Effect of DHA on Lipid and Carbohydrate Metabolism Alterations and Body Fat Distribution in HIV Patients Under HAART.

Hypothesis: DHA treatment would be able to revert, at least partially, the lipid disturbances associated with HAART and to improve or at least not to worse fat redistribution associated with HAART, without inducing further derangements to dendritic cells and monocyte functional ability.

Variables: effects of DHA treatment on total cholesterol and its fractions, triglycerides, insulin, fibrinolysis markers, tumor necrosis factor alpha (TNF-α), and fat distribution assessed by anthropometric measurements, bioimpedance, sonography, DEXA and abdominal CT. Effects of DHA administration on phenotype and functional capabilities of dendritic and mononuclear cells. Objective: To determine if treatment with high doses of highly purified docosahexanoic acid (DHA) is able to revert, totally or partly, the lipid disturbances and the fat redistribution associated with highly active antiretroviral therapy (HAART), and the effects on phenotype and function of dendritic cells and monocytes Data analysis: Baseline data will be analyzed for a good balance. Fisher´s exact test will be used to assess differences between categorical variables, Student's t test for continuous variables, and Mann-Whitney test for ordinal variables. The main efficacy variable will be differences between baseline and final values using MANCOVA model taking baseline value as covariable. Efficacy will be assessed by contrasting adjusted means at the end of the study between both groups of treatment. Intention to treat analysis will be used.

The methodology to perform in the centers where the clinical trail will be conducted are the following: After obtaining informed consent, the patient's baseline will be examined with the collection of demographic data, HIV infection status, pharmacological data, anthropometric parameters, and it will proceed to conduct a complementary examination designed to identify and define body fat composition, body fat distribution and to perform a basal analysis including the parameters previously mentioned. It also will be evaluated the hygienic-dietetic habits of the patient with special attention to alcohol intake, physical activity and concomitant medication. This procedure will be carried out by making retrospective diaries from the last 7 days. The patient's daily caloric intake will be quantified by the Nutrilogic ® software (Bio Logic, Barcelona, España). Physical activity will be quantified by the Minnesota scale (Elosua R, Marrugat J, Molina L, Pons S, The MARATHOM Investigators. Validation of the Minnesota leisure time physical activity questionnaire in Spanish men. Am J Epidemiol 1994, 139: 1197-1209).

Body fat composition and body fat distribution: will be done with the following additional tests:

Densitometry or DEXA: Will be determined with the patient in supine position with legs straight and feet together, on a standardized examination table from a densitometry device (Lunar Prodigy, Madison, WI, USA). In the initial study enerarán more faces of the different regions that will be copied and transferred to the images of subsequent studies to reduce the variability of successive determinations. There will be performed a body fat measurement, body composition and assessment of bone mineral density by dual absorptiometry technique. There will be performed a total body "scan" (body composition) or focused on lumbar spine (trabecular bone) and proximal third of the femur (cortical bone) for the evaluation in the last two cases of bone mineral density and a possible osteopenia or osteoporosis.

Abdominal CT: Will be obtained by performing a spinal digital radiography, latero-lateral view, locating L4 and getting a tomographic section passing through the center of the vertebral body. Subsequently, there will be a digital calculation of intra-abdominal fat and subcutaneous fat, manually selecting the area of interest. The areas measured will be done in cm2.

Lab analysis: blood samples for biochemical determinations will be obtained in Vacutainer tubes of 10 ml without additives. Blood will be clotted at room temperature for 30 minutes and centrifuged at 2500 rpm for 15 minutes. After separating the serum from the clot (about 4-5 ml), will be divided into 1 aliquot of 1.5 ml (for the study of lipids) and 5-7 aliquots of 0.5 ml aliquots which will be frozen at -20 °C until the biochemical analysis are carried out.

Plasma lipids, glucose, glycosylated hemoglobin, insulin and C-peptide: measurements of total cholesterol and triglycerides will be performed using automated trading methods (enzyme colorimetry, Roche Diagnostics) adapted to the automated autoanalyzer Hitachi 911. The determination of apolipoproteins AI, B and C-III, lipoprotein (a) will be held mediate commercial methods (immunoturbidimetric assay, Roche Diagnostics and Wako Chemicals) adapted to the automated autoanalyzer Hitachi 911. The determination of cholesterol lipoprotein fractions (Lipid, ELIP) (VLDL, LDL, HDL) will be performed by the following methods: VLDL separated by ultracentrifugation of flotation, cholesterol HDL determined by a direct method (Roche Diagnostics) and cholesterol LDL determined by difference in total cholesterol and VLDL and HDL cholesterol fractions. Glucose: by a commercial method (enzyme colorimetry, Roche Diagnostics) adapted to the automated autoanalyzer Hitachi 747. Serum insulin concentration will be determined by non-competitive chemiluminescent enzyme immunoassay (Immulite 2000TM, Diagnostic Products Corp., Los Angeles, CA, USA). The serum concentration of C-peptide is determined by non-competitive chemiluminescent enzyme immunoassay (Immulite 2000TM, Diagnostic Products Corp., Los Angeles, CA, USA).

Determining the size of LDL particles: The size of LDL particles will be determined from total plasma using acrylamide gradient gel electrophoresis under nondenaturing conditions. Will be used a standard of four known LDL diameter bands, previously determined by electron microscopy.

Determination of fatty acids in plasma: Treatment compliance will be corroborated by determining the fatty acid composition of plasma lipids by gas chromatography. Plasma total lipids will be extracted by the method of Bligh and Dyer with chloroform: methanol (2:1, V: V). Methyl esters of fatty acids will be obtained by transesterification with boron trifluoride in methanol at 80 °C for 60 min. which allows a high recovery of all the lipidic compounds including PUFAs. Aliquot containing the methyl esters of fatty acids will be analyzed with a gas chromatography device (Hewlett-Packard model 5890) on a 30 m. RTX-2330 column (Restek, Bellefonte, PA) with an internal diameter of 0.25 mm equipped with a flame ionization detector. The carrier gas will be helium at a pressure of 105 kPa. To the total separation of the different compounds it has provided with two temperature settings: 140 °C-200 °C to 3 °C/min or in two stages of 140°C-180°C to 4°C/min and 180°C-210°C at 2°C/min. The temperature of the injector and detector is 260 °C. The linear response of the detector will be tested periodically with standard mixtures. Typically, two internal standards with different molecular weight (13:0 and 23:0 or 27:0) will be used. Peaks will be integrated with an integrator D-2500 (Hitachi Ltd., Tokyo) and identified by comparison of retention times with standards. When necessary, the identification may be confirmed by mass spectrometry (Hewlett-Packard detector model 5970B) at an ionization potential of 70 eV.

Study of markers of inflammation: blood will be withdrawn in 5 ml EDTA tube. After extraction, the tube will be centrifuged, and plasma collected in eppendorf tubes and frozen at -70°C until cytokines determination. The content of TNF-alpha, IL-1 beta and IL-6 in the plasma samples will be measured by ELISA. Each sample will be analyzed in duplicate using ELISA kits "R \& D Systems" following the specifications that commercial supplier.

Intervention: Patients will be randomized to one of two treatment groups (placebo and DHA), using a centralized randomization list. The study will be conducted according to the double-blind trial methodology, in which neither the patient nor the physician are unaware of the active substance administered to the patients. The randomization procedure will be performed assigning either strategy according to a randomization scheme generated by the module PROC PLAN SAS (version 8.2) in multiple of 2 blocks and following a pattern 1:1 stratified by center. A list of random assignment will generated according to the above mentioned, and as from the same will be issued a sealed opaque envelope, identified by a sequential number and the name of the center. The envelopes shall be deposited in the center of central data management and analysis, which will act as coordinator of the centralized randomization assignment by telephone contact.

Patients will be evaluated for inclusion in the study, and their personal data collected in a specific collection data notebook. After the informed consent procedure and if the patient consents to be included in the study, the centralized randomization assignment center will be contacted. When patient data are registered, the envelope will be opened and the strategy assigned to the patient reported to the investigator. The procedure shall be transcribed into a document contained in the envelope, signing and dating this register the person responsible at the central randomization center.

Capsules with identical external appearance containing oleic acid (placebo) and docosahexaenoic acid at doses of 500 mg per capsule will be previously prepared. The capsules are coated with a protective cover which makes it taste imperceptible to the patient. The dosage of DHA will be of 4 g daily. At the same intervals, patients in the placebo group will take 4 g daily of oleic acid.

Data collection: The data of patients will be collected by a physician devoted entirely to guide patients, coordinating tests to perform, and recording information into a database. Data will be collected in a previously designed specific collection data notebook.

• Sample size assessment to specify the number of participants or participant years necessary to demonstrate an effect.

Estimating sample size and data analysis:

It is expected to be able to show a difference in favor of the active treatment in the main variable, namely, reduction in triglycerides values at the end of the follow-up period (48 weeks) of at least 20% in magnitude, in the comparison of the means adjusted by a covariance analysis. According to the estimated standard deviation of 86 from our own data, consistent with literature data (AIDS 1999, 13:1424-1425), and the magnitude of the minimum effect pre-defined as relevant (20% reduction), it would be enough to some 29 evaluable patients per group. It is intended to recruit more patients as forecast of possible expected losses of about 10%, or in other words, 33 subjects per group.

Re-estimation of the sample size:

Since variability is estimated from data from other situations (pre-treatment baseline data), and there is no concrete data on the experimental situation (variability after DHA treatment), it is predefined in this protocol to perform a re-estimation of the sample size when will be obtained a follow up of the 50% after 4 weeks of treatment of the triglycerides values. To this objective data from the main variable, triglyceride levels at baseline and after 4 weeks post-treatment, along with the random assignment but masking the codes will be provided to a second independent statistical group (Catalan Institute of Oncology, Department of Epidemiology and Cancer Registration. Statistician: Dr. Victor Moreno) respect to central statistical analysis of this study (Laboratory of Biostatistics and Epidemiology (LBE), Section Clinical Trials, Autonomous University of Barcelona. Statistician: Dr. Ferran Torres). This new group will perform the analysis, and will only reveal variability data to recalculate the sample size. If the sample size were less than expected, it will be maintained according to the initial design. Under no circumstances significance tests will be performed or the study will be prematurely stopped secondarily to re-estimate variability. According to this method, you do not need any adjustment of multiplicity.

• Plan for missing data to address situations where variables are reported as missing, unavailable, "non-reported," uninterpretable, or considered missing because of data inconsistency or out-of-range results Treatment of missing values

The main analysis by intention to treat will require using all subjects without relevant deviations from the selection criteria, whom they have taken at least one dose of the studied medication and who have evaluation of the main variable prospectively, at least at three months of the follow up. Under this assumption, it is expected that in a 1-year study follow up find subjects with measurements not made. Therefore it is set out the imputation of missing values of the main variable as follows:

• If the lack of data is due to problems of efficiency, or the cause is unknown, shall be attributed the 95% percentile value for the study group.
• In case of missing data for other reasons, will be implemented the LOCF (Last Observation Carried Forward) method provided the availability of a value of at least three months after its baseline measurement.

The detailed procedure will be included in the statistical report and it is planed evaluate the consistency of the method before closing the database at a meeting at which anonymous listings will be facilitated for its evaluation 'Blind Review' where the procedure may be re-evaluated (CPMP/EWP/1776/99: Points to Consider on Missing Data).

Difficulties and limitations of the study:

The number of patients and the peculiar management that they will have, due to the additional tests performed, make the study inevitably complex. However, there will be a physician who permanently coordinates all patients and all tests performed. In addition, the research team is wide with concrete and synergistic functions for each researcher. Finally, researchers constitute a consolidated group and are experienced in previous studies with similar requirements.

• Statistical analysis plan describing the analytical principles and statistical techniques to be employed in order to address the primary and secondary objectives, as specified in the study protocol or plan.

Statistical analysis Descriptive analysis

The following descriptive indexes will be detailed in the statistical report according with the nature of the variables:

• In continuous variables: Mean, CI 95%, SD, minimum, P25, median, P75, max, N and missing. Per group and globally.
• In categorical variables: % total with respect to the column, N in each category. Per group and globally.
• In ordinal variables: They will be described with two tables: one with descriptive parameters of the continuous variables and the other with categorical variables.

In open fields or when are interested to list all cases for one or more variables, will be described by lists.

Inferential analysis The statistical treatment of the primary and secondary efficacy variables will be specifically described in a section below. For all other variables, the appropriate hypothesis test will be applied according to its nature: Fisher exact test for categorical variables, the Student t test for continuous variables, Mann-Whitney U test for ordinal variables.

Basal homogeneity between groups An analysis of the baseline comparability of treatment groups will be established regarding demographic variables, risk factors and clinical characteristics at the time of inclusion.

Main efficacy analysis The main variable of efficacy will be evaluated by comparing the difference between baseline and end values of the different treatments using ANCOVA model with the baseline value of the test as covariate. Adjusted means were calculated and confidence intervals at 95% (95%) by the ANCOVA model. As the main decision criterion of efficacy the contrast between the adjusted means at the endpoint of the study between the two treatment groups will be performed. Additionally and as an exploratory intent, shall be evaluated the compared effect at the different times of observation. No multiplicity adjustments will be made due to the exploratory nature of this second analysis.

In case of relevant deviations in the cases pursuant of the parametric tests, it will be applied a nonparametric approach using transformations to ranges of the corresponding variables. The main analysis of the efficacy will be done with the population intended to treat. However, as evidence of sensitivity and robustness of the results, the analysis under the protocol approach will be provided.

Tolerability and acceptability tests All patients who have been randomized and treated will be included in the safety analysis. The frequency of appearance of adverse events in the two groups will be compared using the Fisher's exact test.

Level of significance In all proposed statistical tests the level of significance required shall be the conventional (p < 0.05 bilateral). No provisions have been made to perform any multiplicity adjust except in the intermediate analysis.