Page 195 - Read Online

Page 4 of 15 Molinari et al. Hepatoma Res 2018;4:56 I http://dx.doi.org/10.20517/2394-5079.2018.71

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).

Covariates used for Cox regression analysis
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 [36-38] . Recipient functional status was reported in the STAR files in 10%
increments with 10% representing a patient who was moribund to 100% who represented a fully active and
normal individual without complaints and no evidence of disease. Patient functional status was used as a
two-level categorical variable: less than 60% and 60% or higher. Recipient educational level was stratified into
six categories: elementary or middle school (grade 1-8), high school (grade 9-12), college or technical school,
associate or bachelor degree, post college or graduate degree. BMI was estimated using the World Health
2
Organization (WHO) formula: weight (kg)/height (m ). The WHO definition of overweight and obesity were
2
used to classify recipients and donors in three categories: normal weight (BMI 18.5-24.9 kg/m ), overweight
2
(BMI 25-29.9 kg/m ), obese (BMI ≥ 30 kg/m ). Obesity was further classified as class I (BMI 30-34.9), class II
2
(BMI 35-39.9) and class III (BMI ≥ 40). Data for different BMI classes were not adjusted for the presence
of ascites as the quantitative contribution of this to the patients’ BMI was not reported in the STAR files.

Statistical analysis
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
[39]
was performed using the Kaplan-Meier method and after assessing that the assumptions of the Cox model
were met, proportional hazard model analysis was used to assess the effect of predictors of survival after LT.
Pre-transplant characteristics utilized in the Cox regression model were selected a priori. Donor variables
used as covariates for proportional hazard model were: age and BMI. Recipient variables used as covariates
for Cox regression model were: age, sex, the presence of type I or II diabetes, need for dialysis prior to LT,
level of education, BMI, time spent on the wait list and functional status. Survival analysis was also adjusted
for cold ischemia time and for the UNOS region where the transplant surgery was performed. The UNOS
region 1 was chosen as the reference category and the follow-up time was restricted to 5 years after LT. Since
previous studies suggested that African Americans had the lowest post LT survival among all the ethnicities,
we compared patients of African descent to patients belonging to other ethnicities.

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.

RESULTS
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