Plasma Adiponectin and Hepatocellular Carcinoma Survival Among Patients Without Liver Transplantation

Materials and Methods

Study subjects, epidemiologic/clinicopathological data and biospecimens' collection. A hospital-based prospective study was conducted at Columbia University Medical Center (CUMC) to recruit HCC patients (>18 years older) from October 2008 to July 2014 who were followed up until March 2016. HCC patients either pathologically diagnosed or who met the criteria of the American Association for the Study of Liver Diseases (18) were recruited, while those who had any previous malignancy within the last 5 years, previously underwent liver transplantation, received any systemic cancer therapy, with CP class C disease or clinically significant ascites were excluded from the current study. Finally, a total of 172 HCC cases were eligible and analyzed for leptin and adiponectin. The research project was approved by the Columbia University Institutional Review Board and written informed consent was obtained from each participant.

All subjects completed a short, self-administered epidemiological questionnaire to collect demographic information on age, gender, race/ethnicity, place of birth of self and parents, education, occupation, active and passive smoking, alcohol consumption and family history of cancer, etc. A physical exam was performed to measure height, weight, waist and hip. Information on hepatitis B virus (HBV) and hepatitis C virus (HCV) infection and clinicopathological features including AFP levels, anti-viral treatment, cirrhotic status, Milan criteria, CP score, tumor stage, tumor size and survival status were obtained from the medical and pathological records. Treatments patients received during the follow-up period were also obtained from the medical records.

A 30 ml fasting morning blood sample was collected at the time of enrollment and then processed by the Biomarkers Shared Resource of Herbert Irving Comprehensive Cancer Center at CUMC according to a standardized protocol. All samples were stored in −80°C freezers connected to a telephone alarm system for future biomarker measurement.

Laboratory assays. Human leptin and adiponectin levels were determined in duplicate using commercial ELISA Kits (ThermoFisher Scientific, Camarillo, CA, USA). Leptin is a hormone produced by the OB gene and secreted by adipocytes (19). Both natural and recombinant forms of leptin can be detected by the ELISA kit according to the manufacturer's protocols. A serial dilution of leptin (from 15.6 to 12,000 pg/ml) was used to generate a standard curve. The intra- and inter-assay coefficients of variations (CVs) were 2.6%, and 3.9%, respectively. Similar processes were used to detect adiponectin with a standard curve covering from 0.5 to 32 ng/ml. The intra- and inter-assay CVs were 4.0% and 6.3%, respectively. Laboratory personnel were blinded to study hypotheses and outcomes. HBV surface antigen (HBsAg) and HCV (anti-HCV) status were determined by ELISA Kits (BioChain Institute Inc., Hayward, CA, USA) if not available from the medical records.

Statistical methods. The primary outcomes of interest were HCC survival status (dead vs. alive) and time from diagnosis to death (years) or date of last contact. Mortality data were obtained from medical records and confirmed with the National Death Index. The main predictors for outcomes were plasma leptin and adiponectin levels that were categorized by the median. Because we completed leptin and adiponectin measurements of all samples at two time points by different assays (140 for assay 1 and 32 for assay 2), we did not consider these two predictors as continuous variables in the statistical analysis. We separately dichotomized them into a single binary variable by the medians in each time point to exclude the potential batch effect. The medians of leptin were, respectively, 7.9 and 17.1 ng/ml for time-points 1 and 2; the medians of adiponectin were, respectively, 13.1 and 11.9 μg/ml for time points 1 and 2. The covariates included age group (≥62.6 vs. <62.6 years), gender (male vs. female), race/ethnicity (Non-Hispanic White vs. others), HBsAg (positive vs. negative) or anti-HCV (positive vs. negative), history of alcohol drinking (yes vs. no), cigarette smoking (ever vs. never), diabetes mellitus (DM) (yes vs. no), body mass index (BMI) (≥30 vs. <30 kg/m²), waist to hip ratio (≥0.953 vs. <0.953), CP class (B vs. A), metastasis (yes vs. no), Barcelona Clinic Liver Cancer (BCLC) stage (B, C vs. 0, A), Milan criteria (outside vs. within), AFP levels (≥48 vs. <48 ng/ml), liver transplantation (no vs. yes).

Categorical variables expressed as frequencies and percent values were analyzed by the Chi-squared test to show the associations between leptin, adiponectin and demographic and clinicopathological parameters, as well as with overall HCC survival and liver transplantation status. The univariate survival curves were calculated using the Kaplan-Meier method and the significance was determined by the log-rank test. In the model, death was the event and follow-up duration was the time variable. The follow-up period (years) for each patient was estimated starting from the date of HCC diagnosis to death or the date of last visit. Variables with significance (p<0.05) from each univariate analysis were entered into the multivariate Cox proportional hazard model to estimate the overall prediction of death adjusted for other variables in the multivariate model. Starting with the full model, stepwise backward elimination was used as a variable selection procedure. A nominal α level of 0.20 was the cut-off for removing variables. Results are expressed as adjusted hazard ratios (aHR) with 95% confidence interval (CI). Because liver transplantation is the most important favorable predictor for HCC survival, few variables can achieve significance in the Cox proportional hazard model if liver transplantation is included as a parameter. Thus, we also conducted additional subgroup analyses among those who could not undergo liver transplantation. All statistical analyses were performed using Statistical Analysis System 9.4 (SAS Institute, Cary, NC, USA).