Probiotics in the Prevention of Hepatocellular Carcinoma in Cirrhosis

Research question Does the modulation of gut microbiome with probiotics prevents the development of hepatocellular carcinoma in patients with cirrhosis? The aim of this study is to evaluate the efficacy of probiotics in reducing the incidence of hepatocellular carcinoma in patients with Child Pugh A-B cirrhosis with portal hypertension.

Methods Study design. Phase III interventional, multicenter, randomized, double-blind, placebo-controlled trial.

The following study will be carried out following CONSORT guidelines. Scenario and temporality. The study will be developed in Argentina including patient care centers with the following inclusion and exclusion criteria, from April 1/2019 (recruitment phase of 12 months) until the last enrolled patient completes the follow-up of 3 years from randomization.

Intervention. Test drug (s). Both probiotic and placebo will be presented in the same container, shape and color.

Name of the active substance Each 50 ml bottle contains Lactobacillus casei 3.3 x 107 CFU / day, Lactobacillus plantarum 3.3 x 107 CFU / day, Streptococcus faecalis 3.3 x 107 CFU / day and Bifidobacterium brevis 1.0 x 106 CFU / day (BIOFLORA®, BIOSIDUS SA, Argentina) .

Administration way: Oral 5 ml orally every 12 hours for 10 consecutive days per month. Duration of treatment: 2 cycles of continuous treatment per year during a 3-year period (6 cycles). Each cycle includes three months of treatment, followed by an off period of three months.

Reference intervention. Name of the active substance: Placebo. Administration way: Oral 5 ml orally every 12 hours for 10 consecutive days per month. Duration of treatment: 2 cycles of continuous treatment per year during a 3-year period (6 cycles). Each cycle includes three months of treatment, followed by an off period of three months.

**Likewise, both arms will explicitly request the non-consumption of alcohol and the consumption of a non-hypercaloric Mediterranean diet throughout the study for all patients.

Methodology. Patients who meet the aforementioned inclusion and exclusion criteria, and have successfully completed all screening tests, will be treated with the standard test or treatment medication according to clinical practice guidelines through a 1: 1 randomization scheme.

The randomization will be in stratified in blocks according to previously identified prognostic variables and in order to avoid potential confounding variables and or effect modifiers distributed by chance in an unbalanced way: Child Pugh score B and platelet count <100,000/mm3. The randomization will be carried out according to the assignment by means of numbers previously made by a blinded researcher and independent from clinical characteristics and contact with the patients. All numbers of the randomization will be stored in closed opaque envelopes and will be assigned consecutively according to the order of inclusion in the protocol.

*All the events developed, both primary and secondary throughout the study, will be evaluated by a blinded and independent committee from baseline variables and randomization allocation. Clinical events will be monitored and re-evaluated by this committee, which will consist of 3 investigators.

Exposure variables.

• Clinical Data
• Peripheral blood samples for the measurement of cytokines, LPS and expression of TLR 4.
• Cytokines: The measurement of cytokines will be made from plasma of peripheral blood samples. The levels of the following cytokines will be quantified in pg/ml: IL-6, TNF-α and endotoxemia or bacterial lipopolysaccharide (LPS). The operators of this measurement will be blinded to any clinical, microbiome and treatment allocation. Samples to be obtained from each individual: two 10ml tubes of whole blood, one with EDTA (sodium ethylenediamine tetraacetate) and another without anticoagulant, to be used in routine laboratory studies and in the measurement of cytokines (IL6 and TNF alpha) and endotoxins (LPS). Quantification of plasma cytokines: Commercial reagents of enzyme-linked immunosorbent assay (ELISA) will be used to measure the levels of cytokines present in the plasmas of patients. The concentrations of TNF alpha and IL-6 will be measured with commercial reagents of eBioscience, affymetrix, or a similar commercial reagent available in the market, strictly following the instructions of the respective manufacturers. For the quantification of the cytokines, the curves will be made with the corresponding concentration standards.
• Quantification of endotoxins in plasma or serum: Measurement of bacterial LPS will be performed by Limulus Amebocyte Lysate Assay (LAL), Biosafe Srl chromogenic endpoint reagent or similar, in parallel with the concentration curve of the standards. The manufacturer's instructions will be followed.
• Expression of TLR 4 in peripheral blood mononuclear cells: the expression of TLR 4 in mononuclear cells will be quantified by real-time PCR technique, using specific primers.
• Stool samples for gut microbiome evaluation.

1. Method of Collection: A sample of spontaneous fecal stool will be collected at any time of the day in a supplied manifold avoiding contact with lavatory/odorless materials. Once the sample is collected, the appropriately capped collector will be placed inside a safety bag and inside a kit box. These samples will be stored frozen at -70º. Each fecal sample must be properly labeled. Later, samples will be aliquoted for final processing.
2. Analysis: Fecal stools will be collected to analyze the gut microbiome using the PCR technique, Ilumina, analyzing the V3-V4 regions of 16sRNA at the biotechnology Institute of Rosario (INDEAR, Rosario, Argentina). The 16S ribosomal subunit of rRNA will be analyzed bacterial, amplifying using PCR method (20 cycles) and then a second round of amplification to incorporate multiplex adapters (Barcodes) (10 cycles). In each procedure, two measurements will be made to avoid amplification information biases. The massive sequencing will be carried out on an illumina MiSeq team from HERITAS institute (CIBIC + INDEAR) in 2x300 mode to finally obtain a minimum of 40,000 sequences per individual. The quality of the raw data will be evaluated by FastQC and the information of pair-ends of Illumina will be combined using PEAR (v0.9.10). The readings are then filtered by Q30 quality and the duplicates and PCR chimeras are eliminated using UCHIME. The filtered sequences will be clustered in Operational Taxonomic Units (OTUs) at 97% similarity with their own strategy implemented in QIIME (v1.91). The operators of this measurement will be blind to the clinical, microbiome, and primary event variables.

• The following measurements will be performed every 6 months in each subject enrolled until the end of the study or development of the primary event: weight and height, concomitant medication, complete physical examination, laboratory tests, Child Pugh score, abdominal ultrasound.
• Longitudinal sampling of gut microbiome and cytokines. A second cytokine measurement will be made from serum from peripheral blood and gut microbiome samples in fecal samples in all patients included in the study at the end of the first treatment cycle (3rd month since randomization or 12 weeks) and at the end of the study (last cycle) using the same methodology described above.

Adherence and Follow-up: All the patients will continue in the protocol until the development of the primary event or until the end of it, having completed a post-randomization follow-up of 3 years. Several measures will be carried out to reinforce adherence to treatment in both arms and strategies to avoid loss of follow-up. Adherence to treatment compliance will be defined in 80% of the tablets delivered for both arms.

Sample size calculation and statistical analysis: A two-tailed statistical value will be taken, α or type I error of 5% (p value <0.05) and a power or error type II or β of 0.20 (power of 80%). The risk of HCC in each arm will be compared using Cox regression analysis with Hazard Ratios (HR) and 95% confidence intervals and the number needed to treat to prevent a clinical outcome will be calculated, done according to the intention-to-treat principle. Since the cumulative incidence of HCC from previous data from the literature is 1-6% per year, a relative risk reduction of HCC development during a 3-year period in the intervention arm was estimated to be clinically relevant with a 30% relative difference, equivalent to a HR of 0.7. For each arm, a sample size of 140 subjects, with a total population to enroll of 280 patients, estimated number of primary events in 28.

All statistical analyzes will be carried out using STATA version 14.0.

Analysis of microbiome metadata. The bioinformatic analysis of the data will be done using a customized QIIME pipeline. For the estimation of differential abundance between the groups we use the DESeq2 package.