Fifteen Years and 382 Extended Right Grafts From In Situ Split Livers in a Multicenter Study: Are These Still Extended Criteria Liver Grafts?

Transcription

1 LIVER TRANSPLANTATION 21: , 2015 ORIGINAL ARTICLE Fifteen Years and 382 Extended Right Grafts From In Situ Split Livers in a Multicenter Study: Are These Still Extended Criteria Liver Grafts? Umberto Maggi, 1,2 Tullia M. De Feo, 3 Enzo Andorno, 4 Umberto Cillo, 5 Luciano De Carlis, 6 Michele Colledan, 7 Patrizia Burra, 8 Nicola De Fazio, 3 and Giorgio Rossi, 9,10 on behalf of the Liver Transplantation and Intestine North Italy Transplant Study Group 1 UO Chirurgia Generale e Trapianti di Fegato, Fondazione IRCCS Ca Granda Ospedale Maggiore Policlinico di Milano, Milan, Italy; 2 Department of Digestive and Hepatobiliary Surgery, AP-HP, U.F.R. de M edecine de l Universit e Paris XII-Cr eteil, Paris, France; 3 North Italy Transplant Program, Immunologia dei Trapianti di Organi e Tessuti, Fondazione IRCCS Ca Granda Ospedale Maggiore Policlinico di Milano, Milan, Italy; 4 UOC Chirurgia Generale e Trapianti d Organo, Genoa, Italy; 5 S.S.D. di Chirurgia Epatobiliare e Trapianto Epatico Az. Osp. Di Padova, Padua, Italy; 6 Chirurgia Generale 2 e dei Trapianti dell A.O. Ospedale Niguarda Ca Granda, Milan, Italy; 7 Chirurgia III e Centro Trapianti di Fegato, Ospedali Riuniti di Bergamo, Bergamo, Italy; 8 Dipartimento di Scienze Chirurgiche e Gastroenterologiche, Universit a degli Studi, Padua, Italy; 9 General Surgery and Liver Transplantation Unit, Department of Surgical Sciences, University of Milan, Milan, Italy; and 10 Fondazione IRCCS Ca Granda, Ospedale Maggiore Policlinico, Milan, Italy In situ split liver extended right grafts (SL-ERGs) are still considered marginal grafts. Our aim was to verify this statement at the present time. From 1997 to 2011, a multicenter, retrospective study based on a prospective database was performed at 9 liver transplantation (LT) centers in northern Italy; it included 382 in situ SL-ERG transplants in adults. There were 358 primary LTs and 24 retransplantations (RETXs). The 1-, 3-, and 5-year overall graft survival rate for LT with in situ SL-ERGs were 73.5%, 63.3%, and 60.7%, respectively, from 1997 to 2004 and 83.5%, 80.3%, and 80.3%, respectively, thereafter (P ). A shorter total ischemia time and fewer RETX grafts were the main differences between the characteristics of Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CI, confidence interval; CIT, cold ischemia time; GGT, gamma-glutamyltransferase; GOT, glutamic oxaloacetic transaminase; GPT, glutamic pyruvic transaminase; HAT, hepatic artery thrombosis; HBV, hepatitis B virus; HCV, hepatitis C virus; HDV, hepatitis delta virus; HR, hazard ratio; ICU, intensive care unit; INR, international normalized ratio; LDLT, living donor liver transplantation; LL, laterolateral; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; NITp, North Italy Transplant program; OR, odds ratio; PNF, primary nonfunction; PVT, portal vein thrombosis; RETX, retransplantation; SD, standard deviation; SL-ERG, split liver extended right graft; SLT, split liver transplantation; TT, terminoterminal; UNOS, United Network for Organ Sharing; UW, University of Wisconsin solution; WG, whole graft. The Liver Transplantation and Intestine North Italy Transplant Study Group includes U. Baccarani (Department of Surgery and Transplantation, University Hospital, Udine, Italy), M. Donataccio (Prima Chirurgia Clinicizzata, Ospedale Civile Maggiore, Universita di Verona, Verona, Italy), A. Risaliti (Universita di Udine, Udine, Italy), E. Regalia (National Cancer Institute, Milan, Italy), M. Vivarelli (Chirurgia dei Trapianti di Fegato, Rene e Pancreas, Universita Politecnica delle Marche, Ospedali Riuniti, Ancona, Italy), and G. Fornoni (Centro Trapianti Fegato, Fondazione IRCCS Ca Granda, Ospedale Maggiore Policlinico di Milano, Milan, Italy). The author contributions were as follows: U. Maggi and G. Rossi designed the research; G. Rossi, U. Maggi, E. Andorno, U. Cillo, M. Colledan, L. De Carlis, U. Baccarani, M. Donataccio, A. Risaliti, E. Regalia, and M. Vivarelli contributed to the liver transplantation surgery; N. De Fazio, T. M. De Feo, and U. Maggi contributed to the statistics and epidemiology; P. Burra and T. M. De Feo contributed to the clinical work; U. Maggi wrote the article; and T. M. De Feo, E. Andorno, and G. Rossi reviewed the article. Potential conflict of interest: Nothing to report. Address reprint requests to Umberto Maggi, M.D., Ph.D., U.O. Chirurgia Generale e Trapianti di Fegato, Fondazione IRCCS Ca Granda Ospedale Maggiore Policlinico di Milano, Via F. Sforza 35, 20122, Milan, Italy. Telephone: ; FAX: ; maggi.umberto@gmail.com DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2015 American Association for the Study of Liver Diseases.

2 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 MAGGI ET AL. 501 the 2 periods. From 1997 to 2011, the 1-, 3-, and 5-year graft survival rates showed a significant difference between the 358 primary LT in situ SL-ERGs and the 24 RETX in situ SL-ERGs (P < 0.001). In a multivariate analysis, the main prognostic factor for 60-day graft survival was a total ischemia time < 8 hours for the 358 primary in situ SL-ERGs. From 2005 to 2011, in 2473 LTs, the 5-year graft survival for 184 in situ SL-ERGs and 2289 whole grafts was 75% and 80% (P 5 0.3), respectively. Univariate and multivariate studies alike failed to indicate that the type of graft was a prognostic factor for graft survival. A donor age > 60 years, RETX grafts, and urgency were the main prognostic factors for failure for all of the grafts. Although caution should be taken regarding the choice of appropriate donors, in situ SL-ERGs should no longer be considered marginal grafts for experienced LT centers. SL-ERGs should not be used in RETX settings, and when SL-ERGs are used as primary grafts, the total ischemia time should be less than 8 hours. Liver Transpl 21: , VC 2015 AASLD. Received July 29, 2014; accepted December 14, One of the main problems in liver transplantation (LT) is the discrepancy between the number of donors and the number of recipients. Efforts have been made to find new surgical techniques that allow the use of a single liver for 2 recipients. After the initial seminal concepts on segmental LT were disseminated in 1969 by Smith, 1 Pichlmayr et al. 2 were the first to report a procedure that split a liver between an adult and a child in Additionally, in 1989, Bismuth et al. 3 split a liver between 2 adults in an emergency setting. The technique of liver splitting for 2 adult recipients quickly appeared to be extremely surgically demanding and produced results that were inferior to those currently expected of living donor transplantation. 4 Still, improving this technique could theoretically expand the adult donor liver pool. Splitting a liver between an adult and a child the conventional splitting technique began to produce better results after its rough start in the 1990s. 5,6 In 1996, an in situ splitting technique 7 for an adult recipient and a child was an early and important step toward a safer procedure. The split liver technique was then able to shorten the pediatric waiting list for LT. 8,9 However, concerns remained about the use of split liver extended right grafts (SL-ERGs) for adults with respect to their lower graft survival in comparison with grafts from whole-organ donations after circulatory death. 10 In large studies, such graft approaches are often considered marginal 11 because of their association with graft failure and their inclusion in the donor risk index; recently, some authors 10 have reported unsatisfying results in terms of graft survival. Others, however, have obtained quite good results. 12 The main aim of this study was to analyze a multicenter experience with in situ split liver transplantation (SLT) over 15 years. The secondary aim was to identify risk factors for early graft survival. Finally, we TABLE 1. Continuous Variables Related to the Donors and Recipients of 358 Primary Transplants and 24 RETXs With SL-ERGs Variables n Minimum Maximum Mean 6 SD Median Donor Age (years) Weight (kg) Height (cm) BMI (kg/m 2 ) ICU days Sodium (meq/l) GGT (IU/L) Recipient Age (years) Weight (kg) Height (cm) BMI (kg/m 2 ) Creatinine (mg/dl) Bilirubin (mg/dl) INR GGT (IU/L) AST/GOT ALT/GPT Albumin (g/dl) MELD score Ischemia time (minutes) Follow-up time (years)

3 502 MAGGI ET AL. LIVER TRANSPLANTATION, April 2015 hoped to evaluate whether split livers should still be defined as marginal grafts by comparing their results with those of whole grafts (WGs). PATIENTS AND METHODS A specific multicenter SLT program was established in 1997 to include the North Italy Transplant program (NITp). The NITp is the main organization for organ retrieval in northern Italy. The donor eligibility criteria for the split procedure included an age less than 60 years, intensive care unit (ICU) stays shorter than 5 days, low inotropic support (dopamine 5 mg/kg/ minute, dobutamine 10 mg/kg/minute, and no epinephrine or norepinephrine), and near-normal liver function tests. 13 Livers derived from split procedures were allocated according to the same clinical and immunological criteria used for the whole livers. The status of the urgency of liver replacement for recipients was classified according to the United Network for Organ Sharing (UNOS) liver status classification (initially implemented in July 1997 and later modified in January 1998 and August 1998; see After the in situ conventional split procedure is adopted, 14 with the separation of the left lateral liver graft (segments 2 and 3) from the right graft (segments 1 and 4 to 8), the policy among NITp centers is to generally maintain the left hepatic artery with the left graft in continuity with the celiac axis. Therefore, the right hepatic artery and the right and main portal veins are retained with the SL-ERG. From January 1997 through December 2011, 9 LT centers participated in the NITp, and this resulted in 382 SL-ERGs being retrieved and transplanted into adult recipients as primary or retransplantation (RETX) grafts. In the 9 centers, the mean number of SL-ERGs per year was The study was presented to the ethics committees of the institutions at which the data were collected (specifically the Comitato Etico della Fondazione IRCCS Ca Granda Ospedale Maggiore Policlinico), but according to Italian law (Gazzetta Ufficiale, serie generale, number 76, page 67-74), no specific requests or patient approval is needed for retrospective studies. In this retrospective study, continuous and categorical data were collected from a prospective database of information on the donors, surgical procedures, and recipients. Missing data were not included in the analyses. For each SL-ERG, the data that were collected included the donor s age, sex, weight, height, and body mass index (BMI); days in the ICU; ABO group; perfusion solution; and plasma sodium and gamma-glutamyltransferase (GGT). Also collected were the recipient s age, sex, weight, height, and BMI; ABO group; plasma creatinine, bilirubin, international normalized ratio (INR), GGT, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and albumin; Model for End-Stage Liver Disease (MELD) at transplant; liver disease and virus-related disease; Child score; UNOS class; total ischemia time; type of anastomosis (caval, arterial, and biliary); immunosuppression; follow-up time; surgical complications after LT [biliary, arterial, and primary TABLE 2. Categorical Variables Related to the Donors and Recipients of 358 Primary Transplants and 24 RETXs With SL-ERGs Variables Frequency % Donor ABO group O A B AB Donor Sex Male Female Perfusion solution UW Celsior Recipient ABO group O A B AB Sex Male Female Indication Cirrhosis Cholestatic Metabolic Acute Tumor Other RETXs Indication restricted Cirrhosis Tumor Others Virus-related cirrhosis as indication (percentages calculated only inside this category) HCV HBV HBV 1 HDV HBV 1 HCV HBV 1 HCV 1 HDV Child score (valid percent) A B C UNOS class a b Urgency UNOS 1-2a UNOS 2b nonfunction (PNF)]; survival; causes of death or graft loss; and RETX. All pretransplant data were intended to be recorded at the time of the LT.

4 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 MAGGI ET AL. 503 TABLE 3. Outcome Data in 2 Different Periods for 358 Primary Transplants and 24 RETXs With SL-ERGs Outcome P PNF No 186 (93.9%) 184 (100%) 370 (96.9%) <0.001 Yes 12 (6.1%) 0 (0%) 12 (3.1%) Biliary complications No 160 (80.7%) 155 (84.2%) 315 (82.4%) Yes 38 (19.3%) 29 (15.8%) 67 (17.6%) 0.36 HAT No 186 (93.9%) 178 (96.7%) 364 (95.3%) Yes 12 (6.1%) 6 (3.3%) 18 (4.7%) 0.24 Short-term graft survival 60 days 40 (20.2%) 19 (10.3%) 59 (15.4%) > 60 days 158 (79.8%) 165 (89.7%) 323 (84.6%) Short-term graft survival (primary grafts) 60 days 27 (15%) 15 (8.4%) 42 (11.7%) > 60 days 153 (85%) 163 (91.6%) 316 (88.3%) 0.05 Short-term graft survival (RETX grafts) 60 days 13 (72.2%) 4 (66.7%) 17 (70.8%) > 60 days 5 (27.8%) 2 (33.3%) 7 (29.2%) , 3-, and 5-year graft survival (overall) 73.5% 83.5% 78.4% 63.3% 80.3% 71.1% < % 80.3% 68.9% 1-, 3-, and 5-year graft survival (358 primary grafts) 79.8% 85.2% 82.5% 68.5% 82.5% 75.1% % 82.5% 73.2% 1-, 3-, and 5-year graft survival (24 RETX grafts) 11.1% 33.3% 16.7% 11.1% No data 12.5% % No data* 12.5% Complications: RETX No 167 (84.3%) 174 (94.6%) 341 (89.3%) Yes 31 (15.7%) 10 (5.4%) 41 (10.7%) *The total number of grafts was 6; 5 grafts were lost within 2 years, and for 1 graft, the Follow-up time was not sufficient. TABLE 4. Causes of Death (n 5 20) in RETX SL-ERG Figure 1. Long-term SL-ERG survival in 358 primary LT cases ( no RETX pre ) and 24 RETX cases from 1997 to 2011 (P < 0.001; 82.6%, 75.3%, and 73.4% for primary SL-ERGs versus 16.7%, 12.5%, and 12.5% for RETX grafts). Cause of Death n n % Sepsis With good liver function 2 Biliary 2 Pulmonary 2 Pulmonary fungal 1 Fungal diffuse 2 Fungal arterial 1 (with rupture) Diffuse 1 Graft malfunction 2 10 (and sepsis) PNF and delayed 4 20 graft malfunction Other* 1 5 HAT 1 5 Tumoral recurrence 1 5 Total *Bronchial hemorrhage.

5 504 MAGGI ET AL. LIVER TRANSPLANTATION, April 2015 TABLE 5. Causes of Graft Loss for 39 Primary SL-ERG Undergoing RETX Cause of RETX n % HAT PNF Recurrence Other 4* 10.2 PVT Chronic rejection Unknown Total *Two hepatic artery ruptures, 1 veno-occlusive syndrome, and 1 hepatic abscess. Figure 2. Long-term RETX survival with 335 WGs and 24 SL- ERGs from 1997 to 2011 (P < 0.001; 60.8%, 54.4%, and 51.2% for WGs versus 16.7%, 12.5%, and 12.5% for SL-ERGs). Data were examined from 2 different periods ( and ), and we were, therefore, able to focus on the differences that occurred over time for SL-ERGs among the clinical and demographic characteristics of patients, indications, PNF, biliary and arterial complications, RETX, and outcomes. The survival of the grafts and the survival of patients undergoing primary transplantation and RETX with SL-ERGs were calculated with the Kaplan- Meier test. Outcome data, including long-term graft survival and surgical complications, were collected and calculated between the 2 groups. Because our main interest was to find a correlation between 60-day graft survival and patient survival with primary SL-ERGs, Considering the characteristics of the donors, surgeries, and recipients we performed univariate analyses with a t test for the continuous data and a chi-squared test for the categorical. Differences with P < 0.05 were considered significant. Those factors that emerged as significant through the univariate analysis were entered into the multivariate logistic regression analysis. Finally, from 2005 to 2011, all cadaveric SL-ERGs and WGs that were used for LT in the NITp area were analyzed, and their graft survivals were calculated. The resultant survival data for several variables were compared with the Kaplan-Meier test. Variables showing a significant association upon univariate analysis (P < 0.05) were included in the multivariate analysis according to the Cox proportional hazards model; first we included RETX SL-ERGs, and then we excluded them to identify any prognostic factors. RESULTS All Split Grafts and 2 Periods In all, 382 right trisegments, including 358 primary LTs and 24 RETXs, underwent transplantation. The donor, surgical, and recipient characteristics are reported in Tables 1 and 2. In RETX, the mean time interval from the previous LT was days (range days); the majority (79%) of retransplants were performed within 90 days of the previous LT. According to graft survival, no center effect was found; no statistical difference in the graft survival curves between all centers was observed (P ). A total of 198 split grafts, including 180 primary LT procedures and 18 RETX procedures, were transplanted from 1997 to 2004 (51.8% of all LT procedures considered), whereas a total of 184 split grafts, including 178 LT procedures and 6 RETX procedures, were transplanted from 2005 to 2011 (48.2%). The differences in the characteristics between the 2 periods were compared. A longer donor ICU stay (P ), a shorter total ischemia time (P ), a higher recipient age (P ), higher percentages of cirrhosis and tumors with fewer RETX grafts (P ), a smaller percentage of UNOS class 1 patients and a higher percentage of UNOS class 2b patients (P ), and a lower percentage of interpositional arterial grafts (P < 0.001) were found in the period from 2005 to In the postoperative period (Table 3), the rates of biliary and hepatic artery thrombosis (HAT) complications did not differ statistically between the 2 periods. The 1-, 3-, and 5-year graft survival rates for the 382 grafts between 1997 and 2011 were 78.4%, 71.1%, and 69.3%, respectively. The 1-, 3-, and 5- year overall graft survival rates were 73.5%, 63.3%, and 60.7%, respectively, from 1997 to 2004 and 83.5%, 80.3%, and 80.3%, respectively, thereafter (P < 0.001). Primary and RETX Grafts The 1-, 3-, and 5-year graft survival rates for the 382 SL-ERGs performed from 1997 to 2011 showed a significant difference between the 358 primary SL-ERGs and the 24 RETX SL-ERGs (P < 0.001; 82.6%, 75.3%, and 73.4% for primary SL-ERGs versus 16.7%,

6 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 MAGGI ET AL. 505 TABLE 6. Prognostic Factors for 60-Day Graft Survival in 358 Primary SL-ERGs: Univariate and Multivariate Analyses Survival Survival Variables 60 Days > 60 Days Univariate P Multivariate P OR (95% CI) Donor Age (years) < > Sex Male Female BMI (kg/m 2 ) > ICU days > Sodium > GGT > ABO Group O A B 4 24 AB 0 1 Perfusion solution Celsior UW Ischemia time 8 hours >8 hours Recipient Age (years) > Sex Male Female Primary disease restricted Cirrhosis Cancer Others 6 40 UNOS Status a b UNOS status/urgency UNOS 1-2a UNOS 2b Child A 3 39 B C MELD >

7 506 MAGGI ET AL. LIVER TRANSPLANTATION, April 2015 Survival TABLE 6. Continued Survival Variables 60 Days > 60 Days Univariate P Multivariate P OR (95% CI) Period Caval anastomosis TT 2 38 LL (Belghiti) Piggyback Arterial anastomosis Direct Interposition graft Biliary anastomosis Hepatojejunal 2 17 Duct-to-duct Initial Immunosuppression Cyclosporine Tacrolimus Other %, and 12.5% for RETX grafts; Fig. 1). The causes of 20 graft losses resulting in patient deaths with RETX grafts are reported in Table 4. Among the 358 primary SL-ERGs, the 1-, 3-, and 5- year survival rates were 80.0%, 68.9%, and 66.1%, respectively, for the 180 grafts in the first period and 85.2%, 82.5%, and 82.5%, respectively, for the 178 grafts in the second period (P ). There was no significant difference in the survival rates for RETX SL-ERGs between the 2 periods. We also performed a comparison of results for RETXs performed from 1997 to 2011 in the NITp area with 334 WGs and 24 SL-ERGs: 1-, 3-, and 5-year graft survival rates showed a statistically significant difference (P < 0.001; 60.8%, 54.4%, and 51.2%, respectively, for WGs versus 16.7%, 12.5%, and 12.5%, respectively, for SL-ERGs; Fig. 2). The patient s 1-, 3-, and 5-year survival rates were 63.5%, 56.8%, and 53.5%, respectively, for WGs and 20.8%, 16.7%, and 16.7%, respectively, for SL-ERGs (P < 0.001). Risk Factors for 60-Day Graft Survival for 358 Primary SL-ERGs Thirty-nine grafts (10.9%) were lost with subsequent RETXs; 23 (59%) were within 60 days of LT, and 16 (4.1%) occurred thereafter. The indications for RETX are reported in Table 5. There was a higher number of cases with HAT (n 5 16 or 41%) and PNF (n 5 8 or 20.5%). However, the rate of RETX decreased from the first to the second period, specifically from 15.7% to 5.4% (P ). Sixty-six patients (18.4%) died, and their grafts were lost without any RETX; 20 (30.3%) were within 60 days of LT, and 46 (69.7%) died thereafter. The causes of death were mainly nontumoral recurrence (n 5 14 or 21.2%) and sepsis (n 5 12 or 18.2.%). Sepsis (n 5 6 or 30%) was again the main cause of death from LT within 60 days, although thereafter, nontumoral disease recurrence (n 5 14 or 30.4%) became more prominent. Risk factors possibly related to the 60-day graft survival rates of the primary SL-ERGs were calculated, and those considered to be important were included in a logistic regression analysis, the results of which are reported in Table 6. In the multivariate analysis, only a total ischemic time > 8 hours (versus 8 hours) maintained a statistical significance. Split Grafts Versus WGs We compared the graft survival of SL-ERGs with that of WGs transplanted into adult recipients in the period from 2005 to 2011 and in the NITp area. During that period of time, LT was performed 2473 times; this included 184 SL-ERGs and 2289 WGs. Data concerning the 2 groups are reported in Table 7. In the univariate analysis of the factors influencing long-term survival, several factors were identified to have prognostic value, but the type of graft was not among them. When the significant factors from the univariate analysis were included in the Cox model multivariate analysis (Table 8), a donor age > 60 years, primary LT versus RETX, and urgency versus no urgency were identified as statistically significant prognostic factors. The 5-year graft survival rate was 75.3% for the 2289 WGs and 79.6% for the 184 SL-ERGs (P ; Fig. 3). In response to the hypothesis that the number of RETX grafts would influence the results of graft survival, a further study was performed that compared the same previous transplants but excluded from the analysis any grafts employed as RETX grafts. A total of 2115 WGs and 178 SL- ERGs were included, and 5-year graft survival rates of

8 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 MAGGI ET AL. 507 TABLE 7. Continuous and Categorical Data Concerning 2293 Primary LT Cases and 180 RETX Cases Performed in the NITp Area From 2005 to 2011 With Whole and SL-ER Grafts WGs n SL-ERG n All n Variables (mean 6 SD) (mean 6 SD) (mean 6 SD) P Value Donor Age (years) <0.001 Weight (kg) Height (cm) <0.001 BMI (kg/m 2 ) <0.001 ICU days Sodium (meq/l) GGT (IU/L) ABO Group O 1013 (44%) 110 (59.8%) 1123 A 949 (41%) 54 (29.3%) 1003 <0.001 B 230 (10%) 20 (10.9%) 250 AB 97 (4%) 0 (0%) 97 Sex Male 1214 (53%) 122 (663%) Female 1075 (47%) 62 (33.7%) 1137 Recipient Age (years) Weight (kg) <0.001 Height (cm) <0.001 BMI (kg/m 2 ) <0.001 Serum creatinine (mg/dl) Hepatic Disease Acute 64 (3%) 2 (1.0%) 66 Cirrhosis 1039 (45%) 107 (56.9%) 1146 Cholestatic 40 (2%) 8 (4.2%) 48 Metabolic 30 (1%) 4 (2.1%) 34 <0.001 Tumor 870 (38%) 49 (26.0%) 919 Other 72 (3%) 8 (4.2%) 80 RETX 174 (8%) 6 (3.1%) 180 Serum bilirubin (mg/dl) INR GOT/AST (IU/L) GPT/AAT (IU/L) Albumin (g/dl) MELD score Total ischemia (hours) ABO group O 926 (40%) 107 (58.1%) 1033 A 970 (42%) 55 (29.8%) 1025 <0.001 B 270 (12%) 21 (11.4%) 291 AB 123 (5%) 1 (0.5%) 124 Sex Male 1790 (78%) 98 (53.2%) 1888 <0.001 Female 499 (22%) 86 (46.7%) 585 UNOS (6%) 6 (3.2%) 145 2a 323 (14%) 19 (10.3%) b 1479 (65%) 125 (67.9%) (15%) 34 (18.4%) 382 RETX graft No 2115 (92%) 178 (96.7%) Yes 174 (8%) 6 (3.2%) 180 Follow-up time (years) <0.001 Follow-up time (years) median <0.001

9 508 MAGGI ET AL. LIVER TRANSPLANTATION, April 2015 TABLE 8. Cox Model Multivariate Analysis, Related to Long-Term Survival, Considering All LT and RETX Cases Performed in the NITp Area From 2005 to 2011 Variables HR P Value 95% CI SL-ERG versus WG Donor age > 60 versus 60 years Donor sodium > 150 versus 150 meq/l ABO compatible versus ABO identity Recipient age > 60 versus 60 years Urgency versus no urgency RETX versus primary LT Ischemia time > 8 versus 8 hours MELD score 15 versus < Recipient creatinine 1.2 versus < 1.2 mg/dl Recipient bilirubin 1.2 versus < 1.2 mg/dl Recipient INR >1.3 versus < Recipient GOT >40 versus 40 IU/L Recipient GPT >40 versus 40 IU/L Recipient GGT >45 versus 45 IU/L Tumor versus cirrhosis Other versus cirrhosis Figure 3. Long-term graft survival of 2289 WGs and 184 SL- ERGs, including RETX grafts, transplanted in into adult recipients (P ). 81.8% for SL-ERGs and 76.5% for WGs (P ) were found (Fig. 4). When the significant factors from the univariate analysis were included in the Cox model multivariate analysis, a donor age > 60 years, urgency versus no urgency, and diseases other than cirrhosis or tumors were found to be important prognostic factors (Table 9). DISCUSSION The history of pediatric LT reached a turning point in 1984 when Bismuth and Houssin 15 severed an adult liver and transplanted the left reduced size liver into a Figure 4. Long-term graft survival of 2115 WGs and 178 SL- ERGs, excluding RETX grafts, transplanted in into adult recipients (P ). child. The right graft was then discarded. This procedure was modified by Pichlmayr et al. 2 in 1988 to include ex situ splitting, and this resulted in 2 feasible portions of the liver that could be transplanted into 1 adult and 1 pediatric recipient. A further improvement was introduced by Rogiers et al. 7 in 1996 because the in situ technique was shown to be superior to the ex situ technique. 8 The surgical technique of splitting a liver along the falciform ligament, which results in a left lateral segment and an extended right graft, has the beneficial result of both decreasing the pediatric waiting list and lowering the pretransplant mortality of children. 8 A recent report 9 from the United States showed that the

10 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 MAGGI ET AL. 509 TABLE 9. Cox Model Multivariate Analysis of Long-Term Survival Considering Only Primary LT Cases Performed in the NITp Area From 2005 to 2011 Variables HR P Value 95% CI SL-ERG versus WG Donor age > 60 versus 60 years Donor sodium > 150 versus 150 meq/l ABO compatible versus ABO id Recipient age > 60 versus 60 years Urgency versus no urgency Ischemia time > 8 versus 8 hours MELD score 15 versus < Recipient creatinine 1.2 versus < 1.2 mg/dl Recipient bilirubin 1.2 versus < 1.2 mg/dl Recipient INR >1.3 versus Recipient GOT >40 versus 40 IU/L Recipient GPT >40 versus 40 IU/L Tumor versus cirrhosis Other versus cirrhosis pediatric waiting list steadily declined between 2008 and The same was true for the pretransplant mortality of candidates who were put on the wait list for a liver-alone transplant, which decreased from 14.3% deaths per 100 wait-list years in to 6.2% in Meanwhile, over the same time period, the amount of pediatric patients who received a whole liver increased from 60.5% to 63.6%, whereas those receiving split livers increased from 12.7% to 16.8%. 9 The results of right grafts are conflicting, although good results have been reported in monocentric studies. 12,16 Hong et al. 16 reported no significant differences in 10-year graft survival among SLT, living donor liver transplantation (LDLT), and WG LT. Some authors 4 have reported better long-term results from split livers versus living donor liver grafts, whereas others, in a series mixed with ex situ grafts, reported 1-, 3-, and 5-year graft survival rates of 84%, 80%, and 71%, respectively. 17 Other data, which were retrieved from 1994 to 2001 from the UNOS scientific registry, suggested that the procedure altered otherwise optimal whole organs (as they were often harvested from young donors) into marginal segmental grafts, which generated worse results than WGs. 18 Some authors observed that there were no long-term results concerning split livers, 16 and thus, a type of uncertainty lingers around all results from conventional split liver techniques; this limits their spread. Otherwise, some authors have made efforts to perform LT with split liver grafts, 19 and recent studies have shown that the risk of failure is actually similar between split liver grafts and WGs. However, this study considered split grafts to be unique and did not specify right in situ split grafts. 20 Comparing our data from and , we observed several statistically significant differences in the more recent period, including a longer donor ICU stay, a more advanced recipient age, a shorter total ischemia time, a higher rate of direct arterial anastomosis, fewer RETX grafts, a smaller percentage of UNOS class 1 patients, and a higher percentage of UNOS class 2b patients. So, despite a more liberal attitude toward the donors and recipients, the use of split grafts in non-retx settings and a shorter total ischemia time seem to favorably influence the recent results: the short-term and 1-, 3-, and 5-year graft survival for both all grafts and primary grafts alone demonstrated improvements. No differences occurred for SL-ERGs used in the RETX setting because bad outcomes were seen in both periods. In our series, it soon became evident that SL-ERGs applied to RETX had a dramatically worse outcome. The 60-day graft survival was 29.2% for RETX grafts and 88.3% for primary grafts (P < 0.001). Moreover, the 1-, 3-, and 5-year graft survival rates were 16.7%, 12.5%, and 12.5%, respectively, for RETX grafts and 82.6%, 75.3%, and 73.4%, respectively, for primary grafts (P < 0.001). On the other hand, when we studied the effect of UNOS classes in primary SL-ERGs and compared 1-2a recipients to 2b-3 recipients, no differences were found. The poor results of SLT in RETX suggest that such situations should probably not rely on these grafts; although primary split grafts are effective in UNOS emergency settings, it can be argued that other factors related to the recipient, such as possessing the appropriate indications or sepsis, are just as important. However, the comparison of specific results of RETX performed from 1997 to 2011 in the NITp area with 334 WGs and 24 SL-ERGs, showing worse 1-, 3-, and 5-year graft survival with SL-ERGs, suggests the importance of the kind of graft in this setting. Another aim of the study was to analyze the factors related to 60-day graft survival in primary SL-ERGs from 2005 to Some articles have highlighted factors related to the outcomes of split grafts (Table 10). Cardillo et al. 13 found that a donor

11 510 MAGGI ET AL. LIVER TRANSPLANTATION, April 2015 TABLE 10. Results of SL-ERG (Most In Situ) in Recent Series % Biliary Complications % HAT % PNF % RETX Factors Influencing Author SL-ERGs (n) Years Outcome Graft Survival Doyle et al. 12 (2013) 18 in situ % (1 years) 89% (5 years) 89% (10 years) Hong et al. 23 (2011) Review MELD > 30; CIT > 10 hours; low volume centers; 20 0 (%) Vagefi 24 (2011) 20 in situ % (1 years) 75% (5 years) 50% (10 years) Hong et al. 16 (2009) 72 in situ RETX; donor age > 45 years; Rage> 60 years; CIT > 10 hours; Sandroussi 25 (2009) 43 (mostly in situ) None 88.4% (3 years) Cardillo et al. 13 (2006) 154 SL-ERG in situ % (3 years) Wilms 26 (2006) 70 SL-ERG in situ % (5 years) 41 0 Corno 27 (2006) 22 in situ % (1 year) 94% (5 years) Spada 28 (2005) 15 in situ 93% (1 year) 93% (3 years) Renz et al. 18 (2004) 152 (in and ex situ) age > 60 years, an ischemia time > 7 hours, a recipient of UNOS class 1 to 2a, and low-volume centers negatively influenced the outcomes of in situ SL-ERGs. Patients with fulminant hepatic failure seemed to represent an exception to the urgent class of patients. 21,22 Renz et al. 18 referred to the high mortality rates (24% and 26%) in UNOS class 1 and 2a recipients. Hong et al. 16,23 identified the following risk factors: recipient MELD score > 30, age > 60 years, donor age > 45 years, cold ischemia time (CIT) > 10 hours, and low-volume centers. In our series, only an ischemia time 8 or> 8 hours maintained a statistical significance after the multivariate analysis [P , odds ratio (OR) , 95% confidence interval (CI) ; Table 6]. The last aim of the study was a comparison of the outcomes of split grafts and WGs. Recently, some authors have found lower survival rates with SL-ERGs versus LDLT or WGs. 10 We analyzed a large number of LT procedures (n ) performed from 2005 to 2011 in northern Italy, which included 184 in situ SL-ERGs and 2289 WGs, and we included and then excluded RETX grafts. The 2 groups were homogeneous, except for the younger age and the thinner constitution of donors and recipients of SL-ERGs. Notably, the MELD score did not differ between recipients of WGs and recipients of in situ SL-ERGs. However, there was a higher rate of WGs than SL-ERGs employed for RETX (8% versus 3%, P ). For transplants performed from 2005 to 2011, the Kaplan-Meier test for graft survival showed no statistical disadvantage for in situ SL-ERGs, regardless of whether RETX grafts were considered. This result differed from those of our previously unpublished data taken from 1997 to 2004, which showed a 5-year survival rate for 2062 grafts of 71% for the 1864 WGs and 61% for the 198 SL-ERGs (P ). In the Cox model for multivariate analysis for long-term survival, several known factors, including a donor age > 60 years, primary LT versus RETX, and urgency versus no urgency, were identified as prognostic factors if RETX grafts were included in the analysis; if they were not, then age and urgency still maintained their importance, but also the prevalence of diseases other than cirrhosis and tumors emerged as important. However, the type of graft still was not found to have prognostic value. Still, using some caution in the choice of appropriate donors for situ SL-ERGs could actually be considered just as safe as WGs. As a last remark, on the basis of our study, we cannot conclude that similar results would apply for back-bench ex situ (in the cold) splitting. This is an important distinction because at many US centers, donor recoveries are more often performed by junior surgeons or trainees who do not have the required expertise to perform in situ splitting. However, whatever the technique, the experience/past learning curve and the volume of the center make the greatest difference. This study has some limitations. The first is missing data for such variables as the MELD score because it

12 LIVER TRANSPLANTATION, Vol. 21, No. 4, 2015 MAGGI ET AL. 511 was introduced into the panel of data only in A second limitation is the absence of data concerning the weight of the right trisegments and the graft-torecipient weight ratio. However, this variable is less important for the right trisegments than for other types of split grafts because right trisegments very often fit well with the weight of recipients. In conclusion, this multicenter study focuses specifically on results of in situ SL-ERGs. The results of 382 LTs with in situ SL-ERGs performed at 9 Italian centers in the last 15 years improved, though a more liberal policy for donors and recipients. They should not be used in RETX. The multivariate analysis of prognostic factors for 60-day graft survival in the primary SL-ERGs revealed the importance of ensuring an ischemic time less than 8 hours. The use of in situ SL-ERGs in adult recipients obtained survival rates equal to the those with WGs transplanted in corresponding periods into adults. For all LTs, the type of graft failed to qualify as a prognostic factor for survival. A donor age > 60 years and urgency were negative prognostic factors for long-term survival. On the basis of this study, with the use of appropriate donors, in situ SL-ERGs should not be considered today marginal grafts at an experienced LT center. REFERENCES 1. Smith B. Segmental liver transplantation from a living donor. J Pediatr Surg 1969;4: Pichlmayr R, Ringe B, Gubernatis G, Hauss J, Bunzendahl H. Transplantation of a donor liver to 2 recipients (splitting transplantation) a new method in the further development of segmental liver transplantation [in German]. Langenbecks Arch Chir 1988;373: Bismuth H, Morino M, Castaing D, Gillon MC, Descorps Declere A, Saliba F, Samuel D. Emergency orthotopic liver transplantation in two patients using one donor liver. Br J Surg 1989;76: Lee WC, Lee CS, Soong RS, Lee CF, Wu TJ, Chou HS, Chan KM. Split liver transplantation in adults: preoperative estimation of the weight of right and left hemiliver grafts. Liver Transpl 2011;17: de Ville de Goyet J. Split liver transplantation in Europe 1988 to 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, Malago 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: Kim JS, Broering DC, Tustas RY, Fischer L, Ganschow R, Burdelski M, Rogiers X. Split liver transplantation: past, present and future. Pediatr Transplant 2004;8: Kim WR, Stock PG, Smith JM, Heimbach JK, Skeans MA, Edwards EB, et al. OPTN/SRTR 2011 annual data report: liver. Am J Transplant 2013;13(suppl 1): Mallik M, Callaghan CJ, Hope M, Gibbs P, Davies S, Gimson AE, et al. Comparison of liver transplantation outcomes from adult split liver and circulatory death donors. Br J Surg 2012;99: Feng S, Goodrich NP, Bragg-Gresham JL, Dykstra DM, Punch JD, DebRoy MA, et al. Characteristics associated with liver graft failure: the concept of a donor risk index. Am J Transplant 2006;6: Doyle MB, Maynard E, Lin Y, Vachharajani N, Shenoy S, Anderson C, et al. Outcomes with split liver transplantation are equivalent to those with whole organ transplantation. J Am Coll Surg 2013;217: Cardillo M, De Fazio N, Pedotti P, De Feo T, Fassati LR, Mazzaferro V, et al.; for NITp Liver Transplantation Working Group. Split and whole liver transplantation outcomes: a comparative cohort study. Liver Transpl 2006; 12: Goss JA, Yersiz H, Shackleton CR, Seu P, Smith CV, Markowitz JS, et al. In situ splitting of the cadaveric liver for transplantation. Transplantation 1997;64: Bismuth H, Houssin D. Reduced-sized orthotopic liver graft in hepatic transplantation in children. Surgery 1984;95: Hong JC, Yersiz H, Farmer DG, Duffy JP, Ghobrial RM, Nonthasoot B, et al. Longterm outcomes for whole and segmental liver grafts in adult and pediatric liver transplant recipients: a 10-year comparative analysis of 2,988 cases. J Am Coll Surg 2009;208: Vigano L, Laurent A, Tayar C, Merle JC, Lauzet JY, Hurtova M, et al. Outcomes in adult recipients of rightsided liver grafts in split-liver procedures. HPB (Oxford) 2010;12: Renz JF, Emond JC, Yersiz H, Ascher NL, Busuttil RW. Split-liver transplantation in the United States: outcomes of a national survey. Ann Surg 2004;239: Emre S, Umman V. Split liver transplantation: an overview. Transplant Proc 2011;43: Cauley RP, Vakili K, Fullington N, Potanos K, Graham DA, Finkelstein JA, Kim HB. Deceased-donor split-liver transplantation in adult recipients: is the learning curve over? J Am Coll Surg 2013;217: Emond JC, Freeman RB Jr., Renz JF, Yersiz H, Rogiers X, Busuttil RW. Optimizing the use of donated cadaver livers: analysis and policy development to increase the application of split-liver transplantation. Liver Transpl 2002;8: Saidi RF, Jabbour N, Li Y, Shah SA, Bozorgzadeh A. Outcomes in partial liver transplantation: deceased donor split-liver vs. live donor liver transplantation. HPB (Oxford) 2011;13: Hong JC, Yersiz H, Busuttil RW. Where are we today in split liver transplantation? Curr Opin Organ Transplant 2011;16: