Effect of Rimonabant on the High-Triglyceride/ Low–HDL-Cholesterol Dyslipidemia, Intraabdominal Adiposity, and Liver Fat
The ADAGIO-Lipids Trial
Background— Rimonabant, the first selective cannabinoid type 1 (CB1) receptor antagonist, improves cardiometabolic risk factors in overweight/obese patients. ADAGIO-Lipids assessed the effect of rimonabant on cardiometabolic risk factors and intraabdominal and liver fat.
Methods and Results— 803 abdominally obese patients with atherogenic dyslipidemia (increased triglycerides [TG] or reduced high-density lipoprotein–cholesterol [HDL-C]) were randomized to placebo or rimonabant 20 mg/d for 1 year. HDL-C and TG were coprimary end points. Intraabdominal (visceral) and liver fat were measured by computed tomography in a subgroup of 231 patients. In total, 73% of rimonabant- and 70% of placebo-treated patients completed the study treatment. Rimonabant 20 mg produced significantly greater changes from baseline versus placebo in HDL-C (+7.4%) and TG levels (−18%; P<0.0001), as well as low-density lipoprotein (LDL) and HDL particle sizes, apolipoprotein A1 and B, HDL2, HDL3, C-reactive protein, and adiponectin levels (all P<0.05). Rimonabant decreased abdominal subcutaneous adipose tissue (AT) cross-sectional area by 5.1% compared to placebo (P<0.005), with a greater reduction in visceral AT (−10.1% compared to placebo; P<0.0005), thereby reducing the ratio of visceral/subcutaneous AT (P<0.05). Rimonabant significantly reduced liver fat content (liver/spleen attenuation ratio; P<0.005). Systolic (−3.3 mm Hg) and diastolic (−2.4 mm Hg) blood pressure were significantly reduced with rimonabant versus placebo (P<0.0001). The safety profile of rimonabant was consistent with previous studies; gastrointestinal, nervous system, psychiatric, and general adverse events were more common with rimonabant 20 mg.
Conclusions— In abdominally obese patients with atherogenic dyslipidemia, rimonabant 20 mg significantly improved multiple cardiometabolic risk markers and induced significant reductions in both intraabdominal and liver fat.
Abdominally obese patients with an excess accumulation of intraabdominal or visceral fat are characterized by a constellation of atherogenic and diabetogenic abnormalities1–4 often referred to as the metabolic syndrome.5 Recent evidence has also established that liver fat is a strong correlate of metabolic risk factors independent of visceral fat, suggesting that both visceral and liver fat affect cardiometabolic risk through partly independent mechanisms.6–8
See accompanying article on page 339
It has been proposed that the mobilization of visceral adipose tissue and liver fat are more clinically relevant therapeutic targets than weight loss.9 In this regard, physical activity/exercise has been shown to selectively reduce visceral adiposity.10 However, for patients who do not respond adequately to lifestyle modification, pharmacotherapy targeting reduction of both visceral and liver fat may be relevant considering the risk associated with this phenotype.
The recent discovery of the endocannabinoid system (ECS) and its impact on visceral fat, liver fat, and associated cardiometabolic risk profile11 has led to the development of a novel approach for targeting visceral obesity and related metabolic complications.12–15 Evidence in humans shows that abdominal obesity associated with excess visceral fat is characterized by an overactivated ECS.16,17 Furthermore, phase III studies conducted with rimonabant, the first cannabinoid type 1 (CB1) receptor antagonist, have shown that antagonism of the ECS reduces body weight, and more importantly, decreases waist circumference while improving several features of the metabolic profile that predict risk of type 2 diabetes (T2DM) and cardiovascular disease.12–15
The ADAGIO-Lipids trial was specifically designed to further study the effects of rimonabant on a comprehensive set of cardiometabolic risk factors in abdominally obese dyslipidemic patients, with plasma high-density lipoprotein cholesterol (HDL-C) and triglycerides (TG) being coprimary end points. Further, for the first time, a computed tomography (CT) substudy was conducted to quantify the effects of rimonabant on visceral adiposity and liver fat.
This randomized, double-blind, placebo-controlled, parallel-group study was conducted in 53 centers in 14 different countries. Male and female patients aged ≥18 years with a waist circumference >102 cm (men)/>88 cm (women) were recruited. Eligible patients were required to have atherogenic dyslipidemia (triglyceridemia ≥1.69 to ≤7.90 mmol/L [≥1.5 to ≤7.0 g/L] or HDL-C <1.04 mmol/L [<40 mg/dL] in men and <1.29 mmol/L [<50 mg/dL] in women). Patients with T2DM could be included if they were on a stable-dose oral antidiabetic for at least 3 months and not treated with glitazones or insulin.
Exclusion criteria included a >5 kg change in body weight within 3 months before the screening visit; presence of clinically significant endocrine or metabolic disorder other than T2DM; any severe medical or psychological illness; or the presence or history (within the previous 5 years) of cancer. Patients were also excluded if they had a history of severe depression (depression which necessitated hospitalization; 2 or more recurrent episodes of depression; or a history of suicide attempt). Pregnant lactating women or women planning on becoming pregnant were excluded, while women of childbearing potential were required to use medically approved contraception. History or current substance abuse, or use of medication altering body weight or appetite, antidepressants, neuroleptics, or thyroid preparations (unless substitution therapy at a stable dose) within 3 months before the screening visit led to exclusion, as did a change in lipid-lowering treatments.
The protocol was approved by the Institutional Review Boards and Independent Ethics Committee at each center, and the study was conducted in accordance with the Declaration of Helsinki and all applicable guidelines of the countries where the study was conducted. All patients provided written informed consent.
The study comprised a 1-week screening period, after which eligible patients were randomized 1:1 to placebo or rimonabant 20 mg/d for 1 year. Tablets were administered with or without food in the morning. Randomization was stratified on the use/no use of concomitant lipid-modifying agents to ensure equal numbers of patients with/without such agents in both treatment groups. During the 12-month treatment period, patients were asked to follow a mildly hypocaloric diet aiming at a deficit of 600 kcal/d in relation to their estimated daily energy needs. A 2-month safety follow-up was performed after the end of treatment or treatment discontinuation.
Patient visits occurred at 2 weeks and 2, 4, 6, 9, and 12 months after randomization; patients were also contacted by phone at 3, 5, 7, 8, 10, and 11 months to maintain compliance and check safety and tolerability.
Assessment of the Metabolic Profile
Measurements of the lipoprotein-lipid profile, glucose, insulin, adiponectin, and high-sensitivity C-reactive protein (hs-CRP) concentrations were performed in central laboratories. Methods are described in the supplemental materials (available online at http://atvb.ahajournals.org).
Imaging Assessments of Visceral and Liver Fat
Abdominal visceral and subcutaneous fat areas and liver fat were assessed by CT using standardized procedures and analyzed centrally using specialized image analysis software (sliceOmatic, Tomovision, Montréal, Québec, Canada). Further details on the methodology substudy are provided in the supplemental materials.
Study End Points
The coprimary efficacy end points were changes from baseline to 12 months in HDL-C and TG. Secondary efficacy end points included changes in visceral and liver fat, fasting lipid parameters (low-density lipoprotein cholesterol [LDL-C], total cholesterol, HDL-C and its subfractions, LDL and HDL particle sizes, apolipoprotein (apo) A1 and B, apo B/apo A1 ratio), adiponectin, glucose, insulin, hs-CRP, body weight, and waist circumference.
Efficacy analyses were performed on data from patients in the intent-to-treat (ITT) population, which consisted of all randomized patients who received at least one dose of study drug and had at least one baseline and one postbaseline assessment. Efficacy end points were analyzed using an analysis of covariance (ANCOVA), with treatment and randomization stratum as fixed effects, and using baseline as covariate. Efficacy analyses on the ITT population excluded assessments obtained after the patients discontinued treatment (up to 1 day after treatment discontinuation), as well as assessments obtained after any lipid-lowering treatment change for the lipid parameters. If a patient discontinued treatment or discontinued the study prematurely or did not have their measurement at the month-12 visit, the last observation carried forward (LOCF) procedure was applied, analyzing the last postbaseline value before the treatment cessation.
The primary statistical comparison was performed at month 12 using appropriate contrast within the framework of repeated measurements ANCOVA. A global 2-sided type I error of 0.05 was used. Categorical efficacy variables were analyzed using the chi-square test. Differences in adverse event rates between rimonabant and placebo groups were tested for significance using the Fisher test.
The authors had full access to the data and take responsibility for its integrity. The coprimary investigators (J.-P.D. and R.R.) drafted the paper, which was then reviewed and approved by all coauthors.
Patient Demographics and Characteristics
In total, 803 patients were randomized, although 4 patients did not receive double-blind treatment; therefore, 799 patients (≈75% whites) were exposed to study treatment, comprising 395 patients receiving placebo and 404 receiving rimonabant 20 mg. A total of 641 patients (80% of those randomized) completed the study (575 patients [72% of those randomized] completed study treatment). Main reasons for treatment discontinuation were adverse events (10.1% with placebo and 14.1% with rimonabant), patient request (13.4% with placebo and 8.4% with rimonabant), and lost to follow-up (3.8% with placebo and 2.5% with rimonabant). The treatment groups had similar baseline characteristics (Table⇓⇓ and supplemental Table I).
After 12 months of treatment, rimonabant 20 mg resulted in significant increases in HDL-C levels and significant reductions in TG concentrations compared with placebo (Table⇑⇑ and Figure 1A and 1B). Whereas a reduction in TG levels was already evident after 2 weeks of rimonabant treatment, the marked increase in HDL-C began to be observed after 6 months of therapy. The early decrease in TG levels was correlated with the magnitude of weight loss (r=0.18, P<0.005 in placebo; r=0.12, P<0.05 in rimonabant) already observed after 2 weeks of treatment (Figure 1C). The total cholesterol/HDL-C ratio was significantly improved in patients receiving rimonabant 20 mg compared with placebo (Table⇑⇑).
Treatment with rimonabant 20 mg led to significant improvements versus placebo in several cardiometabolic risk markers, including apo A1 and apo B levels, apo B/apo A1 ratio, and hs-CRP (Table⇑⇑). In addition, waist circumference and body weight were significantly reduced after 1 year to a greater extent in patients receiving rimonabant 20 mg versus placebo (Table⇑⇑). No change in cholesterol or LDL-C were observed.
After the 12-month treatment period, rimonabant 20 mg resulted in a significantly greater increase in HDL2-C from baseline versus placebo (P=0.0288; Table⇑⇑). Similarly, greater increases in HDL3-C levels were observed with rimonabant 20 mg (P=0.0103; Table⇑⇑). Rimonabant 20 mg was also associated with a greater change in HDL particle size (P=0.0009; Figure 2A). Change in HDL particle size versus placebo was fairly rapid as a substantial difference was already observed between the two treatment groups after 2 weeks.
The distribution of LDL particles shifted toward larger sizes with rimonabant 20 mg versus placebo (P<0.0001; Figure 2B). Rimonabant 20 mg was associated with a significantly lower proportion of small LDL particles (between-group difference: −6.46±1.11%; P<0.0001), and a significantly higher proportion of large LDL particles versus placebo (between-group difference: +4.84±0.98%; P<0.0001; Table⇑⇑). The change in LDL particle size was already observed after 2 weeks of treatment and was correlated with early changes in TG levels (r=−0.35, P<0.0001 in placebo and r=−0.37 in rimonabant, P<0.0001).
HDL-C levels increased to a greater extent in patients who experienced ≥5% or ≥10% weight loss during the study (Figure 3A). Furthermore, patients in the rimonabant group also had greater increases in HDL-C versus patients with the same level of weight loss in the placebo group.
Increases in adiponectin levels from baseline were greater in patients who had achieved higher weight loss during the study (Figure 3B). For any given weight loss, change in adiponectinemia was more pronounced for patients who received rimonabant compared with placebo (Figure 3B).
Visceral Fat and Liver Fat Measured by CT
Rimonabant resulted in a greater reduction in both visceral (16% versus 5.9%; P=0.0003; Figure 4A) and subcutaneous fat (9.7% versus 4.7%; P=0.0043; Figure 4B) compared with placebo after 12 months. Rimonabant was also associated with a greater loss of visceral than subcutaneous fat (P<0.05).
Twenty-six of the 59 (44%) subjects treated with placebo and 27 (53%) subjects treated with rimonabant had a liver/spleen attenuation ratio (fatty liver index: CTL/CTS) less than 1.0, a value used routinely to diagnose fatty liver.18 After 12 months, the CTL/CTS ratio increased by a significantly greater extent with rimonabant versus placebo, with a change of 0.16 versus 0.05 during the study (P=0.0017; Table⇑⇑ and Figure 4C). Almost 3 times (n=13 [48%] versus n=5 [19%]) as many patients on rimonabant who started with a diagnosis of fatty liver increased their fatty liver index to “normal” (>1) values after 1 year. Levels of ALT were also reduced to a significantly greater extent with rimonabant (from 33 to 25 IU/L; −15.4% change) versus placebo (from 33 to 30 IU/L; −3.5% change; difference −11.72±2.52%; P<0.0001).
Plasma Glucose/Serum Insulin Levels and Blood Pressure
Both fasting glucose (difference −0.24 mmol/L; P=0.0009) and insulin (difference −16.7 pmol/L; P=0.0033) levels were significantly reduced compared to placebo with rimonabant therapy (Table⇑⇑). Systolic and diastolic blood pressure were equally reduced (both by 3 mm Hg) with rimonabant, such changes being significantly greater than in the placebo arm (P<0.0001; Table⇑⇑).
Safety and Tolerability
Rimonabant was generally well tolerated. Slightly more patients receiving rimonabant reported treatment-emergent adverse events (TEAEs) than placebo (88.9% versus 84.3%). The adverse events were usually mild or moderate in intensity.
A total of 59 serious TEAEs were reported in 58 patients during the study, the incidence being identical between the rimonabant 20 mg (n=29; 7.2%) and placebo (n=29; 7.3%) groups. The most common TEAEs occurring in ≥2% of patients receiving rimonabant 20 mg are listed in the supplemental Table II. Study treatment discontinuations attributable to TEAEs were slightly higher among patients receiving rimonabant 20 mg (14.1%) compared with placebo (10.1%). Anxiety and depression were the most frequent adverse events leading to premature treatment discontinuation: 9 patients (2.2%) and 4 patients (1.0%) for anxiety; and 8 patients (2.0%) and 5 patients (1.3%) for depression, in the rimonabant and placebo groups, respectively.
Serious adverse events linked to psychiatric disorders were suicidal ideation (placebo n=1; rimonabant 20 mg n=2) and suicide attempt (n=1 in each group). There were no reported deaths during the study.
Results of ADAGIO-Lipids confirm the significant effect of rimonabant on HDL-C and TG levels in abdominally obese patients with an atherogenic dyslipidemia.12,14 Although the difference in HDL-C increase with rimonabant versus placebo was essentially similar to what had been observed in previous RIO studies, the absolute increase from baseline in HDL-C was less than in previous studies.12–15 We had previously proposed that the apparently exaggerated response of HDL-C in both placebo and rimonabant arms observed in the RIO studies was an artifact caused by the design of these studies. Indeed, baseline measurements performed in the phase III rimonabant studies were obtained after a 1-month run-in period during which patients followed the instructions received from study dieticians to reduce their daily caloric intake by 600 kcal. Thus, patients in the RIO program were actively losing weight (about 2 kg) at the time of their baseline measurements. As a consequence, metabolic parameters such as TG and HDL-C had already decreased from the screening period to the time baseline measurements were performed. In phase III studies, the metabolic effect of the weight loss during the run-in period may have altered the absolute response of metabolic parameters to rimonabant as patients’ metabolic profile had already begun to change before the randomization of patients to either placebo or rimonabant. Therefore, ADAGIO-Lipids was specifically designed to assess the net effects of rimonabant (20 mg) when weight-stable abdominally obese dyslipidemic patients are randomized to either rimonabant or placebo.
As opposed to previous phase III studies with rimonabant,12–15 trivial changes in plasma HDL-C (+1.8%) were observed among patients who received a placebo for 1 year. However, rimonabant 20 mg significantly increased plasma HDL-C by 8.7% compared to baseline, with a mean difference versus placebo of +7.4% (P<0.0001). The other coprimary end point, plasma TG, also showed a very small reduction in the placebo group (−2.7%) compared to a reduction of −19.5% in the rimonabant 20 mg group, leading to a −18% difference in favor of rimonabant (P<0.0001).
These favorable changes in the lipid profile were accompanied by significant changes in a comprehensive set of cardiometabolic risk variables, which were also previously examined in some RIO studies but again in the context of the possible confounding effect of the 1-month hypocaloric run-in period. The ADAGIO-Lipids trial confirms that even in the absence of any change in LDL-C, rimonabant 20 mg increases LDL particle size and the proportion of large LDL particles while decreasing the percentage of small LDL particles. This shift in LDL particle size provides another example of the limitation of LDL-C in estimating the concentration of atherogenic lipoproteins in patients with abdominal obesity.2 To deal with this limitation of LDL-C in patients with abdominal obesity and hypertriglyceridemia, it has been proposed to focus on apo B levels.19,20 In ADAGIO-Lipids, rimonabant 20 mg significantly reduced apo B levels and increased apo A1, leading to a significant reduction in the apo B/apo A1 ratio. This ratio, as well as the cholesterol/HDL-C ratio (also reduced after rimonabant therapy), have been reported to be a useful predictors of coronary heart disease risk.21,22
Additional measures of HDL “quality” were performed in ADAGIO-Lipids. Whereas a small but significant effect of rimonabant on apo A1 level was found, a more robust response of HDL subfractions was observed and both HDL2 -C and HDL3-C levels were significantly increased by rimonabant 20 mg, a finding different from the effects of fibrates, which have been reported to preferentially increase the concentration of the HDL3 subfraction.23 HDL particle size assessed by gradient gel electrophoresis was also increased, suggesting that rimonabant may affect to a greater extent HDL particle size than HDL concentration. Such changes in both HDL quantity and quality have also been reported to occur in response to endurance exercise training.24 Further studies examining HDL kinetics and function should determine the physiological implications of such findings regarding well-known properties of HDL including its antiinflammatory, antithrombotic, and antioxidative potential beyond its role in reverse cholesterol transport.25,26
As the high TG–low HDL-C dyslipidemic state of abdominal obesity is very often accompanied by a state of insulin resistance, fasting plasma glucose and insulin levels were also assessed in ADAGIO-Lipids. Significant reductions in these indices of plasma glucose–insulin homeostasis confirm the effect of rimonabant on insulin resistance, which had been shown to improve glucose tolerance,12,15 reduce insulin levels,12,13,15 and improve glycemic control in T2DM patients.14
The results of the CT substudy provides for the first time evidence that rimonabant induces a significant loss of visceral fat. Furthermore, the reduction in visceral fat was accompanied by a decrease in liver fat, a finding compatible with the decrease in ALT previously reported with rimonabant therapy12 and confirmed by the ALT results of ADAGIO-Lipids. Accordingly, results of ADAGIO-Lipids provide for the first time direct evidence in human subjects that CB1 receptor antagonism may represent a relevant approach to specifically target visceral obesity and excess liver fat, a phenotype predictive of a constellation of metabolic abnormalities which include inflammation.6,9 Indeed, ADAGIO-Lipids also provides evidence that pharmacotherapy aimed at visceral obesity/liver fat induces a substantial reduction in plasma hs-CRP levels, a commonly used marker of inflammation predictive of cardiovascular disease risk.27 Furthermore, our observed reduction in systolic and diastolic blood pressure and hs-CRP levels combined with the robust increase in adiponectin concentration produced by rimonabant are consistent with the well-documented effect of weight loss and loss of abdominal fat on blood pressure and inflammation as well as on adipose tissue biology.28,29
Finally, analysis of changes in HDL-C and adiponectin levels across subgroups classified on the basis of magnitude of weight loss provides further evidence of a cardiometabolic effect of rimonabant which cannot be explained by weight loss. For instance, regression analyses revealed that 72% of the response in HDL-C and 67% of the response in adiponectin to rimonabant could not be explained by weight loss. These results are fully concordant with experimental evidence that rimonabant may affect the metabolism of tissues such as the liver (decreased lipogenesis) and adipose tissue (increased secretion of adiponectin).30,31 It is, however, important to point out that the present study did not have sufficient statistical power to address the issue of whether or not the improvements in the cardiometabolic risk profile reported herein are dependent/independent from changes in visceral/ectopic fat. Further trials specifically designed for that purpose will be necessary to address this important question.
The side effect profile of rimonabant has been extensively studied, and most side effects were mild or moderate in intensity with the exception of a low incidence of depressive disorders. No deaths occurred during the trial. Safety results of ADAGIO-Lipids are fully consistent with previous studies, with the exception of the recently published STRADIVARIUS trial which reported higher rates of depressive symptoms than in ADAGIO-Lipids.32 It is important, however, to keep in mind that patients with a history of severe depression were excluded in ADAGIO-Lipids. In STRADIVARIUS almost 20% of patients were treated with antidepressant medications. In ADAGIO-Lipids, only 9 patients (2.2%) and 13 patients (3.2%) in the placebo and rimonabant groups, respectively, started antidepressant medication during the treatment period. Therefore, results of ADAGIO-Lipids provide further evidence that there is a fairly low incidence of depressive disorders leading to discontinuation of treatment when patients with a history of depression are not considered for rimonabant therapy. Furthermore, cardiovascular safety was also confirmed in ADAGIO-Lipids with a significant blood pressure-lowering effect of rimonabant.
Results of the ADAGIO-Lipids trial confirm the well-documented HDL-C–raising properties of rimonabant in overweight/obese patients with atherogenic dyslipidemia and provide novel insight into the significant effects of rimonabant on visceral adiposity and liver fat content. Given the current debate surrounding the benefit/risk ratio of rimonabant and of CB1 antagonists, our results provide evidence that rimonabant should not be prescribed/perceived as an “antiobesity blockbuster” medication. Rather, it is suggested that targeting the ECS with rimonabant remains a promising approach which must be “targeted” to the right patient subgroup: those with excess visceral/liver fat who are at increased cardiometabolic risk.
The authors thank Annaïg Le Halpère, Tian Z. Luo, Laurent Luttenauer, Bruno Orofiamma, Elisabeth Baudin, Soazig Chevalier, Eric Sorel, the sanofi-aventis Clinical Research Units, and the members of the data monitoring committee (A. Leizorovicz, K. Lee, E. Danforth, M. Weintraub, J.E. Gerich, J.-L. Imbs, S. Laporte). The authors express their gratitude to the Global Publication Group at sanofi-aventis for their technical support in the preparation of this paper for publication.
Sources of Funding
This work was supported by sanofi-aventis.
Dr Després is a speaker for Abbott Laboratories, AstraZeneca, Fournier Pharma Inc/Solvay Pharma, GlaxoSmithKline, and Pfizer Canada Inc; has received research funding from INNODIA, Eli Lilly, GlaxoSmithKline, and sanofi-aventis; is on advisory boards for MSD, Novartis, and sanofi-aventis; and has received consulting fees from INNODIA and sanofi-aventis. Dr Ross is a speaker for GlaxoSmithKline and sanofi-aventis; has received research funding from MARS Inc, and sanofi-aventis; is on the advisory board for sanofi-aventis; and has received consulting fees from GlaxoSmithKline, and sanofi-aventis; and receives royalties from Human Kinetics. Dr Boka is an employee of sanofi-aventis.
↵*The investigators and coinvestigators of the ADAGIO-Lipids study are listed in the supplemental materials.
Received September 26, 2008; revision accepted December 15, 2008.
Björntorp P. Metabolic implications of body fat distribution. Diabetes Care. 1991; 14: 1132–1143.
Després JP, Moorjani S, Lupien PJ, Tremblay A, Nadeau A, Bouchard C. Regional distribution of body fat, plasma lipoproteins, and cardiovascular disease. Arteriosclerosis. 1990; 10: 497–511.
Kissebah AH, Krakower GR. Regional adiposity and morbidity. Physiol Rev. 1994; 74: 761–811.
Grundy SM, Brewer HB Jr, Cleeman JI, Smith SC Jr, Lenfant C. Definition of metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation. 2004; 109: 433–438.
Nguyen-Duy TB, Nichaman MZ, Church TS, Blair SN, Ross R. Visceral fat and liver fat are independent predictors of metabolic risk factors in men. Am J Physiol Endocrinol Metab. 2003; 284: E1065–E1071.
Després JP, Lemieux I, Bergeron J, Pibarot P, Mathieu P, Larose E, Rodés-Cabau J, Bertrand OF, Poirier P. Abdominal obesity and the metabolic syndrome: contribution to global cardiometabolic risk. Arterioscler Thromb Vasc Biol. 2008; 28: 1039–1049.
Bluher M, Engeli S, Kloting N, Berndt J, Fasshauer M, Batkai S, Pacher P, Schon MR, Jordan J, Stumvoll M. Dysregulation of the peripheral and adipose tissue endocannabinoid system in human abdominal obesity. Diabetes. 2006; 55: 3053–3060.
Davidson LE, Kuk JL, Church TS, Ross R. Protocol for measurement of liver fat by computed tomography. J Appl Physiol. 2006; 100: 864–868.
Brunzell JD, Davidson M, Furberg CD, Goldberg RB, Howard BV, Stein JH, Witztum JL. Lipoprotein Management in Patients With Cardiometabolic Risk: Consensus statement from the American Diabetes Association and the American College of Cardiology Foundation. Diabetes Care. 2008; 31: 811–822.
Robins SJ, Collins D, Wittes JT, Papademetriou V, Deedwania PC, Schaefer EJ, McNamara JR, Kashyap ML, Hershman JM, Wexler LF, Rubins HB. Relation of gemfibrozil treatment and lipid levels with major coronary events: VA-HIT: a randomized controlled trial. JAMA. 2001; 285: 1585–1591.
Barter PJ, Nicholls S, Rye KA, Anantharamaiah GM, Navab M, Fogelman AM. Antiinflammatory properties of HDL. Circ Res. 2004; 95: 764–772.
Brewer HB Jr. High-density lipoproteins: a new potential therapeutic target for the prevention of cardiovascular disease. Arterioscler Thromb Vasc Biol. 2004; 24: 387–391.
Tchernof A, Nolan A, Sites CK, Ades PA, Poehlman ET. Weight loss reduces C-reactive protein levels in obese postmenopausal women. Circulation. 2002; 105: 564–569.
Bensaid M, Gary-Bobo M, Esclangon A, Maffrand JP, Le Fur G, Oury-Donat F, Soubrie P. The cannabinoid CB1 receptor antagonist SR141716 increases Acrp30 mRNA expression in adipose tissue of obese fa/fa rats and in cultured adipocyte cells. Mol Pharmacol. 2003; 63: 908–914.
Osei-Hyiaman D, DePetrillo M, Pacher P, Liu J, Radaeva S, Batkai S, Harvey-White J, Mackie K, Offertaler L, Wang L, Kunos G. Endocannabinoid activation at hepatic CB1 receptors stimulates fatty acid synthesis and contributes to diet-induced obesity. J Clin Invest. 2005; 115: 1298–1305.
Nissen SE, Nicholls SJ, Wolski K, Rodes-Cabau J, Cannon CP, Deanfield JE, Després J-P, Kastelein JJP, Steinhubl SR, Kapadia S, Yasin M, Ruzyllo W, Gaudin C, Job B, Hu B, Bhatt DL, Lincoff AM, Tuzcu EM, for the SI. Effect of rimonabant on progression of atherosclerosis in patients with abdominal obesity and coronary artery disease: The STRADIVARIUS randomized controlled trial. JAMA. 2008; 299: 1547–1560.