Randomized Study Comparing Switching to Raltegravir-based Antiretroviral Versus Maintaining Any Other Antiretroviral Therapy in HIV Monoinfected Patients Impact on Fatty Liver and Liver Fibrosis Assessed by Noninvasive Diagnostic Methods

Current Non-invasive tools for liver fibrosis and fatty liver/NASH, including Fibroscan/CAP and serum cytokeratin 18 (CK-18), are accurate and ideal for screening and monitoring. To date, there has been no study using these accurate non-invasive tools to prospectively evaluate the effect of switching cART to raltegravir in HIV mono-infected patients with suppressed HIV RNA who have evidence of fatty liver and/or significant liver fibrosis in order to determine whether such a strategy may slow fibrosis and steatosis progression rates.

Once a candidate for screening has been identified, study details will be carefully discussed with the subject. The subject will be asked to read and sign the approved informed consent form prior to any assessments being performed. Self-administered questionnaires will collect information on Alcohol use and Disorders Identification Test (audit C), drug use. chart review, (demographics, medication medical history).

Blood samples will be obtained for clinical laboratory evaluations (as per standard of care) at baseline, and all follow-up visits, plus serum CK-18 that will be obtained at all follow-up visits. Routine hematology and chemistry evaluations will include: complete blood count (CBC) with differential, CD4 and CD8 cell counts, viral load, C-reactive Protein, D-dimer, and measurements of hepatic, pancreatic and renal function, fasting insulin, creatine kinase, glucose and lipid profiles. Plasma HIV RNA levels will be measured with a lower limit of detection of 50 copies/mL. Clinical laboratory evaluations for hematology, biochemistry, CD4/CD8, viral load and serology will be performed at the chest institute as per local practices. Fibroscan will be performed on a fasting patient. The probe transducer will be placed on the skin, between the rib bones at the level of the right lobe of the liver. An experienced operator (> 500 examinations before the study), will perform the Fibroscan and the software will calculate the median . The standard M probe will be used in all patients. The XL probe will be used in obese patients (BMI>30 Kg/m2) or if the M probe fails to provide accurate results. CAP examination will be performed at the same time to diagnose fatty liver. A valid Fibroscan result will be defined by 10 validated measures and IQR < 30% of the median. The following cut-off values will be applied to diagnose fibrosis by Fibroscan: 1) 8 kilopascal (kPa) for significant liver fibrosis; 2) 13 kPa for cirrhosis. Fatty liver by CAP will be diagnosed with the following cut-off values: 1) 237.8 dB/m for mild steatosis; 2) 260 dB/m for moderate steatosis; 3) 292.3 dB/m for severe steatosis.

Subject Identification (ID) numbers will be assigned sequentially to each subject who is eligible to participate in the study. The Subject ID Number will consist of a three digit centre number and a three digit number according to chronological order of entry. If a patient discontinues from the study the Subject ID Number will not be reused. The Subject ID Numbers will be used for identification of subjects in the source documentation, and laboratory samples. This will ensure that subject data and laboratory samples leaving the study sites will be identified and tracked.

Randomization to study groups will be done using a computer-generated random allocation with stratification by gender. Summary statistics (mean, median, standard deviation, interquartile range) will be obtained for clinical and demographic data collected between baseline and study conclusion. The fatty liver and liver fibrosis scores by CAP and Fibroscan, respectively, at month 24 will be expressed as a function of exposure group, adjusting for baseline values of each outcome using an analysis of covariance (ANCOVA) model. The average difference in change of steatosis and fibrosis scores between the two groups and associated hypothesis tests will be obtained from the ANCOVA model. All analyses will be based on an intention-to-treat basis, that is, patients' data will be analyzed as randomized. To assess the sensitivity of inferences to missing data, we will generate plausible values for missing observations using multiple imputations (MI), which is a relatively objective methodology for handling missing data in longitudinal studies. In MI, multiple plausible values are generated for each missing item from the joint distribution of observed data using a Bayesian framework, and the results are pooled using established rules while simultaneously accounting for uncertainty in the imputation process. Multivariate logistic regression analysis will be used to explore factors associated with non-invasive diagnosis of fatty liver/NASH and significant liver fibrosis, including patient-related (age, gender, BMI), drug-induced (time on and type of cART), metabolic (diabetes, insulin resistance, lipid profile, hypertension), HIV-related (HIV viral load, CD4/CD8). Pearson correlation coefficient will be used to measure the association between CK-18 and Fibroscan/CAP values, BMI, transaminases. All hypotheses tests will be conducted at 5% level of significance (2-sided). Analyses will be performed using R program for Windows Release 2.13.1.