HIV Articles  
Pravastatin Improves Fat Loss/Lipoatrophy  
  "Effect of pravastatin on body composition and markers of cardiovascular disease in HIV-infected men-a randomized, placebo-controlled study"
Note from Jules Levin: independent & separate studies have found improvement in lipoatrophy or fat loss occurred after switching to abacavir or tenofovir, use of pioglitazone (without d4T), and use of pravastatin. Perhaps, implementing all three together might result in additive and significant reversal of fat loss. Although these studies showed small, about 10%, improvement in fat loss associated with each approach, taken together perhaps improvement using all three methods simultaneously might improve fat loss 30%, hopefully.
AIDS: Volume 20(7) 24 April 2006 p 1003-1010
Mallon, Patrick WGa,b; Miller, Johnb; Kovacic, Jason Cc; Kent-Hughes, Juliab; Norris, Richardb; Samaras, Katherined; Feneley, Michael Pc; Cooper, David Aa,b; Carr, Andrewb
From the aNational Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, Australia bHIV, Immunology and Infectious Diseases Clinical Services Unit, St Vincent's Hospital, Sydney, Australia cDepartment of Cardiology, St Vincent's Hospital, Sydney, Australia dGarvan Institute of Medical Research, Sydney, Australia.
Objectives: To determine the effect of the 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor, pravastatin, on markers of cardiovascular risk and lipodystrophy in HIV-infected, protease inhibitor (PI)-treated men with hypercholesterolaemia.
Methods: A randomized, placebo-controlled, 16-week study was carried out on 33 HIV-infected, hypercholesterolaemic men (fasting total cholesterol > 6.5mmol/L) on PI-containing therapy. Patients commenced dietary assessment and advice at week 0 and were randomized to 12 weeks pravastatin (40 mg each night) or placebo from week 4. The primary endpoint was the time-weighted change (TWAUC) in total cholesterol from week 0. Secondary endpoints included TWAUC cholesterol from week 4 (start of pravastatin), total and regional body fat, fasting lipids, glucose, insulin, and markers of cardiovascular risk.
Of 33 men randomized (pravastatin n = 16, mean age 48 years), 31 completed the study. Groups were matched for baseline cholesterol and body composition. Although there was no significant between-group difference in TWAUC cholesterol from week 0 (pravastatin -0.6 ± 1.0 versus placebo -0.4 ± 1.0 mmol/L/week; P = 0.8), TWAUC cholesterol from week 4 decreased more in the pravastatin group (-0.8 ± 1.0 versus -0.3 ± 0.9 mmol/L/week; P = 0.04). Neither triglycerides nor dietary intake changed.
Subcutaneous fat increased significantly with pravastatin (+0.72 ± 1.55 versus +0.19 ± 0.48 kg change in limb fat, P < 0.04; +5.2 ± 8.7 versus -1.3 ± 13.7 cm2 change in abdominal subcutaneous fat, P = 0.02).
Apart from homocystine, which decreased in the pravastatin group, there were no significant differences in other cardiovascular, lipid or glucose parameters.
Conclusions: Despite limited effects on cholesterol, 12 weeks use of pravastatin 40 mg each night in HIV-infected men with hypercholesterolaemia resulted in significant increases in subcutaneous fat.

Dyslipidaemia is common in HIV-infected patients receiving HAART [1]. It is characterized by elevations in triglycerides, total cholesterol and low-density lipoprotein (LDL) cholesterol and decreases in high-density lipoprotein (HDL) cholesterol. It is commonly seen as a component of HIV-associated lipodystrophy, a condition of multifactorial aetiology affecting approximately 60% of patients on HAART [2-4].
Several studies have established links between use of certain protease inhibitors (PI) and dyslipidaemia. In healthy volunteers, elevated plasma cholesterol are seen after as little as 2 weeks therapy with the PI ritonavir [5]. The pathogenesis underlying PI-associated dyslipidaemia probably involves abnormalities in several tissues. In hepatocytes, inhibition of the proteasome by ritonavir is associated with accumulation of apolipoprotein B, a major component of plasma lipoproteins [6]. HAART-induced loss of subcutaneous adipose tissue decreases available stores for circulating lipids [7] and may also contribute to increased circulating lipid concentrations. In addition, in a group of seroconverters, HDL cholesterol concentrations decreased postinfection, suggesting that HIV itself also affects cholesterol homeostasis [8].
An important potential consequence of long-term HAART-induced dyslipidaemia is increased cardiovascular disease (CVD). The pattern of dyslipidaemia seen in PI-treated patients is atherogenic [9] and similar to that seen in the metabolic syndrome which has been linked with increased CVD [10]. Changes in endothelial function or peripheral blood atherothrombotic markers, such as fibrinogen and plasminogen activator inhibitor-1 (PAI-1), can act as surrogate markers for cardiovascular risk. Abnormalities in these markers have been shown in HIV-infected patients and increased incidence of myocardial infarction (MI), depending on length of exposure to HAART, has also been observed, with elevated total cholesterol and triglyceride concentrations and decreased HDL cholesterol all being independent risk factors for MI [11]. The increasing age of the HIV-infected population in many developed countries adds to the potential for CVD to become a significant cause of morbidity and mortality [12], making it important to lower cholesterol concentrations, particularly in antiretroviral therapy (ART)-treated patients with multiple non-modifiable risk factors for CVD, such as older age, male sex and a positive family history of CVD.
In the general population 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors or 'statins' lower total and LDL-cholesterol concentrations and increase levels of the LDL receptor, resulting in increased cellular uptake of circulating lipoproteins, and are effective in both the primary and secondary prevention of myocardial infarction [13-15]. Their use in HIV-infected patients is complicated by drug interactions [16-18], as many HMG Co-A inhibitors are metabolized by cytochrome P450 3A4 (CYP 3A4) [19], which is inhibited by PI such as ritonavir. Use of both drugs together increases the concentration of some statins with the potential for increased toxicity. Pravastatin is less reliant on CYP 3A4 for its metabolism and its use has been investigated in the setting of HAART-associated dyslipidaemia in HIV [19].
Studies of statins in the setting of treated HIV infection have shown some beneficial effects [17,20-25]. However, we are not aware of any randomized, placebo-controlled studies of pravastatin for HAART-associated dyslipidaemia. Several studies have shown a beneficial effect of statins on cholesterol [17,20,23] and other markers of cardiovascular risk [24] in patients on PI, but none has examined the use of pravastatin in a cohort receiving predominantly ritonavir-boosted PI and none has assessed the overall effects of this drug class on non-lipid aspects of lipodystrophy, such as body composition.
This is among the first randomized, placebo-controlled studies on the use of statins in the setting of HAART-associated dyslipidaemia that examines all of the relevant clinical and laboratory markers of HIV-associated lipodystrophy, of which dyslipidaemia is an important component. Twelve weeks' treatment with pravastatin resulted in a modest 11% decrease in cholesterol concentrations. Compared to previously published studies of pravastatin use in the setting of HAART, our subjects were slightly older (47 versus 41-44 years [17,20,24]) and all were male (100% versus 77-100% [17,20,23,24]). The effect of pravastatin on total cholesterol in our study (11% reduction) was somewhat less than that reported in previous studies (reductions ranging between 17 and 25% [20,23,24,34]) but is consistent with a recent report of HIV-infected subjects treated with lipid-lowering therapy in which a subgroup receiving pravastatin experienced reductions of 12% [17]. We believe these differences to be related to the higher percentage of subjects in our study receiving HAART regimens containing ritonavir-boosted PI (91%) compared to previous studies (47-61% where data are available [20,23,24,34]) and the higher baseline cholesterol values observed in our subjects [median baseline cholesterol 7.6 (1.7) mmol/l] compared to several similar studies [17,20,24]. Although one could argue that those with greater baseline cholesterol concentrations should have the potential for greater decreases, if these higher baseline cholesterol values reflect a greater perturbation of cholesterol homeostasis, potential therapeutic interventions may be less effective.
An absolute decrease of < 1 mmol/l was observed with pravastatin, and only seven (21%) subjects, spread across both arms, achieved cholesterol values < 6mmol/l with only four subjects, two from each group, achieving a week 16 cholesterol value < 5.5 mmol/l. Together with the absence of non-lipid statin effects, with no significant changes in %FMD or most atherothrombotic markers, these data raise doubt as to the clinical consequences of such small decreases in cholesterol concentrations on the risk of CVD at a population level in HIV-infected patients on PI therapy. However, individual patients with significant numbers of co-existing cardiovascular risk factors may still receive clinical benefit from reductions in total and non-HDL cholesterol, no matter how small.
Homocysteine levels decreased significantly with pravastatin treatment. Although decreases in homocysteine are thought to be cardioprotective and in vitro, statins may modulate release of homocysteine from peripheral blood mononuclear cells [35], decreases in homocysteine with statin use have not been consistently shown in clinical studies of hyperlipidaemia [36] and further investigations are required to confirm this finding in a HIV-infected PI-treated population.
Both the active and placebo groups experienced decreases in cholesterol concentrations during this study. The reasons for the change in the placebo group may be explained either by dietary effect, or a regression to mean, not uncommonly observed in randomized studies. The latter explanation is favoured as no individual dietary component significantly changed over the course of the study.
The changes in subcutaneous fat observed with this study were surprising given that, unlike rosiglitazone, which has been investigated in this setting [37-39], pravastatin use has not been previously reported to have effects on subcutaneous fat mass. However, the finding is consistent across measures (with both CT and DEXA showing changes in subcutaneous and peripheral fat without significant changes in central or visceral fat).
Several mechanisms could underlie the observed increases in subcutaneous fat. As HMG Co-A is metabolized through the proteasome, inhibition of the proteasome by PI could lead to abnormally elevated levels of active intracellular HMG Co-A and blunt the normal negative feedback of accumulated intracellular cholesterol on intracellular HMG Co-A activity. In vitro, depletion of intracellular cholesterol in mouse adipocytes results in upregulation of sterol regulatory element-binding protein 2 (SREBP-2), a nuclear transcription factor which stimulates expression of genes involved in cholesterol syntheses (HMG-CoA synthase) and uptake (LDL receptor) [40]. By directly inhibiting HMG Co-A reductase, rather than relying on proteasomal degradation, cholesterol depletion resulting from statin use could result in increased cell-surface LDL-receptor expression and increased uptake of circulating lipid into tissues such as adipose tissue, causing increase in size of adipocyte depots. Further studies, both in vitro and in vivo are required to test this hypothesis. Although diet may play a potential role, we think this less likely, both for the reasons described above, and as previous studies of dietary intervention alone have been disappointing [41].
Regardless of the mechanism involved, the increases in limb fat observed are clinically significant when compared to other strategies previously used to increase limb fat in treated HIV-infected patients [42]. The strategy of switching away from a tNRTI has resulted in some of the most significant durable changes in limb fat observed to date [43]. When compared to the MITOX study, which observed a mean increase of 0.39 kg over a 24-week period [42], use of pravastatin was associated with a much greater increase, with a median increase of 0.72 kg over a 12-week period of treatment. However, unlike the strategy of switching away from tNRTI to increase limb fat, the use of pravastatin does not place limits on the use of specific antiretrovirals and should be less likely to introduce risks seen with switch strategies such as virological failure or new ART side effects. However, further research is needed to determine if the improvements observed with pravastatin use will continue, or persist in those receiving PI or tNRTI.
This single-centre study does have limitations. Although 12 weeks' therapy with pravastatin had relatively modest effects total cholesterol, it is doubtful whether prolonged exposure to pravastatin would change this outcome, given the rapid effect observed in other hypercholesterolaemic populations [26,27]. The observed changes in limb fat were part of the secondary endpoints and therefore the study was not primarily designed to test the effect of pravastatin on limb fat. This is reflected in the absence of objective body composition data at week 0 when the dietary assessment and advice was started. In addition, that the study was not designed or powered around the secondary endpoints makes it difficult to draw definite conclusions based on specific negative results. Because of the make-up of our clinical population, we did not test the effect of pravastatin in women, non-Caucasian populations, or in those with normal cholesterol concentrations.
Despite this, this study provides valuable information as to the role of pravastatin for the treatment of hyperlipidaemia in HIV-infected men treated with ritonavir-boosted PI. The results suggest that, as a general strategy, the effect of pravastatin on cholesterol concentrations in this patient group is limited. Despite this, the drug is well tolerated and may have a role in CVD prevention in specific at-risk groups, and its effects on limb fat mass, although novel, are encouraging and should be further investigated in larger randomized prospective studies.

Between July 2001 and July 2004, 34 men were screened of which 33 were randomized. Of these, 31 completed the study (Fig. 1). The two post-randomization dropouts occurred in the pravastatin group. One subject withdrew consent at week 4 (prior to start of study drug) and another at week 8 (4 weeks into study medication), both for personal reasons. No subject discontinued study drug due to adverse effects. The groups were well matched for all baseline demographics except age; the pravastatin group being older (Table 1). Of the 33 subjects, 20 (61%) were considered lipoatrophic on physician assessment, of which nine (45%) were randomized to pravastatin. All subjects maintained their antiretroviral regimen and CD4 T-lymphocyte count and HIV RNA did not change between baseline and week 16. With respect to safety analyses, there were no significant changes in serum creatinine, bilirubin, alanine aminotransferase or alkaline phosphatase between weeks 0 and 16 in either group (data not shown).
Lipid and glycaemic parameters
The cohort displayed a dyslipidaemic profile commonly seen with long-term PI-containing HAART, with elevated cholesterol (7.6 ± 1.5 mmol/L) and triglyceride [4.0 (5.3) mmol/l] concentrations. Both groups were well matched for baseline total cholesterol, HDL- and non-HDL-cholesterol (Table 2).
At week 16, serum cholesterol was lower than baseline in both groups (Fig. 2). Although there were no significant between-group differences in TWAUC change from baseline cholesterol [pravastatin -0.6 (1.0) versus placebo -0.4 (1.0) mmol/l; P = 0.8], when measured from week 4 (start of study drug) both total and non-HDL cholesterol decreased significantly in the pravastatin group compared to placebo [-0.8 (1.0) versus -0.3 (0.9) mmol/l; P = 0.04 and -1.0 (0.9) versus -0.4 (0.7); P = 0.01, respectively) (Table 2). Use of pravastatin remained the strongest predictor of TWAUC change from week 4 total cholesterol when corrected for potential confounders [age, AIDS diagnoses, presence of lipoatrophy, current use of non-nucleoside reverse transcriptase inhibitors (NNRTI) or lopinavir, duration of exposure to PI, nucleoside reverse transcriptase inhibitors (NRTI) or thymidine NRTI (tNRTI), and baseline triglycerides, insulin, HOMA and %FMD]. There was no significant correlation between age and TWAUC change in cholesterol. Only seven subjects (four pravastatin, three placebo) achieved week 16 cholesterol levels below 6 mmol/l.
Although HDL-cholesterol, glucose and insulin values were within the normal range, triglycerides were lower in the pravastatin group and, based on the HOMA-IR index, the pravastatin group was more insulin resistant than the placebo group (Table 2). There were no significant, between-group differences in change in HDL-cholesterol, triglycerides, glucose, insulin or HOMA-IR (Table 2) and no correlations between baseline triglycerides or HOMA-IR and AUC change in cholesterol from week 4.
Body composition
At baseline, groups were well matched for body mass index (BMI), total and regional body fat, lean mass and BMI, as assessed by DEXA. On CT, the pravastatin group had more visceral abdominal fat (VAT) than the placebo group, with a correspondingly higher VAT: subcutaneous fat ratio (Table 2). Subjects randomized to pravastatin experienced greater increases in subcutaneous fat compared to the control group (Table 2), with both limb fat on DEXA and subcutaneous fat on CT increasing significantly more in the pravastatin group than control group [+0.72 (1.55) versus +0.19 (0.48) kg; P < 0.04 and +5.2 (8.7) versus -1.3 (13.7) cm2; P = 0.02 respectively). Overall, the pravastatin group experienced 13.7 (27.9)% increase in limb fat versus 4.8 (12.6)% change in the placebo group during the 12 weeks of study drug (Fig. 3). Use of pravastatin remained the strongest predictor of change in limb fat when corrected for potential confounders (age, AIDS diagnoses, presence of lipoatrophy, current use of NNRTI, tNRTIs or lopinavir, duration of exposure to PI, NRTI or tNRTI, time since last exposure to tNRTI and baseline triglycerides, insuslin, HOMA and %FMD). There was no significant correlation between change in limb fat and TWAUC change in cholesterol from either week 0 or week 4, regardless of whether the analysis included the entire cohort or either of the randomized groups. There were no significant differences in change in BMI, trunk fat, VAT or percentage intra-abdominal fat between randomized groups.
Dietary intake
Both groups were well matched at baseline for dietary intake (Table 3). Although carbohydrate intake decreased in the pravastatin group between weeks 4 and 16, there was no significant between-group difference (P = 0.07). There was also no between-group difference in change in other recorded dietary measure at any measured time points (Table 3) and no correlation between AUC change in serum cholesterol and change in any dietary measure (data not shown).
Other markers of cardiovascular risk
Cardiovascular parameters were measured at week 4 and 16. Apart from %FMD and PAI-1, which were lower and higher respectively in the pravastatin group than in the placebo group, the groups were well matched for other measures of cardiovascular risk (Table 2). Week 4 endothelial function, assessed using %FMD, was generally impaired [median %FMD for the cohort, 4.3% (3.6), reference range, 8-11% [32]]. Apart from homocysteine, which decreased more in the pravastatin group compared to placebo, there were no between-group differences in change in %FMD, blood pressure or other serum markers of cardiovascular risk (Table 2). There were no significant correlations between week 4% FMD or PAI-1 and AUC change from week 4 total cholesterol.
Subjects, materials and methods
Population and study design

This was a single-centre, randomized, double-blind, placebo-controlled pilot study. HIV-infected men (age > 18 years) were recruited through the HIV, Immunology and Infectious Diseases Clinical Services Unit, St. Vincent's Hospital, Sydney, which provides ambulatory care to more than 3000 HIV-infected patients. Inclusion criteria included a documented positive HIV antibody test, stable current PI therapy (beginning not less than 12 weeks prior to screening with little likelihood of change to the ART regimen expected during the study) and fasting serum total cholesterol > 6.5 mmol/l (based on the mean of two consecutive samples taken not less than 3 days apart). Exclusion criteria included hypertension, congestive cardiac failure, malabsorption or other serious illness, active AIDS illness, serum lactate > 2.2 mmol/l or concurrent therapy with other lipid-lowering agents, oral hypoglycaemics, anabolic steroids or insulin.
Prior to baseline (week 0), subjects were randomized (1: 1) to 12 weeks of pravastatin 40 mg each night or placebo to begin at week 4. The 12-week duration of treatment was based on previous studies in primary hypercholesterolaemia that showed a cholesterol-lowering effect of pravastatin within 1 week of therapy with maximal responses seen by week 4 [26,27]. The randomization schedule was prepared by the study statistician and randomization was performed by a hospital pharmacist not directly involved with patient care. The study was approved by St Vincent's Hospital Human Research Ethics Committee and subjects provided written, informed consent prior to enrolment.
Clinical assessment
After screening for eligibility, subjects attended at weeks 0, 4, 8, 12, 15 and 16. At each visit, clinical assessment and safety bloods (full blood count, electrolytes, liver enzymes, urea and creatinine) were performed. Fasting lipids (total cholesterol, triglycerides and HDL cholesterol) were performed at each study visit as previously described [28]. Non-HDL cholesterol was derived by subtracting the HDL cholesterol value from the total cholesterol value. Due to difficulties in estimating LDL-cholesterol levels in the setting of elevated triglycerides this parameter was not measured. Subcutaneous and visceral abdominal fat areas were estimated from average values obtained from three single-slice cross-sectional abdominal computed tomography (CT) scans through the bodies of lumbar vertebrae L2-L4 as previously described [28]. Total and regional body fat and lean mass was calculated from whole-body dual-energy x-ray absorptiometry (DEXA) scans as previously described [28]. Both CT and DEXA were performed at weeks 4 and 16. CD4 T-lymphocyte count and HIV RNA were measured at weeks 0 and 16. Insulin resistance was estimated from fasting glucose and insulin values measured at weeks 0 and 16 using the homeostasis model assessment (HOMA; fasting insulin X fasting glucose/22.5).
Dietary assessment
At baseline participants received individualized dietary counselling by an accredited practising dietician and were instructed on how to complete a 3-day dietary record. This method has been shown to be a valid method for assessing dietary intake for up to 5 days [29,30] and is representative of usual diet if repeated [31]. Three-day diet was recorded at week 0, 4 and 16. Participants were reviewed by the dietician at weeks 4 and 16 to review dietary knowledge and clarify information from diet records. The records were analysed using Foodworks Professional Edition software (version 3.00.338, Xyris software; Highgate Hill, Brisbane, Australia).
Other markers of cardiovascular risk
Endothelial function was assessed by measurement of flow-mediated vasodilatation (% FMD) of the brachial artery as previously described [32,33] at weeks 4 and 16. The following peripheral blood markers of cardiovascular risk were measured at weeks 4 and 16; high sensitivity C-reactive protein, PAI-1 (Trinity Biotec, Ireland), homocysetine (Abbott Laboratories, Illinois), and fibrinogen (Dade Behring Diagnostics, Australia). All assays were performed by an accredited laboratory service according to manufacturers' instructions.
Statistical analysis
Due to the potential for changes in diet and other variables to affect plasma lipids during the course of the study, lipids were measured at several time points. To assess changes in cholesterol with time, the primary endpoint was time-weighted area under the curve (TWAUC) change from baseline in total cholesterol from week 0. Secondary endpoints included TWAUC total cholesterol from week 4 (start of pravastatin), individual changes in total cholesterol at each measured time point, TWAUC from week 0 in HDL-cholesterol and triglycerides, change from baseline glucose, insulin, total and regional body fat, endothelial function and peripheral blood markers of cardiovascular risk.
Based on baseline cholesterol values, using a P value of 0.05, 17 subjects in each arm gave 80% power to detect a change in serum cholesterol between groups of 1.45 mmol/l, which represented approximately a 20% reduction in cholesterol, similar to that seen in other studies [20,23,24]. Analysis was based on intention to treat and non-parametric analyses were used to compare groups of variables. We used Wilcoxon Signed Rank test to compare differences between groups of variables, Mann-Whitney U test for between-group differences and Spearman Rank Correlation (ρ) for correlations between variables, with P values < 0.05 considered significant. To determine possible confounding effects of variables on the primary and principal secondary endpoints, regression analysis was used. We did not adjust P values for multiple comparisons. Unless stated, results are presented as median and interquartile ranges.
  icon paper stack View Older Articles   Back to Top