In-Hospital Mortality in Adult Recipients of Living Donor Liver Transplantation: Experience of 576 Consecutive Cases at a Single Center

Transcription

1 LIVER TRANSPLANTATION 15: , 2009 ORIGINAL ARTICLE In-Hospital Mortality in Adult Recipients of Living Donor Liver Transplantation: Experience of 576 Consecutive Cases at a Single Center Toshimi Kaido, Hiroto Egawa, Hiroaki Tsuji, 2 Eishi Ashihara, 2 Taira Maekawa, 2 and Shinji Uemoto 1 Department of Hepato-Biliary-Pancreatic and Transplant Surgery, Kyoto University School of Medicine, Kyoto, Japan; and 2 Department of Blood Transfusion and Immunology, Kyoto University Hospital, Kyoto, Japan Adult living donor liver transplantation (LDLT) was developed against the background of a scarcity of deceased donors and has a number of disadvantages leading to in-hospital mortality, such as marginal donors and grafts and recipients suffering from severe conditions. We have thus developed surgical and medical innovations to overcome these disadvantages. The present study analyzes the causes of death and factors affecting in-hospital mortality in adult recipients of LDLT. Between November 1994 and December 2007, 576 consecutive adult patients underwent LDLT at a single medical center. Overall in-hospital mortality was 18.9%. The peak rate was 55.6% in 1996, and the rate gradually decreased thereafter to 4.4% in The most frequent cause of death was infection (62.5%), which was followed by rejection (15.7%) and nonseptic multiple-organ failure (8.9%). Being intensive care unit bound before the operation, ABO blood type incompatibility, an absence of postoperative enteral nutrition, and a Model for End-Stage Liver Disease score of 25 or higher were independent risk factors for in-hospital mortality. In ABO-identical and ABO-compatible cases, retransplantation and a positive lymphocyte crossmatch test were additional independent risk factors. In conclusion, even aggressive efforts, preoperative conditions such as being intensive care unit bound, a high Model for End-Stage Liver Disease score, retransplantation, and a positive lymphocyte crossmatch test are still risk factors. Enteral nutrition could be a promising strategy to improve adult LDLT. Liver Transpl 15: , AASLD. Received April 16, 2009; accepted June 24, In Japan, living donor liver transplantation (LDLT) was begun for pediatric patients because of the limited availability of deceased donor livers in To further extend the indications for adult patients, the first LDLT procedures in adult recipients were successfully performed with a left lobe graft in and with a right lobe graft in Thereafter, the number of adult-toadult LDLT procedures has dramatically increased, and it exceeded that of pediatric LDLT procedures after 1999 in Japan. LDLT now plays an important role in multimodal treatment options for hepatocellular carcinoma even in patients beyond the Milan criteria 4 and in those with end-stage liver diseases. Despite recent advances in surgical techniques and perioperative management, LDLT has a relatively high in-hospital mortality rate in comparison with other surgeries in the field of hepatobiliary-pancreatic medicine, such as hepatectomy and pancreaticoduodenectomy, because of impaired preoperative conditions and various complications, including acute rejection and infection. We reported that obvious pretransplant encephalopathy, a Model for End-Stage Liver Disease (MELD) score above 30, and a donor age above 50 years were independent predictive factors of graft failure according to a retrospective analysis of the initial 335 consecutive adult LDLT procedures. 5 We also reported that a posi- Abbreviations: FHF, fulminant hepatic failure; GRWR, graft-recipient weight ratio; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; ICU, intensive care unit; LDLT, living donor liver transplantation; LT, liver transplantation; MELD, Model for End-Stage Liver Disease. Address reprint requests to Toshimi Kaido, M.D., Department of Hepato-Biliary-Pancreatic and Transplant Surgery, Kyoto University School of Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto , Japan. Telephone: ; FAX: ; kaido@kuhp.kyoto-u.ac.jp DOI /lt Published online in Wiley InterScience ( American Association for the Study of Liver Diseases.

2 IN-HOSPITAL MORTALITY IN 576 ADULT RECIPIENTS OF LDLT 1421 tive lymphocyte crossmatch test against the donor was a risk factor for LDLT according to an analysis of 585 ABO-identical and human leukocyte antigen mismatched LDLT procedures only by univariate analysis. 6 We introduced postoperative enteral feeding to adult LDLT in 2003, but the outcome has not been analyzed yet. Since the first pediatric case of LDLT in June 1990, 7 we had performed 1282 LDLT procedures at a single center by December 2007, including 576 adult cases. In the present study, we analyze factors affecting in-hospital mortality and the impact of enteral feeding as well as the causes of death in 576 consecutive adult LDLT cases. PATIENTS AND METHODS Between June 1990 and December 2007, 1235 consecutive patients underwent 1297 liver transplantation (LT) procedures at Kyoto University Hospital (Kyoto, Japan). These cases consisted of 1282 LDLT procedures and 15 deceased donor LT procedures, and all of the LDLT cases were the subjects of the present study. There were 575 males and 707 females and 706 children (less than 18 years old) and 576 adults (18 years old and older), with a median age of 13.3 years (range, years). The median MELD score was 20 (range, 2-54). As for ABO compatibility, 195 cases were ABOincompatible, 846 cases were ABO-identical, and 241 cases were ABO-compatible. Selection criteria for the donor and recipient as well as the surgical techniques for both the donor and recipient operations have been described in detail elsewhere. 5,8,9 Four hundred sixty-nine cases received a right lobe graft with or without a middle hepatic vein graft, 68 cases received a left lobe graft with or without a caudate lobe, and 39 cases received another type of graft or transplant, including posterior segment, lateral segment, and auxiliary partial orthotopic LDLT. The immunosuppressive treatment usually consisted of tacrolimus or cyclosporine and low-dose steroids as described elsewhere. 5,10 We started a program for ABO-incompatible cases that included a postoperative portal vein infusion of prostaglandin E1 and steroids in November Thereafter, we implemented a new protocol consisting of 3 treatments starting in June 2004: first, a preoperative anti-cd20 antibody (rituximab, 375 mg/m 2 ) treatment with preoperative plasma exchange to lower the anti-ab antigen titer; second, the postoperative continuous arterial administration of prostaglandin E1 (0.01 g/kg/minute) and steroids (125 mg/day) via a catheter intraoperatively settled in the hepatic artery; and third, postoperative cyclophosphamide (2 mg/kg/ day) followed by mycophenolate mofetil (starting dose, 500 mg/day; maintenance dose, 1000 mg/day) Enteral Nutrition We also started postoperative enteral nutrition after At the time of surgery, a Witzel tube jejunostomy was placed in the proximal jejunum with a 9F enteral tube. Patients received 5% glucose at a speed of 20 ml/hour on the first postoperative day through the jejunostomy. An enteral diet (Erental, Ajinomoto Pharma, Tokyo, Japan) was started at 250 kcal/day on the second or third postoperative day, and 250 kcal/ day was added up to 1000 kcal/day. Enteral feeding was continued until the oral intake was adequate. After the introduction of early enteral feeding, we usually did not provide total parenteral nutrition after LT. In-Hospital Mortality In-hospital mortality was defined as any death within the same hospital admission for LDLT, regardless of the number of days after LDLT. The causes of death were examined. We categorized multiple-organ failure following infection as infection and that without infection as nonseptic multiple-organ failure. Rejection included acute cellular rejection, chronic rejection, and antibody-mediated rejection. Data regarding the following recipient variables for each patient were obtained and analyzed: the ages of the donor and recipient, gender, ABO compatibility, results of a lymphocyte crossmatch test against the donor, graft type (right lobe or left lobe), graft-recipient weight ratio, original diseases for LDLT, pretransplant patient condition, MELD score, postoperative enteral nutrition, and initial transplantation or retransplantation. A lymphocyte crossmatch test against the donor was routinely performed except for emergency cases. The method has been described elsewhere. 6 Statistical Analysis The comparison of categorical variables was performed with the 2 test or Fisher exact test when appropriate. Multivariate analysis was performed with multiple logistic regression models. A P value 0.05 was considered significant. JMP was used for all statistical analyses. RESULTS One hundred ninety-three of 1282 patients died after LDLT within the same hospital admission. Figure 1 shows the number of patients who died in each postoperative period. Seventy-seven of 193 patients (39.9%) died within 30 days after surgery, whereas 116 patients (60.1%) died after 31 or more postoperative days. The overall average in-hospital mortality in all cases was 15.1%. The peak rate was 21.9% in 2000, and the rate gradually decreased thereafter to 4.8% in 2007 (Fig. 2). In-hospital mortality was significantly higher in adult cases (18.9%) than in pediatric cases (11.3%; P 0.001). Therefore, we next focused on adult cases and analyzed the causes of death and various parameters influencing in-hospital mortality. Figure 3 shows the annual number of adult cases and in-hospital mortality. In adult cases, the peak rate was 55.6% in 1996,

3 1422 KAIDO ET AL. TABLE 1. Causes of In-Hospital Mortality in Adult Cases Cause Number of Cases Infection 70 (62.5%) Rejection 11 (15.7%) Nonseptic multiple-organ failure 10 (8.9%) Vascular complications 7 (6.3%) Hepatic artery thrombosis 3 Hepatic artery rupture 2 Portal vein thrombosis 1 Rupture of esophageal varices 1 Cerebrovascular diseases 7 (6.3%) Others 7 (6.3%) Total 112 TABLE 2. Original Diseases in Adult Cases Figure 1. Number of patients and day of death within the same hospital admission for living donor liver transplantation. Number of Disease Cases HCC 173 HBV- or HCV-associated and 166 alcoholic liver cirrhosis Cholestatic diseases 131 Fulminant hepatic failure 69 Other diseases 37 Abbreviations: HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus. Figure 2. Change in in-hospital mortality in all cases. The peak rate was 21.9% in 2000, and the rate gradually decreased thereafter to 4.8% in Figure 3. Annual number of adult cases and in-hospital mortality. In adult cases, the peak rate was 55.6% in 1996, and the rate then gradually decreased thereafter to 4.4% in and the rate gradually decreased thereafter to 4.4% in The most frequent cause of in-hospital mortality was infection, including sepsis, pneumonia, and peritonitis (62.5%), which was followed by rejection (15.7%), nonseptic multiple-organ failure (8.9%), and vascular complications such as hepatic artery thrombosis and portal vein thrombosis (6.3%) in descending order of frequency (Table 1). The original diseases of 576 adult LT cases are shown in Table 2. The following factors were associated with in-hospital mortality on univariate analysis: recipient age under 50 years old, a MELD score of 25 or higher, ABO incompatibility, being intensive care unit (ICU) bound pre-transplant, fulminant hepatic failure, being lymphocyte crossmatch positive, an absence of postoperative enteral nutrition, and retransplantation (Table 3). Those significant prognostic factors were entered into the multiple logistic regression model, which demonstrated 4 independent adverse prognostic factors for in-hospital mortality: a MELD score of 25 or higher, ABO incompatibility, being ICU-bound pre-transplant, and an absence of postoperative enteral nutrition (Table 4). Among these factors, a high MELD score and being ICU-bound pre-transplant reflected the severity of the poor general condition. Since the introduction of arterial infusion to ABO-incompatible LDLT, in-hospital mortality has significantly decreased from 29.9% (n 67 before portal infusion) to 11.3% (n 80 after arterial infusion; P 0.005). Because the ABO blood type barrier has now been overcome and ABO-incom-

4 IN-HOSPITAL MORTALITY IN 576 ADULT RECIPIENTS OF LDLT 1423 TABLE 3. Univariate Analysis of Factors Affecting In- Hospital Mortality in Adult Cases Variable In-Hospital Mortality P Recipient age 18, 50 (n 294) 22.8% (n 282) 15.6% Donor age 20, 50 (n 379) 19.3% (n 197) 18.8% Gender Male (n 293) 16.0% Female (n 283) 22.3% ABO compatibility Identical/compatible (n 492) 17.5% Incompatible (n 84) 29.8% Lymphocyte crossmatch test Negative (n 498) 16.9% Positive (n 20) 30.0% Unknown (n 58) 31.0% Original disease HCC (n 173) 12.1% HBV/HCV/alcoholic (n 166) 21.1% Cholestatic (n 131) 19.8% FHF (n 69) 31.9% Others (n 37) 16.2% Graft type Right lobe grafts (n 469) 17.2% Left lobe grafts (n 68) 21.7% GRWR 0.8% (n 488) 18.9% % (n 88) 20.5% Preoperative condition ICU-bound (n 94) 41.5% Hospitalized (n 278) 18.1% At home (n 204) 9.9% MELD score 25 (n 199) 29.7% (n 377) 13.8% Postoperative enteral nutrition With (n 323) 13.3% Absence (n 253) 26.9% Initial or retransplantation Initial (n 554) 18.4% Retransplantation (n 22) 40.9% Abbreviations: FHF, fulminant hepatic failure; GRWR, graft-recipient weight ratio; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis virus C; ICU, intensive care unit; MELD, Model for End-Stage Liver Disease. patible LT is still not popular in the majority of transplant centers in the world, we further analyzed risk factors and the impact of enteral nutrition in ABOcompatible and ABO-identical LDLT to figure out more general information. This analysis showed that retransplantation, being lymphocyte crossmatch positive, being ICU-bound before the operation, a MELD score of 25 or higher, and an absence of postoperative enteral nutrition were independent risk factors (Table 5). DISCUSSION The present study analyzed risk factors for in-hospital mortality as well as causes of death in a large number of LDLT cases at a single medical center. The overall inhospital mortality (15.1%) and that in adult cases (18.9%) seemed to be higher than that in reports from other Asian countries in which LDLT is more common than deceased donor LT. Lee et al. 14 reported that inhospital mortality was 10.6% in 311 cases of adult-toadult LDLT between 1997 and Surprisingly, Liu et al. 15 reported that in-hospital mortality in 124 adult patients who underwent right lobe LDLT from 2000 to 2004 was 1.6%. However, in-hospital mortality largely depends on the indication for LDLT and the preoperative condition of the recipients. For example, 15.0% of our cases were ABO-incompatible, and ABO incompatibility is a contraindication for deceased donor LT. Because many transplant centers even in Japan have excluded ABO-incompatible cases from their indications for LDLT, many patients have been referred to our center. Recently, ABO-incompatible cases have composed about a quarter of all of our cases. In contrast, the 2 aforementioned reports did not include ABO-incompatible cases. Moreover, we had aggressively accepted patients with severe conditions, such as those who were ICU-bound and had a high MELD score. The median preoperative MELD score in patients between 1990 and 2001 was 22, which was higher than that in other reports from high-volume centers. 16,17 Furthermore, the present study includes all LDLT cases since Therefore, to correctly compare in-hospital mortality with that of other institutes, subgroup analysis of cases showing similar patient backgrounds and operation years would be needed. We previously reported that obvious pretransplant encephalopathy, a MELD score above 30, and a donor age above 50 years were independent risk factors for graft failure in 335 adult LDLT cases between November 1994 and December Because patients with obvious pretransplant encephalopathy are usually admitted to the ICU, the former 2 factors are almost in line with this report. Donor age, however, was not a risk factor for in-hospital mortality in the present study. For this reason, the present report contains 241 adult LDLT cases after 2004, when in-hospital mortality dramatically decreased. Our most recent report examining the influence of donor age showed that an old donor (60 years old or older) was not an independent factor influencing the survival rates of adult LDLT cases. 18 Moreover, other investigators recently reported that cumulative graft survival after LDLT did not differ between older donors (older than 50 years old) and younger donors (50 years old or younger). 17,19 All in all, it can be concluded that donor age does not affect graft or patient survival in LDLT as long as current donor selection criteria are maintained. Being lymphocyte crossmatch positive was not a risk factor in this study in all cases, including ABO-incompatible cases, although it was an independent risk factor in ABO-identical or ABO-compatible cases in line

5 1424 KAIDO ET AL. TABLE 4. Multivariate Analysis of Factors Affecting In-Hospital Mortality in Adult Cases Variable Odds Ratio (95% Confidence Interval) P Being ICU-bound before the operation ( ) ABO incompatibility ( ) Absence of enteral nutrition ( ) MELD score ( ) Abbreviations: ICU, intensive care unit; MELD, Model for End-Stage Liver Disease. TABLE 5. Multivariate Analysis of Factors Affecting In-Hospital Mortality in Adult ABO-Identical or ABO- Compatible Cases Variable Odds Ratio (95% Confidence Interval) P Retransplantation ( ) Being lymphocyte crossmatch positive ( ) Preoperative ICU stay ( ) MELD score ( ) Absence of enteral nutrition ( ) Abbreviations: ICU, intensive care unit; MELD, Model for End-Stage Liver Disease. with our previous report. 6 In fact, most ABO-incompatible cases (97.5%) were lymphocyte crossmatch negative recipients. Moreover, 2 ABO-incompatible and lymphocyte crossmatch positive recipients were discharged from the hospital without lethal complications. These findings might be the reason that being lymphocyte crossmatch positive was not a risk factor in all cases. However, the finding that being lymphocyte crossmatch positive was an independent risk factor in ABO-identical or ABO-compatible cases clearly shows us the necessity of designing a treatment strategy for lymphocyte crossmatch positive recipients in this subpopulation. The fact that ABO-incompatible LT negated the impact of a positive crossmatch suggests the effectiveness of our ABO-incompatible immunosuppressive regimen for a positive crossmatch. A high incidence of infection has been a major concern in the perioperative management of LDLT. In our center as well, infection, including sepsis, pneumonia, and peritonitis, was the most frequent cause of death. In clinical settings, recipients have a high risk of infection due to various factors, including preoperative impaired nutritional status, major surgery with a prolonged surgical duration, and postoperative immunosuppressive treatment. Many investigators have so far reported that there is a close relationship between the nutritional state and postoperative morbidity and mortality Enteral feeding has been associated with decreased postoperative infection rates and fewer metabolic complications after LT in comparison with total parenteral nutrition. 23 Therefore, we started the routine use of postoperative early enteral nutrition via tube jejunostomy in The mean and median enteral periods of feeding were both 21 days. In comparison with a transnasal feeding tube, the greatest advantage of a surgical Witzel tube is the feasibility of long-term feeding without any discomfort. US and European patients generally can eat sufficient food 5 days after transplantation. However, it takes about 2 weeks for Japanese patients to be able to eat sufficient food after LDLT because of appetite loss and general weakness. Hence, we had to maintain a central venous line for at least 2 weeks, and this was one of the causes of sepsis. After we introduced the enteral feeding system, we did not need to maintain the central venous line for more than 5 days after transplantation. Consequently, about 70% of the infection risk was eliminated with the introduction of enteral feeding. Dobhoff transnasal feeding tubes can be conveniently passed through the ligament of Treitz intraoperatively without the need for surgical holes in the bowel; however, these tubes cause patients not a little discomfort and carry a risk of pneumonia. Therefore, we prefer feeding jejunostomy to Dobhoff transnasal feeding tubes, although there might be a risk of infection during the surgical procedure for jejunostomy. We recently reported that the CD8 T cell subpopulation enriched with cytotoxic T lymphocytes is associated with a low survival probability and a high rate of infection in patients undergoing LDLT. 24 We are now investigating the effects of preoperative nutritional support on immunological aspects. To achieve an in-hospital mortality rate of zero is still difficult, especially in a tertiary center such as our institute. Recently, about a quarter of the patients who underwent LDLT at our institute were patients who were refused at other institutes for various reasons, including age, graft size mismatch, ABO incompatibility, and/or severe liver dysfunction. However, we are convinced that overcoming such difficult issues with the aid of a wide range of specialists, including an im-

6 IN-HOSPITAL MORTALITY IN 576 ADULT RECIPIENTS OF LDLT 1425 munologist, a nutritional support team, and an infection control team, is the mission of a high-volume center and will absolutely lead to advances in LT. Our next aim is to overcome a positive crossmatch by B cell desensitization and to improve preoperative conditions by preoperative nutritional support. REFERENCES 1. Nagasue N, Kohno H, Matsuo S, Yamanoi A, Uchida M, Takemoto Y, et al. Segmental (partial) liver transplantation from a living donor. Transplant Proc 1992;24: Hashikura Y, Makuuchi M, Kawasaki S, Matsunami H, Ikegami T, Nakazawa Y, et al. Successful living-related partial liver transplantation to an adult patient. Lancet 1994;343: Yamaoka Y, Washida M, Honda K, Tanaka K, Mori K, Shimahara Y, et al. Liver transplantation using a right lobe graft from a living related donor. Transplantation 1994;57: Mazzaferro V, Regalia E, Doci R, Andreola S, Pulvirenti A, Montalto F, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. New Engl J Med 1996;334: Morioka D, Egawa H, Kasahara M, Ito T, Haga H, Takada Y, et al. Outcomes of adult-to-adult living donor liver transplantation: a single institution s experience with 335 consecutive cases. Ann Surg 2007;245: Ashihara E, Tsuji H, Sakashita H, Haga H, Yurugi K, Kimura S, et al. Antidonor antibody in patients receiving ABO-identical and HLA-mismatched living donor liver transplants: effect on survival. Transplantation 2007;83: Tanaka K, Uemoto S, Tokunaga Y, Fujita S, Sano K, Yamamoto E, et al. Living related liver transplantation in children. Am J Surg 1994;168: Inomata Y, Uemoto S, Asonuma K, Egawa H. Right lobe graft in living donor liver transplantation. Transplantation 2000;69: Ito T, Kiuchi T, Egawa H, Kaihara S, Oike F, Ogura Y, et al. Surgery-related morbidity in living donors of right-lobe liver graft: lessons from the first 200 cases. Transplantation 2003;76: Inomata Y, Tanaka K, Egawa H, Uemoto S, Ozaki N, Okajima H, et al. The evolution of immunosuppression with FK 506 in pediatric living related liver transplantation. Transplantation 1996;61: Egawa H, Ohmori K, Haga H, Tsuji H, Yurugi K, Miyagawa-Hayashino A, et al. B-cell surface marker analysis for improvement of rituximab prophylaxis in ABO-incompatible adult living donor liver transplantation. Liver Transpl 2007;13: Egawa H, Ohdan H, Haga H, Tsuruyama T, Oike F, Uemoto S, Ozawa K. Current status of liver transplantation across ABO blood-type barrier. J Hepatobiliary Pancreat Surg 2008;15: Egawa H, Teramukai S, Haga H, Tanabe M, Fukushima M, Shimazu M. Present status of ABO-incompatible living donor liver transplantation in Japan. Hepatology 2008;47: Lee SG, Park KM, Hwang S, Lee YJ, Kim KH, Ahn CS, et al. Adult-to-adult living donor liver transplantation at the Asan Medical Center, Korea. Asian J Surg 2002;25: Liu B, Yan LN, Wang WT, Li B, Zeng Y, Wen TF, et al. Clinical study on safety of adult-to-adult living donor liver transplantation in both donors and recipients. World J Gastroenterol 2007;13: Yoshida R, Iwamoto T, Yagi T, Sato D, Umeda Y, Mizuno K, et al. Preoperative assessment of the risk factors that help to predict the prognosis after living donor liver transplantation. World J Surg 2008;32: Akamatsu N, Sugawara Y, Tamura S, Kaneko J, Matsui Y, Togashi J, Makuuchi M. Impact of live donor age ( 50) on liver transplantation. Transplant Proc 2007;39: Kuramitsu K, Egawa H, Keeffe EB, Kasahara M, Ito T, Sakamoto S, et al. Impact of age older than 60 years in living donor liver transplantation. Transplantation 2007; 84: Ikegami T, Taketomi A, Ohta R, Soejima Y, Yoshizumi T, Shimada M, Kaehara Y. Donor age in living donor liver transplantation. Transplant Proc 2008;40: Hasse JM. Nutritional implications of liver transplantation. Henry Ford Hosp Med J 1990;38: Campos AC, Matias JE, Coelho JC. Nutritional aspects of liver transplantation. Curr Opin Clin Nutr Metab Care 2002;5: Sanchez AJ, Aranda-Michel J. Nutrition for the liver transplant patient. Liver Transpl 2006;12: Hasse JM, Blue LS, Liepa GU, Goldstein RM, Jennings LW, Mor E, et al. Early enteral nutrition support in patients undergoing liver transplantation. J Parenter Enteral Nutr 1995;19: Tanaka K, Uemoto S, Egawa H, Takada Y, Ozawa K, Teramukai S, et al. Cytotoxic T-cell-mediated defense against infections in human liver transplant recipients. Liver Transpl 2007;13: