SEARCH

SEARCH BY CITATION

Abstract

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

Obesity levels in the United Kingdom have risen over the years. Studies from the United States and elsewhere have reported variable outcomes for obese liver transplant recipients in terms of post–liver transplant morbidity, mortality, and graft survival. This study was designed to analyze the impact of the body mass index (BMI) on outcomes following adult liver transplantation. Data from 1994 to 2009 were retrieved from a prospectively maintained database. Patients were stratified into 5 World Health Organization BMI categories: underweight (<18.5 kg/m2), normal weight (18.5-24.9 kg/m2), overweight (25.0-29.9 kg/m2), obese (30.0-34.9 kg/m2), and morbidly obese (≥35.0 kg/m2). The primary outcome was an evaluation of graft and patient survival, and the secondary outcome was an assessment of postoperative morbidity. Bonferroni correction was applied with statistical significance set at P < 0.012. Kaplan-Meier curves were used to study the effects of BMI on graft and patient survival. A total of 1325 patients were included in the study: underweight (n = 47 or 3.5%), normal-weight (n = 643 or 48.5%), overweight (n = 417 or 31.5%), obese (n = 145 or 10.9%), and morbidly obese patients (n = 73 or 5.5%). The rate of postoperative infective complications was significantly higher in the overweight (60.7%, P < 0.01) and obese recipients (65.5%, P < 0.01) versus the normal-weight recipients (50.4%). The morbidly obese patients had a longer mean intensive care unit (ICU) stay than the normal-weight patients (4.7 versus 3.2 days, P = 0.03). The mean hospital stay was longer for the overweight (22.4 days, P < 0.001), obese (21.3 days, P = 0.04), and morbidly obese recipients (22.4 days, P = 0.047) versus the normal-weight recipients (18.0 days). There was no difference in death-censored graft survival or patient survival between the groups. In conclusion, this is the largest and only reported UK series on BMI and outcomes following liver transplantation. Overweight and obese patients have significantly increased morbidity in terms of infective complications after liver transplantation and, consequently, longer ICU and hospital stays. Liver Transpl 19:551–562, 2013. © 2013 AASLD.

Abbreviations
BMI

body mass index

CI

confidence interval

HR

hazard ratio

ICU

intensive care unit

MELD

Model for End-Stage Liver Disease

WHO

World Health Organization.

Obesity rates are increasing worldwide, and the United Kingdom is no exception. A report published in February 2012 states that 42% and 26% of both men and women in England are overweight [body mass index (BMI) = 25-30 kg/m2] and obese (BMI > 30 kg/m2), respectively.[1] In keeping with this obesity epidemic, recent studies have also reported an increasing prevalence of obesity in the liver transplant population (including both donors and recipients).[2-4] It is well known that obesity is associated with poorer outcomes after both minor and major surgical procedures through associated comorbidities such as cardiovascular and pulmonary diseases and diabetes.[5, 6] Liver transplantation is considered to be one of the most technically challenging procedures, and any associated risk factor such as obesity can further increase postoperative morbidity and mortality.

The impact of recipient obesity on post–liver transplant complications and graft and patient survival remains controversial. Single-center studies that have been published from the United States and other European countries, with sample sizes ranging from 167 to 813 patients, have shown variable results for graft and patient survival.[7-17] Although some single-center studies have shown no effect of obesity on graft and patient survival,[7-11] others have reported reduced graft survival and overall increased mortality in obese and morbidly obese recipients.[12-17]

Three large transplant registry studies of the United Network for Organ Sharing database have been published over the past 10 years.[18-20] These studies have included patient data sets covering the last 20 years, with patient numbers ranging from 18,172 to 73,538, and they have reported similar outcomes for liver transplantation in terms of increased rates of primary nonfunction, increased postoperative complications, reduced graft survival, and overall reduced patient survival in obese and morbidly obese recipients versus normal-weight recipients. Some of these studies have even shown poorer outcomes for underweight recipients (BMI < 19 kg/m2).[19, 20] On the contrary, results from the liver transplantation databases of the Scientific Registry of Transplant Recipients and the National Institute of Diabetes and Digestive and Kidney Diseases suggest no increased morbidity or mortality in obese recipients.[21, 22] Moreover, a retrospective review of the Scientific Registry of Transplant Recipients database by Pelletier et al.[22] that compared survival benefits for obese patients on the waiting list and undergoing transplantation showed significant survival benefits (P < 0.001) for the latter group. However, there are no such studies from the United Kingdom.

The aim of the current study was to analyze the impact of BMI on outcomes following adult liver transplantation and especially look at the effects of high and low BMIs on postoperative complications and graft and patient survival. Additionally, we evaluated the potential risk factors affecting graft and patient survival in a large UK cohort.

PATIENTS AND METHODS

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

Data from 1994 to 2009 were retrieved from a prospectively maintained institutional database. All consecutive adult liver transplants performed during this period were included. Patients with no height or weight data available for BMI estimation were excluded. All the weight estimations were listing weights after those patients with clinical ascites were dry-tapped. Therefore, all weights recorded in our database were dry weights with little or no ascites. Patients were stratified into 5 distinct BMI categories established by the World Health Organization (WHO): underweight (<18.5 kg/m2), normal weight (18.5-24.9 kg/m2), overweight (25.0-29.9 kg/m2), obese (30.0-34.9 kg/m2), and morbidly obese (≥35.0 kg/m2). The primary outcome was an evaluation of graft and patient survival. Secondary outcomes included postoperative complications, the length of the intensive care unit (ICU) stay, the length of ventilation, and the overall length of the postoperative hospital stay.

The collected data included the donor age, sex, BMI, and graft type; the cold ischemia time; the recipient age, sex, BMI, primary etiology of liver disease (alcohol, virus, cholestatic liver disease, or malignancy), Model for End-Stage Liver Disease (MELD) score at transplant, albumin level, and status at the time of transplantation (ventilation before transplantation, renal support, and preoperative sepsis); superurgent transplantation (for patients with a severe form of liver disease needing urgent transplantation if they satisfied 1 of the 9 criteria listed in the National Health Service Blood and Transplant superurgent liver scheme[23]); retransplantation; and the year of transplantation. The recorded clinical outcomes were the need for intraoperative or postoperative blood transfusions (other blood products were excluded), the length of the postoperative ICU stay, the length of ventilation (in the postoperative period only for those who were ventilated before transplantation), the length of the hospital stay until first discharge, and postoperative complications [eg, hemorrhaging, chest infections, wound infections, ascitic or drain fluid infections, urinary infections, culture-proven bacteremia and other unproven sources of infection (opportunistic infections in transplant recipients such as cytomegalovirus and other infections were not included), vascular complications, bile leaks, and biliary strictures].

Statistical Analysis

Categorical variables are expressed as frequencies and proportions, and continuous variables are presented as means and standard deviations. Categorical variables were analyzed with Fisher's exact test or a chi-square test, and continuous variables were analyzed with an unpaired Student t test or a 1-way analysis of variance. For all analyses, we compared 4 categories of interest—underweight, overweight, obese, and morbidly obese patients—to the controls, who were normal-weight recipients. Bonferroni correction was used to adjust the empirical P values obtained from the aforementioned analysis, with the significance level after correction set at P < 0.012 (0.05/4; ie, correction for 4 comparisons in each sample set).

Kaplan-Meier survival curves were studied with log-rank statistics to assess the effects of BMI on death-censored graft survival and patient survival. The 1-, 3-, 5-, and 10-year graft and patient survival rates were extracted from the survival curves. The Kaplan-Meier curves were studied for both overall comparisons of the 5 BMI categories and for individual comparisons with normal-weight recipients after adjustments for potential confounders. The variables that showed statistical significance (P < 0.05) in the Kaplan-Meier univariate analysis were further subjected to a multivariate analysis using a Cox proportional hazards model, and hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated. Potential confounders such as donor age, donor BMI, cold ischemia time, recipient age, recipient sex, and recipient BMI were forced into the multivariate model along with other variables significant in the univariate analysis to strengthen the adjusted survival curves. The variables significant in the multivariate analysis were considered to be independent prognosticators of graft and patient survival. All statistical analyses were performed with SPSS 19.0 for Windows (SPSS, Inc., Chicago, IL).

RESULTS

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

Fourteen hundred adult transplants were identified during the study period; 1325 patients with height and weight data available were included in the study. According to the WHO BMI criteria, there were 47 underweight recipients (3.5%), 643 normal-weight recipients (48.5%), 417 overweight recipients (31.5%), 145 obese recipients (10.9%), and 73 morbidly obese recipients (5.5%).

Donor Characteristics

The mean donor age in the normal-weight group was 41.2 years. The mean donor age was significantly higher for the overweight (44.1 years, P < 0.01), obese (45.2 years, P < 0.01), and morbidly obese recipients (47.4 years, P < 0.01) versus the normal-weight recipients. Livers from male donors were received by 47.0% of the normal-weight recipients, whereas the morbidly obese group predominantly received grafts from male donors (61.6%, P = 0.02). In keeping with their weight, recipients with higher BMIs received grafts from donors with higher BMIs. The mean donor BMI for the normal-weight recipients was 24.0 kg/m2, which was significantly lower than the mean donor BMIs for the overweight (25.4 kg/m2, P < 0.001), obese (26.0 kg/m2, P < 0.001), and morbidly obese recipients (26.1 kg/m2, P < 0.001). There was no difference between the groups in terms of living donor, donation after brain death, or donation after circulatory death grafts. Similarly, there was no difference between the 5 groups in terms of the mean cold ischemia time (Table 1).

Table 1. Donor and Recipient Characteristics Stratified by the Recipient BMI Grade
VariableUnderweight: BMI < 18.5 kg/m2 (n = 47)Normal Weight: BMI = 18.5-24.9 kg/m2 (n = 643)Overweight: BMI = 25.0-29.9 kg/m2 (n = 417)Obese: BMI = 30-34.9 kg/m2 (n = 145)Morbidly Obese: BMI ≥ 35 kg/m2 (n = 73)P Valuea
Normal Weight Versus UnderweightNormal Weight Versus OverweightNormal Weight Versus ObeseNormal Weight Versus Morbidly Obese
  1. NOTE: Continuous variables are expressed as means and standard deviations, and categorical variables are expressed as numbers and percentages. The bolded values are significant.

  2. a

    P values were considered significant at <0.012 (Bonferroni correction: 0.05/4 = 0.012).

Donor age (years)41.0 ± 17.441.2 ± 15.644.1 ± 14.745.2 ± 13.247.4 ± 14.20.93<0.01<0.01<0.01
Donor sex: male [n (%)]23 (48.9)302 (47.0)221 (53.0)78 (53.8)45 (61.6)0.880.060.140.02
Donor BMI (kg/m2)24.3 ± 3.924.0 ± 3.925.4 ± 4.126.0 ± 4.126.1 ± 4.30.61<0.001<0.001<0.001
Living donor graft [n (%)]2 (4.3)3 (0.5)2 (0.5)0 (0)0 (0)0.04>0.990.46>0.99
Donation after brain death graft [n (%)]43 (91.5)610 (94.9)397 (95.2)139 (95.9)71 (97.3)0.310.880.830.57
Donation after circulatory death graft [n (%)]2 (4.3)30 (4.7)18 (4.3)6 (4.1)2 (2.7)>0.990.761.000.76
Cold ischemia time (minutes)594 ± 214582 ± 198558 ± 202585 ± 196630 ± 2070.690.060.870.05
Alcoholic liver disease [n (%)]5 (10.6)91 (14.2)81 (19.4)31 (21.4)15 (20.5)0.660.030.040.16
Hepatitis B virus etiology [n (%)]2 (4.3)24 (3.7)9 (2.2)6 (4.1)4 (5.5)0.690.200.810.52
Hepatitis C virus etiology [n (%)]2 (4.3)63 (9.8)91 (21.8)40 (27.6)16 (21.9)0.19<0.001<0.001<0.01
Cholestatic liver disease [n (%)]22 (46.8)246 (38.3)109 (26.1)37 (25.5)20 (27.4)0.28<0.001<0.010.07
Malignancy [n (%)]1 (2.1)27 (4.2)20 (4.8)6 (4.1)1 (1.4)0.710.65>0.990.35
Recipient age (years)40.4 ± 14.246.6 ± 13.549.4 ± 11.651.1 ± 8.350.5 ± 9.0<0.01<0.001<0.0010.02
Recipient sex: male [n (%)]18 (38.3)311 (48.4)278 (66.7)106 (73.1)41 (56.2)0.23<0.001<0.0010.22
Recipient MELD score18.3 ± 11.620.6 ± 11.718.8 ± 9.418.3 ± 8.919.8 ± 9.90.19<0.010.030.57
Ventilation before transplantation [n (%)]5 (10.6)83 (12.9)30 (7.2)6 (4.1)4 (5.5)0.82<0.01<0.010.09
Dialysis at the time of transplantation [n (%)]5 (10.6)114 (17.7)55 (13.2)10 (6.9)5 (6.8)0.310.048<0.0010.02
Preoperative sepsis [n (%)]1 (2.1)27 (4.2)14 (3.4)3 (2.1)1 (1.4)0.710.52<0.0010.35
Superurgent transplantation [n (%)]6 (12.8)121 (18.8)63 (15.1)10 (6.9)10 (13.7)0.430.13<0.0010.34
Retransplantation [n (%)]6 (12.8)70 (10.9)32 (7.7)7 (4.8)3 (4.1)0.630.090.170.09

Recipient Characteristics

The mean age of the normal-weight recipients was 46.6 years, which was significantly higher than the mean age of the underweight (40.4 years, P < 0.01), overweight (49.4 years, P < 0.001), and obese recipients (51.1 years, P < 0.001). There were more male recipients in the overweight (66.7%, P < 0.001) and obese groups (73.1%, P < 0.001) versus the normal-weight group (48.4%). There were significantly more recipients with hepatitis C virus cirrhosis in the overweight (21.8%, P < 0.001) and obese groups (27.6%, P < 0.001) versus the normal-weight group (9.8%). Cholestatic liver disease was one of the common indications for liver transplantation among the normal-weight recipients (38.3%) in comparison with the overweight (26.1%, P < 0.001) and obese recipients (25.5%, P < 0.01). Among those who underwent liver transplantation for a malignant etiology (n = 55), the diagnoses were hepatocellular carcinoma (n = 31 or 56.4%), cholangiocarcinoma (n = 6 or 10.9%), secondary liver tumors (n = 6 or 10.9%), and other hepatic malignancies (n = 12 or 21.8%; Table 1).

Recipient Status at Transplant

The mean MELD score at transplant was 20.6 for the normal-weight group, which was significantly higher than the mean MELD scores for the overweight (18.8, P < 0.01) and obese recipients (18.3, P = 0.03). More normal-weight recipients were ventilated (12.9%), were on renal support (17.7%), had preoperative sepsis (4.2%), and needed superurgent transplantation (18.8%) in comparison with overweight (7.2%, 13.2%, 3.4%, and 15.1%, respectively) and obese recipients (4.1%, 6.9%, 2.1%, and 6.9%, respectively). There was no difference in retransplant rates between the 5 groups (Table 1).

Graft Survival

The causes of death-censored graft failure are listed in Table 2. There were no difference in death-censored graft survival between the 5 groups in the adjusted Cox proportional hazards model (HR = 0.99, 95% CI = 0.97-1.02, P = 0.75; Table 3 and Fig. 1). Individual comparisons found no differences in graft survival between the normal-weight and underweight patients (P = 0.88), the normal-weight and obese patients (P = 0.97), or the normal-weight and morbidly obese patients (P = 0.51; Table 4). The overweight recipients had better graft survival in the short term and the long term in comparison with the normal-weight recipients (P < 0.01; Table 4).

image

Figure 1. Adjusted Kaplan-Meier curves depicting death-censored graft survival stratified by the recipient BMI grade. The curves were adjusted for all variables significant for graft survival in the multivariate analysis. In addition, donor age, donor BMI, recipient age, recipient sex, recipient MELD score, and cold ischemia time were added to the Cox proportional hazards model.

Download figure to PowerPoint

Table 2. Causes of Death-Censored Graft Failure and Mortality Stratified by the Recipient BMI Grade
 Underweight: BMI < 18.5 kg/m2 (n = 47)Normal Weight: BMI = 18.5-24.9 kg/m2 (n = 643)Overweight: BMI = 25.0-29.9 kg/m2 (n = 417)Obese: BMI = 30-34.9 kg/m2 (n = 145)Morbidly Obese: BMI ≥ 35 kg/m2 (n = 73)Total Failed Grafts or Deaths
  1. NOTE: Some groups of percentages do not exactly equal the totals because of rounding.

  2. a

    Graft-versus-host disease, paracetamol overdose, posttransplant lymphoproliferative disorder, graft hemorrhage, or malignancy elsewhere.

  3. b

    Renal failure, sclerosing peritonitis, pancreatitis, or suicide.

Causes of death-censored graft failure [n (%)]      
Primary nonfunction3 (6.4)14 (2.2)6 (1.4)5 (3.4)3 (4.1)31 (10.3)
Acute/chronic rejection2 (4.3)27 (4.2)10 (2.4)4 (2.8)1 (1.4)44 (14.6)
Biliary complications0 (0.0)5 (0.8)2 (0.5)2 (1.4)0 (0.0)9 (3.0)
Hepatic artery thrombosis0 (0.0)14 (2.2)9 (2.2)0 (0.0)1 (1.4)24 (8.0)
Nonthrombotic infarction0 (0.0)6 (0.9)1 (0.2)1 (0.7)0 (0.0)8 (2.7)
Malignancy within graft3 (6.4)7 (1.1)4 (1.0)1 (0.7)1 (1.4)16 (5.3)
Multiorgan failure/septicemia2 (4.3)9 (1.4)6 (1.4)0 (0.0)5 (6.8)22 (7.3)
Recurrent disease0 (0.0)20 (3.1)14 (3.4)4 (2.8)3 (4.1)41 (13.6)
Other causea0 (0.0)19 (3.0)4 (1.0)3 (2.1)0 (0.0)26 (8.6)
Unknown2 (4.3)45 (7.0)15 (3.6)14 (9.7)4 (5.5)80 (26.6)
Total12 (25.5)166 (25.8)71 (17.0)34 (23.4)18 (24.7)301
Causes of recipient mortality [n (%)]      
Intraoperative death0 (0.0)6 (0.9)3 (0.7)1 (0.7)2 (2.7)12 (2.9)
Cardiac cause0 (0.0)9 (1.4)7 (1.7)3 (2.1)1 (1.4)20 (4.8)
Cerebrovascular accident0 (0.0)6 (0.9)3 (0.7)0 (0.0)0 (0.0)9 (2.2)
Multiorgan failure/septicemia9 (19.1)71 (11.0)41 (9.8)9 (6.2)10 (13.7)140 (33.7)
Respiratory failure1 (2.1)15 (2.3)6 (1.4)1 (0.7)0 (0.0)23 (5.5)
Recurrent disease1 (2.1)16 (2.5)4 (1.0)2 (1.4)1 (1.4)24 (5.8)
Malignancy/posttransplant lymphoproliferative disorder1 (2.1)25 (3.9)14 (3.4)2 (1.4)3 (4.1)45 (10.8)
Pulmonary embolism0 (0.0)2 (0.3)0 (0.0)0 (0.0)0 (0.0)2 (0.5)
Other causeb0 (0.0)8 (1.2)7 (1.7)5 (3.4)0 (0.0)20 (4.8)
Unknown5 (10.6)70 (10.9)26 (6.2)14 (9.7)5 (6.8)120 (28.9)
Total17 (36.2)228 (35.5)111 (26.6)37 (25.5)22 (30.1)415
Table 3. Adjusted Death-Censored Graft Survival and Patient Survival Stratified by the Recipient BMI Grade
 Underweight: BMI < 18.5 kg/m2 (n = 47)Normal Weight: BMI = 18.5-24.9 kg/m2 (n = 643)Overweight: BMI = 25.0-29.9 kg/m2 (n = 417)Obese: BMI = 30-34.9 kg/m2 (n = 145)Morbidly Obese: BMI ≥ 35 kg/m2 (n = 73)Adjusted Cox Proportional Hazards Model
  1. NOTE: The curves were adjusted for all variables significant for death-censored graft survival and patient survival in the multivariate analysis. In addition, donor age, donor BMI, recipient age, recipient sex, recipient BMI, recipient MELD score, and cold ischemia time were added to the Cox proportional hazards model.

Graft survival (%)      
1 year9086908580HR = 0.99
3 years888285706695% CI = 0.97-1.02
5 years8078826450P = 0.75
10 years757746 
Patient survival (%)      
1 year8380848679HR = 1.00
3 years687680787495% CI = 0.98-1.02
5 years5870747066P = 0.63
10 years60726457 
Table 4. Unadjusted (Crude) Kaplan-Meier Log-Rank P Values for Death-Censored Graft Survival and Patient Survival Stratified by the Recipient BMI Grade
Log-Rank ComparisonP Value
Death-Censored Graft SurvivalPatient Survival
  1. NOTE: Bonferroni correction was applied to the log-rank P values for significance (P = 0.05/4 or P < 0.012). The bolded value is significant.

  2. a

    Overweight recipients had significantly better death-censored graft survival than normal-weight recipients, but the corrected P value for the difference in patient survival between the groups was not significant.

Normal weight versus underweight0.880.97
Normal weight versus overweight<0.01a0.03
Normal weight versus obese0.970.27
Normal weight versus morbidly obese0.510.78

Patient Survival

The causes of patient death after transplantation are listed in Table 2. There were no differences in patient survival between the 5 groups according to the adjusted Cox proportional hazards model (HR = 1.00, 95% CI = 0.98-1.02, P = 0.63; Table 3 and Fig. 2). Individual comparisons showed no differences in patient survival between the normal-weight and underweight recipients (P = 0.97), the normal-weight and overweight recipients (P = 0.03), the normal-weight and obese recipients (P = 0.27), or the normal-weight and morbidly obese recipients (P = 0.78; Table 4).

image

Figure 2. Adjusted Kaplan-Meier curves depicting patient survival stratified by the recipient BMI grade. The curves were adjusted for all variables significant for patient survival in the multivariate analysis. In addition, donor age, donor BMI, recipient age, recipient sex, recipient MELD score, and cold ischemia time were added to the Cox proportional hazards model.

Download figure to PowerPoint

Postoperative Complications

The rate of postoperative hemorrhaging needing intervention was 7.5% in the normal-weight group and 0% in the morbidly obese group (P = 0.01). Postoperative chest infections were much more common in the overweight (17.7%, P < 0.001) and obese recipients (15.2%, P = 0.03) versus the normal-weight patients (9.0%). There was no difference in wound infection rates between the 5 groups, but ascitic and drain fluid infections were more common in obese (17.9%, P < 0.001) and morbidly obese (13.7%, P < 0.001) versus the normal-weight recipients (2.2%). There was no difference in urinary infection or bacteremia rates between the 5 groups. The rate of overall infective complications was significantly higher in the overweight (60.7%, P < 0.01) and obese recipients (65.5%, P < 0.01) versus the normal-weight recipients (50.4%). The morbidly obese patients (63.0%) had higher rates of infective complications than the normal-weight recipients, but this was not significant after Bonferroni correction (P = 0.047). The underweight recipients had the lowest rate of infective complications in our cohort (44.7%). There was no difference between the 5 groups in terms of vascular complications, bile leaks, or biliary stricture rates (Table 5).

Table 5. Clinical Outcomes Based on the Recipient BMI Grade
Outcome ParameterUnderweight: BMI < 18.5 kg/m2 (n = 47)Normal Weight: BMI = 18.5-24.9 kg/m2 (n = 643)Overweight: BMI = 25.0-29.9 kg/m2 (n = 417)Obese: BMI = 30-34.9 kg/m2 (n = 145)Morbidly Obese: BMI ≥ 35 kg/m2 (n = 73)P Valuea
Normal Weight Versus UnderweightNormal Weight Versus OverweightNormal Weight Versus ObeseNormal Weight Versus Morbidly Obese
  1. NOTE: Continuous variables are expressed as means and standard deviations, and categorical variables are expressed as numbers and percentages. The bolded values are significant.

  2. a

    P values were considered significant at <0.012 (Bonferroni correction: 0.05/4 = 0.012).

No intraoperative blood transfusion [n (%)]6 (12.8)111 (17.3)93 (22.3)30 (20.7)19 (26.0)0.550.0460.340.08
No postoperative blood transfusion [n (%)]3 (6.4)80 (12.4)64 (15.3)21 (14.5)15 (20.5)0.350.200.490.07
Length of ICU stay (days)1.9 ± 1.83.2 ± 5.13.2 ± 7.13.7 ± 6.04.7 ± 8.90.08>0.990.300.03
Length of ventilation (days)1.5 ± 1.42.4 ± 4.22.5 ± 6.62.5 ± 5.62.7 ± 4.70.140.760.810.57
Length of hospital stay (days)20.0 ± 12.818.0 ± 18.322.4 ± 15.621.3 ± 12.522.4 ± 14.80.46<0.0010.040.047
Postoperative hemorrhage [n (%)]1 (2.1)48 (7.5)29 (7.0)11 (7.6)0 (0.0)0.240.720.860.01
Chest infection [n (%)]6 (12.8)58 (9.0)74 (17.7)22 (15.2)9 (12.3)0.43<0.0010.030.39
Wound infection [n (%)]8 (17.0)127 (19.8)83 (19.9)25 (17.2)11 (15.1)0.85>0.990.560.43
Ascitic fluid/drain infection [n (%)]0 (0.0)14 (2.2)11 (2.6)26 (17.9)10 (13.7)0.610.68<0.001<0.001
Urinary infection [n (%)]1 (2.1)15 (2.3)12 (2.9)5 (3.4)1 (1.4)>0.990.690.39>0.99
Bacteremia [n (%)]3 (6.4)83 (12.9)54 (12.9)16 (11.0)5 (6.8)0.250.920.580.18
Unknown source of sepsis [n (%)]3 (6.4)27 (4.2)19 (4.6)1 (0.7)10 (13.7)0.450.880.04<0.01
Total infective complications [n (%)]21 (44.7)324 (50.4)253 (60.7)95 (65.5)46 (63.0)0.55<0.01<0.010.047
Vascular complications [n (%)]4 (8.5)34 (5.3)13 (3.1)4 (2.8)1 (1.4)0.320.120.200.24
Biliary complications [n (%)]4 (8.5)29 (4.5)18 (4.3)8 (5.5)0 (0.0)0.27>0.990.660.06
Acute rejection [n (%)]16 (34.0)170 (26.4)102 (24.5)29 (20.0)12 (16.4)0.310.520.110.07

The morbidly obese recipients had a longer mean ICU stay than the normal-weight patients (4.7 versus 3.2 days), but this was not statistically significant (P = 0.03). There was no difference in the length of postoperative ventilation between the 5 groups. The postoperative hospital stay was longer for the overweight (mean = 22.4 days, P < 0.001), obese (mean = 21.3 days, P = 0.04), and morbidly obese recipients (mean = 22.4 days, P = 0.047) versus the normal-weight recipients (mean = 18.0 days; Table 5).

Independent Prognostic Factors for Graft and Patient Survival

In the multivariate analysis, ventilation before transplantation (HR = 0.62, 95% CI = 0.42-0.91, P = 0.02), retransplantation (HR = 0.58, 95% CI = 0.41-0.80, P < 0.01), and the year of transplantation (HR = 0.96, 95% CI = 0.92-0.99, P = 0.01) were independent predictors of death-censored graft survival (Table 6). For patient survival, a malignant etiology (HR = 2.09, 95% CI = 1.42-3.05, P < 0.001), retransplantation (HR = 1.59, 95% CI = 1.20-2.11, P < 0.01), and the year of transplantation (HR = 0.95, 95% CI = 0.92-0.97, P < 0.001) were all independent predictors in the multivariate analysis (Table 6).

Table 6. Univariate and Multivariate Analyses of Death-Censored Graft Survival and Patient Survival
 Death-Censored Graft SurvivalPatient Survival
P Value for Univariate AnalysisaMultivariate Analysis†P Value for Univariate AnalysisaMultivariate Analysisb
HR95% CIP ValueHR95% CIP Value
  1. NOTE: The bolded values are significant.

  2. a

    Kaplan-Meier log-rank P value.

  3. b

    Performed with the Cox proportional hazards model.

  4. c

    Being the variable of interest, recipient BMI was forced into the Cox proportional hazards model to test for the survival prognosis.

Donor age (years)<0.011.011.00-1.02<0.010.44   
Donor sex: male0.56   0.44   
Donor BMI (kg/m2)0.13   0.81   
Cold ischemia time (minutes)0.74   0.48   
Alcoholic liver disease0.87   0.88   
Viral etiology0.69   0.32   
Noncholestatic liver disease<0.011.260.96-1.670.09<0.0010.820.65-1.040.10
Malignancy0.020.670.40-1.110.13<0.0012.091.42-3.05<0.001
Recipient age (years)0.10   0.44   
Recipient sex: male0.94   0.88   
Recipient BMI (kg/m2)c0.360.990.97-1.020.750.371.000.98-1.020.63
Recipient MELD score<0.010.990.98-1.010.86<0.0011.000.99-1.020.53
Ventilation before transplantation<0.0010.620.42-0.910.02<0.0011.371.00-1.890.05
Dialysis at the time of transplantation<0.010.900.63-1.290.58<0.0011.150.85-1.560.34
Preoperative sepsis0.52   <0.011.450.95-2.220.09
Superurgent transplantation0.09   0.10   
Retransplantation<0.0010.580.41-0.80<0.01<0.0011.591.20-2.11<0.01
Intraoperative blood transfusion0.06   <0.0011.340.96-1.850.08
Postoperative blood transfusion<0.010.840.65-1.080.84<0.011.130.92-1.390.24
Year of transplantation (1994-2009)0.020.960.92-0.990.01<0.0010.950.92-0.97<0.001

DISCUSSION

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

This is the only reported UK series on BMI and outcomes following liver transplantation. This study has demonstrated that overweight and obese recipients have increased morbidity in terms of infective complications following liver transplantation and, consequently, longer ICU and hospital stays. However, there were no differences in graft or patient survival between the 5 BMI categories. Interestingly, in the individual survival analysis comparing normal-weight and overweight recipients, the latter group had better death-censored graft survival. This study demonstrated no negative impact of being an underweight recipient.

The normal-weight recipients were sicker than the other groups in our study population, had higher MELD scores, and included more patients on ventilation before transplantation and more patients needing preoperative renal support, superurgent transplantation, and retransplantation. Despite the poor preoperative status, the normal-weight patients had significantly fewer complications than the obese recipients, all of whom had a more favorable status at the time of liver transplantation.

None of the registry studies from the United States have reported postoperative complications, and only some of the single-center studies have looked into this in detail.[15-17] In a retrospective review of 277 single-center patients, Sawyer et al.[16] reported increased rates of wound infection (20% versus 4%, P < 0.001) and postoperative death due to multiorgan failure (15% versus 4%, P < 0.001) in severely obese patients (BMI > 35 kg/m2). In a retrospective review of 167 liver transplant recipients, Schaeffer et al.[17] showed higher rates of wound infection (20% versus 4%, P < 0.001), wound dehiscence (40% versus 1.2%, P < 0.001), and incisional hernia formation (30% versus 2.8%, P < 0.001) in severely obese recipients versus obese or nonobese recipients. In keeping with the results of our study, LaMattina et al.[15] in the largest single-center series reported so far (n = 813) looked into the effects of BMI on liver transplantation and showed higher rates of wound infections needing antibiotic treatment (for BMI = 35-40 kg/m2, HR = 4.03, 95% CI = 1.85-8.79, P < 0.001; for BMI > 40 kg/m2, HR = 6.27, 95% CI = 2.81-14.01, P < 0.001), intra-abdominal infections needing antibiotics or interventions (for BMI > 40 kg/m2, HR = 7.21, 95% CI = 1.60-32.45, P = 0.01), and pulmonary infections (for BMI = 25-30 kg/m2, HR = 3.80, 95% CI = 1.10-13.23, P = 0.04; for BMI = 30-35 kg/m2, HR = 3.98, 95% CI = 1.10-14.48, P = 0.04) in recipients with high BMIs.

The increased rate of chest and ascitic/drain fluid infections noted in overweight and obese recipients in our study is not surprising. A high BMI is usually associated with coexistent respiratory illness, an increased risk of atelectasis, poor tissue oxygenation, reduced postoperative mobility, difficult analgesic control, and poor wound management.[22, 24-26] Modifiable risk factors such as cessation of smoking, prophylactic respiratory physiotherapy, early mobilization, adequate analgesic therapy, early removal of drains, and careful wound care in overweight and obese recipients would be of benefit for the reduction of infective complications.

Obesity has been shown to be associated with prolonged postoperative mechanical ventilation and ICU and hospital stays.[27, 28] The current study did not demonstrate any difference in the length of ventilation between obese and nonobese liver transplant recipients, and this is in keeping with the report from Werneck et al.[10] This study noted a trend toward longer ICU and hospital stays with increasing BMI. Most centers have also reported similar increases in lengths of stay and overall increases in resources to manage obese and morbidly obese liver transplant populations.[12, 15, 16] Because of the increased complications and resources needed to treat obese patients, some transplant centers have recommended that obesity should be a relative contraindication for liver transplantation.[18] We do not support this view because the benefits of transplantation in obese patients far outweigh the increased risk of death on the waiting list, and more importantly, our large cohort has shown similar graft and patient survival during long-term follow-up.[21, 22] Although obesity might hinder the rate of improvement of the physical quality of life after liver transplantation, the effects are sustained for up to 30 years after transplantation, and this further strengthens the argument in favor of liver transplantation for obese patients.[29, 30]

In the covariate-adjusted Cox proportional hazards model, there was no difference between the 5 groups in either death-censored graft survival or patient survival. Interestingly, according to an individual comparison of the normal-weight and overweight recipients, 1-, 3-, 5-, and 10-year death-censored graft survival (86%, 82%, 78%, and 75% versus 90%, 85%, 82%, and 77%, P < 0.01) was better for the latter group (Tables 3 and 5). Similarly, the overweight recipients had better 1-, 3-, 5-, and 10-year patient survival (84%, 80%, 74%, and 72%) than the normal-weight recipients (80%, 76%, 70%, and 60%), but the P value was not significant after Bonferroni correction (P = 0.03). The reasons for the better outcomes of the overweight population are likely multifactorial. First, the preoperative evaluation of patients with high BMIs was stringent, and only those who were extremely fit were selected for transplantation. Second, overweight patients were likely to be considered suitable for transplantation if they had no comorbidities or fewer comorbidities than normal-weight patients. Third, BMI was evaluated in this study at listing, and it is possible that a percentage of the overweight patients could have reduced their BMI to normal limits, whereas because of a lack of scrutiny, the normal-weight recipients might have increased their BMI on the waiting list.[31] Fourth, overweight individuals are more likely to reduce their weight after transplantation, whereas normal-weight individuals can put on weight, again because of a lack of medical scrutiny of their weight.[32]

The major strength of our study is that we have used a prospectively maintained database of a large single-center population. This study has made comparisons between normal-weight and other groups for both preoperative characteristics and postoperative outcomes, and this makes the assessment more robust and gives us a meaningful comparison of the outcomes in different BMI groups with respect to standard-weight recipients. Such comparisons with normal-weight recipients are lacking in almost all published studies on this topic. One of the problems with using a database is that the collected information is not necessarily tabulated to the purposes of a particular research question. Along with this, the long study period and the associated protocol changes in patient selection, multiple surgeons involved over the study period, different operative techniques, and postoperative management are likely to have introduced bias into our results. A meaningful comparison of our outcomes with those from different centers and from registry studies is impossible because of the different definitions used by these studies for obesity. In the future, it would be more useful if studies used a standard WHO definition for stratifying these patients instead of using nonstandard cutoffs for BMI stratification. The availability of such pooled information would help in counseling patients preoperatively with appropriate risk figures for both short- and long-term outcomes.

In conclusion, pretransplant obesity is associated with poor post–liver transplant outcomes in terms of increased morbidity, but obese patients have similar graft and patient survival in comparison with normal-weight recipients. Unlike other US registry studies, we did not demonstrate any negative impact of being an underweight recipient. Underweight recipients had lower rates of postoperative infective complications and shorter hospital and ICU stays. The goal of transplantation is not only to obtain better graft and patient survival but also to reduce postoperative morbidity and make appropriate use of the limited available hospital and financial resources. The early identification of overweight and obese patients and the introduction of measures to reduce BMI should be considered to improve outcomes following liver transplantation. Because of the increase in the obese population in the United Kingdom and the burden that it imposes on the National Health Service, it is highly imperative that we develop more specific guidelines to preoperatively optimize and postoperatively reduce morbidity in overweight and obese recipients undergoing liver transplantation.

REFERENCES

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