Hwm YI Yoo, * Ernest0 Molmenti, ' and PuulJ Tbuluvutb"

Transcription

1 The Effect of Donor Body Mass Index on Primary Nonfunction, Retransplantation Rate, and Early and Patient Survival After Liver Transplantation Hwm YI Yoo, * Ernest0 Molmenti, ' and PuulJ Tbuluvutb" Previous studies have suggested that moderate donor liver steatosis is associated with an increased incidence of primary graft nonfunction (PGNF), delayed graft function, early graft loss, and retransplantation rates. The objective of our study was to determine the effect of donor body mass index (dbmi), after adjusting for other known confounding variables, on PGNF, early graft failure, retransplantation rate, and patient survival. The United Network for Organ Sharing (UNOS) database (1987 to 2001) of 22,303 adult patients, excluding patients of pediatric age (age less than 18 years) and those with multiple organ transplantations, living donors, and retransplantations, was used for this study. Based on the BM1 (kg/m2) of the organ donor, transplant recipients were divided into four groups as follows: BM1 < 25 (group 1, n = 11,660), 25 to 29.9 (group 2, n = 7418), 30 to 34.9 (group 3, n = 2301), and 2 35 (group 4, n = 924). Information on donor liver histology was available for 1603 patients who underwent transplantation after 1999, and this subgroup was divided into three groups based on severity of steatosis (group A, < 20%; group B, 20% to 35%; group C, > 35%). Incidence of PGNF and early retransplantation rates were similar in groups 1 to 4 and in groups A, B, and C. Logistic regression analysis showed that dbmi or severity of steatosis was not a predictor of PGNF and early retransplantation. Cox regression analysis, after adjusting for confounding variables, showed similar patient and graft survival at 1 month and 1,2, and 5 years for groups 1 to 4, and at l month and 6 months for groups A, B, and C. Severe donor obesity or moderate steatosis did not influence short-term and long-term outcome liver of transplantation. (Liver Transpl2003;9:72-78.) ncreased prevalence of obesity in the general popu- I lation has led to a higher number of obese donors. Obesity is considered to be a major risk factor for nonalcoholic steatosis.'-5 The fatty infiltration (steatosis) is frequently observed in donor livers,6 and a liver biopsy used to evaluate the degree of steatosis may be the only determinant of whether an organ will be implanted or not. Moderate steatosis (30% to 60%) in the donor liver has been reported to be associated with an increased incidence of delayed or primary graft nonfunction (PGNF), early graft loss, and retransplantati0n.7-l~ One study reported 30% that fatty infiltration was associated with a significantly lower 4-month graft and 2-year patient survival. 1 PGNF and primary dysfunction were the main causes of reduced graft and patient survival, and the investigators suggested that liver allografts that were up to 30% fat were an independent predictor of poor outcome. However, some studies have reported comparable survival in those who received a liver with moderate steatosis.12,13 The outcome of liver transplantation (LT) is dependent on many confounding variables, including donor and recipient age, serum creatinine level, body mass index (BMI), etiology, and severity of liver disease (United Network for Organ Sharing [UNOS] status). Because of small sample size, previous studies have not been able to determine the independent effect of moderate steatosis of donor liver on immediate and late posttransplantation outcome. The objective of our study was to determine the effect of donor body mass index (dbmi), after adjusting for other known confounding variables, on PGNF, early graft failure, retransplantation, and postoperative survival of patients and grafts using the LT database from the UNOS. Patients and Methods The UNOS database (Standard Transplant Analysis and Research [STAR] file) obtained from the UNOS included 50,616 patients who underwent LT between 1988 and We collected the following information regarding the recipients:age,sex,race,height,weight,serumcreatininelevel, etiology of liver disease, AB0 blood type matching status, UNOS listing status, and cold ischemia time. For donors, we collected data on age, sex, race, height, and weight. BM1 of donor and recipient at the time of transplantation was calculated by dividing the weight in kilograms by the height in meters squared. When the histology of donor liver was available, we recorded that information. The outcome variables that were recorded included PGNF, early retransplantation, From the "Division of Gastroenterology and Hepatology, Johns Hopkins University Hospital, and the?department of Surgey, The Johns Hopkins University School of Medicine, Baltimore, MD. Address reprint requests to PaulJ Thuluvath, MD, FRCP, Division of Gastroenterology and Hepatology, Johns Hopkins University Hospital, 1830 E Monument St, Room 428, Baltimore, MD Telephone: ; Fm: ; pjtuluv@jhmi.edu Copyright O 2003 by the American Association for the Study of Liver Diseases 1~ /03/ $35.00/0 doi: IO. 1053~lts Liver Transphntation, Vol9, No 1 (Janus,), 2003: pp 72-78

2 Donor BMI and Outcome Transplantation of Liver 73 and l-month, l-year, 2-year, and 5-year graft and patient ing variables. The variables used in the Cox regression analysis survival. wererecipientage,sex,race,bmi,serumcreatininelevel, Complete data on dbmi and posttransplantation patient UNOS listing status, cause of liver disease, cold ischemia or graft survival information were available only for 22,303 time, donor age, and dbmi. For all analyses, a two-tailed P patients, after excluding 16,409 patients of pediatric age (age value of.05 or less was considered significant. The statistical less than 18 years), those with multiple organ transplantation, analyseswereperformedusing thesoftware SPSS version those who received transplants from living donors, patients (SPSS, Inc, Chicago, IL). who had LT previously, and those without survival information (n = 149). In addition, 12,163 donors were excluded Results from analysis because of incomplete data on recipients or donors and implausible dbmi (< 20 kg/m2 or > 55 kg/m2). The recipients of the remaining 22,303 donor livers were divided into four groups based on the dbmi of the organ donor: group 1, dbmi < 25 kg/m2 (n = 11,660); group 2, dbmi 25 to 29.9 kg/m2 (n = 7418); group 3, dbmi 30 to 34.9 kg/m2 (n = 2301); and group 4, dbmi 135 kg/m2 (n = 924). The definition of UNOS listing status had undergone changes during the study period except for status 1. For the study, UNOS listing status was divided into three categories: UNOS status 1, UNOS status 2A, and others. We classified the cause of liver disease as alcoholic liver disease, hepatitis C, hepatitis B, cryptogenic cirrhosis, primary biliary cirrhosis, and others. Patients with a diagnosis of non-a, non-b hepatitis were included as hepatitis C, combined hepatitis B and C as hepatitis B, and combined alcoholic liver disease and postnecrotic cirrhosis (possible viral hepatitis) as alcoholic liver disease. AB0 matching was divided into three groups, matched (donor and recipient were ofsame AB0 blood type), compatible (donor AB0 bloodtypewas 0 andrecipient AB0 blood type was other than 0, or when recipient AB0 blood type was AB and donor AB0 blood type was other than AB), and mismatched (donor AB0 blood type was A, B, or AB and recipient had another AB0 blood type different from that of the donor). For the exclusion of outliers in our data, recipients with implausible BM1 and creatinine value of 0 mg/dl were considered as missing. The severity of fatty infiltration of the donor liver (liver histology) was recorded in 1603 patients who received LT between 1999 and 200 l. This subgroup was analyzed in more detail. Depending on the severity of fatty infiltration of the donor liver, recipients were divided into three groups: group 1 had 0 to 19% (group A, n = 1363), group 2 had 20% to 35% (group B, n = 193), and group 3 had > 35% (group C, n = 47) steatosis. Further classification of group C was not possible from the database. The short-term and long-term outcomes based on donor liver histology were determined as described earlier. Statistical Analysis The patient characteristics were compared by Chi-squared test and ANOVA for categorical and continuous variables respectively. One-month, l-year, 2-year, and 5-year survival after LT were determined by Kaplan-Meier and Cox regression survival analysis. Log-rank test was used to analyze the differences in survival by Kaplan-Meier analysis. Cox regression analysis was performed to adjust survival for confound- There were statistically significant differences in recipient age, sex, race, UNOS status, etiology of liver disease, BMI, donor age, and donor sex among four groups of patients. However, many of these differences were numerically small except for donor age (Table 1). Serum creatinine level, AB0 matching, and cold ischemia time were similar in groups 1 through 4. There were fewer female recipients, probably because of size match, in the higher dbmi group, but the converse was true for donors. The incidence of PGNF (P =.5) and retransplantation rates (P =.9) was similar in all four groups (Table 2). Logistic regression analysis showed that dbmi was not a predictor of PGNF and early retransplantation (data not shown). Kaplan-Meier survival analysis showed similar survival trends in groups 1 through 4 (Fig. 1). Of 22,303 patients included in the study, 19,425 patients (87.1%) were available for Cox regression analysis and the rest (I 2.9%) had to be excluded because of one or more of missing variables. Independent predictors (P <.OS) of survival on Cox regression analysis were age, sex, race, BMI, serum creatinine level, UNOS listing status, AB0 mismatch, etiology of liver disease for the recipient, and age and cold ischemia time for the donor (data not shown). However, dbmi (reference variable group 1) was not a predictor of short-term or long-term graft and patient survival (Table 3). The proportion of donors with a higher BM1 (groups 3 and 4) increased in recent years, and despite the increase in dbmi, patient and graft survival were better in recent years (Fig. 2). To examine whether morbid donor obesity influenced the outcome, we divided group 4 (dbmi 35 kg/m2) into those with BM1 > 35 to 39.9 kg/m2, BM1 > 40 to 44.9 kg/m2 (n = 157), and BM1 > 45 kg/m2 (n = 104). Cox regression analysis showed similar (short-term and long-term) patient and graft survival in those who received allografts from donors with BM1 > 40 to 44.9 kg/m2 (patient survival HR 0.8, C1 0.6 to 1.2; graft survival HR 0.8, C1 0.5 to 1.2) and BM1 > 45 kg/m2 (patient survival HR 1.16, C1 0.8 to 1.6; graft survival HR 1.1, C1 0.8 to 1.7).

3 74 Yo0 et al Age Cy) Recipient gender (female Yo) Etiology (%) ALD HCV HBV PBC Crypt0 Race (%) White Black Asian Hispanic UNOS listing (%) 1 2A BM1 (kg/m2) Creatinine (mg/dl) Cold ischemia time (h) AB0 matching (Yo) Matched Compatible Mismatched Donor age (y) Donor gender (female %) Group 1 Group 2 Group 3 Group 4 (<25 kg/m2) (25 to 29.9 kg/m2) (30 to 34.9 kg/m2) (235 kg/m2) (n = 11,660) (n = 7,418) (n = 2,301) (n = 924) P Value ? ? ? ? ? ? ? t ? ? t t ? O ? <.oo 1.o Abbreviations: ALD, alcoholic liver disease; HCV, hepatitis C virus; HBV, hepatitis B virus; PBC, primary biliary cirrhosis; crypto, cryptogenic; UNOS, united Network for Organ Sharing; BMI, body mass index. Donor body mass index (kg/m2) <25 25 to to Fatty infiltration (%) to 35 >35 Primary Nonfunction Early Retransplantation Within 30 days n (Yo) After Transplantation (Yo) P=.5 11,660 (3.5) 409 7, (3.9) 2, (3.8) (4.0) P=.8 1, (3.3) (2.6) 47 2 (4.3) P' (3.6) 267 (3.7) 78 (3.6) 33 (3.6) P=.4 62 (4.5) 5 (2.6) 3 (6.3)

4 Donor BMI Outcome and Transplantation of Liver 75 p = m 'E Q) 3 B U p = * Days of Patient Survival Days of Survival Figure 1. Patient (A) and gr& (B) survival by Kaplan-Meier survival analysis based on donor body mass index. Numbers below each survival graph indicate the number of recipients at risk. (Table 4). Incidence of PGNF and early retransplanta- months because there were only a few patients available tion rate were similar in groups A, B, and C (Table 2) for analysis after 6 months (the information on liver and severity of fatty infiltration was not a predictor of histology came from more recent transplantations). PGNF and early retransplantation by logistic regression Kaplan-Meier survival analysis showed similar patient analysis (data not shown). A total of 1319 patients survival in all three groups up to 6 months (P =.l2 by (82.3%) were available (complete data) for Cox regres- log rank test). Six-month graft survival was lower P = sion analysis. Survival analysis was done only up to 6.02) in group A (less than 20% fat) compared with the Group 2 25 to 29.9 kg/m2 PValue (HR, 95% CI) Group 3 30 to 34.9 kg/rn2 PValue (HR, 95% CI) ).90 1 month Patient.l 1 (1.2, ) 09) (1.1, ) 1 year ).93 Patient.28 (1.1, ) 6) (1.1, ) 2 year ).69 Patient.55 (1.0, ) (1.0, OS).63 (1.0, ) 5 year Patient.48 (1.0, ).48 (1.0, ) 8) (1.0, ) *Indicator (reference) variable was patients with body mass index < 25 kg/m2. Abbreviations: HR, hazard ratio; CI, confidence interval. Group kg/rn2 PValue (HR, 95% CI).51 (1.1, ).32 (1.1, ).75 (1.0, ).85 (1.0, ).62 (0.9, ).93 (1.0, ).29 (0.9, ).56 (1.0, )

5 76 Yo0 et al B'.Ol l ).e Y " Days of Patient Survival Days of Survival Figure 2. (A) Donor body mass index has increased during recent years. Cox regression analysis shows a trend toward improved patient (B) and graft (C) survival in more recent years. &MI = donor body mass index (kg/m2). other two groups (Fig. 3), but this difference disap- increase in donor obesity, the transplantation outcomes peared when adjusted for confounding variables by Cox have improved in recent years. The and moderregression analysis (Table 5). ate fatty infiltration did not influence PGNF, early retransplantation rate, or early and late graft and patient Discussion survival. Our data suggest that severity of fatty infiltra- In this study, we have shown &at &MI has increased tion increased with increases in dbmi, and these results in recent years as in the general population. Despite an corroborates previous observations.2.6 Based on our I Table 4. The Relationship Between the Severity of Donor Liver Fatty Infiltration and Donor Body Mass Index I I Donor Body Mass Index <19% (Group A) 20% to 35% (Group B) >35% (Group C) I Group 1 (C25 k//mz) (91633) 1.4% (461633) 7.3% (578/633) 91.3% Group 2 (25 to < 30 kg/m2) 84.4% (480/569) 13.7% (781569) 1.9% (111569) (OR 2.0, C , P =.001) P=.6 Group 3 (30 to C 35 kg/mz) 77.2% (200/259) 17.4% (45/259) 5.4% (14/259) (OR2.7,CI , PC,001) (OR4.0,CI , P =,002) Group 4 (235 kg/rnz) 73.9% (105/142) 16.9% (24/142) 9.2% (13/142) (OR2.6, C , P =.001)(OR7.0,C , P C.001) Abbreviations: OR, odds ratio when compared with Group 1; CI, confidence interval.

6 Donor BMI and Outcome of Liver Transplantation 77 p = Days of Patient Survival A B C v1 Q).z c, 4,.84 Group A 5 U p a Days of SuMval Figure 3. Patient (A) and graft (B) survival by Kaplan-Meier survival analysis based on the severity of fatty infiltration (group A, I 19% fat; group B, 20% to 35%; group C, > 35%). Numbers below each survival graph indicate the number of recipients at risk. study, it is reasonable to suggest that moderate steatosis did not influence patient or graft outcome after LT. Our results are in contrast to those of a previous study that reported a lower 4-month graft survival in 59 patients who received liver allografts with up to 30% fat, compared with 57 patients who received livers without fatty infiltration. In our study, 193 patients had 20% to 35% fat and 47 had > 35% fat. A larger sample size allowed us to adjust the outcome for other known predictors of survival, and this may partly explain the discrepancy in the results. Tabie 5. Patient and Survival by Cox Regression Analysis Based on Severity of Fatry Infilrratian 20% >35% to 35% PValue (HR, CI)* PValue (HR, CI)* 1 month Patient.49 (0.60,O ).30 (0.40, ) (0.67, ).46.l8(0.43, ) 6 month Patient.77 (1.23, ).76 (0.78,O ) (0.86, ) (0.53, ) *men compared with patients with < 20% (group A) fatty infiltration (indicator variable). Abbreviations: HR, hazard ratio; C1 = confidence interval. An argument could also be made that improved preoperative and postoperative patient care, development of a better immunosuppressive regimen, refinement in surgical techniques, and better screening of potential recipients has accounted for some of the improvements in deleterious factors previously attributed to dbmi and moderate steatosis. Another possibility is that the larger liver size in more obese donors may compensate to some extent for moderate steatosis under optimal conditions. Previous studies have shown that severe fatty infiltration (> 60% fat) was associated with poor outcome immediately after LT. Donors used in this analysis were thought to be those with fatty infiltration of less severity because it has been a usual practice to perform liver biopsy from the donor livers before transplantation. Therefore, it is probable that many livers were discarded because of severe fatty infiltration and is it probable that there was a selection bias in our study despite the large sample size. UNOS data did not permit us categorize donors with more than 35% steatosis into further groups based on severity of steatosis. Moreover, we could not determine whether the steatosis was microvesicular or macrovesicular. It has been suggested that microvesicular steatosis is better tolerated.13,1* The limited data on liver histology do not allow us to make any strong recommendations, but it

7 78 Yo0 et a1 seems that it is very safe to use donor livers with steatosis of up to 35%. More importantly, extreme donor obesity (BM1 > 35 kg/m2) should not be considered a contraindication for organ retrieval because most of these organs could be used safely. Our study has many limitations. We had to exclude a substantial number of patients because of inadequate data on dbmi and a variety of other missing information. Nevertheless, the large number of patients allowed us to adjust the outcome for many independent confounding variables. The information on liver histology was available only for recent transplantation, and hence the numbers were limited. More importantly, the histologic grading of severity of steatosis was not standardized. Despite these limitations, we believe that our study is important for many reasons. The general population is becoming increasingly obese. Many of these subjects may have asymptomatic elevation in liver enzymes secondary to hepatic steatosis. It is therefore important to have guidelines for use of these livers with moderate steatosis that could be considered marginal donors by many centers. Liver biopsy is currently the most sensitive, specific, and cost-effective test in the diagnosis and staging of fatty liver. However, currently there are no rigid criteria for staging the sever- ity of steatosis, and the staining techniques and staging criteria may vary from center to center. It is important for transplantation community to develop uniform criteria for staining and staging steatosis. We believe that the definition of moderate steatosis or tolerable amount of fat in the liver may need be to redefined prospectively. Our study suggests that it is very safe to use donor livers with less than 35% steatosis for whole-organ transplantation. It is important to stress that our findings should not be extrapolated to partial LT (living donor) cases, in which the liver mass may not be adequate to compensate for moderate steatosis. References 1. Nasarallah SM, Wills CE, Galambos JT. Hepatic morphology in obesity. Dig Dis Sci 1981;26: Rao ARN, Chui AKK, Shi LW, Tsai L, Leon CD, Sheil AGR. Is donor obesity related to liver steatosis and liver graft dysfunction in liver transplantation? Transplant Proc 2000;32: Youssef W, McCullough AJ. Diabetes mellitus, obesity and hepatic steatosis. Semin Gastrointest Dis 2002; 13: Angulo P, Keach JV, Batts KP, Lindor KD. Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis. Hepatology 1999;30: Rinella ME, Alonso E, Rao S, Whitington P, Fryer J, Abecassis M, et al. Body mass index as a predictor of hepatic steatosis in living liver donors. Liver Transpl200 ;7: Ureiia MAG, Ruiz-Delgado FC, Gonzhlez EM, Romero CJ, Garcia IG, Segurola CL, et al. Hepatic steatosis in liver transplant donors: Common feature of donor population? World J Surg 1998;22: Ploeg RJ, D Alessandro AM, Knechtle SJ, Stegall MD, Pirsch JD, Hoffman RM, et al. Risk factors for primary dysfunction after liver transplantation-a multivariate analysis. Transplantation 1993;55: Chui AKK, Shi LW, Rao ARN, Verran DJ, Painter D, Koorey D, et al. Donor fatty (steatotic) liver allografts in orthotopic liver transplantation: a revisit. Transplant Proc 2000;32: Ureiia MAG, Ruiz-Delgado FC, Gonzdez EM, Segurola CL, Romero CJ, Garcia IG, et al. Assessing risk of the use of livers with macro and microsteatosis in a liver transplant program. Transplant Proc 1998;30: Todo S, Demetris AJ, Makowka L, Teperman L, Podesta L, Shaver T, et al. Primary nonfunction of hepatic allografts with preexisting fatty infiltration. Transplantation 1989;47: Marsman WA, Wiesner RH, Rodriguez L, Batts KP, Porayko MK, Hay JE, et al. Use of fatty donor liver is associated with diminished early patient and graft survival. Transplantation 1996;62: D Alessandro A M, Kalayoglu M, Sollinger HW, Hoffman RM, Reed A, Knechtle SJ, et al. The predictive value of donor liver biopsies on the development of primary nonfunction after orthotopic liver transplantation. Transplantation 199 ;5 1 : Fishbein TM, Fie1 MI, Emre S, Cubukcu 0, Guy SR, Schwart ME, et al. Use of livers with microvesicular fat safely expands the donor pool. Transplantation 1997;64: Zamboni F, Franchello A, David E, Rocca G, Ricchiuti A, LavezzoB, et al. Effect of macrovesicular steatosis and other donor and recipient characteristics on the outcome of liver transplantation. Clin Transplant 2001;15:53-57.