation of both at approximately the same time, with a reduction of the ischemic time compared with sequential transplantation, which is sometime performed in a single center. This can explain the low incidence of PNF in the adult recipients compared with other reported series. 13 With the in situ technique, good hemostasis of the cut surface at the time of allograft procurement was obtained, resulting in reduced intraoperative blood transfusion requirements during transplantation compared with reducedsize liver grafts. Only 1 split-liver recipient (2%) required reexploration for intra-abdominal hemorrhage compared with 13% of the reduced-size liver recipients. It has been claimed that with the in situ split-liver technique, increasing the operation time in the donor hospital might have a negative effect on organ donation. 14 The widespread use of in situ split-liver transplantation in the NITp area did not reduce the number of donor referrals and resulted in a net increase of 6% in the total available liver allografts. 6 The formula adopted to select the pediatric recipients according to DRWR was accurate in predicting the size compatibility of the left lateral segment. In our series, all the LLS split livers chosen according to the formula fit in the recipients abdomens without complications. Regarding the lower size limit of the split graft, we transplanted the liver as a whole liver when the DRWR and the recipient were small because we were not able to find a pediatric transplant center to accept the right graft, not because of the low dimension of the LLS. Moreover, experience with the use of right grafts from split livers for pediatric transplantation is anecdotal, and the only case that we performed developed a fatal vascular complication related to the small caliber of the right branch of the hepatic artery. For large children, we described and adopted a modified splitliver technique that generates 2 grafts more similar in size, both of which are transplantable in adults or large children with a graft-recipient weight ratio of 0.8 or greater. The upper limit of a DRWR of 12 was exceeded in 2 cases. In the first, an LLS from a donor with a DRWR of 17.6 was transplanted into a UNOS status 1 2-month-old girl with acute hepatic failure of unknown origin. Because of the large volume of the graft, the abdomen was closed by means of a polypropylene mesh after further nonanatomic parenchymal reduction. The child developed portal vein thrombosis, respiratory failure, and sepsis and died 17 days after transplantation. The second patient, a UNOS status 2 12-month-old boy with Alagille syndrome received an LLS from a donor with a DRWR of He did not develop major postoperative complications and is well 4 months after the procedure. The reported incidence of hepatic artery thrombosis for pediatric reduced-liver transplantation ranges from 6% to 25%. 13,16,17,19,24 Arterial complications in our patients (9%) were in this range. Recent series reported a lower incidence of arterial occlusion (0% to 6%) obtained with the use of microsurgical techniques. 16,24,25 In our series of 39 split-liver transplantations, there have been 2 cases of hepatic artery thrombosis, 1 in a child with a right graft (segments I and IV to VIII) and the other in a patient with a left lobe. No recipients of conventional split livers experienced hepatic artery thrombosis. This complication occurred in 4 patients with a whole liver, in 1 child with acute recurrence of Budd-Chiari syndrome, and in 2 patients aged younger than 2 years who received whole grafts from a neonatal and a pediatric donor. Published series have documented a greater incidence of arterial thrombosis and inferior survival in small transplant recipients if they received a full-size liver allograft. 25,26 The first split-liver recipient in our series developed venous outflow obstruction caused by stenosis of the left hepatic vein anastomosis. The stenosis was treated successfully by balloon dilatation. After the adoption of the triangulation technique of Emond et al, 11 we did not observe venous outflow obstructions, except in 1 child who underwent transplantation with an LLS graft with 2 independent hepatic veins from segments II and III that did not allow for the creation of a common anastomotic stump. The patient developed acute graft congestion with PNF and underwent successful retransplantation with a reduced-size graft. The incidence of portal vein thrombosis varied in several reported series from 4% to 13%. 13,19,24,25 Portal complications in our series of split-liver transplantations (10%) were in the upper range, all in small children. The triangulation technique adopted for the left hepatic vein anastomosis decreased the graft s rotation to the right, reducing the risk for portal vein

12 426 Spada et al strictures. Moreover, we used to dissect the recipient s extrahepatic portal vein, even if slightly sclerotic, past the primary portal branches, obtaining an adequate length of portal vein to anastomose to the graft left branch of the portal vein. We limited the use of interposition venous grafts to very sclerotic and small portal veins of patients with biliary atresia, performing the anastomosis to the confluence of the splenic vein and superior mesenteric vein. Three of 4 split-liver recipients who developed portal vein thrombosis underwent transplantation for extrahepatic biliary atresia, and 2 patients required an interposition graft. Another possible risk factor in the development of portal vein thrombosis is the volume of the transplanted graft. Tight closure of the abdomen after very large-for-size liver grafts, particularly in small children, can result in impaired perfusion of the graft, high intrahepatic vascular resistance, and portal vein thrombosis. In our series, 1 patient who underwent transplantation for acute liver failure of unknown cause with an LLS from a donor with a DRWR of 17.6 developed portal vein occlusion. The extensive use of delayed abdominal closure adopted by others might help address this issue. 24 The overall incidence of biliary complication in our series was 25%. Whereas only 1 whole-liver recipient (5%) developed a biliary complication, the incidence of biliary complications was greater in reduced-size and split-liver recipients. In the split-liver group, 2 of the 3 patients who received a graft obtained with the modified split-liver technique (segments I to IV) experienced a biliary leak at the cut edge, and 1 of these patients subsequently developed a biliary structure secondary to fibrosis. To date, we adopted the modified split-liver technique in selected cases, and much more experience must be gained with this method to be able to evaluate the type and incidence of biliary complication. After conventional in situ split-liver transplantation (LLS), 9 of 36 patients developed biliary complications. In 1 previously discussed case, biliary complication was related to the ileal loop stenosis rather than to the split-liver technique itself. In another 2 cases (patients 2 and 5 in Table 8), the development of a biliary complication followed an episode of bowel perforation with sepsis. In the reduced-size liver group, hemoperitoneum in 1 patient and bowel perforation with sepsis in another patient preceded the development of 2 anastomotic leaks (patients 12 and 16 in Table 8). Sanchez-Urdazpal et al 27 reported increased bile duct complications in liver transplantation across the ABO barrier. This could be the cause of the biliary stricture observed in reduced-size graft recipients previously discussed (patient 6 in Table 8). The incidence of biliary complications varied in several reports, from 0% to 28% in pediatric split-liver transplantation 14,17,19-21,28 and 10% to 38% in pediatric living related liver transplantation This seems to be related to the technical and anatomic problems associated with the procedure 33 and ischemic injury caused by prolonged cold preservation, 27 especially in the ex situ technique; hepatic artery thrombosis ; sepsis 33 ; and underlying pathological conditions. 32 Several measures have been proposed to reduce the incidence of biliary complications: reduce cold ischemia time, 34 extensively use bench cholangiography to identify anomalous biliary anatomy, 14,19 meticulous ligation of bile duct radicles on the cut surface, 14 routinely use T-tubes, 14 and avoid the dissection of the bile duct bifurcation. 17,19,29 We did not use bench cholangiography because it increases the benching time and needs manipulation of the graft. In our series, preservation time was limited, and we did not observe a difference in cold ischemia time between patients who developed and did not develop strictures of the bile ducts ( v minutes in the entire population, respectively; v minutes in the split-liver group, respectively). We recently started using an ultrasound dissector for parenchymal transection, and since then, we have not observed cut-edge biliary leaks (data not reported). Timely diagnosis and correction of biliary complications is crucial to avoid graft loss or patient mortality. Reichert et al 33 proposed routine planned exploration with open liver biopsy to avert morbidity caused by biliary complications. Our protocol was to closely survey the development of biliary leaks with Doppler ultrasound and perform cholangiography when biliary stricture was suspected, even without significant intrahepatic biliary dilatation. In this way, we were able to properly diagnose biliary complications and treat them without graft or patient loss. The length of hospital stay after transplantation for the patients with split livers was 24 days opposed to 15 days for whole grafts. The greater incidence of biliary complications in the former group does not explain this difference by itself. In the split-liver group, median hospitalization of children experiencing biliary complications was similar to those with no biliary problems (26 v 24 days; P.4). Urgent transplantation of high-risk patients is the dominant factor influencing poor outcome in our experience, as well as that of other investigators. 13,19,23 Nevertheless, split livers are used by us as well as

13 Split Liver for Pediatric Liver Transplantation 427 others 34 in urgent conditions because split-liver grafts provide function similar to whole or reduced-sized grafts. In our experience, all the patients referred to our center with acute or fulminant liver failure (n 4) underwent transplantation with a split liver. One patient previously described died with poor graft function. The other 3 patients showed a normal pattern of graft function recovery. Two patients are alive 1 and 3 months after transplantation with normal liver function; 1 patient died 9 months after transplantation of brain death with normal liver function. Only 1 patient among the adult recipients of a primary right split liver was a high-risk, UNOS status 1 patient. He is well 6 months after transplantation. The split-liver technique has the potential to completely meet the needs of pediatric liver transplantation, and its further technical evolution to generate 2 liver grafts with greater similarity in size could greatly contribute to meet the needs of adult patients. 8 One of the major benefits of living related liver transplantation is to perform transplantation on pediatric patients on an elective basis, early in the course of the original disease. 35,36 In our experience, split-liver transplantation matched this advantage, greatly reducing the recipient waiting time. In conclusion, our report shows that the adoption of a liberal policy of liver splitting provides allografts of optimal quality for pediatric transplantation. The in situ technique of split liver shortened the ischemia times with a low incidence of PNF and reexploration for intra-abdominal hemorrhage and facilitated the sharing of split grafts with other centers. The extensive use of split livers has dramatically decreased the pediatric waiting list time, allowing us to perform transplantation on all the children referred to our center, with no mortality on the waiting list. The extensive use of the alternate technique of in situ split-liver transplantation might help meet the needs of adult patients, further expanding the graft pool. Acknowledgment The authors thank the Nord Italia Transplant for its effort in coordinating the split-liver program and all the liver transplant centers that collaborated with this program. References 1. Bismuth H, Houssin D. Reduced-size orthotopic liver graft for liver transplantation in children. Surgery 1984;95: Emond JC, Whitington PF, Thistlethwaite RJ, Alonso En, Broelsch CE. Reduced-size orthotopic liver transplantation: Use in the management of children with chronic liver disease. Hepatology 1989;10: Broelsch CE, Emond JC, Thistlethwait RJ, Rouch DA, Whitington PF, Lichtor JL. Liver transplantation with reduced-size donor organs. Transplantation 1988;45: Houssin D, Soubrane O, Boillet O, Dousset B, Ozier Y, Devictor D, et al. Orthotopic liver transplantation with the reduced-size graft: An ideal compromise in pediatrics. Surgery 1992;111: Broelsch CE, Emond JC, Whitington PF, Thistlethwaite JR, Baker AL, Lichtor JL. Application of reduced-size liver transplant as split grafts, auxiliary orthotopic grafts, and living related segmental transplants. Ann Surg 1990;212: Centro Trasfusionale e di Immunologia dei Trapianti dell Ospedale Maggiore Policlinico di Milano. Nord Italia Transplant Report Milano: Centro Trasfusionale dell Ospedale Maggiore Policlinico di Milano, Rogiers X, Malagò M, Habib N, Knoefel WT, Pothmann W, Burdelski M, et al. In situ splitting of the liver in the heart-beating cadaveric organ donor for transplantation in two recipients. Transplantation 1995;59: Colledan M, Andorno E, Valente U, Gridelli B. A new splitting technique for liver grafts. Lancet 1999;353: Urata K, Kawasaki S, Matsunami H, Hashikura Y, Ikegami T, Ishizone S, et al. Calculation of child and adult standard liver volume for liver transplantation. Hepatology 1995;21: Deland FH, North WA. Relationship between liver size and body size. Radiology 1968;91: Emond JC, Heffron TG, Whitington PF, Broelsch CE. Reconstruction of the hepatic vein in reduced-size hepatic transplantation. Surg Gynecol Obstet 1993;176: Gridelli B. Programma Split Liver. Proceeding of the Riunione Tecnico Scientifica Nord Italia Transplant. Bergamo, Italy, November 11-12, In press. 13. de Ville de Goyet J. Split liver transplantation in Europe, Transplantation 1995;59: Rela M, Vougas V, Muiesan P, Vilca-Melendez H, Smyrniotis V, Gibbs P, et al. Split liver transplantation. King s College Hospital experience. Ann Surg 1998;227: Rogiers X, Malagò M, Gawad K, Jauch KW, Olausson M, Knoefel WT, et al. In situ splitting of cadaveric livers. The ultimate expansion of a limited donor pool. Ann Surg 1996;224: Goss JA, Yersiz H, Shackleton R, Seu P, Smith CV, Markowitz JS, et al. In situ splitting of the cadaveric liver for transplantation. Transplantation 1997;64: Emond JC, Whitington PF, Thistlethwaite JR, Cherqui D, Alonso EA, Woodle IS, et al. Transplantation of two patients with one liver. Ann Surg 1990;212: Otte JB, de Ville de Goyet J, Alberti A, Balladur P, de Hemptinne B. The concept and technique of the split liver in clinical transplantation. Surgery 1990;107: Houssin D, Boillot O, Soubrane O, Couinaud C, Pitre J, Ozier Y, et al. Controlled liver splitting for transplantation in two recipients: Technique, results and perspectives. Br J Surg 1993;80: Otte JB. Is it right to develop living related liver transplantation? Do reduced and split livers not suffice to cover the needs? Transpl Int 1995;8: Gawad KA, Rogiers X, Malagò M, Gundlach M, Knoefel WT, Izbiki JR, et al. Optimization of donor organ usage with the