Hepatocellular carcinoma (HCC) is the fifth most common cancer in the world[
Treatment modalities for HCC depend on patient age and comorbidities, tumor characteristics and degree of liver disease and portal hypertension in addition to other factors such as local expertise and resources[
Previous studies have reported that in the United States, LT for HCC is performed less frequently in non-Caucasians than in recipients of other ethnicities[
To be listed for a LT in the United States and Europe, patients with HCC must fulfill not only strict oncological criteria[
Since all LT candidates have to satisfy similar inclusion criteria, we hypothesized that there should not be differences in short and long-term outcomes among different ethnic groups, and since studies on ethnicity and outcomes of patients undergoing LT for HCC in the United States are lacking, the primary aim of this study was to assess if African American had short- and long-term outcomes similar to recipients of other ethnic groups.
The United network for organ sharing (UNOS) standard transplant analysis and research (STAR) files were used to identify a retrospective cohort of patients who underwent LT for HCC in the United States between 1 Jan 2002 and 30 Jun 2013. The study was conducted and reported per recommendations from STROBE statement[
There has been some controversy regarding the possible reasons why some ethnic groups have inferior survival than Caucasian recipients after LT[
All adults (age ≥ 18 years) undergoing LT for HCC were candidates for this study. No restriction of race, citizenship or UNOS region were applied. Recipients of LT for other primary and secondary malignancies (e.g., cholangiocarcinoma, hepatoblastoma, hemangiosarcoma, neuroendocrine metastasis) were excluded. Other exclusion criteria were: transplants from grafts recovered from living or donors after cardiac death, split grafts, multi-visceral or redo transplants, and LT performed across ABO incompatible blood groups. Additional exclusion criteria were lack of records on short and long-term outcomes, the absence of HCC in the explanted liver or the presence of variables with values that were deemed implausible for adult recipients or for deceased donor LTs[
Variable collected for LT recipients were age at the time of transplant, sex, donor and recipient body mass index (BMI), ethnicity, presence of renal failure requiring hemodialysis before surgery, history of diabetes (either type I or II), mortality within 30-, 60-, 90-day and 1 year after surgery, main cause of death, date of death or date of last follow up, cold ischemia time (h), UNOS region where patients were transplanted. Additional variables collected for the donors were age, sex, height and weight or BMI.
Recipient overall survival was estimated by the difference between the date of transplantation and the date of death from any cause using the Kaplan-Meier method. Censoring was used for recipients who were still alive on 30 Jun 2013, or who were alive at the time of the last follow-up or if they underwent re-transplantation (date of redo LT surgery).
The presence of renal failure requiring hemodialysis prior to LT and history of diabetes (type 1 or type 2 diabetes) were used as 2-level categorical variables (absent or present). Ethnicity was categorized into five groups: Caucasian, African American, Hispanic, Asian and Multiracial including other minorities such as Hawaiian or Native American. The time on the wait list was calculated from the day of listing for LT to the date of surgery irrespective of the length of time that the patient spent in an inactive state. The waiting time was then categorized into four periods: less than 3 months, 3.1-6 months, 6.1-12 months and longer than 1 year. Recipient functional status at the time of LT was measured using the UNOS classification based on the validated Karnowski performance status[
The sample size of patients was fixed due to the retrospective design of this study. Continuous variables were reported by estimates of central tendency (means or median) and spread [standard deviation and interquartile range (IQR)] while frequency and percentages were used for categorical data. Survival analysis was performed using the Kaplan-Meier method[
For the calculation of the hazard ratios (HR), Caucasian ethnicity, female sex, functional status lower than 60%, waiting time equal or less than 3 months, post college or graduate degree were selected as references. Adjusted HR (AHR) were calculated using Caucasian patients undergoing LT as a reference. All statistical analyses were performed using SPSS Statistics for Windows, Version 24 (IBM Corporation, United States). Statistical significance was defined when P values were equal or less than 0.05, and 2-tailed tests were used for all statistical analyses.
During the study period, 9723 patients were recorded in the STAR files as recipients of a cadaveric LT with HCC being the primary indication for surgery. Cold ischemia time longer than 24 h was logged in 13 recipients and 3019 patients had no HCC in their final surgical pathology report of their explanted livers and were excluded. After the additional removal of 643 recipients who had more than 10% of missing data, we identified a cohort of 6048 LT recipients who represented the study population. The average age of the recipients was 58 years and females represented 20% of the cohort. Most patients were Caucasians (67%), followed by Hispanics (14.2%), African Americans (8.8%) and Asians (8.6%). Detailed demographic and clinical characteristics of the study population are summarized in
Demographic and clinical characteristics of the study population (6048 liver transplant recipients)
Characteristics | Value |
---|---|
Age, years, mean, (SD) | 57.9 (6.9) |
Sex, |
|
Female | 1,224 (20.2) |
Ethnicity, |
|
Caucasian | 4,054 (67.0) |
African American | 531 (8.8) |
Hispanic | 859 (14.2) |
Asian | 522 (8.6) |
Multiracial or others | 82 (1.4) |
Recipient BMI, mean, (SD) | 28.5 (5.0) |
Recipient BMI, category, |
|
Underweight | 3 (0) |
Normal weight | 1,535 (25.4) |
Overweight | 2,420 (41.0) |
Obesity class I | 1,342 (22.2) |
Obesity class II | 545 (9.0) |
Obesity class III | 143 (2.4) |
Donor BMI, mean, (SD) | 27.6 (5.4) |
Presence of renal insufficiency requiring dialysis, |
50 (0.8) |
Presence of diabetes, |
|
No | 4,254 (70.3) |
Type 1 or type 2 | 1,742 (28.8) |
Unknown | 52 (0.9) |
MELD score, mean, (SD) | 12.1 (4.5) |
Hospital stay (day), mean, (SD) | 10.6 (13.0) |
Months spent on the waiting list, |
|
0-3 | 2,370 (39.2) |
3.1-6 | 1,245 (20.6) |
6.1-12 | 1242 (20.5) |
Longer than 12 months | 1,191 (19.7) |
Functional status at the time of transplantation, |
|
Less than 60% | 958 (15.8) |
60% or more | 4,795 (79.2) |
Unknown | 295 (4.8) |
Education, |
|
Elementary of middle school (grade 1-8) | 326 (5.4) |
High school (grade 9-12) | 2,333 (38.6) |
College or technical school | 1,376 (22.8) |
Associate or bachelor degree | 805 (13.3) |
Post-college or graduate degree | 338 (5.6) |
Unknown | 870 (14.4) |
UNOS region, |
|
Region 1 | 274 (4.5) |
Region 2 | 605 (10.0) |
Region 3 | 888 (14.7) |
Region 4 | 734 (12.1) |
Region 5 | 968 (16.0) |
Region 6 | 301 (5.0) |
Region 7 | 549 (9.1) |
Region 8 | 483 (8.0) |
Region 9 | 422 (7.0) |
Region 10 | 512 (8.5) |
Region 11 | 312 (5.2) |
Donors' age, years, mean, (SD) | 41.1 (15.9) |
Cold ischemia time, hours, mean, (SD) | 6.7 (2.5) |
BMI: body mass index
When compared to all other ethnic groups, the cohort of African American recipients had a higher percentage of women (26.2%
Demographic and clinical characteristics of the study population stratified by recipient ethnicity: African Americans (8.8%)
Characteristics | Other Ethnicities (no. 5,517) | African American (no. 531) |
|
---|---|---|---|
Age, years, mean, (SD) | 58.0 (6.9) | 57.1 (6.9) | 0.005 |
Sex, |
|||
Female | 1,085 (19.7) | 139 (26.2) | ≤ 0.001 |
Body mass index, mean, (SD) | 28.5 (5.0) | 28.3 (5.2) | 0.593 |
Recipient BMI, Category, |
|||
Underweight | 3 (0.1) | 0 (0.0) | 0.273 |
Normal weight | 1,398 (25.3) | 137 (25.8) | |
Overweight | 2,253 (40.8) | 227 (42.7) | |
Obesity class I | 1,230 (22.3) | 112 (21.1) | |
Obesity class II | 508 (9.2) | 37 (7.0) | |
Obesity class III | 125 (2.3) | 18 (3.4) | |
Donor BMI, mean, (SD) | 27.6 (5.4) | 27.3 (5.6) | 0.260 |
Presence of renal insufficiency requiring dialysis, |
38 (0.7) | 12 (2.3) | ≤ 0.001 |
Presence of diabetes, |
|||
No | 3,867 (70.1) | 387 (72.9) | 0.226 |
Type 1 or type 2 | 1,600 (29.0) | 142 (26.7) | |
Unknown | 50 (0.9) | 2 (0.4) | |
MELD score, mean, (SD) | 12.1 (4.4) | 12.1 (5.0) | 0.794 |
Hospital Stay (day), mean, (SD) | 10.5 (13.1) | 11.5 (12.1) | 0.110 |
Months spent on the waiting list, |
|||
0-3 | 2,141 (38.8) | 229 (43.1) | 0.022 |
3.1-6 | 1,125 (20.4) | 120 (22.6) | |
6.1-12 | 1,142 (20.7) | 100 (18.8) | |
Longer than 12 months | 1,109 (20.1) | 82 (15.4) | |
Functional status at the time of transplantation, |
|||
Less than 60% | 881 (16) | 77 (14.5) | 0.413 |
60% or more | 4,372 (79.2) | 423 (79.7) | |
Unknown | 264 (4.8) | 31 (5.8) | |
Education, |
|||
Elementary of middle school (grade 1-8) | 310 (5.6) | 16 (3.0) | 0.010 |
High school (grade 9-12) | 2,110 (38.3) | 223 (42.0) | |
College or technical school | 1,251 (22.7) | 125 (23.5) | |
Associate or bachelor degree | 750 (13.6) | 55 (10.4) | |
Post college or graduate degree | 314 (5.7) | 24 (4.5) | |
Unknown | 781 (14.2) | 88 (16.6) | |
UNOS region, |
|||
Region 1 | 256 (93.4) | 18 (6.6) | ≤ 0.001 |
Region 2 | 485 (80.2) | 120 (19.8) | |
Region 3 | 804 (90.5) | 84 (9.5) | |
Region 4 | 685 (93.3) | 49 (6.7) | |
Region 5 | 927 (95.8) | 41 (4.2) | |
Region 6 | 295 (98.0) | 6 (2.0) | |
Region 7 | 514 (93.6) | 35 (9.5) | |
Region 8 | 439 (90.9) | 44 (9.1) | |
Region 9 | 382 (90.5) | 40 (9.5) | |
Region 10 | 454 (88.7) | 58 (11.3) | |
Region 11 | 276 (88.5) | 36 (11.5) | |
Donors’ age, years, mean, (SD) | 44.2 (15.9) | 42.6 (16.0) | 0.030 |
Cold ischemia time, hours, mean, (SD) | 6.7 (2.5) | 6.6 (2.5) | 0.393 |
BMI: body mass index
The median follow-up of the cohort was 7.6 years (95% CI: 7.5-7.8). During this period, 2079 patients had died (34.3%), 3762 were censored (62.2%), and 207 patients (3.4%) were lost at follow-up. Overall 30-, 60-, 90-day and 1-year mortality was 1.7%, 2.3%, 3.0% and 8.8% respectively with no statistically significant differences between African Americans and other ethnicities
Analysis of the frequency of postoperative mortality observed in African American patients
Primary cause of death after cadaveric liver transplantation for hepatocellular carcinoma by recipient ethnicity
The primary cause of death, |
Other ethnicities, |
African American, |
|
---|---|---|---|
Cardiovascular | 101 (7.7) | 19 (11.2) | 0.185 |
Graft failure | 150 (11.4) | 28 (16.6) | 1.115 |
Cerebrovascular complications | 16 (1.2) | 3 (1.8) | 0.622 |
Pulmonary complications | 46 (3.5) | 7 (4.1) | 0.766 |
Renal insufficiency | 11 (0.8) | 2 (1.2) | 0.664 |
Multiorgan failure | 95 (7.2) | 26 (15.4) | 0.001 |
Infections | 119 (9.0) | 15 (8.9) | 0.810 |
Hemorrhagic complications | 31 (2.3) | 2 (1.2) | 0.305 |
Malignancy | 408 (31.1) | 26 (15.4) | 0.001 |
Unknown | 333 (25.4) | 41 (24.3) | 0.554 |
Total number (%) | 1310 (100) | 169 (100) | - |
Kaplan-Meier survival function showed that the 5-year probability of survival for all patients who underwent LT for HCC was 69%
Kaplan Meier survival function representing the 5-year overall survival of all patients undergoing liver transplantation in the United States from 1 Jan, 2002 to 30 Jun, 2013
Kaplan-Meier survival functions of patients undergoing liver transplantation for hepatocellular carcinoma in the United States stratified by ethnicity. The probability of 5-year survival was 81% for patients belonging to multiracial or other minorities, 79% for Asians, 73% for Hispanics, 68% for Caucasians and 56% for African American (
At univariate Cox regression analysis, ethnicity, age, history of diabetes and functional status at the time of transplantation were independent predictors of survival after LT. At multivariate analysis, African American ethnicity remained the strongest independent predictor for increased mortality in comparison to Caucasian recipients (reference group) (HR = 1.524; 95% CI: 1.283-1.803;
Univariate and Multivariate Cox proportional hazard model of mortality of patients undergoing cadaveric liver transplantation for hepatocellular carcinoma. The adjusted Hazard Ratio was casculated by including both clinical and socio-demographic characteristics
Chracteristics | Unadjusted HR | 95% CI |
|
Adjusted |
95% CI |
|
---|---|---|---|---|---|---|
Recipient ethnicity | < 0.001 | < 0.001 | ||||
Caucasian (reference) | 1 | 1 | ||||
African American | 1.484 | 1262-1.746 | < 0.001 | 1.524 | 1.283-1.803 | < 0.001 |
Hispanic | 0.799 | 0.679-0.939 | 0.007 | 0.785 | 0.656-0.940 | 0.008 |
Asian | 0.602 | 0.483-0.751 | < 0.001 | 0.618 | 0.485-0.787 | < 0.001 |
Multiracial or Other Ethnicities | 0.610 | 0.360-1.033 | 0.066 | 0.733 | 0.431-1.246 | 0.251 |
Donor age (year) | 1.010 | 1.007-1.014 | < 0.001 | 1.010 | 1.007-1.014 | < 0.001 |
Recipient age (year) | 1.020 | 1.012-1.028 | < 0.001 | 1.019 | 1.010-1.027 | < 0.001 |
Recipient sex (female as reference) | 1 | 1 | ||||
Male | 0.935 | 0.822-1.063 | 0.306 | 0.988 | 0.861-1.134 | 0.697 |
Donor BMI | 1.006 | 0.997-1.016 | 0.193 | 1.002 | 0.991-1.012 | 0.766 |
Recipient BMI | 1.002 | 0.992-1.013 | 0.714 | 0.996 | 0.985-1.008 | 0.788 |
Cold ischemia time (hour) | 1.018 | 0.998-1.038 | 0.082 | 1.013 | 0.992-1.034 | 0.233 |
Presence of diabetes (Type 1 or 2) | 1.182 | 1.029-1.270 | 0.013 | 1.065 | 0.567-2.000 | 0.844 |
Dialysis prior to transplant | 1.263 | 0.731-2.181 | 0.420 | 1.109 | 0.612-2.009 | 0.734 |
MELD score | 1.005 | 0.994-1.017 | 0.370 | 1.002 | 0.989-1.014 | 0.812 |
Functional status at the time of transplantation | < 0.001 | < 0.001 | ||||
Functional status < 60% (reference) | 1 | 1 | ||||
Functional status |
0.752 | 0.617-0.918 | 0.005 | 0.696 | 0.602-0.806 | 0.044 |
UNOS region | 0.165 | 0.149 | ||||
Region 1 (reference) | 1 | 1 | 1 | |||
Region 2 | 1.27 | 0.955-1.690 | 0.101 | 1.216 | 0.897-1.649 | 0.207 |
Region 3 | 1.181 | 0.900-1.551 | 0.230 | 1.216 | 0.903-1.637 | 0.198 |
Region 4 | 0.904 | 0.679-1.203 | 0.489 | 1.010 | 0.742-1.374 | 0.951 |
Region 5 | 0.824 | 0.623-1.089 | 0.174 | 0.906 | 0.669-1.225 | 0.521 |
Region 6 | 0.913 | 0.661-1.262 | 0.582 | 1.097 | 0.778-1.546 | 0.598 |
Region 7 | 0.978 | 0.728-1.315 | 0.883 | 0.989 | 0.723-1.354 | 0.947 |
Region 8 | 0.882 | 0.647-1.204 | 0.430 | 0.983 | 0.708-1.365 | 0.918 |
Region 9 | 1.264 | 0.942-1.695 | 0.118 | 1.171 | 0.854-1.604 | 0.326 |
Region 10 | 1.139 | 0.852-1.522 | 0.381 | 1.217 | 0.887-1.668 | 0.223 |
Region 11 | 1.279 | 0.930-1.758 | 0.130 | 1.286 | 0.909-1.820 | 0.155 |
Waiting time (month) | 0.390 | 0.430 | ||||
0-3 months (reference) | 1 | 1 | ||||
3.1-6 months | 1.050 | 0.914-1.206 | 0.489 | 1.110 | 0.955-1.291 | 0.172 |
6.1-12 months | 0.908 | 0.782-1.054 | 0.206 | 0.995 | 0.842-1.177 | 0.956 |
> 12 months | 0.966 | 0.835-1.118 | 0.643 | 1.085 | 0.922-1.278 | 0.324 |
Education | 0.064 | 0.173 | ||||
Elementary of middle school (grade 0-8) | 0.964 | 0.750-1.240 | 0.512 | 1.146 | 0.877-1.497 | 0.317 |
High school (grade 9-12) | 0.915 | 0.790-1.060 | 0.182 | 0.969 | 0.827-1.135 | 0.697 |
College or technical school | 0.819 | 0.693-0.967 | 0.013 | 0.882 | 0.738-1.053 | 0.164 |
Associate or bachelor degree | 0.831 | 0.686-1.007 | 0.014 | 0.891 | 0.728-1.091 | 0.082 |
Post college or graduate degree (reference) | 1 | 1 |
The adjusted HR (*) was calculated including clinical and sociodemographic variables. Clinical characteristics used for the adjustment were: donor and recipient age, recipient sex, recipient body mass index (BMI), MELD score, history of diabetes and dialysis, functional status. Social characteristics used for the adjustment were the highest level of education obtained by the recipient. The surgical characteristic used for the adjustment was the cold ischemia time. Other characteristics used for the adjustment of the HR were the UNOS region where the transplant occurred and the length of waiting time
Over the past decades, there has been an increasing awareness that cancers have unique mutations in signaling pathways[
Socio-economic conditions are difficult to define and may fluctuate over time[
In a retrospective analysis of 754 patients with HCC eligible for LT at Mount Sinai Hospital in New York between 2003 and 2013, Sarpel
More recently, Moylan
Despite these positive changes, other investigators continued to report that African Americans have the lowest survival rate among all LT recipients for benign conditions[
Therefore, we analyzed only LT recipients with confirmed HCC with the main intent of testing the null hypothesis that after adjusting for clinical and socio-economic factors, African Americans should have short and long-term outcomes comparable to other ethnic groups. When compared to other ethnicities, we found that African Americans had lower education level, were more frequently affected by renal dysfunction requiring dialysis (2.3%
Only 56% of African Americans were alive after 5 years
Patients who undergo LT are only a fraction of the number of patients who are referred but fail selection due to insufficient social support, inability to travel to transplant centers or lack of resources including health-care insurance. And, since most of the transplant centers in the United States use comparable criteria for screening patients with inadequate socio-economic resources, and use the Milan criteria for staging HCC irrespective of patient ethnicity, we advanced the hypothesis that unless there were biological reasons, there should not be significant ethnic differences in outcomes after LT.
Overall the results of this study are not novel, yet there are several methodological differences that distinguish our study from others. First of all, we included only patients who had documented HCC in their explanted livers. Confirmation that all recipients in this study had HCC is important because up to 11% of patients who are diagnosed with HCC by imaging tests without biopsy prior to LT end up having no pathological evidence of neoplastic lesions in their explanted livers[
To the best of our knowledge, our study is also the very first to explore if the causes of death after LT were different between African Americans and other ethnic groups. We found that the primary causes of death were similar between African Americans and other ethnic groups except that African Americans had a two-fold risk dying of multiorgan failure (15.4%
Besides the retrospective design of this study, there are several other limitations that are worth mentioning. Although the STAR files have the advantage of containing data on a very large number of transplant recipients, it does not provide enough granularity on the type of insurance, socio-economic status and other personal information that might be important when trying to analyze the impact of socio-economic factors on recipients outcomes and it is subject to data entry errors and miscoding. It is well known that the introduction of random errors reduces the reliability of studies making significant findings less likely[
The effect of ethnicity on the pharmacokinetic of commonly used immunosuppressive agents is often underestimated. In a study on immunocompetence between African Americans and Caucasians, Nagashima
In conclusion, the findings of our study are several. The first is that the short-term outcomes of African American recipients of cadaveric LTs for HCC are similar to patients belonging to other ethnicities. Second, we confirme that African Americans have the lowest 5-year survival rate among all the ethnic groups after adjusting for several clinical and socio-demographic characterstics. Third, that African American ethnicity and poor functional status at the time of LT are the two strongest predictors of inferior survival.
Previous investigators have suggested that differences in the socioeconomic status might be responsible for the lowest survival observed among African Americans. We recognize that there are many factors that were not accounted in our analysis such as type of health care insurance, household income, serum alpha-feto-protein, number and size of the largest tumor, cellular differentiation and vascular invasion. However, due to similar oncological and socio-economic criteria equally applied across all ethnicities during the evaluation and selection of LT recipients, there might be biological reasons, rather than socio-economic factors responsible for the survival differences observed among ethnic groups undergoing LT for HCC.
Michele Molinari designed the study, performed the statistical analysis and wrote the manuscript, Allan Tsung reviewed the statistical analysis and the manuscript, Subhashini Ayloo designed the study, reviewed the manuscript and the statistical analysis, Patrick Bou Samra revised the manuscript and performed the review of the literature, Naudia Jonassaint designed the study, reviewed the manuscript and the statistical analysis.
Data and materials are available from the corresponding author on reasonable request.
None.
The authors declare that there are no conflicts of interest related to this study.
Not applicable.
Not applicable.
© The Author(s) 2018.