We carried out a cross-sectional study in 97 patients with advanced HCV-related cirrhosis who were selected from the ESCORIAL cohort (see Acknowledgements), which is a prospective cohort of patients with advanced HCV-related cirrhosis initiating anti-HCV therapy with all-oral DAAs at four tertiary referral hospitals in Madrid, Spain. All patients were enrolled between January and December 2015.
The inclusion criteria of the ESCORIAL cohort were: 1) plasma HCV RNA detectable by polymerase chain reaction; 2) one or more clinical criteria related to advanced cirrhosis (prior history of liver decompensation (ascites, bleeding esophageal varices, hepatic encephalopathy), or liver stiffness measurement (LSM) ≥25 kilopascals (kPa), or hepatic venous pressure gradient (HVPG) ≥10 mmHg, or CTP ≥7); 3) Initiation of all-oral DAA therapy; 4) a biological sample to carry out immunological assays. The exclusion criteria were: i) the previous diagnosis of hepatocellular carcinoma, ii) hepatitis B virus coinfection. The presence of HIV infection was not an exclusion criterion for the study.
In the present study, we only included patients with advanced HCV-related cirrhosis at baseline, when they had not yet started the HCV treatment. The ESCORIAL study included 112 patients, but 15 of them did not have a plasma sample at baseline, leaving only 97 patients available for the study (32 HCV-monoinfected patients and 65 HIV/HCV-coinfected patients.
The ESCORIAL study was conducted according to the Declaration of Helsinki, and the Research Ethics Committee of the Instituto de Salud Carlos III (CEI PI 41_2014) approved this study. Written informed consent was obtained by all the participants in the study.
Clinical and laboratory data were recorded using a standard database via an online form within each center, which satisfied local requirements of data confidentiality. This process was monitored to verify that all the information in the database was consistent with the patient’s records.
LSM was evaluated by trained operators by transient elastography (FibroScan, Echosens, Paris, France), as we previously described24, and results were reported in kPa, with a range of 2.5 to 75 kPa. The CTP score was calculated from five factors (total bilirubin, albumin, international normalized ratio, ascites, and encephalopathy) and range between 5 and 15 points25. CTP values serve to classify the patient into one of three severity classes of liver cirrhosis: A – Least severe liver disease (5–6 points), B – Moderately severe liver disease (7–9 points), and C – Most severe derangement (10–15 points). All HIV/HCV-coinfected patients were on ART and had undetectable plasma HIV viral load (<50 copies/mL) at least one year before the study.
Enzyme-linked immunosorbent assays
The Spanish HIV HGM BioBank collected plasma samples, which were stored until use at –80 °C. We evaluated plasma biomarkers by ProcartaPlex multiplex immunoassay (Bender MedSystems GmbH, Vienna, Austria) according to the manufacturer’s specifications using a Luminex 200 analyzer (Luminex Corporation, Austin, TX, United States). The plasma biomarkers measured by multiplex ELISA were: i) inflammatory response: IFN-γ-inducible protein 10 (IP-10), monocyte chemoattractant protein-1 (MCP1), IL-8, IL-1β, IL-18, IL-6, tumor necrosis factor-alpha (TNF-α), interleukin-1 receptor antagonist (IL-1RA), soluble receptor activator of nuclear factor- kappaB ligand (sRANKL) and osteoprotegerin (OPG); ii) endothelial dysfunction: soluble vascular cell adhesion molecule 1 (sVCAM-1), soluble intercellular cell adhesion molecule 1 (sICAM-1); and soluble tumor necrosis factor receptor 1 (sTNF-R1); iii) coagulopathy: plasminogen activator inhibitor-1 (PAI-1) and d-dimer; iv) angiogenesis/fibrosis: vascular endothelial growth factor A (VEGF-A) and soluble receptors for vascular endothelial growth factor (sVEGF-R1). In these assays, a high proportion of the analyzed samples were below the lower limit of detection (LOD), and the analysis software censored calculated biomarker levels. The measured fluorescence intensity (FI) values are an alternative to alleviate the concern of determining levels of LOD. Because of this, we did the data analysis using the raw FI values, without subtracting blank, as a relative quantification of the analyte abundances26. With this approach, there were no missing values, it was not necessary to specify a LOD, and we can analyze the low FI signals, which added more statistical power to the data analysis26,27. All measured FI (arbitrary units, a.u.) values were normalized using log10 transformation (log10 transformed).
The plasma biomarkers measured by simple ELISA were lipopolysaccharide-binding protein (LBP; (R&D Systems, Minneapolis, USA), sCD14, and fatty acid-binding protein 2 (FABP-2) (Raybiotech, Georgia, USA)). The lipopolysaccharide was evaluated by a Limulus amebocyte lysate chromogenic endpoint ELISA (LPS; Hycult Biotech, Uden, The Netherlands).
The statistical analysis was performed with Stata 15.0 (StataCorp, Texas, USA) and Statistical Package for the Social Sciences (SPSS) 22.0 (SPSS INC, Chicago, IL, USA). All p-values were two-tailed, and statistical significance was defined as p < 0.05.
For the descriptive analysis, categorical variables were analyzed by the chi-squared test or Fisher’s exact test, as required, and the Mann-Whitney test was used to analyze continuous data.
In this study, the outcome variables were the severity of liver cirrhosis, evaluated with the CTP score, and the presence of severe cirrhosis with Child-Pugh B (CTP 7–9). For the statistical association analysis, Generalized Linear Models (GLM) with a gamma distribution (log-link) were used to analyze the relationship among plasma biomarkers and the CTP score. Besides, GLM with binomial distribution was used to analyze the differences in plasma biomarkers between study groups. These tests give us: i) the arithmetic mean ratio (AMR) and the odds ratio (OR), and ii) significance levels (p-values), which were corrected for multiple testing using the false discovery rate (FDR) with Benjamini and Hochberg (q-values) procedure to reduce the risk of spurious results. GLM models were also adjusted by clinical and epidemiological co-variables: age, gender, smoker, alcohol intake, intravenous drug user (IVDU), previous IFNα therapy, statins treatment, HCV genotype, and log10 HCV RNA. Each plasma biomarker was included by forced entry (Enter algorithm), and the most significant co-variables were selected by a stepwise algorithm (at each step, co-variables are considered for entry with a p-value <0.20), allowing to avoid the over-fitting of the regression.
The accuracy of the biomarkers to separate the study groups was evaluated by the area under the ROC curve (AUC-ROC). Youden’s index was used to select the best cut-off.
Ethics approval and consent to participate
The study was conducted in accordance with the Declaration of Helsinki and patients gave their written consent. The Institutional Review Board and the Research Ethic Committee of the Instituto de Salud Carlos III (ISCIII) approved the study.